CacheHitRatio
in [Server Configuration]
|
From: Abba on 7 Apr 2010 00:11 Hello, [SQL2005] We use a monitoring software that checks various system resources and alerts us based on the threshold values. As part of this, it tracks the CacheHitRatio where it is set as to error when the value goes down 95% and warn when it does down 98%. Of late we notice that the alert triggers off frequently and when we check the memory, it shows as fine. On a day this parameters fluctuates a lot and now it is very close to being tagged as a False alarm and thus we plan to remove/disable it. What is your experience with CacheHitRatio? Is it a good measure to figure out the memory issues for a server? If not, how do you check if there is memory crunch in your SQLServer? tia, AbbA
From: Greg Linwood on 7 Apr 2010 03:30 Cache hit ratio is total baloney when it comes to measuring cache performance. As you've seen, it's nowhere near transient enough to derive any meaningful analysis. Unfortunately, measuring cache / memory usage in SQL Server is a big fat mess, but you'll probably get a lot more mileage with the Buffer Manager's Page Life Expectancy counter than with Cache Hit Ratio. Cheers, Greg Linwood SQL Server MVP "Abba" <sql_help(a)aaa.com> wrote in message news:uMl3Zig1KHA.264(a)TK2MSFTNGP05.phx.gbl... > Hello, > > [SQL2005] > > We use a monitoring software that checks various system resources and > alerts us based on the threshold values. As part of this, it tracks the > CacheHitRatio where it is set as to error when the value goes down 95% and > warn when it does down 98%. > > Of late we notice that the alert triggers off frequently and when we check > the memory, it shows as fine. On a day this parameters fluctuates a lot > and now it is very close to being tagged as a False alarm and thus we plan > to remove/disable it. What is your experience with CacheHitRatio? Is it a > good measure to figure out the memory issues for a server? If not, how do > you check if there is memory crunch in your SQLServer? > > > tia, > AbbA > > > >
From: Abba on 8 Apr 2010 00:59 Thank you Greg! "Greg Linwood" <g_linwoodho(a)tmaildotcom> wrote in message news:uPGRzQi1KHA.2028(a)TK2MSFTNGP05.phx.gbl... > Cache hit ratio is total baloney when it comes to measuring cache > performance. As you've seen, it's nowhere near transient enough to derive > any meaningful analysis. > > Unfortunately, measuring cache / memory usage in SQL Server is a big fat > mess, but you'll probably get a lot more mileage with the Buffer Manager's > Page Life Expectancy counter than with Cache Hit Ratio. > > Cheers, > Greg Linwood > SQL Server MVP > > "Abba" <sql_help(a)aaa.com> wrote in message > news:uMl3Zig1KHA.264(a)TK2MSFTNGP05.phx.gbl... >> Hello, >> >> [SQL2005] >> >> We use a monitoring software that checks various system resources and >> alerts us based on the threshold values. As part of this, it tracks the >> CacheHitRatio where it is set as to error when the value goes down 95% >> and warn when it does down 98%. >> >> Of late we notice that the alert triggers off frequently and when we >> check the memory, it shows as fine. On a day this parameters fluctuates a >> lot and now it is very close to being tagged as a False alarm and thus we >> plan to remove/disable it. What is your experience with CacheHitRatio? Is >> it a good measure to figure out the memory issues for a server? If not, >> how do you check if there is memory crunch in your SQLServer? >> >> >> tia, >> AbbA >> >> >> >> >
|
Pages: 1 Prev: Unable to script database Next: sp_who2 and system background processes |