Date: Mon, 29 Jun 2026 08:45:37 -0700
From: Mark Wong <markwkm@gmail.com>
To: Olaf Bohlen <olbohlen@eenfach.de>
Cc: buildfarm-members@lists.postgresql.org
Subject: Re: hake failing on pgsql-Git kinda "random"
Message-ID: <akKTIQs-5grW0bts@ltdrgnflg2>
References: <2F0B1C7F-A27D-42E2-A303-CC99D1E43CD0@eenfach.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <2F0B1C7F-A27D-42E2-A303-CC99D1E43CD0@eenfach.de>
Archived-At: 
 <https://www.postgresql.org/message-id/akKTIQs-5grW0bts%40ltdrgnflg2>
Precedence: bulk

Hi!

On Mon, Jun 29, 2026 at 03:06:32PM +0200, Olaf Bohlen wrote:
> Hi,
> 
> I'm running the hake animal and after a server loss in April, everything should go smooth again. But now I have the problem that at random times hake failes in pgsql-Git. It works for days, and then out of the blue it stops.
> 
> I have the following in my crontab:
> 
> $ crontab -l | egrep -v "^#"                     
> 15 3,10,16,21 * * * ( cd /export/home/elmer/c20x && PATH=/usr/gnu/bin:$PATH ./run_branches.pl --run-all --verbose >/export/home/elmer/c20x/cron.out 2>&1 )
> 
> and the cron.out says:
> 
> $ cat cron.out                                             
> Mon Jun 29 10:15:02 2026: buildfarm run for hake:REL_14_STABLE starting
> hake:REL_14_STABLE [10:15:03] checking out source ...
> Missing checked out branch bf_REL_14_STABLE:
> fatal: not a git repository (or any of the parent directories): .git
> hake:REL_14_STABLE [10:15:10] failed at stage pgsql-Git
> Mon Jun 29 10:15:10 2026: buildfarm run for hake:REL_15_STABLE starting
> hake:REL_15_STABLE [10:15:10] checking out source ...
> Missing checked out branch bf_REL_15_STABLE:
> fatal: not a git repository (or any of the parent directories): .git
> hake:REL_15_STABLE [10:15:13] failed at stage pgsql-Git
> Mon Jun 29 10:15:13 2026: buildfarm run for hake:REL_16_STABLE starting
> hake:REL_16_STABLE [10:15:13] checking out source ...
> Missing checked out branch bf_REL_16_STABLE:
> fatal: not a git repository (or any of the parent directories): .git
> hake:REL_16_STABLE [10:15:16] failed at stage pgsql-Git
> Mon Jun 29 10:15:16 2026: buildfarm run for hake:REL_17_STABLE starting
> hake:REL_17_STABLE [10:15:16] checking out source ...
> Missing checked out branch bf_REL_17_STABLE:
> fatal: not a git repository (or any of the parent directories): .git
> hake:REL_17_STABLE [10:15:19] failed at stage pgsql-Git
> Mon Jun 29 10:15:19 2026: buildfarm run for hake:REL_18_STABLE starting
> hake:REL_18_STABLE [10:15:19] checking out source ...
> Missing checked out branch bf_REL_18_STABLE:
> fatal: not a git repository (or any of the parent directories): .git
> hake:REL_18_STABLE [10:15:21] failed at stage pgsql-Git
> Mon Jun 29 10:15:22 2026: buildfarm run for hake:HEAD starting
> hake:HEAD          [10:15:22] checking out source ...
> Missing checked out branch bf_HEAD:
> fatal: not a git repository (or any of the parent directories): .git
> hake:HEAD          [10:15:24] failed at stage pgsql-Git
> 
> Once I delete the buildroot directory it starts working again for some days.
> 
> Does anyone here has seen the same or an idea?

I get that a lot on a number of the VM's I use with POWER and ARM.
Andrew Dunstan has suggested I set "use_vpath => true".  I think that
has helped...

Regards,
Mark