Microsoft Virtualization Discussions

Capturing Performance Stats for a Filer


I would like to know if thee is a method for capturing performance stats for a filer.  Primarily I am looking to capture Disk I/O, CPU, and Network statistics.



There are all sorts of ways.

Probably the best way is by setting up a NetApp Operations Manager server and using Performance Monitor to view the data you want to see.

If you need something a little less involved, there are a few quicker ways.  You can use the NetApp ONTAPApi toolkit for Powershell and the Get-NaPerfCounter, Get-NaPerfData, Get-NaPerfInstance and get-NaPerfObject cmdlets in a Posh script.

I've also queried occasional stats via VBScript and using SSH to run the stats command on the filer console.

Even more directly, you can log onto the filer and run sysstat -u 1 to see realtime stats on the console itself.

The best approach all depends on what you want to do with the stats you gather.


does netapp capture how much gets sent to certain clients? 

We have operations manager and I've seen the performance advisor stats, which are great, but I need something more detailed.

I have a pair of filers that are in our main building(location a).  We also have a datacenter one town over(location b) that has all of our other netapps.  We're planning on migrating all of the data from this netapp cluster, to the one at our datacenter.  The datacenter has hundreds of machines that potentially use the netapp in location A, and we have many employees/servers in location A that may also use this filer.  I know that the filer sends a lot more outbound network traffic than it receives, and I'm trying to gauge what precentage of the traffic is going to location A vs location B.

The good thing is the locations are 2 very different networks.  A = 172.x.x and B = 10.x.x so it should be easy to filter... I just don't know where I should be looking

Anyone have any thoughts in how I could go about looking into this?




I prefer Performance Advisor - which comes with Operations Manager.  However, the next best thing is the stats command.  The following blog entry does a fantastic job of explaining how to use this CLI tool to collect performance stats including disk I/O, CPU and network stats: