Are their logs I could peruse that would confirm that I am having an issue with the scsi limit on the proxy?
as far as the hotadd restriction, yes I have 2 datastores that are using VMFS 3 out of the 30 associated to the cluster so I guess those default to NBD as a result. I increased the management vswitch to account for throughput. It was previoulsy 500 MB. The proxy is 4 GB.
The issue we see with the skipped VM's is when there is an error with snapshot removal on the host and the host begins to not respond disconnecting from the vcenter the consolidation completes but is skipped in the VM job.
I believe this could be more about VSA management of the job. Even if you set the readers to 2 or 3 you cannot tell the VSA agent to not perform more then 1 backup per ESX host at any time. So I see that the same host is cleaning up snapshots 2 or 3 at a time since I have it set to 3 readers to get through 120 VM's before business starts.
Are there any reg keys or anything that would allow me to enforce a rule to limit VM backups per host?