On Wed, 2014-12-17 at 11:42 +0100, Carsten Ziegeler wrote: > Am 17.12.14 um 11:17 schrieb Bertrand Delacretaz: > > On Wed, Dec 17, 2014 at 10:25 AM, Carsten Ziegeler <[email protected]> > > wrote: > >> ... I also have the feeling that many jobs fail > >> just because of some Jenkins wiredness or whatever.... > > > > Sometimes maybe, but not that often AFAICS. > > > > Looking at > > https://builds.apache.org/view/S-Z/view/Sling/job/sling-trunk-1.6/ > > for example, right now I see one blue build that passed, three red > > which probably indicate (possibly mysterious) Jenkins failures and > > then 7 consecutive yellow ones which mean tests failing. > > > > Looking at the details it's the same 3 integration tests which fail in > > these 7 builds - just broken tests, failing every time, no mystery. > > > > I'm not against improving notifications, but apart from that there's > > no big mystery - fixing whatever broke those 3 tests would already go > > a long way ;-) > > > Sure, do we know what broke it? > > And still 3 mysterious fails out of 11 runs is a lot :)
Well, after looking at the three 'red' builds I see that all of them are red due to IT failures, mostly rooted in org.apache.sling.testing.samples.integrationtests.serverside . So no mistery here, we can't blame Jenkins for any of the sling-trunk-1.6 problems. Robert > > Carsten
