On Mon, Feb 16, 2009 at 1:29 AM, Tatsuo Ishii <[email protected]> wrote:
> Changing log_connections should not produce message 'S'. My guess is > your /etc/init.d/postgresql reload actually does stopping and > starting, or you changed something other than log_connections which > will produce message 'S'. See the ostgreSQL manual to know which > directive produce the message 'S'. The executed command was "pg_ctlcluster 8.3 main reload". In the PostgreSQL log I can see: LOG: incomplete startup packet LOG: received SIGHUP, reloading configuration files This "incomplete startup packet" message was discussed in the PostgreSQL mailing lists a few weeks ago, and it had been discussed before. Here you are what Tom Lane said about it: "The probable cause is something connecting to port 5432 and then disconnecting without sending anything. [..] I've heard that there's stuff out there that makes a cheesy probe that Postgres is "up" by just seeing if it can connect to 5432." http://archives.postgresql.org/pgsql-jdbc/2008-09/msg00099.php Also by Tom Lane: "Broken client software, or possibly something portscanning your machine. It's reporting a bogus connection attempt." http://archives.postgresql.org/pgsql-admin/2006-01/msg00184.php I will try to erase and rewind, see if I can reproduce it and post more details here. -- Jaume Sabater http://linuxsilo.net/ "Ubi sapientas ibi libertas" _______________________________________________ Pgpool-general mailing list [email protected] http://pgfoundry.org/mailman/listinfo/pgpool-general
