We only have 5 build servers in the "community QA". Red Hat QA has
many more but these are firewalled, not useful at all as contributors
would get links to reports they can't see.
So yes if a build takes ages we'd rather have it marked as a failure
than block a build server for days.. Not least because if it takes
much longer than expected there is probably something wrong going on
(a resource leak? wrong heap sizing?).
But if the build regularly takes - for example - 88 minutes, and
occasionally crosses the 90 minutes timeout.. then we can change the
timeout of course to give it a bit of flexibility. But it shouldn't
take 88 minutes!
Sanne
On 1 March 2016 at 12:53, Vlad Mihalcea <mihalcea.vlad(a)gmail.com> wrote:
Hi,
I noticed that the hibernate-orm-master-h2 keeps on failing from time to
time because of a timeout:
Build timed out (after 90 minutes). Marking the build as failed.
Build was aborted
[CHECKSTYLE] Collecting checkstyle analysis files...
FAILED
I wonder why do we even have a timeout during the build process?
Is there something that could potentially take so long that we cannot
afford running the build for too long?
Vlad
_______________________________________________
hibernate-dev mailing list
hibernate-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev