pg_repack use SeqScan for read table. Can we use Parallel Seq Scan like the Command Copy does?

[VladMak]

We have a table and we run pg_repack on it. We noticed that the access to the table occurs through SeqScan. And the COPY command uses Parallel SeqScan. The table goes through pg_repack in 45 minutes, and if you copy it using COPY, then in 15. Is there an opportunity to add Parallel SeqScan to pg_repack? And if not, why?

[Melkij]

I'm not keen on reinventing insert .. select in pg_repack. In the postgresql core, the implementation of parallel INSERT SELECT has a long history, once it was even committed, but then reverted.

#251 will helps is some cases (probably in most cases). I forgot about this patch, to be honest.

[bz007]

@Melkij would you update patch in MR #251? Right now it is marked "having conflicts".