On Thu, Dec 16, 2010 at 12:03 PM, William Muriithi < william.muriithi@xxxxxxxxxxxxxxxxxxx> wrote: > The other problem is, if there is an exhaustion of lower memory, OOM killer > will attempt to clear memory and once in a while, that will result to the > database being the candidate. That can result in data inconsistency and a > rough night. So why use a 32 bit kernel when the 64 bit kernel has long > been so stable and reliable? Seriously, I can not think of a single reason > why it would be preferable to use a 32 bit RHEL5. Unless you really have > fun fighting unnecessary fires > > This bit from Note 452000.1 may also be helpful: *The Kernel is out of LowMem (32-bit architectures only)* "LowFree" in /proc/meminfo is very low, but HighFree is much higher. OOM killer will take action under that situation. *The specific workload may benefit from being run on a 64-bit architectures *or other methods (See Note 452326.1<https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&id=452326.1>) . Jared Still Certifiable Oracle DBA and Part Time Perl Evangelist Oracle Blog: http://jkstill.blogspot.com Home Page: http://jaredstill.com