[pgpool-hackers: 627] Re: child_life_time doesn't work in 3.4beta

Muhammad Usama m.usama at gmail.com
Wed Oct 29 19:07:07 JST 2014


Hi Yugo.

Thank for pointing out the issue and the cause. I have pushed the fix
for it. The problem was the "connected" flag in do_child(), was being
rested to zero at every loop iteration. Which was wrong.

Kind regards
Muhammad Usama



On Wed, Oct 29, 2014 at 6:55 AM, Yugo Nagata <nagata at sraoss.co.jp> wrote:
> Hello,
>
> I found that child_life_time doesn't work in 3.4.
> You can confirm this by setting child_life_time shorter (e.g. 5),
> connecting to pgpool by psql and disconnecting immediately,
> and then monitoring pg_stat_activity at backend. The connection
> from pgpool remains even after 5 sec.
>
> child_lief_time is checked in do_child() at child.c;
>
>    if(pool_config->chil_life_time > 0 && connected)
>     {
>         ...
>     }
>
> However, "connected" is always 0 at this line.
>
> --
> Yugo Nagata <nagata at sraoss.co.jp>
> _______________________________________________
> pgpool-hackers mailing list
> pgpool-hackers at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-hackers


More information about the pgpool-hackers mailing list