On Wed, 30 Aug 2023 at 01:01, Jim Jones <jim(dot)jones(at)uni-muenster(dot)de> wrote:
> However, pgbouncer closes with a segmentation fault, so I couldn't test the result of pg_basebackup itself - but I guess it isn't the issue here.
Oops it indeed seemed like I made an unintended change when handling
database names that did not exist in pgbouncer.conf when you used
auth_type=hba. I pushed a fix for that now to the replication-support
branch. Feel free to try again. But as you said it's unrelated to the
postgres patch.