Comment 17 for bug 1817484

I looked at the situation at a different angle:

1) MAAS was up the whole night without regiond restarts (had enough time to settle) on two surviving nodes;
2) Stored procedures that issue pg_notify are present on the new master: https://pastebin.canonical.com/p/SfYWjpzc2v/

https://github.com/maas/maas/blob/2.5.0/src/maasserver/triggers/websocket.py#L937-L955 (rrset_sipaddress_link_notify)
https://github.com/maas/maas/blob/2.5.0/src/maasserver/triggers/websocket.py#L1268-L1270 (register_websocket_triggers)

3) When I manually execute updates via `maasdb# update maasserver_dnsresource_ip_addresses set staticipaddress_id=<new_staticip_id> where id=<dnsresource_id>` the surviving region controllers do not get any notifications.

4) There is no way to list all current listeners via pg_listening_channels for all postgres sessions (only for the current session):
https://pastebin.canonical.com/p/Kc63bJr23T/

So I cannot say if there are any registered listeners.

5) `maas <profile> dnsresources read` returns the same value as stored in the DB at any given moment (so the connectivity to the DB is fine).

6) SELECT pg_notify('domain_update', "<domain_id>"); does not trigger updates.

So, the patched MAAS region still does not get notifications.