Bug 183682

Summary: OOM killer, then NMI watchdog, on FC5-post-test3 x86_64
Product: [Fedora] Fedora Reporter: Matt Domsch <matt_domsch>
Component: kernelAssignee: Kernel Maintainer List <kernel-maint>
Status: CLOSED RAWHIDE QA Contact: Brian Brock <bbrock>
Severity: medium Docs Contact:
Priority: medium    
Version: 5CC: bstretch, matt_domsch, wtogami
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-10-12 16:44:14 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
fc5-oom.txt none

Description Matt Domsch 2006-03-02 19:50:32 UTC 
Description of problem:
Running Fedora Core 5 rawhide tree as of 02-March-2006.  Kernel
2.6.15-1.1996_FC5 x86_64 on a Dell Precision 380.  This system is used to run
'mock' for building packages.  System has 1GB RAM, 2GB swap, and one socket with
a dual-core Intel(R) Pentium(R) D CPU 3.00GHz.

Often (not always, not terribly rare though), while running mock to build the
libsmbios .src.rpm packages for various releases, the kernel will invoke the OOM
killer, and subsequently, the NMI watchdog will fire.  The box is then hung.  A
trace of the failure was captured.

This has been happening on kernels from -test3 to version noted.  If multiple
mock builds are run in parallel, failure happens sooner.

Version-Release number of selected component (if applicable):
2.6.15-1.1996_FC5 x86_64

How reproducible:
fairly often, several times a day

Steps to Reproduce:
1. mock -r fedora-4-x86_64.cfg libsmbios-$version.src.rpm
2. wait
Comment 1 Matt Domsch 2006-03-02 19:50:32 UTC 
Created attachment 125561 [details]
fc5-oom.txt
Comment 2 Brian Stretch 2006-05-27 15:38:42 UTC 
I'm having similar problems on my dualcore AMD64 machine. 
http://www.redhat.com/archives/rhl-list/2006-May/msg04862.html
It sounds like you're a little better at diagnosing these things than me, maybe
my notes will give you some ideas?
Comment 3 Brian Stretch 2006-05-31 13:08:48 UTC 
Hmm, my problems were fixed by reinstalling FC5.  I have no idea what the
original problem was.  Prior to that, swapping in a singlecore CPU failed to help.
Comment 4 Matt Domsch 2006-10-12 16:44:14 UTC 
Not seen by me in quite a while with newer kernels.  Closing.