Having spent some time working with vSphere Replication I came across a number of issues trying to get my vSphere Replication Appliances to talk to each other and then to get vSphere Replication working.
The moral to this blog post is DNS and Networking!
DNS
Contrary to popular belief the DNS settings in the vSphere Replication Appliance appear to do err nothing.
I was receiving the Error Code ‘vSphere Replication Generic Server Error: No Route To Host’
After confirming my vCenter servers could resolve each other and also my vSphere Replication Appliances (as I had entered in A records for them) and the fact that I could ping everything, I decided to hop straight onto the vSphere Replication Appliances to test they could ping each other directly.
This ended in an epic fail as they didn’t have any DNS names for each other, so to resolve this I edited the host files on both vSphere Replication Appliances by entering the following commands:
vi /etc/hosts
i
172.19.144.149 VCT01.domain.local VCT01
172.19.146.149 VCT02.domain.local VCT02
(Press Escape Key)
:wq
After doing this my vSphere Replication Appliance could ping each other and the connection between the Appliances formed.
Networking
When I came to replicate the VM’s, a folder would be created for the VM and a VMDK file, however the VMDK would always remain at 0.00KB and when I tried to perform a manual synchronisation, I would receive the helpful error:
‘Call “HmsGroup.OnlineSync” for object “GID” on Server “” failed. An unknown error has occurred.’
After much head scratching, I realized we have two different default gateways, so I changed the default gateway on the VM which was being protected to the one being used by vSphere Replication, same issue occurred.
I then went over all of my default gateways for the following items:
- vCenter Server
- vSphere Replication Appliances
- ESXi Hosts
The last one was key, when I changed the default gateway on the ESXi Hosts to match the vSphere Replication Appliances, everything fell into place.
VMFocus 
Changing the default gateway of ESXi host as that of the appliance means that the replication works only on one particular subnet or vlan only???
Yep that’s my conclusion, I had the ESXi Hosts on the same VLAN as the VRA but using a different default gateway and it failed as the VRA uses the Management Network. You could introduce an alternative Management Network and then use the vSphere Web Client to enable vSphere Replication Traffic on this, but after this you still need to add manual routes in on the ESXi Hosts.
When I try to configure my virtual appliance on either side, I am getting this warning:
Self-signed certificate installed successfully.
Warning: Bad service status: execv() arg 2 must contain only strings.
Sounds like a DNS issue, does it work via IP Address?
Same warning with IP Address.