From: Charles G. <ch...@de...> - 2004-08-12 15:08:18
|
Aah, yes - It succeeds on a server running a 2.6 kernel. thanks! The kernels on the 3 boxes i initially tested against are 2.4.20 2.4.20-8smp 2.4.20-31.9smp 2.4.20-30.9 I'm at my wits end. The reason I am running these older kernels is I am having other problems with newer kernels. These are brand new msi boxes with high end 3ware and LSI raid systems. They can handle a kernel compile in a while /bin/true look with make -j8, running mprime, and ltp all at once for days, but as soon as we put some cpanel accounts on it, the box locks up. This happens on both of them. It seem i/o related (although not i/o load related). All services stop responding and nothing in the logs. If you have a shell open, you can do operations that don't touch disk, but as soon as you do something that accesses disk (say ls) it locks up the session. We have tried the latest 2.4 and 2.6 kernels but it seems like an issue with the raid code in newer kernels. So we have reverted to an older kernel and software raid as it is reported to occur less frequently in sw raid. Can anyone tell me if these system calls that cause the oops are commonly used, and since the box keeps running after the oops are they safe to ignore? The box is going to be used as a virtual hosting server using cpanel. thanks charles On Aug 12, 2004, at 10:25 AM, Paul Larson wrote: > On Wed, 2004-08-11 at 20:03, Charles Galpin wrote: >> It must be failing on test 2. Can anyone tell me what this means? Can >> I > Looks like you found a kernel bug, but you didn't provide any details > about the kernel you were running. I haven't seen that bug before, but > chances are, moving up to a more recent kernel will fix it for you. > > Thanks, > Paul Larson |