Re: Logical replication is missing block of rows when sending initial sync?

On Fri, Nov 10, 2023 at 12:44:17PM +0100, hubert depesz lubaczewski wrote:
> > I think it'd be interesting to look at the WAL using pg_waldump, and see
> > how it relates to the LSN used for the tablesync slot.
> 
> well, I can try to dump, and show data, but it will most likely be
> a lot. Are you more interested in looking into wal around the second
> where data was inserted, or when table was added to publication?

I tried to find it based on relfilenode of assignments table, and ctid,
but I can't find it in waldump :(

I'm probably doing something wrong, or basing on wrong assumptions, but
I got 33GB of wal, dumped it all, found relfilenode (41441) and grepped
for this, and block number (ctid       │ (24192,1), so i assumed block
24191, 24192, or 24193 - nothing like this exists in wal dump.

Best regards,

depesz