Message-ID: From: "papegaaij (@papegaaij)" To: "pgjdbc/pgjdbc" Date: Wed, 09 Jul 2025 12:48:37 +0000 Subject: Re: [pgjdbc/pgjdbc] issue #3724: Regression: FATAL: simple query "BEGIN" arrived before ending an extended query message In-Reply-To: References: List-Id: X-GitHub-Author-Login: papegaaij X-GitHub-Comment-Id: 3052550670 X-GitHub-Comment-Type: issue_comment X-GitHub-Issue: 3724 X-GitHub-Repo: pgjdbc/pgjdbc X-GitHub-Type: comment X-GitHub-Url: https://github.com/pgjdbc/pgjdbc/issues/3724#issuecomment-3052550670 Content-Type: text/plain; charset=utf-8 I totally get the issue with limited time, especially on open source projects like this and I really do not judge the decisions made. I think the PostgreSQL community is great, responsive, open minded and very helpful in many cases. See it from the positive side: the discussion-like ticket has now become a user-facing issue, so it will probably attract some more attention 😄 The problem is that at the moment pgpool-II users are a bit stuck between a rock and a hard place. I've also raised this issue at the pgpool mailing list. IMHO, a decision needs to be made to either support missing syncs or forbid them. In the first case, pgpool needs to be adjusted, in the later case pgjdbc needs to be changed. Unfortunately, I'm not in the position to make this decision, nor can I fix pgpool or pgjdbc as I'm not familiar enough with the details on the protocol. From the perspective of pgjdbc: would it hurt to do a sync when in extended mode and a simple query is about to be executed? The discussion in the thread on pgsql-hackers ended with 'it would probably be a good idea to throw an error from the server side'. This is what pgpool is currently doing, but apparently this was never implemented in PostgreSQL itself, otherwise this would most likely have been addressed already in pgjdbc.