[pgpool-hackers: 2020] feature request: neat recovery after bad shutdown

[Sebastian Stellingwerff]

Hi,

I'm using pgpool2 3.5.4-1.pgdg80+1 on debian jessie and am using the
watchdog functionality.

I wanted to know what happened after a bad 'shutdown' pgpool (for
instance by an oomkiller) so did kill -9 on it. Recovery from it was a
bit tricky. Would it be possible for pgpool2 to do the following in a
new release?

1) if any socket it needs still exist, check if it actually still in use
and if not remove them before creating a new one.

2) if it knows it's not the master in the pgpool pool, remove it's VIP
from it's interface.

Thanks!

-- 
Sebastian Stellingwerff
06-49136664