Collected from the PG bugs email list.
|Reported by||Mircea Pirv|
Body of first available message related to this bug follows.
The following bug has been logged on the website: Bug reference: 16094 Logged by: Mircea Pirv Email address: (redacted) PostgreSQL version: 12.0 Operating system: OSX Description: Hello, We've updated our database from 10.6 to 12.0 recently and we keep encountering an error which says that the database is entering recovery mode. Our setup has replication enabled, with a replication slot, and using a replication subscription. On our development machines replication is setup to use the same machine. The problem is, that with replication enabled, changing a column value using an update statement, causes the database to enter recovery mode. This happens only when replication is enabled. We could not find any pattern, like what kind of columns might be the culprit. It happens with all kinds of columns, regardless of the fact that we have constraints, triggers, defaults on the columns. With some columns it happens randomly, with some it happens all the time. In the logs the only error we see, is a segmentation fault, when replication tries to run. We tried with both a local server, and a dockerized version (12.0). It only reproduces in 12.0. 10.6 and 11.5 do not have this problem, we could not reproduce it. Any ideas on what may have changed in 12.0 to cause this behavior ? Thanks, Mircea
|2019-11-04 14:50:45+00||PG Bug reporting form||BUG #16094: Database entering recovery mode|
|2019-11-04 15:03:45+00||Tom Lane||Re: BUG #16094: Database entering recovery mode|
|2019-11-06 13:37:31+00||Mircea Pirv||Re: BUG #16094: Database entering recovery mode|
|2019-11-07 02:35:51+00||Michael Paquier||Re: BUG #16094: Database entering recovery mode|
|2019-11-07 08:07:42+00||Mircea Pirv||Re: BUG #16094: Database entering recovery mode|