IRC Logs for #crux Thursday, 2014-07-31

*** sh[4]rm4 has joined #crux00:01
*** jdolan has joined #crux00:02
*** SiFuh_ has quit IRC00:03
*** sh4rm4 has quit IRC00:03
*** tilman has quit IRC00:03
*** tilman has joined #crux00:05
*** c0x has quit IRC00:11
retardjaeger: i have used drbd pretty extensively00:13
retardtimcowchip: i was just struggling with building firefox myself, the build consumed at least 3 gigs of ram and about 7 gigs of disk space00:16
*** c0x has joined #crux00:16
Worksteri have suggested DRBD but i haven't used it myself00:17
retardi like it00:18
*** sh[4]rm4 has quit IRC00:19
Worksteri am using mdadm + lvm2 though00:20
retardi use drbd on top of that00:21
Worksteruploading firefox
retardi finally went for it and got a package building chroot with a pkg-get repo up and running for myself and some friends00:23
jaegerretard: ever have any problems with it randomly losing connection to the other node without apparent networking troubles? PingAck timeout messages00:25
Worksteri'm only offering them large compile time ports.00:25
retardjaeger: no, and i've abused it pretty badly, running 15+ volumes at the same time over a gigabit link.. what are the nics and switches like?00:26
retardi struggled pretty badly when i was on some shitty desktop onboard stuff00:26
retarddon't notice how shitty those can be until you stress them a little00:26
jaegerIt's odd since the cluster stack never blinks but drbd sometimes fails00:27
retardwell, i'd assume drbd is what stresses the network the most00:28
jaegerVery low load in this case, clustered postgresql server without high traffic00:29
jaegerVirtualized servers (vSphere 5.5), I switched the NICs from vmxnet3 to e1000 this evening to see if it dies any time tonight00:30
retardthe only time i had a problem like that on decent hardware i had misconfigured something00:30
retardtrying to remember what it was00:30
retardwhich protocol are you running?00:30
retardthe least forgiving one :]00:31
retardi have never ran drbd between VMs00:31
pidsley-afkWorster -- is the method of building FF on CRUX significantly better than simply downloading the binary from Mozilla?00:31
retardmy coworkers who attempted it didn't seem too enthused00:32
jaegerThis is the first time I've used it, C seemed to be a common suggestion in the docs. I'll give another protocol a try if the NIC change has no effect00:32
Worksterworkster... tab completion00:32
retardC is the most paranoid mode00:32
pidsley-afkyeah thanks ;)00:32
retardalso slowest00:32
Worksterpidsley-afk, well ours uses system libs where the nightly does not so much00:32
jaegerThe networking between the two VMs is 1GbE and both hosts are connected to the same physical switches00:32
jaegeryeah, paranoid but slowest seemed like an OK choice for this when I was researching00:33
retardi'm concerned with the virtualized devices00:33
pidsley-afkI don't use the nightly, just the majors00:33
Worksterso expect a prebuilt firefox off mozilla to consume more ram00:33
jaegerwhy? Far more demanding workloads run virtualized than this :)00:33
jaeger90% of my infrastructure is virtual, has been for years00:33
Worksterany off the mozilla site infact wont use system libs except for glibc00:33
Worksterand maybe a few others..00:34
jaegerThe host hardware are all dell/oracle servers with dual processors, 8-12 cores each, and 128 or 155 GB RAM00:34
pidsley-afkthat's OK with me, considering the prebuilts use 60% of the memory it was using on Arch00:34
pidsley-afkit just gets annoying because it's such a major build, and they update so often00:35
jaegerstupid fingers00:35
retardvm clustering is what i use drbd for00:35
Worksterpidsley-afk, then use my prebuilt one for crux if you wish00:35
pidsley-afkthank you00:35
Workstersee if it's any faster/smoother/better. be keen to see if there is any real difference00:36
Workster46% uploaded00:36
pidsley-afkit's hard to tell with FF, because people use it so differently -- some have 30 tabs open at a time00:37
pidsley-afkwhat they are doing, I have no idea00:37
*** jdolan has quit IRC00:38
pidsleywhen i used gentoo, i just held FF and didn't update it00:38
*** jdolan has joined #crux00:38
retardjaeger: drbd is specifically designed the way it is in order to efficiently interface with the network stack00:38
retardit is very much built for running on actual hardware00:39
retardthat's the whole reason for it being in kernel00:41
jaegerno doubt, but the virtual stack is good enough for other similar things in my experience00:41
*** DaViruz has quit IRC00:41
retardi wouldn't run drbd inside vms in production, and i'd stick with  protocol a00:42
jaegerI think it's a bug in ubuntu server 12.04/kernel 3.whatever in this case but not sure00:42
retardi have the same experience, but drbd is.. a very special case00:42
retardabstracting the hardware away from it will definitely have adverse effects00:43
retarddeep in network stack/block device mojo :)00:43
jaegerI considered that, actually, but it's also not hard to find people doing the same thing with VMs via some searching :/00:44
retardparticularly in this case, because any trouble will be amplified00:44
jaegerso I'm not sure00:44
jaegergoing to keep playing with it00:44
retardon account of it eing synchronous00:44
retardi consider drbd something that belongs "at the bottom"00:44
*** jdolan has quit IRC00:44
Worksteranything from debian that cherypicks commits is gonna end up creating a problem that upstream does not have.00:44
Worksterwhich is why crux does not do this unless it's serious breakage or security related.00:45
jaegerIt'd be a huge waste for me to dedicate two servers to this, heh. That would be sad00:45
retardin the setups where i use drbd i use the drbd volumes as backing devices for virtual machines00:46
retardrunning on top of lvm running on top of mdraid00:47
retardno live migration, but in these cases there is no need00:47
retardi run some fairly huge xfs volumes over it with good results00:48
retardselectively chosen server grade hardware to keep stability up and price down, has worked out really well00:49
retardi love those wonderful ecc supporting amd desktop cpus00:49
retardif you have the possibility of replication at a lower level i would avoid drbd00:51
jaegerIf drbd completel fails me in this case I could fall back to vsphere FT. Just didn't want to do that as it has some strict limitations00:51
retardplease do not hold its performance in a virtualized environment against it though00:51
jaegerI kinda do, just because everything else I've experienced in a virtual environment has been rock solid. I do get, though, that is isn't really intended use00:52
retardit is really nifty, it just has a fairly narrow set of circumstances00:52
retardwhere it's appropriate00:52
retardthe appropriate use in this case would be setting up drbd on the actual virtualization servers00:53
retardbut then that probably does not make sense in your setup00:53
jaegerwhich isn't an option in this case, yeah00:53
*** jdolan has joined #crux00:53
retardi have actually used a passive drbd layer in some standalone servers00:54
retardletting me migrate vms out with minimal downtime00:54
retardjust replicate out, shut down on the old and boot up on the new hardware00:54
*** DaViruz has joined #crux00:55
retardjaeger: generally i agree very much with you, though, there are a lot of myths around applications being unfit for virtualization00:57
retardclassically databases00:57
retardwhich is fud00:57
jaegerSome of it was true, technically. Not for years, though.00:57
jaegerI was at VMworld last year, saw some insane Oracle deployments 100% virtualized00:58
retardsure, it's just kinda.. lingered00:58
jaegerall this reminds me, though, that I tested that same drbd postgresql cluster setup in VMs here at home and didn't see that behavior00:59
jaegerfar smaller deployment, only one host. less network traversal00:59
retardi run it over separate nics, and have actually done cabling directly between servers for it01:02
jaegeryeah. I could also do that but then I lose lots of the benefits of virtualization01:03
retardthe combination of drbd and virtualization has allowed me to GREATLY increase the time i get to spend without pants on01:03
retardwhich is directly proportional to the time i am truly happy01:03
retardbut for more than two nodes i would of course want to go with some san stuff01:04
retardhooray multipath hell :)01:04
retarddrbd is just... cheap01:04
jaegernow on to more important stuff... upgrading my minecraft server to 3.101:06
retardkeeping your priorities straight are important01:06
jaegerI worked for 11 hours today, heh01:07
retardi am sorry01:07
jaegernot too bad, I work 4x1001:07
joacimI hate working 7.5 hours :/01:07
jaegerFridays off, every week01:07
retard4x5 sounds better01:09
retardmaaaybe 5x401:12
*** jdolan has quit IRC01:35
*** jdolan has joined #crux01:38
*** mavrick61 has quit IRC02:43
*** mavrick61 has joined #crux02:44
*** SiFuh has joined #crux02:46
*** timcowchip has quit IRC02:47
*** diverse has joined #crux02:52
*** jdolan has quit IRC03:11
*** Rotwang has quit IRC04:34
*** pidsley has quit IRC05:56
*** c0x has quit IRC06:58
*** c0x has joined #crux06:58
*** arduo has quit IRC07:35
*** timcowchip has joined #crux07:41
timcowchipthis is interesting, firefox#31.0-1.pkg.tar.xz here is 54Mb07:43
timcowchipwhile firefox#31.0-1.pkg.tar.xz here is 71Mb07:44
timcowchipI wonder what the 17Mb diff is all about07:45
timcowchipassuming they wre both made from the same port07:47
*** timcowchip has left #crux (" - Chat comfortably. Anywhere.")07:49
frinnstis is interesting, firefox#31.0-1.pkg.tar.xz here is 54Mb       Romster07:55
frinnstoops, sorry07:55
frinnstwrong clipboard07:55
frinnstThe Virtual Disk service terminated unexpectedly.  It has done this 1 time(s).  The following corrective action will be taken in 60000 milliseconds: Restart the service.07:55
frinnsthow informative07:55
joacimI had this magic utility that synchronised teh different clipboards07:57
*** Rolinh has joined #crux07:58
*** Rolinh has quit IRC08:44
*** q5sys has quit IRC09:01
joacimjust something i had in my .xinitrc09:37
frinnstI find two separate clipboards to be quite useful09:45
frinnstwell, most of the time anyways :)09:45
teK_jaeger: are you using  Fujitsu Servers/Storage, by chance?10:04
frinnstI think they use mostly sun storage10:20
*** SiFuh has quit IRC12:02
*** dkoby has joined #crux12:06
*** SiFuh has joined #crux12:11
*** rexich has joined #crux12:17
*** sh4rm4 has joined #crux12:29
*** SiFuh has quit IRC12:33
*** rexich has quit IRC12:37
*** SiFuh has joined #crux12:37
*** rexich has joined #crux12:38
*** jdolan has joined #crux13:10
*** jdolan has quit IRC13:11
*** jdolan_ has joined #crux13:14
*** dedmanwlk has quit IRC13:15
*** nwe has quit IRC13:15
*** jaeger has quit IRC13:15
*** syncn has quit IRC13:15
*** nwe has joined #crux13:15
*** syncn has joined #crux13:15
*** dedmanwlk has joined #crux13:16
*** jaeger has joined #crux13:16
*** Romster has quit IRC13:16
jaegerteK_: guessing I got disconnected before my answer went through.13:19
jaeger08:14 <@jaeger> teK_: no, sorry13:19
jaeger08:15 <@jaeger> Oracle/Dell servers, Oracle/Isilon storage13:19
*** rexich has quit IRC13:27
*** Romster has joined #crux13:43
*** SiFuh has quit IRC14:00
*** SiFuh has joined #crux14:18
*** rexich has joined #crux14:18
*** u7nv9h has joined #crux14:36
u7nv9hhow do i make my laptop suspend when the lid is closed?14:45
*** SiFuh_ has joined #crux14:47
jaegeru7nv9h: acpid plus pm-utils is probably easiest14:48
*** SiFuh has quit IRC14:49
u7nv9h can the third answer work?14:50
jaegerIt might be ubuntu-specific, don't know14:51
*** SiFuh_ has quit IRC14:51
jaegeryou can always try it and see14:52
*** ph4n70m4s has joined #crux14:52
*** phant0mas has quit IRC14:55
jaegerI use acpi event scripts for that. <-- /etc/acpi/events/default14:55
jaeger <-- /etc/acpi/handler14:55
u7nv9hpm-suspend is a part of pm-utils?14:58
u7nv9his pm-suspend a part of pm-utils?*14:58
u7nv9hdefault and handler needs to chmod?15:01
jaegerdefault is 0644, handler needs to be executable (0755 or 0700 perhaps)15:05
u7nv9hchmod a+x default && chmod a+x handler ?15:06
jaegerdefault doesn't need x but if you want to, go ahead15:08
u7nv9hoh, sorry, i didn't notice the "default is 0644"15:09
u7nv9hnow i need to restart or will it work without reboot?15:10
jaegeryou should only need to restart the acpid service15:23
u7nv9hhow? and about services, how do i add one?15:24
jaegeradd "acpid" to the services=() array in /etc/rc.conf15:24
jaegerthen /etc/rc.d/acpid restart15:24
u7nv9hbash: /etc/rc.d/acpid: No such file or directory15:25
jaegerdid you install acpid?15:30
u7nv9hdamn, no, i installed acpi instead15:32
u7nv9hnow i install it15:32
u7nv9hinstalled acpid and restarted the service, but still nothing, it doesn't suspend15:34
u7nv9hbut in /etc/acpi now there are two new files, and events15:35
jaegerremove those and use the custom ones instead15:36
u7nv9hstill nothing15:38
u7nv9hso, i have the custom default in /etc/acpi/events, handler in /etc/acpi and i deleted the that was in /etc/acpi15:39
jaegerdid you restart acpid after deleting those files?15:41
jaegerdo you see anything in the system logs when you try to trigger acpi events?15:41
*** u7nv9h_ has joined #crux15:42
u7nv9h_oh, totally forgot about restarting, all works, thanks!15:43
*** u7nv9h has quit IRC15:44
u7nv9h_if i wanted to add aa service that starts mpd, clears the playlist (mpc clear) and starts mpdscribble what should i do?15:44
jaegernot a clue there, I don't use mpd or mpc15:45
u7nv9h_ah, last question, how do i set xrandr permanently? i have a 1366x768 display, but x starts everytime in 1368x768 and i have to set 1366x768 manually, is there a way to set 1366x768 permanently'15:47
u7nv9h_(sorry for the bother)15:47
jaegerrun the xrandr command in your .xinitrc probably15:49
*** dkoby has quit IRC15:49
*** u7nv9h_ has quit IRC15:53
*** u7nv9h has joined #crux15:55
u7nv9hi really don't know how to use it, i tried to read the --help and search something on the net, but i really can't understand how do i set it (i use arandr, a graphical frontend)15:56
u7nv9hwhen i run 'xrandr --output LVDS --mode 1366x768' i get15:58
u7nv9hwarning: output LVDS not found; ignoring15:59
u7nv9hxrandr: Failed to get size of gamma for output default15:59
jaegerrun 'xrandr' by itself to see your outputs16:00
*** Rotwang has joined #crux16:00
jaeger$ xrandr | grep " connected"16:00
jaegerDVI-I-1 connected 1920x1200+0+0 (normal left inverted right x axis y axis) 519mm x 324mm16:00
jaegerDP-0 connected 1920x1200+1920+0 (normal left inverted right x axis y axis) 519mm x 324mm16:00
jaegerso I could use DVI-I-1 or DP-016:00
u7nv9hxrandr: Failed to get size of gamma for output default16:01
u7nv9hdefault connected 1366x768+0+0 0mm x 0mm16:01
u7nv9hnow works16:11
u7nv9hthanks for all the help16:11
*** jdolan_ has quit IRC16:20
*** u7nv9h has quit IRC16:26
*** rexich has quit IRC16:45
*** deus_ex has joined #crux17:40
deus_excongrats on 3.1, devs, great release.17:42
deus_exi see crux is gnu/linux now ;)17:43
deus_exquestion: when nvidia-uvm module loads, device node in /dev is not created.problem with udev rule, or?17:45
deus_exif I reload nvidia-uvm after boot (modprobe -r, modprobe), device node is created, but not on boot.what am I missing here?17:48
*** rexich has joined #crux17:52
*** u7knv9h has joined #crux18:11
u7knv9hmy webcam doesn't work, and i think i have the support compiled in the kernel (Device drivers -> Multimedia support -> Cameras video grabbers support, right?), should i install some packages in order to get it working or i missed something in the kernel?18:12
rexichsee if you've enabled support for UVC video in the kernel (at the same place as Cameras you've mentioned)18:14
rexichwithout drivers enabled, having the support enabled does nothing :D18:14
rexichand the grabbers are for TV cards, not for web cameras18:14
u7knv9hthanks! anyway, i can't find "UVC video" in the same place, may it be in another place?18:15
rexichjust a sec18:16
deus_exu7knv9h: Media USB Addapters section18:17
u7knv9hmust i enable just UVC or the others too?18:18
rexichDevice Drivers -> Multimedia support -> Media USB Adapters (MEDIA_USB_SUPPORT [=y]) -> USB Video Class (UVC)18:18
rexichfirst, is it a laptop webcam?18:18
rexichintegrated in a laptop?18:19
rexichwhich laptop is in question?18:20
rexichis it less than five years since you've bought it?18:20
rexichbecause most of the cameras today are UVC, the rest are custom chips, codecs and drivers18:21
u7knv9hi bought it almost 2 years ago18:22
rexichthen we can safely assume it's UVC. enable UVC video support and you're good to go. :)18:22
u7knv9hanyway thanks! now i try with uvc only and if i have issuess i'll try the rest18:22
rexichno problem :)18:22
deus_exsince eudev complains 'input group not found', add one to /etc/group or remove mentions of the group from 50-default.rules?18:35
rexichdidn't some updates to CRUX 3.1 add the "input" group? do "rejmerge"18:37
u7knv9hah, a little question, do i need to add the user to a group in order to use the webcam (like with audio)?18:38
rexichyes, add him to the video group18:46
rexichhim -> *the user18:46
u7knv9hok, thanks!18:48
*** u7knv9h has left #crux ()18:48
deus_exrexich: you are right, there is input group in core/filesystem group file.18:53
deus_exi forgot to merge it with the existing one, it seems18:54
deus_exrexich: thanks for the tip.18:54
rexichno problem. remember to do rejmerge after updates, just in case.18:55
deus_exrexich: i did, I missed that one :)18:57
tilmanyou might want to add $user to the tty group, too if you want to use /dev/pts18:58
rexichyes, don't forget that or you may have trouble with GUI terminal emulators18:58
pedja'udevadm monitor' is for debugging udev rules, right?it's been a while.19:01
rexichpedja: jesi li iz ex-yu prostora? :D19:04
pedjarexich: da :)19:05
rexichpedja: mali je svet :)19:05
pedjarexich: vala bas :)19:06
rexichpedja: raduje me što ima nas CRUX korisnika i na ex-yu prostorima. baš kul. :D19:07
pedjarexich: jos od 1.0 verzije, ili tako nekako.19:08
*** timcowchip has joined #crux19:08
rexichpedja: super, ja od 2.8. draži mi je CRUX or Arch-a, nema systemd ovde i interesantnije je za konfiguraciju, da ne pominjem što tera korisnika na neku vrstu discipline i razvija veštine samoobuke :)19:10
*** timcowchip has left #crux (" - Chat comfortably. Anywhere.")21:11
*** timcowchip has joined #crux21:15
timcowchipappend to /etc/profile21:16
timcowchipgets rid off the chromium "Google API keys are missing" message21:16
timcowchipnp :)21:19
*** timcowchip has left #crux (" - Chat comfortably. Anywhere.")21:24
*** timcowchip has joined #crux21:30
*** Pingax has quit IRC21:57
*** SiFuh has joined #crux22:14
*** timcowchip has quit IRC23:17
*** timcowchip has joined #crux23:17
*** SiFuh has quit IRC23:23
*** SiFuh has joined #crux23:24
*** SiFuh has quit IRC23:29
*** SiFuh has joined #crux23:30
*** vlnx has quit IRC23:31
*** vlnx has joined #crux23:32
*** SiFuh has quit IRC23:35
*** SiFuh has joined #crux23:38
*** SiFuh has quit IRC23:43

Generated by 2.11.0 by Marius Gedminas - find it at!