On Fri, Jun 04, 2010 at 12:46:17PM -0400, Pittigher, Raymond - ES wrote: > Started to see these messages in the log files. After a while the slab cache uses all the memory then OOM. Memory leak someplace? > > > May 27 19:48:06 suoserv4 kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disable > s this message. > May 27 19:48:06 suoserv4 kernel: nfsd D ffff81000103f1a0 0 4562 1 > 4563 4561 (L-TLB) > May 27 19:48:06 suoserv4 kernel: ffff810c0d43bac0 0000000000000046 ffff810c0d43ba10 000000 > 0000000000 > May 27 19:48:06 suoserv4 kernel: ffff810c0d43ba58 000000000000000a ffff810c0e3a67a0 ffff81 > 0c2fce5100 > May 27 19:48:06 suoserv4 kernel: 0000142538fcf800 000000000000863d ffff810c0e3a6988 000000 > 07885df478 > May 27 19:48:06 suoserv4 kernel: Call Trace: > May 27 19:48:06 suoserv4 kernel: [<ffffffff885ef19d>] :linuxshield:doScan+0x8e7/0xcff > May 27 19:48:06 suoserv4 kernel: [<ffffffff800a1ba4>] autoremove_wake_function+0x0/0x2e > May 27 19:48:06 suoserv4 kernel: [<ffffffff885ee4ce>] :linuxshield:cacheLookup+0x96/0x152 > May 27 19:48:06 suoserv4 kernel: [<ffffffff885f0332>] :linuxshield:SC_doScan+0x108/0x115 > May 27 19:48:06 suoserv4 kernel: [<ffffffff885ed625>] :linuxshield:openHook+0x138/0x1b7 > May 27 19:48:06 suoserv4 kernel: [<ffffffff885de9a3>] :lshook:invokeCallbacks+0x32a/0x4f7 my first guess would be 'linuxshield'. This isn't actually an oom, it's complaining because a task stayed in kernel context for a really long time. This can happen if for eg, an nfs server goes away for a while. Or it could just be a side-effect of the way that mcaffee does their file scanning. On the surface, this doesn't look like a linux bug. Dave _______________________________________________ kernel mailing list kernel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/kernel
