08:04:02 <yasufum> #startmeeting tacker
08:04:02 <opendevmeet> Meeting started Tue Nov 16 08:04:02 2021 UTC and is due to finish in 60 minutes.  The chair is yasufum. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:04:02 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:04:02 <opendevmeet> The meeting name has been set to 'tacker'
08:04:53 <esto-aln> Hello Everyone, Takahashi-san is in another meeting and cannot participate today.
08:06:10 <yasufum> OK, thanks.
08:06:35 <yasufum> There are three topics on etherpad today.
08:06:43 <yasufum> #link https://etherpad.opendev.org/p/tacker-meeting
08:07:25 <yasufum> I am not sure who is the owner of the first item.
08:08:03 <h_asahina> Sorry, it's me.
08:08:17 <h_asahina> I forgot to write my name.
08:08:43 <yasufum> and it is helpful to change the color of hightlight.
08:09:03 <yasufum> Anyway, could you start your topic?
08:09:48 <h_asahina> ok.
08:10:32 <h_asahina> As I wrote here, I recently found the bug that VNF Instantiaion with Legacy API fails when heat tries to bind `net_mgmt` port to a server.
08:11:47 <h_asahina> First of all, I'd like to confirm does anyone have the same problem
08:13:59 <h_asahina> If nobody face this problem, I'd like to ask you whether we should fix this soon or not.
08:14:36 <h_asahina> As it's a legacy API, I guess there aren't so many people using it.
08:16:54 <yasufum> Have you confirmed the error is caused in takcer, or underlying component, might be heat or neutron?
08:17:22 <h_asahina> I think the problem happens between Nova and Neutron
08:17:38 <yasufum> Is it neutron considering the error log?
08:17:45 <yasufum> OK, i see.
08:17:52 <h_asahina> yes
08:18:13 <h_asahina> I've confirmed it by manually runnning `openstack server ...` command
08:19:19 <h_asahina> So, it might be a matter of the OpenStack deployment, and not Tacker's bug.
08:20:16 <h_asahina> but, at least, the environment using local.conf in the tacker repository incurs this problem in my environment.
08:22:47 <yasufum> matter of the OpenStack deployment > I think so...
08:23:30 <yasufum> Anyway, I will try the same on my env.
08:24:05 <yasufum> so that we will discuss later.
08:24:21 <h_asahina> sounds good. thanks
08:24:57 <yasufum> Any other comment, or go to the next topic?
08:25:26 <ueha> Do you mean that it can happen with SOL compliant APIs with similar settings?
08:26:16 <h_asahina> No. This problem doesn't happen when using SOL complicant APIs.
08:27:13 <h_asahina> though I'm not sure the reason why...
08:27:48 <ueha> Okay I see, thanks.
08:28:11 <h_asahina> thanks
08:29:44 <yasufum> Thanks for your comment.
08:31:20 <yasufum> The second topic is mine.
08:31:59 <yasufum> We have found that fenix plugin is not activated in the examples of local.conf with k8s
08:32:30 <yasufum> while the error of previous fenix was fixed by ueha.
08:33:21 <yasufum> Fenix is supported in tacker officially, so it should be included in the example basically.
08:33:48 <yasufum> I would like to hear your comment if you opposite this change.
08:34:38 <yasufum> Do you think there is any problem if fenix is included in k8s example?
08:36:23 <ueha> I thinks it should be included and it may be no problem.
08:36:48 <yasufum> OK, thanks. I will do that.
08:37:01 <yasufum> any other comment?
08:38:21 <yasufum> good
08:38:49 <yasufum> go to the final topic
08:39:05 <yasufum> ueha: can you share your topic?
08:39:34 <ueha> Sure, this is about the CI error we discussed last week.
08:40:01 <ueha> The FT error that I reported last week has been resolved in patch [1], but there is another error in gate job..
08:40:07 <ueha> [1] https://review.opendev.org/c/openstack/tacker/+/817183
08:41:00 <ueha> Failure in "tox -e docs" that showing following error: "undefined label: types_typedecorator"
08:41:55 <ueha> This error caused by updating SQLAlchemy version to 1.4.27 on Nov 11.
08:42:39 <ueha> This error of tox-docs job seems to be problem between SQLAlchemy and Sphinx, so I think there are following resolutions:
08:43:15 <ueha> 1. Error point is db/types.py. Therefore this error can be fixed by excluding db/types.py from sphinx-apidoc. and I posted the patch [2] and test is passed.
08:43:23 <ueha> [2] https://review.opendev.org/c/openstack/tacker/+/817953
08:44:18 <ueha> 2. Exclude "SQLAlchemy 1.4.27" from tacker's requirements and openstack/requrements.
08:45:16 <ueha> Since the gate of many patches is stopped, first of all, I think that 1. should be implemented as a workaround in tacker.
08:45:47 <ueha> Do you have any other good idea or comment?
08:47:25 <ueha> If there is no problem with solution 1 in particular, I would like you to review patch [2].
08:47:51 <yasufum> Is something dropped from API docs if you add 'db/types.py' in apidoc_excluded_paths?
08:48:18 <yasufum> for [2]
08:49:53 <ueha> I think it dropped the tacker.db.types pages of documentation -> https://docs.openstack.org/tacker/latest/contributor/api/tacker.db.types.html
08:53:42 <yasufum> My another concern is there is an possibility the same problem is happened in other projects although nothing on any projects from my short survey.
08:54:23 <yasufum> [2] looks good to me as an temporary and local fix for tacker.
08:54:46 <ueha> sure
08:55:43 <ueha> I think it should be fixed with Sphinx or SQLAlchemy.
08:55:46 <yasufum> But I also think it does not take times so much to fix global requirements in ‘requirements’ project.
08:56:02 <yasufum> If the team agree to do so.
08:56:16 <yasufum> It’s just adding the excluded version.
08:57:08 <yasufum> Although I am still not sure it is ideal way for fixing.
08:58:25 <h_asahina> I'm concerned that the problem happens again in the future when the SQLAlchemy is updated.
09:00:34 <yasufum> It can be avoided by using lower thant 1.4.27, but we cannot use any other updates in the later versions…
09:00:35 <ueha> Yes I'm concerned same things. If it isn't fixed in SQLAlchemy or Sphinx, we are going to keep updating requirements.
09:02:39 <ueha> and as yasufum says, we can't use the latest version SQLAlchemy... uum..
09:03:36 <h_asahina> I guess we might have to contact with SQLAlchemy or Sphinx community to deal with this problem. Is it possible?
09:05:17 <yasufum> I think you can do that, but it is something difficult to ask to stop using `:ref:` attribute.
09:07:05 <yasufum> Anyway, in my opinion, [2] is acceptable although we will lost a API doc for `tacker.db.types`
09:07:43 <yasufum> because our top priority is to recover the gate test.
09:08:10 <yasufum> as a temporary solution.
09:08:22 <h_asahina> I agree
09:08:30 <ueha> I think so. thanks
09:10:50 <yasufum> thanks
09:11:32 <yasufum> please continue to discuss for the details on gerrit review if you have any comment
09:13:15 <yasufum> It is the end of the time for today’s meeting.
09:13:42 <yasufum> So, I would like to wrap up this meeting if you have no more comemnts.
09:14:07 <ueha> Sorry, I have one thing.
09:14:13 <yasufum> ok
09:14:42 <ueha> next week (Nov 23), Japan is national holiday.
09:15:14 <ueha> so, is next weekly meeting cancel?
09:17:03 <yasufum> Considering not so many people joining from other than
09:17:13 <yasufum> japan, it seems OK
09:17:49 <ueha> I see, thanks
09:18:42 <yasufum> I would like to make an announcement for that because of “my absence”.
09:19:58 <yasufum> and ask someone to host the meeting if anyone want to have a meeting at the day.
09:21:06 <yasufum> thanks for your good suggestion!
09:21:31 <ueha> :)
09:21:37 <yasufum> Is there any other comment?
09:22:11 <yasufum> good
09:22:19 <yasufum> Let’s close this meeting.
09:22:27 <yasufum> Bye!
09:22:30 <ueha> thanks, bye!
09:22:31 <masaki-ueno> bye
09:22:37 <manpreetk> bye
09:22:39 <esto-aln> bye
09:22:39 <h_asahina> bye
09:22:56 <yasufum> #endmeeting