If you want to avoid using the SDK, you can use the dfm cli to display data that you normally see in the perf advisor:

The cmd sets are:

dfm perf data retrieve

Example for total ops of all vols:

dfm perf data retrieve -o <filer name> -C volume:total_ops -M nov -s 3600 -m mean

Regards

Andreas

Yes, I was going to suggest that too.  If you download our SDK, there is a Windows tool called zexplore.exe (for ZAPI explorer, named for Zephyr – the code name for the original API project).  This will let you connect to a NetApp controller and … explore.  Be careful.  It actually executes APIs, so things like volume-destroy do exactly as the name suggests.  perf-object-* are the equivalent APIs to the stats command Rich Payne mentioned.  Any halfway decent OO perl hacker could whip up a script to pull the objects/instances/counters you want and make them all pretty.  Or you could look at the PowerShell tools we’ve posted on Communities.

Share and enjoy!

Peter

I’m not much on coding, but it sounds like the NetApp Manageability SDK may be what you are looking for.  http://communities.netapp.com/docs/DOC-1152

I asked a similar question on the Netapp community.  I got several hundred reads but not a single response.  Disappointing.

My question specifically was how to get the Perf Advisor data from DFM via the cli, specifically for scripting and throwing the data into other applications like mrtg.  Perf Advisor collects not only total iops but type of iop as well; read, write, or meta.  Very nice info but using PA all the time is too heavyweight for me.

If all you want is total iops, the sysstat parse, as already mentioned, on a periodic basis is simple enough to do.

I’m hoping a Netapp lurker might have an answer to my query though….