MyqueWooMiddo
commented
4 years ago how do you build the multi-master cluster ? I 've built a 3-node cluster . all nodes are among master ,named A , B ,C A subscribes to B & C B subscribes to A & C C subscribes to A & B
If I insert a record in A , I can see that from B & C , but I update that record in B , there's no effect in A & C , while there's huge log flushed out like ' apply UPDATE from remote relation public.tbl in commit before 5/38C01B40' in all nodes
all nodes set pglogical.conflict_resolution to 'last_update_wins' already
So, could you teach me how to construct an workable multi-master cluster ?
We have set up a group of databases in a multi-master configuration using pglogical.
Are there any queries that can be used to monitor replication status, and provide us with an alert when replication is not working? A text-based query is fine; we can run the query periodically and pump the results into icinga2.
An example of what I'm looking for: we had an incompletely applied schema update a few weeks ago which blocked replication, but the problem was not caught for several hours. If this is not the best way to approach the problem, any alternatives would be appreciated.