On 12/02/15 09:34 -0800, Chris St. Pierre wrote:
Yeah, that commit definitely disables the file-backed queue -- it certainly
*looks* like we want to be rid of it, but all of the code is left in place and
even updated to support the new format. So my confusion remains. Hopefully Zhi
Yan can clarify.

Link added. Thanks.


Hi Chris,

I touched bases with Zhi Yan and my understanding is right. Since
Juno, we switched to using a queue based on database instead of file
and the file queue is considered redundant and on its way to be
deprecated.

I'll also reply on the review,
Thanks for bringing this up,
Flavio


On Thu, Feb 12, 2015 at 12:59 AM, Flavio Percoco <[email protected]> wrote:

   On 11/02/15 13:42 -0800, Chris St. Pierre wrote:

       I recently proposed a change to glance to turn the file-backed scrubber
       queue
       files into JSON: https://review.openstack.org/#/c/145223/

       As I looked into it more, though, it turns out that the file-backed
       queue is no
       longer usable; it was killed by the implementation of this
       blueprint: https://
       blueprints.launchpad.net/glance/+spec/image-location-status

       But what's not clear is if the implementation of that blueprint should
       have
       killed the file-backed scrubber queue, or if that was even intended.
       Two things
       contribute to the lack of clarity:

       1. The file-backed scrubber code was left in, even though it is
       unreachable.

       2. The ordering of the commits is strange. Namely, commit 66d24bb
       (https://
       review.openstack.org/#/c/67115/) killed the file-backed queue, and
       then,
       *after* that change, 70e0a24 (https://review.openstack.org/#/c/67122/)
       updates
       the queue file format. So it's not clear why the queue file format
       would be
       updated if it was intended that the file-backed queue was no longer
       usable.

       Can someone clarify what was intended here? If killing the file-backed
       scrubber
       queue was deliberate, then let's finish the job and excise that code.
       If not,
       then let's make sure that code is reachable again, and I'll resurrect
       my
       blueprint to make the queue files suck less.

       Either way I'm happy to make the changes, I'm just not sure what the
       goal of
       these changes was, and how to properly proceed.

       Thanks for any clarification anyone can offer.


   I believe the commit you're looking for is this one:
   f338a5c870a36e493f8c818fa783942d1e0565a4

   There the scrubber queue was switched on purpose, which leads to the
   conclusion that we're moving away from it. I've not participated in
   discussions around the change related to the scrubber queue so I'll
   let Zhi Yan weight in here.

   Thanks for bringing this up,
   Flavio

   P.S: Would you mind putting a link to this discussion on the spec
   review?





       --
       Chris St. Pierre


       
__________________________________________________________________________
       OpenStack Development Mailing List (not for usage questions)
       Unsubscribe: [email protected]?
       subject:unsubscribe
       http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



   --
   @flaper87
   Flavio Percoco
__________________________________________________________________________
   OpenStack Development Mailing List (not for usage questions)
   Unsubscribe: [email protected]?subject:unsubscribe
   http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev





--
Chris St. Pierre

__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


--
@flaper87
Flavio Percoco

Attachment: pgp66NNP94Yfg.pgp
Description: PGP signature

__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to