Re: Regression tests failures on Windows Server 2019 - on master at commit # d816f366b

Hi,

On 2022-03-04 09:57:35 -0800, Andres Freund wrote:
> On 2022-03-04 09:46:44 -0800, Andres Freund wrote:
> > On 2022-03-04 09:30:37 -0800, Andres Freund wrote:
> > > I wonder if we're missing some steps, at least on windows, to make pg_ctl
> > > start independent of the starting shell?
> >
> > Sure looks that way. On windows, if I do pg_ctl start, then hit ctrl-c, the
> > server shuts down.
>
> Short term the easiest fix might be to start postgres for those tests as a
> service. But it seems we should fix whatever the cause of that
> terminal-connectedness behaviour is.
>
> I'm out for ~2-3h. I started a test run with using a service just now:
> https://cirrus-ci.com/task/5519573792325632 but I very well might have typoed
> something...

That fixed the immediate problem, but dblink, postgres_fdw tests failed:

https://api.cirrus-ci.com/v1/artifact/task/5519573792325632/log/contrib/dblink/regression.diffs
FROM dblink(connection_parameters(),'SELECT * FROM foo') AS t(a int, b text, c text[])
WHERE t.a > 7;
- a | b | c
----+---+------------
- 8 | i | {a8,b8,c8}
- 9 | j | {a9,b9,c9}
-(2 rows)
-
+ERROR: could not establish connection
+DETAIL: connection to server at "localhost" (::1), port 5432 failed: FATAL:
role "SYSTEM" does not exist

https://api.cirrus-ci.com/v1/artifact/task/5519573792325632/log/contrib/postgres_fdw/regression.diffs
+ERROR: could not connect to server "loopback"

and it also seems to redirect logging to the event log without further
configuration...

The dblink and fdw failures can presumably be fixed by passing current_user as
the username. That seems like a good idea anyway?

Greetings,

Andres Freund