Certainly your latency numbers are off the charts.  But what is so strange to me is the magnitude of this latency given how little IO traffic you're seeing.  Single digit IOPS are insignificant.

What does the array its service time is?  Are you seeing the same latency to the same volume from other ESX hosts?  How is the storage configured for the volume showing those high latencies?

Scott

Do you use iSCSI or FC?

I have seen this kind of latency with iSCSI and then the problem was with iSCSI adapter.

hi

this is a pretty large sata rg with a lot of luns on it. I am using fc.

could it be because I have too many luns on the rg and the drives are slow?

do you have navi analyzer ? If yes, what does it show you for Response time and Total IOPs for that LUN ?

no i dont have that. i have generated naz files and sent them to support. they said my sata drives are too slow and only capable of 80 iops

it's interesting that your graphs do not show that many IOPS, so could it be that other systems (non-VM) are using the same RG and generate a lot of workload ?

you would need to check with esxtop & compare with nar data.

perf. reporting (latency) on esx5 does not work: (which also implies any other reporting tool sitting on top of vcenter...)

http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=2012309&sliceId=1&docTypeID=DT_KB_1_1&dialogID=289804019&stateId=1%200%20289982769

you can capture the esxtop output to .csv if required:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1033346