On Thu, 17 Aug 2017 17:43:13 +1000
Dan Callaghan <dcallagh(a)redhat.com> wrote:
Sorry for my very slow reply Petr...
Excerpts from pschindl's message of 2017-08-08 17:24 +02:00:
> Hi Dan,
> I have few questions regarding a beaker.qa.fp.org
. What is the
> state of the project right now? I now work with DesktopQA team on
> upstreaming their tests. And because they run their tests on
> internal beaker so I'd like to mimic their workflow. So I'd like to
> try to run tests on our Fedora's beaker.
> So my questions. Does the beaker.qa.fp.org
works at all? Could you
> give me access to at least one machine? Can I help you somehow with
> make it work? What can I do to put F26 or Rawhide trees to Beaker?
> Thank you for your answers. Have a nice day. With regards
> Petr Schindler (Fedora QA)
I've cc'ed the Fedora qa-devel list. Tim Flink did a lot of work
getting the Fedora Beaker instance set up.
If you wanted to get access to use it, I'm sure we could do that.
would just need to add you to this group:
As far as I know, Beaker itself is fully up and running now (we had
some issues with Fedora authentication but they were all sorted out a
while back). However it looks like the Dell machines attached to
Beaker are not successfully booting into the installer right now. Tim
might know more about what is missing/broken with them. I'm not
entirely sure what's wrong because I don't think I have access to
their serial console.
The biggest problem I'm hitting right now is that I can't import repos.
There's something about the setup we have inside Fedora infra that
beaker's repo importing mechanism can't handle.
I keep meaning to get around to figuring out what's wrong and fix it
but as beaker is a pretty low priority right now, I never seem to get
around to it.
That reminds me, one thing we never did set up properly is a
Conserver instance so that Beaker can scrape the serial console logs.
That would certainly make it easier to see why the Dell machines
aren't netbooting successfully.
For now, you can find me and ask me to look at the machine itself. That
doesn't scale but for now, I'm not aware of other options.
I should try and see about writing a playbook to deploy Conserver...
I'd love to see that happen :)