Re: Add 'worker_type' to pg_stat_subscription

Nathan Bossart <nathandbossart@gmail.com>

From: Nathan Bossart <nathandbossart@gmail.com>
To: Peter Smith <smithpb2250@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-09-07T22:28:34Z
Lists: pgsql-hackers

Attachments

On Thu, Sep 07, 2023 at 12:36:29PM +1200, Peter Smith wrote:
> Modified as suggested. PSA v3.

Thanks.  I've attached v4 with a couple of small changes.  Notably, I've
moved the worker_type column to before the pid column, as it felt more
natural to me to keep the PID columns together.  I've also added an
elog(ERROR, ...) for WORKERTYPE_UNKNOWN, as that seems to be the standard
practice elsewhere.  That being said, are we absolutely confident that this
really cannot happen?  I haven't looked too closely, but if there is a
small race or something that could cause us to see a worker with this type,
perhaps it would be better to actually list it as "unknown".  Thoughts?

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

Commits

  1. Add worker type to pg_stat_subscription.

  2. Simplify determining logical replication worker types.