Hi,
I do have one of my MS-SQL servers that it is having a LUN/VMware data storage with extremely poor and high write latency values. This is the only LUN that is showing this problem. Average write latency is around 500ms, even more, is horrible. All the mdf files on that LUN exhibit the same high write latency values.
Now, here's the interesting part. I moved the database client to another of my virtualized SQL boxes. That new server has two LUNs with a 1.5TB database on one of them. After I moved the 50GB client to the new server, all of the sudden the new LUN and mdf file started exhibiting the same poor writes that it was showing on its previous server.
The storage I/O control is disabled. It is properly formatted for MS-SQL (64KB cluster size) and it's fixed. The file system is VMFS 5.54 with a 1MB block size. And I should add that these specs are identical on all other servers.
Now, I am a professional MS-SQL DBA. I am not the VMware/SAN admin. But the latency values on this particular LUN does not look right to me.
Any ideas of what I should check or suggest to my VMware admin so he can check for me? Could be that the path to all these new LUNs or something is not properly configured? What should I check?
Thanks in advance,
PS: The SAN is a DELL COMPELENT, auto tier system. The size of the LUN is 2TB.