Please consider kernel update to 2.6.32-220.4.2.el6 on RHEL6 based server.
The current old rhel6.2 kernel includes sched_clock() overflow problem which the system does crash when uptime exceeded 208.5 days. This may cause critical situation on the lustre servers if the lustre servers do not reboot occasionally.

Redhat announced and released security update kernel.

http://rhn.redhat.com/errata/RHBA-2012-0124.html

This is not only RHEL problem, but also it looks like this codes were landed since 2.6.28.