I have been battling what I believe is a memory issue and I have been trying to figure out why the page life expectancy on this server is all over the place.
We are running sql 2000/sp4 on windows 2003 sp2. The server HP DL585 with 4 dual core cpu's, 48GB RAM (maxserver mem = 44GB) mirrored OS, RAID 10 (6 drives) for log files, EMC SAN LUN for data files (RAID 5 w/5drives), LUN for tempdb and separate LUN for paging file.
AWE is enabled and the patch for ...not all memory available...has been applied.
We also use Full-text search extensively and use 15 databases that drive our websites. There are a few others but they are not used much.
On any given day it can vary from a low of around 150 to in the thousands.
The Buffer Cache Hit Ratio is a rock solid 99.5+
SQLServer:Cache Manager(_Total)\Cache Hit Ratio is 85+
AVG. \PhysicalDisk(SAN DATA Files)\% Idle Time over a 24hr period is around 80%
I've run profile traces looking for long running queries but nothing really sticks out. We also get occassional CPU spikes over 80% - 95% that may last longer than 5min. However there is no real pattern to when the spikes occur.
We run reindexing on the tables at least once a week and one of the databases that gets written heavily to is done nightly.
Any suggestions on where to start looking or what to try next.
As always thanks.
We are running sql 2000/sp4 on windows 2003 sp2. The server HP DL585 with 4 dual core cpu's, 48GB RAM (maxserver mem = 44GB) mirrored OS, RAID 10 (6 drives) for log files, EMC SAN LUN for data files (RAID 5 w/5drives), LUN for tempdb and separate LUN for paging file.
AWE is enabled and the patch for ...not all memory available...has been applied.
We also use Full-text search extensively and use 15 databases that drive our websites. There are a few others but they are not used much.
On any given day it can vary from a low of around 150 to in the thousands.
The Buffer Cache Hit Ratio is a rock solid 99.5+
SQLServer:Cache Manager(_Total)\Cache Hit Ratio is 85+
AVG. \PhysicalDisk(SAN DATA Files)\% Idle Time over a 24hr period is around 80%
I've run profile traces looking for long running queries but nothing really sticks out. We also get occassional CPU spikes over 80% - 95% that may last longer than 5min. However there is no real pattern to when the spikes occur.
We run reindexing on the tables at least once a week and one of the databases that gets written heavily to is done nightly.
Any suggestions on where to start looking or what to try next.
As always thanks.