Source and _replica name collision?

[B.F.]

Greetings,

Been experimenting with VM replica's and Failover's. One thing that I can't quite figure out. How is it when one does a Failover Now, the _replica is then powered up, and there is no name collisions the network? I now have both source and _replica up, they both have the same computer name, different IP range of course, and yet they both run fine as if they are different boxes? About the only thing that I could maybe see is that NETBIOS is disabled on _replica? When doing a NBTSTAT -A on the _replica IP, no name comes up but. However, I do see the name come up fine on the sources IP.

Just curious.

Thanks

[Vitaliy S.]

Hello,

I now have both source and _replica up, they both have the same computer name, different IP range of course, and yet they both run fine as if they are different boxes?

How does your current routing work in this configuration? Do you see all packets still going to the original server? How long have you been running both servers with the same name on your network?

About the only thing that I could maybe see is that NETBIOS is disabled on _replica?

As far as I remember the "Duplicate name exists" error shows up only when NetBIOS name of the Windows computer conflicts with some other name on the network. Since it is disabled, you do not see any warning messages.

Thank you!

[B.F.]

How does your current routing work in this configuration? Do you see all packets still going to the original server? How long have you been running both servers with the same name on your network?

Not quite sure how to answer the routing question. Both locations are within the same OU in the domain. What specifics are you looking for as far as routing? Both servers have been up for several hours now. I can see both addresses in DHCP with the same name however the MAC addresses are different. When I ping the name of the box, it always resolves to the source VM. I turned off the source for ~15 minutes, the ping requests would always resolve to the source.

As far as I remember the "Duplicate name exists" error shows up only when NetBIOS name of the Windows computer conflicts with some other name on the network. Since it is disabled, you do not see any warning messages.

Ahh, so Netbios is disabled. That would explain why no name comes up for the _replica.

Typically when doing Failover's, should the source be powered off?

[tsightler]

Typically when doing Failover's, should the source be powered off?

Yes, not doing so will lead to unpredictable behavior. If you want to test replicas they should be power-on in an isolated network using a SureBackup Lab using the SureReplica functionality.

Veeam basically makes the assumption that if you do a "failover now" that the original machine is offline so you're currently running with two machines with the same name. Based on how your infrastructure is configured this could be catastrophic or no big deal. For example, if the server can't be reached by clients, or cant' reach a DNS server to register it's new IP address then it probably won't cause any issues, at least temporarily, however, if it can do this, and clients can reach the new server, there's nothing to keep the replica from registering it's new IP with the same name as the original server and you may end up seeing client connections go to the replica which is probably not what you want.

[Vitaliy S.]

Yes, source VM should be powered off or located in isolated environment when you run a failover operation. As to the routing question, I was just wondering to what server DNS server resolves the network packet query.