08:02:11 <yasufum> #startmeeting tacker
08:02:11 <opendevmeet> Meeting started Tue Apr 22 08:02:11 2025 UTC and is due to finish in 60 minutes.  The chair is yasufum. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:02:11 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:02:11 <opendevmeet> The meeting name has been set to 'tacker'
08:02:38 <yasufum> #link https://etherpad.opendev.org/p/tacker-meeting
08:03:08 <yasufum> So, let's start today's meeting.
08:04:07 <yasufum> It's just one item on the etherpad
08:04:27 <yasufum> PTG Suggestions On Optimizing Test Execution Efficiency and Resource Consumption in Zuul CI
08:04:57 <yasufum> It' a recap the discussion of the previous vPTG
08:07:21 <yasufum> shivam: could you elaborate on today's discussion point?
08:07:32 <shivam> yes
08:08:48 <shivam> This discussion proposal is recap of proposal discussed in Tacker PTG recently and there are some suggestions from PTG discussion.
08:09:31 <shivam> so, currently tests like tacker-ft-v1-userdata-vnflcm and tacker-ft-v1-compliance-sol are split into multiple jobs to reduce individual execution time, but this significantly increases the number of parallel jobs and total node consumption (up to 128 nodes per patch).
08:10:55 <shivam> There is some feedback from the PTG such that tacker-ft-v1-userdata-vnflcm can be shifted to weekly execution
08:11:17 <shivam> tacker-ft-v1-compliance-sol test, which ensures SOL v1 API compliance, can be moved to local-only execution, since the spec is finalized and no longer evolving.
08:12:00 <shivam> tacker-ft-v2-compliance-sol test (yet to be implemented) is more relevant due to active development of the v2 API, and can be executed periodic-weekly.
08:12:28 <shivam> On-demand or experimental jobs are proposed for tests that don't need to run in every patchset, helping save resources while maintaining coverage.
08:13:25 <shivam> Hence, I would like to know the community members opinion on these proposed suggestions.
08:13:32 <shivam> thanks
08:14:08 <yasufum> Thanks
08:14:54 <yasufum> LGTM for your proposal.
08:15:12 <takahashi-tsc> As yasufumi san said, this is recap and we would like to start community discussion, such as hear the opinions of KDDI etc.
08:15:32 <takahashi-tsc> In addition, also want to discuss "on-demand" testing.
08:15:58 <yasufum> yes
08:20:45 <yasufum> Sorry, I've disconnected shortly.
08:21:52 <yasufum> you mean how we define such a experimental job or so?
08:22:11 <yasufum> about "on-demand" testing.
08:23:17 <yasufum> I basically agree to not to run every time tests for such a experimental features.
08:23:38 <takahashi-tsc> I thnk yes. Shivam, what do you think about it? Any opinions?
08:26:02 <shivam> yes, I also think it is ok to not run every time  test for experimental
08:26:57 <yasufum> Thanks
08:28:49 <yasufum> IMO, we should run unit tests for such a experimental ones, but no need to implement FTs without if the feature is expected to run in a end-to-end usecase.
08:32:19 <takahashi-tsc> I would like to decide on a concrete first step, but it is not urgent.
08:32:41 <yasufum> yes
08:33:32 <yasufum> I think we don't have any plan for developing such a experimental in this cycle.
08:34:28 <takahashi-tsc> High priority work is discussion on a strategy of periodic testing...
08:34:39 <yasufum> sure
08:35:04 <takahashi-tsc> I will personally ask KDDI members to check today's discussion and share their opinion on etherpad.
08:35:18 <yasufum> thanks
08:36:25 <yasufum> Although I think it doesn't look experimental for proposal from KDDI this cycle, it's helpful.
08:38:32 <yasufum> Any other comment?
08:39:55 <yasufum> good
08:40:37 <yasufum> shivam: Thank you for the progress.
08:41:07 <shivam> thank you for the support
08:42:07 <yasufum> OK, the topic is done. So, let's close this meeting.
08:42:16 <yasufum> Thank you for joining, bye!
08:42:57 <shivam> thanks, bye
08:43:06 <takahashi-tsc> thanks
08:43:06 <yasufum> #endmeeting