We're using 3.5 update 1 (brand new install) on both hosts and cold migration / cloning of VMs fails with an Operation Timed Out error. Then when I look at the datastore of the destination host I can see the folder with the VM and the vmname-flat.vmdk file but none of the other files are there.
1. Am I likely to have a working/stable VM if I just scp those remaining files over and register the vm on the destination host?
2. Why is it failing? Same ESX version (same cd even) - same subnet pings ok. There is no VMKernel set up but all our hosts have their own datastores (no SAN) so we have to do a cold migration.
TIA
Check if you can ping from one host to the other host.
Are there local Storages or NFS? I think you need vkernel for transfering VMs within the management server.
But you can unregister the vm on one side, transfer it with scp to the other host and register it there.
Thanks for the prompt reply.
Yes I can ping from source to dest. Both hosts have local storage. Unfortunately I have no control over the VC server but have been given (I think) full control over our Datacentre.
I have not had a DNS entry created for my hosts - is this required to migrate? I didn't need it for ESX 2.5.3 -- I'm still very new to VI:3.5.
You should make an DNS entry or edit the host files of both esx servers, just to take it out of the error list.
Looks like I must edit my /etc/hosts with all my machines' names and IPs in order for the migration to work because just having the entries in DNS didn't seem to be enough.
Thanks all. It's working now.
Edit:
Oops, Sorry Packwhatever I meant to click correct answer on your last one. lol