------- Additional Comments From tdickson@xxxxxxxxxxx 2003-10-12 17:03 PDT
Created an attachment (id=92)
output from /proc/slabinfo during mount xfs and tree command
Here is some output from running some tests. I wasn't able to get the oom
killer to come into play on this one (I don't think I have enough files). I'm
going to try again tomorrow with less RAM and see what happens. As for now, I'm
retrying with ext3 to see if the issue is in the vfs layer.
xfs_check still gets terminated (out of memory). Hmph.
------- Additional Comments From dizzy@xxxxxxxxx 2003-11-12 03:37 PDT -------
For me it happened just as you said , on a production machine if I run any
command that whould walk a very large tree (I have some >500k of files/dirs on
that partition) then OOM started to kill random processes. I havent seen it
since I upgraded to 2.4.23 but maybe its just a coincidence (I havent tried to
reproduce it either but I noticed it doesnt happen anymore when the machine runs
doing its normal stuff like it happened before).
Another thing to note... Is it true that you dont have swap active ? (at least I
noticed that enabling swap made the OOM killer cool down on normal server
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.