Red Hat Bugzilla – Bug 163658
kernel oopses under load in __find_get_block_slow
Last modified: 2015-01-04 17:21:00 EST
Description of problem:
Server machine oopses & freezes from time to time under load - easily
reproducible with disabled unmaskirq and 32bit transfer for HDDs accessed by
samba, but then does not print Oops. When unmaskirq and 32bit transfer are
enabled kernel produces oops before freezing up.
Version-Release number of selected component (if applicable):
Access to samba over the gigabit LAN with 10 computers in a rendering farm.
Steps to Reproduce:
1. Setup samba
2. Trigger heavy load on it
3. Wait for the crash
Machine freezing up under heavy loads.
Stable under heavy loads.
Attached are oops traces from /var/log/messages, and dmesg logged to a file.
Please let me know if you think oopses are HW related so that we could start
switching HW parts in the box one by one.
Created attachment 116953 [details]
oops traces from "/var/log/messages" and dmesg output
Can you try and reproduce this problem with the latest updates please ?
If it still reoccurs, I recommend filing a bug in http://bugme.osdl.org (the
upstream kernel bugzilla), as reiserfs is unsupported by Fedora.