15:00:36 <MattCrees[m]> #startmeeting blazar 15:00:36 <opendevmeet> Meeting started Thu Mar 20 15:00:36 2025 UTC and is due to finish in 60 minutes. The chair is MattCrees[m]. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:36 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 <opendevmeet> The meeting name has been set to 'blazar' 15:00:45 <MattCrees[m]> #topic Roll call 15:00:55 <priteau2> o/ 15:00:56 <markpowers> o/ 15:01:24 <MattCrees[m]> Hi both, thanks for joining 15:01:40 <MattCrees[m]> #topic Epoxy release 15:02:33 <MattCrees[m]> The release process is ongoing now, most of the generated patches have been fine. 15:02:46 <priteau> There is an issue with the blazar-nova patches 15:02:52 <MattCrees[m]> I do see however that we're still having issues with blazar-nova not finsing the stable/2025.1 branch 15:03:24 <priteau> Maybe something is not yet complete on the opendev side 15:03:37 <priteau> Or we are cloning nova the wrong way 15:05:19 <priteau> I have just done another recheck 15:05:26 <MattCrees[m]> Perhaps, we did change how we were cloning recently. 15:05:31 <priteau> It's weird that it worked for py312 but not py39 15:06:29 <MattCrees[m]> Indeed 15:06:43 <priteau> I was wondering if we are missing a required-projects config 15:08:57 <MattCrees[m]> Do you mean in the Zuul config? 15:09:02 <priteau> yes 15:09:52 <MattCrees[m]> It would be good to try if the CI fails again :) 15:10:55 <MattCrees[m]> Other than that, I think we've been on top of all the release tasks. 15:11:40 <MattCrees[m]> #topic PTG 15:13:41 <MattCrees[m]> I am going to reserve us a slot for the PTG, I was thinking our usual of two hours from 4pm UTC on Thursday 10th. Would this work for you markpowers? 15:14:47 <markpowers> That works for me 15:15:40 <MattCrees[m]> Great stuff. Please invite anyone else from your side who may be interested, I'll do the same on ours 15:15:46 <priteau> As mentioned I probably won't be available 15:17:19 <MattCrees[m]> I plan to keep the default etherpad name also, so we should use https://etherpad.opendev.org/p/apr2025-ptg-blazar. I'll populate this with the items to review from last time, feel free to also add any new topics. 15:19:53 <MattCrees[m]> That's all from me around the PTG, I'm looking forward to hosting our session :) 15:20:32 <MattCrees[m]> #topic Open discussion 15:20:46 <wncslln> o/ 15:20:49 <wncslln> can I get a discussion? 15:20:57 <wncslln> sorry for not be in the meeting on the first minutes 15:21:08 <MattCrees[m]> Sure, go for it 15:21:09 <wncslln> I'm a newcomer contributor 15:21:20 <MattCrees[m]> Welcome :) 15:22:15 <wncslln> pierre told me that Blazar currently does not support resource overcommit 15:22:35 <wncslln> and I need this feature in my environment (only for virtual machines) 15:22:49 <wncslln> I was thinking that I can contribute with that if you guys are ok with it 15:23:18 <wncslln> MattCrees[m]: thanks! 15:24:15 <MattCrees[m]> I'd be ok with that, it sounds like a useful feature to have. Are you familiar with contributing to other OpenStack projects? 15:25:15 <wncslln> I've already made some contributions to other projects 15:25:20 <wncslln> so yes :) 15:25:39 <wncslln> but I'm not all familiar with blazar source code 15:26:00 <MattCrees[m]> Cool, no need for me to explain the contribution process then :) 15:26:48 <markpowers> I may not be fully aware of what is needed for resource overcommiting, but I thought there is accounting for this in the flavor plugin via `allocation_ratio`. 15:27:12 <priteau> We track the allocation_ratio in the DB, but I don't think we do anything with it 15:27:39 <markpowers> It is used to calculate free slots: https://opendev.org/openstack/blazar/src/commit/896ef7f935573dbf297eb47e504487fad3180012/blazar/plugins/flavor/flavor_plugin.py#L185-L186 15:28:17 <priteau> yeah I see that now 15:28:19 <wncslln> maybe here also https://github.com/openstack/blazar/blob/master/blazar/plugins/instances/instance_plugin.py#L230 15:28:31 <priteau> Maybe it works. I don't think it has been tested though 15:29:54 <markpowers> I think it works for flavor reservations, but I am unsure of the instance reservation status 15:31:37 <wncslln> what it is the main difference about flavor and instance reservation? or I use the flavor reservation to create instances? 15:34:47 <MattCrees[m]> With instance reservations, you reserve specific resource amounts. vcpus, disk_gb, etc. It is the older/long-running way of a reservation. Flavor reservations are a newer feature which allow you to reserve instead based on an existing flavor on your system. 15:36:08 <wncslln> oh I see 15:36:37 <wncslln> I'm using blazar in zed release, so the only types that I have is physical host and virtual instances 15:37:21 <wncslln> maybe for instances may I implement this same logic but for VMs? 15:41:55 <MattCrees[m]> I'd need to look closer at how it's handled in the flavor plugin. But the flavor plugin was based on the instance one I believe, so the logic may be transferable. 15:43:05 <wncslln> cool. for the next weeks I may be able to take a look on this, so I can take some updates in next weekly. what do you think? 15:44:11 <MattCrees[m]> That would be great. If you get far enough to get a patch up, I'd be happy to review it. 15:44:36 <MattCrees[m]> We have fortnightly meetings, so the next one is April 3rd 15:45:00 <wncslln> I'll put it on my agenda. Thanks 15:45:03 <wncslln> that's all for me 15:45:21 <wncslln> if I a have some questions during the process, I'll ping here 15:45:46 <MattCrees[m]> Sounds good, I'll be available here :) 15:46:00 <wncslln> :) 15:46:19 <MattCrees[m]> priteau, markpowers anything else from you today? 15:47:10 <markpowers> nothing from me 15:49:26 <MattCrees[m]> I know Pierre is in another meeting now, so let's end here. Thanks for joining everyone! :) 15:49:30 <MattCrees[m]> #endmeeting