topic Re: unidentified IOPS in StoreVirtual Storage

unidentified IOPS

oikjn — Fri, 18 May 2012 17:57:33 GMT

I see this every once and a while when deleting snapshots or restriping, but is there any way to figure out what is causing nodes to go under load?

I am monitoring each LUN and the cluster in general and don't see anything over 200 IOPs (including the total cluster), but I see an average of ~1500 IOPs on each node. I've found sometimes I just turn off the manager on the coordinating node and the "load" goes away, but what I want to know is if this is a load that is something from a node gone wild or if this is just typical background activity done while the cluster is under light activity?

Re: unidentified IOPS

David_Tocker — Wed, 23 May 2012 05:03:43 GMT

Sounds like typical activity to me - how are you monitoring this?

If it is from the CMC then it sounds pretty insignificant..

Re: unidentified IOPS

oikjn — Thu, 24 May 2012 14:00:31 GMT

its through CMC. my concern is that there might be some process that is crashing or otherwise stuck and generating a large load on each VSA. I know its typcail to have these high background loads during restripes and snapshot activity, but if that isn't obviously the case, how do I know if the load is normal or not?

If I turn off the manager for the coordinating node and then turn it back on I get all of this background load to stop and it doesn't come back... does that mean I"m killing some imporatnt function in the background? If not, why have that background load at all unless maybe thats just a brackground consistency scan or something, but it would be nice to have that documented.