[pgpool-general: 6190] pgpool killed with -9 signal refuses to start again

[Mathieu Ignacio]

Hi,


I'm writing a test plan for a customer and I found a weird behavior (to 
me). When pgpool runs and is killed with "killall -9 pgpool" it refused 
to start again (and there is no more pgpool processes).

Messages in log :
DETAIL:  bind on "/tmp/.s.PGPOOLWD_CMD.9000" failed with reason: 
"Address already in use"
DETAIL:  bind socket failed with error: "Address already in use"

If I delete .s.PGPOOLWD_CMD.9000, pgpool accepts to start again. Even a 
reboot does not permit pgpool to start again because Centos7 does not 
purge /tmp/ at boot.

pgpool version : pgpool-II version 3.7.5 (amefuriboshi)
OS : Centos 7
Package used : pgpool-II-10-3.7.5-1.rhel7.x86_64

Do I miss some configuration parameters, or should I open a bug ?


Thanks,


-- 
Mathieu Ignacio
NFrance Conseil
9 rue Ritay 31000 Toulouse - France
http://www.nfrance.com/