Sorry, I was away for some other business for a while. I've fixed the build slave. The root cause is we were hitting device mapper 100G pool size limit, hence there's no space left for docker containers to run. I see the build is currently running and it's good so far. It's super late here, so please leave message and ping me if any additional issue happened.
2015-12-12 15:38 GMT+08:00 Konstantin Boudnik <[email protected]>: > Ok, I think the slaves are back now and I have enabled the nexus job, but > still the package builds are failing because of unknown id nexus. > > Not sure what's next and it is almost midnight here, so my brain doesn't > function well ... > > Cos > > On Fri, Dec 11, 2015 at 11:43PM, Olaf Flebbe wrote: > > Hi, > > > > I changed the build of bigtop-trunk-packages to have a pre job > > bigtop-trunk-packages-nexus-start to be run on all node via a > label-plugin and a bigtop-trunk-packages-nexus-stop to clean up afterwards. > > > > unfortunately the docker daemon on docker-slave-06 is unresponsive , so > I decided to disable bigtop-trunk-packages-nexus-start > > > > If someone with a root passwd please reboot the docker-slave-06 node and > enable bigtop-trunk-packages-nexus-start ... > > > > Olaf > > >
