Memory corruption regression with kernel 3.0.101-0.35 and using pagecache limit

This document (7015898) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 11 Service Pack 3 (SLES 11 SP3)
SUSE Linux Enterprise Server for SAP Applications

Situation

After updating to kernel 3.0.101-0.35 systems are susceptible to crashes if Pagecache limit is set. Corruption of memory pointers happens during file cache reclaim resulting in null pointers and other exceptions. Limiting the page cache ensures that the reclaim process will be run often and thus ensures the system will crash. Stack traces vary so are not referenced.

The system _will_ crash if the following conditions are met.

PRE.cjk { font-family: "Droid Sans Fallback",monospace; }P { margin-bottom: 0.08in; }A:link { }

- kernel version is 3.0.101-0.35
- pagecache limit is set (vm.pagecache_limit_mb)

P { margin-bottom: 0.08in; }A:link { }

Note that limiting the kernel pagecache is only supported on SLES for SAP.

PRE.cjk { font-family: "Droid Sans Fallback",monospace; }P { margin-bottom: 0.08in; }A:link { }

Example of checking to see if the pagecache limit is set (a non-zero value means it is).

myhost:~ # sysctl -a | grep vm.pagecache_limit_mb
vm.pagecache_limit_mb = 4096

Resolution

Update to kernel 3.0.101-0.40 to resolve the issue available here.

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.