[pgpool-general: 3105] [pgpool 3.3.3-4] - defunct processes while failover

[Guy Meler]

Hey all :)
I`ve 4 pgpool running in active-active mode (no virtual ip), with
master/slave backends.
Sometimes, when I initiate a failover , none of the pgpool servers runs the
failover.sh script and they go, as best to describe it, into a deadlock
state. In each of everyone of them, all pgpools child processes are in
defunct state, I can't kill them by `service pgool stop`, and I can't make
any new connections to them. The only thing left is to kill them manually
and clean up status files (/var/run/pgpool,/var/lock/subsys/pgpool, /tmp
sockets).

In the log files, all pgpool servers are trying periodically to connect to
the failed node - "check_replication_time_lag - couldn't connect to DB 2".

Thank you very much.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20140817/df67b93a/attachment.html>