December 21, 2018 at 3:03 pm
I've created a new SQL server VM at work, and decided to run some benchmarks on it. HammerDB reports around 500K TPM.
I decided to see how it compared to our existing SQL server VM. To my surprise, the existing server only reports around 45K TPM.
How can i troubleshoot what the bottleneck is? Both VMs are located on the same ESXI host and use the same SSD datastores.
They're both SQL Server 2012, the new server is SP4, the old server is SP1.
UPDATE - To add to the confusion, after-hours I ran the same test and got around 400K TPM which is more normal.
I didn't see CPU, disk, etc usage that would have created the botteneck earlier in the day.
December 24, 2018 at 8:31 am
You'd need a way to measure the disk udsage for all VMs/processes at the same time. When you have a hypervisor in the way, it's possible that other processes are using some of your bandwidth.
December 27, 2018 at 7:36 am
I thought about that, but both VMs use the same datastore, so if it was disk contention they should both be effected. It seems to be related to the VM. I haven't been able to reproduce the results again yet.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply