Dimitri Fontaine

I'm unclear if you tested with use_sudo = False in your pg_staging.ini setup?

Could you run the unpatched pg_staging in debug mode and paste the output?

That's a good comment, thanks. Most of the data mentioned is not available yet in pgcopydb though, and require extra setup and computations to fetch: - size of data transferred...

But it does not work that way. The amount of bytes sent through the network is not the same as the `pg_relation_size()` on the source, because of bloat (tuples on-disk...

Hi @branecc ; what happens is that the wal2json output plugin includes enough information (pkey) with each message for the client (pgcopydb here) to re-build the SQL command whereas test_decoding...

At the moment unfortunately neither test_decoding nor wal2json has the wanted feature where “it just works”. Consider building wal2json from sources to benefit from https://github.com/eulerto/wal2json/pull/255 and then use `pgcopydb clone...

The `pgcopydb list ...` commands are now in a different place as they were before the SQLite catalogs implementation. Because the catalogs are kept on-disk as a cache that we...

> Regarding the catalog initialization using `copydb_fetch_schema_and_prepare_specs`, it's worth noting that we also update the `s_table_parts` catalog. Conseqently now `--split-tables-larger-than` should be passed to `list tables`. Yes. > And `list...

> I think implementing a proper undo log will become tricky to unless we resort to a transactional file write. Maybe opening the latest file and appending a ROLLBACK(if last...

Hi @ulo23k ; thanks for opening this issue. Is it possible that table public.audition is filtered-out in your setup? That could explain why it's not in our internal catalogs, and...