cxo
Contributor
Contributor

Disk Performance Overview Charts with NAS/NFS

I have recently migrated my shared storage from iSCSN SAN to NFS NAS. Since then I lost the vCenter Performance Overview Chart for Disk utilization for all VMs. I am currently running virtual Center 2.5 update 3 (to be upgraded to update next week) and ESX 3.5u3 hosts. I understand this goes from a "storage" to a "network" device, but curious if there is anything already out there I have missed or some other tool to get the same functionality as I was accustomed to with iSCSI storage (new version of ESX/vCenter, plug-in, etc)?

Thank you,

Charlie

0 Kudos
3 Replies
FranckRookie
Leadership
Leadership

Hi Charlie,

Did you found any solution for your needs? We also have NAS storage, on NetApp bays. Whatever vCenter version we use (2.5 u3, u4, u5, even 4.0) we don't have any information on virtual disks performance.

We followed two ideas: retrieve statistics from the bay or use command line tools inside the system console. NetApp tools do not give a very granular information if you have several vmdk on the same Qtree. And vScsi statistics tool is not simple to use on a day-to-day basis.

Please, let us know if you found any solution.

Thanks

Regards

Franck

0 Kudos
bulletprooffool
Champion
Champion

What are you using for NFS?

~You caqn use ESXTOP to view real time performance

Your NFS device could have its own monitoring tool for viewing overall LUN performance (certainly, the bigger players like Netappdo)

You could use a 3rd party tool like Cacti (free) to monitor individual VMs

One day I will virtualise myself . . .
0 Kudos
cxo
Contributor
Contributor

Franck, Bullet,

No I have not got a resolution. Indeed I can use esxtop, vscsi, or other LUN perfromance metrics (although I am concerned with NFS now), but it is not the same as the ol' iscsi performace overview chart. ESXTOP is useful for point in time things, but not on actual disk used (more for io wait). vscsi stuff is not exactly the same as from the performance overview metrics and it is cumbersome to run and anaylze and it takes a hit on the ESX host as well (plus I can't see history for a specific VM right now from two weeks ago, etc).

I am using NetApp but do not have the granularity to see information on a VMDK/VM level.

Cacti is nice (we use it here for many things), but will be cumbersome to have it retrieve information from many 100's of VM, etc.

Would be nice if that metric could be added back into vCenter (vSpher 4 update 1???).

Thoughts?

Charlie

0 Kudos