02:00:03 <wanghao> #startmeeting zaqar
02:00:04 <openstack> Meeting started Tue Mar 20 02:00:03 2018 UTC and is due to finish in 60 minutes.  The chair is wanghao. Information about MeetBot at http://wiki.debian.org/MeetBot.
02:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
02:00:07 <openstack> The meeting name has been set to 'zaqar'
02:00:22 <wanghao> hello everyone
02:03:04 <wanghao> yangzhenyu: hi
02:03:38 <yangzhenyu> hi
02:03:47 <wanghao> seems now only you and me....
02:04:03 <yangzhenyu> e...
02:04:03 <wxy> o/
02:04:11 <wanghao> wxy: hi
02:05:30 <yangzhenyu> wanghao, What are we going to discuss?
02:05:33 <wanghao> okay let rock
02:05:39 <flwang1> o/
02:05:49 <wanghao> #topic rocky scheduler
02:05:50 <yangzhenyu> hi feilong
02:06:01 <wanghao> flwang1: hi
02:06:11 <wanghao> https://releases.openstack.org/rocky/schedule.html
02:06:36 <wanghao> here is the rocky schedule
02:07:17 <wanghao> just reminder guys to follow the time line.
02:07:40 <flwang1> wanghao: got, thanks
02:07:48 <wanghao> flwang1: np
02:07:53 <flwang1> btw, I would suggest to change our meeting from this channel to #zaqar
02:08:03 <flwang1> i think it would be easier for people to join
02:08:17 <wanghao> en that's good too.
02:08:33 <wanghao> I will update our wiki for this change.
02:08:42 <wanghao> everyone else is ok with this?
02:08:53 <yangzhenyu> I am ok
02:08:57 <flwang1> wanghao: you may also need a change for the meeting repo, IIRC
02:09:23 <wanghao> wow, thanks for this tip.
02:10:01 <wanghao> ok. let's jump to next
02:10:13 <wanghao> #topic Rocky tasks
02:10:21 <wanghao> https://etherpad.openstack.org/p/zaqar-rocky-plan
02:11:50 <wxy> is any feature under development? Then we can target them to R-1 IMO.
02:12:19 <wanghao> queue filter is one from me.
02:13:29 <wanghao> and others like delete message with clain id also can target R-1 too.
02:14:25 <wanghao> so guys pls push those task if you're interest in it when you have time.
02:14:26 <yangzhenyu> ok
02:14:35 <wanghao> I know everyone is very busy.
02:14:55 <wxy> some specs should be updated first.
02:15:28 <wanghao> yes
02:16:01 <wanghao> let's enhance our review for this specs and let them merged soon before R-1
02:17:06 <wanghao> do anything else want to discuss for those tasks?
02:17:34 <yangzhenyu> Websocket enhancement
02:18:25 <yangzhenyu> I can add some functional like delay queue or dead letter queue
02:18:36 <yangzhenyu> for websocket
02:18:50 <wxy> we need a zuul job first
02:19:20 <wxy> now our tempest test jobs are all based on wsgi.
02:20:05 <yangzhenyu> alright, I got it.
02:20:15 <wanghao> em,  have any idea to add the test job?
02:20:56 <wxy> base on zuul v3, it's easy to add now.
02:21:07 <wxy> update the .zuul.yaml
02:21:35 <yangzhenyu> any docs?
02:21:53 <wxy> finding.
02:22:02 <wanghao> yangzhenyu: I think you can list the detailed jobs including the functional adding and test job. we can follow those tasks.
02:22:14 <wanghao> wxy: thanks
02:22:50 <yangzhenyu> ok
02:25:29 <wanghao> and we also need to see how to implement the tempest test for websocket.
02:25:45 <wanghao> not very clear for me at least.
02:27:04 <wanghao> will try it.
02:27:38 <yangzhenyu> First of all, what tests are needed for websocket? tempest? unit? functional? interg?
02:28:56 <wanghao> I think tempest is ok
02:29:39 <wxy> can't find the exact url. I'd do it later.
02:30:34 <wanghao> wxy: okay
02:31:07 <wanghao> yangzhenyu: I think we can add some tempest tests fist, build a basic job in zuul.
02:32:36 <yangzhenyu> en ok, first add tempest tests, then improve unit test and functional test. Finally, improve zaqar function, I think.
02:32:53 <wxy> yangzhenyu: ++
02:33:29 <wxy> I can start an original patch for zuul jobs
02:33:52 <wxy> then feel free to update it. :)
02:34:26 <yangzhenyu> ok
02:34:48 <yangzhenyu> Does this need to register bp?
02:35:29 <wanghao> yangzhenyu: I'm thinking if ut is necessary, but others I agree too.
02:35:40 <wxy> good to have
02:35:47 <wanghao> +1
02:36:34 <wanghao> wxy: you can register one, thanks
02:37:03 <wxy> sure
02:37:34 <wanghao> okay, next topic
02:37:40 <wanghao> #topic free discuss
02:37:56 <wanghao> yangzhenyu: I want to discuss your patch: https://review.openstack.org/#/c/517901/
02:38:17 <yangzhenyu> ok
02:39:07 <flwang1> wanghao: as I suggested before, I would recommend to talk to united stack (tong fang cloud) to get some feedback from them about zaqar
02:40:00 <wxy> flwang1: any guys we can connect to ?
02:40:06 <wanghao> flwang1: yes,  recently I am finding the united stack guys from webchat
02:40:52 <wanghao> I know a easycloud guy who may know some guys.
02:41:54 <wanghao> I have joined some webchat group, want to get some feedback.
02:42:14 <flwang1> I just posted the contact card in our wechat group
02:42:26 <flwang1> you guys can talk to him
02:42:39 <wanghao> flwang1: sure thanks
02:44:26 <wanghao> yangzhenyu: back to your patch,  I see your issue.
02:45:36 <wanghao> yangzhenyu: but I feel the ttl that zaqar modifies,  the new ttl will mean the time that this message could be lived more.
02:46:04 <yangzhenyu> wanghao, this patch to solve bug/1730335. Our msg ttl is unchanged, But when the msg is claim status, it may change to smaller.
02:46:15 <wanghao> So,  ttl=claim_ttl+claim_grace make sense
02:46:26 <wanghao> yes, it be smaller
02:46:46 <yangzhenyu> But ttl is unchanged at the beginning.
02:47:47 <yangzhenyu> always unchange, so it can not mean the msg could be lived more.
02:48:24 <yangzhenyu> It is more like a total time to live.
02:48:43 <wanghao> the meaning of ttl has been changed after modify it.
02:48:54 <yangzhenyu> Not a remaining time.
02:49:45 * wxy leave for some internal things
02:49:52 <wanghao> but the total time to live is not the actual time that message keep in zaqar
02:50:11 <wanghao> it's just a limit.
02:51:59 <wanghao> if we use ttl=ttl+claim_ttl+claim_grace, that means we didn't need claim_ttl and claim_grace any more.
02:52:14 <wanghao> just give a defined value for ttl.
02:52:59 <wanghao> maybe we can split ttl to two parts.
02:53:17 <yangzhenyu> This situation only occurs in the claim state and the message expires soon.
02:53:57 <wanghao> yes,  if user didn't claim the message,  so the ttl will be what?
02:54:08 <wanghao> ttl  or ttl+claim_ttl_claim_grace?
02:54:15 <yangzhenyu> example, ttl
02:54:20 <yangzhenyu> ttl
02:55:03 <wanghao> so,  if we defined the  ttl is unchanged value, we just can use ttl now.
02:55:34 <yangzhenyu> This is also possible
02:56:00 <wanghao> and we add a new metadata,  if a message be claimed and will expires soon,  we add some time to it.
02:56:33 <wanghao> tell the user,  "your message's ttl is expired, but we allow it to live a little more"
02:56:38 <yangzhenyu> However, it may affect the message get because the claim status has not yet ended and the theory cannot delete the message.
02:57:11 <yangzhenyu> ok
02:57:17 <yangzhenyu> I got it
02:57:56 <yangzhenyu> But I can't remember if the msg get will be affected.
02:59:12 <yangzhenyu> When the msg claimed, and expires soon, The message should be accessible.
03:00:13 <wanghao> en  we can discuss it in zaqar channel
03:00:20 <yangzhenyu> ok
03:00:23 <wanghao> time is up now.
03:00:35 <wanghao> #endmeeting