[OpenStack-DefCore] Some Governance Patches To Be Aware Of

Mark Voelker mvoelker at vmware.com
Fri May 6 20:42:20 UTC 2016


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512


On May 6, 2016, at 4:28 PM, John Dickinson <me at not.mn> wrote:



On 6 May 2016, at 10:03, Mark Voelker wrote:

Hi Folks,

For those of you who didn’t catch it elsewhere, here are a couple of DefCore-related things that the TC is currently considering that you may want to be aware of:

"add resolution explaining which tests we think defcore should use"
https://review.openstack.org/#/c/312718/

Should this resolution pass, what incentive remains to write a Tempest plugin?

Currently the object storage project is working on a Tempest plugin so that our in-tree functional tests may be used as part of the DefCore validation suite. That work seems like needless code churn now. If in-tree tests are specifically discouraged for being used in DefCore compliance checks, this removes the incentive to use Tempest's plugin capabilities. Does this resolution thus imply that the Tempest plugin capability will be going away at some point?

- --John




1.)  I’d encourage you to raise these sorts of questions in the review if you haven’t already as I think the TC would be interested in hearing these sorts of concerns. =) 

2.)  I’ll let the RefStack folks speak for themselves, but I’d speculate that it’s unlikely RefStack would eliminate the ability to run tests via the plugin interface completely.  This is partly because there’s been interest from several parties in the ability to author third-party “Guidelines” that would be unofficial and not kept in the DefCore repository, but would be useful to individuals.  For example: if I know I’m going to use my cloud to run particular sorts of apps that need particular projects or particular API’s that are a superset of what DefCore currently covers, I could write my own JSON file, add it to my personal account in RefStack, and use it to see which products suit my needs.  As there’s no TC mandate to move all tests to Tempest, RefStack would still need to have a second interface to handle that sort of need.

3.)  As far as whether or not the Tempest plugin interface itself would be going away: here again I suspect not, but the QA folks are probably better suited to answer that than I am. 

At Your Service,

Mark T. Voelker


-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - https://gpgtools.org

iQIcBAEBCgAGBQJXLQGrAAoJELUJLUWGN7Cb4MEP/R9vBOW1LzNyCeBEGgFEUosh
a+TXab6C0ldb1rtmmWOucNCseC1G+sWBVyVhXJj8F8u0cD4Za7b340qoQZooZqys
Pg8yv1Z95IKa/Yc3zd+e5sV8IdwiuF150ItQnAse520di1kF3ZadMJwpcD7GsHXB
R526jfgNKUi9g3f1a8e0earX+DQcQLcsbIyLLqzSwDioj4Axhy+LphD4FtbRYtw7
6y9aSu2a6yfBkHXLTmr3lw9dnO1Khxx6On4z26nFNmrskY8lHQpLTKrF7Ih5U91y
iJEcd5mQPKw8uhX46vXqjgwB/4LU5nTe49TOG9nDbtVGEGF27wZs2G1d3zpvBYj6
EkZcQeTFh/5XLvXAssC4fvl6zpRP0wR2G7F2Cnv8oiotSs305ttr3BNxn5csIc2T
KeQU8zWZ7ziGPUM52bYf5ya1W2xz4Y3bDkAvGibIQASLVPPSk7b6yx7lNNl+BFdZ
ftoHIVdWxWIrD++ovdVUHCnkug6gv4sNX3FujIHBdCA6dcdbFeo3Sja4HqMImyG3
yDhGpFGciyzb0rFTATyy3NVDwWTIW11dFqhkq++p6cH8jaaNbha59sIiQmirau/h
jZ6LdF0GnIu3iXWksfkETkzlsgwm7q/r/pPdlsjqfPa+/gycPUC/JtvId4faHR+t
WGc+gyzNYblUSsstVXAe
=sr9x
-----END PGP SIGNATURE-----


More information about the Defcore-committee mailing list