--- Log opened Wed Jun 06 00:00:46 2018 00:03 < BlueProtoman> lnnb: nvidia is useless, they forwarded me to my manufacturer 00:03 < BlueProtoman> leopard: No one with the requisite knowledge is in #ubuntu right now 00:09 < lnnb> BlueProtoman: if you don't call it's probably some guy or AI handling multiple convos 00:10 < lnnb> maybe try calling intel tech support after nvidia real tech support shuts you down 00:14 < DLange> BlueProtoman: https://devtalk.nvidia.com/default/topic/989530/cuda-setup-and-installation/cuda-installation-issues-ubuntu-16-04-geforce-gtx-860m/ looks useful. You can also ask in that forum if you look to use cuda. 00:14 < mawk> gammu is nice bls indeed 00:14 < mawk> now I can turn my raspberry pi into a SMS center 00:21 < pankaj_> Even wikipedia cannot define what does service management means in linux. 00:27 < djph> see init? 00:39 < uplime> pankaj_: why would it mean anything more than "managing services"? 00:40 < Dagmar> because consultants have to sell training hours 00:40 < uplime> oh silly me 00:40 < uplime> serverless cloud based service management 00:40 < uplime> in docker 00:40 < meyou_> ...as a service 00:41 < meyou_> scbsmidaas 00:41 < uplime> perfect 00:41 < Dagmar> I see VC dollars in your future 00:42 < uplime> tls5.0 encryption 00:42 < lnnb> not enough blockchain 00:43 < meyou_> still waiting for a blockchain social network 00:43 < uplime> it also needs to support anonynimity 00:43 < meyou_> i feel like that could get around the whole data privacy thing, everybody hosts it themselves as nodes and nobody gets paid except the users' isps and power co's 00:44 < uplime> everybody is everybody else's ISP 00:44 < Dagmar> Don't forget the MPAA 00:44 < Dagmar> They'll want their cut as well 00:45 < meyou_> i guess you'd have the issue of freeloaders 00:45 < meyou_> popular people making the other nodes do work for their social media influence 00:45 < Dagmar> Oh I thought you were talking about congress 00:47 < meyou_> i still think we can replace the federal government with a smartphone app but that plan is a little more complex 00:49 < lnnb> hmm maybe, lets say it's torrent-like. your friends could all host eachother's content, only problem is what happens when no one is online, you have to fallback to some probably swamped super-node that hosts everyone. 00:51 < meyou_> lnnb, yeah i think you'd need some compensation for nodes 00:51 < meyou_> so if you host a node but you're not very active on social media, you might get a few bucks compensation for your node 00:51 < meyou_> and if you're super active but your node sucks, you might get a bill 00:52 < meyou_> but i'm afraid this kills the social network, nobody wants to pay when they can let facebook use their data and get likes for free 00:52 < djph> meyou_: I think they did that on an episode of "The Orville" 00:55 < lnnb> maybe you could leverage email services to store "offline" updates 00:55 < djph> lnnb: what're you making today 00:55 < djph> ? 00:55 < lnnb> facebook killah 00:56 < djph> replace github :) 00:59 < lnnb> if anyone wants to pay me to make it happen PM away :P 01:00 < Psi-Jack> djph: GitLab, BitWeaver, Gogs and Gitea already have. :p 01:00 < Psi-Jack> BitBucket I mean. 01:01 < meyou_> i wonder how long before the option to log into github with your MS account will pop up 01:01 < meyou_> and then some time long after that it will become mandatory 01:01 < meyou_> likely with at least one deadline postponement 01:01 < djph> Psi-Jack: obviously we need a fifth option 01:03 < Psi-Jack> http://www.reactiongifs.com/r/but-why.gif 01:06 < djph> https://xkcd.com/927/ 01:07 < meyou_> there's always a relevant xkcd 01:08 < Psi-Jack> heh 01:09 < mattfly> is it possible to use xrdp like a vnc server 01:09 < mattfly> connecting to a x sessiosn im not using 01:09 < lnnb> the only real question is how do we make money, and i ask you how much do you sell the IoT appliances for? 01:14 < djph> 9.99 01:16 < Psi-Jack> And yet, it only cost you $6.57 to make. :) 01:16 < Psi-Jack> Good profit! :D 01:16 < lnnb> maybe in some decades then, it's going to have to be able to transcode videos without igniting 01:16 < meyou_> we gotta get those margins up 01:18 < superkuh> Get in on the ground floor of the ICO for my bespoke artisanal IoT whitespace RF blockchain written in rust compiled to ECMAScript running on node on an rpi, http://thingshit.io/ 01:24 < lnnb> for advertising we will tap old war memes, to remind every citizen they must "do their part" in the fight against the great menace. 01:26 < Guest80> hi there 01:26 < GunqqerFriithian> Hello 01:27 < Guest80> need some help 01:27 < GunqqerFriithian> Details? 01:27 < Guest80> i try to install csh package in redhat 6.5 and i get this error during decompress 01:27 < Guest80> unpacking of archive failed on file /bin/csh: cpio: rename 01:28 < djph> superkuh: URL is wrong. should be shitthings.io 01:29 < superkuh> Yeah, but sharing the s of two words makes my url more hip. 01:29 < lnnb> very minimal of you 01:36 < djph> superkuh: like expertsexchange? 01:37 < meyou_> i prefer amateursexchange 01:38 < Iven> installed firefox long term support, changed which addon was enabled, one of the switches was from session manager that worked with new to the one that works with old. My firefox browser history and the tabs that are suppose to autoload when I start, are gone. What might have caused this and is there a way to recover them? 01:38 < bls> Iven: might want to try ask the firefox channel on irc.mozilla.org 01:40 < Iven> but this is also a linux problem 01:40 < Iven> I doubt they can tell me much about how to recover files on linux 01:41 < djph> meyou_: hahahaha 01:41 < meyou_> djph, i was scared to tack .com onto it 01:41 < meyou_> lest the site exists 01:43 < mattfly> im using a xrdp server and trying to open a opengl program doesnt work 01:43 < mattfly> yes i have a nvidia driver and im on ubuntu bionic 01:44 < mattfly> no chance this would work w 01:44 < Dagmar> You're using XRDP. What you get is what you get 01:44 < Dagmar> Typically these things do not play well with GL applications. 01:49 < mattfly> damn... 01:54 < tona> hello everyone 01:54 < tona> if [ "${development}" == "details" ]; then development = "The web service in Development Environment is working"; else development = "The web service in Development Environment not working, Please attend this issue inmediately"; fi i am geting one error on else , could someone advice me 02:04 < uplime> ftr, assignments can't have spaces around them, which was tona's issue 02:05 < uplime> also, == doesn't work in [ 02:05 < uplime> (or at least, don't expect it to) 02:44 < sinatrablue> what distro is the most comfy 02:44 < uplime> armchair linux 02:45 < sinatrablue> didnt even know 02:46 < GunqqerFriithian> hannah montana linux best linux 02:48 < jim> debian is definitely one of the most stable... and it's more toward setting stuff up for you (as opposed to arch and lfs, where you have to largely set stuff up yourself, and have to learn how at the same time (if you don't already know) 02:50 < GunqqerFriithian> Ubuntu is pretty good also, but I suggest to use kubuntu (ubuntu with KDE) as KDE is better than gnome IMO 02:53 < trhr> I asked one of our IT guys today if he preferred debian or centos for a server environment, and he said "CentOS, but I'm really more of a Windows guy." Should I fire him? 02:54 < djph> 50kv straight to the nipples 02:56 < GunqqerFriithian> what we playing CaH now? 03:01 < TinyTimmyTokyo> i'm wondering which option would be best for a laptop, if I want to try something other than ext4: ZFS, XFS, or JFS? 03:01 < TinyTimmyTokyo> the laptop has 4GB of RAM 03:02 < Elladan> I'd suggest XFS 03:02 < kerframil> timeoperator[m]: ext4 and XFS are the best mainstream filesystems 03:02 < kerframil> oops 03:02 < devxxx> ntfs 03:02 < Elladan> ZFS is more for file servers and isn't integrated well due to licensing issues. 03:02 < Elladan> JFS is kind of old and less supported. 03:02 < kerframil> TinyTimmyTokyo: I wouldn't bother using any others unless you have a very clear need for the features that they provide (COW, data integrity/scrubbing, integrated snapshots etc). 03:03 < epicmetal> scrubbing and dedup would be nice 03:03 < Elladan> Yeah, XFS isn't really going to give you extra features. 03:03 < Elladan> Your best bet right now for fancy-features is btrfs. However, it has a large number of caveats. 03:03 < kerframil> right. XFS is about scaleability and reliability. bells and whistles, not so much. 03:07 < jim> 4gb ram is pretty good, what kind of net card does it have? 03:08 < TinyTimmyTokyo> intel wireless 03:08 < jim> devxxx, nope, ntfs won't work for a linux system... best you can do with that is mount it 03:09 < jim> TinyTimmyTokyo, then you might need firmware for it 03:09 < jim> there's a version of the debian netinstall image that comes with the firmware you might need 03:10 < TinyTimmyTokyo> i'm a gentoo user 03:10 < jim> ok, do you want to put gentoo on the laptop? 03:11 < TinyTimmyTokyo> I already have it on 03:11 < TinyTimmyTokyo> but I want to start over with a new partition set 03:11 < TinyTimmyTokyo> but I need to pick a filesystem type, though 03:11 < jim> if you do that, you may or may not want to back up your /home first 03:12 < GunqqerFriithian> nah `rm -rdf ~/` 03:12 < GunqqerFriithian> :P 03:12 < jim> GunqqerFriithian, careful, destructive code posting isn't permitted here 03:14 < jim> besides if he doesn't care what's in his laptop home dir, he can just delete the partitions in the installer 03:25 < nekoseam> t1\ 03:29 < gbellinoz> I'm getting "kernel: [643158.352717] Corrupted low memory at ffff880000009000..." in my logs. Googling tells me it's no big deal - just BIOS doing something it shouldn't. Problem is, these messages are new. No chance of that being malware? 03:36 < TinyTimmyTokyo> I already have what's in /home backed up to a USB flash storage device 03:36 < TinyTimmyTokyo> so I can safely nuke the partitions (all of them) and start afresh 03:37 < matkatmusic> anyone know where the root of http://localhost would be on my HD? 03:37 < matkatmusic> like /var/www/ or something? 03:48 < phogg> matkatmusic: that's one option; some people choose it 03:48 < phogg> matkatmusic: I also see the use if a directory in /home. Some even use a directory in /srv. 03:49 < phogg> s/if/of/ 03:49 < oerheks_> it would be the place www-data has access to 03:50 < phogg> Not necessarily. For many sites the webserver only needs read access and it's possible that any user on the system has that. 03:50 < matkatmusic> i'm trying to figure out why chrome won't connect to a url i added to the hosts file that points to a local ip, yet visiting http://localhost works and now I'm curious where the file that localhost points to is, because on my screen it says "It works!" when i visit http://localhost 03:51 < phogg> matkatmusic: It might not even be a file (though that one probably is). You can work out where it is by looking at your webserver config. Are you using apache? 03:51 < matkatmusic> I'm on OS X, and the ##php people said stop asking those types of questions there and ask them here so.. :-/ 03:52 < phogg> matkatmusic: then why are you asking in ##linux? 03:52 < matkatmusic> I'm using a VirtualBox running linux 03:52 < matkatmusic> the box is registered at 192.168.56.101 03:52 < phogg> matkatmusic: Is the web server running on Linux? 03:52 < matkatmusic> the web server is running on the virtual box 03:53 < matkatmusic> so, yes 03:53 < phogg> matkatmusic: Okay. 03:53 < phogg> matkatmusic: Do you know what webserver software you are using? Is it apache? 03:56 < phogg> matkatmusic: it's possible your web server will tell you what it is if you ask: curl -sI localhost 03:56 < phogg> Look for the Server: line 03:57 < matkatmusic> I can connect to that IP directly in chrome and it displays what I expect it to display 03:58 < matkatmusic> so I think the problem is with chrome 03:59 < kerframil> visit from where? a user-agent running in macOS? 04:00 < kerframil> if so, and unless you've deliberately hijacked the resolution of "localhost", that won't be content served by your VM 04:00 < kerframil> and, should all of the above be so, the answer is probably /Library/WebServer/Documents 04:01 < matkatmusic> i have this in my /etc/hosts file: 192.168.56.101 matkatmusic.com 04:01 < matkatmusic> on OS X 04:01 < kerframil> that has nothing to do with localhost, after which you were enquiring 04:02 < matkatmusic> I was of the understanding that etc/hosts affects the resolution of addresses I type in to the address bar in the browser 04:02 < kerframil> localhost resolves to an address of the loopback interface (127.0.0.1 or ::1). if your user-agent is running under macOS, that's the loopback interface owned by macOS. 04:03 < matkatmusic> I'm not connecting to http://localhost 04:06 < matkatmusic> curl -v http://mytest.site works as expected 04:10 < kerframil> as for chrome, it has its own DNS resolver and cache. clear its internal cache, at least (chrome://net-internals/#dns). 04:14 < k_sze[work]> Is there a way to check whether a certain binary (executable or shared library) is running? 04:14 < k_sze[work]> or being used? 04:14 < kerframil> matkatmusic: also, disable and options that pertain to dns "pre-fetching", "predictive" network loading or any other such nonsense. those are somewhere in the conventional settings page, iirc. 04:14 < kerframil> any* options 04:15 < k_sze[work]> iirc, Linux processes don't necessarily keep an open file handle to executables and shared libraries? 04:19 < k_sze[work]> And I need a way that will follow symlinks as well. e.g. I want check if /usr/bin/python2.7 is being run, but it can be symlinked as /usr/bin/python -> /usr/bin/python2 -> /usr/bin/python27. So if it's being run as /usr/bin/python, I want to know as well. 04:21 < matkatmusic> thanks kerframil 04:25 < lnnb> k_sze[work]: /proc/self/exe /proc/self/maps 04:25 < k_sze[work]> lnnb: is `lsof` good enough? 04:26 < lnnb> for maps, the path doesn't mean file hasn't been overwritten you maybe have to map memory and compare 04:26 < lnnb> path matching 04:27 < k_sze[work]> And it makes me wonder: how do Linux distros manage to update something like bash or opensshd when they are still running? 04:27 < lnnb> hopes and prayers until reboot 04:27 < k_sze[work]> lnnb: but no reboot is required for updating bash! 04:28 < k_sze[work]> Somehow you just don't get an updated bash until you start a new bash instance. 04:28 < pnbeast> k_sze[work], I'm just into the channel and don't know what's passed, already, but writing a new file is not prevented by having an old one in use (or memory, at least). So, you can update sshd a bazillion times. 04:29 < pnbeast> The sshd (or bash, or whatever) *in* memory keeps running, of course, until it's closed or restarted. 04:30 < k_sze[work]> ok 04:32 < kerframil> k_sze[work]: try replacing an ELF binary that is still open and undergoing execution, then running lsof +L1. 04:32 < pnbeast> I.e., k_sze starts bash ("old one"), then updates bash. The inode that points to old bash is gone, although the data probably does remain on disk and definitely in RAM. (this is why deleting giant files that are still in use doesn't free up disk space) 04:32 < kerframil> k_sze[work]: the prior instance is not purged until the link count drops to zero 04:32 < kerframil> k_sze[work]: not even from the filesystem 04:32 < pnbeast> So, if k_sze starts bash again, he gets the newly installed one. 04:33 < pnbeast> Yeah, what kerframil says! 04:33 < k_sze[work]> kerframil: I see. 04:35 < k_sze[work]> So if I have updated bash and still have an old bash instance running, can `lsof` tell the difference between the two? 04:35 < k_sze[work]> Or, if it does, how do *I* tell which is which? 04:35 < lnnb> does it tell you inode # 04:36 < lnnb> if it's not also bind mounted from another fs inode # might be good enough 04:37 < k_sze[work]> Yeah, I was going to ask about bind mounts, too. 04:42 < kerframil> k_sze[work]: hmm, actually, I think that incantation of lsof only works for files that are explicitly unlinked. this is something that I use quite often: http://www.schwarzvogel.de/software/lib_users.html 04:44 < kerframil> the key thing is that something can be marked for deletion but still mapped. the stale data can be deleted once it is no longer referenced. this, along with mandatory file locking not being a thing in Linux, is what facilitates relatively smooth in-situ upgrades - at least, compared to some other platforms. 04:44 < aclaivi> Hello all. 05:02 < swift110> hey aclaivi 05:04 < k_sze[work]> kerframil: thanks for the explanation! 05:19 < ayecee> also bonus points for working in-situ in there. 05:19 < brun1n> How do I change a USB COM port for a USB device in ubuntu? 05:19 < brun1n> I need to change it to COM port 1 05:20 < brun1n> lsusb is 'Bus 001 Device 010: ID 2341:0043 Arduino SA Uno R3 (CDC ACM)' 05:20 < ayecee> why 05:20 < brun1n> For arduino project 05:20 < ayecee> ok, but why 05:20 < brun1n> To communicate with my arduino and load sketches ayecee 05:20 < ayecee> why must it be com1 05:20 < brun1n> Because that was the IDE is mapped to 05:21 < ayecee> can the IDE config be changed 05:21 < brun1n> Not sure, I cant find a way to change it 05:22 < ayecee> i'm sure you can change the IDE config. however, the lazy way to work around this would be to manually create a symlink 05:22 < ayecee> pointing from the device you need to the device you have 05:23 < brun1n> Ah found it lol, yeah can change it in interface cheers anyways 05:25 < ayecee> :) 05:31 < annihilator> Is there a reason why linux only detects 1.8 tb of a 4tb raid? 05:31 < ayecee> yes 05:32 < annihilator> Is the a way around that? 05:32 < ayecee> probably yes 05:32 < annihilator> Without removing the raid 05:33 < ayecee> most common cause would be using mbr partition scheme, with its 2tb limit. the way around that is to reformat and use gpt instead. 05:33 < annihilator> But it was gpt it's set when I installed windows 05:33 < kerframil> the other cause would be creating an md array with the deprecated 0.90 format superblock 05:34 < kerframil> obviously, if you're using something other than md, that doesn't apply 05:34 < annihilator> I'll have to look 05:36 < Siva_Machina> I assume tthere may be a chance it got set back to mbr when you switched to Linux? 05:37 < annihilator> I checked dskprt and it showed gpt still 05:38 < jim> are both oses booting the same mode (legacy or uefi)? 05:38 < annihilator> Uefi 05:39 < jim> ok, which are you running at the moment? 05:39 < annihilator> Neither 05:39 < annihilator> I was running windows 05:39 < jim> oh? is there a third os? or the computer is shut down? 05:39 < annihilator> But just wipped all petitions thinking it was that 05:39 < annihilator> Wiped* 05:40 < ayecee> partitions* 05:40 < jim> oh, so nothing is installed at the moment/ 05:40 < jim> ? 05:40 < annihilator> Yea spell check hates me 05:40 < annihilator> Correct 05:40 < annihilator> Windows was installed but it showed 1.8 TB so I wiped all partitions 05:41 < kerframil> what is "a 4tb raid" exactly? 05:41 < annihilator> To see if it fixed the issue but nope. 05:41 < jim> ok... me still confused :) what do you want to do next with this machine, if anything? 05:41 < annihilator> Install linux 05:41 < jim> ok 05:41 < annihilator> Using my ssd for root and my raid for my home 05:42 < jim> and, do you want to dual boot something else? or linux only? 05:42 < annihilator> kerframil: 4tb raid is a raid setup using hard drives totalling 4tb 05:42 < annihilator> Linux only 05:42 < jim> ok 05:42 < annihilator> I'm going to do qume for my windows needs as it's only for a few games 05:42 < cmj> boy, i wouldn't mind 4tb 05:43 < kerframil> annihilator: that's somewhat nebulous. what raid level? orchestrated by a raid controller or implemented in software? how did you gauge that the capacity appears to be short of what is expected? 05:43 < jim> are you stuck somewhere? or, can we help in some way? 05:43 < ayecee> may be a good time to let him drive the problem solving 05:43 < annihilator> I'm stuck on it not showing total space 05:44 < annihilator> Windows shows 4tb but linux only shows 1.8 both when using uefi booted USB drive 05:44 < jim> ok.. thanks, I think I understand 05:45 < jim> so you have a pair of 2tb drives? 05:45 < annihilator> I no I have 4 1tb drives 05:45 < jim> oh ok, that sounds better 05:46 < annihilator> And raid controller shows all drivers are healthy and online 05:46 < jim> you say you want to put the raid on your home dir... before you do that, do you want to run a database (or other thing that likes truely random access)? 05:47 < jim> how big is the ssd? 05:47 < annihilator> 250gb 05:47 < jim> ok 05:48 < annihilator> I don't need a database its more for storage and games 05:48 < jim> ok 05:48 < jim> that's helpful info for when they decide the raid level 05:48 < annihilator> Its raid 0 05:48 < jim> ok 05:49 < ayecee> raid y0l0 05:49 < jim> you 0nly live 0times? 05:49 < annihilator> You only lose once? 05:49 < ayecee> you only lose one, i guess 05:49 < aclaivi> I wish I only lived 0times 05:50 < annihilator> Lol 05:51 < annihilator> If I have too I can setup 2 raid 0s but only if linux will detect all of it and I can dedicate one raid to qemu 05:51 < kerframil> annihilator: which method are you using to gauge the capacity of the array under Linux? 05:51 < ayecee> asking the important questions! 05:52 < annihilator> As stated before I'm trying to install linux 05:52 < kerframil> without much effect, it seems 05:52 < annihilator> So that would indicate what I'm using 05:52 < ayecee> well, no 05:52 < annihilator> If I'm installing linux I would be in the installer 05:53 < ayecee> it's not clear what method the installer uses, nor which installer it is. 05:53 < annihilator> So then it should be which distro 05:53 < annihilator> Not what method 05:53 < ayecee> couldn't hurt, but it still won't answer the question 05:54 < kerframil> does this installer let you open or drop to a terminal? 05:54 < ayecee> because it's not clear what method the installer uses 05:54 < annihilator> I'm installing linux and the installer is showing 1.8tb 05:54 < ayecee> showing it how 05:54 < ayecee> take a picture 05:54 < storge> this is excrutiating. 05:54 < annihilator> I'm in the kubuntu 18.04 installer 05:54 < ayecee> (also, pretty sure we're not talking about md raid here) 05:54 < kerframil> does this installer let you open or drop to a terminal? 05:55 < ayecee> kerframil: if yes, then what? 05:55 < annihilator> I will have to reboot to the live CD portion but I can 05:55 < kerframil> ayecee: identify the block device node then use blockdev --getsize64 to get the actual size in bytes 05:56 < ayecee> good. alt-ctrl-f1 should get to the console. 05:56 < kerframil> ayecee: so we can determine whether this is really is an issue at the raid driver/controller level or one of the partitioning scheme 05:56 < ayecee> good call. 05:56 < annihilator> But why would linux detect it differently than windows or the raid controller? 05:56 < swift110> What do you guys think of Solus? 05:56 < ayecee> annihilator: that is what we must find out ;) 05:56 < ayecee> swift110: never heard of it 05:56 < epicmetal> swift110: the gui crashed on me while updating the os in a gnome terminal 05:57 < swift110> oh wow ayecee 05:57 < epicmetal> swift110: and the budgie taskbar needs work 05:57 < swift110> epicmetal: when did you last try it? 05:57 < epicmetal> swift110: a few months ago 05:57 < swift110> oh 05:57 < epicmetal> i also tried void which had its own insanities 05:57 < swift110> epicmetal: I have had it installed on my t420 since fall 2016. I really like it 05:58 < ayecee> check that. i've heard the name, and that's about it. 05:58 < storge> epicmetal: what insanities? 05:58 < epicmetal> storge: locking up when switching VC 05:58 < epicmetal> storge: ssh listening by default with a known username/password 05:58 < storge> why would they config that?? 05:58 < storge> yikes 05:58 < epicmetal> storge: i tried telling the dev, he wouldn't listen to reason 05:58 < annihilator> I'll do it but I have to get the dev location so give me a minute 05:59 < RustyJ> google_autneticator TFA is pretty awesome. 05:59 < RustyJ> easy setup and secures shells to the next level, i'm impressed 06:00 < ayecee> in ubuntu it's called libpam-google-authenticator 06:00 < annihilator> kerframil: 1800976728064 06:00 < ayecee> and yes, it's pretty awesome 06:00 < kerframil> annihilator: I see. which device node was that? 06:01 < RustyJ> i ahve a client that has an ex-developer who messes with their servers... this seemed to stop him cold....... for now 06:02 < storge> RustyJ: that's great it's working--for now. just remember, violence is the only solution. 06:04 < annihilator> Node? 06:04 < kerframil> annihilator: for instance, if you used blockdev just then, what did you point it at (/dev/something)? 06:05 < annihilator> Oh it was /dev/mapper/ and I forgot the last part I restarted comp 06:05 < kerframil> annihilator: I'm guessing its dmraid then 06:05 < ayecee> almost certainly 06:06 < annihilator> O 06:06 < annihilator> Ok 06:06 < ayecee> there may still be a protective mbr on there that's messing up the results 06:06 < kerframil> that wouldn't affect blockdev 06:06 < ayecee> erm, no, that's probably not it. 06:06 < kerframil> blockdev just reports the raw block device size 06:07 < ayecee> does dmraid have size limits? 06:07 < kerframil> I've heard of dmraid playing up in this fashion before but I don't know the reason 06:07 < kerframil> annihilator: I forgot, is this going to exclusively run Linux? 06:07 < annihilator> Kde partition manager picks up all hdds 06:07 < annihilator> Yes 06:07 < ayecee> windows is already installed though 06:08 < annihilator> Windows is only in qemu 06:08 < annihilator> Was installed 06:08 < annihilator> I wiped all partitions 06:08 < kerframil> in that case, I think that the path of least reistance would be to de-activate the array so it's back in JBOD (just a bunch of disks) mode, and use Linux's RAID implementation 06:08 < ayecee> would be good to destroy the raid in the bios, and make the raid within the installer 06:08 < kerframil> right 06:08 < kerframil> it's almost certainly more reliable anyway (in so far as RAID0 can ever be called reliable) 06:09 * storge marks this one [SOLVED] 06:09 < storge> ...too soon? 06:09 < kerframil> one thing though 06:09 < ayecee> yes 06:09 < annihilator> I was already taught never use software raid lol 06:09 < annihilator> But then again it was regarding windows 06:10 < ayecee> old knowledge, a carryover from when hardware raid was much more capable than software raid 06:10 < ayecee> the tides have turned on that one. 06:10 < annihilator> I still say software raid in windows is unreliable 06:11 < ayecee> that may well be, i haven't had reason to use it :) 06:11 < storge> Get The Facts 06:11 < ayecee> but if you're using raid0, reliability isn't the focus :P 06:11 < Happyhobo> Hi storge ayecee 06:11 < ayecee> oy, this guy 06:11 < storge> what's up Happyhobo 06:12 < storge> Happyhobo: how goes it in happy hoboville 06:12 < Happyhobo> Remember me? 06:12 < storge> kinda 06:12 < annihilator> As windows is unreliable it's why I'm switching to linux. The only reason I used windows was for gaming but with GPU pass through it's no need for windows native install 06:12 < ayecee> no, but maybe if you hum a few bars of it 06:12 < Happyhobo> I got caught in a dependency loop with Mepis several several years ago. 06:12 < ayecee> annihilator: do you have more than one gpu? 06:12 < annihilator> Yea 06:12 < storge> MEPIS 06:12 < ayecee> ok 06:12 < storge> what's shakin nowadays Happyhobo 06:12 < annihilator> I'm going to run linux of a 780 and use my 1060 for windows 06:12 < Happyhobo> I used to be bad for changing my repos from stable to unstable. 06:13 < ayecee> living la vida loca 06:13 < storge> testing repos are stable enough--at least in my experience in debian 06:13 < Happyhobo> I now run Antergos and my current fight is getting wireless to work by installing a new card and external antennas on my laptop. 06:13 < kerframil> annihilator: I would recommend this course of action: 1) backup any data from the current array if needed 2) destroy the array 3) boot back into linux and verify that each of the four 1TB disks is now exposed as a discrete device 4) run wipefs -a on each of these disks to wipe any residual raid/partition/filesystem metadata 5) carry on with your install and setup RAID from there 06:13 < annihilator> So the advice is to switch bios from raid to achi and let linux raid them 06:13 < ayecee> annihilator: exactly yes 06:13 < storge> Happyhobo: what's antergos based on? 06:13 < Happyhobo> Mepis had its own reiterations to throw into that mix. 06:14 < annihilator> Ok and I already backed upwindows on an external and wiped all hdds 06:14 < Happyhobo> Arch, it's a simplified version with no support that a lot of folks say will die in 6 months. 06:14 < storge> brb rebooting 06:14 < annihilator> Kde shows all hdds so that's not a problem I'll go change bios hdd driver 06:14 < kerframil> annihilator: the mere fact that it is currently an array means that the disk is not clean. do not trust your raid controller to wipe the metadata clean just because you delete or deconstruct an array. 06:15 < jim> Happyhobo, more likely it'll stay alive due to folk's preference 06:15 < ayecee> that is the sound of experience right there 06:15 < ayecee> this guy raids 06:16 < kerframil> annihilator: this is exactly what wipefs is designed for. it recognises all signatures that libblkid is aware of and wipes them. that can prevent nasty situations later where the Linux host can be confused as to exactly how the devices should fit into the storage stack overall. 06:16 < Happyhobo> I hope so because I really enjoy it jim other than the fact that I can't do anything cli other than dd ls cp a single file LOL 06:16 < annihilator> Not clean like that I used windows to wipe the partitions windows created to see if linux would recognize the raid properly 06:16 < kerframil> annihilator: I've even seeen bootloaders get confused over failing to wipe a residual signature. helped a guy out a few months ago with this issue.l 06:17 < annihilator> As in there is no data for me to care about 06:17 < kerframil> well, whatever 06:17 < kerframil> the advice stands 06:17 < ayecee> that's the spirit 06:17 < annihilator> Or need to back up 06:17 < ayecee> dd if=/dev/zero all the things 06:17 < annihilator> I'm going to follow that as the raid data is still there if when raid is related 06:17 < annihilator> Deleted 06:17 < storge> Antergos. your linux fresh. never frozen. 06:17 < storge> that's a heady vapor there 06:17 < ayecee> oh, sorry, that was me 06:18 < Happyhobo> I still don't see how you can dd back on to the drive you're dding from. 06:18 < annihilator> I learned that the hard way kerframil 06:18 < storge> Happyhobo: well cool, how do you like it? 06:18 < ayecee> Happyhobo: how do you mean? 06:18 < kerframil> annihilator: as long as the individual disks (/dev/sda, /dev/sdb ... etc) in JBOD mode are rendered clean by wipefs -a, you're good to go. just make sure you run it on the right devices! don't go picking your SDD or flash drive by mistake. 06:18 < annihilator> I installed windows 7 after recreating a raid and fed it the hell up 06:19 < Happyhobo> storge it's great there is never a dependency issue and it is always up to date 06:19 < ayecee> Happyhobo: if you're talking about /dev/zero, those zeros aren't on the disk, they're manufactured on the spot 06:19 < kerframil> annihilator: actually, you could do your SDD if that's going to be the root filesystem but, anyway, just be careful 06:19 < ayecee> fresh blocks of zeros, just for you 06:20 < annihilator> I was going to let linux raid my hdds and leave my ssd alone and let ot setup root on it outside of the raid 06:20 < storge> fresh zeroes. never frozen. 06:20 < kerframil> sounds good 06:20 < annihilator> Similar to what I did with wi dows 06:20 < Happyhobo> OH, I was referring to dd if=/home/me/homiedidthisonce.iso of=/dev/sda and it ate itself 06:20 < annihilator> Os on ssd and user data on raid 06:20 < ayecee> is zfs an option for kubuntu? zfs has nice options for tiered storage 06:21 < Happyhobo> Zelda fries shrimp 06:21 < epicmetal> zebra foot soldier 06:22 < Happyhobo> zoo fondling skank 06:22 < ayecee> zoroastrianism feeds on sacrifices 06:22 < storge> hmm, i'd hoped zfs made it into the kernel by now 06:22 < ayecee> (zoroastrianism, great scrabble word if you ever get to play it) 06:23 < lunaphyte> why can't an unprivileged user see their own info with lsof? : http://dpaste.com/1CQYGMP.txt 06:23 < nekoseam> GitLab is switching from azure to google cloud 06:23 < nekoseam> test 06:23 < annihilator> kerframil: how do I let linux raid software wise? 06:23 < ayecee> nekoseam: we see you. we just don't care. 06:23 < nekoseam> ayecee: I had a conflicting hotkey 06:23 < storge> zoroastrianism, source the of the christian ritual of drinking wine as jesus blood; they drank a distillation of a plant called the son of god, presto, drinking the blood of the son of god. given to the west, via hashishim 06:24 < ayecee> it's all a rich tapestry 06:24 < kerframil> annihilator: if you're using a mainstream distro like ubuntu, I think that the installer should present the options. but you could set it up from a terminal if you prefer (with mdadm). 06:24 < jim> unfortunately I misread that as "I had a conflicting monkey" 06:25 < annihilator> I'm using kubuntu 18.04 cause I'm too lazy to setup the ability to install steam on lfs 06:25 < storge> haven't we all 06:25 < ayecee> storge: looks like zfs is native on debian and derivatives now 06:25 < kerframil> annihilator: in fact, for the kind of setup you're going for, you wouldn't even need to partition the four member disks. they could be integrated into the array as 'whole' block devices. 06:25 < jim> yeah, didn't make sense to me either 06:25 < kerframil> annihilator: the installer-guided method may or may not expect that they be partitioned though 06:25 < storge> ayecee: but not zfs support in kernel? 06:25 < ayecee> that is what native would imply 06:25 < storge> root@T520-debian:/usr/src/linux-4.17# zgrep -i zfs /proc/config.gz 06:25 < storge> root@T520-debian:/usr/src/linux-4.17# 06:25 < ayecee> oh wait 06:25 < kerframil> annihilator: either way is fine so go with the flow, I suppose 06:26 < ayecee> "Debian GNU/kFreeBSD systems" :P 06:26 < storge> ah 06:26 < ayecee> the part after debian was on the next line 06:26 < ayecee> so deceptive 06:26 < storge> maybe someday 06:27 < kerframil> annihilator: actually 06:27 < ayecee> i was big on btrfs as a challenger until i discovered that its load balancing is pid-based - a given pid will read from only one of the disks in a mirrored raid. 06:27 < kerframil> annihilator: as the root filesystem isn't even going to be part of the array, you could just install ubuntu to the ssd first then set up the array after you've booted into it 06:27 < ayecee> though it will write to both, of course. 06:27 < Happyhobo> night folks 06:27 < tvm> grr, someone recommend me diagram drawing tool that doesn't suck 06:27 < storge> ayecee: i wonder why they did that 06:28 < storge> pencil, paper 06:28 < tvm> yeah, probably best solution 06:28 < storge> seriously 06:28 < Happyhobo> I hope my new antennas have enough cable to make the antennas stick ouot like bunny ears 06:28 < ayecee> storge: i would think because it's very easy to do, and does reasonable load balancing in multithread or multiprocess scenarios. 06:28 < storge> if i really care that much, i'll duplicate a diagram in cad the next time i use it, but i just draw them. luckily i'm a line art artist 06:28 < ayecee> where the number of threads and processes reading is high 06:29 < storge> ayecee: i see 06:30 < ayecee> might even favor more multithreaded design 06:31 < ayecee> however i have a bitcoin daemon that is not multithreaded 06:31 < Hashtag> I'm having some problems with PATH. I don't know if I should ask here because it's on a mac terminal 06:31 < ayecee> it hammers one disk and leaves the other virtually untouched 06:31 < ayecee> Hashtag: probably not then 06:32 < Hashtag> ayecee: Well, it's pretending to be linux, isn't it? 06:32 < ayecee> Hashtag: sure. be sure to bring your parallels problems to a mac channel too. 06:32 < ayecee> err, windows channel 06:33 < ayecee> (is parallels still a thing?) 06:33 < Siva_Machina> proprietary OS channel 06:33 < ayecee> Hashtag: in any case, it would be better to describe the problem with the caveat, than describe the caveat without the problem 06:34 < Hashtag> So my PATH according to env is PATH=/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin but when I type $PATH it says "No such file or directory" at the end 06:34 < Dagmar> THat's because the command "/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin" makes no sense 06:35 < Hashtag> Missing a space? 06:35 < ayecee> Hashtag: what should it do instead? 06:35 < Dagmar> ayecee: GOod question 06:36 < ayecee> because that is exactly what typing $PATH alone would do in linux 06:36 < Dagmar> What on earth must he have expected that would o 06:36 < Dagmar> er do 06:36 < Hashtag> ayecee: Normally is just lists everything in the path, right? The PATH= and then anything appended to it 06:36 < Dagmar> No. 06:36 < kerframil> no 06:36 < Dagmar> Absolutely not 06:37 < ayecee> Hashtag: no 06:37 < Hashtag> Seems like strange default behavior 06:38 < ayecee> it seems like a fundamental misunderstanding of how variable expansion works 06:38 < Elladan> ayecee, there... are so many things that are questionable with btrfs, I don't know that the load balancing is where I'd start 06:38 < Elladan> I mean it works, it's just not the fastest. 06:38 < Dagmar> Maybe something like IFS=: for blerp in $PATH; find -maxdepth 1 -type f -mode +0111 -exec echo {} \; done 06:38 < ayecee> Elladan: and yet, that's where i started 06:38 < Dagmar> It's still insane 06:38 < ayecee> Elladan: it was a problem i noticed. 06:38 < Elladan> ayecee, fair enough. 06:38 < ayecee> there's a lot of btrfs problems that i wouldn't notice. 06:38 < Hashtag> I see the problem. I wanted 'echo $PATH' 06:39 < Dagmar> Hashtag: When you type something into bash that has a variable in it, the variable is substituted 06:39 < ayecee> Hashtag: that would echo the contents of the variable, yes 06:39 < ayecee> but not list the programs within that path 06:39 < Elladan> ayecee, here's another one: if you have one thread doing continuous streaming writes, and another thread doing small synchronous writes (e.g. pwrite+fdatasync, as a VM would), the second thread will experience unbounded latency >300s. 06:40 < ayecee> makes sense 06:40 < Elladan> ayecee, in other words if you write a large file all the VMs hosted on the filesystem may freeze indefinitely and crash. 06:40 < ayecee> would be better if vms weren't doing synchronous writes for every transaction 06:40 < ayecee> in qemu, using writeback instead of none for cache. 06:40 < Dagmar> Ya'll gotta be doing something really weird there 06:41 < Elladan> ayecee, they have to do synchronous writes or it's incorrect. 06:41 < annihilator> Ok I did lsblk and it shows dmraid.... 06:41 < ayecee> only if physical machines have to do synchronous writes for everything 06:41 < ayecee> which they don't 06:41 < Elladan> ayecee, I mean you're free to set it up to be incorrect, but then you should expect the VMs to corrupt on a reboot. 06:41 < Dagmar> I do not experience anything like that over here 06:42 < Elladan> ayecee, that's not true. 06:42 < ayecee> but it is 06:42 < Elladan> *shrug* as you like. 06:43 < ayecee> on physical machines, there are write barriers that ensure that critical data is flushed, while most writes are not critical. 06:44 < Elladan> Indeed. 06:44 < kerframil> annihilator: that's to be expected 06:44 < annihilator> I can't use mdadm because it says it only recognize 2 not the 4 06:45 < annihilator> But I can see all 5 disks 06:45 < Elladan> On VMs, there are also write barriers for the same reason. The hypervisor implements them by doing a synchronous write (more accurately, a forced stable write). 06:45 < annihilator> 4 hdds and one ssf 06:46 < Elladan> If you turn off that feature in the hypervisor, the VM will be subject to the same corruption a physical machine is if its write barriers aren't honored by the hardware. 06:47 < annihilator> Got it had to restart the live cd 06:48 < ryanjg> mod_rewrite 06:48 < ryanjg> is causing me endless frustration rn 06:50 < ayecee> Elladan: that's accomplished by using writeback cache rather than none cache in qemu 06:51 < ayecee> as far as i can tell 06:52 < ayecee> at least, the performance for none cache looks like a sync mount, and writeback looks like typical write barriers. maybe i'm misunderstanding that. 06:53 < ayecee> i should probably check that 06:53 < annihilator> kerframil: thanks finally got it to be able to be installed the way I want it. 06:53 < ayecee> this is suggestive of that: https://lists.nongnu.org/archive/html/qemu-devel/2009-05/msg00212.html 06:53 < nekoseam> Is ryzen 3's iGPU supported well by kernel 4.16? 06:54 < ayecee> may not reflect current development 06:54 < kerframil> annihilator: which option did you go for in the end? setting up raid as early as possible or will you defer until the install is done? 06:54 < ayecee> this suggests same: https://www.suse.com/documentation/sles11/book_kvm/data/sect1_1_chapter_book_kvm.html 06:55 < annihilator> I did before install 06:55 < annihilator> I had to reboot live cd after the wipefs and it listed all hdds correctly. 06:55 < Elladan> ayecee, I believe both writeback and none implement barriers correctly in qemu, but I don't know all the qemu settings off the top of my head. 06:55 < annihilator> Should I do a 2gb swap or 4? 06:56 < ayecee> cache=unsafe acts like cache=writeback except that flush requests are not honored 06:56 < annihilator> I have 24gb of ram 06:56 < ayecee> annihilator: 2 06:56 < annihilator> Ok 06:57 < Elladan> ayecee, IIRC virtualbox defaults to not implementing barriers (i.e. fsync) and going full writeback, which means that many IO benchmarks are incorrect. 06:58 < ayecee> Elladan: from the last url, it looks like writeback and none both honor flushes, but the cache is limited in that case to the cache on the disk, not on the host. 06:58 < Elladan> ayecee, yeah cache=unsafe would be similar to the virtualbox setting. 06:58 < storge> show of hands, who uses initrd 06:58 < ayecee> err, in the case of no cache, it'd be limited to the drive's cache. 06:58 < annihilator> Its been a while doing 2 disk setup with linux. / Is on ssd and /home is on raid. Is there anything else I should throw on the raid for data? Like steam games? 06:59 < ayecee> annihilator: the ideal would be to use bcache or dmcache to treat the ssd as a cache for the raid. 07:00 < Elladan> ayecee, yeah from Linux's point of view qemu is just issuing read/write/f*sync calls with or without O_DIRECT. 07:02 < ayecee> as such, using either the host page cache or the disk cache as storage between flushes 07:02 < ayecee> should be no appreciable difference in data integrity, and a big performance boost for using writeback 07:04 < annihilator> ayecee: what you mean? 07:04 < ayecee> i mean what i said 07:04 < annihilator> Why not using ssd as is drive? 07:04 < annihilator> Os* 07:05 < ayecee> annihilator: because the benefits of ssd would only be for data on the ssd, rather than all the data. 07:05 < ayecee> if the ssd is a cache for the hdd raid, most reads and writes can be accelerated by the ssd 07:05 < hexnewbie> SSD cache for a RAID isn't really as great as having /home on the SSD - performance is worse, with dmcache write patterns can make it slower than hard drive (causing processes to wait for cache flushes), and if SSD is not in RAID you actually have *less* redundancy making RAID pointless.. Caveat: I used dmcache, not bcache, and bcache is superior. I spent all of last Friday migrating "SSD cache for RAID" to proper SSD RAID, because it's a 07:05 < hexnewbie> completely idiotic setup. 07:06 < ayecee> man, you really gotta take more time to edit 07:06 < ayecee> hexnewbie: on average, it's better. for specific cases, it's worse. 07:07 < Elladan> annihilator, I didn't understand your setup. Do you mean you have 1 SSD and 2 rust drives in a RAID? 07:07 < ayecee> yes, that 07:07 < hexnewbie> Or, well, was a pretty idiotic setup for me (based on using it for two years) 07:07 < ayecee> well, 4 rust drives 07:07 < annihilator> Rust drives lmao 07:07 < ayecee> :D 07:07 < annihilator> But yes I have 4 hdds and 1 ssd 07:08 < ayecee> ideal scenario for tiered storage 07:08 < Elladan> Personally I'd suggest you get a second SSD and RAID them. 07:08 < annihilator> Under windows I used my ssd for only startup my games and user data was on raid 07:08 < ayecee> comeon, he's using raid0, data integrity is not a big concern. 07:08 < annihilator> I have an external for important stuff 07:09 < ayecee> well done 07:09 < Elladan> Maybe your SSD has better failure characteristics than rust, but it seems kind of weird to have only one SSD if you're worried about losing disks 07:09 < ayecee> he's not. 07:09 < ayecee> the disks are in raid0. 07:09 < Elladan> Ah ok. 07:09 < annihilator> I'm wanting fast boot time and mass storage 07:09 < Elladan> Sorry missed that. 07:09 < ayecee> raid yolo, "you only lose one" 07:09 < annihilator> My game load times I dont care about 07:09 < ayecee> XD 07:10 < Elladan> In that case, I'd probably put / and /home on the SSD, and made the raid /big or something. 07:10 < annihilator> I been running MMOs ony raid for a while its because windows is bugging out for my programming that I'm going to linux 07:10 < hexnewbie> I'd go for /, /home and swap on SSD, /data on the RAID 0. But I'd also go for SSD RAID-1 and HDD RAID-10, so one's mileage may vary. 07:11 < ayecee> raid all the things 07:11 < annihilator> On windows I threw my user folder on raid 0 07:11 < Elladan> Using automatic tiering e.g. bcache is another option, I wouldn't say it's necessarily "ideal" but it's certainly good. 07:11 < annihilator> And the OS with only startup programs such as steam itself was on ssd 07:11 < ayecee> it extends the benefits of ssd to bulk storage, without having ssd bulk storage 07:12 < ayecee> usually enough, given that the working set for a given user is usually smaller than the ssd. 07:12 < Elladan> It really depends on your use case whether automatic tiering works well for you or not. 07:12 < Elladan> There's nothing stopping you from doing both at once though. 07:12 < annihilator> And that is the only setup I really care about if there is a way to improve on that then cool other than that home is where I store my data so it's why I was putting it on the raid 07:12 < Elladan> You can just make a partition on the SSD for bcache and also have /, /home, and swap partitions as needed. 07:13 < annihilator> Will bcache speed up boot times? 07:13 < annihilator> Its the only speed I care about lmao 07:13 < ayecee> i have a freenas-based tiered storage setup at the office, and over the long term 50% of requests are serviced by a 120gb ssd. 07:14 < Elladan> Not if you have your OS and the files it uses to boot on the SSD. 07:14 < ayecee> for some 20ish users 07:14 < annihilator> Which is faster bootimg from ssd or bcache 07:14 < ayecee> ssd 07:14 < Elladan> annihilator, putting /home on the SSD will help boot times somewhat, in as much as you won't need to use the rust to read in your user config files etc. 07:14 < annihilator> It may be crazy but is all I care about lmao 07:14 < iflema> sleep 07:15 < iflema> hibernate 07:15 < ayecee> shutdown now -h 07:15 < annihilator> Well I'm used to that with windows as my desktop icons and files are in rust 07:15 < Elladan> In general bcache / auto-tiering is meant to help average performance, not rare events like boot time. 07:16 < annihilator> Rare events 07:16 < annihilator> Lmao 07:16 < annihilator> I guess I'm to used to windows 07:16 < hexnewbie> Yeah, moving your icon theme to SSD visibly increases performance in KDE. I suspect something similar may be true for .desktop files, but those may have cache in some DB somewhere. 07:16 < annihilator> No 07:16 < annihilator> .desktop and all that on windows is not db stored.... 07:16 < Elladan> annihilator, yeah my laptop's been up 35 days. 07:16 < Dagmar> Okay that's kind of sad if a simple ram cache makes it go faster 07:16 < annihilator> Its read from hdds dieing every boot 07:17 < annihilator> During 07:17 < annihilator> I know linux is stable it's why I'm switching since I have 2 viable gpus 07:17 < hexnewbie> During boot is OK. KDE reads icon theme during *menu opening*, meaning that if you're not on SSD you wait some time for the menu to pop up. 07:18 < Dagmar> So... if you've got spacious RAM, force it into filesystem cache 07:18 < Elladan> Also, there's a prefetch system that Linux distros often use for boot, which actually helps quite a bit with spinning disk boot time. 07:18 < annihilator> I have 24gb I was going to give qemu 12 07:18 < Elladan> However it doesn't usually include the bits you load in during account login, so it's more to get you to the login screen fast. 07:18 < iflema> i hate fast boot... no boot splash 07:19 < iflema> atleast no time to see it 07:19 < ayecee> the boot splash that says "you got time, go get some coffee" 07:19 < Elladan> Heh 07:19 < annihilator> So for me as long as I'm not doing the windows crash on linux caching would be better? 07:20 < Elladan> My desktop actually spends whatever boot would normally take (a few seconds) + 90 seconds right now, because of some bug in how I have initrd + cryptsetup configured. LOL. 07:20 < ayecee> hard to interpret that question 07:20 < annihilator> Sorry 07:20 < Elladan> ... but I reboot it every few months so not really worth fixing. 07:21 < Dagmar> That's the spirit. 07:21 < Elladan> annihilator, are you asking about bcache / dm-cache when you mean caching? 07:21 < annihilator> If I'm not "bsod" on linux (since I'm prolly not having top reboot every 2 hours like windows), caching would be better? 07:21 < Elladan> Er, say caching rather. 07:21 < annihilator> Yes ell 07:21 < Elladan> Ah. 07:21 < ayecee> still a weird question, but caching would still be better even if you were bsoding. 07:22 < Elladan> I'd recommend you put / and swap on the SSD regardless. 07:22 < ayecee> journalling ftw 07:22 < Elladan> How big is your SSD? 07:22 < Dagmar> I would be deeply concerned about more than one "bsod" a month 07:22 < annihilator> 250gb 07:22 < ayecee> that's kind of personal 07:22 < iflema> Dagmar: my windows is like that. it doesnt shut down but it reboots so into linux it goes... no issues =) 07:22 < annihilator> And do I need a swap other than aesthetics? 07:22 < annihilator> I have 24gb ram 07:22 < Elladan> 250 G is big enough that if you don't put /home on the SSD or use it as cache, it's going to be mostly wasted. 07:23 < ayecee> annihilator: some swap can still improve performance by making more room for disk cache. 07:23 < Elladan> Yes, you should have swap. 07:23 < Dagmar> iflema: THat means you get to enjoy all that time spent wonder why the eff Microsoft thinks it's okay to spend an hour doing visibly nothing and rebooting twice for a simple update 07:23 < ayecee> too much swap could lead to thrashing if you ever have a runaway memory-hungry process. 07:23 < annihilator> I figured if I bump up 32gb ram I could do a ram swap drive with 2gb 07:24 < Dagmar> Easier than that 07:24 < ayecee> ram swap makes no sense, though zram swap could 07:24 < Elladan> Was about to say that yes. 07:24 < Elladan> zram helps on systems with memory pressure, I'd imagine it's less useful with 24+ GB 07:25 < Elladan> annihilator, the other thing I'd make sure to suggest is to use LVM. 07:25 < annihilator> Lvm>md? 07:25 < iflema> Dagmar: i had to wait for two reboots to told the update failed not so long ago for about month... it fixed itself 07:26 < Elladan> It's much easier to resize volumes and all that later if you use LVM. 07:26 < ayecee> still useful in that rarely used pages are compressed, freeing up more ram 07:26 < Dagmar> iflema: I would rather that crap happen when I don't have to watch it 07:26 < iflema> Dagmar: i was getting tea or coffe... worked out well 07:26 < iflema> =) 07:26 < annihilator> To make my life simple I'm just going to use ssd for caching so I dont accidentally do something stupid with /home 07:27 < ayecee> cool! i haven't tried that myself on desktop system. i usually have bulk storage on some random server in the basement. 07:27 < Elladan> annihilator, beware that bcache/dm-cache are not really as stable as using separate filesystems, so people sometimes have issues and lose all their data. 07:28 < Elladan> You may wish to use bcache in "writethrough" rather than "writeback" mode for safety, but you'll get less in the way of performance benefits that way. 07:29 < ayecee> current desktop system just has the one 500gb ssd. 07:31 < annihilator> Elladan: so doing / ssd is safer but slower doing / and /home and /swap is better and doing bcache is good for lazy people but risks data? 07:32 < storge> i kinda like that idea, right now in this laptop i have an hdd and an ssd, ssd has the system which i like but it's good to consider another way 07:32 < Elladan> annihilator, I'd always do / on ssd regardless. 07:32 < annihilator> Ok 07:32 < annihilator> So / and bcache would be less risky? 07:33 < annihilator> Or just say scew it and play till I find what I like? Lol 07:33 < Elladan> bcache is somewhat risky-ish regardless. It's more whether you use writethrough or writeback mode. 07:33 < storge> iterate to satisfaction 07:33 < ayecee> i figure / _on_ bcache or dmcache 07:33 < ayecee> but i haven't tried it 07:33 < Elladan> Yeah honestly since you're new to this I'd really just suggest you put /, swap, and /home on SSD, make the RAID /data or something, and use LVM so it's relatively easy to move things around. 07:33 < ayecee> would be kind of like a sshd with a really big cache. 07:34 < ayecee> sure, that's a good conservative approach 07:34 < Elladan> You can figure out bcache later and move files into it if you like it. 07:34 < annihilator> I'm completely new to bcach 07:34 < ayecee> it's kind of hard to retrofit bcache onto an existing system 07:34 < storge> i haven't used it either 07:34 < ayecee> you'd have to plan for it 07:35 < annihilator> I'm used to old school raid all on one back up on external 07:35 < storge> i would think you need a plan! i'd have to tread very carefully 07:35 < Elladan> That's somewhat true, yes. I believe there's a tool to convert the backing store to bcache but it's somewhat questionable. 07:35 < ayecee> planning for it would involve setting it up as a bcache but without a cache. 07:35 < Elladan> I'm not sure about dm-cache. 07:36 < Elladan> Just using lvm is probably enough. 07:36 < Elladan> Assuming you have enough free space, you can just have 2 partitions. 07:36 < ayecee> so many options! 07:36 < Elladan> ... or have one and shrink it when you want to convert. 07:38 < storge> anyone here using zram? 07:38 < ayecee> no, you are the first 07:38 < Elladan> I've used it in the past. 07:38 < storge> surely not the only one 07:39 < ayecee> was intended to be a facetious response to a silly question 07:39 < ayecee> if you have a question about zram, ask a question about zram 07:39 < storge> well i wasn't really asking among the entire user list, just the people talking 07:39 < ayecee> don't poll 07:39 < storge> i'm sure i'm sorry 07:39 < nekoseam> With x86 being found out to be utter shit I've been looking at some old PowerBooks. Apparently Gentoo has great PPC support but I'm afraid the early C2D equivalent CPU would have too little horsepower for it to be worth it 07:40 < ayecee> nekoseam: mind the language please 07:40 < ayecee> nekoseam: also, wtf. 07:40 < nekoseam> >mind the language 07:40 < nekoseam> >wtf 07:40 < storge> gentoo has great ppc support, what does that mean 07:40 < ayecee> ppc development is pretty much dead end. 07:40 < Elladan> If you really don't like x86, the path of least resistance these days would be some little ARM box. 07:40 < nekoseam> Compiling a kernel on my 3rd gen i5 already takes forever 07:41 < ayecee> do you know why that is? 07:41 < storge> nekoseam: localmodconfig 07:41 < Elladan> i.e. a raspberry pi type thing. 07:41 < nekoseam> POWER9 looks really good 07:41 < ayecee> the answer is, no, i've never checked 07:41 < storge> why compile for the universe 07:41 < ayecee> never known where the quantum uncertainty drive will take you 07:42 < storge> i remember (rather fondly) putting walnut linux on some old ppc miniboards 07:42 < luke-jr> nekoseam: more or less 07:42 < luke-jr> I should update my kernel anyway 07:43 < ayecee> compiling your own kernel is a neat way to explore options currently in development 07:45 < storge> quite a lot of new drivers in 4.17 that i compiled tonight 07:45 < storge> it's interesting to see if if it's just N N N N 07:45 < storge> even if* 07:45 < ayecee> huh. this says the ZFs kernel module is provided by default in ubuntu: https://wiki.ubuntu.com/ZFS 07:46 < storge> where is that patch :) 07:47 < storge> zfsonlinux hmm 07:47 < ayecee> :) 07:47 < ayecee> i think that's the fuse package 07:47 < ayecee> not the native package 07:48 < storge> libzfs2linux - OpenZFS filesystem library for Linux 07:48 < storge> oh 07:49 < luke-jr> doh, 4.17 not in gentoo-sources yet 07:50 < ayecee> i wonder how much work it would be to become a gentoo package maintainer 07:50 < storge> wget -c https://cdn.kernel.org/pub/linux/kernel/v4.x/linux-4.17.tar.xz 07:50 < storge> :) 07:51 < ayecee> who does kernel.org cdn these days? cloudflare? fastly? 07:51 < storge> dunno, but it's fast 07:51 < ayecee> ah, fastly 07:52 < ayecee> (so says nslookup) 07:52 < storge> it's somewhat recent i think, i ran a script that had www that failed 07:52 < storge> somewhat meaning maybe a year since i ran the script 07:54 < storge> it got me thinking, when i saw the initrd size and stripped of debug, that i might want to try without it, for hobby sake. not like size matters. 07:57 < ayecee> one way to decrease the size of the initrd is to reduce the number of drivers compiled into it 07:57 < ayecee> one way to increase the size of the kernel is to compile everything into it instead of the initrd. 07:57 < storge> yep 08:01 < storge> i run a localmodconfig that it's pretty trim but i must have enabled some debug that was making initrd larger, so i used install mod strip 08:01 < ayecee> strip all the things 08:02 < storge> now i'm finding whatever debug i set. 08:04 < horseface> i love my linux system it's soo good 08:04 < horseface> :) 08:04 < horseface> yay 08:04 < storge> +1 08:05 < horseface> :) 08:05 < iflema> +3/3 08:08 < horseface> hey what is a good ebook cataloger? 08:09 < Sveta> jabref 08:09 < storge> bash 08:09 < horseface> i want one which kind of mimics the Books application on apple iOS devices, like shows them in view on shelves and allows you to choose and read them... 08:10 < horseface> ill have a look at jabref 08:10 < storge> me too, i'm not psyched about what i've tried so far 08:11 < storge> jabref - graphical frontend to manage BibTeX and BibLaTeX databases 08:11 < storge> thanks for asking, answering 08:17 < annihilator> Lvm with lvcache lol 08:18 < annihilator> I must be doing something wrong tho cause mdadm is easier for me to use than lvm 08:20 < annihilator> If I'm using my raid as just data then couldn't I go real old school and just do /, /boot, /swap 08:27 < jim> annihilator, if you want you could use the raid for lvm PVs 08:28 < Furai> Hey, some weird thing with vsftpd, service cannot start, it says it started vsftpd but it's inactive when doing status check. 08:29 < storge> status check? from within vsftpd? 08:29 < Furai> From systemctl, ubunt 16.04 08:29 < RustyJ> 'ps uxa'? 08:29 < Furai> No running process. 08:29 < Furai> When running directly, works. 08:29 < Furai> Just the init/service fails for some reason. No error output whatsoever. 08:30 < Furai> directly i.e. /usr/sbin/vsftpd 08:30 < Furai> I'm really puzzled here. 08:30 < ayecee> i wonder if there's some logs that could shed light onto this 08:30 < RustyJ> not to be argumentative but why vsftpd? is sftp out of your usecase? 08:31 < ayecee> could be that journalctl could say something useful 08:33 < Furai> Clients wants to use that, I'm just contractor that fixes their linux instances. 08:33 < Furai> Ok, will look at journalcts 08:33 < ayecee> hard to find good help these days 08:34 < storge> and you get mad when people swear 08:34 < Furai> hah 08:34 < ayecee> zing 08:34 < Furai> Btw, thanks for helping me. journalctl has no additonal output 08:35 < ayecee> what output does it have? 08:35 < Furai> It must be something ec2 related. Dunno, maybe network.target isn't satisfied? 08:35 < Furai> Just starting vsftpd and then started vsftp 08:35 < Furai> But there is no proccess running. 08:35 < Furai> Again, running directly works like a charm. 08:36 < ayecee> grep -ilr ftp /var/log/* , any output? 08:36 < storge> start it with verbosity? 08:36 < ayecee> also, how are you starting the service, aside from manually? 08:37 < jim> sometimes there is a port displatcher thing, inetd, that listens for various ports as it's configured to, and (if conrigured so) dispatches the ftpd process it's configured to 08:37 < ayecee> but probably not 08:37 < Furai> "[ ok ] Starting vsftpd (via systemctl): vsftpd.service. " when run via init.d script 08:37 < jim> yeah it's been awhile since I've heard of it used 08:37 < Furai> And nothing is there. 08:38 < ayecee> Furai: how about after "service start vsftpd" ? 08:38 < Furai> That's how I was running it till now. 08:38 < jim> Furai, try ftping to it 08:38 < Furai> Tried init just now 08:38 < ayecee> what happens when you do that? 08:38 < Furai> Can't ftp, port is closed. 08:38 < Furai> ps auxw | grep ftp shows no running process 08:38 < jim> so you get connection refused? 08:38 < Furai> yes 08:39 < ayecee> what shows up in journalctl when you do that? 08:39 < Furai> But when I run it manually by hand, like runing the executable it works 08:39 < ayecee> immediately after "service start vsftpd" 08:39 < jim> ok, yeah, so it's not running 08:39 < epicmetal> Is there some nice program that tunes your system for desktop use, e.g. "don't ever skip youtube audio when copying files to USB" 08:39 < jim> can you tail /var/log/daemon.log while you try to stert it 08:39 < epicmetal> I'd rather not do it all manually 08:40 < Furai> ayecee, like I said before 08:40 < ayecee> epicmetal: i can't think of any reason why youtube audio would be skipped when copying files to usb. 08:40 < Furai> Starting vsftpd FTP server... 08:40 < Furai> Started vsftpd FTP server. 08:40 < Furai> Only these 2 lines. Nothing else. 08:40 < Furai> It like exits somewhere early. 08:40 < epicmetal> ayecee: i.e. the sound cuts out/in frequently during the copy operation 08:40 < ayecee> Furai: okay, how about the output from that grep line? 08:40 < Furai> Nothing is syslog either. 08:40 < jim> ok 08:41 < jim> I'm out of ideas... 08:41 < epicmetal> ayecee: i figure it's a bunch of kernel settings 08:41 < ayecee> epicmetal: ah, the description is misleading. 08:41 < epicmetal> i don't care if it slows down the copy a bit 08:42 < epicmetal> audio stutter in 2018 is embarassing 08:42 < Furai> ayecee, https://file.furai.pl/2018-06-06-08-41-56.txt 08:42 < ayecee> it's not that you want to skip copying youtube audio, but that you want to prevent youtube audio from skipping while you copy. 08:42 < epicmetal> ayecee: yes 08:42 < epicmetal> ayecee: i mostly don't want my coworkers to laugh at my os choice ;) 08:42 < storge> #priorities 08:43 < storge> i totally get it 08:43 < epicmetal> storge: i have them 08:43 < horseface_> i like the look of bookworm but it doesn't wory as well as it looks, can someone recommend me something like it? 08:43 < storge> i just want to look l33t in the airport with my hacker desktop 08:43 < ayecee> i wonder if there's a way to prioritize the interrupt from the sound device. 08:43 < epicmetal> i keep getting told to install win10 for various reasons (not this reason in particular) 08:44 < epicmetal> storge: nothing like being hauled off by security for using a tilingn wm 08:44 < epicmetal> tiling* 08:44 * rouji uses a tilting wm 08:44 < ayecee> Furai: grep ftp /var/log/syslog, what does that return? 08:44 < storge> epicmetal: while your android's plugged in with adb in a terminal blasting traffic up the screen 08:44 < epicmetal> rouji: sounds fun 08:44 < ayecee> jiggle the tiling vm until it tilts 08:45 < Furai> Just the start stop messages. 08:45 < Furai> Anyway, trying dpkg-reconfigure 08:45 < Furai> Could locales be a problem if not set properly? 08:45 < ayecee> Furai: bummer. would be handy to find some way to make it more verbose, but i have nothing offhand. 08:46 < ayecee> maybe? seems unlikely. 08:46 < ayecee> might have to hit up the vsftpd service file to make it more verbose 08:46 < jim> you could look at the startup scripts, sometimes there's a flag whether to start it... but to have it off doesn't seem like a reasonable default 08:47 < jim> do you have any files /etc/default/*ftp*? 08:47 < storge> as i understand it, journalctl doesn't prevent normal syslogging if you want, yes? 08:48 < ayecee> logs from journalctl are typically copied to syslog 08:48 < ayecee> to the syslog service, that is 08:49 < jim> that's from Oregon? 08:49 < ayecee> wat 08:49 < ayecee> you have died from dysentry 08:49 < Dagmar> heh 08:49 < jim> hmm! if I died, I probably shouldn't post! 08:50 < ayecee> it would be rude 08:50 < Dagmar> At least dwarves never die of dysentary 08:50 < storge> so i have an idea 08:50 < ayecee> i am skeptical, but open-minded 08:50 < jim> they probably feed on it 08:50 < Dagmar> storge: Wash your hands after touching it 08:50 < ayecee> lol 08:50 < storge> look at 'man vsftpd.conf' and search for xferlog_enable 08:51 < Furai> jim, nope nothing. 08:51 < Furai> Anyway, guys, thanks for all the tips. I'll figure it out eventually. :) 08:51 < storge> oh that's for transfers 08:51 < storge> crap 08:51 < ayecee> (plot twist - vsftpd on osx) 08:51 < storge> does vsftpd really default to no logs? 08:52 < storge> look at 'man vsftpd.conf' and search for vsftpd_log_file 08:52 < storge> and syslog_enable 08:52 < ayecee> no u 08:52 < jim> Furai, ok, best of luck 08:52 < storge> maybe those aren't 08:52 < jim> please spell out u as you, it helps people (particularly new english speakers) to understand, at least, most of what's going on "P 08:52 < storge> Furai: see what i mentioned? 08:53 < storge> Furai: you might be able to force vsftpd to log to where you want 08:53 < ayecee> you better not have that macroed 08:53 < jim> :P 08:53 < storge> anyway good luck 08:54 < Furai> storge, ok, thanks again. 08:54 < Furai> :) 08:54 < Furai> I really start to think it's some problem with init script. 08:54 < storge> this is ubuntu? 08:54 < Furai> But it's fresh system installation and just apt installed vsftpd 08:54 < Furai> ubuntu 16.04 08:54 < Furai> ec2 instance 08:55 < Furai> Maybe amazon is doing something weird. 08:55 < storge> hmm, if it was install from apt it should be putting init scripts that match LSB like debian wants 08:55 < storge> but who knows 08:55 < Furai> According to client they have other exactly same setup and it works there. 08:56 < Furai> Maaaybe I'll go through bash history. 08:56 < storge> i know only as much about ubuntu as i can from running it's very first release and then only debian since then 08:56 < Furai> Who knows what they did there. 08:56 < Dagmar> Are you sure you're actually logging the facility vsftpd is writing to? 08:57 < ayecee> how to check? 08:57 < Dagmar> Looking at syslog's configuration 08:57 < Dagmar> Easy mode: STart logging *.* to something 08:57 < jim> Furai, see if you can get proftpd working on it, or some other ftpd, and tell the customer the good and bad news 08:58 < Dagmar> I usually have that pointed at a fifo on anything I'm actively still working on 08:58 < ayecee> Dagmar: i like it! 08:58 < ayecee> what is that, /etc/rsyslog.conf ? 08:58 < Dagmar> It's not entirely safe on non-Linux hosts, and it used to not be safe on Linux hosts, but it gets the job done 08:59 < ayecee> not safe how? 08:59 < ayecee> logging loops? 09:00 < Furai> chmod -Rf 777 orders/ uploads/ 09:00 < Furai> gotta love finding stuff like taht 09:00 < ayecee> shotgun fix 09:02 < storge> shotgun fix as leverage to argue for a better solution 09:03 < ayecee> problem is no one patches the wall behind afterwards 09:04 < ayecee> also the other non-target entities 09:08 < Dagmar> No, if the syslog daemon isn't aware of what a fifo is, it may well *block* when it writes if there's no listener 09:08 < ayecee> i see 09:09 < Dagmar> Hence the curious requirement in /etc/syslog.conf of preceding the filename with the actual pipe character 09:09 < Dagmar> Because it's SUPER FUN to be locked out of your own machine 09:09 < ayecee> my sarcasm meter just exploded 09:09 < ayecee> i hope you're happy 09:10 < Dagmar> When one has failed at this, the symptom is generally that you try to login, and if more than about three minutes has passed since the machine booted up, _you can't_ 09:10 < Dagmar> It'll try to log that you logged in, syslog will block and not return, and your console will just sit there. 09:10 < ayecee> the voice of experience 09:10 < Dagmar> Yes. 09:10 < ayecee> so it's better now? 09:10 < Dagmar> It's maddening to figure out, too 09:11 < storge> sounds SUPER FUN 09:11 < ayecee> goddammit 09:11 < Dagmar> Now syslogd can be told explicitly about the pipe so it will handle things appropriately if the fifo is in fact full 09:11 < ayecee> there goes another 09:11 < storge> i had no idea, thanks for explaining that 09:11 < Dagmar> No problem 09:17 < storge> Note that the fifo must be created with the mkfifo(1) command 09:17 < storge> before rsyslogd(8) is started. 09:17 * storge sees bubbles rise as he starts in over his head 09:21 < sauvin> Eh, mkfifo ain't hard. 10:03 < interrobangd> hello, ... i've set my gpu freq. with udev and now i want to reset to default. i removed that rule, but frequency are not changed back to default value 10:03 < interrobangd> and i i've rebooted 10:04 < ayecee> do a cold boot. shut down the pc, then start it again. 10:08 < interrobangd> power off and on again will change that? i dont belive that 10:08 < interrobangd> its same like reboot 10:08 < ayecee> except it's not 10:09 < interrobangd> o_O 10:09 < ayecee> there's a reason that we have the terms "warm boot" and "cold boot" 10:09 * Armand gives ayecee a cold boot up the rump! 10:10 < Armand> Morning! 10:10 * ayecee yipes 10:10 * storge blinks 10:10 < Armand> ^_^ 10:10 < interrobangd> can i see when /sys/class/drm/card0/gt_max_freq_mhz ware written? 10:10 < ayecee> interrobangd: like, a timestamp? 10:10 < ayecee> no. 10:10 < interrobangd> maybe its done by a shell script 10:11 < interrobangd> *did 10:11 < ayecee> done was correct 10:11 < interrobangd> ok^^^ 10:11 < ayecee> would be it's, but that's minor 10:12 < ayecee> "maybe it is done by a shell script" 10:12 < interrobangd> i will try a COLD boot 10:13 * storge silences the false alarm of the Pedant-O-Meter 10:13 < interrobangd> by 10:13 < DevAntoine> Hi 10:14 < jim> hi 10:14 < DevAntoine> Do we agree that with a compress tarball I can't update/replace a file in it? I have to extract the whole .tar.gz, update my file and then recreate it right? 10:14 < mujjingun> hi 10:14 < ayecee> DevAntoine: yes 10:14 < jim> DevAntoine, maybe not directly 10:14 < ayecee> uh oh 10:15 < jim> hi 10:15 < ayecee> ambiguous answer detected 10:15 < ayecee> you will now be subject to his questioning 10:15 < jim> no, we know what he wants 10:15 < ayecee> do we? 10:16 < DevAntoine> jim: not directly? :,D 10:16 < interrobangd> back again, cold boot changed nothing 10:16 < cart_man> Hi all .. how do I run a cron job as root ? Also how do I test it if its weekly or hourly? 10:17 < DevAntoine> cart_man: sudo crontab -e 10:17 < ayecee> interrobangd: how did you set the frequency originally? 10:17 < interrobangd> some day i've set /sys/class/drm/card0/gt_max_freq_mhz to 600 using UDEV rule, then changed to another... then removed that rule an now i am again at 600 o_O 10:17 < jim> DevAntoine, you'd have to decompress it so that you can do stuff with tar, then compress the resulting tar and replace the old one 10:18 < DevAntoine> jim: yes, exactly what I said then ;) 10:18 < jim> that's probably not the only way 10:18 < ayecee> interrobangd: what was it before you set it? 10:18 < interrobangd> ayecee, writing to /sys/class/drm/card0/gt_max_freq_mhz 10:18 < DevAntoine> jim: well, after some search I found no other way to do it 10:18 < interrobangd> using root 10:19 < cart_man> DevAntoine -> No crontab for root : / 10:19 < ayecee> interrobangd: what was it before you set it? 10:19 < interrobangd> 1050 10:19 < jim> there;'s a file /etc/crontab 10:19 < cart_man> DevAntoine Ahh nbm 10:19 < interrobangd> ayecee, that are default 10:19 < ayecee> i see 10:21 < interrobangd> i dont now why its now at 600 instead of 1050 when i remove that rule- i remeber that i've set it someday to 600 10:21 < ayecee> i wonder if you could grep for 600 in files in /etc 10:21 < interrobangd> ..should i do that? 10:22 < ayecee> might as well 10:23 < interrobangd> sudo grep -r 600 /etc/* ? 10:24 < ayecee> sure 10:24 < cart_man> DevAntoine Is this a legitimate crontab entry for something that needs to run HOURLY? -> 0 * * * * node /truvelo/LXSyncManager/LXSync.js > ~/text.txt 10:24 < DevAntoine> cart_man: what? 10:26 < interrobangd> /etc/sysfs.conf:# devices/system/cpu/cpu0/cpufreq/scaling_setspeed = 600000 10:26 < cart_man> DevAntoine That Entry ... does it look like it will work? 10:26 < ayecee> the smoking gun! 10:26 < cart_man> That is a Crontab entry 10:26 < DevAntoine> cart_man: you have no clue what you're doing right? 10:26 < cart_man> DevAntoine Not with Crontabs no... my first one 10:27 < interrobangd> ayecee, but its not related to GPU 10:27 < ayecee> ah, still strange that it would show up 10:27 < DevAntoine> cart_man: looks good to me 10:27 < ayecee> anyhow, i'm out of ideas 10:28 < cart_man> DevAntoine To Run hourly right? 10:28 < DevAntoine> man... 10:28 < DevAntoine> http://crontab.guru 10:29 < cart_man> DevAntoine Whoaa thats awesome ! Thanks 10:31 < cart_man> DevAntoine Also the spaces between the starts does not matter right? 10:35 < gurrkiin> generally, is it faster to send an uncompressed archive and unpack or compress and then uncompress on the other side? 10:36 < Sveta> depends on the bandwidth 10:36 < gurrkiin> i suppose it depends on your compression rate 10:36 < ayecee> gurrkiin: generally, "it depends" 10:36 < Sveta> oh, sorry; depends on the bandwidth and on your compression rate :-) 10:36 < gurrkiin> so pretty case by case really? 10:37 < Sveta> yes, i think so 10:37 < ayecee> and on the compute time required for compression/decompression, and on the compute time of the background tasks running parallel. 10:37 < gurrkiin> and does anyone know how much of an overhead uploading via SSH gives? 10:38 < gurrkiin> assuming this is all on a modern connection 10:38 < ayecee> the encryption is normally negligible compared to other factors. 10:38 < gurrkiin> you mean encryp/decrypt time ? 10:38 < ayecee> yes 10:39 < gurrkiin> I would have thought that to be one of the longer tasks in the pipeline 10:39 < ayecee> you would be mistaken on modern hardware 10:39 < gurrkiin> archiving/unarchiving is pretty instant also, although I havent worked with *massive* archives 10:42 < HailTheRobots> Hello! I am looking for good free web hosting control panel for Ubuntu/Debian, which can control NGINX + Apache + PHP-FPM + MYSQL + Postgres, and also have backup and letsencrypt control function. Just for personal use for few personal sites, so I hope its lightweight. Does anyone knows good? 10:48 < cart_man> How can I test my crontab entry without waiting an entire hour? 10:49 < iflema> if you worried reload cron, if you think sytax is wrong look again 10:49 < iflema> it aint hard 10:49 < ayecee> well obviously it is 10:49 < ayecee> that's why there are websites like crontab.guru 10:50 < iflema> lol 10:50 < iflema> will the earth explode if it doesnt trigger 10:51 < ayecee> can't take that chance 10:51 < iflema> ask you mum 10:51 < jim> cart_man, add another crontab entry after it that emails you a hello world, and does it every 5 mins... does that one work? 10:51 < ayecee> that escalated quickly 10:51 * Armand shoves ayecee into an elevator 10:52 < KnightsOfNi> How come when I type $ ls -laht | head -5 I see only 3 files? 10:52 < sauvin> I use cron so rarely that when I DO have to set something up, I have to look at the man page, throw up a script to launch that logs, and have cron call it every couple of minutes or so until I'm happy it'll work, then I'll change the crontab entry to put in the appropriate execution times. My cron jobs *always* launch some kind of script. 10:52 < jim> KnightsOfNi, what happens when you just run ls? 10:52 < iflema> lool 10:53 < ayecee> KnightsOfNi: because you touch yourself at night 10:53 < KnightsOfNi> jim it lists all the files 10:53 < deusstultus> KnightsOfNi: use -A instead of -a then 10:53 < jim> KnightsOfNi, more than 5? 10:53 < iflema> it tells you right there when you edit 10:54 < jim> wait, are you sure it's not 5? 10:54 < KnightsOfNi> deusstultus, then I get 4 10:54 < jim> do you see . and .. as 2 of the 5? 10:54 < KnightsOfNi> only the . 10:55 < KnightsOfNi> but also a total size 10:55 < KnightsOfNi> as first line 10:55 < KnightsOfNi> I guess that counts as a line 10:55 < KnightsOfNi> I just want the files, not hte total size 10:55 < deusstultus> head doesn't care what content is, it prints n lines so ya 10:56 < jim> then you'd have to grep it out, or see if ls has an option to suppress it from the long listing 10:56 < KnightsOfNi> ok 10:57 < jim> KnightsOfNi, but then because you're feeding ls to stuff, you're effectively parsing ls, and that's something you should probably avoid 10:58 < jim> try this: find . -maxdepth 1 10:58 < KnightsOfNi> or I could just do head -6 11:03 < jim> KnightsOfNi, try this: find . -maxdepth 1 -ls 11:04 < KnightsOfNi> I did, what about it Jim? 11:05 < jim> that output is a little different from ls -l, and also doesn't have the total line 11:07 < KnightsOfNi> True 11:16 < interrobangd> ayecee, maybe you have a answer... after running "sudo update-initramfs -u" my actualy settings for /etc/class/drm/card0 are used for default 11:17 < interrobangd> so i cant setup a udev rule or change values and running initramfs to set as default - i dont no why o_O 11:20 < interrobangd> the question is, waht did update-initramfs 11:20 < interrobangd> *what 11:50 < snatcher> (tar --ignore-failed-read -czf backup.tar.gz -P -- /path/to/dir), why tar still exiting with 1 "file changed as we read it" in such a case? 11:51 < djph> something in /path/to/dir changed 11:51 < djph> it's not a "failed read" 11:53 < snatcher> what? failed read != file changed as we read 11:54 * epicmetal wonders how badly a musl-based system would break him 11:55 < dadabidet> hello, imagine I launch a compile command, so I have a lot of errors, is there some way to automatically scroll to the beginning of that output? tail ? 11:58 < djph> snatcher: IIRC, '--ignore-failed-read' is to deal with the inability to read off the media; not the OS / filesystem telling tar "oh hey, thisfile changed" 11:58 < BluesKaj> Hey folks 12:03 < momomo> i know maybe not related, but no better place to ask, anyone knows the simplest csv format for contacts that always works to import .... like old school SIM CARD format ? I only need a name and a number ... the formats on these new devices are killing me ... anyone has an example? 12:05 < djph> import into what? 12:11 < realbadhorse> does .config/autostart work on all mainstream distros out of the box? 12:11 < repys> I have a broken mountpoint in fstab and the VM is not starting anymore. could be that the reason? 12:12 < repys> I mean if the mountpoing is not correct in fstab, will linux start or not? 12:12 < realbadhorse> probably not 12:14 < repys> how can I get it starting if that is not correct? 12:14 < epicmetal> realbadhorse: it's more a question of your graphical session supporting that standard 12:14 < realbadhorse> oh.. so it doesent work on tty login? epicmetal 12:14 < epicmetal> realbadhorse: not by default 12:15 < epicmetal> realbadhorse: (no) 12:15 < realbadhorse> anything i could use that works for tty+graphical login ootb? 12:15 < epicmetal> realbadhorse: not that I know of 12:16 < realbadhorse> .profile? 12:16 < epicmetal> realbadhorse: that's for your shell' 12:16 < realbadhorse> hmm 12:21 < repys> how can I start linux with fstab not correct 12:21 < repys> ? 12:21 < Triffid_Hunter> repys: pass init=/bin/bash on your kernel commandline 12:21 < Triffid_Hunter> repys: then mount -o remount,rw / and fix your fstab 12:27 < fire2199> what command can i use to find the list of files and directories used by a given process? 12:27 < alexandre9099> hi, is it possible to create an user on linux that can be remotely acceded but can only do offline tasks (no internet access) 12:27 < alexandre9099> ? 12:27 < lopid> lsof 12:28 < Triffid_Hunter> fire2199: ls -l /proc/`pidof process`/fd perhaps 12:30 < lopid> lsof -p 12:30 < Triffid_Hunter> alexandre9099: you can block packets by uid with iptables 12:31 < Triffid_Hunter> alexandre9099: problem is you'll need an exception for ssh or so ;) 12:32 < alexandre9099> Triffid_Hunter, hmm, since ssh would run as system service cant i just drop everything from that user? 12:32 < alexandre9099> (also, can i limit what can be runned? do i need a different shell?) 12:33 < jim> alexandre9099, for example, what would you limit? 12:34 < alexandre9099> jim, only allow certain binaries to be executed, in this case i want to *rent* my gpu but i want to be sure that only the binaries that are going to use the gpu are allowed 12:35 < Triffid_Hunter> alexandre9099: ssh*d* is a system service, but the handler for each individual connection runs as the logged in user 12:35 < jim> hmm... 12:35 < Triffid_Hunter> alexandre9099: I suppose you could block new,unrelated packets using conntrack and allow established.. 12:36 < djph> sounds like you might want a chroot, with only those applications you want to allow people to run ... 12:36 < jim> I guess I'd start by putting the binaries they're allowed to run in a particular dir 12:36 < dzove855> alexandre9099: i would say look at restricted bash 12:36 < alexandre9099> oh how could i forget about chroots :) 12:37 < alexandre9099> about iptables, any recomended reading? 12:39 < fire2199> Triffid_Hunter : yeah that seems fine. It shows symlinks pointing to the what i was looking for 12:39 < fire2199> lopid : yeah lsof worked 12:40 < lopid> yeah 12:42 < mawk> alexandre9099: you can use cgroups, or plain matching by UID in iptables 12:42 < mawk> just read man iptables-extensions 12:42 < mawk> cgroups is maybe more robust 12:43 < mawk> for instance if the user can execute /bin/ping it will get execute as another user, and escape the UID matching 12:44 < mawk> the net_cls cgroup will assign an identifier to a process and all its descendants, and you can match on that in iptables 12:44 < Reventlov> Hi. 12:44 < mawk> but you can also use some lightweight container to isolate the network, and setup your host as a real router to decide what to let pass or not; it's probably way easier 12:44 < mawk> hi 12:45 < Reventlov> I want to check if and when a wireless driver function is called, is there a prefered way to "send" messages from inside the driver code? 12:45 < Reventlov> (I guess a simple printf won't do it) 12:46 < mawk> printk ? 12:46 < mawk> it will appear in dmesg 12:46 < jim> alexandre9099, you might want to look at the difference between iptables and nftables (note: iptables in its original form does not exist in the kernel anymore... the reason you can still write iptables rules, is there is a compatibility layer, so that when you write an iptables rule, it becomes an nftables rule 12:47 < alexandre9099> mawk, i might go with plain iptables and DROP *everyhing* from that user 12:47 < alexandre9099> jim, oh, nice, didn't knew about that change :) but is it like init.d / systemd or is it a kernel thing? 12:47 < Dagmar> ipchains/ipfw have been done away with 12:47 < Dagmar> iptables has always been the interface for netfilter 12:48 < mawk> alexandre9099: using a network namespace is even easier 12:49 < mawk> you put the user in the new network namespace, and...that's it 12:50 < alexandre9099> i'll check about namespaces 12:50 < mawk> you can try manually: sudo unshare -n 12:50 < mawk> then look at the network interfaces with ip addr 12:52 < repys> how can I boot my linux even if my mountpoint is not connected ? 12:52 < alexandre9099> mawk, hmm so unshare is *like* a chroot? 12:52 < repys> I mount a usb disk via fstab but I want to boot linux even if the disk is not connected 12:52 < mawk> not really alexandre9099 12:52 < mawk> unshare creates a new namespace, and starts a process in it 12:52 < mawk> if you combine unshare with a chroot, then you've got a little container 12:53 < Reventlov> mawk: Ok, thanks, will try. 12:53 < alexandre9099> mawk, so unshare isolates resources and chroot isolates the filesystem? 12:55 < mawk> yeah in a sense 12:56 < mawk> well chroot is a bad isolation by itself 12:56 < alexandre9099> can some user *break* a chroot? 12:56 < mawk> the filesystem isn't isolated, it's still the same one and you have tricks to get back to the old root; but with a little more work using chroot and unshare you have a real container like docker or lxc 12:57 < alexandre9099> nice :) 12:57 < interrobangd> why send "systemd-lodingd" boost request to my GPU????? my GPU ist always at max speed! 13:02 < winsoff> Is it possible for a desktop to look good, crisp, and simple without taking up very many resources? Perhaps it's a matter of font and color choice. 13:03 < iflema> i3 13:03 < winsoff> not awesomewm? D= 13:03 < iflema> 9 13:03 < iflema> I like twm 13:04 < iflema> once you set up its madness 13:04 < winsoff> Also, to clarify, if I install a window manager, it's only taking up space on the disk while other window managers run, right? Can I break something by installing too many? 13:05 < adrian_1908> winsoff: only taking up disk space unless it runs, yes. 13:05 < iflema> yeah... i doubt it 13:05 < winsoff> adrian_1908, I know that sounds like a stupid question, but I don't know how anything works with systemd. 13:05 < winsoff> Is there a "how linux works since systemd" manual I can RTF out of? 13:06 < iflema> systemd manual? 13:06 < ankk> hi. 13:06 < iflema> o/ 13:07 < iflema> i3 with compton is pretty crisp 13:07 < adrian_1908> I barely ever touch systemd, but I don't see how it's inherently difficult to grasp or worthy of attention unless you're a server admin. 13:07 < winsoff> iflema, does it explain how a linux distro that uses systemd will work? I installed awesomewm, and now none of my typical MATE startup items start (I wonder why!). With that in mind, does MATE's wm need to be running for me to properly run other applications? 13:07 < cart_man> Is there a way I can get the eth0s Mac Address via CMD? 13:08 < iflema> winsoff: distro that uses systemd is doing it wrong... but hey too late now 13:08 < adrian_1908> winsoff: using a very hand-on distro? generally, the packages should take care of that. 13:08 < djph> cart_man: ip a eth0 (IIRC) 13:08 < cart_man> cause with Ubuntu the eth0 has a new name and it is the MacAddress 13:08 < cart_man> djph ? 13:08 < iflema> winsoff: no idea 13:09 < adrian_1908> winsoff: or is it, because you want to mix stuff that's not packaged together (MATE and awesomewm)? 13:09 < djph> cart_man: oops, it's "ip a show eth0" 13:09 < winsoff> adrian_1908, i primarily just want to run nm-applet and get it docked somewhere, lol 13:09 < epicmetal> winsoff: have you configured awesomewm to start ~/.config/autostart things 13:09 < winsoff> i haven't 13:09 < cart_man> djph Well I need to be able to use it in /etc/network/interfaces 13:09 < winsoff> i didn't know that was a thing. interesting. 13:10 < epicmetal> winsoff: just use the native awesomewm way of starting things, which I don't remember what it is... probably some lua config 13:10 < adrian_1908> winsoff: can't help you regarding the wm itself, but with systemd you'd basically create a "unit" file and make it depend on a "target" after which it should run. 13:10 < epicmetal> i'd say his problem has nothing to do with systemd 13:10 < djph> cart_man: what? 13:10 < iflema> maybe huge things that do every thing will be ok in a world with stoopids and AI 13:11 < iflema> who can tell 13:11 < djph> cart_man: what are you trying to do? 13:11 < Mavericks> what's a good link to download man pages ? 13:12 < Mavericks> chm / pdf whatever format ? 13:12 < epicmetal> Mavericks: why not just install/use manpages natively 13:12 < azarus> Mavericks: man -Tpdf 13:12 < jim> alexandre9099, this is a change in the kernel, there are no iptables modules, only nftables modules, and a compatibility layer 13:12 < cart_man> djph I have Ubuntu Mate on a raspberry pi. I want to move the SD card to another Raspberry PI BUT upon doing that the network settings change since the ETH0 interface gets a new name everytime 13:12 < djph> cart_man: oh, fix udev to just name it something sane 13:13 < cart_man> djph How ? 13:13 < Mavericks> i'm on windows. i'd like to go through it and have it as a reference doc. i wish it was there on devdocs.io but they don't 13:13 < Mavericks> epicmetal: 13:13 < Mavericks> azarus: o 13:13 < Mavericks> * ok 13:14 < adrian_1908> djph: you can revert to the old naming scheme if that's what you're after. just look it up online, I think it's a simple toggle. 13:14 < adrian_1908> err cart_man I mean. 13:14 < cart_man> adrian_1908 I have tried it before... it was not that easy 13:15 < adrian_1908> Hmm, I see. 13:16 < djph> cart_man: IIRC, add/edit the "/etc/udev/rules.d/??-persistent-net.rules file 13:16 < djph> (where "??" is a number) 13:17 < djph> cart_man: put two entries in there, call rpi(1) eth0, and rpi(2) eth1, then set /etc/network/interfaces properly for each host. 13:17 < djph> or rather, the MAC address of the ethernet interface 13:18 < repys> what are used for the files in /etc/udev/rules.d? 13:18 < djph> telling udev how to do things 13:18 < djph> cart_man: syntax is SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:20:19:52:d3:c0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth1" 13:18 < Mavericks> epicmetal: azarus: I saw https://www.linux.org/docs/ but it can be better 13:18 < azarus> Mavericks: linux.die.net ? 13:20 < Mavericks> azarus: i'm desperate to visit that url but is there an euphemistically pleasing name that does not have the words linux and die close to the other ? unless DIE is an acronym that means something else 13:20 < sie> How do applications interface with pulse or alsa? Through which apis? 13:21 < azarus> Mavericks: a "die" is something you use to play jahtzee 13:21 < azarus> "to die" != "die" 13:21 < azarus> one's a verb, one can be a noun 13:21 < Mavericks> azarus: ok! great! phew 13:23 < rumpel> sometimes it's even an article 13:25 < Reventlov> https://unix.stackexchange.com/questions/448162/iwlwifi-dvm-debugfs-and-fixing-rate If someone has an idea 13:28 < Mavericks> azarus: any other site besides that - plan to bookmark it 13:28 < azarus> Mavericks: any linux system near you(TM) 13:28 < Mavericks> azarus: heh. got it :) 13:31 < likcoras> Is there any straightforward way to pass on any SIGTERMs as SIGINT to children? An application shuts down gracefully on sigint but hangs on sigterm, and I can't really patch it. 13:41 < Mavericks> azarus: is there a bot in this channel where I can do or and get the info. just for myself ? or a channel like that ? 13:41 < azarus> Mavericks: dunno 13:41 < azarus> Hey everyone, are the Maas? (Man pages As A Service) :P 13:41 < Dagmar> likcoras: Have the applications signal handler simply absorb SIGTERM and when it sees that signal send the SIGINT to the children itself 13:42 < Mavericks> azarus: heh 13:42 < Mavericks> lol 13:42 < rebelmoon> Does anyone know a better way to compile a kernel for a certain architecture than editing the Makefile to add -march in every instance of any type of CFLAGS or CXXFLAGS. It's a real pain in the ass, but exporting those variables in the shell does nothing, make overrides them with -march=x86_64 13:42 < likcoras> Dagmar: ah, makes sensee. 13:43 < Dagmar> rebelmoon: Perhaps you should familiarize yourself with the kernel build docs. What you describe should simply not be necessary 13:43 < rebelmoon> Dagmar: Oh, OK, thanks. 13:50 < ALowther> I just repartitioned a HD to make it smaller, it says I *may* need to update /etc/fstab. How would I update it & how do I know if I need to update it? 13:51 < cousin_luigi> Greetings. 13:51 < cousin_luigi> Can you recommend a Qt front-end for tesseract? 13:51 < Dagmar> You don't need to change your fstab unless you renumbered or re-ordered your partitions 13:52 < Dagmar> i.e., you have to _move_ something, not just make it a bit bigger 13:52 < ALowther> Dagmar: Ah okay. Thank you. 13:53 < HaZrD> hi all 13:53 < HaZrD> i see the #slackware channel has gone away - anyone here know how to get hold of them on irc these days? 13:53 < Ridout> ##slackware 13:55 < pyro_> hi i created echo "hello world" > test1; ln -s test1 test2 13:55 < rocketmagnet> hi everyone, can someone please tell me how i can set permanent kernel parameters ? (without grub) 13:55 < pyro_> if i do ls -l 13:55 < pyro_> the size slightly differs from two files 13:55 < pyro_> why is that 13:58 < HaZrD> Ridout much appreciated, thank you! 13:58 < Ridout> You're welcome. 13:59 < alexandre9099> can i lock a console session without losing what i am doing? 14:00 < heftig> alexandre9099: vlock 14:01 < maboc> @alexandre9099, tmux ???? 14:01 < maboc> @alexandre9099 it depends on what you are actually trying to accomplish. 14:15 < rocketmagnet> re: kernel parameters ?? (debian) 14:15 < alipoor90> Hi, in nftables, how can i use wildcards in a list? i tried this and it gives me error when loading it: https://paste.fedoraproject.org/paste/c9iaM5QwD~GkDG8vhADoOg ,moving each interface to a separate line fixes the problem, but i wanted to know if it is possible to use wildcards in a list as well or not 14:17 < jim> rocketmagnet, as you know, one way is to change the "kernel command line", which is what you present to the kernel at the time you start/boot it 14:17 < jim> I -guess- another way would be to compile them in? 14:18 < repys> better nfs or SMB for file sharing across vpn ? 14:20 < jim> repys, I also like sshfs 14:22 < repys> sshfs is ok for large files? 14:23 < heftig> alipoor90: no, I don't think that's possible; make multiple rules 14:24 < msdsos> hello good people :) 14:24 < msdsos> i was wondering if anybody else have tested the Red Star OS? i got it to work and changed the lang. but was wondering if anybody else have tested it? :) 14:26 < Pentode> why would anyone want to? lol 14:27 < compdoc> doesnt sound interesting to me 14:27 < Pentode> im not real big on being held back and spied on. 14:28 < lunaphyte> why can't an unprivileged user see their own info with lsof? : http://dpaste.com/1CQYGMP.txt 14:29 < blackgatonegro> Hi, I already know the file extension/file type a bunch of "*.chk" files should have, how I recover them using linux terminal? * The type is video and webm 14:33 < alexandre9099> maboc, leaving my pc alone (for example with root account logged in[i know, not the best example]) 14:36 < section1> lunaphyte, crazy...ssh-agent process makes weird symbolic links 14:37 < lunaphyte> section1: in /proc/? 14:37 < section1> yes, i think its because of suid binarie 14:38 < section1> binary* 14:45 < deego> On debian stable, any invocation of webmagick, including --help yields this: Can't use 'defined(%hash)' (Maybe you should just omit the defined()?) at /usr/bin/webmagick line 1329. 14:45 < deego> Any help appreciated 14:46 < maboc> alexandre9099: try running in the background, or an at-job, or indeed something like tmux. 14:46 < alexandre9099> maboc, hmm so then i could just exit tmux and then logout? 14:47 < maboc> You can detach from tmux. Later on you can re-attach to tmux. It's worth a try :-) 14:50 < alexandre9099> nice, thanks, i have been using screen on ssh sessions, i'll give a look at tux 14:50 < alexandre9099> tmux 14:50 < maboc> alexandre9099: :-) GO 14:59 < repys> can you share ext4 partition over samba? 15:00 < Tadgy> repys: Samba is a file sharing protocol; the underlying filesystem doesn't matter. 15:01 < GunqqerFriithian> What disk cloner software do you guys reccomend? I'm getting an SSD and will need to move my HDD to it 15:01 < redredhathat> dd? 15:02 < GunqqerFriithian> will that work with the file system mounted and in use? 15:02 < redredhathat> I don't think so 15:02 < redredhathat> but you could live USB 15:02 < GunqqerFriithian> I could 15:03 < GunqqerFriithian> Well that's certainly an option 15:16 < strixdio> anyone know if it's possible to make qemu-kvm/spice default to a specific resolution? All my VMs currently default to a 4:3 ratio, I'd like to make it 16:9 if possible. 15:16 < compdoc> how do you connect to them? 15:16 < compdoc> from the console? 15:17 < strixdio> graphical console. 15:19 < compdoc> I connect remotely and the client sets the rez 15:20 < drz3k> Something happened and now my program doesn't work :( 15:21 < strixdio> compdoc: connect remotely, meaning? 15:22 < drz3k> I was working on a c++ program that uses the IO ports on a linux system, had to call iopl(3), but in order to do so I either 1) had to be root 2) had to set some capabilities. I was struggling to install new version of my IDE today and even though the capabilities are set, it returns with error 15:25 < Pentode> what's the error? 15:37 < drz3k> iopl return -1, errno is set to 1 - EPERM 15:37 < drz3k> the compiled binary can be started normaly (as non-root) 15:37 < drz3k> the /bin/bash and gdb also have capabilities 15:41 < Pentode> hmm. eperm essentially means "you can't do that unless you have special privileges." 15:43 < drz3k> Pentode, yes 15:43 < Pentode> sorry i'm not much help ;) 15:48 < RustyJ> pentode? eperm in linux? 15:48 < Pentode> RustyJ, gcc 15:48 < RustyJ> ahhhh 15:49 < RustyJ> been up all night coding crap and setting stuff up... i'm about as sharp as a bowling ball 15:49 < Pentode> make stronger coffee 15:50 < RustyJ> blah... i just have to remove some stuff and call it secure and i'm going sleepy time 15:53 < Leishan> Anyone with Oracle VM 3.3 experience? tried my luck at #oracle to no avail. I'm trying to find how to export/backup a VM in 3.3, but there's no export option (in 3.4 we have export VM to virtual appliance), any help? thanks :) 15:56 < lupine> > oracle 15:56 < drz3k> funny thing, I installed back the older version of my IDE (codeblocks, had sources for different versions), it starts normally. Guess I'll have to tell the creators about possible bug 16:10 < hailhydra> I'm trying to install Ubuntu 18.04 server on a dell poweredge. I'm stuck at (initfamfs) can anyone help me? 16:10 < GunqqerFriithian> can you give more info? 16:17 < hailhydra> GunqqerFriithian: yes what would you like to know 16:18 < GunqqerFriithian> just more info on the sitation, what messages you're getting, etc 16:19 < hailhydra> GunqqerFriithian: I just get the (initramfs) prompt with a list of commands if I type help 16:19 < GunqqerFriithian> so have you installed it and are trying to boot up now? 16:20 < hailhydra> GunqqerFriithian: I originally installed the OS when it was connected to a PCIe riser. But then I moved it over to the RAID array in slot 2. Slot 0 and 1 have HDDs 16:20 < GunqqerFriithian> so the OS is installed on the RAID array? 16:21 < hailhydra> GunqqerFriithian: yes I am trying to boot. But I think since I moved the drive after the OS was installed it is having errors 16:22 < GunqqerFriithian> if you move it back does it work? 16:22 < hailhydra> GunqqerFriithian: I don't know whats on drives 0 and 1 but 2 is the SSD with Ubuntu 18 server 16:22 < GunqqerFriithian> Hmm 16:22 < hailhydra> I've never got it to boot off the PCI cause I can't make a VD in Dells BIOS 16:23 < hailhydra> All the drivers are old and there is no LifeCycles installed. 16:23 < GunqqerFriithian> can you get the system to boot from a live USB? 16:23 < GunqqerFriithian> if you can you can check the disks 16:24 < hailhydra> GunqqerFriithian: nope no live USB. Error for this PowerEdge model. Cost me 3 days. 16:24 < hailhydra> I have a server DVD 16:25 < GunqqerFriithian> so no error messages? 16:25 < fire2199> what tool other than chkdsk can be used to repair a fat32 filesystem? I tried photorec and i recovered some data but i know this partition is repairable, either dosfsck or chkdsk in windows are truncating the beginning of some directories area and consequently the data is lost. 16:25 < hailhydra> cat /proc/modules lists a ton of stuff 16:27 < BCMM> fire2199: how do you "know" it's repairable? 16:33 < fire2199> BCMM : actually everything is repairable, it all relies on some complexity. I can mount the partition but when doing ls, directory contents are not displayed and "Input/output error" is displayed instead. 16:39 < plexigras> what is the fastest way to check if a word is in a wordlist in a shellscript? 16:39 < revel> grep 16:39 < storge> grep -i if you want to ignore care 16:39 < storge> grep -i if you want to ignore case 16:40 < plexigras> i want to do it repedatly so i would think grep would not be the best 16:40 < lopid> rerecord not fade away 16:40 < storge> you mean another script that's checking repeatedly? 16:41 < lopid> grep -F wordlist shellscript 16:41 < lopid> -f, rather 16:43 < uplime> plexigras: what are you trying to do? 16:43 < plexigras> i want to implement a simple spell checker 16:44 < kerframil> plexigras: fastest? use >=bash-4.0 and make the list an associative array 16:44 < uplime> well if you want speed you shouldn't be considering shell code 16:44 < uplime> otherwise, grep or an associative array should be perfectly fine 16:44 < uplime> although an associative array might eat up more memory 16:45 < hailhydra> simple spell checker lol. 16:45 < kerframil> yeah. that's the trade off. 16:45 < uplime> right 16:46 < prussian> grep -F word$'\n'list$'\n'etc... file should be fast enough 16:46 < prussian> or -qF if you just want to know 16:47 < plexigras> if i understand this right grep is reading line by line which cant be fast 16:48 < kerframil> GNU grep is faster than you think 16:48 < revel> plexigras: It's actually highly optimized. 16:48 < prussian> at the end of the day you have to scan whole strings no matter what. 16:49 < lopid> it is faster than you can think 16:49 < kerframil> plexigras: https://lists.freebsd.org/pipermail/freebsd-current/2010-August/019310.html 16:50 < kerframil> even if it weren't so well optimised, it wouldn't be true where using either -q or -m1 16:51 < repys> is the user used in samba mapped to the user in linux? 16:51 < BCMM> plexigras: hailhydra has a point. natural language is one of those things where you start by ironing out a few obvious edge-cases and then two years later you realise it's obvious edge-cases all the way down 16:51 < kerframil> repys: yes 16:54 < hailhydra> your essentially greping against a dictionary in real time. 16:55 < repys> how can I enable encryption with samba? is there any easy doc? 16:56 < kerframil> repys: are your clients all using >=SMB2? 16:56 < repys> I use mac os and windows 10 and linux with samba4.5 16:57 < kerframil> repys: add "server_min_protocol = SMB3_00" and "smb_encrypt = required" to your global section. 16:58 < kerframil> repys: restart, re-establish sessions, check with smbstatus 16:58 < uplime> what does samba use for encryption? just tls? 16:59 < kerframil> uplime: AES-128-CCM 16:59 < uplime> ah 16:59 < uplime> neat 16:59 < kerframil> uplime: that's SMB3 only. packet signing and sealing has been required since SMB2 also, I believe. 16:59 < repys> nothing has changed 17:00 < repys> I added "server_min_protocol = SMB3_00" and "smb_encrypt = required" in the global section and restarted smbd 17:00 < repys> after that I re established the connection and checked with smbstatus 17:00 < repys> I see - in the encryption column output from smbstatus 17:02 < repys> unkwnon parameters :) 17:03 < aaa__> . 17:03 < kerframil> I'm not following 17:03 < aaa__> hi 17:03 < kerframil> repys: you should see AES-128-* in both the encryption and signing columns 17:03 < aaa__> with ecryptfs, encryption key are in memory ? if someone make a dump of the memory i can recover the encryption key? 17:04 < aaa__> he* 17:07 < kerframil> aaa__: yes 17:07 < aaa__> ok ty 17:07 < kerframil> aaa__: which can include swap 17:09 < aaa__> ok 17:09 < oiaohm> aaa__: its why hybernate to disc unencrypted on a encrypted system is kind of epic fail. 17:10 < aaa__> lol 17:10 < aaa__> ok 17:10 < oiaohm> aaa__: sounds like lol until you find out its been found a few thousand times in security audits. 17:10 < oiaohm> aaa__: on laptops. 17:11 < repys> maybe I need some extra packages for encryption 17:11 < repys> ? 17:11 < repys> if I enable smb encrypt = required . samba doesn't share anything 17:14 < kerframil> repys: nothing extra is required. define "doesn't share anything". 17:14 < uplime> /b 6 17:14 < oiaohm> repys: you would most likely need to ask that in the #samba channel. 17:15 < repys> smbd start but if I can't connect to it anymore 17:15 < repys> it doesn't ask me for credentials at all 17:16 < oiaohm> repys: https://www.samba.org/samba/docs/current/man-html/smbclient.1.html you are using the -e flag for encrypted. 17:16 < kazdax> hello users 17:17 < kerframil> repys: connect to it with what? 17:17 < repys> ubuntu gui 17:17 < oiaohm> repys: I don't think the ubuntu gui provides the encrypted flag option. 17:18 < kerframil> repys: you only mentioned macos and windows 10 earlier 17:19 < jeffspeff> how can i append a string to the output of echo $HOSTNAME | tr "[:lower:]" "[:upper:]" ? I am trying to get result like HOSTNAME-FOOBAR 17:19 < kerframil> repys: even so, the gvfs smb implementation should support smb3. at least, it does here (gnome-3.28 on arch). 17:20 < oiaohm> kerframil: encrypted requires setting extra flag. https://blogs.msdn.microsoft.com/openspecification/2015/09/09/smb-3-1-1-encryption-in-windows-10/ << even under windows. 17:20 < kerframil> repys: if your system doesn't, you can try dropping smb min protocol to, say, smb2 - but you would then have to not force mandatory encryption. 17:20 < revel> jeffspeff: `echo string | sed 's/$/-something/'` works. 17:20 < oiaohm> kerframil: mistake samba support encryption on smb2 using unix extentions of course this means no windows access at all. 17:20 < kerframil> oiaohm: I use smb3 with crypto exclusively, and absolutely no modificaiton is required in windows 10, macos High Sierra nor gnome-3.28 on my arch box. 17:21 < kerframil> oiaohm: it justworks(tm) 17:21 < jeffspeff> revel, thanks! that worked perfect 17:22 < kerframil> oiaohm: in case you missed the earlier discussion, that's with smb3 as the min protocol and encryption mandates, both in the global section of the server's smb.conf. 17:22 < kerframil> mandated* 17:22 < oiaohm> kerframil: https://www.samba.org/samba/docs/current/man-html/smbclient.1.html read smbclient 17:23 < kerframil> oiaohm: I don't need to read about smbclient because I don't use smbclient 17:23 < oiaohm> kerframil: POSIX extensions via GSSAPI. << This is how samba does it on smb2/smb1 17:24 < kerframil> I use win10, high sierra and gvfs-smb-1.36.2. they work fine. there's not much more to be said. 17:24 < oiaohm> kerframil: of course windows clients completely fail if they see that. 17:26 < oiaohm> kerframil: running on Linux/BSD/OS X systems you have option of encrypted SMB on 1 and 2 as well as 3 but if you want cross platform and avoid some strange broken I would say smb3. 17:27 < repys> is there any app like filezilla to connect to samba with encryption? 17:27 < kerframil> oiaohm: win10 (1703) negotiates smb3 with crypto extensions fine here, no matter whether unix extensions is enabled or not. and, yes, that's why I had repys set min_protocol to smb3. he _only_ mentioned the use of win10 and macos initially, so it should be fine. 17:29 < Exagone313> Hi, do you know how could I list files (recusively) that have only one hard link? I haven't found an argument for that in GNU find, and if I have to use ls piped to grep it's not really awesome (in this case I'd write a small C program because I like to complicate things). Thanks for your help! 17:31 < kerframil> repys: all I can tell you is that the current version of gvfs-smb works. a possible alternative is to use the cifs client build into the kernel (mount -t cifs -o vers=3 ...). 17:32 < kerframil> repys: see the mount.cifs(8) man page for more details 17:35 < repys> if I use samba in an encrypted vpn, is not necessary to enable encrytion in samba as well? 17:35 < oiaohm> repys: https://help.ubuntu.com/community/Samba/SambaClientGuide normally I end up using smbclient as diagnostics on the server for diagnostics. Because I have had work on the server and not anywhere else and it turn out to be a firewall rule blocking a port. 17:35 < kerframil> repys: I believe you need at least a 4.11 kernel for cifs to work 17:35 < kerframil> repys: depends on your attack model 17:36 < repys> what do you mean? 17:36 < kerframil> repys: are you worried about wire sniffing? using wi-fi? those would be compelling reasons. 17:37 < repys> wire sniffing yes 17:37 < repys> or wifi also 17:38 < djph> repys: always encrypt everything 17:38 < djph> repys: also, don't use samba 17:40 < Reventlov> If you have any idea: https://unix.stackexchange.com/questions/448162/iwlwifi-dvm-debugfs-and-fixing-rate 17:40 < annihilator> Is lvmcache good or stick with bcache? 17:41 < repys> why no samba? 17:44 < Siecje> What is the difference between Ctrl+D and enter+~+. to disconnect from an SSH session? 17:44 < bls> Siecje: Ctrl+D generates and End-Of-File, which causes your shell to exit 17:45 < Siecje> bls: So is the latter a client side disconnect? 17:45 < bls> Siecje: ~. is handled by ssh to issue a disconnect from the local end instead of the remote 17:46 < Siecje> I'm asking becausee I ran into an issue with a program using 100% CPU. http://forum.tinycorelinux.net/index.php/topic,21346.msg133429.html 17:49 < bls> if the program can't handle being disconnected from, either initiate your disconnections remotely or fix the program 17:51 < Siecje> How does the program know it was disconnected? 17:53 < bls> it should receive a SIGHUP 17:57 < Siecje> So the client disconnects, and then eventually something (what is it?) sends SIGHUP? And htop is not being killed with a SIGHUP signal? 17:59 < Elladan> The client is connected to the ssh server via a TCP socket. When the client disconnects, it will send a TCP close message, which may or may not be delivered. 17:59 < bls> and the SIGHUP is sent when the pseudo-terminal is closed 18:00 < Elladan> If it's delivered, the ssh server process for that socket will run through a shutdown sequence, which results in the SIGHUP (not sure if it sends it directly or the shell does). 18:00 < Elladan> Or the pseudo-terminal 18:00 < Elladan> If it's not delivered, then eventually the socket or ssh will hit some sort of timeout condition and shut down through a different path, possibly hours later. 18:04 < mnovelo> mnovelo 18:06 < Siecje> bls: Elladan Thanks, I understand it now. 18:06 < Siecje> However I can't reproduce the issue anymore. 18:06 < za1b1tsu> Anybody with a MacBook Pro (Retina, 15-inch, Mid 2015) managed to make their webcam work? 18:07 < uplime> za1b1tsu: what os is your pro running? 18:08 < za1b1tsu> uplime: I tried ubuntu gnome live, everything worked perfectly except webcam 18:08 < za1b1tsu> want to make sure everything works before the big ascend 18:08 < Elladan> Siecje, ignoring SIGHUP and other signals, as well as going into infinite loops, is a common bug. 18:09 < alexandre9099> anyone using lshell? how safe it is? 18:10 < Siecje> Would the SSH service send SIGHUP to the shell or each sub process? 18:12 < lpyr> i have set the sticky bit using chmod +t dirname; touch dirname/hello1; sudo touch dirname/hello2 18:12 < lpyr> now if i do rm -f hello1; rm -f hello2; both the files are getting deleted 18:15 < kerframil> lpyr: if the user that ran `rm -f hello2` effetively has +w on the directory, that's perfectly normal - as surprising as it may seem. 18:15 < kerframil> the containing directory, I should say 18:16 < alexandre9099> but isn't sticky bit like "you can read/write every file but you can only delete your own"? 18:17 < lpyr> kerframil: even in tmp directory we have drwxrwxrwt. 7 root root 103 Jun 6 16:07 /tmp 18:17 < kerframil> alexandre9099: pay particular attention to the "has +w" part. 18:17 < kerframil> alexandre9099: look at /tmp as a counter-example. it should be owned by root:root, which doesn't fulfil that requirement. 18:17 < lpyr> however the file was not able to delete the hello2 which was owned by root 18:18 < kerframil> ^ 18:19 < lpyr> kerframil: if the dir is owned by root, then other root accounts no one can create a file unless the permissions are given 18:20 < kerframil> lpyr: can you pastebin the output of `id; stat dirname;` 18:21 < kerframil> where dirname is the dir that you're currently testing 18:24 < lpyr> kerframil: https://paste.linux.community/view/71116865 18:25 < kerframil> lpyr: ok, yeah. you simultaneously own the directory and have +w rights. you will be able to delete its entries, even if they are owned by another account. +t won't prevent that from being possible. 18:25 < mattfly> rpd software recommendations? 18:26 < mattfly> im using remmina but it uses 100% of my cpu 18:26 < mattfly> after a while 18:26 < kerframil> lpyr: in short, you'll get the behaviour you expect if you chown root:root test4 && chmod 1777 test4 (like /tmp) 18:26 < kerframil> lpyr: the key thing there is to change the owner to root 18:31 < kerframil> lpyr: you can think of it this way: directories contain a list of entries (names and inode numbers), and +w on a dir allows you to modify that list 18:31 < kerframil> lpyr: many don't realise but you only need +w on a dir to be able to delete its immediate entries 18:32 < kerframil> thus, the combination of the mode with the associated user:group is very important 18:40 < yakiza> guys wtf 18:40 < GunqqerFriithian> ? 18:40 < yakiza> my /OPT directory got deleted when extracted the nessus agent on / 18:42 < Trel> I'm trying to figure out how to format a grep command I want to find files that have lines that match multiple patterns (can be different lines, as long as all patterns match at least once in the file) 18:43 < Trel> If grep can't do that, I'm open to other options btw 18:44 < bls> that will likely require awk, perl, or python 18:44 < birbie> Greetings everyone 18:44 < matkatmusic> is it ok to ask apache-related questions here? 18:44 < bls> matkatmusic: you can try, although #httpd might be better 18:44 < matkatmusic> thanks 18:44 < matkatmusic> join #httpd 18:45 < kerframil> Trel: /msg greybot multigrep <-- that's one way 18:45 < kerframil> Trel: oops, I mean !multigrep 18:48 < kryptynasium> I am creating an RPM package for RHEL 7 on RHEL 6. Except for a couple of soft link changes in my RPM spec file - should there be any other changes I need to be aware of? I tested. RPM installs fine. 18:48 < kryptynasium> Or ideally an RPM for RHEL 7 should only be created in RHEL 7 ?! 18:49 < Dagmar> If you're having to ask this, you probably shouldn't be doing it. 18:50 < Dagmar> If, for example, what's going into the RPM happens to link against any shared libraries, the program is *not* going to behave properly when installed on a different version of RHEL 18:50 < bls> yeah, not sure why you wouldn't just package it on the same release it's intended for 18:51 < Dagmar> A lack of the necessary qualifications 18:52 < kryptynasium> Dagmar: nice! True - client describes an issue which I don't see. So yeah this sounds like a bad idea. The whole process of creating a dev environment on RHEL 7 is making me wince, But if it is what it is. I will roll up my sleeves! 18:53 < Dagmar> I didn't need to know this was for a paying client 18:53 < Dagmar> Why I don't just blow up the earth is beyond me 18:54 < kerframil> a lack of the necessary qualifications? (joke) 18:55 < kryptynasium> My college prof used to say you cannot be a programmer if you are not (good) lazy. Otherwise you won't devise shortcuts! Or find a better/faster way to do it. I am not talking of a kludge 18:57 < hexnewbie> kryptynasium: Is a common platitude. Does lazy mean writing software that does the work for you? Does it mean not overengineering it? Does it mean underengineering it? Does it mean writing everything as a one-line hack that only works on one computer in the whole world? 18:57 < Trel> kerframil: I'll give that a look 18:57 < Trel> thanks 18:58 < ntd> https://blog.cloudflare.com/welcome-hidden-resolver/ 18:58 < ntd> i mean, great, but trusting cloudflare after their many years of anti-tor shenanigans? 18:58 < Dagmar> kerframil: Some people would expect that perhaps a person should know how compiled software works, and then none of this would be even remotely hard 18:59 < uplime> /b 5 18:59 < uplime> ffs 18:59 < uplime> i promise im not doing that on purpose 18:59 < GunqqerFriithian> $10 he's doing it on purpose 19:00 < SuperSeriousCat> But he promised 19:01 < uplime> i am pretty attention-seeking 19:01 < uplime> so maybe 19:01 < SuperSeriousCat> What drink do you recommend, uplime? 19:02 < uplime> oh 19:02 < uplime> whiskey and coke 19:02 < SuperSeriousCat> A "yay, weekend tomorrow" drink to get me sleepy and tipsy 19:02 < uplime> for sure 19:02 < SuperSeriousCat> Not a fan of whiskey 19:03 < uplime> a bar i go to makes a drink called gold rush tahts pretty good 19:03 < uplime> oh but its got whiskey as well 19:03 < realbadhorse> how do i find out which line triggered the except? 19:03 < uplime> i used to not like whiskey/tequila/vodka until i started doing shots at work 19:03 < SuperSeriousCat> I think I just take a couple of beers 19:03 < uplime> beers are nice 19:03 < realbadhorse> ooh nvm.. wrong channel 19:03 < uplime> shinerbock is good 19:03 < SuperSeriousCat> np, realbadhorse 19:04 < uplime> oh 19:04 < uplime> this is not ##chat 19:04 < uplime> i was very confused 19:04 < SuperSeriousCat> I just asked since you are the official Freenode bartender 19:06 < SuperSeriousCat> How are ##chat nowadays? Still have the little kid thinking he know it all as op and the middle aged attention seeking UK woman repeating herself every day? :P Got bored if that channel pretty quick 19:06 < uplime> lol 19:06 < uplime> its under new management TM 19:13 < annihilator> Is 50gb enough for /home if I'm using raid 0 for my data 19:13 < GunqqerFriithian> I don't suggest using raid 0 19:15 < Pentode> 50gb ought to be enough for anybody 19:15 < Pentode> and raid0 is asking for trouble 19:15 < alex88> Hello everyone, having a raspberry pi with an attached usb->eth adapter, the naming eth0 and eth1 for the network interfaces will ever change? for example right now eth0 is the integrated one and eth1 the usb adapter 19:15 < annihilator> I have 4 1tb hdds used for data 19:15 < annihilator> All important data is on an external 19:16 < Dagmar> alex88: If you install a different distribution, they could change to enp0s1 or somesuch 19:16 < alex88> I was wondering if there is a way to tell the os to never change the interface name so it's not possible for them to be swapped 19:16 < annihilator> I'm setting up linux as I did windows 19:16 < Dagmar> alex88: Otherwise, they're not going to change around 19:16 < revel> Pentode: 50GB? Enough for anybody? Guess I'm nobody <_< 19:16 < kerframil> annihilator: the thing is, nobody else can project your space requirements 19:16 < annihilator> OS on ssd data like games are on my data raid0 19:16 < Pentode> all that space and only 50gb for home? 19:16 < GunqqerFriithian> 50gb def not big enough for me 19:17 < kerframil> annihilator: if you can't decide, LVM is a potential option (allowing for future growth). 19:17 < alex88> Dagmar: oh ok, since I'm NATing a vpn connection on eth1 and I want to be 100% clients connected to it will never go through the other interface but with the vpn only 19:17 < annihilator> Kerf its why I was going to put /home on raid so I don't have to worry 19:17 < SuperSeriousCat> He said he use other disk for data. So im guessing his /home will be only config files. 50gb is enough in that chase 19:17 < Dagmar> alex88: ...and you don't see it as a problem that everything would just stop working entirely if the interfaces swapped? 19:17 < annihilator> But you said put it on ssd lol 19:18 < Dan39> i suggest raid0 :D 19:18 < Dan39> just not for anything you wouldnt mind losing 19:18 < GunqqerFriithian> raid 0 is asking for shit to break 19:18 < annihilator> All that is on external or on my onedrive account 19:19 < annihilator> I been running raid 0 on windows with it running off of one drive and external. All my important stuff will be saved game files 19:19 < alex88> Dagmar: no that would be less a problem, it would be much worse if I connect to my company VPN (which I tunnel through this VPN) from the country I'm currently in 19:19 < annihilator> And those saved game files will be on my external 19:19 < revel> I've got damn near a terabyte in my $HOME 19:20 < annihilator> I know with running qemu and windows I need more 19:20 < Dagmar> alex88: Anyway, the interfaces won't swap. There's already a mechanism in place to prevent it, plus, the things that would cause them to swap do not happen when both components are soldered onto the board 19:20 < alex88> Dagmar: ok! thank you very much for your help! 19:20 < Psi-Jack> Hmmm. 19:21 < Psi-Jack> Not entirely true, though. 19:21 < Psi-Jack> ethX naming convention is only "static" if you have udev rules in place to insure it. 19:21 < Psi-Jack> Debian < 8 used to save a file automatically for this purposes, but has since been removed due to predictable naming. 19:22 < Psi-Jack> Check /etc/udev/rules.d 19:23 < buoyantair> #offtopic -> open source mailing list manager? 19:23 < Dagmar> Raspbian, dude 19:23 < Psi-Jack> Which is based on Debian. :p 19:23 < buoyantair> anyone? 19:23 < Dagmar> ...which happens to give a damn about one and only one platform, so they configure them properly. 19:23 < Dagmar> Out Of The Box. 19:23 < Psi-Jack> Not from what I saw. 19:23 < Dagmar> Go buy a couple of them and use them 19:24 < Psi-Jack> I did. 19:24 < Psi-Jack> And do. 19:24 < Dagmar> Sounds like a personal problem to me then 19:25 < kerframil> just to be clear, the kernel has no formal mechanism to guarantee that device names don't change in the long term. you use udev rules to designate names (not in the 'eth[0-9]*' namespace) - like udev does by default - or you simply take your chances. 19:25 < alex88> btw before I was reading about the /etc/udev/rules.d/70-persistent-net.rules file which isn't there on raspbian, but maybe creating it might set the name to a fixed one 19:25 < Psi-Jack> heh 19:25 < Psi-Jack> Dagmar: I rest my case. ^ 19:25 < Psi-Jack> Is NOT there. 19:26 < kerframil> alex88: that's legacy. older versions of systemd-udev used to generate and populate your /etc/udev directory with such a file. that mechanism was dropped because it causes problems. 19:26 < kerframil> alex88: you can sitll write your own rules, if you prefer 19:26 < Psi-Jack> kerframil: Oh? Causes what problems? 19:26 < alex88> kerframil: what's the non legacy way now to ensure they'll never change? set the rules and don't use eth* names? 19:27 < Psi-Jack> alex88: Using eth* names is fine, but you can get more creative by giving them more useful names. 19:27 < Psi-Jack> Like, lan0, vpn0 19:27 < kerframil> Psi-Jack: the problem wasn't the generation aspect, but because the rules that were generated used a naming scheme that overlaps with the kernels ("eth" followed by digits). the result - in some cases - was that udev would race with the kernel as it was in the very process of handing out names, leading to very bizarre results. 19:27 < alex88> so creating those "rules" is still the recommended way to make them fixed? 19:27 < Evidlo> does termite not have a keyboard-based way to copy urls? 19:27 < kerframil> Psi-Jack: I ran into that problem exactly once, which lead me to a bug report where Kay Sievers explained. 19:27 < Psi-Jack> Ahh yes, the renaming effect was sometimes an issue. 19:28 < Psi-Jack> kerframil: Yep. I hit it a couple times. :) 19:29 < kerframil> Psi-Jack: so, while you're right that the 'predictable' scheme put a nail in the coffin, the rules generator was also dropped because of that, with Kay saying that he had no intention of its return. it could have been made to work if it chose a different naming scheme but with predictable names, they just didn't see the point. which is fair enough. 19:29 < sysrpl> hey can someone please help me resolve strange problem? at both lightdm and gdm3, i can move my mouse, but left click don't ativate anything. they only cancel. as a reutls i cannot change my xsession. also, relate in cinnamon, the task panel doesn't not respond to left clicks as well 19:29 < kerframil> the user gets to write their own rules if they want - so it's all good 19:29 < sysrpl> but i can use left lciks everywhere lese 19:30 < Dagmar> ANy way you slice it, there's no issue with having the devices swap around between reboots anymore 19:30 < Psi-Jack> kerframil: I kinda still hate the "predictable names" 19:30 < comet23> do people know this shit by heart? 19:30 < comet23> https://www.howtoforge.com/tutorial/install-mongodb-on-ubuntu/ 19:30 < Psi-Jack> comet23: Yes, and mind your language. 19:30 < comet23> or is that something you always copy and paste 19:30 < kerframil> Psi-Jack: I have mixed feelings about it. lately, I've tended not to bother fighting it on systems with one or two interfaces. for anything more complex, I like to disable the default scheme and just write my own rules file. 19:30 < sysrpl> anyone/ 19:30 < Dagmar> ...and certainly not when one of them is part of the main chipset 19:30 < uplime> comet23: smart people use config management so they only have to do it once 19:30 < kryptynasium> hexnewbie: to me lazy translates to MVP https://en.wikipedia.org/wiki/Minimum_viable_product iterate fast. Apple now embraces it: iterations > innovation 19:31 < kryptynasium> *iteration 19:31 < kerframil> Psi-Jack: the greatest criticism I can level at the scheme is that they are no predictable in the generally accepted sense of the word 19:31 < Dagmar> kryptynasium: Iteration means everyone rushes out to buy the new shiny because they're Apple fanatics 19:31 < kerframil> Psi-Jack: persistent, yes. but predictable in advance? not exactly 19:31 < Psi-Jack> kerframil: Exactly that. 19:31 < Evidlo> sysrpl, try a new mouse, or see if you can use Tab to selection the session icon 19:32 < sysrpl> Evidlo, it does the same thing with the touchpad 19:32 < sysrpl> and when i tab to the session icon, it requires the mouse to select an item. the keyboard cannot slect from the lsit 19:34 < kryptynasium> Dagmar: fan boys disappointed WWDC18 - no new hardware launches - just better software. Woohoo for memoji :) 19:36 < Evidlo> sysrpl (IRC): I just tried with lightdm. You can access the menu with F10, then use arrow keys 19:38 < Evidlo> actually F9-F12 are the different menus 19:40 < Evidlo> sysrpl (IRC): also, if graphical stuff breaks for whatever reason, you can hit Ctrl+Alt+F[1-7] to access a terminal 19:41 < royal_screwup21> Is it possible to create a new "instance" of ubuntu with super user privileges? I want my friend to use my pc for the linux experience, but since mine obviously has my credentials and stuff, I have to figure out a way to create a clean, separate instance of ubuntu. I know I can make a partition, but I'm not sure I want to go through the hassle just 19:41 < royal_screwup21> for this one time 19:41 < Dagmar> install virt-manager and run a virtual machine 19:42 < angelo_ts> hi, is there any tool able to repair a HFS+ disk from linux ? 19:43 < sysrpl> Evidlo, yes i know that 19:45 < dka> Are SHA512 password always equal for the same string ? or they can accept multiple string combination for the same ? 19:51 < Evidlo> a fixed input to SHA512 produces fixed output 19:51 < uplime> thats why you salt it, no? 19:52 < Evidlo> its possible that two inputs produce the same output, but very unlikely 19:52 < uplime> right 19:52 < uplime> google proved that with sha1 a few months back 19:54 < Evidlo> that should be true for any function which maps from big thing -> smaller thing 19:55 < Evidlo> lots of things -> less things, I mean 19:55 < hexnewbie> dka: Do you mean equal hashes for same password, or equal passwords for same hash? Equal passwords do get the same hash, which is why uplime suggests you salt it, so they get a different one. Different passwords/data sequences would have the same hash (all hash functions have collisions), but the chances are astronomical. So unlikely ZFS is using SHA256 for data dedup without verification 19:55 < hexnewbie> dka: Er, s/would have/could have/ 20:02 < hexnewbie> dka: https://en.wikipedia.org/wiki/Birthday_attack If you create a SHA512 hash of *every* kilogram of matter in universe, there is a chance *less than* 1 in 1 American quintillions that there would be *1* collision of two of the hashes. For SHA384, the chance would be approximately 1 in a trillion. 20:20 < CrazyTux> hello, can we install and use "Timeshift" in other distros? It is the main feature of Mint now. 20:21 < Kl0nEz> hello 20:35 < CrazyTux> hello.... 20:36 < Artemis3> CrazyTux, can we use google? https://itsfoss.com/backup-restore-linux-timeshift/ 20:37 < CrazyTux> Artemis3, thanks 20:38 < annihilator03> Device /dev/sda not found (or ignored by filtering). can someone help? 20:40 < littlebean> Hello, which channel is more suited too data wiping context? 20:40 < littlebean> from a hardware level, boot up into a bootable software too wipe everything 20:40 < littlebean> I mean, perhaps even a linux live distro could do it? 20:41 < littlebean> I've been trying things like Ultimatebootcd and MBAM with no success 20:41 < Dagmar> littlebean: https://tiptopsecurity.com/how-to-securely-wipe-your-hard-drive-with-dban-erase-your-data-for-good/ 20:42 < littlebean> thank you 20:43 < littlebean> <3 20:44 < windsurf_> Does linux (debian) ever spontaneously reboot? I'm looking at syslog and am seeing a startup boot sequence without any indication of a shutdown I think. 20:44 < windsurf_> I don't have a kernel log for the day that it occurred so no luck looking with dmesg 20:44 < windsurf_> device is battery powered but power level is still high 20:45 < littlebean> Dagmar: I share something back as thanks, https://www.reddit.com/r/TronScript/comments/8l3lry/tron_v1051_20180521_bugfixes_and_misc_updates/ 20:45 < Dagmar> Why would a server spontaneously reboot without there being a massive problem? 20:45 < littlebean> I got a new laptop and haven't even booted into it yet 20:45 < oerheks_> powerdrop.. 20:45 < windsurf_> Dagmar was that directed at me? 20:45 < Artemis3> obviosuly servers need to reboot at random 20:45 < Dagmar> windsurf_: Yes. Your question was a bit silly. 20:45 < littlebean> it came with windows preinstalled which I am going to dban, then put a none affiliated w10 on, then run TronScript asap 20:46 < Dagmar> We wouldn't be building entire enterprises on Linux if it just randomly decided to reboot for no reason 20:46 < windsurf_> fair enough. just need to figure out why it rebooted then 20:47 < qrvpzvb> udisks2 isn't letting me create files on my external HDD 20:47 < windsurf_> I have no reason to believe it was a powerdrop 20:47 < qrvpzvb> the thing is, everything in there belongs to my user 20:47 < qrvpzvb> with write permissions of course 20:47 < uplime> Dagmar: its a feature TM 20:47 < Dagmar> qrvpzvb: It's still mounted read-write? 20:47 < qrvpzvb> so I don't see what is blocking me 20:48 < qrvpzvb> Dagmar: yes 20:48 < Dagmar> qrvpzvb: ALl it takes is one write error from the drive for the kernel to mark the entire device read-only as a safety measure 20:48 < Dagmar> This will not be reflected in the mount status of the filesystems impacted. 20:48 < Dagmar> You will get a message in the syslog about it happening, and that's about it 20:48 < qrvpzvb> oh wait, it's using the kernel ntfs "driver" 20:48 < qrvpzvb> not ntfs-3g 20:48 < Dagmar> Ouch 20:49 < phogg> randomly rebooting for no reason is more of a Windows trait 20:50 < Artemis3> Dagmar, don't you need mount option errors=remount-ro for that? 20:50 < Dagmar> No. 20:50 < Dagmar> It's the default behaviour 20:50 < Dagmar> It happens all the time with SAN environments if the network is interrupted 20:50 < Dagmar> The network wobbles for a moment, and *BOOM* hundreds of mounts are now read-only 20:51 < Dagmar> This is why SAN stuff is followed to closely by multipath routing improvements 20:51 < Dagmar> s/to/so/ 20:51 < Dagmar> Once there's a failure to write to a disk, the kernel has _no_ idea what state the disk is in 20:51 < Dagmar> The only thing likely to happen from that point forward is _data corruption_ 20:52 < Dagmar> There's basically no choice but to immediately flag the device as read-only and hope a human can sort it out 20:52 < qrvpzvb> now fuse fails with "device or resource busy" 20:53 < Dagmar> qrvpzvb: Cd out of the mount point 20:53 < Artemis3> "errors={continue|remount-ro|panic} Define the behavior when an error is encountered. (Either ignore errors and just mark the filesystem erroneous and continue, or remount the filesystem read-only, or panic and halt the system.) The default is set in the filesystem superblock, and can be changed using tune2fs(8)." 20:53 < Dagmar> You can quote whatever you like, it's not going to change things 20:53 < phogg> Panic on write error. What a good idea. 20:53 < Dagmar> Well, at least someone notices _that_ right away 20:54 < Dagmar> Usually it's "our logs haven't been getting written for a month now and we'd like you to figure out why" 20:54 < qrvpzvb> oh yeah, thanks Dagmar !! 20:54 < phogg> Dagmar: Try telling PHP developers something like that. 20:55 < Dagmar> The only thing I'm going to tell PHP developers will probably invove a small mallet 20:55 < annihilator03> i install linux on the hdd not the cache partition? 20:58 < Dagmar> You probably shoudln't install it on any odd socks you've lying around 20:59 < dka> I have debian, I want to a windows software 20:59 < dka> I have just formated because installing whine previously destroyed my kernel 20:59 < dka> Cause I type yes do as i said 20:59 < Dagmar> dka: You've wandered into the wrong neighborhood, boy-o 20:59 < dka> I want to still have windows just for one software 20:59 < Dagmar> Oh you want to install some Windows software under WINE 21:00 < dka> I am thinking of using a VM like vmware 21:00 < dka> Yeah but wine did fuck my linux 21:00 < dka> previously 21:00 < dka> I want something more soft 21:00 < dka> that doesnt require me to update all my apt libs 21:00 < dka> I rather keep concern separate 21:00 < dka> whats the best way to do so? 21:00 < Dagmar> Well, if your CPU supports virtualization and you're not using a terrible distro, install virt-manager and spin up a VM using KVM/qemu 21:01 < Dagmar> There's no sense in paying money to VMware for something like this 21:01 < phogg> dka: use Wine 21:01 < phogg> dka: wine cannot break your Linux install. If it broke then the cause is something else. 21:01 < dka> phogg it ask me to change massive lib from my system 21:01 < dka> which is clean 21:01 < dka> I dont want to 21:01 < Dagmar> Use the distro-provided packages for WINE. 21:01 < phogg> dka: apt-get install wine # this should be fine 21:02 < dka> This is not 'sudo dpkg --add-architecture i386 ' 21:02 < Dagmar> phogg's entirely right 21:02 < Dagmar> dka: Uh, it doesn't replace the 64-bit libs 21:02 < dka> ok i try 21:02 < phogg> dka: if you have both 386 and amd64 architectures enabled you get libs from BOTH 21:03 < Dagmar> 32-bit and 64-bit versions of libraries can sit on the same machine just fine (assuming a reasonably well made distribution like Debian) 21:03 < Dagmar> Debian developers will have built things properly so they libraries, um, "never touch" 21:04 < Dagmar> s/they/the/ 21:05 < dka> Ok so now how can I install and execute my PokerStars.exe ? 21:05 < dka> wine PokerStars.exe ? 21:05 < phogg> dka: yes 21:05 < Dagmar> Ugh yes 21:05 < dka> or wine64 21:05 < dka> wine64-stable 21:05 < dka> wineboot ? 21:05 < Dagmar> I unfortunately know first-hand that one works under WINE 21:05 < annihilator03> LOL 21:06 * dka spew cash 21:06 < Dagmar> All that work done to get my dad off of bloody Windows 21:06 < Dagmar> s/of/off/ 21:06 < phogg> Dagmar: you had it right the first time. 21:06 < Dagmar> Ugh. Not enough caffienation yet 21:06 < Dagmar> Pardon while I just drip some straight into my eyes 21:16 < plexigras> how do i count the leading whitspaces in bash? 21:17 < kurahaupo> plexigras: leading whitespace in what? 21:17 < phogg> plexigras: leading whitespace where? 21:17 < plexigras> a variable 21:17 < kurahaupo> plexigras: remove from the first non whitespace to the end, then count what's left 21:18 < kurahaupo> x=${var%%[![:space:]]*} ; echo ${#x} 21:20 < plexigras> substitution failed 21:21 < plexigras> nvm got it to work 21:21 < useless-eater> going to put 4x 3T disks in a raid 10, should I use software raid + lvm or btrfs? Winchester disks. What would generally be best? 21:22 < useless-eater> mostly static storage, but a couple of busy nfs shares also maybe 21:23 < useless-eater> btrfs is new and exciting, but sw raid + lvm has never failed me. 21:23 < phogg> I don't trust the longevity of btrfs, personally. LVM seems safer. 21:24 < meyou_> it's all about commercial sans presented to ZFS presented to btrfs 21:25 < Cyrum> is btrs a write/copy model? 21:25 < Cyrum> btrfs* 21:25 < Cyrum> i cant remember 21:27 < SuperSeriousCat> It is CoW by default if thats what youre asking 21:30 < nicolas17> how does badblocks -n (non-destructive read-write mode) work? 21:31 < klinux> hi guys 21:31 < nicolas17> does it read and then re-write the existing data, or does it write something else and then restore the original? the latter would lead to lost data if eg. there's a power loss while it's running... 21:34 < klinux> I have this problem with linux https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=684425#17 21:34 < useless-eater> hmm, how come General Failure is reading my hardd rives? 21:34 < Neobenedict> how can i test udp packet loss 21:35 < Neobenedict> and get a number 21:35 < moony> question: is there a way for a process to ask the kernel to put a block of memory into swap? 21:35 < moony> (that it owns) 21:35 < uplime> Neobenedict: you can't really, 21:35 < uplime> unless you run packet captures on both peers 21:36 < nicolas17> moony: I don't think so 21:36 < nicolas17> moony: why do you need that? 21:36 < moony> just curious 21:36 < moony> i dont need it lol 21:38 < nicolas17> if you mmap a file, you can say "I won't need this, remove it from RAM and re-read it from the file if I need it again" using madvise with MADV_DONTNEED 21:38 < nicolas17> but I'm pretty sure that won't work for anonymous memory with no backing file 21:42 < magyar> hi, how can I collect the free smime cert from comodo? 21:43 < magyar> the buggers only support windows or apple osx 21:52 < alex88> so after I've named the device nat0, I'm trying to let devices connected to nat0 go trough the vpn device to access internet, this is what I have right now https://gist.github.com/alex88/1a7f32daf2ea7da64822046bf33d26bd.. do you see any problem here? it's strange because DNS works fine and I can see packets going back and forth the 2 interfaces (nat0-tun0) but https for example doesn't work 21:52 < alex88> I see the packets going in from nat0 but not going to to tun0 22:00 < camsn0w> How do I set a powerline font in my ttys? 22:00 < dka> phogg, Dagmar it is not working 22:00 < dka> sudo wine PokerStars.exe 22:00 < dka> wine: Bad EXE format for Z:\home\dka\Downloads\Downloads-next\PokerStarsInstall.exe. 22:00 < dka> I want to be able to run .exe on Debian stretch 22:00 < dka> how to do? 22:00 < revel> As root? @_@ 22:00 < revel> dka: What does `file PokerStars.exe` say? 22:01 < prussian> uh. run a valid interpreter for your PE? 22:01 < nicolas17> dka: why on earth are you using sudo for that? 22:01 < prussian> probably a PE32, vs whatever 64 bit PE format exists 22:01 < dka> because it asked me to 22:01 < nicolas17> it could be .NET 22:01 < prussian> also true 22:02 < dka> Logs; https://paste.gnome.org/pylb8xafy 22:02 < dka> If I type : ` wine PokerStarsInstall.exe `, then I have 22:02 < dka> it looks like wine32 is missing, you should install it. 22:02 < revel> Get 32-bit wine then. 22:02 < revel> Do what it's telling you to do .-. 22:02 < nicolas17> it told you to install wine32 as root 22:02 < prussian> ok. well, then you need i686 wine and equivalent libs 22:02 < nicolas17> not to run wine as root 22:04 < alex88> ok found something, sites like facebook and others works.. google for some reason doesn't.. 22:05 < alex88> https://gist.github.com/alex88/1a7f32daf2ea7da64822046bf33d26bd#file-gistfile2-txt if you look here, seems routing doesn't use the vpn gateway but the home router 22:05 < alex88> however input interface is nat0 which should use the vpn table which should use the vpn gateway 10.12.10.5 22:10 < phogg> dka: https://wiki.debian.org/Wine 22:10 < phogg> dka: follow the directions on that page 22:10 < dka> where does Wine install things ? 22:10 < dka> I have run Wine Pokerstars.exe 22:11 < dka> but my pwd was ~/Download/Downloads-next 22:11 < phogg> dka: It doesn't. You choose where at install time, or the installer does (just as on Windows) 22:11 < dka> is it an issue ? 22:11 < revel> phogg: The output he pasted even tells him "execute "dpkg --add-architecture i386 && apt-get update && apt-get install wine32" 22:11 < dka> the Installer choose in C:/Program Files :D 22:11 < revel> A whole 6 times, in fact. 22:13 < dka> where does Wine install things ? 22:13 < phogg> dka: *it doesn't choose* 22:13 < revel> I think it's usually ~/.wine 22:13 < phogg> dka: The installer you're running chooses, usually it prompts you. 22:13 < phogg> if you use the fake C: drive then that's likely in ~?.wine, but could be otherwhere defending on your config 22:13 < dka> thanks revel 22:14 < dka> so my entrypoint is the .exe ? 22:14 < dka> even for a program already installed ? 22:14 < revel> Uhh. If the .exe is an installer, then no. 22:14 < jim> it would be something the installer installs? 22:15 < phogg> dka: where did you *tell* it to install to? Did it make a shortcut on your desktop, maybe? 22:15 < Psi-Jack> Also, maybe #winehq is a better channel for this. 22:16 < jim> that does sound like something to try 22:16 < dka> ty 22:17 < jim> (not saying you can't be here or you can't ask here) 22:18 < jim> also, please expand ty as thank you (channel policy) 22:19 < dka> Is it normal that it is installed in /root/.wine/drive_c/ ? 22:19 < jim> maybe... that's within root's home dir 22:20 < dka> Well, why root is having .wine related thing? 22:20 < dka> why not me ? 22:20 < Psi-Jack> No, it's not normal at all to be using /root/ for anything user-based. 22:20 < jim> could it be you ran the installer as root? 22:20 < dka> I wonder jim 22:20 < phogg> dka: you should not run wine as root 22:20 < Psi-Jack> ^ 22:20 < dka> It's done now, should I rm -rf /root/.wine ? 22:20 < dka> and reinstall ? 22:21 < phogg> dka: yes 22:21 < jim> well before removing stuff, try installing as a plain user first 22:21 < toothe> My kernel is having trouble with something called MMIO 22:21 < toothe> Is there a way to disable that? 22:21 < Psi-Jack> toothe: Vague input. 22:21 < toothe> pardon? 22:21 < Psi-Jack> input is vague. 22:22 < phogg> toothe: mmio is a big topic 22:22 < toothe> My question? :) 22:22 < Psi-Jack> You should provide more details on what actual problem you're having. 22:22 < jim> could mmio be memory mapped i/o? 22:22 < toothe> yes...le tme get the exact detail from journalctl. 22:23 < TonyWonder> Did somebody say 'Wonder?' 22:23 < Psi-Jack> No. 22:23 < jim> TonyWonder, well, you just did... 22:23 < Psi-Jack> Rotating door? 22:24 < hexnewbie> Generally, something dropping love letters in logs is not reason enough to rush to disable things, especially if you don't know what they are. 22:25 < toothe> ahh, so a driver is having trouble. Journalctl says: nouveau 0000:01:00.0: bus MMIO read of 0000000 FAULT at 3c7e3c IBUS 22:26 < Psi-Jack> heh 22:26 < Psi-Jack> toothe: So, no. You can't "disable" that. :) 22:26 < plexigras> how can i have a bash regex like this that replaces all things that are not matched with spaces https://regex101.com/r/RI8twK/1 22:26 < hexnewbie> I have so little idea where ‘MMIO’ stands exactly in the land of memory mapping. Wouldn't disabling be akin to disabling the entire computer? 22:27 < GunqqerFriithian> CPU has error -> disable CPU 22:27 < NGGJamie> plexigras, If you mean doing it in a bash script, sed might be your friend 22:27 * Psi-Jack watches GunqqerFriithian shut down. 22:27 < toothe> Psi-Jack: yeah, that was dumb question. 22:27 < GunqqerFriithian> computer throws error -> disable computer 22:28 < toothe> I think this is a kernel module or something that's causing the system to fail to boot. 22:28 < hexnewbie> Well, joking aside, going headless will get rid of the error. 22:28 < NGGJamie> plexigras, https://linux.die.net/man/1/sed 22:28 < Psi-Jack> GunqqerFriithian: Ssh. You're already stopped. :) 22:28 < GunqqerFriithian> Can't have an error if your computer isn't on *taps head* 22:28 < Psi-Jack> hexnewbie: Le-gasp! Off with his headf? 22:28 < Psi-Jack> head* 22:28 < Psi-Jack> toothe: Or... You have bad hardware. 22:28 < plexigras> right now i have `${2//[[:punct:][:digit:]]/ }` but that removes the punctuation that is inside words 22:29 < hexnewbie> toothe: That's the last logged error before the boot process grinds to a halt? Is the kernel panicked (like blinking keyboard LEDs)? 22:29 < hexnewbie> toothe: And what's the last thing you see on the screen during the failed boot? 22:29 < Psi-Jack> plexigras: What.... Are you actually trying to do in the end? 22:30 < toothe> hexnewbie: so 22:30 < toothe> I actually don't see that happen when I am booting. 22:30 < plexigras> i want `I'm the best.` turn into `I'm the best ` 22:30 < toothe> I get this vague Booting UTMP message. 22:30 < toothe> my coworker checked journalctl and saw this message. 22:30 < toothe> so, we suspect its a driver issue? 22:30 < toothe> these are brand new computers. 22:31 < Psi-Jack> plexigras: Into a file? Into a variable within bash? 22:31 < plexigras> yes a variable inside bash 22:31 < hexnewbie> toothe: Computer*s*? Do all of them fail in the same way? Did you try a different distro? 22:32 < toothe> Yes, I did. We ran it on Debian, no problem. 22:32 < Psi-Jack> plexigras: Have you asked in #bash? 22:32 < toothe> only Kali has this issue. (damn, i hate Kali) 22:32 < toothe> so, I dunno what else might be different. 22:32 < bitSt0rm> Is it possible to use lightdm as display manager and lxde as WM and meanwhile not use lxdm? 22:32 < plexigras> Psi-Jack: that's a good idea 22:32 < Psi-Jack> toothe: Kali is not for desktop use. 22:32 < NGGJamie> That's weird, considering Kali is just essentially an Ubuntu spin 22:32 < Psi-Jack> NGGJamie: Not really. 22:32 < NGGJamie> But I wouldn't install it permanently anyway 22:32 < GunqqerFriithian> So we should use kali as a daily driver :P 22:32 < toothe> Psi-Jack: I know. 22:32 < Psi-Jack> It's a locked down specifically customized fork. 22:32 < toothe> This isn't my choice. 22:32 < lnnb> toothe: you can get more verbose info by setting the drm.debug level on kernel cmdline 22:32 < toothe> Psi-Jack: I don't agree with this decision at all. 22:32 < Psi-Jack> toothe: Sure it is. 22:32 < Psi-Jack> 100% completely your choice. 22:32 < toothe> no, its not. 22:33 < toothe> its 0% my choice. 22:33 < toothe> lnnb: Let me try that. 22:33 < NGGJamie> Sounds like he might be doing it for a company or some likes 22:33 < toothe> its the government. 22:33 < Psi-Jack> And yet, he still has a choice. :p 22:33 < toothe> Once its in policy, nothing can change it. 22:33 < hexnewbie> toothe: Well, since Kali doesn't work on the computers, using Kali is no longer *a* choice, regardless whose ;) 22:33 < phogg> toothe: the *government* lets you run *Kali*?! Now that's surprising (and irresponsible)! Which government? 22:33 < dka> I am on cinnamon and I try to create a shortcut in my menu. My command is '/usr/bin/wine ~/.wine/drive_c/Program\ Files/PokerStars/PokerStars.exe' , I have check `Run in terminal` to see why it was not starting, and I see the terminal closing straight away. How can I do a shortcut for wine? 22:34 < phogg> dka: you don't need or want to run that in a terminal 22:34 < phogg> it's not a terminal program 22:34 < Psi-Jack> dka: #winehq, and look into PlayOnLinux 22:34 < upgrade> Hi all, have some errors in dmesg. intel-spi intel-spi: unrecognized JEDEC id bytes: c2, 25, 38 22:34 < upgrade> and EDAC pnd2: Failed to register device with error -22. 22:35 < upgrade> In lts-version dont have any. 22:35 < phogg> dka: you make a standard .desktop file as you would or non-wine, only you run wine and give it a program argument. Or, you set up binfmt-misc to exec wine for PE executables (but then you can't really choose alternate WINEPREFIXes) 22:35 < phogg> s/or non-wine/for non-wine/ 22:36 < za1b1tsu> What tilling WMs work great with gnome? 22:36 < Psi-Jack> za1b1tsu: None 22:36 < GunqqerFriithian> nothing 22:36 < za1b1tsu> thanks for saving some big headaches lol 22:36 < za1b1tsu> *me 22:36 < za1b1tsu> too bad 22:37 < Psi-Jack> You can run gnome apps, but gnome as a whole, not so much. :) 22:37 < dka> phogg, I have created a launcher using cinnamon-menu-editor, I have added this Command to launcher: /usr/bin/wine ~/.wine/drive_c/Program\ Files/PokerStars/PokerStars.exe 22:37 < dka> I have uncheckd Open in terminal 22:37 < dka> and it is not starting at all 22:37 < dka> I dont want a desktop icon 22:37 < Psi-Jack> dka: #winehq, and look into PlayOnLinux 22:37 < dka> I rather hide it on my desktop 22:38 < NGGJamie> za1b1tsu, I can vouch for i3, it's a pretty nice tiling WM. 22:38 < GunqqerFriithian> PlayOnLinux is very good 22:38 < NGGJamie> Especially if you're fond of keyboard shortcut navigation 22:38 < GunqqerFriithian> i3 does have a learning curve 22:38 < dka> Ok i solve it 22:38 < dka> I could use ~ 22:38 < dka> *could not 22:38 < phogg> dka: of course not, that's a shell construct 22:38 < za1b1tsu> Psi-Jack: I tried multiple DEs on a MacBookPro from live usbs, with gnome everything works except webcam, the only DE that comes close is Budgie. I dont like DEs but setting up everything for mac seems like a nightmare 22:38 < Psi-Jack> Gnome has a learning curve. KDE has a learning curve. i3, fvwm, windowmaker, lxqt, linux, etc, all have learning curves. 22:39 < upgrade> So, no one knows? :c 22:39 < lupine> best to toss the mac really 22:39 < za1b1tsu> NGGJamie: you run it with gnome> 22:39 < GunqqerFriithian> i3 has a higher learning curve than gnome or unity 22:39 < Psi-Jack> Not really. 22:39 < GunqqerFriithian> Well if you're used to windows or OS X yes it does 22:39 < lnnb> theres some firmware loader support the kernel needs for "iSight" 22:39 < phogg> anything that is not point-and-clicky has a higher learning curve due to a lack of discoverability 22:39 < Psi-Jack> Same with Gnome or KDE. 22:39 < NGGJamie> za1b1tsu, I run Ubuntu MATE and everything I've ran on i3 displays fine. 22:39 < NGGJamie> MATE being gnome-based. 22:40 < Psi-Jack> MATE is Gnome 2 on failing life support. :p 22:40 < za1b1tsu> NGGJamie: Tried MATE on macbookpro, had some issues 22:40 < NGGJamie> I have it customized to hell, but you're not wrong Psi-Jack 22:40 < GunqqerFriithian> I personally use KDE Plasma 22:40 < za1b1tsu> hibernation does not work for me in KDE 22:41 < GunqqerFriithian> ? 22:41 < Psi-Jack> Yep. Time to stop using a MacBook Pro to run Linux. 22:41 < phogg> KDE is nice except for the CPU usage, occasional weirdness and the part where it uses an excessive number of X clients 22:41 < NGGJamie> KDE in my experience uses more resources than I'd like. Though I do like the look. On this particular laptop I'm working with some decent limitations 22:41 < Psi-Jack> Heh, KDE is awful these days. :9 22:41 < annihilator> Can I change the mount point / home after install? 22:41 < za1b1tsu> Psi-Jack: it's from work, I don't give money to Apple 22:41 < GunqqerFriithian> KDE does use a bit of resources, but I personally like it 22:41 < NGGJamie> Dual core with 3 something GB RAM, I don't have much to spare there. 22:41 < GunqqerFriithian> ah, even less than what I have 22:41 < phogg> annihilator: You can change anything you like 22:42 < Psi-Jack> zaibitsu: And, are you authorized to be running Linux on their hardware? 22:42 < phogg> annihilator: that said moving / to another device is not easy 22:42 < annihilator> I only care about moving /home as kubuntu is failing install 22:42 < za1b1tsu> Psi-Jack: so-so, I work remote, but I can't have issues with webcam, mic etc 22:43 < Psi-Jack> So that's.... A no. 22:43 < alexandre9099> is it possible to see for how long an interface has been up (assuming that is since when ifconfig starts counting the traffic)? 22:43 < NGGJamie> GunqqerFriithian, Dual core 2GHz processor, that's what it was. 3403MB RAM 22:43 < GunqqerFriithian> honestly with a MBP I suggest just sticking with OS X and use WMs made for OS X 22:43 < GunqqerFriithian> whew, NGGJamie that's low. Almost as low as my old laptop :P 22:44 < za1b1tsu> Psi-Jack: my plan is to have a common partition between mac and linux with all the projects that I have to work. I shit hits the fan on linux, I just switch to OSX 22:44 < za1b1tsu> *If 22:44 < NGGJamie> It didn't run Windows so well, and I wanted a primarily linux machine. It's been running various linux distros beautifully for years now. 22:44 < GunqqerFriithian> nothing runs windows well :P 22:44 < Psi-Jack> za1b1tsu: You'd better get proper authorization from your company before you even begin to consider that plan. 22:44 < Psi-Jack> Most likely it will be rejected. 22:44 < NGGJamie> I've even got some VMs under it that function decently for what I need them for. 22:45 < Psi-Jack> A big reason companies issue out Mac Book Pros is because of the remote management capability of it, securely. 22:45 < za1b1tsu> Psi-Jack: What could go so wrong though? Could I burn the Mac or what? 22:45 < Evidlo> How is it that I can copy and paste the ⏎ character, but I can't find its unicode code-point in /usr/include/X11/keysymdef.h ? 22:45 < za1b1tsu> I got it with factory settings, nothing to monitor me 22:45 < Psi-Jack> za1b1tsu: Criminal tampering. :) 22:45 < GunqqerFriithian> then there's my school who has imacs but it isn't locked down at all except an admin password lol 22:46 < __heisenbug> Only hipsters use MacBooks 22:46 < GunqqerFriithian> easy boot to single user and I have all the power 22:46 < NGGJamie> za1b1tsu, It's that it's an expensive piece of company equipment, and if the OS were replaced they would lose any potential warranty/support they have from Apple most likely. 22:46 < Psi-Jack> Also true. 22:46 < za1b1tsu> Really? You loose macbook warranty if you install linux???? 22:46 < GunqqerFriithian> F 22:46 < Psi-Jack> Well, d'uh. 22:46 < NGGJamie> Well, yeah. It wasn't made for that. 22:47 < za1b1tsu> God I hate Apple so much 22:47 < hexnewbie> Evidlo: Unicode codepoints are in the unicode standard, not in /usr/include/X11/keysymdef.h. Debian has them in /usr/share/unicode (look in UnicodeData.txt), but those are not necessarily up-to-date, and may lack certain characters (CJK) 22:47 < NGGJamie> Similarly to how you lose your warranty for jailbraking an i(phone|pad) 22:47 * Psi-Jack whips out his iPhone 8, while poking at his iPad Pro 10" 22:48 < notmike> Some dude said to me yesterday "well, with c++ now c is basically irrelevant" 22:48 < __heisenbug> Psi-Jack: I dont like that iTunes stuff 22:48 < NGGJamie> notmike, That's not true at all 22:48 < notmike> So you know I slapped his ass reeeal fast. 22:48 < lnnb> lol criminal tampering for installing linux? 22:48 < GunqqerFriithian> yeah with c++ c sharp is redundent 22:48 < notmike> Now I need somebody to bail me out :/ 22:48 < Psi-Jack> lnnb: For tampering with company property. 22:48 < __heisenbug> notmike: Everything you can do in C you can do in C++ too :D 22:48 < NGGJamie> GunqqerFriithian, C# has it's uses as a RAD platform. 22:48 < GunqqerFriithian> *whoosh* 22:49 < lnnb> Psi-Jack: that was issued to you 22:49 < Psi-Jack> Issued != given or owned. 22:49 < notmike> C is far superior to C++ 22:49 < GunqqerFriithian> it's prob company things, apple won't care (or know) if you installed linux on your computer 22:49 < __heisenbug> notmike: Why? 22:49 < NGGJamie> Depending on your country, that likely wouldn't be criminal. But you would most definitely not hold your job if they found out 22:49 < lnnb> i think you're a bit paranoid, legally 22:49 < notmike> You can do oop in C. And C++ breeds bad programmers. 22:49 < prussian> lol. C++ having the standard dsl of C++ alone makes it better than C 22:49 < GunqqerFriithian> IT def wouldn't be happy 22:50 < Psi-Jack> lnnb: Noe. This is very much reality. ;) 22:50 < za1b1tsu> meh, I work remote in another country 22:50 < notmike> Women will not date you if you say you're a C++ dev. C is an instant panty dropper. 22:50 < prussian> oooook 22:50 < NGGJamie> notmike, "i use printf instead of cout" 22:50 < __heisenbug> notmike: C is not designed for OOP. Every language breeds bad programmers. C++ is just a bit more complicated 22:51 < Evidlo> hexnewbie (IRC): All the python libraries I try have trouble sending arbitrary unicode strings to the keyboard 22:51 < Hdphn> Github is now owned by microsoft 22:51 < hexnewbie> notmike: You misspelled Haskel and Scheme. 22:51 < Hdphn> suck it open source people 22:51 < prussian> yikes 22:51 < __heisenbug> Hdphn: What do you gonna do now? It kinda sucks :/ 22:51 < __heisenbug> GitLab? 22:51 < lrvick> Anyone know a way to ensure every user can't use suid binaries in their home directories other than making /home a dedicated partition mounted with "nosuid" ? 22:51 < za1b1tsu> the only way I could get in trouble is if I say: Sorry guys my webcam does not work cause I am on linux, let me reboot in to OSX 22:51 < notmike> __heisenbug: kys. C was around before OOP was even a thing. You can still easily do OOP in C. There are whole bodies of work on the topic. 22:51 < Hdphn> __heisenbug: no. I use windows 10 with github. linux is dying. 22:52 < Psi-Jack> za1b1tsu: Also not true. 22:52 < hexnewbie> lrvick: Don't... let the users become root? 22:52 < Hdphn> opensource in general is dying 22:52 < notmike> __heisenbug: fossil? 22:52 < Hdphn> :) 22:52 < lnnb> Psi-Jack: what country/state does za1b1tsu live in? 22:52 < GunqqerFriithian> Begon ~~thot~~ troll! 22:52 < __heisenbug> Hdphn: You are allowed to leave the channel. 22:52 < Hdphn> __heisenbug: WSL + github integration. windows 10 is rock solid 22:52 < Psi-Jack> Intel Management Engine, which yes, is on your MacBook Pro, allows your company to fully remotely access that laptop. 22:52 < za1b1tsu> Romania :P 22:52 < notmike> I'm not a thot 22:52 < NGGJamie> za1b1tsu, I'd recommend checking up on your company's policies on tech use 22:52 < revel> __heisenbug: He's also been forcibly made to leave a couple of channels I'm on as well. 22:52 < Hdphn> Psi-Jack: intel management engine is on linux laptops too :) 22:52 < __heisenbug> revel: lol :D 22:53 < Psi-Jack> Hdphn: It's on anything that has Intel Management Engine hardware. :p 22:53 < Hdphn> yea :D 22:53 < za1b1tsu> Psi-Jack: What? my company can spy on me with Intel Management Engine ? 22:53 < lrvick> hexnewbie: well abuses of suid binaries in clever ways are often how people privesc or escape jails. By preventing them from being executed on that partition at all it avoids a wide range of possible attacks 22:53 < notmike> Kickbanning is the ultimate beta move. 22:53 < GunqqerFriithian> Yeah check company policy to see if you can get fired, figure out what you're allowed to do, and be nice to IT 22:53 < Psi-Jack> za1b1tsu: D'uh yep yep. 22:53 < Hdphn> so.. unless you use x86 thinkpads you cant remove Intel ;) 22:53 < za1b1tsu> Psi-Jack: even if got this mac with factory settings brand new, nobody touched it> 22:54 < __heisenbug> notmike: C is still not made for OOP. Even if you can make it OOP. Being published before OOP was a thing is not a valid argument I guess 22:54 < Psi-Jack> za1b1tsu: They issued it, right? 22:54 < __heisenbug> Hdphn: I like Thinkpads 22:54 < hexnewbie> lrvick: Of course, that's why there's nosuid, and it may fit home. But the users shouldn't be able to create setuid binaries in their home directories, regardless of nosuid. 22:54 < prussian> __heisenbug: it's a valid argument for its disuse 22:54 < notmike> Wouldn't surprise me to learn __heisenbug doesn't know how to read. Read one page of a Dr. Suess book and I'll donate 75k to a charity of your choice. 22:54 < Evidlo> hexnewbie (IRC): ok, I see the code I want in UnicodeData.txt, but why do these libs only handle certain characters? 22:54 < Hdphn> Psi-Jack: I would rather trust macbooks with their regular firmware updates than other windows/linux laptop whose bios is not updated after 6 months of its release date. 22:54 < za1b1tsu> Psi-Jack: they bought it 22:54 < Hdphn> __heisenbug: good for you 22:55 < Hdphn> tbh chromebooks and macbooks are the only ones updating their bios even 1 year later 22:55 < NGGJamie> IF you NEED OOP, then I would say you may as well go with C++ instead of attempting to self-implement it. It's always better to go with the peer-reviewed solution if it's available 22:55 < lnnb> as far as new york law is concerned, A person is guilty of criminal tampering in the third degree when, having no right to do so nor any reasonable ground to believe that he has such right, he tampers with property of another person **with intent to cause substantial inconvenience to such person or to a third person.** 22:55 < Hdphn> of their release 22:55 < __heisenbug> notmike: Also OOP is not the only thing C and C++ differ in 22:56 < lupine> the only thing of value in C++ is default arguments to methods 22:56 < Psi-Jack> za1b1tsu: IME is part of the hardware, not software. 22:56 < lupine> the rest can be tossed 22:56 < NGGJamie> C and C++ have entirely different cultures to how code is written. 22:56 < notmike> OOP is JUST a way of THINKING about and StRUctUrinG your programming 22:56 < rascul> oops 22:56 < user3> I use Thunderbird for email. how can I arrange to be notified at the command line ("You have new mail") when I receive new mai? Thunderbird's command line options seem irrelevant here 22:56 < hexnewbie> Evidlo: Which libraries would that be? Apart from some programs/libraries having issues with non-BMP characters (usually cosmetic-ish), I don't know of any libraries that don't handle all of unicode. Displaying characters is a different matter altogether though, as they added like daily now. 22:56 < za1b1tsu> Psi-Jack: and what kind of info could they get? 22:56 < lrvick> hexnewbie: sure. I am just looking for an alternate way to keep the failsafe security mounting with nosuid offers, without having a dedicated partition. Wondering if I can do some nosuid equivalent with filesystem namespaces or some other feature. 22:56 < Hdphn> za1b1tsu: the info linux sends to linus torvalds and its sponsors 22:56 < __heisenbug> notmike: Yes, that is true. 22:56 < notmike> If you don't use camel case, then what are you doing really? 22:57 < Psi-Jack> za1b1tsu: Literally, everything. 22:57 < lrvick> I want to be able to use filesystem mounting security features for user directories without needing real partitions. 22:57 < Hdphn> za1b1tsu: such as AWS 22:57 < Hdphn> AWS is on linux btw 22:58 < za1b1tsu> I don't know what I should be worried more, the massive amount of visists to porn sites or having linux installed lol 22:58 < sauvin> notmike: a more concise way of saying that: "it's just a paradigm". 22:58 < hexnewbie> lrvick: Bind mounts can be made read-only, perhaps they can be made nosuid (I rule it unlikely, but it would be very easy for you to test). Other than that, 1) it should be easy to make it a dedicated partition, 2) you may stop worrying about it - if a setuid binary appeared in a user's home directory, you're already in trouble for whatever caused it to appear there 22:58 < Hdphn> za1b1tsu: btw microsoft pays linus for its work 22:58 < Hdphn> za1b1tsu: can you visit porn sites on linux? does linux support browser? 22:58 < Psi-Jack> Hdphn: You should stop. 22:58 < Hdphn> with multimedia capabilities? 22:59 < rascul> hexnewbie users can create suid binaries at will 22:59 < Evidlo> hexnewbie (IRC): give me a minute to code up a sample 22:59 < sauvin> Hdphn, why don't you load linux up into a VM and see? 22:59 < jeffree> is there a was to know if a package was installed from from a non-repo source with dpkg -i ? 22:59 < hexnewbie> rascul: Er, I suspect lrvick means setuid root 22:59 < rascul> ok 22:59 < Hdphn> sauvin: last time I used linux, it didnt support sound card and wifi card I have 22:59 < rascul> users certainly can't create suid root binaries 22:59 < Hdphn> lol 22:59 < sauvin> Hdphn, get another computer. 23:00 < GunqqerFriithian> CLI-porn must be a thing, right? 23:00 < lrvick> hexnewbie: using real dedicated partitions in my target systems are not an option. virtual servers where I can't easily repartition, so I need some kind of virtual filesystem abstraction perhaps 23:00 < GunqqerFriithian> honestly wouldn't be surprised 23:00 < NGGJamie> Hdphn, That's massively become less of an issue in the past few years 23:00 < Hdphn> sauvin: didnt require getting old specialised hardware for windows :) 23:00 < lrvick> hexnewbie: vps host does not support external volumes etc 23:00 < rascul> lrvick lvm? 23:00 < `Koyaanisqatsi> so who's moving to gitlab? 23:00 < Hdphn> NGGJamie: ofcourse. microsoft is contributing to linux now it seems 23:00 < NGGJamie> Yeah, it was a bitch in ~2010-2012, but post 2016? Not really. 23:00 < hexnewbie> lrvick: rascul raises a good point, are you worried about root setuid binaries, or about setuid binaries owned by a user (allowing one user to execute with another's permissions)? 23:00 < __heisenbug> jeffree: dpkg-query -l 23:00 < lrvick> rascul: was not provisioned with lvm. not my call either. 23:00 < sauvin> Hdphn, when I buy a new computer at Walmart, I don't pay any attention to what it is or what it's got. Linux gets installed on it anyway. Get yourself another computer. 23:00 < lnnb> `Koyaanisqatsi: gnome and gimp 23:01 < NGGJamie> Yeah, Microsoft is creating their own Linux distribution 23:01 < NGGJamie> BlueSphere or whatever. 23:01 < Hdphn> sauvin: why would I pay for another computer just to install inferior monolithic kernel on it? 23:01 < `Koyaanisqatsi> so crazy that some people are trusting microsoft 23:01 < lrvick> hexnewbie: either. shared public unix system, and I want untrusted users to stay in their lane. 23:01 < rascul> sauvin do you typically buy computers from walmart? 23:01 < GunqqerFriithian> embrace, extend, exterminate :P 23:01 < `Koyaanisqatsi> are they sellouts or just ignorant? 23:01 < Hdphn> sauvin: future is fuchsia and other microkernels 23:01 < prussian> sellouts 23:01 < sauvin> rascul, on occasion. The point was generic. 23:02 < rascul> `Koyaanisqatsi there are a lot of people in the tech world nowadays who weren't around for the ms/linux wars 23:02 < __heisenbug> GitHub got bought by MS. What are you guys gonna do? 23:02 < rascul> __heisenbug i'm not going to do anything different 23:02 < GunqqerFriithian> ^ 23:02 < GunqqerFriithian> I don't have any repos so 23:02 < `Koyaanisqatsi> the linux kernel is not github 23:02 < Hdphn> `Koyaanisqatsi: well if you dont trust microsoft then you shouldnt use linux since microsoft has contributed the most to open source world in recent years 23:02 < sauvin> Neither is the sabayon community, rascul, unless MS changes github's terms of service or something. 23:02 < Hdphn> cheers 23:02 < hexnewbie> All is lost. Switching to Windows 10 seems to be the only possible recourse now. 23:03 < Hdphn> ^ 23:03 < revel> `Koyaanisqatsi: It's mirrored on it, at least. 23:03 < `Koyaanisqatsi> Hdphn, they're not in charge of the kernel 23:03 < rascul> i haven't really been using github for years, ms buying it won't directly affect me 23:03 < jeffree> __heisenbug: That tells that it is installed but not the source 23:03 < sauvin> `Koyaanisqatsi, here, "linux" means GNU/Linux. 23:03 < hexnewbie> Or least when I figure out how to replace LEDE with it. ;) 23:03 < rascul> ms buying github has indirectly affected me, because all the people switching to gitlab.com has made it slow :( 23:03 < `Koyaanisqatsi> look I'm open minded, but I'll believe it when I see it. Open source directx and we'll talk 23:03 < GunqqerFriithian> do you not self-host? 23:04 < __heisenbug> jeffree: Oh, I misread. Sorry 23:04 < notmike> sauvin: I'm working on being more humble about how autistic I am 23:04 < NGGJamie> I don't foresee MS giving GitHub much trouble. They most likely want github's infrastructure to become part of the microsoft culture to sell companies more stuff 23:04 < lrvick> hexnewbie: It seems like from googling a bit more that docker uses virtual volumes via their virtual filesystem drive that use nosuid, noexec etc. I guess I need to dig into how that is implemented. 23:04 < Hdphn> `Koyaanisqatsi: if you were realist and open minded, you wouldn't be using this crap at all 23:04 < Hdphn> no offense 23:04 < lnnb> if you want someone to blame, blame the linux foundation for getting cozy with them 23:04 < Hdphn> ^ 23:04 < Hdphn> that too 23:04 < rascul> GunqqerFriithian i don't generally bother hosting my git stuff, my servers can go away forever with no notice and i need my git repos to stay around 23:04 < Psi-Jack> And now, thank you for helping the troll troll more. :op 23:04 < `Koyaanisqatsi> Hdphn, what crap? 23:04 < Hdphn> lnnb: linux foundation needs the cash 23:04 < Hdphn> cash is the king 23:04 < GunqqerFriithian> The thing is, everything M$ has bought has only gotten worse or stayed the same so I'm not optimistic 23:04 < lrvick> I -could- just throw every single user in a container, but I would rather implement jailing with linux features I understand myself instead of "use docker". 23:05 < `Koyaanisqatsi> yeah, he's obviously a troll 23:05 < __heisenbug> jeffree: https://askubuntu.com/questions/8560/how-do-i-find-out-which-repository-a-package-comes-from?utm_medium=organic&utm_source=google_rich_qa&utm_campaign=google_rich_qa 23:05 < Hdphn> microsoft has cash that could buy and produce 20 more linus torvalds lol 23:05 < rascul> GunqqerFriithian dos got better 23:05 < Hdphn> s/20/20K 23:05 < GunqqerFriithian> everything modern 23:05 < sauvin> `Koyaanisqatsi, there's the even more frightening possibility this guy actually BELIVES what he's spewing. 23:05 < rascul> xenix probably got better also ;) 23:05 < GunqqerFriithian> ^^ 23:05 < lnnb> Hdphn: code is king 23:05 < lrvick> The only possible thing microsoft could do to keep me on Github, is open source it. ATM my 150+ repos and I are going elsewhere. 23:05 < `Koyaanisqatsi> sauvin, I can't prove it and I believe what I can prove 23:05 < lnnb> code > cash 23:05 < Hdphn> lnnb: code is useless. product is the king 23:06 < sauvin> I don't believe in dismissing possibility. 23:06 < NGGJamie> Either way, open source isn't really dying out, if anything it's seeing surges with linux becoming a more popular operating system as years go by. 23:06 < GunqqerFriithian> but... the product uses code... 23:06 < NGGJamie> Certainly it's in a better place now than where it was years ago 23:06 < __heisenbug> lrvick: GitLab? Or where are you going? 23:06 < `Koyaanisqatsi> I'll believe it when they open source directx, until then these linux people drinking the koolaid are crazy 23:06 < Hdphn> NGGJamie: it wont die out. microsoft is after it 23:06 < Hdphn> it has funded and contributed more to open source than redhat crap 23:06 < lrvick> __heisenbug: I am mirroring to gitlab for starters so I have my repos anywhere else at all, but I'll probably land on my own gitea instance as my personal source of truth 23:06 < `Koyaanisqatsi> or I think it's more likely that those people were posers all along 23:06 < sauvin> But but but but Kool-Aid has sugar in it! 23:06 < GunqqerFriithian> unless "extend embrace exterminate" rings true 23:07 < hexnewbie> lrvick: This worked for me: mount -o bind /home /home; mount -o remount,nosuid /home 23:07 < NGGJamie> I'm not precisely sure how they would manage to exterminate open source, open source software is a community, not a singular entity. 23:07 < lrvick> hexnewbie: is your /home a dedicated partition or just a folder under your root partition? 23:07 < rascul> hexnewbie nosuid doesn't work with the initial bind mount? 23:07 < Evidlo> hexnewbie (IRC): http://termbin.com/k4wm 23:07 < `Koyaanisqatsi> I do believe there is a small branch in MS that wants to embrace open source and give up on windows, but they are not microsoft. they are there as false validation if anything. sort of how like foxnews puts on some "liberal" actors. 23:07 < Psi-Jack> Hdphn: What are you doooooing here? Besides trolling? 23:07 < phogg> NGGJamie: not knowing how they might try is why everyone treats anything they do with suspicion 23:08 < GunqqerFriithian> Psi-Jack: that's all he's doing 23:08 < Hdphn> Psi-Jack: preaching the facts 23:08 < hexnewbie> lrvick: No dedicated partition. I actually tested with a newly created directory named /test in the / mountpoint 23:08 < NGGJamie> Even if Torvalds did abandon the work he's done over the past 20+ years for some cash, we're at a point where I'm pretty sure things would survive anyway. 23:08 < Psi-Jack> Hdphn: Please leave. 23:08 < lnnb> at least he's not telling people installing linux on their work laptop is criminal 23:08 < sauvin> Hdphn, more like "pissing off the ops". 23:08 < GunqqerFriithian> lol 23:08 < Psi-Jack> lnnb: I never said that. 23:08 < `Koyaanisqatsi> windows garbage is not going anywhere and they will never move against themselves to phase it out 23:08 < lrvick> hexnewbie: whoa. I really didn't think that was possible. So it should then be feasable to do an fstab incantation of that to bind mount /home on boot that way 23:09 < rascul> installing linux on the work laptop might be criminal, in some jurisdictions, in certain cases 23:09 < Psi-Jack> lnnb: I said, altering company property might be. There's a huge difference. 23:09 < `Koyaanisqatsi> so what if it's criminal 23:09 < `Koyaanisqatsi> bank fraud is also criminal 23:09 < `Koyaanisqatsi> and obstruction of justice 23:09 < GunqqerFriithian> walking across a street can be criminal 23:09 < Hdphn> accept the fact or be sold to your master (microsoft). cheers 23:09 < oo_miguel> Is there some mechanism that will allow me to create some kind of symlink to an existing file but additionally run some commands (as sed) over it.. 23:09 < phogg> GunqqerFriithian: jaywalking is a civil offense 23:10 < Hdphn> lnnb: thanks for being realist 23:10 < GunqqerFriithian> phogg technically true 23:10 < phogg> oo_miguel: first create a symlink, then run the other command 23:10 < rascul> it's a crime to tell people about mormonism in mississippi 23:10 < oo_miguel> phogg: I do not want to change the original file. 23:10 < hexnewbie> lrvick: I believe bind,nosuid will work in fstab's options. At least something is making it work for bind,ro even though ‘mount -o bind,ro’ doesn't work. However, I don't know if something won't get be confused by /home /home :) 23:10 < Psi-Jack> rascul: LOL, is that true? 23:10 < `Koyaanisqatsi> they can turn jaywalking into resisting arrest whenever they feel like it, and that is criminal 23:10 < phogg> oo_miguel: so you don't want a symlink. Make a copy. 23:10 < rascul> Psi-Jack something to that effect, yes 23:10 < GunqqerFriithian> I don't have a problem with that one, rascul :P 23:11 < oo_miguel> phogg: I want changes to the original file be reflected in the resulting (link) without having to rerun the command each time 23:11 < NGGJamie> rascul, I'm surprised they got away with outlawing that, doesn't that directly violate the first amendment? 23:11 < rascul> NGGJamie one would think, but i also doubt it's ever been tried 23:11 < GunqqerFriithian> ^ 23:11 < Psi-Jack> rascul: Heh. My parents were mormons. So, I can't live in Mississipi then. :) 23:11 < lnnb> sure it could be, maybe if you live somewhere backwards like australia 23:11 < GunqqerFriithian> a lot of stuff is like that, technically unconstitutional but it hasn't been challenged ever 23:11 < rascul> Psi-Jack you just can't explain to others what it means to be a mormon ;) 23:11 < Psi-Jack> heh 23:11 < phogg> oo_miguel: you have two choices: two names for one file, or two names for different files. They can't be both the same file and a different file 23:12 < `Koyaanisqatsi> the first amendment has limitations, it all depends on the interpretation 23:12 < sauvin> "This girl here... um... wait... is she my wife, my niece or my daughter? I forget...." 23:12 < lrvick> hexnewbie: this is a super interesting use of bind mounts. I will for sure be playing with this. thanks! 23:12 < NGGJamie> Clearly they would have to use some wonked out interpretation, like under what circumstances you can tell someone about mormonism. Getting rid of the door-knocking via some law would be the easiest way 23:13 < NGGJamie> Essentially confining them to public spaces 23:13 < oo_miguel> phogg: so there seem to be no solution for my requirement. 23:13 < sauvin> Mormons don't go door-knocking that I know of. 23:13 < Evidlo> hexnewbie (IRC): anyway, both PyUserInput and pynput can't handle ⏎ 23:13 < NGGJamie> I've had it happen to me a couple times. 23:13 < Psi-Jack> Hmmm. I wonder... Does SwagArch and/or Anarchy have their own overlay reposititories? 23:13 < phogg> oo_miguel: your requirement is nonsensical. WHat are you tryiung to do? 23:13 < GunqqerFriithian> if anyone knocks at my door with out my damn package Imma be mad 23:13 < Psi-Jack> sauvin: Actually they do 23:13 < Psi-Jack> sauvin: Missionaries. 23:14 < NGGJamie> Couple teenagers in nice clothes with tucked in polos, ties, the works. Asked to come in and talk about being mormon and invited me to their church. I just shoo'd them off as politely and quickly as possible and went about my day. 23:14 < sauvin> Oo... missionaries! I heard a rumour they go GREAT with a basil asiago sauce! 23:14 < Psi-Jack> lol 23:14 < GunqqerFriithian> What you do is out jesus them 23:15 < GunqqerFriithian> make em think you're the crazy one 23:15 < sauvin> No, what I DO is answer the door wearing nothing but a smile. 23:15 < hexnewbie> Evidlo: Is that Python 3.3 or newer? 23:15 < GunqqerFriithian> sauvin that's how you get on a registry 23:15 < NGGJamie> sauvin, if they're teenagers like I encountered that's how you answer the door to the police a bit later 23:15 < oo_miguel> phogg: I wanted to put some config files in my gitrepo but strip out some variables as passwords (replacing them with placeholders). Having done this I wanted to create links to the git versions of this files, filling the sensitive information back in the stuff via sed or something similar. 23:15 < rascul> Psi-Jack correction, explaining polygamy is illegal, not mormonism https://law.justia.com/codes/mississippi/2010/title-97/29/97-29-43/ 23:16 < oo_miguel> phogg: s/the/stuff// 23:16 < Psi-Jack> rascul: Ahhhh... That's different. 23:16 < rascul> indeed 23:16 < phogg> oo_miguel: Definitely cannot be done. You need instead to copy the files out of git and modify them during the transfer or afterwards. 23:16 < Evidlo> hexnewbie (IRC): I'm on 3.6.5 23:16 < rascul> i had misremembered 23:16 < phogg> oo_miguel: you can script that, though, so it can become automatic 23:16 < sauvin> GunqqerFriithian, NGGJamie, it's exactly what I did for the JW crowd. I just don't open the door all the way, and I explain why. They tend to leave rather quickly. 23:16 < NGGJamie> rascul, the minimum fine of $25 tells me that's a super old law 23:16 < Psi-Jack> rascul: Remembered by association and the polygamy situation that was labeled "mormons", yet weren't mormons. 23:17 < rascul> mormons aren't even polygamists anymore, that officially stopped before any of us were born 23:17 < Psi-Jack> Yep. 23:17 < GunqqerFriithian> sauvin PRAISE THE LORD JESUS CHRIST! Instead of them being the crazy religous nuts you be it so they get off your case and never, ever come back 23:17 < `Koyaanisqatsi> most religious salesmen will leave you alone if you tell them you're agnostic 23:18 < phogg> rascul: the US government told them to stop, then suddenly god told their leaders to stop. Mysterious! 23:18 < GunqqerFriithian> I praise satain daily 23:18 < `Koyaanisqatsi> "maybe there is a god but I don't believe in religion" 23:18 < Psi-Jack> I once had the mormons come to my door, and greet me by my name, before I even offered it. LOl 23:18 < Dagmar> They also don't come back if you answer the door naked and covered in sweat (among other things) 23:18 < sauvin> ALL of them leave me alone when I can't speak any English. :> 23:18 < rascul> phogg funny how that works sometimes 23:18 < GunqqerFriithian> no especo englise 23:18 < twainwek> Psi-Jack: this is ##linux related chat. Please stay on topic 23:18 < oo_miguel> phogg: I will have a look at git hooks then. Or maybe some watchdogs that will recreate the resulting files as soon as the "templates" change 23:18 < Psi-Jack> I was rather pissed, actually, because my mother specifically was told never to tell her church my location. So the church actually specifically looked me up and sent their minions. 23:18 < sauvin> twainwek, why don't you let me decide when the off-topic has gone too far off? 23:18 < NGGJamie> twainwek, to be fair, the entire channel has kinda railroaded a little bit. Probably should try to get back on 23:19 < rascul> choo choo 23:19 < Psi-Jack> sauvin: Heh. twainwek at something again? 23:19 < GunqqerFriithian> most chats should be able to go off track a bit 23:19 < phogg> oo_miguel: hooks is probably easiest, but watching with inotify might suit you too 23:19 < oo_miguel> phogg: Also I still do not think that my initial idea is completely nonsensical.. (craeting deriviate files from existing ones) 23:19 < Psi-Jack> Ahh... I see. 23:20 < Psi-Jack> Yeah, twainwek seems to constantly target me, randomly, pointlessly, noisilly. 23:20 < phogg> oo_miguel: it would require a custom filesystem (via fuse) or otherwise cooporation from the filesystem layer 23:20 < hexnewbie> Evidlo: Hm, Python 3.3+ is fully UTF-32 so no troubles with non-BMP characters, and ‘⏎’ is BMP. Maybe that library literally presses keyboard keys? There's no keyboard key for emojis. 23:20 < oo_miguel> phogg: hopes something like this might be supported on filesystem level 23:20 < Psi-Jack> Had to check znc logs to see. :) 23:20 < twainwek> my favorite thing is when someone pretends to have ignored someone on irc, but doesn't actually do it 23:20 < phogg> oo_miguel: maybe on the HURD 23:20 < GunqqerFriithian> I've done that so many times, easy way to piss peopel off 23:20 < oo_miguel> phogg: yeah I agree. my idea/hope would require some fs level functionaltiy 23:20 < Dagmar> hexnewbie: Give it time. There's keyboards for traditional Chinese 23:21 < Dagmar> Someone will make an emoji keyboard if they haven't already 23:21 < oo_miguel> phogg: I guess I will stick to hooks/inofify. thank you for your support 23:21 < Evidlo> hexnewbie (IRC): I think it tries to map them to keysyms and uses xlib 23:21 < `Koyaanisqatsi> we won't need keyboards in the matrix, and it will run linux 23:22 < hexnewbie> Dagmar: There are? I mean, I have played with CJK input methods for awhile, but my impression is that they all used complex input methods, not a literal keys for the characters. But I could be wrong. 23:22 < Dagmar> hexnewbie: It's always complex, but there's a company that makes a 600-key keyboard for that 23:23 < `Koyaanisqatsi> hexnewbie, you are right 23:23 < hexnewbie> (It's very funny experience to convince someone studying Japanese that they can install Slackware on their computer, and everything will work fine for them on it) 23:23 < NGGJamie> hexnewbie, That seems super specific 23:23 < Dagmar> hexnewbie: see also https://www.reddit.com/r/MechanicalKeyboards/comments/3l63qr/if_made_mechanical_i_imagine_this_chinese/ 23:25 < Psi-Jack> But yeah, soon as they addressed me by name, those mormons. I told them pretty much very clearly that they have 30 seconds to vacate my property or the cops would be out. Whom lived literally eye-shot distance away accross the street (the police station) 23:25 < Psi-Jack> Never seen a mormon at my door since. :D 23:26 < Evidlo> hexnewbie (IRC): Is it not related to the fact that my character is not in /usr/include/X11/keysymdef.h 23:26 < hexnewbie> NGGJamie: Well, yeah. I then had to then make it work. Also how I discovered that switching to an UTF-8 locale fixes G_BROKEN_FILENAMES (along with fixing kinput2 for said student of Japanese) 23:26 < Dagmar> Somehwat obviously I didn't have police anywhere nearby 23:27 < Dagmar> ...but the people who live there now tell me the missionaries cross the street rather than even walk in front of the house, just in case Satan comes back 23:29 < hexnewbie> Evidlo: Very possible. If you're trying to pass strings by keysyms, your choice may be very limited. Have you considered trying to (programmatically) place it in CLIPBOARD and pasting it by sending Ctrl+V or Shift+Insert to the program? I know that sounds like a bad idea, but I've seen at least two popular programs do it that way. 23:29 < hexnewbie> Evidlo: Or better yet, reconsider what you're doing. WHy are you trying to send symbols to programs using keypresses? 23:30 < Evidlo> hexnewbie (IRC): its a CLI password manager and I need to be able to send usernames/passwords to any sort of program 23:31 < hexnewbie> So, KDE 4's paste widget does it - it generates a password, and pastes is in another program. Have no idea if it pastes it using Ctrl-V, Shift-Insert or middle button. 23:31 < NGGJamie> Most "universal" password managers of the sort just use the clipboard method 23:31 < hexnewbie> Evidlo: Then stop using non-ASCII in your passwords 23:31 < NGGJamie> Also yeah using non-ASCII in passwords sounds like a bad idea. 23:31 < GunqqerFriithian> ^ 23:32 < Evidlo> someone submitted an issue with me and I'm trying to figure out how to accomodate them 23:32 < phogg> hexnewbie: it could also generate fake keystrokes. 23:32 < NGGJamie> You never know how the hashing algorithm/server on the other side might handle that sort of input, because I doubt many developers really considered it. 23:33 < Evidlo> using clipboard sounds possible, but possibly dangerous 23:33 < hexnewbie> In Unicode's defence, with PHP's popularity on the Internet, you may have more trouble with \, ' and " than with the weirdest Unicode codepoints. 23:34 < GunqqerFriithian> depends how paranoid you are, as long as your computer is secured (encrypted drive, etc) and you lock your computer when you leave you could use clipboard but depending on what your clipboard remembers it could be more or less dangerous 23:34 < NGGJamie> I'm moreso concerned with hashing issues because of what implications that might have against the hash algorithm, depending on which one is being employed and how. 23:35 < motte> whats the best cloud software to run on a home server? 23:35 < NGGJamie> Instead of storing permanently on the clipboard, either save the clipboard contents, replace them, then put them back when you're done, or on a timer, or just clear it after X time 23:35 < NGGJamie> motte, For what particular purpose? 23:35 < Evidlo> I guess I'll just add an option that copies the username/password to the clipboard 23:35 < hexnewbie> I mean, if you try using a\bcdef as password on a modern web site, your password often ends up being either a\\bcdef, abcdef or acdef. 23:36 < Loshki> I used to have a password with a space in it. Guess how many ways that breaks stuff... 23:36 < revel> hexnewbie: So, you could use that as a possibility to have a blank password? :D 23:37 < Evidlo> but there's probably no way to trigger a paste, right? 23:37 < GunqqerFriithian> you can have an email with anthing in it as long as it's in quotes 23:37 < NGGJamie> Evidlo, I have no experience in doing it on linux, but what about simulating keypresses? 23:38 < Evidlo> that wont work because there's no way to generate the character I'm after with keypresses 23:39 < motte> NGGJamie: i'd like to be able to sync local directories with the cloud service, i tried seafile but that doesn't work too well 23:39 < NGGJamie> No, like Ctrl+V or so 23:39 < rascul> motte what issues did you have with seafile? 23:40 < Evidlo> I think that's application specific 23:40 < motte> rascul: i want to import a large directory of media files, but seafile uses blocks so it makes a copy of everything which doubles the disk space used 23:40 < rascul> oh 23:41 < NGGJamie> I'm a pretty simple guy and just used Samba when I wanted file syncing, though that's probably not what you're after. 23:41 < motte> there seems to be an option for extension for FUSE but i can't figure out if i can use it like that 23:41 < NGGJamie> I'm a stickler about not liking to install new software 23:41 < motte> i use samba and sshfs, the cloud service would mostly be for my girlfriend 23:41 < likcoras> Evidlo: haven't read the backlog, but have you looked into stuff like xdotool? 23:41 < rascul> motte have you looked at nextcloud? 23:42 < likcoras> It simulates keypresses, so it can type stuff into websites/applications/etc. 23:42 < motte> rascul: does it support importing local directories? 23:42 < rascul> motte iirc yes 23:42 < Evidlo> I suppose I can fall back to that. I know xdotool isn't available everywhere 23:43 < T-Rog> So ever since doing a dup from OpenSUSE 42.3 to 15 I get this when trying to run BG:EE 23:43 < T-Rog> https://pastebin.com/h839wk2d 23:45 < rascul> T-Rog might be you need to update/reinstall/reconfigure your graphics driver stuff 23:45 < hexnewbie> Evidlo: Why not just stick to ASCII characters? 23:45 < T-Rog> and Diablo II looks completely fucked up in Wine 23:46 < NGGJamie> hexnewbie, From what he said earlier, it's a client asking for the functionality. 23:46 < hexnewbie> I apologise 23:47 < alexandre9099> hi, can i log commands runned inside a chroot? 23:47 < alexandre9099> (in a way that the chroot user can't delete it) 23:47 < NGGJamie> I don't blame you for not seeing, it was before we jumped onto the religion track 23:47 < GunqqerFriithian> lol choo choo 23:47 < hexnewbie> alexandre9099: No. Also chroot is not a security measure. But you can look into auditd. 23:47 < annihilator> would qemu be better from source or repo? 23:47 < lnnb> alexandre9099: send it through a pty and log all input 23:47 < hexnewbie> alexandre9099: security measure on its own, I mean 23:47 < Evidlo> I dont understand why clipboard hacks are necessary. I can paste unicode fine, so why can't xlib provide a mechanism for handling unicode? 23:48 < rascul> chroot is just a fancy cd ;) 23:48 < GunqqerFriithian> but I don't have an optical drive? 23:49 < alexandre9099> isn't chroot running as a "no permissions" user be safe enough? (using ssh's ChrootDirectory) 23:49 < alexandre9099> from what i read, as root, it is possible to break the chroot 23:49 < NGGJamie> If you're root, sure 23:49 < lnnb> depends how ssh implements chroot call 23:50 < Evidlo> this document indicates that xlib should support some of this already: http://www.cl.cam.ac.uk/~mgk25/ucs/xorg.html 23:50 < alexandre9099> lnnb, what is a "good" chroot implementation? 23:52 < T-Rog> well I fixed my baldur's gate issue 23:52 < lnnb> it depends on too many variable to summarize quickly. basically it's a minefield that you trust ssh devs to have cleared 23:52 < alexandre9099> heading to #openssh 23:52 < lnnb> but with no privileges it's less trouble. most devs are smart enough to change CWD at least 23:52 < lnnb> which is the easiest way to escape 23:53 < Loshki> Have bsd jails still not been ported? 23:53 < alexandre9099> lnnb, but without root it would not be possible, right? 23:54 < lnnb> depends on too many variables 23:54 < lnnb> to say for certain 23:56 < Dagmar> The sshd devs _have_ cleared that minefield. 23:57 < Dagmar> With respect to what's a "good" chroot implementation, the answer is chroot(). 23:57 < Dagmar> The challenge that remains beyond that is manifold: 23:58 < Dagmar> 1. The chroot() must be into a directory that is the root of a _separate filesystem_. Make one with dd mkfs and mount it via loopback if you have to. 23:58 < KeyboardNotFound> Is there a way to delete some screen resolution from showing in "Select resolution menu" ? 23:58 < Dagmar> 2. Generally the thing should be _empty_ but for the chroot to be secure, everything that's going to be running in it _also_ has to be bound within the chroot, which is why one has to make those special devices like /dev/null and so on --- Log closed Thu Jun 07 00:00:06 2018