16:01:25 <priteau> #startmeeting blazar
16:01:26 <openstack> Meeting started Thu Mar 26 16:01:25 2020 UTC and is due to finish in 60 minutes.  The chair is priteau. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:29 <openstack> The meeting name has been set to 'blazar'
16:01:33 <priteau> #topic Roll call
16:01:49 <priteau> Hi jakecoll
16:02:38 <jakecoll> Hi priteau
16:02:38 <priteau> #topic Ussuri priorities
16:02:54 <priteau> One of the priorities for Ussuri is to implement network reservation
16:03:20 <priteau> Thank you jakecoll for resubmitting an updated patch
16:03:22 <priteau> #link https://review.opendev.org/#/c/668749/
16:03:44 <priteau> I've started to review but need more time on it to complete
16:04:10 <jakecoll> No problem. I can update when you are done.
16:04:12 <priteau> I have two pieces of advice when updating patches
16:04:52 <priteau> 1) use grep (or ripgrep/ack/ag) to find code that might be relevant to comments
16:05:22 <priteau> for example I pointed out the vfc code a few days ago but noticed today that there are still a couple of vfc references in the latest one
16:05:55 <priteau> and 2) check back to the spec to make sure it matches the API we agreed on
16:06:42 <jakecoll> Yea, I see the vfc code in the models now. I'll fix that.
16:07:01 <jakecoll> Are the differences in the spec API based on the allocations patch I added?
16:07:22 <priteau> No, I think we just made some minor changes to the format
16:07:30 <priteau> Like, using segmentation_id instead of segment_id
16:09:01 <jakecoll> oh, I see
16:09:25 <priteau> It would be really good if we could have a tempest scenario for this as well, but do you feel like you could write one?
16:10:01 <jakecoll> I've never written one
16:10:18 <jakecoll> Where are tempest scenarios for blazar?
16:10:55 <priteau> https://opendev.org/openstack/blazar-tempest-plugin/src/branch/master/blazar_tempest_plugin/tests/scenario
16:11:23 <priteau> Maybe if I can start a template for a new one with the right config, you could extend with some tests
16:12:25 <priteau> It will need a custom config to enable the network plugin, plus make sure the neutron config is compatible
16:12:31 <jakecoll> I could potentially. I get a lot of pushback from Kate these days on contributing upstream because of the time commitment.
16:13:09 <jakecoll> What is tempest?
16:13:44 <priteau> Testing framework for OpenStack
16:13:45 <priteau> https://docs.openstack.org/tempest/latest/
16:14:25 <jakecoll> It's an integration tool. Maybe this is something we can add to our CI/CD pipeline on Chameleon in the future.
16:14:36 <priteau> Then ignore it for now, I'll ping you if I get something started where you could contribute a few more tests
16:15:03 <jakecoll> ok, thanks
16:15:05 <priteau> Yes, you can run it against an OpenStack cloud. I would advise using Rally to manage it
16:15:20 <priteau> Rally: https://docs.openstack.org/rally/latest/
16:15:53 <priteau> Though it might not play well with Chameleon because it wouldn't know that it needs reservations to launch instances
16:16:00 <priteau> Anyway, that's getting off topic
16:16:30 <priteau> Anything else that you would like to highlight? I know there are specs from Jason that I need to finish reviewing
16:16:56 <jakecoll> I pushed a bug fix recently
16:17:11 <priteau> A very good one indeed
16:17:31 <priteau> Did you see tetsuro's comment?
16:17:38 <priteau> > Is it difficult to reproduce the bug in the unit tests by mocking the exception?
16:18:12 <jakecoll> Yes
16:19:22 <priteau> Do you think that's feasible?
16:19:24 <jakecoll> I imagine you could throw an exception with the unit test and then see if it fails. However, I haven't had time to play around with it. Not sure how the tests handle looping calls.
16:20:58 <jakecoll> probs not actually
16:21:00 <jakecoll> def test_start(self):        # NOTE(starodubcevna): it's useless to test start() now, but may be in        # future it become useful        pass
16:21:44 <jakecoll> We don't even test the start function where the fix is implemented.
16:22:03 <priteau> That is some old code!
16:23:42 <priteau> OK, more thought needed
16:24:08 <priteau> Have you deployed the fix to prod?
16:24:45 <jakecoll> Yep, we haven't run into the issue since
16:26:01 <priteau> An alternative would be to make sure the exception is handled correctly
16:26:14 <priteau> Did you figure out where the exception came from exactly?
16:29:01 <jakecoll> It was a DBConnectionError
16:31:04 <priteau> So something that could be thrown by any db code really
16:31:29 <priteau> OK, thanks for the patch, I will check again with Tetsuro if he can allow it without tests
16:31:34 <priteau> Anything else?
16:32:10 <jakecoll> Nope. Did you get a chance to review Jason's specs?
16:32:16 <jakecoll> He's back in office tomorrow
16:32:31 <jakecoll> or... online
16:33:06 <priteau> I reviewed one of them, left some comments
16:33:12 <priteau> Still need to look at the other one
16:33:26 <priteau> Good to know that he's back soon
16:34:52 <jakecoll> I believe that is all on my end
16:35:19 <priteau> Same for me. Thanks for your time!
16:35:39 <jakecoll> :+1:
16:35:46 <jakecoll> thanks for the comments
16:35:50 <priteau> Bye
16:35:51 <priteau> #endmeeting