On Thu, 2018-01-18 at 12:46 +0100, Nicolas Chauvet wrote: > 2018-01-11 16:53 GMT+01:00 Nicolas Chauvet <kwiz...@gmail.com>: > > 2018-01-04 23:31 GMT+01:00 Nicolas Chauvet <kwiz...@gmail.com>: > > > 2018-01-04 15:22 GMT+01:00 Kevin Kofler <kevin.kof...@chello.at>: > > > > Hi, > > > > > > > > Nicolas Chauvet wrote: > > > > > For this specific issue, I (one?) can probably find the > > > > > option to > > > > > raise the koji task timeout. > > > > > > Probably because the build was resubmitted before i've set the > > > change in kojid > > > So I've resubmitted the build and verified that the option was > > > correctly propagated from kojid.conf to the mock site- > > > default.conf > > > Thx Dan btw ;) > > > > > > While at it, I've also migrated the remaining f24 armv7 builders > > > to > > > current f26. But it will not affects the qt5-qtwebengine- > > > freeworld > > > build (as it already failed on f26 builders). > > > > So to sum-up the current situation. > > The last qt5-qtwebengine 5.10 build succeeded and took about 2 days > > and 2,5 hours to build. > > Compaired with last 5.9 build which took 1,5 day to build on arm, > > it's > > a rather big jump but it's comparable with the x86_64 ratio. > > > > The way forward is to build armv7 on aarch64 (using guest vm), But > > this requires to have an armv7 fedora cloud image (and if our cloud > > provider for aarch64 can run armv7 guests..). > > So there is obviously an issue with this build or target or else... > The builder seems to crash at some point in the compilation. > > There is no message written to disk (builders are using nbd), so not > much content for debug. > What I plan is to submit a scratch build for armhfp on my "in-house" > builder (running a fedora kernel), then recompose a koji build to fix > that as a workaround for now. > I will be travelling this week-end, so I will do this next week. > > @Leigh > Can you cancel the qt5-qtwebengine armhfp build and restart the > builder ? It will give you more ressources for other ffmpeg rebuild > tasks.
F27 build of qt5-qtwebengine is much slower , like Kevin state. Watching the logs [1] the build is in 14618 task of 20361 tasks which means 71% , so maybe needs one day, if timeout is lower than that , it is better cancel it . Anyway apart of previous state , don't see any problem with the build. [1] http://koji.rpmfusion.org/koji/watchlogs?taskID=189480 > > -- > - > > Nicolas (kwizart) > _______________________________________________ > rpmfusion-developers mailing list -- rpmfusion-developers@lists.rpmfu > sion.org > To unsubscribe send an email to rpmfusion-developers-le...@lists.rpmf > usion.org -- Sérgio M. B. _______________________________________________ rpmfusion-developers mailing list -- rpmfusion-developers@lists.rpmfusion.org To unsubscribe send an email to rpmfusion-developers-le...@lists.rpmfusion.org
