Just spent days troubleshooting vSphere Replication in the lab. Installation of the VRMS and VRS servers went without issue. Configuring replication on a VM completed without error. But when I would view the replication status, it always said "Not Active" and never replicated. When I hit "Synchronize Now" it would pop up the following error:
Call "HmsGroup.OnlineSync" for object "GID" on Server "" failed. An unknown error has occurred.
Investigating the vmkernel logs on the ESXi host with the source VM I found this:
2012-07-12T21:30:06.452Z cpu1:2049)WARNING: Hbr: 3855: Failed to establish connection to [10.11.49.44]:31031(groupID=GID-dc6e77d9-f516-408b-9d11-8ea3b315e253): VR: The set of disks on the server doesn't match up
Finally called support and they eventually found a relevant bug report.
Apparently if you use NetApp's mbralign tool to align your disks it can cause a miss-match between the protected and recovery sides for the disks and this error results. You also seem to have to have more than one disk attached to the VM for this to happen.
The fix is to edit the vmdk settings and add 5 MB to the disk size. You don't have to re-partition to actually use the 5MB, you just have to grow the disk by 5 MB. Then magically replication starts working.
That was a weird one that I find nothing on anywhere, so thought I would share. I'm sure VMware will put out a KB at some point.
