Closed super-cache-money closed 2 years ago
Kudos on this library. I'm currently wrapping pg_recvlogical
, but it's difficult to build that out of the pg source tree, so it limits users to running it on a system with postgrestsql installed.
See pg_recvlogical
wrapped: https://github.com/JarvusInnovations/lapidus
I find the re-connection behavior of pg_recvlogical to be correct as Lapidus has been running for years in production without issue. Perhaps if I can add additional tests (that will fail if I switch to pg-logical-replication) to verify the connection handling I can provide some feedback.
The behavior I like here is if either process dies we don't continue to acknowledge/advanced the LSN position for the slot.
Hello, @super-cache-money
Check the server log to see if there are any messages below.
LOG: terminating walsender process due to replication timeout
Check out this document. https://www.postgresql.org/docs/current/static/runtime-config-replication.html#GUC-WAL-SENDER-TIMEOUT
Could you notify me if the same problem occurs after changing the wal_sender_time
value?
I have the same problem, but the stream only keeps working when using the proc
function provided in the example, which handles automatic restarting. Setting wal_sender_timeout
to 0 fixes the problem, but doesn't that mess up disconnection detection?
I have submitted a PR that should resolve this issue https://github.com/kibae/pg-logical-replication/pull/8
Firstly, thank you very much for making this library. I can't believe more people aren't capitalising on logical replication.
We have gotten your example working, but we keep getting the following errors, with
pg@6.2.2
:This error keeps throwing around once a minute, but the streaming still works and I can still see the DB changes.
Edit: The same occurs when we use
pg@7.4.0
.