Horton

Re: Multiple hosts in connection string failed to failover in non-hot standby mode

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Justin Pryzby <pryzby@telsasoft.com>
Cc: Hubert Zhang <zhubert@vmware.com>, tsunakawa.takay@fujitsu.com, pgsql-hackers@postgresql.org, Andreas Seltenreich <seltenreich@gmx.de>, Bruce Momjian <bruce@momjian.us>
Date: 2021-05-06T17:22:27Z
Lists: pgsql-hackers 
Justin Pryzby <pryzby@telsasoft.com> writes:
> 52a10224 broke sqlsmith, of all things.

> It was using errmsg as a test of success, instead of checking if the connection
> result wasn't null:

>      conn = PQconnectdb(conninfo.c_str());
>      char *errmsg = PQerrorMessage(conn);
>      if (strlen(errmsg))
>          throw dut::broken(errmsg, "08001");

> That's clearly the wrong thing to do, but maybe this should be described in the
> release notes as a compatibility issue, in case other people had the same idea.
> Clearing errorMessage during success is an option..

Hm.  I'd debated whether to clear conn->errorMessage at the end of
a successful connection sequence, and decided not to on the grounds
that it might be interesting info (eg it could tell you why you
ended up connected to server Y and not server X).  But perhaps
it's too much of a compatibility break for this small benefit.

I'm curious though why it took this long for anyone to complain.
I'd supposed that people were running sqlsmith against HEAD on
a pretty regular basis.

			regards, tom lane

Commits

  1. Clear conn->errorMessage at successful completion of PQconnectdb().

  2. Avoid ECPG test failures in some GSS-capable environments.

  3. Try next host after a "cannot connect now" failure.

  4. Uniformly identify the target host in libpq connection failure reports.

  5. Allow pg_regress.c wrappers to postprocess test result files.

  6. In libpq, always append new error messages to conn->errorMessage.