This is new, I haven't seen this before. One of the new servers running Gentoo, has just died with error on the screen, PANIC Out of Memory, no killable processes available. System is 32Bit has 2GB of RAM / 4GB of SWAP. This didn't used to happen using Slackware 9.1 and 2.6.6 Kernel, discovered since reinstalling with Gentoo 2004.2 and 2.6.8-r3 & genkernel. running with 24 NFSD Threads. Possible memory leak? I would update to r4 but it doesn't compile. Reproducible: Always Steps to Reproduce: To my knowledge the kernel should release the memory after is dumps what it has to disk. Either way the whole server crashes with a kernel panic out of memory. and on a NAS solution that isn't a good thing :(
Can you reproduce this on development-sources-2.6.8.1 ?
I am trying it with 2.6.8.1 developement sources as you suggested.. will know in a week.
cdrdao sets up an out-of-memory condition with 2.6.8-gentoo-r6 and -r3 I don't know if that's related. Googling pops up some comments re a known bug in kernel. Behavior is that once cdrdao runs out-of-memory creeps up until everything killed off.
re kernel out of memory post cdrdao/cdrecord see http://www.ussg.iu.edu/hypermail/linux/kernel/0407.3/0220.html
That sounds like another issue to me. Unless you are burning CD's on your machine Darren?
Nope no CD Burning.. it's a dedicated NFS system. and no 2.6.8.1 didn't resolve the issue, system crept up until there was no killable processes. Unless what your saying is that cdrdao starts chewing memory right off the bat in which case where do I not select it for compilation.
2.6.8.1 apparently has an XFS fault in it's source so now I am stuck between a rock and a hard place. 2.6.8.1 has an XFS problem 2.6.8-r4 doesn't compile correctly 2.6.8-r3 appears to have a wierd memory leak Can you guys include the XFS_FREEZE patch for the kernel since cannot run at all the xfs_freeze command to create snapshots correctly on the XFS volume. http://www.uwsg.iu.edu/hypermail/linux/kernel/0404.0/0483.html here's some information I recieved via a mailing list with regards to the XFS issue: >Thanks for your bug report. The only xfs patch that is applied to the >Debian kernel is an ioctl patch which is pending a merge upstream, so I >guess this is a bug in the XFS code as present in Linus' tree. I have >CCed the linux-xfs mailing list in the hope that someone there can help >you out. > >linux-xfs people, please feel free to either include or not include the >debian bug tracking system on any resulting thread by CCing >274988@bugs.debian.org as you feel fit. I have attached the one patch >that is applied to the debian 2.6.8 kernel for referance. Please let me >or the bug's address know if there is a resolution as I am not on the >linux-xfs list.
I am now trying out 2.6.8-r7 in a vague hope that it will resolve all my present issues. *sigh*
Any chance you could try 2.6.9-final ? The bug is much more likely to have been fixed there.
2.6.8-r7 appears to have an issue with NFSD in a sense that for no reason apparent it decides to stop answering NFSD calls from anywhere but can still look at its self.. very wierd. Different bugs on different kernels all I wanted was snapshots.. :( Am now trying 2.6.9
Sorry that this bug is dragging on so long. Hows it going with 2.6.9?
So far 2.6.9 is holding up quite nicely.
Great. Assuming this is fixed. Please reopen if it isn't :)