| gouthamr | gmann: the srbac goal-update seems like a good one to merge; i thought about asking for more votes.. but i changed my mind re-reading the house rules.. | 17:55 |
|---|---|---|
| gouthamr | i workflowed it now | 17:55 |
| gouthamr | JayF++ gmann++ "tox -e check-review-status" is a real powerful specter :D | 17:56 |
| JayF | yeah, setting topic on governance changes is important for that :D | 17:57 |
| gouthamr | true | 17:57 |
| gmann | gouthamr: ++ | 17:59 |
| gmann | yeah this script is really helpful | 17:59 |
| gouthamr | #startmeeting tc | 18:00 |
| opendevmeet | Meeting started Tue May 7 18:00:39 2024 UTC and is due to finish in 60 minutes. The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot. | 18:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 18:00 |
| opendevmeet | The meeting name has been set to 'tc' | 18:00 |
| noonedeadpunk | o\ | 18:00 |
| gouthamr | #topic Rollcall | 18:01 |
| gmann | o/ | 18:01 |
| dansmith | o/ | 18:01 |
| noonedeadpunk | I forgot to mark myself as out today, but on openinfra Sweden event, so really more out then it | 18:01 |
| gouthamr | ack noonedeadpunk | 18:02 |
| gouthamr | this meeting is happening simultaneously on Zoom | 18:02 |
| gouthamr | #link https://us06web.zoom.us/j/87108541765?pwd=emlXVXg4QUxrUTlLNDZ2TTllWUM3Zz09 | 18:02 |
| slaweq | o/ | 18:02 |
| JayF | o/ | 18:03 |
| gouthamr | courtesy ping: spotz[m] frickler | 18:04 |
| gouthamr | gtema | 18:04 |
| gtema | O/ | 18:05 |
| gouthamr | #topic DPL model improvement proposals and TC Liaisons | 18:05 |
| gouthamr | #link https://review.opendev.org/c/openstack/governance/+/916822 (Move DPL model into doc) | 18:05 |
| gmann | gtema: today is video call #link https://us06web.zoom.us/j/87108541765?pwd=emlXVXg4QUxrUTlLNDZ2TTllWUM3Zz09 | 18:05 |
| gouthamr | #link https://review.opendev.org/c/openstack/governance/+/916833 (Add DPL model & liaison reset policy) | 18:05 |
| gouthamr | #link https://review.opendev.org/c/openstack/governance/+/917516 (TC liaisons for requirements, release management and ; freezer is the fourth) | 18:05 |
| gtema | Right, forgot about video, will join in a min | 18:06 |
| gouthamr | gmann is presenting the idea behind https://review.opendev.org/c/openstack/governance/+/916833 and https://review.opendev.org/c/openstack/governance/+/917516 | 18:06 |
| gouthamr | we're not in a hurry to merge this | 18:08 |
| gouthamr | but we're interested to know if there are any major disagreements | 18:08 |
| gouthamr | please add your comments to these gerrit changes | 18:08 |
| gouthamr | tkajinam brought up the concern that the initial DPL model proposal suggests that the model is suggested for low activity projects | 18:09 |
| gouthamr | JayF suggested we change the language to clarify this | 18:10 |
| gouthamr | we're suggesting spreading the load for the TC liaisons for the DPL projects | 18:11 |
| gouthamr | tc-members please feel free to claim any projects you'd wish | 18:11 |
| gouthamr | gtema is seeking if we have a written procedure to know when a project is active/not | 18:12 |
| JayF | #link https://governance.openstack.org/tc/reference/emerging-technology-and-inactive-projects.html | 18:12 |
| JayF | I think this is it? gtema | 18:12 |
| gouthamr | JayF++ | 18:12 |
| gouthamr | project meetings aren't a strict requirement to judge an active project... | 18:13 |
| gouthamr | but a holistic/subjective approach that works instead to judge a project's activity/inactivity seems to warrant a TC liaison's judgement | 18:14 |
| gouthamr | (or the TC's judgement in case of PTL model projects) | 18:14 |
| slaweq | helper script which may give some data about project's health: https://github.com/openstack/governance/blob/master/tools/project_stats_check.py | 18:15 |
| gouthamr | its really hard to automate this sort of thing | 18:15 |
| JayF | given how projects in openstack are mature enough to be mostly "done", it's hard to use statistics alone to measure project activity | 18:16 |
| gouthamr | gtema says project users/operators need to know which projects are active | 18:17 |
| gouthamr | JayF brought this up at the TC PTG | 18:18 |
| gouthamr | #link https://etherpad.opendev.org/p/apr2024-ptg-os-tc (TC PTG etherpad) | 18:18 |
| gouthamr | there was no conclusion to make a proposal | 18:19 |
| gouthamr | can we break up the problem to make a proposal - for example, if the security concerns are at the top of our mind, can we have a proposal just isolating the security requirements for projects | 18:21 |
| gouthamr | we'd need to change the project requirements to accommodate this | 18:21 |
| gouthamr | the security-sig and VMT has less volunteers though; so we'd need to do this simultaneously by growing the security-sig/VMT | 18:22 |
| gouthamr | will a yellow/green/red marking on our existing project list help? in addition to having https://governance.openstack.org/tc/reference/emerging-technology-and-inactive-projects.html and | 18:22 |
| JayF | (I said) We have a limited amount of resources, any approach would have to acknowledge that and use prioritization rather than just adding additional requirements | 18:23 |
| gouthamr | slaweq suggests adding a badge in repo | 18:23 |
| gouthamr | people are discovering projects using multiple sources: project documentation, project navigator, git repositories | 18:24 |
| opendevreview | Merged openstack/governance master: Move openstack/kuryr and openstack/kuryr-libnetwork to Zun https://review.opendev.org/c/openstack/governance/+/917517 | 18:24 |
| opendevreview | Merged openstack/governance master: Add timeline to remove enforce_scope in RBAC goal https://review.opendev.org/c/openstack/governance/+/915179 | 18:24 |
| gouthamr | gmann brought up the "openstack map" would also need to accommodate this | 18:24 |
| gouthamr | #link https://object-storage-ca-ymq-1.vexxhost.net/swift/v1/6e4619c416ff4bd19e1c087f27a43eea/www-assets-prod/openstack-map/openstack-map-v20240401.pdf (current "openstack landscape" or map | 18:25 |
| gouthamr | #action gtema will make a proposal to the governance repo regarding tagging inactive projects more prominently | 18:26 |
| gouthamr | #topic Next steps on PyPi Maintainers cleanup (gmann) | 18:27 |
| gouthamr | the existing state is recorded in this etherpad | 18:27 |
| gouthamr | #link https://etherpad.opendev.org/p/openstack-pypi-maintainers-cleanup (Notes on PyPi maintainership of OpenStack packages) | 18:27 |
| gouthamr | we've to break the package list into two: one where openstack-ci is a co-maintainer/owner; and other where openstack-ci doesn't figure in the maintainers list | 18:28 |
| gouthamr | #action gouthamr will work on this list | 18:29 |
| gouthamr | #topic 2024.2 TC Tracker | 18:31 |
| gouthamr | #link https://etherpad.opendev.org/p/tc-2024.2-tracker (Technical Committee activity tracker) | 18:31 |
| gouthamr | we discussed items that needed more TC review atm | 18:34 |
| gouthamr | re: dropping eventlet vs replacing eventlet with asyncio | 18:34 |
| gouthamr | #link https://review.opendev.org/c/openstack/governance/+/902585 (Migrate eventlet usages to asyncio) | 18:35 |
| gouthamr | #link https://review.opendev.org/c/openstack/governance/+/916546 (Goal: Remove eventlet requirements from shared libraries) | 18:35 |
| gouthamr | #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/PY6NUNYFU63XOIZHDDBONJDHRXJMZF5J/ ([all][oslo][neutron][glance][ironic] Deprecating eventlet based WSGI server) | 18:35 |
| gouthamr | projects seem to be concerned that the big leap from eventlet to asyncio needs to be figured out via https://review.opendev.org/c/openstack/governance/+/902585 | 18:37 |
| gouthamr | and we may be impeding individual project progress | 18:37 |
| gouthamr | JayF suggests that we should clarify that the TC doesn't stand in the way of projects exploring options to replace eventlet | 18:38 |
| gouthamr | we're not going to have a one-size-fits-all solution | 18:38 |
| gouthamr | JayF suggests communicating this to the ML is required now | 18:39 |
| gouthamr | ^ this seems like a general consensus in the TC | 18:43 |
| gouthamr | hberaud ^ | 18:44 |
| gouthamr | we'd continue chatting on the gerrit change.. | 18:44 |
| JayF | I will take an action to chat with hberaud and try to find a soft landing for that proposal and communicate that to the larger community (via gerrit/ML) | 18:47 |
| gouthamr | #topic Open Discussion | 18:49 |
| gouthamr | JayF++ | 18:49 |
| gouthamr | why does devstack take a longer time on zuul jobs vs github actions | 18:50 |
| gouthamr | clarkb brought this up multiple times in the past | 18:51 |
| gouthamr | #link https://review.opendev.org/c/openstack/devstack/+/676016 (OSC as a service on devstack) | 18:53 |
| gouthamr | ^ this proposal is ancient in our release timeline | 18:53 |
| gouthamr | the underlying issue seems to be the token caching | 18:54 |
| gouthamr | in OSC | 18:54 |
| gouthamr | thank you all for joining! | 18:59 |
| gouthamr | lets catch up here in a week! | 18:59 |
| gouthamr | #endmeeting | 18:59 |
| opendevmeet | Meeting ended Tue May 7 18:59:57 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 18:59 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/tc/2024/tc.2024-05-07-18.00.html | 18:59 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/tc/2024/tc.2024-05-07-18.00.txt | 18:59 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/tc/2024/tc.2024-05-07-18.00.log.html | 18:59 |
| clarkb | gtema: https://zuul.opendev.org/t/openstack/build/3b50d468adfb4946adc28f3ca8e8b145/log/controller/logs/devstacklog.txt#21104-21120 is the recorded timing output | 19:04 |
| clarkb | osc is the largest time block and announted time is actually pretty small now. Thats nice | 19:04 |
| clarkb | the runtime there was closer to 20 minutes than 40 fwiw | 19:05 |
| clarkb | so half not a quarter. But still pretty bad | 19:05 |
| clarkb | though maybe need to cross check against the wall clock on the job too | 19:05 |
| gtema | clarkb: on GH from what I see osc is half of that (239s), but also pip_install is half of the time | 19:06 |
| clarkb | 17:10:46.019046 is roughly when devstack stack.sh starts and 17:29:53.817236 is when it ends so that accounting seems to be rouhgly accurate with the wall clock too | 19:10 |
| clarkb | gtema: another thing to keep in mind is if you have tempest disabled that prevents a lot of tempest setup from running which may speed your jobs up | 19:10 |
| gtema | right | 19:10 |
| clarkb | gtema: re python startup costs I think one of the things that made it so insidious is that everyone running python on their laptops never really noticed since they likely have small numbers of packages installed into a venv on an ssd or nvme drive. But when you get to clouds with metered iops it is really apparent. I also did comparisons between spinning drives and ssds back | 19:18 |
| clarkb | when and that was very noticeable too | 19:18 |
| clarkb | but it gets worse the more stuff you have installed and teh more versions you have installed beacuse pkg_resources/importlib also sort all the packages and their versions in addition to the filesystem scan | 19:18 |
| gtema | Clarkb, I perfectly know that since I was investigating and tracing this heavily. Actually because of that I more or less gave up with the current OSC/SDK in py and decided to focus on rust | 19:24 |
| gtema | It's not repairable without huge redesign, but in that case it makes no sense to stick to py (for me personally) | 19:25 |
| clarkb | to be fair its repairable in the same way as an implementation in most other languages though: don't use dynamic plugin loading and install all the code upfront | 19:26 |
| gtema | A pure "import openatack" is already taking 0.9sec on my desktop | 19:26 |
| clarkb | yes python will still be generally slow, but a big part of startup cost would be removed that way | 19:26 |
| gtema | Big - yes. But I am still not convinced in performance of py | 19:27 |
| spotz[m] | Sorry all! In a different time zone and missed joining | 19:55 |
| gouthamr | spotz[m]: np; i didn't mark absence in the official meeting note; but dansmith mentioned that you're at the RH summit | 20:04 |
| dansmith | *mentioned that he thought she said she was...at summit | 21:13 |
| spotz[m] | I put it in channel before I left so he spoke fact.:) I was OpenStacking:) | 21:42 |
| * gouthamr missed it | 21:43 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!