Hi,
I've been having kernel panics on a not-so-regular basis (server was
running fine for 30 days, then had a panic, again fine for 2 months or
so then panic'ed again). Problem is, the machine is far away and I
don't have access to the screen. So far I've changed syslog-ng to
record emergencies not to /dev/tty8 and console(root) but to files
<snip>
destination emergency { file("/var/log/emergencylog"); };
destination else_all { file("/var/log/evrelse"); };
(...)
filter f_emergency { level(emerg); };
(...)
log { source(src); filter(f_emergency); destination(emergency); };
log { source(src); destination(else_all); };
#line above ends all the log statements so it's basically a catch-all
</snip>
Is this enough to record the msg dumps when it would panic again?
Also, I have no idea as to what's causing it, since it's so irregular.
I'm running Linux undine 2.6.23-hardened-r7-undine
(Just had to recompile kernel and tried to fix around stuffs + remove
the things I didn't need hoping it'd be the panic cause)
grsec is enabled and logs most things, with most security enabled
(minus the parts for TCP connections)
PaX is also enabled
---------------------
Hieu Luu Danh
---------------------
Freelance Web Designer
E: <[EMAIL PROTECTED]>
N: United Kingdom, SW15 & East London
--
gentoo-user@lists.gentoo.org mailing list
