High Disk Latency / How to check HBA saturation?

[JESX35]

Hello,

Currently we are having troubles with one of our ESXi 5 clusters.  From time to time the VM's in these clusters slow down to a crawl, however when looking at the performance of the VM's and host all seems well within the normal except the Disk Latency.  We see very high disk latency but low usage.  We have been looking at the SAN as the possible problem, however some members of our team think it may have something to do with the HBA's as well.  What is the best way to find out if the HBA's on the system are maxing out or running above normal ?  Also what is a high number for commands per second on a HBA ? for instance anything over 100ms in disk latency for extended periods is never good, what would a threshold be for HBA'a ?

Thanks

[mcowger]

Look at the throughput numbers for the HBA - if its 80% of what the HBA is capable of, you may be getting near saturation.

Theres no equivalent commands/sec # for an HBA - they are effectivly passthru for that stuff, and can handle hundreds of thousands.  In all honesty, I've never seen an HBA get saturated before the array - I'd look at the arrray first, given what you've said.

[JESX35]

Which counters are you looking at when you are looking at the HBA's throughput numbers the read rate / write rate ?

[mcowger]

Exactly.