March 2, 2006 at 10:14 pm
Hi,
Last month i faced a problem that i want to share with u all and need suggestion,
The client have a clustered server raid 5(clustring not Enabled) 4 CPU with 32 GB RAM Database Size (+500 GB).
700 users.
The boot.ini is configured to use maximum RAM.
AWE is enabled in SQL Server.
SQL Server set to use Dynamic Memory.
As a DBA i use to pin the tables in Buffer.
for fast performance.
In last one year the server never reached to 100% CPU Hike.
Few daysback the Server reached to 100% CPU HIKE and remained same for long time and now it often gets the same hike.
I checked the server for CPU BOTTLENECKS , Page Faults but got nothing as a culprit.
I run SQL Profilor to see what is wrong, but i did not found anything except few long queries.
They took long time to execute but that is normal as i compared the CPU ,read time of last month with this month.
I looked for locks and latches but there also i get nothing. normal locking no dead locks.
I run dbcc traces but still all things are normal.
I run DBCC commands for memory status and datacaching but i did not get any information which proves that due to the reason the server is slow.
Pls help me where or in which area should in look to get the culprit for slow performance.
one thing i like to tell more is the target server memory and total server memory counter have the same values.
from
Killer
March 6, 2006 at 8:00 am
This was removed by the editor as SPAM
March 9, 2006 at 9:44 am
hi,Can you check on the size of your table? I experience this before due to the size of the table is too big but after the partition, the CPu time gone down.
sisi
March 9, 2006 at 9:16 pm
Dear Sisi,
I checked the Statistics it is ok as i scheduled a job for reindexing.
Also executes DBCC checkdb so if there is so problem with database.
It is fine. no Errors.
When all these things are ok then why u suggest to do partion of the table as the system have 32 GB of RAM
From
Killer
March 9, 2006 at 9:23 pm
SOrry I dont know why some other user name is attached to my user name.in the post
Dear Sisi,
I checked the Statistics it is ok as i scheduled a job for reindexing.
Also executes DBCC checkdb so if there is so problem with database.
It is fine. no Errors.
When all these things are ok then why u suggest to do partion of the table as the system have 32 GB of RAM
From
Killer
March 10, 2006 at 7:05 am
What do you mean you scheduled a job for reindexing? Do you reindex your database daily?
If you are using DBCC DBREINDEX that could be the problem. That command DROPS the indexes...so any query being run can't use the indexes until they are rebuilt. Further, it is an offline operation, so users can't use the tables while the table's index(es) are being rebuilt.
-SQLBill
March 14, 2006 at 2:37 am
Hi Bill,
I create a stored procedure and scheduled a job to reindex the indexex.
I scheduled the job twice a month.
I know the time the reindex in progress user cant use the tables.
But the problem exists before the table reindex and after the reindex.
can u pls tell me is Lsass.exe is a virus or a system file.Read in a website that LSASS is a virus but Windows help say it is a system file because it is using 35% of the CPU.
I cant afford to format the server.
from
Killer
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply