sön 2012-05-27 klockan 06:28 +0200 skrev Mattias Ellert: > lör 2012-05-26 klockan 12:36 +0200 skrev Julien Cristau: > > On Sat, May 26, 2012 at 04:26:11 +0200, Mattias Ellert wrote: > > > > > The failure om armel is the same that happened for a previous version > > > (2.4.28-1.1). That version built successfully when given back. > > This one completed now. Thank you. Sorry for mixing things up in my last mail. It was the other way around - it was the mips build that completed and the armel build that failed. > > > The mips build failed during make check. Looking at the log I see no > > > obvious error messages, so I hope this was a temporary issue. The text below is really about armel... > This one failed again. But on a different test. Looking closer at the > logs, they both contain a line lika: > > kill: No such process > > This makes me suspect that the builds did not actually fail, but rather > that they ware killed by either some watchdog script or some impatient > build manager before they were expected to finish. > > The first build was killed after 3H29M. A recent successful build on the > same host (alwyn) completed in 3H39M. > > The second build was killed after 13H17M. There have been no recent > builds on this host (argento), but scaling the build time on this host > of an older version (7H12M - ver 2.4.23-1) with the ratio of the build > times of similar versions on a different host (3H4M - ver 2.4.28-1.1, > 1H38M - ver 2.4.23-5, both in host ancina) gives an estimated build time > of the current version on this host of 13H31M. > > So I would like to ask for it to be given back again - and please don't > kill the build before it is expected to finish. Also the third attempt on armel was killed before it had time to complete. I have now tried to build the package on the armel portedbox abel.debian.org, and it built without problem there. Who or what is killing the builds on the armel buildds? Mattias
Attachment:
signature.asc
Description: This is a digitally signed message part