| *** openstack has joined #openstack-zaqar | 00:16 | |
| *** zigo has quit IRC | 00:17 | |
| *** david_cu has quit IRC | 00:17 | |
| *** vipul has quit IRC | 00:17 | |
| *** zigo has joined #openstack-zaqar | 00:24 | |
| *** david_cu has joined #openstack-zaqar | 00:24 | |
| *** vipul has joined #openstack-zaqar | 00:24 | |
| *** david_cu has quit IRC | 00:28 | |
| *** wanghao has joined #openstack-zaqar | 00:37 | |
| *** harlowja has joined #openstack-zaqar | 01:04 | |
| *** Kevin_Zheng has joined #openstack-zaqar | 01:30 | |
| flwang2 | wanghao: did you see my comments on the service queue spec? does that make sense for you? | 02:04 |
|---|---|---|
| *** yangzhenyu has quit IRC | 02:06 | |
| *** yangzhenyu has joined #openstack-zaqar | 02:06 | |
| *** lei-zh has joined #openstack-zaqar | 02:12 | |
| flwang2 | lei-zh: hello | 02:22 |
| lei-zh | flwang2, hi | 02:24 |
| flwang2 | lei-zh: did you get a chance to test the poc of service queue? | 02:25 |
| flwang2 | does that work for you? | 02:25 |
| flwang2 | if we can get this feature done in pike-3, do you think the zaqar publisher feature can be merged in searchlight as well? | 02:26 |
| lei-zh | yeah, I did some tests on the api, it works fine | 02:28 |
| lei-zh | but there lacks some fine grained role control, like access control by a user's role | 02:30 |
| flwang2 | could you please explain it a bit? | 02:33 |
| flwang2 | what do you mean access control by a user's role? | 02:33 |
| flwang2 | can searchlight get the user's role based on the notification? | 02:33 |
| lei-zh | there are some admin only fields in the message searchlight are going to publish, we don't want non admin users see these fields | 02:35 |
| flwang2 | lei-zh: ah, yep, i see. so there are some notifications are for admin user, and we don't want to expose them to the user's queue, right? | 02:37 |
| flwang2 | do you have any idea for that? | 02:37 |
| lei-zh | flwang2, yes, of course searchlight could create a dedicated queue to store these messages, but there is concern that non admin user could also see the messages. | 02:38 |
| flwang2 | lei-zh: for those notifications, can't you just create queues in the service tenant? | 02:39 |
| flwang2 | if the notifications are not allowed to be shown to end user, then there is no reason to create queues in customer tenant | 02:40 |
| lei-zh | flwang2, sorry, I have to pick up a call, will be back later | 02:41 |
| flwang2 | ok | 02:41 |
| lei-zh | flwang2, ping | 03:01 |
| flwang2 | yes | 03:01 |
| lei-zh | it's kind like api query, there are some fields only expose to user with admin role | 03:02 |
| lei-zh | if you are a admin role user, you will see them | 03:03 |
| lei-zh | searchlight api just mimic behavior of openstack service apis like nova | 03:04 |
| flwang2 | lei-zh: ok, for that case, I can't see zaqar can do much things | 03:05 |
| flwang2 | in zaqar, you can't manage the message at the field level | 03:05 |
| lei-zh | this is something needs to have a look at, at first maybe searchlight could choose not to expose these admin fields until we find a solution for this case | 03:06 |
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Update subscription confirm doc https://review.openstack.org/475629 | 03:08 |
| wanghao | flwang2: yes, I see your comments, that make sense to me | 03:08 |
| wanghao | flwang2: trust way seems a little complex for service | 03:08 |
| flwang2 | lei-zh: yep, that's what I'm thinking | 03:09 |
| flwang2 | for the case of Horizon, we don't really need the 'admin' fields | 03:09 |
| flwang2 | horizon just cares about the notification to update resource status | 03:09 |
| wanghao | lei-zh: yes, I get your issue too, in nova or cinder, that will be controllers by project before showing them to user. | 03:10 |
| flwang2 | and we can definitely improve later | 03:10 |
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Update subscription confirm doc https://review.openstack.org/475629 | 03:10 |
| *** wxy has joined #openstack-zaqar | 03:11 | |
| lei-zh | flwang2, I thought horizon is going to use the notification for display purpose, to get rid of api pollings | 03:13 |
| lei-zh | if so, they may need these admin fields | 03:13 |
| flwang2 | lei-zh: why? IMHO, they just care about the new status and the resource id | 03:14 |
| flwang2 | why do they need the admin fields? | 03:14 |
| lei-zh | flwang2, ok, if so that's fine | 03:14 |
| lei-zh | I may misunderstand, because what searchlight publishes is not simply a notification | 03:15 |
| lei-zh | it's enriched resource information, like what you query from openstack service api | 03:15 |
| flwang2 | interesting, you mean, the data forwarded to zaqar queues' are not notifications, but the whole resource object info, like we run 'nova show xxx'? | 03:17 |
| wanghao | you mean some messages are notifications, some are resource obejct info? | 03:18 |
| lei-zh | yeah, that's searchlight trying to do, index all the openstack resources | 03:19 |
| lei-zh | wanghao, searchlight listen on mq for notifications, and turn notifications into resource info | 03:19 |
| lei-zh | and we index and publish resource info | 03:20 |
| wanghao | lei-zh: ok i see, like user stop a vm, so the message is not only the acton notification, but also including the vm infor. right> | 03:20 |
| wanghao | right? | 03:20 |
| lei-zh | yes | 03:20 |
| flwang2 | lei-zh: but that doesn't block the user case of horizon i think | 03:21 |
| flwang2 | if horizon can get the latest status, then should be fine | 03:21 |
| wanghao | em argee | 03:21 |
| flwang2 | no matter it's a notification or a full object | 03:22 |
| wanghao | horizon can just take what it want from the message. | 03:22 |
| flwang2 | exactly | 03:22 |
| flwang2 | lei-zh: did you update your patch based on wanghao's poc patch? | 03:22 |
| lei-zh | flwang2, if horizon only wants latest status, that's fine, if they also want resource detail, the admin case matters | 03:23 |
| flwang2 | lei-zh: as far as know, for now, there are 2 user cases | 03:23 |
| flwang2 | 1. update resource status | 03:23 |
| flwang2 | 2. a central notification center | 03:23 |
| flwang2 | like a message box on dashboard | 03:23 |
| flwang2 | and using zaqar as the backend | 03:24 |
| wanghao | I got a question, if horizon need some admin fields to show, that means user could see them. Then it's ok we post those fileds to queue. | 03:26 |
| wanghao | if horizon didn't want user to see those, them shouldn't post those filelds to queue. | 03:27 |
| lei-zh | but these fields should be exposed to admin role users, if searchlight posts them to the queue, is there a potential security issue that non admin users of same project could get the message from the queue | 03:28 |
| lei-zh | flwang2, I haven't update my patch, it's based on zaqarclient to operate queues, I'll continue works on it | 03:30 |
| wanghao | lei-zh: yes, I see this, could we ensure what admin fields are, and to check if horizon need them? | 03:31 |
| lei-zh | wanghao, sure, I'll ping them to figure it out | 03:31 |
| flwang2 | i don't want to go further before we figure out the exact user case | 03:36 |
| flwang2 | lei-zh: wanghao: so let's contact with horizon team and get feedback from them instead of thinking a loud :D | 03:36 |
| lei-zh | flwang2, ++ | 03:40 |
| flwang2 | lei-zh: so could you please reply the thread and said you have tested the wanghao's patch, and it works. now we need feedback from Horizon team to move on | 03:44 |
| wanghao | lei-zh: flwang2: agree with you | 04:21 |
| *** flwang has quit IRC | 05:11 | |
| *** pcaruana has joined #openstack-zaqar | 05:14 | |
| *** pcaruana has quit IRC | 05:30 | |
| *** tesseract has joined #openstack-zaqar | 06:40 | |
| *** pcaruana has joined #openstack-zaqar | 06:49 | |
| *** rcernin_ has joined #openstack-zaqar | 06:53 | |
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Update subscription confirm doc https://review.openstack.org/475629 | 07:00 |
| *** lei-zh has quit IRC | 07:30 | |
| *** lei-zh1 has joined #openstack-zaqar | 07:30 | |
| *** pcaruana|afk| has joined #openstack-zaqar | 07:35 | |
| *** pcaruana|afk| has quit IRC | 07:35 | |
| *** pcaruana has quit IRC | 07:36 | |
| *** wanghao_ has joined #openstack-zaqar | 07:36 | |
| *** wanghao has quit IRC | 07:39 | |
| *** pcaruana has joined #openstack-zaqar | 07:40 | |
| *** wanghao_ has quit IRC | 07:51 | |
| *** wanghao has joined #openstack-zaqar | 07:51 | |
| *** gengchc has joined #openstack-zaqar | 07:54 | |
| *** wanghao has quit IRC | 07:56 | |
| openstackgerrit | wangxiyuan proposed openstack/zaqar master: Update subscription confirm doc https://review.openstack.org/475629 | 07:57 |
| *** gengchc_ has joined #openstack-zaqar | 07:58 | |
| *** gengchc has quit IRC | 07:58 | |
| *** gengchc_ has quit IRC | 08:02 | |
| *** wanghao has joined #openstack-zaqar | 08:05 | |
| wanghao | wxy: ping | 08:08 |
| wanghao | wxy: plz have a look this bug: https://review.openstack.org/#/c/473651/ | 08:08 |
| wanghao | wxy: thanks | 08:08 |
| wxy | wanghao: Done. | 08:14 |
| wanghao | thanks | 08:14 |
| openstackgerrit | wanghao proposed openstack/python-zaqarclient master: Fix the creation issue when special meanings words in queue name https://review.openstack.org/473651 | 08:31 |
| openstackgerrit | Merged openstack/python-zaqarclient master: Add subscription other properties https://review.openstack.org/474936 | 08:36 |
| openstackgerrit | Merged openstack/zaqar master: Fix some typos in zaqar https://review.openstack.org/474818 | 08:39 |
| *** openstackgerrit has quit IRC | 08:48 | |
| *** wanghao has quit IRC | 09:08 | |
| *** wanghao has joined #openstack-zaqar | 09:08 | |
| *** lei-zh1 has quit IRC | 09:28 | |
| *** wanghao has quit IRC | 09:33 | |
| *** Aaaa has joined #openstack-zaqar | 09:46 | |
| *** Aaaa has quit IRC | 10:07 | |
| *** flwang has joined #openstack-zaqar | 10:09 | |
| *** lhx has joined #openstack-zaqar | 10:32 | |
| *** lhx is now known as Guest42538 | 10:32 | |
| *** openstackgerrit has joined #openstack-zaqar | 11:55 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-zaqarclient master: Updated from global requirements https://review.openstack.org/456013 | 11:55 |
| *** flwang has quit IRC | 11:55 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/zaqar master: Updated from global requirements https://review.openstack.org/474699 | 11:56 |
| *** therve_ is now known as therve | 12:41 | |
| *** Guest42538 has quit IRC | 12:54 | |
| -openstackstatus- NOTICE: restarting gerrit to address slowdown issues | 12:57 | |
| openstackgerrit | Houpeng Liu proposed openstack/zaqar master: Fix html_last_updated_fmt for Python3. https://review.openstack.org/475825 | 14:31 |
| *** mpanetta_ has joined #openstack-zaqar | 14:34 | |
| *** rcernin_ has quit IRC | 14:47 | |
| *** pcaruana has quit IRC | 16:19 | |
| *** tesseract has quit IRC | 17:00 | |
| *** yangzhenyu has quit IRC | 20:03 | |
| *** flwang has joined #openstack-zaqar | 21:07 | |
| *** yangzhenyu has joined #openstack-zaqar | 22:06 | |
| *** openstack has joined #openstack-zaqar | 22:15 | |
| *** yangzhenyu has quit IRC | 22:17 | |
| *** yangzhenyu has joined #openstack-zaqar | 22:30 | |
| *** yangzhenyu has quit IRC | 22:55 | |
| *** mpanetta_ has quit IRC | 23:07 | |
| *** harlowja has quit IRC | 23:19 | |
| *** wanghao has joined #openstack-zaqar | 23:43 | |
| *** mpanetta has joined #openstack-zaqar | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!