[Pgpool-general] Weird end of failover/failback scripts execution

[Jaume Sabater]

Hi everyone!

I have set up a cluster with two nodes of PostgreSQL 8.3 and pgpool-II
2.1, with replication, connection pooling and load balance. It is
working fine and I can do online recovery with no problem.

I was using failover_command and failback_command in the form of "echo
something" and then I decided to call an external script as I wanted to,
first, log to the syslog, and second, be able to do something else, like
send an email or similar.

Here you are the parameters:

failover_command = '/var/lib/postgresql/8.3/main/pgpool-failover %d %h
%p %D %m %M'
failback_command = '/var/lib/postgresql/8.3/main/pgpool-failback %d %h
%p %D %m %M'

And this is the /var/lib/postgresql/8.3/main/pgpool-failover script (the
pgpool-failback command is identical but it echoes failback instead of
failover to the syslog):

#! /bin/sh
LOGGER="/usr/bin/logger -i -p local0.info -t pgpool"
BASENAME=`/usr/bin/basename $0`
ID=`/usr/bin/id -un`
$LOGGER "Executing $BASENAME as user $ID"
$LOOGER "Failover of node $1 at hostname $2. New master node is $5. Old
master node was $6."
exit 0

This is the relevant part of the syslog when shutting down the
PostgreSQL in the second node and doing an online recovery:

Nov 4 15:45:15 xen01 pgpool: 2008-11-04 15:45:15 LOG: pid 29203: execute
command: /var/lib/postgresql/8.3/main/pgpool-failover 1 10.0.33.253 5432
/var/lib/postgresql/8.3/main 0 0
Nov 4 15:45:15 xen01 pgpool[29245]: Executing pgpool-failover as user
postgres
Nov 4 15:45:15 xen01 pgpool:
/var/lib/postgresql/8.3/main/pgpool-failover: 15: Failover of node 1 at
hostname 10.0.33.253. New master node is 0. Old master node was 0.: not
found

[..]

Nov 4 15:47:13 xen01 pgpool: 2008-11-04 15:47:13 LOG: pid 29203: execute
command: /var/lib/postgresql/8.3/main/pgpool-failback 1 10.0.33.253 5432
/var/lib/postgresql/8.3/main 0 0
Nov 4 15:47:13 xen01 pgpool: 2008-11-04 15:47:13 LOG: pid 29239:
recovery done
Nov 4 15:47:13 xen01 pgpool[29366]: Executing pgpool-failback as user
postgres
Nov 4 15:47:13 xen01 pgpool:
/var/lib/postgresql/8.3/main/pgpool-failback: 15: Failback of node 1 at
hostname 10.0.33.253. New master node is 0. Old master node was 0.: not
found


Now my question is: what is that ": not found" at the end of the script
execution doing there? I've been trying to figure it out but no luck. It
did not happen when the failover_command and failback_command looked
like "echo something > file.log".

Any hints? Thanks in advance.

-- 
Jaume Sabater
http://linuxsilo.net/

"Ubi sapientas ibi libertas"