| *** inteq has joined #openstack-zaqar | 00:01 | |
| openstackgerrit | Merged openstack/zaqar: Make websocket run in correct logging mode https://review.openstack.org/233746 | 00:12 |
|---|---|---|
| *** flwang has quit IRC | 00:23 | |
| *** flwang has joined #openstack-zaqar | 00:23 | |
| *** Qiming_ has joined #openstack-zaqar | 00:24 | |
| flwang | vkmc: ping | 00:34 |
| vkmc | flwang, pong | 00:34 |
| flwang | sorry, I missed your email on 14 Oct | 00:35 |
| vkmc | np! | 00:35 |
| flwang | my thounderbird is very slow | 00:36 |
| flwang | would you mind summarizing it at here? :) | 00:36 |
| vkmc | sure thing | 00:37 |
| flwang | :-D | 00:37 |
| vkmc | so, the javascript client in the Horizon side is working as expected | 00:37 |
| vkmc | at least, it receives messages from the queue | 00:38 |
| vkmc | the response to auth and subscription create at least | 00:38 |
| vkmc | in the Ceilometer side is like wise | 00:38 |
| vkmc | the publisher starts, recovers all the events and sends them to the queue | 00:38 |
| vkmc | now, when the messages reach horizon queue | 00:38 |
| vkmc | for some reason Zaqar doesn't notify Horizon | 00:39 |
| flwang | really? | 00:39 |
| vkmc | I'm suspecting of the subscription being created several times | 00:39 |
| flwang | it doesn't make sense | 00:39 |
| vkmc | I dunno how to show you... I could share my screen but its not a good idea if you have a slow connection | 00:40 |
| flwang | did you debug it? | 00:40 |
| vkmc | yeah I was trying to | 00:40 |
| vkmc | I tried to create a subscription with another client (just curling to the api) | 00:40 |
| flwang | i think the key part is you need to make sure there is/are subscriptions for the queue | 00:40 |
| vkmc | but I'm hitting errors that I believe are because I'm using it wrongly | 00:41 |
| flwang | and make sure when there is a msg posted to the queue, notifier.py is called to check the subscriptions and iterate the loop | 00:41 |
| flwang | what's the errors? | 00:41 |
| vkmc | 404 :| | 00:42 |
| vkmc | are the subscriptions linked to the pools or something like that? | 00:42 |
| vkmc | because I deactivated the pools | 00:43 |
| vkmc | curl -i -X GET http://localhost:8888/v2/subscription -H "X-Auth-Token: $TOKEN" <- is this ok to list the existing subscriptions? | 00:43 |
| flwang | subscriptions | 00:43 |
| flwang | not subscription | 00:43 |
| * vkmc tries again | 00:44 | |
| flwang | vkmc: any luck? | 00:51 |
| vkmc | devstack running | 00:52 |
| flwang | ok | 00:54 |
| vkmc | vkmc@trusty:~$ curl -i -X GET http://192.168.122.167:8888/v2/subscriptions -H "X-Auth-Token: $TOKEN" | 00:58 |
| vkmc | HTTP/1.1 404 Not Found | 00:58 |
| *** achanda has quit IRC | 00:59 | |
| vkmc | http://paste.openstack.org/show/476573/ | 01:00 |
| flwang | hmm | 01:00 |
| vkmc | and that is the uwsgi log | 01:00 |
| flwang | a | 01:01 |
| flwang | it's not correct | 01:01 |
| vkmc | now... when subscription_create is issued in the javascript | 01:01 |
| flwang | the url should be /v2/queues/{queue_name}/subscriptions | 01:01 |
| vkmc | it works | 01:02 |
| vkmc | ohh | 01:02 |
| vkmc | let me try that | 01:02 |
| *** achanda has joined #openstack-zaqar | 01:02 | |
| vkmc | vkmc@trusty:~$ curl -i -X GET http://192.168.122.167:8888/v2/queues/horizon/subscriptions -H "X-Auth-Token: $TOKEN" | 01:02 |
| vkmc | HTTP/1.1 200 OK | 01:02 |
| vkmc | content-length: 101 | 01:02 |
| vkmc | content-type: application/json; charset=utf-8 | 01:02 |
| vkmc | super | 01:02 |
| vkmc | :D | 01:02 |
| flwang | i'm going to have breakfast, leave msg at here | 01:03 |
| flwang | brb | 01:03 |
| vkmc | k | 01:03 |
| vkmc | enjoy | 01:03 |
| vkmc | :) | 01:03 |
| flwang | back | 01:12 |
| *** achanda has quit IRC | 01:15 | |
| flwang | any findings? | 01:20 |
| vkmc | so... authenticated, subscription created | 01:24 |
| vkmc | launching and instance to see where it stops | 01:25 |
| flwang | cool, you can add break points at notifier.py and webhook.py | 01:25 |
| vkmc | yeah | 01:28 |
| vkmc | I just got a bunch of notifications | 01:28 |
| vkmc | maybe my environment had something weird going on at the moment | 01:28 |
| vkmc | now I have pooling enabled again | 01:28 |
| flwang | so it works? | 01:28 |
| vkmc | http://paste.openstack.org/show/476574/ | 01:33 |
| vkmc | not really | 01:33 |
| vkmc | I just suspended the instance and that happened | 01:34 |
| vkmc | I keep testing :) | 01:42 |
| flwang | you said you can see notifications, at that moment, you didn't enable pooling, right? | 01:43 |
| flwang | and then you tried to enable pooling to see if the notifications works or not, is it? | 01:44 |
| vkmc | yeah | 01:45 |
| vkmc | (more on the same http://paste.openstack.org/show/476577/) | 01:45 |
| flwang | each time when your restart the horizon page, current demo code will create a new subscription | 01:48 |
| vkmc | yeah | 01:48 |
| vkmc | I thought that might affect | 01:48 |
| vkmc | I was right? | 01:48 |
| flwang | and the 'subscriber' is using the protocol id | 01:48 |
| flwang | that means | 01:48 |
| flwang | when you refresh the page, the old connection is gone | 01:49 |
| flwang | that's why you are seeing the 'Connection aborted' in webhook's log | 01:49 |
| vkmc | ohhh | 01:49 |
| vkmc | so we should check if the subscription exist | 01:50 |
| *** Eva-i has quit IRC | 01:50 | |
| flwang | not really | 01:50 |
| flwang | based on current design of notification websocket | 01:50 |
| flwang | the subscriber for each connection is different | 01:51 |
| vkmc | I see | 01:51 |
| flwang | so it's hard to verify if the subscription is existing | 01:51 |
| flwang | but | 01:51 |
| vkmc | I remember we discussed about this some time ago | 01:51 |
| flwang | it shouldn't impact the demo | 01:51 |
| flwang | let me repeat the work flow to see if we missed anything | 01:52 |
| flwang | 1. create a queue, like 'horizon' | 01:52 |
| flwang | 2. open the instances page of horizon, which will create subscrption for the 'horizon' queue | 01:52 |
| flwang | at this moment, do not refresh the page | 01:53 |
| flwang | check if the subscription has been created successfully | 01:53 |
| flwang | 3. suspend an instance | 01:53 |
| flwang | 4. the publisher for zaqar should be able receive the msg and 'forward' it to the 'horizon' queue | 01:54 |
| flwang | 5. zaqar gets the msg of suspend and trigger the notification | 01:54 |
| vkmc | yeah, the problem is that horizon forces a refresh when you perform an action | 01:55 |
| flwang | 6. the javascript of horizon will be notified by the websocket | 01:55 |
| flwang | oh | 01:55 |
| flwang | do that with cli? | 01:55 |
| flwang | :) | 01:55 |
| vkmc | I wonder if we could stop horizon to refresh | 01:55 |
| vkmc | but that requires a lot more knowledge about it | 01:56 |
| *** amitgandhinz has joined #openstack-zaqar | 01:56 | |
| flwang | what do you mean 'more knowledge'? | 01:56 |
| flwang | just run 'nova suspend <id>' | 01:57 |
| flwang | pls keep there is only one instance to make it easy | 01:57 |
| vkmc | oh I meant that if we wanted to change the refresh rate for Horizon we would need to know Horizon better | 01:59 |
| vkmc | let me try what happens if I perform a nova suspend | 01:59 |
| vkmc | ok the token expired and spoiled me the test heh | 02:02 |
| vkmc | but seems to work | 02:02 |
| * vkmc updates the token | 02:02 | |
| *** Eva-i has joined #openstack-zaqar | 02:05 | |
| flwang | vkmc: hmm... did you remove/comment the line 'horizon.table.update()'? | 02:06 |
| vkmc | I did yes | 02:06 |
| flwang | ok | 02:06 |
| flwang | if so, horizon shouldn't refresh it automatically i think | 02:07 |
| vkmc | hmm, that just affects the table | 02:07 |
| vkmc | I just ran the cli command with the token updated | 02:08 |
| vkmc | and it does the same | 02:08 |
| flwang | can zaqar get the msg? | 02:09 |
| flwang | can you see the subscription? | 02:09 |
| vkmc | so I should be able to remove the subscription | 02:10 |
| vkmc | and create a new one | 02:10 |
| vkmc | right? | 02:10 |
| *** amitgandhinz has quit IRC | 02:12 | |
| flwang | yes | 02:12 |
| flwang | i would suggest to add a breakpoint/log when creating the subscription via websocket to know the subscriber/proto id | 02:13 |
| flwang | so that you can check if the subscription is created successfully or not | 02:14 |
| vkmc | so I added a breakpoint in the websocket side | 02:15 |
| vkmc | sorry, javascript side | 02:15 |
| vkmc | so it shows me when the subscription is created | 02:15 |
| flwang | there are 2 checkpoints: 1. if zaqar get the message from ceilometer publisher 2. if the subscription is created with correct subscriber | 02:16 |
| flwang | if both checkpoints passed, the javascript of horizon should be able be notified | 02:16 |
| vkmc | this is the request/response when the page is refreshed | 02:16 |
| vkmc | http://paste.openstack.org/show/476578/ | 02:16 |
| vkmc | I don't see a proto id though | 02:18 |
| vkmc | it would be very useful to have this running on a public cloud so you can log in there :@ | 02:19 |
| flwang | can you see the subscription? | 02:21 |
| *** achanda has joined #openstack-zaqar | 02:21 | |
| flwang | the proto id is the end string of subscriber | 02:21 |
| vkmc | http://paste.openstack.org/show/476579/ | 02:22 |
| vkmc | several subscriptions | 02:23 |
| flwang | http://trusty:36153/d25d636b-c004-47c0-b6b7-dc287895cc90 | 02:26 |
| flwang | d25d636b-c004-47c0-b6b7-dc287895cc90 is the proto id | 02:26 |
| vkmc | yeah | 02:26 |
| flwang | now when you post a message to the queue, can you see the notification on the horizon javascript side? | 02:27 |
| flwang | or can zaqar get the message and trigger the notification? | 02:27 |
| flwang | no matter if it's a nova suspend msg | 02:29 |
| flwang | you can add an alert() on the javascript side, so that you can see the msg when horizon get it | 02:30 |
| vkmc | yeah, until that point we are good | 02:30 |
| vkmc | maybe I have to test things in a clean env | 02:31 |
| vkmc | right now its failing to get the subscription controller | 02:31 |
| vkmc | bbl, very late dinner | 02:31 |
| vkmc | I have some ideas to try later | 02:31 |
| vkmc | thanks for shed some light on the issue :D | 02:32 |
| flwang | vkmc: my pleasure | 02:32 |
| flwang | i can give it a try if you can let me know where is your code | 02:33 |
| Eva-i | Hello. I noticed there was almost no activity in chat during previous weekend. Most of you rest from zaqar at weekends, right? | 02:36 |
| flwang | Eva-i: yep? :) | 02:45 |
| flwang | Eva-i: anything i can help? | 02:45 |
| Eva-i | flwang: yes | 02:47 |
| flwang | Eva-i: do you have any question? | 02:50 |
| Eva-i | flwang: I had, but I just figured it out now | 02:51 |
| flwang | Eva-i: cool | 02:51 |
| *** flwang has quit IRC | 02:59 | |
| *** Eva-i has quit IRC | 03:05 | |
| *** Eva-i has joined #openstack-zaqar | 03:07 | |
| *** exploreshaifali has joined #openstack-zaqar | 05:51 | |
| *** exploreshaifali has quit IRC | 06:36 | |
| *** exploreshaifali has joined #openstack-zaqar | 06:40 | |
| *** achanda has quit IRC | 06:41 | |
| *** achanda has joined #openstack-zaqar | 06:45 | |
| *** achanda has quit IRC | 07:07 | |
| *** tmu has quit IRC | 07:12 | |
| *** exploreshaifali has quit IRC | 07:16 | |
| *** mdnadeem_ has joined #openstack-zaqar | 07:16 | |
| *** exploreshaifali has joined #openstack-zaqar | 07:29 | |
| *** dynarro has joined #openstack-zaqar | 08:01 | |
| *** exploreshaifali has quit IRC | 09:30 | |
| *** dynarro has quit IRC | 10:00 | |
| *** achanda has joined #openstack-zaqar | 11:07 | |
| *** achanda has quit IRC | 11:07 | |
| *** exploreshaifali has joined #openstack-zaqar | 11:49 | |
| *** mpanetta has joined #openstack-zaqar | 12:53 | |
| *** mpanetta has quit IRC | 12:57 | |
| *** exploreshaifali has quit IRC | 13:01 | |
| *** dynarro has joined #openstack-zaqar | 13:20 | |
| *** david-ly_ has quit IRC | 14:40 | |
| *** itisha has joined #openstack-zaqar | 14:59 | |
| *** davideagnello has quit IRC | 16:02 | |
| *** Qiming_ has quit IRC | 16:48 | |
| *** achanda has joined #openstack-zaqar | 17:06 | |
| -openstackstatus- NOTICE: Gerrit will be offline for project renames starting at 1800 UTC. | 17:10 | |
| *** ChanServ changes topic to "Gerrit will be offline for project renames starting at 1800 UTC." | 17:10 | |
| *** dynarro has quit IRC | 17:11 | |
| *** achanda has quit IRC | 17:26 | |
| -openstackstatus- NOTICE: Gerrit is offline for project renames. | 18:02 | |
| *** ChanServ changes topic to "Gerrit is offline for project renames." | 18:03 | |
| *** achanda has joined #openstack-zaqar | 18:37 | |
| *** ChanServ changes topic to "OpenStack Queuing and Notification Service || Smile :D || Meetings every Monday @ 21:00 UTC/15:00 UTC || Wiki: https://wiki.openstack.org/wiki/Zaqar || Paste: http://paste.openstack.org/ || Send messages and make some noise :D" | 18:39 | |
| -openstackstatus- NOTICE: Gerrit is back online. Github transfers are in progress and should be complete by 1900 UTC. | 18:39 | |
| *** openstackgerrit has quit IRC | 18:46 | |
| *** openstackgerrit has joined #openstack-zaqar | 18:46 | |
| *** mpanetta has joined #openstack-zaqar | 19:01 | |
| *** mpanetta has quit IRC | 19:05 | |
| *** itisha has quit IRC | 19:21 | |
| *** openstackgerrit has quit IRC | 19:31 | |
| *** openstackgerrit has joined #openstack-zaqar | 19:31 | |
| *** achanda has quit IRC | 19:49 | |
| *** david-lyle has joined #openstack-zaqar | 20:13 | |
| *** achanda has joined #openstack-zaqar | 21:10 | |
| *** mpanetta has joined #openstack-zaqar | 21:41 | |
| *** mpanetta has quit IRC | 21:46 | |
| *** davideagnello has joined #openstack-zaqar | 22:08 | |
| *** davideagnello has quit IRC | 22:09 | |
| *** davideagnello has joined #openstack-zaqar | 22:09 | |
| *** davideagnello has quit IRC | 22:11 | |
| *** achanda has quit IRC | 22:32 | |
| *** flwang has joined #openstack-zaqar | 23:07 | |
| *** flwang has quit IRC | 23:30 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!