VMware Cloud Community
Subversive
Contributor
Contributor

Sql 2000 resource usage issues

Hi, we're having some issues with a Windows 2000 SP4 VM running SQL 2000. It is a very busy database, though i don't have actual numbers.It has 4GB of RAM with PAE enabled, and 4 vCPU's. Lately it's basically been running wild,and we've been having timeouts and such when trying to access the DB. However, when I check task manager from inside the VM, it only shows 3-8% CPU usage, but the resource usage from the VI client shows it from 80-95%. Can anyone explain why this is?

Also, if anyone can point me to a good SQL troubleshooting spot, that would be much appreciated.

Thanks.

0 Kudos
2 Replies
OlivierTremolie
Enthusiast
Enthusiast

Hi,

When I use SQL Server onVM, I take care on :

  • Number of CPU (I suppose you've got at least 4 cores)

  • SAN configuration (I suppose you're using streaming)

The difference between task manager and VI is due to reservation. Task manager will show you CPU used based on maximum cpu ressources (3Ghz for exemple). But VI show you cpu used based on cpu reservation (250Mhz for exemple). This is why you have got difference between both. If you have got cpu ressources enough, you can make this test : setup minimum cpu reservation on your VM. By this way, your VM always got enought ressources and don't wait ESX to give it more ressources, and so, SQL performance will be better. But remember, a VM with 4 cpu will use 4 physical different cores. So, if your ESX has got only 4 cores, all other VMs will be impacted. On other side, if you've got 8 or more cores, other VMs will be a minimum impacted.

Olivier

###############

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

Olivier ############### If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!
0 Kudos
Adam-ProInteg
Contributor
Contributor

We too are experiencing timeouts running SQL2000 on Windows 2000 ADV /wSP4

We experience a network timeout about every 15-50 minutes and even the console is locked. We did a live snapshot and reverted to the image so I'm wondering if this could of did something bad to the VM, per some articles I read. We are running ESX 3.5 update 1. We used CPU affinity to assign 4 cores on our 8 core box. We bumped the memory share priority to high and reserved 7GB of RAM. We looked at everything we could but the problem is persistent.

All of the other VM's on either of our 2 ESX server are running fine.

Thanks in advance for your help,

Adam

0 Kudos