On Tue, Jun 29, 2010 at 08:17:01AM -0400, Patrick LeBoutillier wrote: > Hi all, > I'm running kernel 2.6.32.6 with patch 2.3.0.36.28 and I > think there's a concurrency issue in the /proc /filesystem. sounds interesting ... > Don't know if that makes a difference, but I'm also using > Jacques' tools. would suggest to test with util-vserver (only), just to rule out that the issue is tool related (i.e. incomplete isolation or similar) i.e. please create and start a test guest with util-vserver and see if you can recreate the issue ... > On a multicore (8 cores) server, if I run the following script: > while /bin/true ; do > ls -ld /proc/self > /dev/null > done > twice in parallel (i.e. sh test.sh & sh test.sh) inside a guest, I get > a lot of errors like this: > > ls: cannot read symbolic link /proc/self: No such file or directory > ls: cannot read symbolic link /proc/self: No such file or directory > ls: /proc/self: No such file or directory > ls: cannot read symbolic link /proc/self: No such file or directory > ls: cannot read symbolic link /proc/self: No such file or directory > ls: cannot read symbolic link /proc/self: No such file or directory > ls: /proc/self: No such file or directory > However, if I run the same script directly in the root server, I can't > reproduce the issue. > Here's how the various procs are mounted: > none on /proc type proc (rw) # root /proc > none on /vtmp/test/proc type proc (rw) # guest /proc > Does anyone have a clue on how to fix this? once we can reproduce it, we can fix it :) > I looked at the patch, but it's a bit out of my league... I tried here (to recreate the issue) with a dual core cpu (kernel 2.6.33.x) but couldn't reproduce .... It might be a good idea to update to a recent kernel or at least recent patch as well, just to make sure that you are not seeing a (mainline) issue which was already fixed .... thanks in advance, Herbert > Thanks, > Patrick > > -- > ===================== > Patrick LeBoutillier > Rosemère, Québec, Canada