Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Rafael Laboissiere raf...@debian.org [2009-05-05 11:25]: That would be great. One question: is the bug deterministic or does it happen randomly? I'm not sure. I created an account for you yesterday and tried it again and now I cannot reproduce the problem anymore. I don't know if this is because the bug doesn't show up all the time (I thought it did) or because of the change in your latest upload. -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
On Tue, Apr 28, 2009 at 06:31:32PM +0200, Martin Michlmayr wrote: Just for the record, I see the same problem on an amd64 system (Intel EM64T chip). Is this a normal installation or as part of a build process? Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Thomas Weber thomas.weber.m...@gmail.com [2009-05-05 11:10]: On Tue, Apr 28, 2009 at 06:31:32PM +0200, Martin Michlmayr wrote: Just for the record, I see the same problem on an amd64 system (Intel EM64T chip). Is this a normal installation or as part of a build process? build process. It installs fine in my normal sid chroot. BTW, I should be able to give people root on this machine if you haven't been able to track down the issue yet. -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Martin Michlmayr t...@cyrius.com [2009-05-05 11:19]: * Thomas Weber thomas.weber.m...@gmail.com [2009-05-05 11:10]: On Tue, Apr 28, 2009 at 06:31:32PM +0200, Martin Michlmayr wrote: Just for the record, I see the same problem on an amd64 system (Intel EM64T chip). Is this a normal installation or as part of a build process? build process. It installs fine in my normal sid chroot. BTW, I should be able to give people root on this machine if you haven't been able to track down the issue yet. That would be great. One question: is the bug deterministic or does it happen randomly? -- Rafael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
I am revisiting this buildd problem conerning the postinst script of the octave3.0. I see in the buildd status page for mipsel [1], the following: State: Building (16) tbb (26d 00:44, 3/7, mayer) octave-communications (18d 00:24, 10/10, rem) octave-financial (18d 00:24, 11/11, rem) octave-fixed (17d 17:32, 11/11, mayer) octave-nan (14d 21:44, 11/11, mayer) octave-linear-algebra (14d 21:44, 11/11, mayer) octave-plot (14d 19:05, 11/11, rem) octave-sockets (11d 23:34, 11/11, rem) octave-specfun (11d 23:34, 11/11, rem) zoneminder (10d 07:11, 10/10, mayer) dballe (10d 03:16, 4/9, mayer) protobuf (9d 16:24, 8/9, mayer) nikwi (9d 06:14, 5/6, mayer) usplash-theme-debian (9d 01:54, tried 7 times, 4/6, mayer) synce-trayicon (1d 15:16, tried 3 times, 10/10) gurlchecker (1d 10:10, tried 3 times, 10/10) [1] https://buildd.debian.org/~luk/status/architecture.php?suite=a=mipsel If I interpret this correctly, there are at least eight octave-* packages with the build stalled right now, aren't there? Either we kill the processes or we take oportunity to debug the problem. I think developers have not access to either rem or mayer, right? Otherwise, if you discover anything interesting, please drop me a note. -- Rafael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
Rafael Laboissiere wrote: I am revisiting this buildd problem conerning the postinst script of the octave3.0. I see in the buildd status page for mipsel [1], the following: State: Building (16) tbb (26d 00:44, 3/7, mayer) octave-communications (18d 00:24, 10/10, rem) octave-financial (18d 00:24, 11/11, rem) octave-fixed (17d 17:32, 11/11, mayer) octave-nan (14d 21:44, 11/11, mayer) octave-linear-algebra (14d 21:44, 11/11, mayer) octave-plot (14d 19:05, 11/11, rem) octave-sockets (11d 23:34, 11/11, rem) octave-specfun (11d 23:34, 11/11, rem) zoneminder (10d 07:11, 10/10, mayer) dballe (10d 03:16, 4/9, mayer) protobuf (9d 16:24, 8/9, mayer) nikwi (9d 06:14, 5/6, mayer) usplash-theme-debian (9d 01:54, tried 7 times, 4/6, mayer) synce-trayicon (1d 15:16, tried 3 times, 10/10) gurlchecker (1d 10:10, tried 3 times, 10/10) [1] https://buildd.debian.org/~luk/status/architecture.php?suite=a=mipsel If I interpret this correctly, there are at least eight octave-* packages with the build stalled right now, aren't there? Either we kill the processes or we take oportunity to debug the problem. I think developers have not access to either rem or mayer, right? Otherwise, if you discover anything interesting, please drop me a note. There is only one of them really stalled on the buildd, the others just did not get retried. It being stalled is to debug the problem and it's kind of waiting on your input how to debug it AFAIK. Cheers Luk -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Luk Claes l...@debian.org [2009-05-02 14:19]: There is only one of them really stalled on the buildd, the others just did not get retried. It being stalled is to debug the problem and it's kind of waiting on your input how to debug it AFAIK. I am afraid, I have no idea on how to debug this, specially if I do not have access to the machine. Ideas are welcome. -- Rafael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Martin Michlmayr t...@cyrius.com [2009-04-28 18:31]: Just for the record, I see the same problem on an amd64 system (Intel EM64T chip). Thanks for your report. This bug is driving me crazy. It seems to be a random problem and I have no clue on how to debug this. Out of desperation, I just upload octave3.0_3.0.5-3 with some changes in the postinst. I am not very optimistic but let us see what happens. -- Rafael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Stephen Gran sg...@debian.org [2009-04-21 23:55]:
This one time, at band camp, Rafael Laboissiere said:
* Martin Zobel-Helas zo...@ftbfs.de [2009-04-21 21:17]:
here some more infos:
21029 ?S 0:00 \_ sh -c /usr/bin/sudo /usr/bin/apt-get
--purge -o
Dir::State::status=/home/buildd/build/chroot-unstable/var/lib/dpkg/status
-o DPkg::Options::=--root=/home/buildd/build/chroot-unstable -o DPkg::R
21030 ?S 0:03 \_ /usr/bin/apt-get --purge -o
Dir::State::status=/home/buildd/build/chroot-unstable/var/lib/dpkg/status
-o DPkg::Options::=--root=/home/buildd/build/chroot-unstable -o
DPkg::Run-Directory=/ho
22519 pts/0Ss+0:00 \_ /usr/bin/dpkg
--root=/home/buildd/build/chroot-unstable --force-confold --status-fd 14
--configure libmagic1 file gettext-base libkeyutils1 libkrb5support0
libk5crypto3 libkrb5-3 libgss
23967 pts/0S+ 0:00 \_ /bin/sh -e
/var/lib/dpkg/info/octave3.0.postinst configure
23976 pts/0R+ 1987:42 \_ octave-3.0.5 --silent
--no-history --no-init-file --eval pkg ('rebuild');
strace -p 23976
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
ioctl(0, TIOCNXCL, {0x1000 /* B??? */ -opost -isig -icanon echo ...}) = 0
[...] runs infitive
Thanks for the info. Is the hanging due to some file Octave is trying to
open/read/close? Could you please run lsof on the octave-3.0.5 process
and post the results? In the case above, it would be:
lsof -p 23976
The process has been killed, so we no longer have that information for
you.
Yes, I know, but could you try to launch the offending command:
octave-3.0.5 --silent --no-history --no-init-file --eval pkg ('rebuild');
in one of the machines where it fails and see what happens?
--
Rafael
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
* Thomas Weber thomas.weber.m...@gmail.com [2009-04-20 00:11]: On Sun, Apr 19, 2009 at 05:47:11PM +0200, Luk Claes wrote: Package: octave3.0 Version: 1:3.0.5-2 Severity: grave Hi The postinst of octave3.0 seems to hang on at least the mipsel buildds (for unstable) [1]. This issue blocks the buildds from building any other package till the octave3.0 process (mentioned in the build log) is killed manually. This has already caused delays for security updates and transitions... Installation of octave3.0 in a mipsel (qemu) environment works without a problem. Is there something specific about the mipsel buildds? Indeed, this seems to be a random problem on the mipsel buildd. Some octave-* packages built correctly on mipsel recently, like octave-econometrics on 2009-04-17 [1] and octave-data-smoothing on 2009-04-14 [2]. Both were built against octave3.0 3.0.5-2, so the postinst script ran fine, without hanging. [1] https://buildd.debian.org/fetch.cgi?pkg=octave-econometricsver=1%3A1.0.7-3arch=mipselstamp=1239954430file=log [2] https://buildd.debian.org/fetch.cgi?pkg=octave-data-smoothingver=1.1.1-2arch=mipselstamp=1239721086file=log It will be very hard for us to debug this random bug on the buildd. Any suggestions on how to proceed? -- Rafael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#524745: [Pkg-octave-devel] Bug#524745: Bug#524745: Postinst seems to run forever on mipsel
Rafael Laboissiere wrote: * Thomas Weber thomas.weber.m...@gmail.com [2009-04-20 00:11]: On Sun, Apr 19, 2009 at 05:47:11PM +0200, Luk Claes wrote: Package: octave3.0 Version: 1:3.0.5-2 Severity: grave Hi The postinst of octave3.0 seems to hang on at least the mipsel buildds (for unstable) [1]. This issue blocks the buildds from building any other package till the octave3.0 process (mentioned in the build log) is killed manually. This has already caused delays for security updates and transitions... Installation of octave3.0 in a mipsel (qemu) environment works without a problem. Is there something specific about the mipsel buildds? Indeed, this seems to be a random problem on the mipsel buildd. Some octave-* packages built correctly on mipsel recently, like octave-econometrics on 2009-04-17 [1] and octave-data-smoothing on 2009-04-14 [2]. Both were built against octave3.0 3.0.5-2, so the postinst script ran fine, without hanging. [1] https://buildd.debian.org/fetch.cgi?pkg=octave-econometricsver=1%3A1.0.7-3arch=mipselstamp=1239954430file=log [2] https://buildd.debian.org/fetch.cgi?pkg=octave-data-smoothingver=1.1.1-2arch=mipselstamp=1239721086file=log It will be very hard for us to debug this random bug on the buildd. Any suggestions on how to proceed? Maybe this is related to the kernel bug for the Broadcom hardware which is triggered by the new glibc (put Aurelien in Cc)? Cheers Luk -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
