January 29, 2008 at 2:54 pm
Does anybody know a good way to identify what is fragmenting my MemToLeave area?
I am using LiteSpeed SQL 2005 on an SQL Server 2005 Enterprise SP2 default installation running on Windows 2003 Enterprise SP2. Total RAM is 16 GB. I am using AWE, PAE, and 3GB. The min and max server memory are fixed at 14595 MB. The startup parameter -g is 512.
I just discovered an extended stored procedure (xp_memory_size) that LiteSpeed installs to show the largest contiguous block of MemToLeave is very low 4 mb.
Litespeed backup and restore is taking long time. we upgraded sql server 2000 to sql server 2005 inplace. before upgrade we removed lite speed and installed after upgrade. before upgrade there is no issue.
All ideas and suggestions will be appreciated.
Thanks.
January 29, 2008 at 11:28 pm
Do you have any linked servers in use? Using CLR?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 30, 2008 at 3:27 pm
There is 2 linked servers and no clr.
January 30, 2008 at 11:15 pm
Linked server drivers use the MemToLeave area. I've had a case of a buggly driver leaking memory in MemToLeave and causing slow backups and eventually a system crash.
The thread stacks are also in the MemtoLeave area.
Are you running x86, x64 or IA64 architecture? What drivers do those two linked servers use?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 31, 2008 at 10:44 am
Thanks for reply. We are using X86 and linked server we are using sql native client provider. we did inplace upgrade from sql server 2000 to sql server 2005. Automatically updated providers also.
January 31, 2008 at 10:55 am
Than I'm stumped. Sorry. Maybe someone else here has an idea. If you're got a support contract with MS, consider logging a case with PSS.
I've had very bad experiences with drivers for the AS400 on an IA64 box, but SNAC in x86 should be fairly well behaved.
What you could try is logging the results of xp_memory_size at regular intervals after a server restart. If you can find out when the memory gets fragmented, maybe you can figure out what's causing it.
Maybe unrelated, but if your max and min memory settings are the same, maybe reduce the min by a couple MB. I've heard (but it may just be urban legend) that setting max and min memory to the same value stops a couple of memory management routines from running.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 31, 2008 at 3:11 pm
I've had that issue with Idera SqlSafe before...their extended stored procedures fragmented it. I'm sure they have improved since I used it, but at the time the only solution they were able to give me was to bounce the service occasionally. Perhaps LiteSpeed is similar?
The Redneck DBA
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply