On Sat, Aug 23, 2014 at 08:50:58PM -0500, Steve M. Robbins wrote: > On August 23, 2014 11:13:48 AM Steve M. Robbins wrote: > > On August 22, 2014 05:47:24 PM Kurt Roeckx wrote: > > > On Fri, Aug 22, 2014 at 09:09:47AM -0500, Steve M. Robbins wrote: > > > > I am assuming this is the "out of memory" killer at work. Maybe I'm > > > > wrong > > > > in this assumption? But if I'm right: how do we solve it? Is there an > > > > autobuilder with more memory that can handle this? Or maybe it is > > > > overloaded during the build; can we make it less loaded for building > > > > ITK? > > > > > > I checked earlier for that on the buildd but couldn't see it in > > > dmesg. But this time it's in it. > > > > > > I've given it back and excluded it on brahms. > > > > Thanks. The build completed this time, albeit unsuccessfully :-( > > The build failures in the log are actually test failures and at least some of > them appear to be caused by lack of space on the disk. The build server > started out with 37 GB free. However, I just did a build today on my amd64 > machine and 43 GB was consumed. > > So how best to proceed? Since I have a build here, maybe the most expedient > option is to just upload it, assuming I can figure out how to make it upload > only the amd64 debs.
Just remove the rest from the .changes file? > But for the future: how can we ensure ITK is built on an autobuilder with > enough disk space and enough memory? The buildds now have from 30 to 50 GB available, most are around 30. I can ask DSA if they have more available or not, or limit it to those buildds that do have enough disk space. Kurt -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: https://lists.debian.org/[email protected]
