[Pgpool-general] on line recover script possible design flow

[Tatsuo Ishii]

Hi,

While testing pgpool-II 2.1 beta, I think I found a design flow in
pgpool-II's on line recovery script interface. Currently it defines
three arguments:

- database cluster path of master node
- host name of recovery target node
- database cluster path of recovery target node

Since user supplied recovery scripts want to execute "SELECT
pg_switch_xlog()" or some such on the master node, the script has to
know the port number on which the master node's postmaster is running
on.

Any thought?
--
Tatsuo Ishii
SRA OSS, Inc. Japan