All my VM's are getting very bad performance running Disk Bench at 1024KB blocks and beyond
What could cause this? Below is a snapshot of a SQL server which has been performing pretty poorly lately.
Storage is iSCSI StoneFly SAN. These benchmarks used to be very consistent.
Is the disk and file system blocks aligned? This may help http://www.vmware.com/vmtn/resources/608
~Ernie
VCP3,VCP4, MCSE, CCA
I have just realized that this is just happening on one particular ESX host...very strange
Any ideas?
compare the cluster size for your raid volume between the poorly and decently performing hosts.
I don't think it would be the storage raid block size since it happens to the same VMs but when I Vmotion them off to another ESX host they're fine. When I vmotion back, they'r effed again.
I'm thinking bad cabling or something particular to that ESX host. Bad NIC maybe?? Whats weird is that it happens over both paths to the storage (I have 2 nics and multipathing setup for iSCSI) I disable one NIC and and run the test and its bad...I disable the other run again and still bad.
Will do some more digging when I get to work tomorrow.
Any other suggestions? What else could it possibly be?