On Sun, Oct 12, 2014 at 11:14:45PM -0700, Craig Rodrigues wrote:
> I have created this Jenkins job, which you can see a graphical
> representation of:
It is noted below stuff is being done this for current as well and in
> (1) does a buildworld/buildkernel on amd64 when someone checks new
> code into the stable/10 branch
Is not this excessive? I would suggest every N commits or M hours if
there were not enough commits to reach given threshold. Of course no
work in there were no commits whatsoever, but then do the work after the
first which gets in.
> (2) Creates a bootable UFS image with makefs
any chance zfs will be used as well?
> (3) Boots the image under bhyve
> (4) Runs these commands inside the bhyve VM:
> cd /usr/tests
> kyua test
> kyua report-junit --output=test-output.xml
would be nice to run some kind of stress testing. buildworld with a high
-j is an example of a general purpose test. This could be done with
different frequency than regular tests.
> (5) Shuts down the bhyve VM
Do you have crashdumps configured in case stuff goes wrong?
> (6) imports test-output.xml into Jenkins.
> You can see a full test report here:
> We already do the same thing for CURRENT.
> Hopefully by running the tests regularly, we can help improve the quality
> of FreeBSD.
> firstname.lastname@example.org mailing list
> To unsubscribe, send any mail to
Mateusz Guzik <mjguzik gmail.com>
email@example.com mailing list
To unsubscribe, send any mail to