Buildfarm failure in rolenames.sql

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Postgres hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Buildfarm failure in rolenames.sql
Date: 2018-06-21 07:24:19
Message-ID: 20180621072419.GH1679@paquier.xyz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi all,

Some of your may have noticed, but the buildfarm has thrown up the
following failure today:
https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=gull&br=HEAD

The failure is the following one:
-- ALTER TABLE OWNER TO
\c -
! \connect: FATAL: could not read block 3 in file "base/16384/2662": read only 0 of 8192 bytes

I would have not bothered mentioning that as it could be perfectly be a
spurious failure, but I have actually seen the *same* failure for the
*same* block number in the *same* test on my laptop just two days ago.
So I suspect that we are up to something here. I saw the failure on my
laptop with gcc (7.3) and Debian sid, and gull is using Debian 9.4,
which is too much for a simple coincidence. The thing is hard to
reproduce only with the main regression test suite.

Thoughts?
--
Michael

Browse pgsql-hackers by date

  From Date Subject
Next Message Masahiko Sawada 2018-06-21 07:49:34 Re: [Proposal] Table-level Transparent Data Encryption (TDE) and Key Management Service (KMS)
Previous Message Amit Langote 2018-06-21 07:19:01 Re: bug with expression index on partition