IRC Logs for #crux-devel Wednesday, 2018-03-07

*** pedja has joined #crux-devel00:09
*** _________mavric6 has quit IRC03:25
*** _________mavric6 has joined #crux-devel03:27
*** dlcusa has quit IRC03:37
*** j_v has joined #crux-devel07:22
*** john_cephalopoda has joined #crux-devel09:06
*** fun- has quit IRC10:19
*** fun has joined #crux-devel10:19
*** groovy2shoes has joined #crux-devel10:38
*** pedja has quit IRC12:33
*** pedja has joined #crux-devel13:11
pedjathis can't be related to xorg freezing, surely? http://dpaste.com/1W7YYND13:13
pedjaI removed xorg's libinput driver and switched to evdev, see if it happens again13:15
j_vpedja: I've never seen that error in my xorg logs (even xsession output that I save from startx stdout/stderr output) and I've been using evdev with no libinput for long time.13:34
pedjaj_v, that's why I switched back to evdev13:35
j_vahhh, sometimes I wonder if I ever actually read what I read... "the light are on, but no one's home" symptoms I have sometimes is just incredible13:36
pedja-ENOTENOUGHCOFEE :)13:37
j_vfprintf("%s: you are not operating at optimal percentage, use a bigger coffe mug", strerror(ENOTENOUGHCOFFE));13:39
j_vsee, now I know it's true. I forgot it have 'stderr' before the print format string.13:40
j_vI'm skeptical of mouse debouce errors locking up xorg. Have you checked dmesg for drm driver errors around time or approaching time of freeze?13:42
pedjanothing in both 'var/log/kernel' or messages, hence my confusion :)13:44
j_vdunno, sorry. I know I'm holding my breath for an upcoming patch that may (wishful thinking perhaps) help an issue I've been seeing with nouveau.13:46
pedjainterestingly enough, when xorg froze and HD started grinding, the only thing that worked was the mouse :)13:47
pedjacomplete lockup, apart from it13:48
j_vHD grinding... kinda makes me wonder what's behind that.13:48
j_vcan't even ssh in from another device?13:48
pedjadon't have another device13:48
j_vdamn13:49
j_vthat's the kind of thing I have acpid installed for and set up the power button to do a normal shutdown so the disks don't get trashed. but if the kernel itself is lockup, that probably won't help either.13:51
pedjaI do have a netbook, but it's broken. waiting for nuke and install of...something. new Leap maybe13:51
j_vthat's the new suse thingy, right? i've never run suse, though my wife had to for a job a while back.13:52
pedjathat's their LTS. Tumbleweed is rolling release.13:55
pedjaluckily, I still have TW snapshot on a usb key, came in handy last night when ext4 ate some of mine root dir :)13:57
pedjagrub from '/boot', some stuff from root's home, some kernel modules13:58
j_vwow, I've never seen ext4 do that. but I've read bug reports of something like it, but man, not cool when the fs itself eats your data.13:59
pedjait happens from time to time.sometime it eats a bit less :)14:00
pedjaI keep of backup of a few system directories for the occasion14:00
j_vI've mostly used ext4 for a long time and it's actually saved me when power loss and ups took a crap14:01
j_vweird how it can be so different14:02
pedjanot in my experience. other, JFS filesystems, were, so far, much more robust. ext4 just craps itself :)14:05
pedjaI'll have to check what FS can grub2 use as a root FS, and migrate to one of them14:06
pedjaI wish I could migrate all of it to, say, XFS, but can't atm. being unemployed sucks big time14:07
j_vyeah, you and me both14:08
j_vwhat little $ I have go out pretty fast14:08
j_vhttps://www.gnu.org/software/grub/manual/grub/html_node/Features.html#Features14:11
j_vhow much of a drive would you need?14:15
pedjaI have 2 1Tb drives, from the times when Samsung still made spinning rust14:19
j_vcould this be related to your issue? https://www.systutorials.com/linux-kernels/94656/ext4-fix-data-corruption-in-datajournal-mode-linux-4-9-14/14:21
pedjaext4 is mounted data=ordered here (defaults), but I do wonder if that landed in 4.14.24 that I run14:26
j_vreading through that article, it looks older than I thought, that patch is over a year old. sorry, i should have looked it over better before pasting it.14:34
pedjahow do you mean? the article is from yesterday14:36
j_vyeah, but i looked up the patch date, it's from Jan 2017... more that a year ago, and author of article is stating it was a part of 4.9.14 release (old now), so is very old news even if it says it was written yesterday... weird if you ask me14:38
pedjaheh, it *is* weird14:42
darfopedja: hard disk grind during a lockup might be a core dump. Got any core files?16:12
pedjadarfo, I looked all over, can't find any16:26
j_vCould it be swap related?16:39
pedjahm. system was relatively idle, just FF open :)16:45
pedjano compile, no nothing16:46
j_vcron jobs that could have run around then?16:46
pedjayeah, locatedb16:47
j_vdisk intensive...16:48
pedjathat might be it. and my main hard disk is old, so something like that would push it just enough to crash.16:50
pedjaI removed locatedb from cron jobs, so I'll see if it happens again.16:51
pedjaon the plus side, it forced me to refresh my memory on grub and grub-rescue usage :)16:53
j_vno loss without some gain16:56
*** onodera has joined #crux-devel18:21
juefrinnst: wrt FF/PA, I did some more tests but it seems that the only reliable way is to run PA continiously, until FF get fixed, hopefully18:30
juefrinnst: I have the following in my .xinitrc -> 'pulseaudio --start --exit-idle-time=-1 --log-target=syslog &'18:31
pedjajue, what do you mean by 'continuously'?18:59
pedjaI am curious, why so aggressive exit idle time?19:04
juenormaly the PA daemon exits after 20 secs idle time19:04
pedjayes19:05
jueand if you restart FF sound will not work, because FF don't start PA :(19:05
pedjaah.19:06
jue--exit-idle-time with a negative value means don't exit at all19:06
juepedja: you are running some kind of DE?19:10
pedjaXFCE19:10
jueyou don't see the problem because something else, probably a volume control, keeps the PA daemon running19:11
pedjaI think my issue with FF and PA is that PA sets the default output device wrong (headphones instead of Line Out)19:11
jueif you try it with a plain wm you will see what we mean19:12
juewe = frinnst and me19:12
pedjaI have to start PA with 'start-pulseaudio-x11' before xfce pulse plugin pick it up19:13
pedjapicks*19:13
pedjait would probably work automagically if I used one of those login manager thingies19:14
pedjaI could never wrap my head around that leet tiling WM's :)19:16
pedjaI am old school, plain DE is all I need19:17
pedjaand Xfce is perfect for that19:17
jaegerjust for grins, if you guys haven't done so already, want to try my way?19:19
jaeger(regarding pulseaudio start)19:19
jaeger$ cat .xinitrc19:19
jaegerexec ck-launch-session i319:19
juewhat is ck-launch-session?19:20
jaegerconsolekit's session launcher19:20
jueah19:20
juedoes this solve the FF/PA issue?19:22
jaegerJust curious if that makes a difference... I haven't really noticed any issues with starting pulse or it dying randomly19:22
jaegerThat's why I wondered if someone else would test. I've yet to see any problems but I've been using ck-launch-session for a long time19:23
jueno, it dies not randomly, it dies after exit-idle-time, which has a default of 20 sec19:23
jueand because FF don't start PA at all I have the above line in my .xinitrc19:24
juebut, of course, I'll try your suggestion19:25
pedjaI used to have this in .xinitrc 'ck-launch-session dbus-launch --exit-with-session startxfce4', but startxfce4 does that on its own now19:27
juejaeger: doesn't help with the FF/PA issue19:29
pedjaspeaking of consolekit, any interest at all in switching to consolekit2?19:32
pedjait's a drop in replacement, but actively maintained19:33
pedjaafaik, at least, haven't checked consolekit proper lately19:34
jueI'm interested in not beeing forced to install it ;)19:36
jaegerjue: ok, was just curious. Thanks for testing19:43
juenp20:08
jaegerfor reference, my asound.conf only contains these:20:10
jaegerpcm.default pulse20:10
jaegerctl.default pulse20:10
jueyep, same here20:10
jaegerI've not had problems with audio at all, though I use a binary firefox release rather than opt/firefox20:14
crash_i'm using the binary firefox aswell20:42
*** dlcusa has joined #crux-devel20:44
frinnstjue: cheers. I'll give it a spin21:09
*** onodera has quit IRC23:11

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!