|
sql
newsgroups
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
SQL Server 2005 Performance
We're seeing a strange thing occur with a process that only reads the
database via views and stored procs. After a certain point in the process a significant slow down occurs. We're not seeing anything sginificant in profiler to indicate a problem but setting the log file to simple and back to full either right before or while the process is running fixes the performance (we stumbled upon this workaround accidently). Does anyone know what happens with the database and/or engine when switching between full and simple? Thanks, sjg Are you taking frequent log backups in FULL Recovery? If not then you don't
need the database in FULL Recovery. Changing from FULL to SIMPLE truncates the active portion of the log and makes it immediately available for reuse. What is the value of the log_reuse_wait_desc for the database in question when the problem occurs? select name, recovery_model_desc, log_reuse_wait_desc from sys.databases Show quoteHide quote "sjg" <s**@discussions.microsoft.com> wrote in message news:AF89BA2C-2140-4166-9757-13502C1EBC13@microsoft.com... > We're seeing a strange thing occur with a process that only reads the > database via views and stored procs. After a certain point in the process > a > significant slow down occurs. We're not seeing anything sginificant in > profiler to indicate a problem but setting the log file to simple and back > to > full either right before or while the process is running fixes the > performance (we stumbled upon this workaround accidently). Does anyone > know > what happens with the database and/or engine when switching between full > and > simple? > > Thanks, sjg Hi Jonathan,
Yes, we have a transactional system so we are taking log backups every 30 or 60 minutes. The database does need to be in FULL mode. However, we've seen the problem when the database is set to SIMPLE mode and the problem goes away when we set it back to FULL. It seems that making any change to the recovery model is what speeds the process back up. I will try and find out the log_reuse_wait_desc sometime today. Thanks, Sarah Show quoteHide quote "Jonathan Kehayias" wrote: > Are you taking frequent log backups in FULL Recovery? If not then you don't > need the database in FULL Recovery. Changing from FULL to SIMPLE truncates > the active portion of the log and makes it immediately available for reuse. > What is the value of the log_reuse_wait_desc for the database in question > when the problem occurs? > > select name, recovery_model_desc, log_reuse_wait_desc > > from sys.databases > > > -- > Jonathan Kehayias > SQL Server MVP > http://jmkehayias.blogspot.com > http://www.sqlclr.net > > > "sjg" <s**@discussions.microsoft.com> wrote in message > news:AF89BA2C-2140-4166-9757-13502C1EBC13@microsoft.com... > > We're seeing a strange thing occur with a process that only reads the > > database via views and stored procs. After a certain point in the process > > a > > significant slow down occurs. We're not seeing anything sginificant in > > profiler to indicate a problem but setting the log file to simple and back > > to > > full either right before or while the process is running fixes the > > performance (we stumbled upon this workaround accidently). Does anyone > > know > > what happens with the database and/or engine when switching between full > > and > > simple? > > > > Thanks, sjg > > >
Other interesting topics
testing 2000 to 2005 update
Query is very slow then fast bak file = 5gb, attempted restore claims its 100Gb bcp delimiter Transactional data fun... Performance Dashboard giving incorrect missing index using DATEADD to report informaiton from 60 days ago SQL 2005 Management Studio truncate log Trigger returning DBCC input buffer statement being cut-off |
|||||||||||||||||||||||