INFO: task blocked for more than 120 seconds.

beta990 · 2014-08-08 21:04:08

At the moment I'm receiving a lot of these messages:

[19200.187328] INFO: task kworker/u4:12:7581 blocked for more than 120 seconds.
[19200.187380]       Not tainted 3.15.8-1-ARCH #1
[19200.187415] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.

[19200.188032] INFO: task chmod:1000 blocked for more than 120 seconds.
[19200.188081]       Not tainted 3.15.8-1-ARCH #1
[19200.188116] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.

It sometimes locks my system, and I need to (force) restart the PC.

Specs:
ASUS C60M1-I
Archlinux x64 (Linux 3.15.8-1-ARCH #1 SMP PREEMPT Fri Aug 1 08:51:42 CEST 2014 x86_64 GNU/Linux) with latest updates
Samba, NZBGet, syncthing
HDD's are encrypted (luks) and on top are using btrfs (4 HDD's (/home), 1 SSD (/))

What could cause the issues? Is it maybe a faulty HDD? I've done the SMART test, but it hasn't found any error (yet).
Could it be that the system is to slow to process multiple actions? (e.g. downloading, sync, smb, etc.)

Hope you can help me,

Thanks.

Last edited by beta990 (2014-08-08 21:05:47)

Rexilion · 2014-08-09 19:20:05

Did this happen after an upgrade?

anatolik · 2014-08-09 19:43:10

Post kernel stack trace for the stuck processes. They should be in your dmesg. If they are not, then post content of /proc/$PID/stack where $PID of stuck process (7581 and 1000 in your examples).

What filesystem do you use? UPDATE: oh, sorry now I see it is btrfs. There are similar issues reported by btrfs users:
https://bbs.archlinux.org/viewtopic.php?pid=1439605
http://www.spinics.net/lists/linux-btrfs/msg36267.html

Last edited by anatolik (2014-08-09 19:46:57)

beta990 · 2014-08-09 20:06:00

Thanks for the replies!

@Rexilion: To be honest I can't remember, I daily update the system.
I didn't notice the warnings in dmesg, only till now.

@anatolik: Yesterday I tried a few things:
- Installed Linux-ck (ck-bobcat)
- Removed some btrfs mount flags, now only: noatime,compress=zlib

It seems indeed to be btrfs related, because I don't have the issues (anymore) at the moment.
I cannot find the link at the moment, but I found a post on the Arch bug reports that it may be an issue when using compress=lzo instead of compress=zlib.

So at the moment no problems, but going to try out if I can reproduce the problem.

Thanks and keep you guys posted!

PS. Should I report this bug?

Last edited by beta990 (2014-08-09 20:06:38)

anatolik · 2014-08-09 20:19:03

beta990 wrote:

PS. Should I report this bug?

This sounds like an upstream btrfs bug. Once you hit it again - collect as much information as possible and post it to btrfs project maillist. This is the best way to get the bug fixed.

davidm · 2014-08-10 20:38:43

There are deadlock issues with the btrfs code showing on kernels 3.15.x and 3.16.x.

https://www.mail-archive.com/linux-btrf … 36238.html

Based on what I have read it is expected to be fixed and in the kernel around when 3.17 hits (see linkedmessage from "Duncan" on the mailing list).

Last edited by davidm (2014-08-10 20:42:53)

Arch Linux

#1 2014-08-08 21:04:08

INFO: task blocked for more than 120 seconds.

#2 2014-08-09 19:20:05

Re: INFO: task blocked for more than 120 seconds.

#3 2014-08-09 19:43:10

Re: INFO: task blocked for more than 120 seconds.

#4 2014-08-09 20:06:00

Re: INFO: task blocked for more than 120 seconds.

#5 2014-08-09 20:19:03

Re: INFO: task blocked for more than 120 seconds.

#6 2014-08-10 20:38:43

Re: INFO: task blocked for more than 120 seconds.

Board footer