0

We are having major performance and latency issues with SQL Server 2019. We are running on a "quiet" VMware instance connected to a vSAN.

According to MSFT documentation, best performance is obtained if the drives are formatted at 64k block size vs 4k. Doing some research, VMware publishes that you should NOT use 64k but 4k block size if on a vSAN.

So which one is it and what is the science behind this?

Improve this question
4
  • 2
    Typically you want to follow guidance from your SAN vendor about this. VMWare is just an abstraction over everything and can't provide best practice guidance for all vendors. Even if you do this perfectly, it is not likely going to solve your performance issues, though. Commented Feb 5, 2023 at 19:17
  • 1
    Seconded Eric on "Even if you do this perfectly, it is not likely going to solve your performance issues". Rare it is due to hardware or hardware configuration that one's SQL Server instance experiences "major performance and latency issues" (unless there's something extreme like a failing disk). More likely you probably have an issue on the software side such as how the database is architected or how the queries are designed. Commented Feb 6, 2023 at 2:40
  • AFAIK the NTFS cluster size primarily makes a difference for allocations, not reads and writes, which always go by the allocation size not cluster size. So what makes you think that the issues are to do with the cluster size? More likely it's something else entirely, such as your storage network (switches, NICs etc). Commented Feb 6, 2023 at 13:22
  • Seems like block size and alignment that VMware refers to specifically are two different things. I have cleared the DMV and adjusted the Cost Threshold and MaxDOP and that solved most of the issues so far... Commented Mar 5, 2023 at 22:53

0

Reset to default

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.