Im running a CentOS 6.6 VM on ESXi 6.0.
The VM is experiencing extreemly high wait states and very little throughput.
When the VM was running on ESXi 5.5, we had no problems with throughput.
I need some help diagnosing what's going wrong with the throughput.

Device: rrqm/s wrqm/s r/s w/s rMB/s wMB/s avgrq-sz avgqu-sz await svctm %util
sda 0.00 0.20 0.00 0.60 0.00 0.00 10.67 0.00 2.33 1.33 0.08
sdb 0.00 863.60 0.00 4.40 0.00 2.20 1024.00 147.50 90122.23 227.27 100.00
dm-1 0.00 0.00 0.00 0.80 0.00 0.00 8.00 0.00 2.25 1.00 0.08
dm-2 0.00 0.00 0.00 870.40 0.00 3.40 8.00 18880.87 58314.25 1.15 100.00

The disk(s) are raw-mapped luns from a Clariion disk array. From what I can tell with other hosts attached to the Clariion, the problem is not on the array. I do not have any array diagnostics capability, but other hosts are not seeing any performance issues, only the VM's which are running on ESXi 6.0.
Even a simple 'sync' command is taking minutes to run...

Quote:

other hosts are not seeing any performance issues, only the VM's which are running on ESXi 6.0.

All the companies VMS running on ESX 6 are having IO issues or just this VM on ESX 6 is having an IO issue?
Are you mapping the LUN over storage fabric or ethernet? Have you compared the mappings between the two adapters
Do you have multipath or powerpath running? Do you have the most recent version of powerpath (assuming you're using powerpath as a replacement for multipath considering this is a ClARiiON array)
What kind of adapters are you using? Have you tried building new ones and failing over to them?

The ESXi 5.5 VM's are running fine, only the ones running on ESXi 6 are having issues.
If I move a VM from ESXi 6 back to ESXi 5.5, then it runs fine.
The multipathing is being done by ESXi itself, not the VM.
Both ESXi boxes were at 5.5 previously, and I upgraded one to ESXi 6 and started seeing the problem.

Im hoping someone can point to a log file, or metric which can be measured to tell me if it's specifically ESXi 6 which is the problem or the box that ESXi 6 is running on.

If I can't make any headway, I may try rolling back to ESXi 5.5 and see if the problem goes away. That's a shot in the dark though, not based on any real metrics.

I assume you did a clone or simply imported the vm to esxi instead of a clean install?? Try a clean install maybe to see if it changes.

and of course, make sure your VMWare tools are current....