When you clone from VIC 2 what happens behind the scenes?

[dlevens]

I am trying to find out why my clones are going so slow recently. In the past I was getting 2GB per minute. I am not getting .1 to .2 GB a minute.

I am using the clone process to migrate 100s of VMs from 2.5.x to VI3. All my vms are on a 0808 Clariion and are going to a Clariion CX3-80.

I go into virtual infrastructure client 2.0.1 Patch 2 and choose clone, pick the new esx host and new lun and click go. The first few nights we were flying, now we are crawling.

EMC is asking what happens when I clone, does the command get passed to the console and then the console does a pull or push? Anything I can run to check performance?

I do know that nothing seems to pass over copper, it all appears to be going over the hba cards. My guess is LUN to LUN over fiber, but I have no idea if one lun pushes or pulls and where the command is run.

Dennis

[acr]

Ive always believed its simply running vmkfstools -i..??

[dlevens]

Possible, then the VC would handle the vmx portion and config changes. But how to confirm this. I need to track this down to what WWPN is talking to what WWPN to make it happen.

[bretti]

You might want to check your write cache status on your Clarion's. If that gets disabled, performance will go in the tank. Something as simple as the disk array not seeing the UPS hooked up can disable the write cache. (at least on the AX-150's)

As far as what is behind the scenes, it's always been my belief that vcenter issues the commands to the host and monitors their progress. All the data moving happens on the hosts. If the hosts in question can see both the source and target disks, then it should all be over FC, which would rule out any type of networking issue.

You may also want to check your SAN multipath settings, with the Clarion lines you want to use MRU instead of fixed multipath policies.