[OE-core] maintaining sumo (was Re: [PATCH][thud] cve-check: backport rewrite from master)

[Ryan Harkin]

Hi Mikko,

On Mon, 11 Nov 2019 at 08:08, <Mikko.Rapeli at bmw.de> wrote:
>
> Hi,
>
> On Thu, Nov 07, 2019 at 10:41:42AM +0000, Ryan Harkin wrote:
> <snip>
> > Not sure if it helps, but I have a Jenkins job that tests sumo on a
trigger
> > (there is one for Warrior also):
> >
> > https://ci.linaro.org/job/warp7-openembedded-sumo/
> >
> > eg. it was triggered when Armin's patch was merged yesterday.
> >
> > This builds Sumo, based on Linaro's OE-RPB distro for NXP WaRP7
> > (imx7s-warp). It then runs the build in our LAVA lab (although the
boards
> > have gone down recently, they're normally up). Once the boards are up
> > again, I'll add ptest to the job, to give it a more thorough workout.
I'll
> > also add the sumo-next branch to the list of build configurations.
>
> Could it make sense to also test sumo-next using this job?

Once the boards are recovered in the lab, I'll create sumo-next and
warrior-next trigger jobs and add a few more tests.

> Also, I would not mind
> seeing test trigger and result emails on this list as long as access to
logs and details
> is open to everyone.

I believe this is public. At least, I can see everything from a private
browser window where I'm not logged in. It's trivial for me to add this
mailing list to the job, although I wonder if it'll bounce due to the email
address not being subscribed? I don't have access to the inbox where the
LAVA emails are sent from (lava at validation.linaro.org). It may be possible
for the lab team to subscribe to the list, I'd have to raise a support
request.

>
> I checked the jenkins build job config and looks quite straight forward.
Did not see
> the actual target testing logs though.

That's because the boards have gone "bad" in the lab. Builds 16 and 17 are
still queued, waiting for a "good" board to run on. I've raised a support
ticket to recover the boards.

Build 15 was the last Sumo job to run, and it's been purged from Jenkins
since it's over 30 days old. However, our SQUAD and LAVA instances keep
their data forever (at the moment). The jobs are still available, if you
know where to find them.

The binaries have not been purged yet (pinned-manifest.xml will tell you
what was built):
  https://snapshots.linaro.org/openembedded/warp7/sumo/imx7s-warp/15/rpb/

SQUAD data for Sumo on WaRP7:
  https://qa-reports.linaro.org/warp7/warp7-bsp/build/16a83e5/

SQUAD results for Sumo build 15 on WaRP7:

https://qa-reports.linaro.org/warp7/warp7-bsp/build/16a83e5/testrun/1937355/

LAVA job for Sumo build 15 on WaRP7:
  https://validation.linaro.org/scheduler/job/1937355

Testing wise, I'm only really running boot testing at the moment. The test
job makes sure the board boots to the commandline. Then it attempts to run
"memtester", which isn't installed, so fails. I did that on purpose
originally, because I wanted to make sure my memtester test failed properly
when not installed. I should update that now I'm sure it works. Next it
runs badblocks across the rootfs partition (/dev/mmcblk1p2) in read-only
mode.

Regards,
Ryan.

>
>
> Cheers,
>
> -Mikko
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20191111/5e917284/attachment.html>