| *** lei-zh has joined #openstack-zaqar | 00:29 | |
| *** wanghao has joined #openstack-zaqar | 00:45 | |
| *** lei-zh1 has joined #openstack-zaqar | 00:51 | |
| *** lei-zh has quit IRC | 00:51 | |
| openstackgerrit | wanghao proposed openstack/zaqar master: Notification Delivery Policy https://review.openstack.org/477724 | 00:54 |
|---|---|---|
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Refactor the doc tree https://review.openstack.org/480066 | 01:04 |
| *** harlowja has quit IRC | 01:13 | |
| *** lhx__ has joined #openstack-zaqar | 01:56 | |
| *** lei-zh1 has quit IRC | 02:04 | |
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Refactor the doc tree https://review.openstack.org/480066 | 02:13 |
| wxy | flwang: Feel free to co-author the patch to make the index page better. :) | 02:13 |
| *** lei-zh has joined #openstack-zaqar | 02:14 | |
| flwang | wxy: cool | 02:17 |
| flwang | wanghao: around? | 02:26 |
| wanghao | yep | 02:26 |
| flwang | i'm revisiting the service queue feature | 02:26 |
| flwang | and i'm thinking if we should turn off the lazy queue feature for service queue feature | 02:26 |
| flwang | that said, if the service user want to post messages to a service queue, the queue must be created by the tenant user | 02:27 |
| flwang | otherwise, the queue can't be created automatically | 02:27 |
| flwang | does that make sense? | 02:27 |
| wanghao | emm, I think it should | 02:27 |
| flwang | what do you mean 'it should'? | 02:28 |
| wanghao | I mean it should turn off | 02:28 |
| flwang | with that way, we can leave an option for the user | 02:28 |
| flwang | to totally turn off the feature | 02:28 |
| wanghao | the control for service queue should be in user's hand. | 02:28 |
| flwang | for him/her and for the tenant | 02:28 |
| wanghao | you mean leave an option to turn off the lazy queue. | 02:29 |
| flwang | no, i mean, give the tenant user an option to turn off the feature though the feature is enabled globally | 02:29 |
| wanghao | emmm, let me see, just turn off the lazy queue in service queue, right? | 02:35 |
| wanghao | so besides the switch of service queue, and there will be one more option that can turn off the lazy queue when using service queue. | 02:37 |
| flwang | if current queue is not service queue, then everything is normal | 02:39 |
| flwang | and I'm thinking if we can totally get rid of the prefix :D | 02:40 |
| flwang | at line 183 https://review.openstack.org/#/c/469343/13/zaqar/transport/wsgi/v2_0/messages.py | 02:40 |
| flwang | we will get the queue metadata, so that said, we can tag the queue if it's a service queue by metadata | 02:41 |
| flwang | wxy: ^ | 02:41 |
| flwang | wanghao: pls help review this one https://review.openstack.org/#/c/480066/13 the doc patch from wxy | 02:48 |
| wanghao | flwang: okay, I will review it. | 02:49 |
| flwang | does above make any sense for you? | 02:49 |
| flwang | we can discuss more later | 02:49 |
| wanghao | flwang: about the prefix, I think we can use metadata to tag it, but I feel it's like the prefix. | 02:52 |
| wanghao | flwang: like metadata _service_queue = True, TBH, I prefer to use prefix, it's more simpler. | 02:55 |
| wanghao | for those user who didn't turn on the service queue feature, service.xxx is a normal queue, just like other queues. | 02:56 |
| wanghao | if user open the service queue feature, that is not different between using service.xxx or _service_queue=True to indicate this queue is a service queue. | 02:58 |
| flwang | " for those user who didn't turn on the service queue feature, service.xxx is a normal queue, just like other queues." how? | 02:58 |
| wanghao | I mean user can name the queue like service.xxx, and use it as normal, no need to worry service project will send messages to it. | 03:00 |
| flwang | how can an user turn off the service queue but still keep the prefix? | 03:00 |
| flwang | why? | 03:00 |
| flwang | I can't follow, sorry | 03:00 |
| wanghao | I got your point, if we use metadata, that use can turn off the single service queue as he want. | 03:03 |
| wanghao | My mean is now we have a global switch to turn off the service queue, if user didn't want use the service queue, then you just don't create a queue named service.xxx | 03:06 |
| flwang | "if user didn't want use the service queue, then you just don't create a queue named service.xxx", that's is based on initial suggestion, right? turn off lazy queue when enabling the service queue, right? | 03:10 |
| flwang | wxy: ping | 03:11 |
| wanghao | flwang: yes | 03:13 |
| flwang | ok | 03:13 |
| wanghao | flwang: so I suggest by default we turn off lazy queue for service queue | 03:15 |
| wxy | flwang: hi | 03:16 |
| wanghao | flwang: about the remove the prefix, I will talked to lei-zh to see if there is some big change to searchlight. | 03:16 |
| wxy | actually, I don't like the reserved metadata usage. How about add more properties to queue? Any way, this work should be done after Pike. | 03:20 |
| wxy | If so, we can use _service_queue = True now. And in Queue, move it to a new queue property. | 03:20 |
| wanghao | there is an issue I think, if we turn off the lazy queue feature, let user to create the service queue, how do they know what name shoud be used? | 03:23 |
| wanghao | now the searchlight will create the queue by itself, use the names that only it knows. | 03:25 |
| wanghao | As current design, if operator closes the service queue feature, searchlight can't create service queue in user's tenant. Just can create a queue in service's project. | 03:28 |
| flwang | wanghao: that's a good question. hence why i think we could use metadata instead of name to tag a service queue | 03:32 |
| wxy | wanghao: What if a normal user create a queue named like "service.xxx"? | 03:34 |
| wanghao | wxy: if the name is same with searchligh want to post, this queue will recevie the notificaiton messages. | 04:26 |
| wanghao | wxy: if the name is not same with searchlight want to post, it's just like a normal queue, nothing will happen. | 04:27 |
| wanghao | flwang: so basically, I think we should not let user to create the service queue by himself, since it's hard to choose the right name. But user can choose close the service queue if he want by using metadata. | 04:29 |
| wanghao | flwang: and we can remove the prefix restriction， searchlight or other service project can create any name they want by using lazy queue feature. | 04:31 |
| *** harlowja has joined #openstack-zaqar | 04:31 | |
| *** harlowja has quit IRC | 04:47 | |
| *** lhx__ has quit IRC | 04:55 | |
| *** lhx__ has joined #openstack-zaqar | 05:02 | |
| *** flwang1 has quit IRC | 05:06 | |
| *** harlowja has joined #openstack-zaqar | 05:46 | |
| *** rcernin has joined #openstack-zaqar | 05:50 | |
| *** wanghao_ has joined #openstack-zaqar | 06:06 | |
| *** wanghao has quit IRC | 06:09 | |
| *** lhx__ has quit IRC | 06:17 | |
| *** lhx__ has joined #openstack-zaqar | 06:24 | |
| *** harlowja has quit IRC | 06:42 | |
| openstackgerrit | gengchc2 proposed openstack/zaqar master: Support dead letter queue for redis https://review.openstack.org/479242 | 07:12 |
| openstackgerrit | gengchc2 proposed openstack/zaqar master: Support dead letter queue for redis https://review.openstack.org/479242 | 07:15 |
| openstackgerrit | gengchc2 proposed openstack/zaqar master: Support dead letter queue for redis https://review.openstack.org/479242 | 07:17 |
| *** tesseract has joined #openstack-zaqar | 07:17 | |
| openstackgerrit | wanghao proposed openstack/zaqar master: Notification Delivery Policy https://review.openstack.org/477724 | 07:59 |
| *** lei-zh has quit IRC | 08:03 | |
| *** lei-zh has joined #openstack-zaqar | 08:04 | |
| *** yangzhenyu has quit IRC | 08:23 | |
| *** lhx__ has quit IRC | 08:25 | |
| *** lhx__ has joined #openstack-zaqar | 08:25 | |
| openstackgerrit | Hangdong Zhang proposed openstack/zaqar master: Update and optimize documentation links https://review.openstack.org/485089 | 08:27 |
| *** yangzhenyu has joined #openstack-zaqar | 08:40 | |
| *** openstackgerrit has quit IRC | 08:49 | |
| *** lei-zh has quit IRC | 08:51 | |
| *** lei-zh has joined #openstack-zaqar | 08:52 | |
| *** lhx__ has quit IRC | 09:03 | |
| *** lhx__ has joined #openstack-zaqar | 09:03 | |
| *** wanghao_ has quit IRC | 09:26 | |
| *** openstackgerrit has joined #openstack-zaqar | 09:40 | |
| openstackgerrit | Merged openstack/zaqar master: Refactor the doc tree https://review.openstack.org/480066 | 09:40 |
| openstackgerrit | gengchc2 proposed openstack/zaqar master: Support dead letter queue for redis https://review.openstack.org/479242 | 09:44 |
| *** lei-zh has quit IRC | 09:53 | |
| *** lhx__ has quit IRC | 10:16 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/zaqar master: Updated from global requirements https://review.openstack.org/484596 | 13:03 |
| *** rajathagasthya has joined #openstack-zaqar | 15:46 | |
| *** wxy- has joined #openstack-zaqar | 16:05 | |
| *** rcernin has quit IRC | 16:09 | |
| *** rcernin has joined #openstack-zaqar | 16:23 | |
| *** harlowja has joined #openstack-zaqar | 17:25 | |
| *** harlowja has quit IRC | 17:25 | |
| *** harlowja has joined #openstack-zaqar | 17:26 | |
| *** tesseract has quit IRC | 17:46 | |
| *** boris-42___ has joined #openstack-zaqar | 18:17 | |
| *** flwang1 has joined #openstack-zaqar | 18:55 | |
| *** rajathagasthya has quit IRC | 19:03 | |
| *** rajathagasthya has joined #openstack-zaqar | 19:46 | |
| *** rajathagasthya has quit IRC | 20:07 | |
| *** rajathagasthya has joined #openstack-zaqar | 20:27 | |
| *** flwang1 has quit IRC | 20:45 | |
| *** boris-42___ has quit IRC | 21:41 | |
| *** rcernin has quit IRC | 23:34 | |
| *** flwang1 has joined #openstack-zaqar | 23:37 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!