[pgpool-general: 3222] PgPool with Virtual IP

[Jose Baez]

Hello!

I have deployed Pgpool 3.3.4 with 2 Psql nodes in a master-master
configuration.

Pgpool is installed in both Psql nodes and only one instance of Pgpool is
active at same time (with pgpool-HA, corosync/pacemaker)

When a psql-node fails, what's the best way to run an automatic
online-recovery? (I know it is not recommended)

Should I run "pcp_recovery_node" as a "failover_command" in pgpool.conf ?

Should I run the automatic online-recovery within Pacemaker? Or within an
individual bash script?


Another question, would there be a specific situation in which if Pgpool
starts in node-2 it sees the psql nodes in differente status as the Pgpool
which ran previously in node-1? Would it be a good idea to store
"pgpool_status" in a shared drive, from where each Pgpool instance loads
the same file?

Would a master-slave streaming replication configuration be more versatile
in a High-Availability environment?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20141014/7fc5d832/attachment.htm>