June 15, 2004 at 1:39 pm
I have a Win2k box running SQL7.0 and I am trying to monitor several counters...of interest are %disk time and Buffer Cache Hit Ratio.
Right now, my drives are averaging 371% with a high of 3,224. Typically, I do not see such numbers...but when the drives are being hit hard, I see some really high counters.
Also, my Buffer Cache Hit Ratio been at 99.100 ALL DAY LONG. It has never changed yet.
Are there any system level setting or Reg changes that need to be made to bring these counters under control (or make them work at all)?
TIA
Ryan Hunt
June 18, 2004 at 8:00 am
This was removed by the editor as SPAM
June 20, 2004 at 6:52 pm
99.1% Cache Hit Ratio sounds good, in itself. The data being requested is found already in memory 99.1% of the time.
Given the high cache hit ratio, perhaps high disk usage is a little surprising. How many disks are there ? I think basically you can get 100% per disk adding together, possibly even bigger spikes.
It is not a contradiction if there is a high cache hit ratio most of the time, while disk activity jumps up now and then.
If both are occurring together, one possible explanation is a lot of table scans. Because they use 'read-ahead', it is treated as if it was in memory before it was 'requested', even though it is being read for the query.
Are you trying to measure (a) whether the application is usnig the database efficiently, or (b) whether your hardware capacity is adequate for the application. These measures address b-capacity rather than a-efficiency. Cache hit ratio is an indicator of adequate memory. Do you have some specific problem with application response time ?
Have you seen:
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply