#1076017 purity-off: autopkgtest regression on arm64: output keeps growing

#1076017#5
Date:
2024-07-09 15:25:13 UTC
From:
To:
Dear maintainer(s),

Your package has an autopkgtest, great. However, on arm64 it recently
started to fill the entire disk with its output file in $AUTOPKGTEST_TMP
(in testing and unstable, I haven't checked stable). On an otherwise
empty host, there's 63 GB free, a watchdog kicks in at 95% disk usage
and prevents most damage, but the current scheduled job for purity-off
on arm64 never finishes. Can you please investigate the situation and
fix it? The output only shows the first test passes: OK: 100.

The release team has announced [1] that failing autopkgtest on amd64 and
arm64 are considered RC in testing.

For now, I have added purity-off to our reject-list on arm64 to avoid
more jobs to get scheduled.

More information about this bug and the reason for filing it can be
found on
https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation

Paul

[1] https://lists.debian.org/debian-devel-announce/2019/07/msg00002.html

#1076017#10
Date:
2024-07-10 21:23:01 UTC
From:
To:
I am not sure of how I can investigate this, since it works fine on my
system.
The test is written in shell and Perl, so I do not expect it to be
architecture-dependent. The package is even "Architecture: all".

md:purity-off$ sadt -bv
cannot parse package relationship "@", returning it raw
----------------------------------------------------------------------
everything
----------------------------------------------------------------------
O: OK: 100
O: OK: 1500
O: OK: 400
O: OK: 500
O: OK: dabney
O: OK: new100
O: OK: pt100
----------------------------------------------------------------------
everything: PASS


OK (tests=1)
md:purity-off$

#1076017#15
Date:
2024-07-11 19:16:24 UTC
From:
To:
Hi Marco,

You have an arm64 system? If yes, good to know it's not systematic and
apparently only happening on the ci.d.n infrastructure. It would be
interesting to figure out what the differences in setup (hardware) are.

If you're interested, I can give you access to a testbed on ci.d.n
infrastructure to investigate. We'd need to align on the time, because I
need to be around while you're testing.

I realized that. But apparently only arm64 is broken. So it's probably a
(indirect) dependency that broke your test on arm64.

Paul

#1076017#20
Date:
2024-07-12 00:17:09 UTC
From:
To:
Yes. It's a Banana Pi M5 and I cannot see how this could be
hardware-dependent.
I should be able to check on the ci infrastructure myself in two weeks.

#1076017#25
Date:
2024-07-13 19:37:05 UTC
From:
To:
Hi Marco,

On the ci.d.n infrastructure our nodes run bookworm and use the lxc
backend. Do you do that too?

Paul

#1076017#30
Date:
2024-07-14 05:17:49 UTC
From:
To:
Hi Marco,

I just ran the test on stable, and there it passes. So there is
definitely something in the archive that make the difference in our setup.

Paul

#1076017#35
Date:
2024-07-14 12:47:26 UTC
From:
To:
No, I just run it on bare metal.
#1076017#40
Date:
2024-07-26 09:55:39 UTC
From:
To:
We believe that the bug you reported is fixed in the latest version of
purity-off, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 1076017@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Marco d'Itri <md@linux.it> (supplier of updated purity-off package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)
Format: 1.8
Date: Fri, 26 Jul 2024 11:31:56 +0200
Source: purity-off
Architecture: source
Version: 0-7
Distribution: unstable
Urgency: medium
Maintainer: Marco d'Itri <md@linux.it>
Changed-By: Marco d'Itri <md@linux.it>
Closes: 1076017
Changes:
 purity-off (0-7) unstable; urgency=medium
 .
   * Fix the autopkgtest regression on arm64. (Closes: #1076017)
     Probably this was caused by purity(6) calling TIOCGWINSZ without
     checking for errors, and the undefined results changing only for
     this architecture.
Checksums-Sha1:
 278ac167fd1f0c76510e62cad4542813b5eaf1a5 1173 purity-off_0-7.dsc
 d1ed586795f5a68f34395d86cea1bb6b92b8d5f9 1992 purity-off_0-7.debian.tar.xz
 1d432656d2ff336bd1200700d11c8528df9de3e5 5155 purity-off_0-7_amd64.buildinfo
Checksums-Sha256:
 b6355ab74b866b7eeba08672cdc07e65b391428075c2c3d66a832c6e31fac180 1173 purity-off_0-7.dsc
 3efdffa88a65f38160606cadc5c7c159c5fa9c73e4f6179827f81ce3da25f2d4 1992 purity-off_0-7.debian.tar.xz
 10fcd878b4401488255922500bacf63ca1f2197fc4fb11587e743f34da63755e 5155 purity-off_0-7_amd64.buildinfo
Files:
 8df53a33eef12d30fb7cf475f4085037 1173 games optional purity-off_0-7.dsc
 8a6c8c90dbf047ef4c7f464d55e2a54c 1992 games optional purity-off_0-7.debian.tar.xz
 abf6a660315802412c9f0df787f72398 5155 games optional purity-off_0-7_amd64.buildinfo
-----BEGIN PGP SIGNATURE-----

iHUEARYIAB0WIQQnKUXNg20437dCfobLPsM64d7XgQUCZqNtGwAKCRDLPsM64d7X
gcF4AQC7PeU80ZJ51XfXBsRfuqIrMAn2zaPh0vugn/Szsfg/+QD6Akg8yGEBNGb5
XqhVX1iZvgfMcJ2h7/djVP4iFXUEiw8=
=T7DD
-----END PGP SIGNATURE-----