Disk Latency

[bsmitty23]

Hi All,

Not sure if this is the right spot for this question, so appologies if it isnt. I've just started moving some data over to Netapp from our HP SAN, and I've just been checking out the performance statistics of my controller via Netapp System Manager. I really don't know what are normal numbers, but my disk latency seems rather high, and I'm not sure if it is, or if its normal. I only have one LUN created for VMWare, which is a 600GB LUN with a 350GB Volume behind with Dedupe turned on. Do these numbers look accurate, or would this indicate a performance problem? I have 47 disks in the aggregate, all Fibre 146GB 15K dirves. I've attached the image of the graph. Like i said, I'm just looking for information on whether this is normal, or if that sort of latency would indicate a problem.

Thanks,

Brent

[cedric_renauld]

Hi,

First question in VMware environement,

Have you mbralign block on your VMware LUN ?

Second for you: The DEDUP is in auto configuration or in sched configuration ?

[bsmitty23]

We have not done the MBRALIGN, and we've been getting really good dedupe rates since I guess all of our VM's are misaligned. For dedupe, I have it set to scheduled, with the default settings.

Thanks for the help

[amiller_1]

Have you setup Operations Manager in your environment? More important than Operations Manager for something like this is Performance Advisor (which comes with it) as that lets you see very fine-grained latency, disk busy, aggr busy, IOps per volume/LUN/aggr, etc. and set up thresholds and alerting on pretty much everything you could want.

[kvinay]

Hi Brent,

              The latency that you are seeing in System Manager is protocol latency , and if you are thinking it is something to do with issue in disks i suppose it is not .It is indicating the response of the underlying protocol that handles data, if you point against the latency dashboard will show the details of the protocol .

Regards,

Vinay