Thread

Logical Streaming Replication stuck at specific LSN

Kristjan Mustkivi <sonicmonkey@gmail.com> — 2023-08-20T10:22:24Z

Hello,

I am doing a major version upgrade from Postgres standard version
11.15 (Debian 11.15-1.pgdg90+1) to Cybertec PGEE version 14.8_EE_1.1.5
(Ubuntu 14.8ee1.1.5-1.cybertec22.04+1). While I am waiting for the
support case to be processed, I was hoping to get advice from the
Community about how to check the situation meanwhile on my own.

The initial copy of the tables completed nicely and from Aug 18 6pm
until Aug 19 6am the streaming flowed without issues. And then
suddenly the WAL sending/write/flsh/reply got stuck:

 application_name │   state   │    sent_lsn    │   write_lsn    │
flush_lsn    │   replay_lsn
──────────────────┼───────────┼────────────────┼────────────────┼────────────────┼────────────────
 sub_whupgrade    │ streaming │ 3F5B3/7FBDDCD0 │ 3F5B3/7FBD49E0 │
3F5B3/7FBD49E0 │ 3F5B3/7FBD49E0

I restarted the subscriber but the only thing that happened was that
the LSN got now stuck with status "catchup" at exactly the same place.
I upgraded PGEE to version 14.9 EE 1.1.6 (Ubuntu
14.9ee1.1.6-1.cybertec22.04+1) but that also did not change the
situation.

With best regards,
-- 
Kristjan Mustkivi

Email: kristjan.mustkivi@gmail.com