On Thu, Aug 03, 2017 at 01:35:18PM +1000, Dan Callaghan wrote:
Excerpts from Don Zickus's message of 2017-07-26 14:37 -04:00:
> Hi Dan,
>
> I am playing with some KernelCI stuff and one of the topics is fast
> booting/installs of a new kernel. The thought is waiting 20-25 minutes to
> install Anaconda is too long, when you are resource constrained and need to
> run 30-40 tests/day on the same hardware.
>
> An idea came up about creating a custom kernel and initrd with a testsuite
> builtin. Currently I don't think beaker supports that, only pre-loaded
> distro kernels/initrds.
>
> How hard is it to add that to the LabController? Or some other alternative?
Bill actually floated a similar idea many years ago, to make it
faster/non-destructive to run the Beaker hardware scanning script. Just
build it into an initramfs, boot it up and let it run without installing
anything to disk.
I guess you would be looking at something similar here?
Yes.
It would be a nice piece of functionality although you are right, it
doesn't really fit with Beaker's current assumptions about installing
a distro.
There are probably two big pieces you would need in place. One is the
ability to boot a particular installer tree (or in your case, no
installer, just kernel+initrd) by feeding Beaker the exact URL.
Basically bypassing the distro library. If you are producing 30-40
different kernel builds per day (plus corresponding initramfs image) you
don't really want to register each of them with Beaker's distro library.
That would be covered by this RFE:
https://bugzilla.redhat.com/show_bug.cgi?id=911515
Ok.
More importantly, we would also need a way for a Beaker recipe to skip
the installation step. We have discussed a similar idea in the past for
image-based "distros" like RHEV-H where there is no installation step,
it just boots straight into the environment. Atomic has a new-ish
"PXE-to-live" feature that works the same way too. It's kind of touched
on in this Beaker RFE:
https://bugzilla.redhat.com/show_bug.cgi?id=1254889
Yeah, this is something similar.
In your case, it might be enough to just make Beaker skip the kickstart
generation (and hide the relevant bits of the UI that task about the
recipe installation) assuming that the initramfs will contain a harness
that can report back to Beaker.
Right and it should.
The question is how to move this forward? Any phone call discussion?
Cheers,
Don