Bug 463372

Summary: Kernel 2.6.27.7 SuSE file monitoring serious bug (inotify)
Product: [openSUSE] openSUSE 11.1 Reporter: Johan Persson <johanp>
Component: KernelAssignee: Jeff Mahoney <jeffm>
Status: RESOLVED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: bluedzins, forgotten_kVkGRNwR1h, forgotten_oT5fNj878H, forgotten_vXTZVacoSi, JC.Polanycia, jeffm, kai.schmidt, meissner
Version: Final   
Target Milestone: ---   
Hardware: i686   
OS: openSUSE 11.1   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Johan Persson 2009-01-03 15:18:29 UTC
There's a bug in linux kernel 2.6.27.7 that is used in SuSE 11.1 that creates unkillable, runaway processes when file monitoring is enabled.  This is especially critical when trying to run an IMAP service (which needs file monitoring for efficiency reasons) on SuSE 11.1 since it will force the entire server to be hard reset several times a week and makes SuSE 11.1 unsuitable as an IMAP server

Hence it is important that the latest stable 2.6.27.10+ kernel can be packaged and distributed as an update. When this update is done notification also needs to go to affected graphic drivers (e.g. NVidia) so they can release appropriate update drivers that match the new kernel and modules.

There's a kernel patch [1], and it's supposedly fixed in 2.6.27.10.

1: 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=711a49a07f84f914aac26a52143f6e7526571143
Comment 1 Forgotten User oT5fNj878H 2009-01-07 03:08:08 UTC
I believe I have encountered this bug also. I am running dovecot on systems ranging from Celeron 450MHz to dual-core Athlon, and these machines freeze between once several hours to once per day and were not usable as mail servers.

I will try turning off dovecot on my machine to see if it helps. I'm also happy to test any kernel updates you might issue.
Comment 2 Forgotten User kVkGRNwR1h 2009-01-07 14:33:51 UTC
I too believe I have encountered this but.  I run Dovecot.
Comment 3 Greg Kroah-Hartman 2009-01-07 21:07:56 UTC
Already fixed, 2.6.27.10 is in the queue, and you can pick it up from the "kernel-of-the-day" site right now if you want to test and verify it's resolved for you.
Comment 5 Johan Persson 2009-01-14 11:11:40 UTC
I can confirm that KOTD 2.6.27.10 fixes this problem
Comment 7 Marcus Meissner 2009-01-14 16:04:27 UTC
JFYI, we are planning to release a 11.1 update soon, including this patch.
Comment 8 Forgotten User oT5fNj878H 2009-01-15 00:41:42 UTC
Me too, no appearance of this issue with KOTD 2.6.27.10, or Factory 2.6.27.10. Thanks guys.
Comment 9 Kai Schmidt 2009-02-18 15:42:24 UTC
hmmm, this bug is marked as resolved about one month ago, but until now, it isn't possible to fix running systems with an normal command like zypper up..... browsing through the different available repos is more confusing then helping.

on 2009-01-07 Greg Kroah-Hartman wrote: Already fixed, 2.6.27.10 is in the queue ...

but until now not avaibale through a normal update. and the hint to get it over kotd is not very helpfull, because Repos like http://download.opensuse.org/repositories/Kernel:/SL111_BRANCH/openSUSE_11.1/ did not deliver 2.6.27.10 i only get 2.6.27.8 :( 

hello opensuse/novell wake up. this is an very important bug......
Comment 10 Greg Kroah-Hartman 2009-02-18 15:49:58 UTC
I agree.  Jeff, what's the status on pushing the 11.1 kernel out to the update servers?
Comment 11 Jeff Mahoney 2009-02-18 15:55:44 UTC
We're still sorting out the kABI issues.

Kai, in the interim, if you're willing to deal with kABI breakage again (only important if you're using third-party modules), you can use a fixed kernel from http://ftp.suse.com/pub/projects/kernel/kotd/SL111_BRANCH/
Comment 12 Forgotten User vXTZVacoSi 2009-02-26 21:41:07 UTC
Kernel update available, could you please test if issue is solved? It should be, just for clarity and then close as fixed please =)
Comment 13 Forgotten User oT5fNj878H 2009-02-27 01:06:58 UTC
Linux ken 2.6.27.19-3.2-default #1 SMP 2009-02-25 15:40:44 +0100 x86_64 x86_64 x86_64 GNU/Linux

Speaking for myself, so far so good, after running with it for 4 hours. I use dovecot.
Comment 14 JC Polanycia 2009-03-06 19:36:11 UTC
I have updated as well.  I was seeing this problem with the samba service.  It's been about 2 weeks and have not seen the runaway, unkillable processes since upgrading the kernel.
Comment 15 Forgotten User kVkGRNwR1h 2009-03-06 20:15:13 UTC
I've been running the 2.6.27.19-3.2-pae kernel on my DoveCot server without any problems for two days.
Comment 16 Marcus Meissner 2009-03-06 20:43:35 UTC
tahnks for the reports! -> fixed :)
Comment 17 Lubos Lunak 2009-04-22 11:45:01 UTC
*** Bug 471410 has been marked as a duplicate of this bug. ***