On 22 November 2017 at 23:53, Michal Skrivanek
<michal.skrivanek(a)redhat.com> wrote:
On 22 Nov 2017, at 13:56, Barak Korren <bkorren(a)redhat.com> wrote:
> I suppose you are talking about RHEL here. As a developer, when you
> add dependencies, its your responsibility to check that they exist on
> your target platforms and avoid adding them if they don’t.
True, though the point of CI is to make sure it actually happens, right:)
We I can't argue with that...
> We cannot
> really use RHEL on oVirt's CI, the developer's free license is only
> good for one laptop.
It does have a solution though. You work for Red Hat, you have a free
license for machines in private network, you can place the jenkins slave
there, and make it run RHEL test publishing results back to oVirt
infra/gerrit
Slave/master connections are not good at traversing firewalls, and it
would also mean we'll need to include all kinds of
internal-network-only bits in the oVirt source repos.
But we actually do have a way to do something like this, now that I
think of it. I hinted at that in the internally published email where
we announced we build a new internal CI system based on the oVirt CI
code. We can have the internal system connect directly to oVirt
Gerrit, and then run RHELx jobs.
Anyone volunteers his project to be a guinea pig for that? I'd rather
not try this on the big and complex engine and vdsm first...
Also I worry that people might confuse this with the RHV CI...
--
Barak Korren
RHV DevOps team , RHCE, RHCi
Red Hat EMEA
redhat.com | TRIED. TESTED. TRUSTED. |
redhat.com/trusted