[pgpool-hackers: 583] Re: [pgpool-committers: 2070] pgpool: Change connect() timeout longer (10 seconds) for flaky network.

[Tatsuo Ishii]

Hi,

I have changed connect() timeout (actually it's the timeout of right
after select()).  This is necessary for flaky network such as AWS.

This is for the known problem described in:
http://pgpool.net/mediawiki/index.php/FAQ#I.27m_running_pgpool.2FPostgreSQL_on_Amazon_AWS_and_occasionaly_I_get_network_errors._Why.3F

The fix is back patched to 3.3 and 3.2. 3.3 or before does not need
the patch since they do not use non-block connect().

> Change connect() timeout longer (10 seconds) for flaky network.
> 
> Change connect() timeout in connect_inet_domain_socket_by_port() from
> 1 second to 10 seconds for flaky network such as AWS. This should help
> reducing frequent failover. Pgpool-II 3.4 will have switch to control
> the timeout value, rather than fixed value. However our policy does
> not allow to add new feature like that. So I decided to just tweak the
> fixed value.
> 
> Branch
> ------
> V3_2_STABLE
> 
> Details
> -------
> http://git.postgresql.org/gitweb?p=pgpool2.git;a=commitdiff;h=1f4528e117d9906a052ceb3de81b156a1ff42ddc
> 
> Modified Files
> --------------
> pool_connection_pool.c |    4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
> 
> _______________________________________________
> pgpool-committers mailing list
> pgpool-committers at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-committers