Controlling LUN queue depth throttling

[Mike-S]

Hi,

I have an ESXi cluster connected to two storage arrays. HP EVA 8400 and HP XP 24000. Recently, I have been having performance issues with the cluster and a check of the logs shows me the EVA storage array is having I/O congestion and returning a Queue Full status. I have found a KB http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&externalId=1008113 on controlling LUN queue depth throttling and am considering implementing the settings.

However, the KB indicates if I enable this setting and a Queue Full status is detected it will throttle all storage arrays. I don't have any performance issues with the XP 24000 so I am concerned this change will make things worse.

Any advice on this topic would be good. And maybe some other suggestions on how to address this Queue Full staus without throttling?

Thanks

[kopper27]

something at the array level? I mean your EVA I would first check your hardware-firmware-raid level-spindles...try to contact your storage vendor

once I read that changing queues is recommended when your working with your storage when you implement your infraestructure so you can see

increase performance when changing queues but I think is not recommended to do it whe you are in troubles and also in your case you have

2 storages...so this change can impact both