[Pgpool-general] what does "master node" mean in context of replication in pgpoolII?

[Tatsuo Ishii]

> When I enable replication in pgpool II 3.0.1, does the "master node"
> have any special significance? I have found (by experiment) that INSERTs
>  get propagated to all nodes instantly as far as I can tell, so how does
> one node differ from another node?

The idea "master node" is neccessary to keep consistency among
multiple DB nodes. When updating a DB, pgpool starts a transaction and
performs an update on master first, then does on other nodes. When
commiting the transaction, pgpool commits none master node first,
lastly master node. This way, we can avoid data inconsistency by
failure on a DB node by just aborting the transaction.

> There is a caveat, but rather small one:
> 
> "caution: Stop autovacuum on the master node (the first node which is up
> and running). Autovacuum may change the contents of the database and
> might cause inconsistency after online recovery if it's runnung."
> 
> Other caveats?
> 
> And how do I find out which node is master node from viewpoint of pgpool
> anyway?

Just the first live DB node in pgpool.conf.
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp