Thursday, 2020-06-18

openstackgerritBrin Zhang proposed openstack/python-cyborgclient master: Remove translation sections from setup.cfg
*** brinzhang has joined #openstack-cyborg01:02
*** brinzhang_ has joined #openstack-cyborg01:59
*** brinzhang has quit IRC02:02
*** brinzhang0 has joined #openstack-cyborg02:26
*** Yumeng has joined #openstack-cyborg02:26
*** brinzhang_ has quit IRC02:29
*** xinranwang_ has joined #openstack-cyborg02:43
*** brinzhang_ has joined #openstack-cyborg02:47
*** brinzhang0 has quit IRC02:50
*** chenke has joined #openstack-cyborg03:01
*** brinzhang_ is now known as brinzhang03:01
chenke#info chenke03:01
Yumenghi all03:01
Yumeng#startmeeting openstack-cyborg03:01
openstackMeeting started Thu Jun 18 03:01:45 2020 UTC and is due to finish in 60 minutes.  The chair is Yumeng. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.03:01
*** openstack changes topic to " (Meeting topic: openstack-cyborg)"03:01
openstackThe meeting name has been set to 'openstack_cyborg'03:01
brinzhang#info brinzhang03:02
Yumeng#topic Roll call03:02
*** openstack changes topic to "Roll call (Meeting topic: openstack-cyborg)"03:02
Yumeng#info Yumeng03:02
Yumeng#topic Agenda03:02
*** openstack changes topic to "Agenda (Meeting topic: openstack-cyborg)"03:02
Yumengpls check the agenda for today.03:03
brinzhangIt seems not many people for today meeting...03:03
Yumengbrinzhang: have you checked that if your browser blocked the captcha?03:04
brinzhangYumeng: I dont know why, and I cannt save it03:05
brinzhangI want to add the agenda is: #nova-cyborg-interaction(server operations):
songwenping_hi all03:06
brinzhangThe firewall?03:06
Yumengpossibly the popup blocking03:07
xinranwang_Hi all03:09
YumengHi xinranwang_ and songwenping_03:09
openstackgerritMerged openstack/python-cyborgclient master: Remove translation sections from setup.cfg
xinranwang_#info xinranwang_03:09
Yumengbrinzhang: do you wanna bring something about  #nova-cyborg-interaction(server operations)?03:10
songwenping_#info songwenping_03:10
brinzhangYumeng: No, just some progress, because some works, I have not add the UT for the new patch03:11
*** wangzhh has joined #openstack-cyborg03:11
Yumengok. fine. just take your time.03:12
Yumeng#topic smartnic update03:12
*** openstack changes topic to "smartnic update (Meeting topic: openstack-cyborg)"03:12
Yumengsmartnic integration update
brinzhangand I will talked with nova team, accelerate the integration of Sundar and sean patch03:13
Yumengwhich patch do you mean by "Sundar and sean patch"?03:14
brinzhangfrom the ML,about the smartnic, there are some different opinions, right? xinranwang_?03:14
brinzhangYumeng: the patch in the link I was pasted above03:15
Yumengbrinzhang: yes. I would suggest that we start with a nova-spec to continue discuss the smartnic integration.03:17
brinzhangan optimization patch for songwenping_, I think it's good to go, please review, it has been a long time03:17
brinzhangYumeng:agree, we need a nova-spec, and show our views in the spec03:18
brinzhangjust in ML, I dont think it can reach our goal.03:18
xinranwang_brinzhang: yes, sean mooney has diffirent opinions.03:19
Yumeng brinzhang:xinranwang_ did the overall proposal, and she is preparing the nova-spec.03:20
xinranwang_Yes, I am preparing a nova-spec03:21
Yumengbrinzhang: but we may not have time to attend nova team meeting every time, pls sync with us if nova has questions or new ideas.03:22
brinzhangxinranwang_: cool ^^03:22
Yumengif necessary, we can join their meeting .03:23
brinzhangYumeng: Sorry, I am also not attended every time for nova meetting, the meetting time is UTC 14:00, it's too later for us03:23
xinranwang_If you have some other opinions or preferation,  please let us know, and we will consider more.03:23
brinzhangbut, we can review the meeting irc log03:23
Yumengyes. that's fine. no worries.03:23
brinzhangbefore the meetting, send the idea to gibi, or others core03:23
*** shaohe_feng has joined #openstack-cyborg03:25
Yumengok,good. let's go to next topic.03:25
YumengRelease schedule and storyboard update
YumengDoes anyone wants to bring up anything about the milestones or storyboard?03:26
Yumenghi shaohe_feng03:28
shaohe_fengYumeng: morning03:28
Yumengshaohe_feng: do you wanna Attribute API be implemented? I not yet created a story, just worrying about if we can finish.03:30
shaohe_fengWe can create one03:30
wangzhhI can take it. :)03:31
shaohe_fenggreat, thanks. wangzhh03:31
Yumengwangzhh: cool! thanks03:31
wangzhhHaha Noop03:32
wangzhhAny context before? And do we need a proposal first?03:32
shaohe_fengwangzhh: yes.03:33
shaohe_fengone scenarios is that:03:33
Yumengwangzhh: yes,I think a spec is necessary since it is a API change.03:34
wangzhhYumeng: Yep03:34
shaohe_fengthe FPGA driver discovery the resource, report them to cyborg DB and placement.03:34
brinzhangneed to bump a microversion03:34
shaohe_fengbut the function may change, we can get the function id, it is machine readable03:35
shaohe_fengwe want to set a alias for it to make it human readable.03:36
shaohe_fengso we want to a attribute API for us to do it.03:37
wangzhhIt makes sense. shaohe_feng.03:38
YumengREF: from L20003:38
wangzhhYumeng, Cool03:39
YumengThanks wangzhh.03:40
Yumengand Added a new community goal:
YumengI will investigate if we need do something and talk about it later.03:40
shaohe_feng (y)03:41
wangzhhLet me make a proposal this week.03:41
Yumengjust wants to mention in case anyone is interested.03:41
shaohe_fenggood (y)03:41
Yumenglast topic03:42
Yumeng#topic review patches03:42
*** openstack changes topic to "review patches (Meeting topic: openstack-cyborg)"03:42
YumengWe will hit the milestone1 tomorrow. So pls review and merge patches planned for mileston1 including 'remove OPAE dependency' patches and ussuri legacy patches.(patch topics include: devstack-remove-opae, unittest.mock, drop_future_imports, drop_mock,reset_conf, tox etc.)03:43
brinzhangThese patch ready to go03:43
Yumengyes mock patches are ready to W+103:43
brinzhangYumeng: the remove opae patch, is it necessary backport to train?03:44
shaohe_fengHi all There is a issue about glance API in cyborg.03:44
brinzhangIMO, backport to ussuri is enough03:44
xinranwang_I agree to backport to U03:45
shaohe_fengI want to get the image information from glance and download it for program.03:45
shaohe_fengbut I find the keystone adapter can not help to get the glance endpoint03:46
Yumengok. I'm fine with to u release.03:46
brinzhangYumeng: and the backport pach need use cherry-pick, I found you dont do that, or I missed some thing?03:46
YumengI did cherry-pcik to Ussuri.03:46
brinzhangbut you lost the (cherry picked from commit 7ee7de1ec9f546f7ff45b51f74e54a52ad6083f7) in commit message03:47
Yumengbut not cherry-pick to train because train is quite different, there is conflict.03:47
brinzhangit should keep in commit message03:47
shaohe_feng^ I can not get endpoint here.03:48
Yumengbrinzhang: not sure if that's the case. I didn't delete anything in commit message.03:48
shaohe_feng^  I also can not get endpoint here.03:48
brinzhangYumeng: I cherry-picked again, and it's ok now, the same patach03:49
Yumengok. great.03:50
shaohe_fengFile "/opt/stack/cyborg/cyborg/common/", line 213, in get_endpoint03:50
shaohe_feng  "Could not find requested endpoint for any of the following "03:50
shaohe_fengkeystoneauth1.exceptions.catalog.EndpointNotFound: Could not find requested endpoint for any of the following interfaces: ['internal', 'public']03:50
brinzhangshao_feng: which patch?03:50
shaohe_fengbrinzhang the upstream code03:51
shaohe_fengwe used it test for our FPGA03:51
shaohe_fengwe need to get the glance endpoint before download image03:52
shaohe_fengbut it failed03:52
shaohe_fengthe information is above03:52
shaohe_fengso what's wrong?03:52
shaohe_fengdo we need extra config for glance?03:52
brinzhangyou can paste you progress in, and you can try to talk with alex_xu, may can get some help03:53
brinzhangafter you paste the progress, I will look into03:54
Yumengshaohe_feng:    do you know what does efried's note mean?  # NOTE(efried): Id9bd19cca68206fc64d23b0eaa95aa3e5b01b676 may also do the03:55
Yumeng# a list.  (It could also happen with a real auth if the endpoint isn't03:55
brinzhangI dont similar this scenario for you test error now :(, sorry shaohe_feng03:55
Yumengis shaohe_feng still around?03:58
shaohe_fengI have read efried note03:58
shaohe_fengthe glance image handle the endpoint is different from others03:59
shaohe_fengit will join version twice03:59
shaohe_fengsuch as: /image/v2/v2/balabala04:00
Yumengsounds interesting..04:00
shaohe_fengthe first time efried write a special code to get endpoint for image04:01
shaohe_fengif error, it will try another way to get the endpoint04:01
Yumengshaohe_feng: does this need a real image to test? can we test this with fake image and fake driver?04:02
shaohe_feng these two logic I have past in above04:02
Yumengaha. cool.04:02
shaohe_feng^ Yumeng, you can have a look at it04:03
brinzhangI remember download image need to steps, should combine the glance api, but I am not sure04:03
shaohe_fengthe logic it not complex.04:03
shaohe_fengYumeng let me find the bug description about image04:04
openstackLaunchpad bug 1707995 in Glance Client "Can't use versioned endpoints" [Low,Triaged]04:05
shaohe_feng^ it is here04:05
shaohe_fengBut it think it is not the root cause fro my issue.04:05
shaohe_fengnot sure04:06
shaohe_fengI can not get the endpoint from keystone adaptor04:06
shaohe_fengmaybe I need dive into it04:06
shaohe_fengjust want to know anyone know it04:07
Yumengok. I will take a look. but I don't have an real env to test the program.  :(04:07
shaohe_fengor who introduce glance API into  cyborg04:07
shaohe_fengmay know better about it.04:07
Yumengmaybe you can also discuss with shogo. he may know something.04:08
YumengI will  also take a look at this issue. ^^04:09
Yumengjust ping in wechat, so that we all can see that.04:09
Yumengops. we seems run out of time.04:10
Yumeng#topic AOB04:10
*** openstack changes topic to "AOB (Meeting topic: openstack-cyborg)"04:10
Yumengshaohe_feng: or ML is better? glance guys can see it and discuss04:11
shaohe_fengYumeng thanks04:11
Yumengis there anything else for today?04:11
Yumengif nothing else let's wrap up today's meeting. Thank you all.04:12
*** openstack changes topic to "Pending patches (Meeting topic: openstack-cyborg)"04:12
openstackMeeting ended Thu Jun 18 04:12:29 2020 UTC.  Information about MeetBot at . (v 0.1.4)04:12
openstackMinutes (text):
*** links has joined #openstack-cyborg05:46
*** links has quit IRC06:07
*** links has joined #openstack-cyborg06:12
*** chenke has quit IRC06:18
*** links has quit IRC06:32
*** links has joined #openstack-cyborg07:01
*** links has quit IRC08:00
*** songwenping__ has joined #openstack-cyborg08:03
*** songwenping_ has quit IRC08:06
*** links has joined #openstack-cyborg08:10
efriedYumeng, shaohe_feng, brinzhang: If you need help with that glance endpoint thing, I might be able to remember the details, if you catch me during US Central daytime.09:26
*** links has quit IRC09:27
efriedBut IIRC what happened here was that cyborg copied in a bunch of nova code, warts and all.09:27
efriedI can't remember if we since fixed up the nova side in this area. We might have. That would be the first thing to check. Then you could assimilate whatever nova is doing to get around this.09:27
shaohe_fengefried thanks you09:28
efriedwhat would be better is if cyborg used openstacksdk instead.09:28
shaohe_fengYes,  need your help09:28
*** links has joined #openstack-cyborg09:28
efriedIf you jump in #openstack-sdks, they might be able to help you get that spun up.09:29
shaohe_fengdo you means use openstacksdk instead of keystone adapter?09:29
efriedWe started doing that cutover in nova. It was a pretty simple transition.09:29
efriedWe did it for placement, and got most of the way there for ironic.09:29
efriedlet me find those patches...09:30
shaohe_fengsounds good.09:30
shaohe_fengcurrently, I can not get  the image endpoint from keystone adapter09:31
shaohe_fengAnd for the service_catalog is empty.09:32
efriedHere's a good place to start poking:
shaohe_fenglet me read up ti.09:35
shaohe_fengis it convenience to change the current keystone adapter to openstacksdk09:36
shaohe_fenghow many effort?09:36
shaohe_fengsuch as how many days?09:37
efried"It depends" :P09:37
efriedThe transition from ksa should be very easy, ~1d (of course blown out across many calendar days for review cycle etc)09:38
efriedWhen we cut placement over, the transition involved very little code.09:38
shaohe_fengsounds good. does placement call glance?09:39
shaohe_fengI see your ksa code, there are some special handle for image?09:40
efriedNo, sorry, what I mean is:09:40
efriedFor a given $service that $your_project calls via the API, if today you're using ksa to get a handle on that API, transitioning to calling $service via sdk is easy.09:40
efriedWe did this for $service=placement from $your_project=nova.09:40
efriedI'm suggesting you do it for $service=glance from $your_project=glance09:40
efriedI'm suggesting you do it for $service=glance from $your_project=cyborg09:41
efried <== Here's what I was referring to earlier. Nova got rid of that chunk of ugliness. Your shortest path may be to subsume the same change.09:41
efriedYou'll need to bump your ksa minimum to 3.3.0 (which I guess had already happened by the time ^ went in)09:43
shaohe_fengOK, any special extra configure in conf file for it?09:44
efriedI don't think so. In fact, I think you should be able to leave the [glance] section *empty* and it should "just work"09:45
efriedHere's the commit that cut nova over to using the sdk for placement:
efriedIt was built on top of the commit(s) that introduced get_sdk_adapter, which I linked above. But you can see, once that was in place, this change was very simple.09:45
shaohe_fengthat's good. this is more easy for deployer with less configure.09:46
efriedIn fact, on the code side, it consisted only of s/get_ksa_adapter/get_sdk_adapter/ -- the rest was just test fixups.09:46
shaohe_feng^ in your early code, you do some extra handle for image.09:47
efriedOh, yeah, that's super stale. See the commit above for things you can do at a minimum to clean that up.09:47
shaohe_fengafter sdk, can we remove it?09:47
efriedYeah, if you cut over to sdk, you wouldn't need get_endpoint *or* the things that call it.09:48
shaohe_fengsounds good. let me try.09:49
efriedHey, cyborg already has get_sdk_adapter :P
efriedyou may want to just make sure it's in sync with what's in nova.09:49
efriedby the way, long-term, you'll want to use the higher-level interfaces that sdk provides to talk to glance, rather than the REST primitives. Depending how much bandwidth you have, and how disruptive whatever you're working on is going to be, you may be able to do some of that now.09:50
efriedAgain, the folks in #openstack-sdks would be delighted to help you with this.09:52
shaohe_fenggood thanks09:53
efriedmordred (US) and gtema (EU I think) would be good resources there.09:53
shaohe_fengIs sdk compatible with ksa?09:53
efriedYes, at its lowest level, sdk is simply a wrapper around ksa. That's why the placement thing worked (s/get_ksa_adapter/get_sdk_adapter/). But the sdk builds on that so you can do things more powerfully than just the REST primitives.09:54
efried...on a per-service basis.09:55
shaohe_fengyes, we should use sdk.09:55
efriedIf you're familiar with the legacy python-*client projects -- openstacksdk is supposed to replace all of those and be a one-stop shop for that kind of usage, with a (mostly) common interface across all the services.09:55
efriedThat was a multi-year mission that I was a part of before I had to leave. mordred should still be excited about it, I think.09:56
shaohe_fengDoes that mean SDK is more easy to use than python client?09:57
shaohe_fengOK, thanks.,09:57
shaohe_fenghave a nice day.09:57
efriedyou too09:57
efriedgood luck!09:57
*** Yumeng has quit IRC09:58
*** shaohe_feng has quit IRC10:29
*** wangzhh has quit IRC10:31
openstackgerritShogo Saito proposed openstack/cyborg master: Programming support (v2 Deployable API)
openstackgerritPratik Raj proposed openstack/cyborg master: optimize size and time using --no-cache-dir
openstackgerritPratik Raj proposed openstack/cyborg-tempest-plugin master: optimize size and time using --no-cache-dir
*** links has quit IRC15:44
*** songwenping_ has joined #openstack-cyborg16:22
*** songwenping__ has quit IRC16:25
*** songwenping__ has joined #openstack-cyborg18:41
*** songwenping_ has quit IRC18:44
*** songwenping_ has joined #openstack-cyborg19:00
*** songwenping__ has quit IRC19:03
*** gmann is now known as gmann_afk19:22
*** xinranwang_ has quit IRC19:33
*** gmann_afk is now known as gmann19:37
*** gmann is now known as gmann_afk20:10
openstackgerritMonty Taylor proposed openstack/cyborg master: Remove api_servers references
openstackgerritMonty Taylor proposed openstack/cyborg master: Remove deprecated
openstackgerritMonty Taylor proposed openstack/cyborg master: Remove api_servers references
openstackgerritMonty Taylor proposed openstack/cyborg master: Remove deprecated
openstackgerritMonty Taylor proposed openstack/cyborg master: Removed unused image api methods
*** gmann_afk is now known as gmann22:04
*** tetsuro has joined #openstack-cyborg23:47

Generated by 2.17.2 by Marius Gedminas - find it at!