[pgpool-general: 4828] i think i've found a bug in config parser.

[Tema Zelikin]

Hello.

So, usually you describe backend in pgpool.conf like that:
backend_hostname0 = 'IP'
backend_port0 = 5432
backend_weight0 = 1
backend_data_directory0 = '/var/lib/postgresql/9.5/main'
backend_flag0 = 'ALLOW_TO_FAILOVER'

Which works fine. But then, when you try to describe backend like that:
backend_hostname0 = 'IP'
backend_port0 = '5432'
backend_weight0 = 1
backend_data_directory0 = '/var/lib/postgresql/9.5/main'
backend_flag0 = 'ALLOW_TO_FAILOVER'

- you'll get following problems:
1. pgpool wont see any backends. it wouldnt even try to connect to them
without any messages.
2. in case you're using watchdog with delegate_IP - first pgpool node
will start (but with problem no.1), but the second one will crash with
the following errors:
Jul 27 13:41:00 localhost kernel: [ 3648.050867] pgpool[12264]: segfault
at fffffffffffffff0 ip 0000000000458433 sp 00007ffe523e24a8 error 5 in
pgpool[400000+134000]
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12260:
WARNING:  watchdog child process with pid: 12264 was terminated by
segmentation fault
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12302:
DETAIL:  connect on "/var/run/postgresql/.s.PGPOOLWD_CMD.9000" failed
with reason: "Connection refused"
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12302:
ERROR:  get node list command reply contains no data
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12301: LOG: 
IPC socket path: "/var/run/postgresql/.s.PGPOOLWD_CMD.9000"
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12301:
FATAL:  failed to create watchdog command server socket
Jul 27 13:41:00 localhost pgpool: 2016-07-27 13:41:00: pid 12301:
DETAIL:  bind on "/var/run/postgresql/.s.PGPOOLWD_CMD.9000" failed with
reason: "Address already in use"