[Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

whoops · 2013-03-23 09:11:32

edit: adjusted thread title because "stuff got weirder" in post #6. Due to the newly discovered random aspect of the problem I can't really tell any more, how long it has been occurring (or been there undetected) - most information given by me in the OP might be useless.

Hibernating worked fine and was reliable for several months. Don't remember having changed any system files lately and nothing that directly has to do with hibernating at all since I switched to systemctl.

Hibernating looks normal first, but the machine doesn't power down - just shows black screen and doesn't do anything. I've been able to switch to a unused tty but couldn't login (it just freezes). Machine did not react to power acpi button event... after turning it off, it booted just normally (no resume from hibernate).

Mar 10 23:59:26 arch polkitd[23396]: Registered Authentication Agent for unix-process:20270:5128224 (system bus name :1.59 [/usr/bin/pkttyagent --notify-fd 5 --fallback], object path /org/freedesktop/PolicyKit1/AuthenticationAgent, locale de_DE.UTF-8)
Mar 10 23:59:35 arch polkitd[23396]: Operator of unix-session:2 successfully authenticated as unix-user:root to gain TEMPORARY authorization for action org.freedesktop.login1.hibernate-multiple-sessions for unix-process:20270:5128224 [systemctl hibernate] (owned by unix-user:user)
Mar 10 23:59:35 arch systemd[1]: Starting Sleep.
Mar 10 23:59:35 arch systemd[1]: Reached target Sleep.
Mar 10 23:59:35 arch systemd[1]: Starting Hibernate...
Mar 10 23:59:35 arch polkitd[23396]: Unregistered Authentication Agent for unix-process:20270:5128224 (system bus name :1.59, object path /org/freedesktop/PolicyKit1/AuthenticationAgent, locale de_DE.UTF-8) (disconnected from bus)
Mar 10 23:59:35 arch kernel: PM: Hibernation mode set to 'platform'
Mar 10 23:59:35 arch systemd-sleep[20285]: Hibernating system...
Mar 10 23:59:36 arch kernel: PM: Marking nosave pages: [mem 0x0009e000-0x000fffff]
Mar 10 23:59:36 arch kernel: PM: Marking nosave pages: [mem 0xcff90000-0xffffffff]
Mar 10 23:59:36 arch kernel: PM: Marking nosave pages: [mem 0xc4000000-0xc7ffffff]
Mar 10 23:59:36 arch kernel: PM: Basic memory bitmaps created
Mar 11 08:31:12 arch systemd-journal[158]: Allowing runtime journal files to grow to 398.8M.
Mar 11 08:31:12 arch kernel: Initializing cgroup subsys cpuset
Mar 11 08:31:12 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 11 08:31:12 arch kernel: Kernel command line: BOOT_IMAGE=/boot/vmlinuz-linux-ck root=/dev/sda1 ro ipv6.disable=1 drm_kms_helper.edid_firmware=edid/edid-todvi.bin resume=/dev/sda2 init=/bin/systemd elevator=bfq loglevel=5

(Kernel line above looks the same as always, too... double-checked anyway: fstab says "/dev/sda2 swap swap defaults 0 0")

I guess the last time it worked was on Mar 08:

Mär 08 00:00:06 arch systemd[1]: Starting Sleep.
Mär 08 00:00:06 arch systemd[1]: Reached target Sleep.
Mär 08 00:00:06 arch systemd[1]: Starting Hibernate...
Mär 08 00:00:06 arch kernel: PM: Hibernation mode set to 'platform'
Mär 08 00:00:06 arch systemd-sleep[6915]: Hibernating system...
Mär 08 00:00:07 arch kernel: PM: Marking nosave pages: [mem 0x0009e000-0x000fffff]
Mär 08 00:00:07 arch kernel: PM: Marking nosave pages: [mem 0xcff90000-0xffffffff]
Mär 08 00:00:07 arch kernel: PM: Marking nosave pages: [mem 0xc4000000-0xc7ffffff]
Mär 08 00:00:07 arch kernel: PM: Basic memory bitmaps created
Mär 08 09:02:24 arch kernel: PM: Syncing filesystems ... done.
Mär 08 09:02:24 arch systemd[1]: Time has been changed
Mär 08 09:02:24 arch systemd-sleep[6915]: System thawed.

Haven't changed the kernel (or anything important I could remember/find) since that:

Mar 04 00:03:03 arch systemd-sleep[4283]: Hibernating system...
Mar 04 09:55:50 arch systemd-sleep[4283]: System thawed.
Mar 04 16:30:23 arch systemd-sleep[20466]: Hibernating system...
Mar 04 17:18:41 arch systemd-sleep[20466]: System thawed.
Mar 05 00:01:54 arch systemd-sleep[10753]: Hibernating system...
Mar 05 08:29:27 arch systemd-sleep[10753]: System thawed.
Mar 05 11:55:26 arch kernel: Linux version 3.7.10-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Thu Feb 28 12:54:41 CET 2013
Mar 05 13:07:34 arch systemd[1]: Starting Reboot...
Mar 05 13:08:03 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 05 23:59:19 arch systemd-sleep[1135]: Hibernating system...
Mar 06 08:52:51 arch systemd-sleep[1135]: System thawed.
Mar 06 10:22:20 arch systemd[1]: Starting Reboot...
Mar 06 10:22:45 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 06 23:54:44 arch systemd-sleep[16726]: Hibernating system...
Mar 07 08:30:06 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 07 14:40:01 arch systemd-sleep[17829]: Hibernating system...
Mar 07 16:44:32 arch systemd-sleep[17829]: System thawed.
Mar 08 00:00:06 arch systemd-sleep[6915]: Hibernating system...
Mar 08 09:02:24 arch systemd-sleep[6915]: System thawed.
Mar 08 23:59:48 arch systemd-sleep[32107]: Hibernating system...
Mar 09 08:30:10 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 10 00:00:41 arch systemd-sleep[1210]: Hibernating system...
Mar 10 09:44:57 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013
Mar 10 23:59:35 arch systemd-sleep[20285]: Hibernating system...
Mar 11 08:31:12 arch kernel: Linux version 3.8.2-1-ck (user@arch) (gcc version 4.7.2 (GCC) ) #1 SMP PREEMPT Tue Mar 5 12:51:38 CET 2013

There were kernel updates earlier (mar 05+6), but unless I read something wrong I rebooted afterwards and hibernated successfully, so that can't be it I guess...? Tried downgrading nvidia, but that didn't change anything. Didn't see anything else in there that arose my suspicion:

[2013-03-05 09:05] upgraded opencv (2.4.3-5 -> 2.4.4-1)
[2013-03-05 09:05] upgraded xfdesktop (4.10.0-5 -> 4.10.1-1)
[2013-03-05 13:02] upgraded linux-ck (3.7.10-1 -> 3.8.2-1)
[2013-03-05 13:02] upgraded linux-ck-headers (3.7.10-1 -> 3.8.2-1)
[2013-03-05 13:03] upgraded dosfstools (3.0.15-1 -> 3.0.16-1)
[2013-03-05 13:06] upgraded google-earth (7.0.2.8415-1 -> 7.0.3.8542-1)
[2013-03-05 13:06] upgraded nvidia-ck (313.18-4 -> 313.18-5)
[2013-03-05 13:07] upgraded linux-ck (3.8.2-1 -> 3.8.2-1)
[2013-03-05 13:07] upgraded linux-ck-headers (3.8.2-1 -> 3.8.2-1)
[2013-03-05 20:06] upgraded crypto++ (5.6.1-4 -> 5.6.2-1)
[2013-03-05 20:06] upgraded hexedit (1.2.12-5 -> 1.2.13-1)
[2013-03-05 20:06] upgraded linux (3.7.9-2 -> 3.7.10-1)
[2013-03-05 20:06] upgraded linux-docs (3.7.9-2 -> 3.7.10-1)
[2013-03-05 20:06] upgraded linux-headers (3.7.9-2 -> 3.7.10-1)
[2013-03-05 20:06] upgraded mercurial (2.5.1-1 -> 2.5.2-1)
[2013-03-05 20:07] upgraded scons (2.2.0-1 -> 2.3.0-1)
[2013-03-06 10:04] upgraded ethtool (1:3.7-1 -> 1:3.8-1)
[2013-03-06 10:04] upgraded lib32-libgpg-error (1.10-1 -> 1.11-1)
[2013-03-06 10:04] upgraded nvidia-utils (313.18-1 -> 313.26-1)
[2013-03-06 10:04] upgraded lib32-nvidia-utils (313.18-5 -> 313.26-1)
[2013-03-06 10:04] upgraded lsof (4.86-1 -> 4.87-1)
[2013-03-06 10:04] upgraded nvidia (313.18-3 -> 313.26-1)
[2013-03-06 10:04] upgraded tzdata (2012j-1 -> 2013a-1)
[2013-03-06 10:04] upgraded xfdesktop (4.10.1-1 -> 4.10.1-2)
[2013-03-06 10:20] upgraded nvidia-ck (313.18-5 -> 313.26-1)
[2013-03-06 14:19] upgraded file (5.11-2 -> 5.13-1)
[2013-03-06 15:44] upgraded chromaprint (0.7-3 -> 0.7-4)
[2013-03-06 17:46] upgraded ntfs-3g (2013.1.13-1 -> 2013.1.13-2)
[2013-03-07 11:18] upgraded pyqt-common (4.9.6-2 -> 4.10-1)
[2013-03-07 11:18] upgraded sip (4.14.3-1 -> 4.14.4-1)
[2013-03-07 11:18] upgraded python2-sip (4.14.3-1 -> 4.14.4-1)
[2013-03-07 11:18] upgraded python2-pyqt (4.9.6-2 -> 4.10-1)
[2013-03-08 18:50] upgraded calibre (0.9.21-3 -> 0.9.22-1)
[2013-03-08 23:35] upgraded gparted (0.14.1-1 -> 0.14.1-2)
[2013-03-08 23:35] upgraded libreoffice-de (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-en-GB (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-common (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-base (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-calc (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-draw (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-impress (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-math (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded libreoffice-writer (3.6.5-1 -> 4.0.1-1)
[2013-03-08 23:35] upgraded meld (1.6.1-1 -> 1.7.0-1)
[2013-03-08 23:35] upgraded serf (1.1.1-1 -> 1.2.0-1)
[2013-03-08 23:35] upgraded vim-colorsamplerpack (8.03-3 -> 2012.10.28-2)
[2013-03-08 23:35] upgraded vim-taglist (45-5 -> 46-1)
[2013-03-09 09:07] upgraded blender (6:2.66-2 -> 7:2.66a-1)
[2013-03-09 09:07] upgraded libnl (3.2.19-1 -> 3.2.21-1)
[2013-03-09 09:07] upgraded iw (3.8-2 -> 3.8-3)
[2013-03-09 09:07] upgraded crda (1.1.3-1 -> 1.1.3-2)
[2013-03-09 09:07] upgraded firefox (19.0-1 -> 19.0.2-1)
[2013-03-09 09:07] upgraded firefox-i18n-de (19.0-1 -> 19.0.2-1)
[2013-03-09 09:07] upgraded libpcap (1.3.0-2 -> 1.3.0-3)
[2013-03-09 09:07] upgraded libsecret (0.12-2 -> 0.14-1)
[2013-03-09 09:07] upgraded pavucontrol (1.0-1 -> 2.0-1)
[2013-03-09 09:07] upgraded wpa_supplicant (2.0-2 -> 2.0-3)
[2013-03-09 09:25] upgraded proggyfonts (0.1-7 -> 0.1-8)
[2013-03-09 11:51] upgraded inputproto (2.2-1 -> 2.3-1)
[2013-03-09 11:51] upgraded xorg-server-common (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server-devel (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server-xdmx (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server-xephyr (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server-xnest (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 11:51] upgraded xorg-server-xvfb (1.13.2.901-1 -> 1.13.3-1)
[2013-03-09 12:23] upgraded xine-lib (1.2.2-6 -> 1.2.2-7)
[2013-03-09 19:13] upgraded gparted (0.14.1-2 -> 0.14.1-4)

"Free" looks fine, too, I guess (no way the ram image thing could get too big for swap):

             total       used       free     shared    buffers     cached
Mem:          7,8G       2,7G       5,1G         0B        64M       1,3G
-/+ buffers/cache:       1,3G       6,5G
Swap:          31G         0B        31G

The latest kernel update (I think there were two since I started investigating this problem) did not change anything. Same with arch stock kernel.

I've been trying to fix that for a while now... Ran out of ideas a few days ago. Any hints?

Last edited by whoops (2013-04-03 18:37:58)

illusionist · 2013-03-23 10:17:48

The only thing I can think of right now is try and use the lts kernel to see if that narrows down the problem.

Last edited by illusionist (2013-03-23 10:19:17)

whoops · 2013-03-23 21:57:52

Thanks, that sort of... narrowed down... stuff... I think.

Turned out, hibernate not only works after rebooting with the LTS kernel - it works after rebooting with my current kernel too.

Guess some user-space thing that always happens sooner or later while I'm using the machine must be breaking hibernate. Hibernating directly after (re)booting always seems to work. I tried again after starting the DE and various software and it still worked. The next time I tried (6 hours later) it failed again.

Have not been able to figure out what exactly makes it break. But it reliably breaks EVERY single time I use the machine for a longer time before hibernating (which is the only situation where I normally use hibernate).

whoops · 2013-03-25 09:04:03

So I...:
- Booted (linux-ck, as usual)
- Started ALL programs that I normally use over the course of a day (loading files that I was working on / using where possible)
- Also a random game with wine because my ram was only 3GB full
- compiled a small code snipped to fill the rest of my memory (http://superuser.com/questions/278855/a … ses-memory)
- Started some random torrents and amule too, just to make sure that there are network connections.
- started a (limited) semi-forkbomb to get some more random useless processes + cpu usage

Hibernating was slow and took a while, but worked fine (tried 10 times, rebooting and doing everything again 3 times in between).

Then 6 hours of normal use later, when I really wanted to hibernate, it froze again, just like it did every single day since [Mar 08 23:59:48 / see logs in OP]. (no hd acitivty. Waited 15 Minutes anyway - nothing happened).

Can't try to run kernel-LTS for a whole day, because I need that machine to do stuff and not everything works with kernel LTS. So... fresh out of ideas.

Last edited by whoops (2013-03-25 09:07:22)

illusionist · 2013-03-26 01:51:56

Have you tried "tuxonice" patched kernel ? For some strange reason new kernels are not working for me. I am sure it is not my setup or hardware or anything else form my side. But *-lts kernel is working fine. if you don't want/like to use lts then IMO changing the hibernation back-end may have some benefits.

Last edited by illusionist (2013-03-26 01:54:19)

whoops · 2013-04-02 07:52:42

I'd rather try to figure out what's the problem than "abandon" the default method for hibernating.

So... it gets weirder and weirder:
While trying around more, I noticed that it's possible to make hibernate work by hibernating directly after reboot and while setting up my DE several times (I did that primarily to figure out what breaks hibernate). The first hibernation directly after boot always works. It's possible that there's a random aspect (meaning it might be decided at boot or shortly thereafter whether hibernate will keep working) to this or that the order in which I start my stuff plays a role here (p.E. maybe one application / script has to be started before another initializes).

Once I have everything running and hibernate works, it keeps working reliably, for days, no matter what I do (as far as I can tell by now, that only can happen if I hibernate directly after reboot at least once) until I have to _reboot_ at which point the problem occurs again (unless I repeat the whole "hibernating after reboot multiple times between starting applications" thing again).

This is way to weird for me to figure it out just by trial and error... any more suggestions on how to locate the problem?

whoops · 2013-04-03 18:35:23

Problem might have something to do with nvidia hdmi sound / pulseaudio.

I noticed that it occurs, if I remove the nvidia sound device ("echo 1 > /sys/devices/pci0000:00/0000:00:02.0/0000:01:00.1/remove") AFTER I start Xorg (/xfce4 etc).

Maybe pulseaudio (or something else that's started by the session) makes hibernate freeze if it fails to remove a sound card that was there when it was started?

Haven't tested further yet. For now a workaround is simply: making sure that the annoying extra sound device I don't wand to use is always removed BEFORE I do anything else (using a "rc-local.service").

Not sure if I'm going to find the time to figure out what exactly the problem was and write a bug report somewhere (because I already broke something else in the meantime and need to fix that ), but I'll try. Any ideas what I should be taking a closer look at first?

Last edited by whoops (2013-04-03 18:40:30)

heaven · 2014-08-11 07:45:19

whoops wrote:

Problem might have something to do with nvidia hdmi sound / pulseaudio.
I noticed that it occurs, if I remove the nvidia sound device ("echo 1 > /sys/devices/pci0000:00/0000:00:02.0/0000:01:00.1/remove") AFTER I start Xorg (/xfce4 etc).
Maybe pulseaudio (or something else that's started by the session) makes hibernate freeze if it fails to remove a sound card that was there when it was started?
Haven't tested further yet. For now a workaround is simply: making sure that the annoying extra sound device I don't wand to use is always removed BEFORE I do anything else (using a "rc-local.service").
Not sure if I'm going to find the time to figure out what exactly the problem was and write a bug report somewhere (because I already broke something else in the meantime and need to fix that ), but I'll try. Any ideas what I should be taking a closer look at first?

Hi, it seems I am suffering from exactly the same problem and seems it started once I installed pulseaudio.

I am using arch for many years and never had such a problem but never used pulse also. Have Nvidia 560TI and an integrated into the motherboard sound card. Sometimes, when hibernate fails, there's a line about an nvidia related error/warning (will make a photo next time). All the other symptoms are the same.

Will try to remove the sound card before the hibernate next time. Also killing pulseaudio may help as well (will try also).

heaven · 2014-08-11 07:57:28

A similar issue: https://bugs.launchpad.net/ubuntu/+sour … ug/1301601

Arch Linux

#1 2013-03-23 09:11:32

[Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#2 2013-03-23 10:17:48

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#3 2013-03-23 21:57:52

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#4 2013-03-25 09:04:03

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#5 2013-03-26 01:51:56

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#6 2013-04-02 07:52:42

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#7 2013-04-03 18:35:23

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#8 2014-08-11 07:45:19

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

#9 2014-08-11 07:57:28

Re: [Workaround/unsolved] Hibernate freezes sometimes (pulseaudio/nvidia?)

Board footer