Date: Tue, 02 Nov 1999 20:58:27 -0800
From: Rich Ryan <postgres@weblynk.com>
Subject: Inifinite socket send from backend to java servlet
To: pgsql-interfaces@postgreSQL.org
Message-id: <00ab01bf25b8$1192b260$9befaace@usedcars.com>
MIME-version: 1.0
Content-type: text/plain; charset="iso-8859-1"
Content-transfer-encoding: 7bit
References: <381978E3352.179CSAKAIDA@smtp.psn.ne.jp>

I'm running RH Linux 5.2, Jdk 1.1.7v3 on machine A, and RH Linux 6.1,
postgres 6.5.2 on machine B. Some queries made from A to B never return. It
looks like the java servlet on machine A skipped out on the query or the
thread died or got interrupted or something, and machine B is infinitely
waiting for it to read query results. Doing a ps -x will show several
postmaster processes doing SELECT's, but just sitting there, ( not SELECT
waiting), but not consuming any cpu time either. When I restart the web
server on machine A and then immediately look at the postgres log file on
machine B, I get a ton of these messages...

FATAL: pq_endmessage failed: errno=32
pq_flush: send() failed: Broken pipe

Within a half day or so, my database connection pool is exhausted because
there taken up by these infinite waits. Of course, I could write a db
connection timeout thread that closes these lazy connections, and I suspect
that the postmaster processes on machine B would detect the broken
connection and go away also, but still... seems like pretty icky behavior.
Any comments?
Rich