On 13 December 2016 at 22:56, Nir Soffer <nsoffer(a)redhat.com> wrote:
Barak, do you think we can change the script so setup and cleanup failures are
not treated as build failures but build errors?
Doing that means assuming that the cleanup script, that runs
flawlessly 100** times a day for all patches in all projects suddenly
failed for reasons that have nothing to do with the patch that was
just tested.
I think we err on the right side of caution now...
In travis such failure seem to start another build automatically,
making developers life much nicer.
And in VMware there is no SPM, you can mix local and remote storages
on the same nodes, and upload images to the storage domain from the
GUI, making the admin's life much nicer. What is your point again?
On a more serious note, we may do auto-rerunning at some point, but we
need to re-engineer most of the standard-CI system to make that
happen, so it'll take a while.
--
Barak Korren
bkorren(a)redhat.com
RHCE, RHCi, RHV-DevOps Team
https://ifireball.wordpress.com/