Hi Riku, thanks for the information. Am Mittwoch, den 23.11.2011, 07:23 +0200 schrieb Riku Voipio: > On Tue, Nov 22, 2011 at 09:54:06PM +0100, Joachim Breitner wrote: > > looking at > > https://buildd.debian.org/status/logs.php?pkg=pandoc&arch=armel which > > very conveniently by now lists the builder repsonsible for each builds, > > there is a pattern that the builds fail on ancina, but succeed on the > > other arches. The error message are not helpful – some intermediate > > assembly file is broken. > > > Is there anything that makes ancina different from the other buildds > > that could help us debug the problem? > > Ancina is same machine as the rest, but has only 512MB of RAM while the > other builders have 1.5GB. The other difference is that other buildds > use lvm snapshots, while ancina still purges chroots. This could lead > to some lefteovers (say, conffiles in /etc) that disrupt builds. I don’t think that leftover can affect it, but the memory consumption can be the cause: If you look at https://buildd.debian.org/status/logs.php?pkg=haskell-xss-sanitize&arch=armel you see that the others are much faster, probably due to swapping. So it seems it is a minor bug in ghc that it gives some misleading messages about bad assembly when it is likely that just some process OOMed. This might also affect xss-sanitize on mipsel, although the picture is not as clear: https://buildd.debian.org/status/logs.php?pkg=haskell-xss-sanitize&arch=mipsel Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata
Attachment:
signature.asc
Description: This is a digitally signed message part