On Thu, 2013-01-10 at 09:25 -0500, Gary Gregory wrote: > On Thu, Jan 10, 2013 at 7:31 AM, Oleg Kalnichevski <[email protected]> wrote: > > > On Thu, 2013-01-10 at 12:04 +0000, sebb wrote: > > > On 10 January 2013 11:54, Continuum@vmbuild <[email protected]> > > wrote: > > > > Online report : > > http://vmbuild.apache.org/continuum/buildResult.action?buildId=25720&projectId=128 > > > > > > > > Build statistics: > > > > State: Failed > > > > Previous Build: No previous build. > > > > > > That was me. > > > > > > I've updated the build definitions to use deploy rather than install, > > > and that seemed to confuse the parent build. > > > Not sure if it will be OK next time; we'll have to see (Continuum > > > build definitions are complicated) > > > > > > The intention was to provide automated snapshot builds, which appear > > > in the snapshots repo, e.g. > > > > > > > > https://repository.apache.org/content/groups/snapshots/org/apache/httpcomponents/httpcore/4.3-alpha2-SNAPSHOT/ > > > > > > I can try adding the 4.2 branch later if it is thought useful. > > > > > > > Sebastian > > > > Jenkins can easily do snapshot publishing and we had it enabled for some > > while. The problem is that more often than not ASF snapshot repository > > is out space, which causes too many false build failures. I ended up > > disabling automatic shapshot publishing for Jenkins jobs for that > > reason. > > > > We probably need to think of some other way to publish snapshots. > > > > It's sad to me that we have an official way to publish snapshots and that > it seems broken. > > The space issue should be fixed, we should not have to find some > workaround. How do other projects deal with this? Can old snapshots be > removed? Or is there just one for each project? > > Gary >
I never felt irritated strongly enough to have raised this issue with infra. I suspect they are well aware of it, though. To me, the main problem is not the issue with storage but the fact that a failure to publish a snapshot due to an I/O error causes the build to fail rather than a warning. Oleg --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
