Am Sonntag, 31. Januar 2010 23:53:59 schrieb Lubos Lunak:
Hello,
right now the OBS is mostly idle, trying to repeatedly build home:prusnak:scout/scout. The full log I managed to capture is at http://pastebin.ca/1773632, the interesting parts seem to be: ===== ... preinstalling bash... ... preinstalling libreadline6... ... /bin/bash: error while loading shared libraries: libreadline.so.5: cannot open shared object file: No such file or directory Kernel panic - not syncing: Attempted to kill init! =====
I'm not sure why exactly this got broken, but home:prusnak:scout includes also a bash package, so I assume the preinstall somehow used packages also from there, and it got somehow messed up.
The real problem however is that the OBS was repeatedly trying to dispatch this build, it failed almost immediately and the OBS retried again, and again, and again. It managed to dispatch about 1 other build per minute.
Yes, that is a new behaviour. OBS is handling this as "Bad build host", because even the bash inside of the VM is not able to start and to write any status. (it was a change from me 1-2 weeks ago).
I think this is a consequence of the recent fix that retries builds that failed because of the build host itself instead of the package. Would it be possible to adjust it so that such failed builds are put at the end of the queue instead of being handled again right away?
No, that makes not really sense, because it either is a bad host, than it should not affect the priority. Or it is one and the server should give up. But I will add some support that the dispatcher is giving up after 3 tries with "bad host". bye adrian
-- Adrian Schroeter SUSE Linux Products GmbH email: adrian@suse.de -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-buildservice+help@opensuse.org