Tuesday, 2022-01-25

*** hemna9 is now known as hemna07:38
yasufumHi tacker team08:00
yasufum#startmeeting tacker08:01
opendevmeetMeeting started Tue Jan 25 08:01:56 2022 UTC and is due to finish in 60 minutes.  The chair is yasufum. Information about MeetBot at http://wiki.debian.org/MeetBot.08:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.08:01
opendevmeetThe meeting name has been set to 'tacker'08:01
yasufum#link https://etherpad.opendev.org/p/tacker-meeting08:03
yasufumesto-aln: thx for updating the etherpad08:04
yasufumand we have two topics today.08:04
yasufumFirst one is from edagawa_kc08:05
yasufumDo you have any trouble for the gate test, right?08:05
yasufumcould you share your topic?08:06
edagawa_kcAt first, thank you for your reviewing the spec of this feature.08:06
edagawa_kcCurrently I'm in functional testing phase and facing two issues.08:07
edagawa_kcIssue 1: Functional test of paging records causes TIMED_OUT in Zuul due to handling large number of pages.08:07
edagawa_kcIssue 2: Appropriate default number of records contained in a page hasn't been decided (currently defined "100" as tentative).08:07
edagawa_kcThe detail is described in etherpad. I would like to ask your opinions/ideas to resolve them.08:08
edagawa_kcRegarding issue 1, I suggested five options to avoid issue. For my side, option 1' or 2 is better way, but I want to ask your opinion.08:11
yasufumBefore that, there are several failures not only in ft but also others such as tox-cover08:12
yasufumDon't we need to take care about them now?08:12
edagawa_kcYes, I have already noticed that and will fix it.08:12
edagawa_kcThat' all from my side.08:15
uehaAs for option 2, you wrote "it affects other existing test cases using tacker.conf.", what kind of effect does it have?08:19
yasufumAlhtough I don't how much time cost for the test itself, is opt 3 available? 08:20
uehaIf paging does not work in other tests, I think there is no effect. right?  If so, I think Option 2 is simple and good.08:22
edagawa_kcAs for the effect of option 2, for example, if we want to verify new test case with using value "3" for future, it needs changing tacker.conf definition.08:26
edagawa_kcI'm afraid the word "existing" is somehow incorrect. It means "the other tests".08:27
uehaThanks, I got it. If we use option 1' or 3, do you mean that we can test according to the value?08:31
edagawa_kcAs for issue 3, current behavior is that instantiation and termination repeats many times and each operation needs time to be completed.08:34
edagawa_kcI intend to fix this behavior more and shorten TAT so as not to become TIMED_OUT.08:35
edagawa_kcFor example, preparing large records as pre-condition, using FAILURE process on purpose, etc.,08:38
yasufumIt seems so expensive comparing with our purpose of the test...08:40
edagawa_kc"If we use option 1' or 3, do you mean that we can test according to the value?" -> In option 1', actual paged behavior is only once.08:41
uehaThe time out occur because it have to do a lot of LCM operations to prepare the data to test the paging query..08:41
yasufumI think the feature is not so important actually as introducing a new FT.08:41
yasufumueha: What do you think?08:41
uehaYes, I think so.08:42
edagawa_kcYou are mentioning Option 4, aren't you?08:44
yasufumIf it's reasonable for the purpose08:44
yasufumIn my conclusion, it's ok to implement is as UT, but how to test should be discussed later, on gerrit possibly.08:45
yasufumedagawa_kc: Is the issue 2 also about the FT?08:48
edagawa_kcIssue 2 affects entire paging feature, but we can also discuss this issue on gerrit.08:50
yasufumgot it08:50
yasufumI've just wondered if we can skip the issue2 because we don't so much time today.08:51
yasufumSo, go to the next topic.08:51
yasufumhirogumi-noguchi: can you share your topic?08:52
hirofumi-noguchiMy topic is about proposal to abolish default vim.08:52
hirofumi-noguchiCurrent Tacker can register default VIM, which is used when some API requests and Grant response does not include vimConnectionInfo > vimId.08:52
hirofumi-noguchiThough this implementation has advantages, there are also problems with vim's policy such as an issue which we have discussed last week in “About multi-tenant policy (ma-ooyama)”. 08:53
hirofumi-noguchiWhen instantiation is executed without specifying VIM, the default VIM of the user who executes instantiate is selected.08:53
hirofumi-noguchiI think vim should be specified in the API request or Grant response basically.08:54
hirofumi-noguchiSo I don't think the lack of a default vim is a problem.08:54
hirofumi-noguchiIn my opinion, it is better that Tacker will return error in cases where the default vim was previously required.08:54
hirofumi-noguchiSince default vim is outside the scope of ETSI NFV SOL, I think it is an implementation issue.08:55
hirofumi-noguchiI would like to hear your opinions.08:55
hirofumi-noguchiThat' all from my side. Thank you.08:55
yasufumI'm not sure the reason why default vim was introduced, I agree with hirofumi08:57
yasufumbecause we don't need to have difficulties discussed as previous meeting08:58
yasufumif we drop the feature08:58
yasufumI think it's simple and better.08:58
hirofumi-noguchiI think the default vim is inherited from the legacy code.09:01
yasufumMy only concern is that there is any operator expectig to use default vim.09:01
hirofumi-noguchiRegarding vim registration, SOL code also uses a legacy implementation.09:01
yasufumIs there any comment for the suggestion?09:03
takahashi-tscI also agree that we do not need features for default VIM, And I also want to know operator's opinion.09:04
ueha+1, sorry for late because my computer was freezed..09:08
yasufumI think it's almost unlikely that operators want to keep default vim in tacker personally.09:09
yasufumThere are no objections right now.09:11
yasufumhirofumi: Could you upload a blueprint for the topic as first step,09:11
hirofumi-noguchiI got it.09:11
ma-ooyamaI don't have clear opinion about whether default vim is needed.09:12
ma-ooyamaBut it is possible that some user may use NFVO that use default VIM because of limitation of its implimentation, but sorry I don't know the case.09:12
yasufumOK, thanks.09:12
yasufumPlease continue to discuss because it's no urgent topic.09:14
yasufumWe can have discussion in the next ptg if it's needed.09:15
hirofumi-noguchiOK. I got it. Thank you.09:17
yasufumma-ooyama: I'd appreciate if you give us any futher update, especially about NFVO stand point.09:17
yasufumAny other comment?09:19
ma-ooyamaSure, I will.09:19
yasufumma-ooyama: thanks09:20
yasufumseems nothing more09:20
yasufumSo, wrap up this meeting09:21
yasufumThank you for joining, bye!09:21
takahashi-tscthanks, bye!09:21
uehaThanks, bye.09:21
ma-ooyamathanks, bye09:21
esto-alnthanks, bye09:21
hirofumi-noguchiThanks, bye.09:21
opendevmeetMeeting ended Tue Jan 25 09:21:44 2022 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)09:21
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tacker/2022/tacker.2022-01-25-08.01.html09:21
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tacker/2022/tacker.2022-01-25-08.01.txt09:21
opendevmeetLog:            https://meetings.opendev.org/meetings/tacker/2022/tacker.2022-01-25-08.01.log.html09:21
*** soniya29 is now known as soniya29|afk11:56

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!