Re: [v15 beta] pg_upgrade failed if earlier executed with -c switch

On Tue, Jun 07, 2022 at 08:30:47AM +0900, Michael Paquier wrote:
> If we don't split by the millisecond, we would come back to the
> problems of the original report.  On my laptop, the --check phase
> that passes takes more than 1s, but the one that fails takes 0.1s, so
> a follow-up run would complain with the path conflicts.  So at the end
> I would reduce the format to be YYYYMMDDTHHMMSS_ms (we could also use
> a logic that checks for conflicts and appends an extra number if
> needed, though the addition of the extra ms is a bit shorter).

So, attached is the patch I would like to apply for all that (commit
message included).  One issue I missed previously is that the TAP test
missed the log files on failure, so I had to tweak that with a find
routine.  I have fixed a few comments, and improved the docs to
describe the directory structure.

We are still need a refresh of the buildfarm client for the case where
pg_upgrade is tested without TAP, like that I guess:
--- a/PGBuild/Modules/TestUpgrade.pm
+++ b/PGBuild/Modules/TestUpgrade.pm
@@ -140,6 +140,7 @@ sub check
          $self->{pgsql}/src/bin/pg_upgrade/log/*
          $self->{pgsql}/src/bin/pg_upgrade/tmp_check/*/*.diffs
          $self->{pgsql}/src/bin/pg_upgrade/tmp_check/data/pg_upgrade_output.d/log/*
+         $self->{pgsql}/src/bin/pg_upgrade/tmp_check/data/pg_upgrade_output.d/*/log/*
          $self->{pgsql}/src/test/regress/*.diffs"
     );
     $log->add_log($_) foreach (@logfiles);
--
Michael