October 25, 2007 at 11:34 am
We had a ping/post process, external webservices and internal clients all hitting the same database and server.
It would take 8 seconds for one query to run on the server and there was quite a bit of conention.
We setup replication to another server and replicated a few articles to the other server.
We then moved the ping/post process from the original server to the new server.
The ping/post process is working great.
However, the same proc that took 8 seconds to run on the original server, is now taking 30-40 seconds.
At first I thought it may have been the replication, but we stopped and disabled replication, but still had the same results. As soon as we point the ping/post back to the original server, the execute time goes back down to 8 seconds or less.
Why would things slow down AFTER you move some of the processes to another server.
I ran a profiler for memory, cpu and disk i/o, but didn't see anything out of the ordinary.
Could it be an issue with the caching? How can I track that?
Thanks in advance for any help
October 25, 2007 at 12:32 pm
Any chance there's an issue in the network between the 2 servers so that when you moved and were usign the process across both servers the nics may have been out of synch or something (I.E both set to auto instead of set to a specific value).
October 26, 2007 at 6:56 am
Are the indexes the same between the original and new server? By default, nonclustered indexes are not copied when a table is replicated.
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
October 26, 2007 at 7:34 am
Also - are the network links between the two slow? It may at least partially be due to network latency. If so - you might consider have a vlan and/or an additional connection set up between the two (secondary NIC, etc..)
Still - 30 seconds is a lot or network latency....
----------------------------------------------------------------------------------
Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?
October 26, 2007 at 2:55 pm
Are the indexes the same between the original and new server? By default, nonclustered indexes are not copied when a table is replicated.
I did add the indices, but the server B we replicating to isn't the issue.
I also removed replication from ServerA and still experienced the same slow downs
It's AFTER we moved process of server A that ServerA was slower than before.
Could it be a caching issue? Now we aren't using the server so much, nothing is left in cache?
Not sure about how to sync a nic card...
October 28, 2007 at 6:27 am
Are you making you connections by ip, hostname etc? If it's by hostname, make sure you are resolving the ip address of the new server correctly. Make sure that your server isn't just failing DNS/Wins and dropping to NetBios name resolution.
-Luke.
October 29, 2007 at 12:25 am
SG (10/26/2007)
It's AFTER we moved process of server A that ServerA was slower than before.Could it be a caching issue? Now we aren't using the server so much, nothing is left in cache?
Oops. I missed that point. I don't see how it can be caching..
Can you post some of the code please? The stuff that's performing eratically and the ping/post process if possible?
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
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply