**Reported in Launchpad by Mrten last update 30-10-2013 19:51:07

It has been recommended ( https://groups.google.com/forum/#!topic/codership-team/1fyLt83jmBc , other places ) that the server-id in a PXC cluster is the same for all members if the cluster is a replication master.

However, this means you cannot run pt-heartbeat on all members of the cluster (which would be nice for redundancy, ease of config and checking of replication lag). (I tried, and it creates transaction errors).

I noticed a 'server-id'-column in the heartbeat table, shouldn't there also be a kind of 'PXC-node-id' column? Perhaps the wsrep_node_name in there?