MY QUESTIONS ARE:
1) Why did these boxes not reboot after the kernel panic?
2) Are there 2 kinds of kernel panic? A "mini-panic-don't-reboot" as
well as a full-out kernel panic?
3) Is the "file-max limit 4096 reached" error indicative of the reason
why the kernel for some reason couldn't reboot?
----------------------------------
DETAILS & EXPLANATION:
I have a number of 2.2.14 boxes running some custom software. The
loadlin bootup parameters include:
[...] panic=1
And to be extra certain, rc.local calls 'echo "1"
>/proc/sys/kernel/panic'. I can confirm the value of "panic":
[root@localhost /root]# cat /proc/sys/kernel/panic
1
So all seems fine ***EXCEPT*** that recently I've observed the
following situation twice on these boxes:
[root@localhost /root]# Kernel panic: Free list corrupted
VFS: file-max limit 4096 reached
This box had been sitting mostly idle for a few days before I got
around to checking the console. Doing a usenet search on "Kernel
panic: Free list corrupted" returns much information, but not much that
seems to be helpful.
The first question is why did these boxes not reboot after the kernel
panic?
Strangely enough, I found that even after this kernel panic I still had
access to the console. However, I couldn't run anything, as evidenced
by this cut-and-paste:
[root@localhost /root]# cat /proc/sys/kernel/panic
bash: fork: Resource temporarily unavailable
[root@localhost /root]# ls -la
bash: fork: Resource temporarily unavailable
Are there 2 kinds of kernel panic? A "mini-panic-don't-reboot" as well
as a full-out kernel panic? Or is the "file-max limit 4096 reached"
error indicative of the reason why the kernel for some reason couldn't
reboot?
A few technical details:
- vanilla 2.2.14 kernel
- FAT fs on 1st hard disk (mounted periodically to backup some files)
- ext2 on 2nd hard disk (Linux on this drive)
- 512 MB ram
- PCnet/FAST III (pcnet32.c) ethernet
Thanks in advance for any assistance provided,
St?phane Charette