On Tue, Jan 12, 2021 at 10:50:16AM +0800, James(王旭) wrote: > You mean PG_13_202007201 was created by PG 13 ? No, there isn't any other > clusters, maybe it was created by my previous failed attempts of upgrading. Yes, it had to be created by PG 13 --- it starts with PG_13, followed by the system catalog version. Yes, it might be left over. You have to delete the old cluster and its tablespaces when you re-try. > So it seems it should be ok that I issue a command like this > > find . -name "PG_13*"|xargs rm -rf > > and then do upgrade again? I suggest you do more research than just run that --- at least I would. -- Bruce Momjian <bruce@xxxxxxxxxx> https://momjian.us EnterpriseDB https://enterprisedb.com The usefulness of a cup is in its emptiness, Bruce Lee
