Netty's backups have the wrong the nodeID

[borges]

Hi,

In trunk, Replicating Netty servers are failing to notify that the backup is up. The exception is always reproducible in, say, NettyReplicatedFailoverTest. It boils down to the Netty backup/live pairs not having the same nodeID.

There isn't anything in place to force these to have the same nodeID.

Is there any straightforward way for the backup to determine the nodeID of its live? By that I mean, is there a way to get the live's nodeID before assigning a nodeID to the backup itself?

Such as having the backup connecting as a client to the live, and then determining the nodeID, would that work?

[clebert.suconic]

There isn't anything in place to force these to have the same nodeID.

Right... synchronization should set the nodeID on the backup. as we talked earlier today on irc