I have a bunch of web servers all running RedHat 6.0 that use 2.2.10-13
kernels. The servers are a collection of PPro, PII, and PIII's with
different motherboards, scsi, network and video cards. Since moving to
2.2.x kernels, these servers have not been very stable under load. Back
with 2.0.x kernels, I would have 90+ days of uptime.
But after doing fresh installs of Redhat 6.0 and rebuilding the kernels, I
only get uptimes of a week or less before the different boxes lock up
solid. When one locks up, I don't see any Oops or kernel panics or
anything else that might give a hint as to what happended.
About the only thing common between all of these servers is that they all
run serial consoles with this line in lilo.conf
append="console=tty0 console=ttyS1,19200n8"
Now I haven't tested this extensively yet, but it appears that when I
reverse the ordering or drop the serial consoles, the servers appear to be
more stable.
Is this just a coincident or could there be a problem somewhere in the
kernel that would cause the system to be less stable when a serial console
is used as the default for /dev/console? Or am I just missing something
else more obvious? I ran memtest on everything and they all come back
error free, also all these boxes ran 2.0.x kernels without problems.
I'm just not sure what else to try, any suggestions?
David
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/