Message-ID: <gh-pgjdbc-pgjdbc-194-c577024605@github.com>
From: "ringerc (@ringerc)" <noreply+ringerc@github.com>
To: "pgjdbc/pgjdbc" <noreply+pgjdbc-pgjdbc@github.com>
Date: Wed, 22 Jan 2020 06:10:59 +0000
Subject: Re: [pgjdbc/pgjdbc] issue #194: PgJDBC can experience client/server deadlocks during batch execution
In-Reply-To: <gh-pgjdbc-pgjdbc-194@github.com>
References: <gh-pgjdbc-pgjdbc-194@github.com>
List-Id: <gh-pgjdbc-pgjdbc.github.com>
X-GitHub-Author-Login: ringerc
X-GitHub-Comment-Id: 577024605
X-GitHub-Comment-Type: issue_comment
X-GitHub-Edited-At: 2020-01-22T06:14:08Z
X-GitHub-Issue: 194
X-GitHub-Repo: pgjdbc/pgjdbc
X-GitHub-Type: comment
X-GitHub-Url: https://github.com/pgjdbc/pgjdbc/issues/194#issuecomment-577024605
Content-Type: text/plain; charset=utf-8

Actually, we can probably break the deadlock by using `Socket.setSoTimeout(int timeout)`. A whole lot of blocking calls would need wrapping, so it might make sense to start by doing this only in places where a deadlock is known to be possible. On send timeout try to consume input from the socket then resume the send attempt.

I still wonder if it's better to just switch to NIO and SocketChannel. It's probably not as hard as I thought above. We might need to wrap the rx channel in a buffer so we can push unconsumed input back onto it, but that's not too bad.