When checking the log files, the following entries kept appearing every second: synapse.federation.sender - 1030 - INFO - wake_destinations_needing_catchup-0- Destination DOMAIN has outstanding catch-up, waking up.

These are often notifications to old servers that can no longer be reached.
This spams everything, especially if you have been travelling in larger groups.

The first thing to do is to fix the cause in homeserver.yaml.
This limits the number of attempts and, above all, the maximum duration of the attempts:

federation:
  destination_min_retry_interval: 10m
  destination_retry_multiplier: 5
  destination_max_retry_interval: 14d

After restarting the service, this is effective, but only for new messages. So let's take care of the old ones, which can be done manually in the database:

Firstly, make a backup of the tables:

# pg_dump -h HOST -U USER -W DATABASE -t TABELLE > BACKUP.sql
pg_dump -h 10.0.0.11 -U synapse_user -W synapse -t destinations > synapse_destinations.sql
pg_dump -h 10.0.0.11 -U synapse_user -W synapse -t destination_rooms > synapse_destination_rooms.sql

The two tables can then be emptied:

psql -h 10.0.0.11 -U synapse_user -W synapse
DELETE FROM destinations;
DELETE FROM destination_rooms;
\q

The BSDBox is my IT playground. Professional services around network, server and structure building I offer with my company computing-competence. If you find this content valuable and useful, I welcome feedback via matrix, email or follow me on Mastodon.

Previous Post