smudger
Enthusiast
Enthusiast

SRM 4.1: High Host CPU on HBA Rescan

We recently implemented v4.1 of SRM. Great product and looking forward to upgrading sometime soon to v5 (pending vCenter upgrade).

I had a question re having 4 or more VMs in one Protection Group and while testing failover or real failover, all hosts CPU hit 100% CPU while the rescan operation is going on. The cluster is running vSphere 4.1.1, 10x hosts, 20x 1TB LUNs running 300x VMs with approx 15 of those inside SRM.

Underlying storage is a 3PAR T800 array

Has anyone else seen this?

Thanks

0 Kudos
4 Replies
msatya
Enthusiast
Enthusiast

hi are testing one by one or all at a time

-------------------------
please award points if you see fit.Thanks

http://imagicon.info/cat/5-59/vbsmile.png

0 Kudos
smudger
Enthusiast
Enthusiast

Hi,

One recovery plan with 4x VMs is being tested. This is the only recovery plan being run at that time.

Thanks

0 Kudos
mal_michael
Commander
Commander

Hi,

Do you see high CPU usage during manual rescan also?

Michael.

0 Kudos
smudger
Enthusiast
Enthusiast

Hi, during a manual rescan the host behaved as normal ie.no high cpu usage/alarms.

I am double checking with our storage admins to see if there were any LUN presentations (or other storage operations) going on when the failovers were taking place.

Thanks

0 Kudos