I was wondering if anyone has ever had an issue where processor utilization goes through the roof after it is vMotioned? We are getting crushed when this happens, but the issue resolves when the VM is rebooted. So how do I get the system to work properly without having to reboot?
Wow, not sure what you did here but this has never been a problem with the 20 odd SQL servers I had running at IndyMac Bank. Make sure you are using RDM LUN's for the DB and a minimum of 2 CPU's if this is production.
Thank you for responding. I just had the oppurtunity to speak with the Scott Drummonds on this issue. Apparently there is known bug for this and if I apply some updates the problem should go away. I guess it was just a matter of finding the right person to ask.
Had the same problem, added the following from a posting i found. Go to your cluster settings, under VMWare DRS, go to Advanced options and add this "MemOverheadGrowth" the value is "4".
Would you mind sharing the info on the updates? I occasionally see this on my VC VM when it is migrated. Sometimes the CPU stays pegged for up to an hour after the VMotion. I've 'resloved' this by stopping/starting the VC service, but even that takes a while to perform when the CPU is at 100%.
According to Scott Drummonds, a VMware employee who deals with technical marketing and is a regular poster to the performance community, part of the last service pack for VI3 was a patch to deal with this issue. He told me that the test group experienced the same issue. Scott told me to start with the updates first before opening a support call.
Hi Josh, just to let you know, we are having the same problem. I hit up VM at the Tech Bar while I was at VMWorld and I'm working with Keith Farkas on trying to resolve the issue. My issue started once I upgraded to 3.5 and I didn't have the problem before. I could vMotion all day long and my performance would not be affected.
