On Wed, Dec 30, 2009 at 11:48 PM, Rodery, Floyd A Mr CIV US DISA CDB12 <[email protected]> wrote:
> We've noticed some repetitive error messages in regards to a memory > issue on one of our SLES 10 SP2. We notice the error below every > morning around the same time, with multiple oom-kills of oracle, perl, > etc. Anyone have any thoughts as to why this might be happening with > the amount of memory this guest has? If you have any thoughts or need > anymore information, I would certainly appreciate it. So you filled memory and then ran out of swap space? Your first check would be to see whether SGA and PGA fit comfortably in the guest without swapping. If you're having swappiness still to the default of 60, I would recommend to get that down to 0 (certainly if you don't run Oracle on raw devices). Your ESAUCD2 report would show whether the page cache was far more than the SGA (happens with high swappiness) and whether configured PGA size could be handled. When you believe that SGA and PGA would fit, then you may be bitten by a bug in Oracle where it forgets about PGA setting and allocates far more. I can dig up the reference for that... Rob ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
