Debian Bug report logs - #1080198
h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Display info messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Graham Inggs <ginggs@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Date: Sat, 31 Aug 2024 12:33:26 +0000

Source: h5py
Version: 3.11.0-5
Severity: serious
User: debian-ci@lists.debian.org
Usertags: regression

Hi Maintainer

Since the upload of mpi-defaults 1.17, the autopkgtests of h5py have
regressed on 32-bit architectures [1]:  I've copied what I hope is the
relevant part of the log below.

Regards
Graham


[1] https://ci.debian.net/packages/h/h5py/testing/armhf/


158s system msg for write_line failure : Bad file descriptor
158s [cli_2]: PMIU_write error; fd=15 buf=:cmd=abort exitcode=-1
message=PMI_Init failed
158s :
158s system msg for write_line failure : Bad file descriptor
158s [cli_2]: PMIU_write error; fd=15 buf=:cmd=get_appnum
158s :
158s system msg for write_line failure : Bad file descriptor
158s Fatal error in internal_Init_thread: Other MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xff82d538) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s [cli_2]: PMIU_write error; fd=15 buf=:cmd=abort
exitcode=404350991 message=Fatal error in internal_Init_thread: Other
MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xff82d538) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s :
158s system msg for write_line failure : Bad file descriptor
158s cmd=init pmi_version=1 pmi_subversion=1
158s cmd=abort exitcode=-1 message=PMI_Init failed
158s cmd=get_appnum
158s Fatal error in internal_Init_thread: Other MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xfffca0a8) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s cmd=abort exitcode=739895311 message=Fatal error in
internal_Init_thread: Other MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xfffca0a8) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s [cli_3]: PMIU_write error; fd=19 buf=:cmd=init pmi_version=1
pmi_subversion=1
158s :
158s system msg for write_line failure : Bad file descriptor
158s [cli_3]: PMIU_write error; fd=19 buf=:cmd=abort exitcode=-1
message=PMI_Init failed
158s :
158s system msg for write_line failure : Bad file descriptor
158s [cli_3]: PMIU_write error; fd=19 buf=:cmd=get_appnum
158s :
158s system msg for write_line failure : Bad file descriptor
158s Fatal error in internal_Init_thread: Other MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xffee5d38) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s [cli_3]: PMIU_write error; fd=19 buf=:cmd=abort
exitcode=404350991 message=Fatal error in internal_Init_thread: Other
MPI error, error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xffee5d38) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s :
158s system msg for write_line failure : Bad file descriptor
158s PASSED [ 75%]
158s tests/test_ros3.py::test_ros3_temp_token SKIPPED (AWS S3 access
toke...) [100%]Fatal error in internal_Init_thread: Other MPI error,
error stack:
158s internal_Init_thread(48394): MPI_Init_thread(argc=(nil),
argv=(nil), required=3, provided=0xff8e4e68) failed
158s MPII_Init_thread(195)......:
158s MPIR_pmi_init(149).........:
158s (unknown)(): Other MPI error
158s autopkgtest [12:30:09]: test python3-mpi-network: -----------------------]

Message #12 received at 1080198@bugs.debian.org (full text, mbox, reply):

From: Drew Parsons <dparsons@debian.org>

To: Debian Bug Tracking System <1080198@bugs.debian.org>

Subject: Re: h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Date: Mon, 16 Sep 2024 19:11:37 +0200

Source: h5py
Followup-For: Bug #1080198

As far as I can read the test log, the mpich (32-bit) tests are not
actually failing.  It looks like the error conditions is occuring
afterwards during shutdown.  Perhaps some test files need to be
explicitly closed before shutdown.

Message #17 received at 1080198-close@bugs.debian.org (full text, mbox, reply):

From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>

To: 1080198-close@bugs.debian.org

Subject: Bug#1080198: fixed in h5py 3.11.0-6

Date: Mon, 14 Oct 2024 14:49:57 +0000

[Message part 1 (text/plain, inline)]

Source: h5py
Source-Version: 3.11.0-6
Done: Drew Parsons <dparsons@debian.org>

We believe that the bug you reported is fixed in the latest version of
h5py, 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 1080198@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Drew Parsons <dparsons@debian.org> (supplier of updated h5py 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)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Mon, 14 Oct 2024 16:35:39 +0200
Source: h5py
Architecture: source
Version: 3.11.0-6
Distribution: unstable
Urgency: medium
Maintainer: Debian Science Maintainers <debian-science-maintainers@lists.alioth.debian.org>
Changed-By: Drew Parsons <dparsons@debian.org>
Closes: 1080198
Changes:
 h5py (3.11.0-6) unstable; urgency=medium
 .
   * debian/tests: run all MPI tests directly via pytest --with-mpi
     instead of h5py.runtest(), which crashes with MPICH (32-bit).
     Closes: #1080198
Checksums-Sha1:
 127fdb9710653032dee60302b0a5a80092133318 2831 h5py_3.11.0-6.dsc
 388ef8cb18e1c160bd1f52d2691e0b8e244fac57 21476 h5py_3.11.0-6.debian.tar.xz
 93bb04286258ccf8f8af559ce60314001f2044f5 14309 h5py_3.11.0-6_source.buildinfo
Checksums-Sha256:
 27b0d8db38450e084fd99ea367a45dc86b81d0a36dd6a4245154fecb06db413f 2831 h5py_3.11.0-6.dsc
 1c26281bed38e76283bab226f417778c8ed406e8b6552d272c1ab33da2c861ce 21476 h5py_3.11.0-6.debian.tar.xz
 c516d43df78583fc539bb5f497a57e153799e6e2d67d4b9df1d3f907b112feb0 14309 h5py_3.11.0-6_source.buildinfo
Files:
 4bccfdc79c7f2f5e1ec4725e59a0a277 2831 python optional h5py_3.11.0-6.dsc
 eed53c7f22b1ca31185f5e0cd8f40af1 21476 python optional h5py_3.11.0-6.debian.tar.xz
 be7fe498bdeb8a170ce6d71657b0d0a6 14309 python optional h5py_3.11.0-6_source.buildinfo
-----BEGIN PGP SIGNATURE-----

iQJIBAEBCgAyFiEEI8mpPlhYGekSbQo2Vz7x5L1aAfoFAmcNLF4UHGRwYXJzb25z
QGRlYmlhbi5vcmcACgkQVz7x5L1aAfqP7A/9FlTkA/u7Q5WHZjW1k6lCb+0Jshom
WYpW4eqByJYeLVIQdyLLQj5SvRHmBgGYyqX/r6p6jLZRLziCY9wX8te9kvhT6QLQ
KQ8mfHvzjXH8C+E3LOD5WIHGCiyyeRHZg8cF69sXYenfqEbXJOegG644xyrmVeqc
sqd8pqoN9UOn+ssU1pv/uFnl8788mu+mI/ObQXUX+m7+lYw+zuR/upz3xwkZ6iKw
hVRdfwDk0ujnuBwjE6dUISIUh7LY+gHiU/84yE7P0TCYsZxv64ozF05kCoY8j0Rb
x2DzVUOzdk/sQ+QNdpu9VnimpPEt43oHXFRlWWf3tvUSr+1w4vexfvgIZmvejYvb
wLArdJgQpgUPjw0HuaRfLBLbROwWmP46R8BhtmY0CNJnyC6VUto6/DBKi0x/GUYV
H2UNDa4SUlNVGBL4iS34QHZtjHpGUX1voiip72brJpbL36nrq9DH7u71KqgSnV/e
cTIjEUrpRfvIVogiaqMMJG629C/NQXIvGKzSKBZM33Q+wl9iAQTvj3tb19aCFzq1
pl5eBFm+y7AMZvnQiZfXvV7Icaj7PhQhZW8KVJUS6gQQogugKrsDD/1I4ZOamlky
cukOeIHgMhEropt836V37/GTZdqecTFtdanVERUSG73oIMwQ9eWWVNcwe+VgdG57
XkYOdk6LhqY548o=
=03fT
-----END PGP SIGNATURE-----

[Message part 2 (application/pgp-signature, inline)]

Message #22 received at 1080198@bugs.debian.org (full text, mbox, reply):

From: Chris Hofstaedtler <zeha@debian.org>

To: Graham Inggs <ginggs@debian.org>, 1080198@bugs.debian.org, python-meshio@packages.debian.org, dolfin@packages.debian.org

Subject: Re: Bug#1080198: h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Date: Sat, 9 Nov 2024 12:00:40 +0100

On Sat, Aug 31, 2024 at 12:33:26PM +0000, Graham Inggs wrote:
> Since the upload of mpi-defaults 1.17, the autopkgtests of h5py have
> regressed on 32-bit architectures [1]:  I've copied what I hope is the
> relevant part of the log below.

Pinging this bug as 3.11.0-8 doesn't migrate, due to autopkgtest
regressions in python-meshio:

autopkgtest for python-meshio/7.0.0-really-5.3.5-3: amd64: Pass, arm64: Regression or new test ♻ (reference ♻), armel: Pass, armhf: Pass, i386: Pass, ppc64el: Regression or new test ♻ (reference ♻), riscv64: Regression or new test ♻ (reference ♻), s390x: Regression or new test ♻ (reference ♻)

Looks like there is an instabillity problem with regards to
python3-dolfin:

 23s The following packages have unmet dependencies:
139
 23s  python3-dolfin : Depends: python3-dolfin-real but it is not installable or
140
 23s                            python3-dolfin64-real but it is not installable
141
 23s  satisfy:command-line : Depends: python3-dolfin-real but it is not installable
142
 23s E: Unable to correct problems, you have held broken packages.
143
 23s test-meshio          FAIL badpkg

Message #27 received at 1080198@bugs.debian.org (full text, mbox, reply):

From: Francesco Ballarin <francesco.ballarin@unicatt.it>

To: Debian Bug Tracking System <1080198@bugs.debian.org>

Subject: Re: h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Date: Sat, 09 Nov 2024 12:52:33 +0000

Source: h5py
Followup-For: Bug #1080198
X-Debbugs-Cc: francesco.ballarin@unicatt.it, zeha@debian.org

Hi Chris,
please give us a grace period of further 30 days before pinging us again because
the current migration issue is due to openmpi 5 being broken in unstable on all
the arches on which you see the failure.

There is basically nothing we can do until that gets sorted out.

Also note that the original report was about 32 bit arches (which use mpich), 
while the current migration problems are on 64 bit arches. In this sense,
it is still correct that this bug should be remained marked as done, since the
current failed migration is unrelated to the previous ones.

Cheers,
Francesco

Message #32 received at 1080198@bugs.debian.org (full text, mbox, reply):

From: Chris Hofstaedtler <zeha@debian.org>

To: Francesco Ballarin <francesco.ballarin@unicatt.it>, 1080198@bugs.debian.org

Subject: Re: Bug#1080198: h5py: autopkgtest regressions with mpich as default provider on 32 bit architectures

Date: Sat, 9 Nov 2024 14:29:35 +0100

Hi Francesco,

* Francesco Ballarin <francesco.ballarin@unicatt.it> [241109 13:57]:
> please give us a grace period of further 30 days before pinging us again because
> the current migration issue is due to openmpi 5 being broken in unstable on all
> the arches on which you see the failure.
> 
> There is basically nothing we can do until that gets sorted out.

I'll note that without the ping, h5py would have been autoremoved on
Nov 13.

The autoremoval process in Debian is more or less designed in a way
where individual package maintainers can ignore neither their
dependencies, or their reverse-dependencies.

> Also note that the original report was about 32 bit arches (which use mpich), 
> while the current migration problems are on 64 bit arches. In this sense,
> it is still correct that this bug should be remained marked as done, since the
> current failed migration is unrelated to the previous ones.

Not doubting it, but it still is unfixed in testing.

Chris

Send a report that this bug log contains spam.