| *** felipemonteiro has joined #openstack-horizon | 00:01 | |
| *** felipemonteiro_ has joined #openstack-horizon | 00:02 | |
| *** felipemonteiro has quit IRC | 00:06 | |
| *** felipemonteiro_ has quit IRC | 00:21 | |
| flwang1 | ying_zuo: I mean this patch https://review.openstack.org/510392 | 00:52 |
|---|---|---|
| *** gyee has quit IRC | 01:08 | |
| *** gary-smith_ has quit IRC | 02:07 | |
| *** catintheroof has joined #openstack-horizon | 02:20 | |
| *** dave-mccowan has joined #openstack-horizon | 02:20 | |
| *** dave-mcc_ has quit IRC | 02:22 | |
| *** catintheroof has quit IRC | 02:22 | |
| openstackgerrit | Feilong Wang proposed openstack/horizon master: Make image name on instances panel configurable https://review.openstack.org/510392 | 02:48 |
| *** dave-mccowan has quit IRC | 02:51 | |
| *** erlon has quit IRC | 02:54 | |
| openstackgerrit | Feilong Wang proposed openstack/horizon master: Add white list for Nova extension https://review.openstack.org/511095 | 03:23 |
| *** yamamoto has quit IRC | 03:25 | |
| openstackgerrit | Feilong Wang proposed openstack/horizon master: Rename "Import Key Pair" to "Import Public Key" https://review.openstack.org/505431 | 03:30 |
| *** flwang1 has quit IRC | 04:04 | |
| *** ratailor has joined #openstack-horizon | 04:05 | |
| *** ratailor_ has joined #openstack-horizon | 04:07 | |
| *** ratailor has quit IRC | 04:11 | |
| *** yamamoto has joined #openstack-horizon | 04:13 | |
| *** ratailor__ has joined #openstack-horizon | 04:21 | |
| *** jtomasek has joined #openstack-horizon | 04:22 | |
| *** ratailor_ has quit IRC | 04:22 | |
| *** ratailor_ has joined #openstack-horizon | 04:32 | |
| *** gyee has joined #openstack-horizon | 04:34 | |
| *** ratailor__ has quit IRC | 04:36 | |
| *** yamamoto has quit IRC | 04:43 | |
| *** gyee has quit IRC | 05:10 | |
| *** ratailor__ has joined #openstack-horizon | 05:23 | |
| *** ratailor_ has quit IRC | 05:26 | |
| *** lajoskatona has joined #openstack-horizon | 05:30 | |
| *** udesale has joined #openstack-horizon | 05:38 | |
| *** yamamoto has joined #openstack-horizon | 05:43 | |
| *** yamamoto has quit IRC | 05:49 | |
| *** yamamoto has joined #openstack-horizon | 05:54 | |
| *** yamamoto has quit IRC | 05:58 | |
| *** yamamoto has joined #openstack-horizon | 06:02 | |
| *** yamamoto has quit IRC | 06:04 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/horizon master: Imported Translations from Zanata https://review.openstack.org/510758 | 06:09 |
| *** hoonetorg has quit IRC | 06:17 | |
| *** belmoreira has joined #openstack-horizon | 06:18 | |
| *** yamamoto has joined #openstack-horizon | 06:21 | |
| *** spectr has quit IRC | 06:27 | |
| *** spectr has joined #openstack-horizon | 06:28 | |
| *** hoonetorg has joined #openstack-horizon | 06:31 | |
| *** ratailor_ has joined #openstack-horizon | 06:32 | |
| *** ratailor__ has quit IRC | 06:35 | |
| *** openstackgerrit has quit IRC | 07:03 | |
| *** rcernin has joined #openstack-horizon | 07:05 | |
| *** pcaruana has joined #openstack-horizon | 07:13 | |
| *** tesseract has joined #openstack-horizon | 07:19 | |
| *** ratailor__ has joined #openstack-horizon | 07:21 | |
| *** ratailor_ has quit IRC | 07:24 | |
| *** nikhil has quit IRC | 07:34 | |
| *** yamamoto has quit IRC | 07:35 | |
| *** yamamoto has joined #openstack-horizon | 07:36 | |
| *** yamamoto has quit IRC | 07:36 | |
| *** ratailor__ has quit IRC | 07:36 | |
| *** jpich has joined #openstack-horizon | 07:39 | |
| makowals | Good morning everyone | 07:47 |
| makowals | Question about performance measurements & tuning | 07:48 |
| *** openstackgerrit has joined #openstack-horizon | 07:48 | |
| openstackgerrit | Ferenc Cserepkei proposed openstack/horizon master: Add "prev" link to instance page list pagination https://review.openstack.org/498018 | 07:48 |
| makowals | I have a problems with instance list taking 15 seconds to load. What I have checked so far, the problem is definitely not with the get_data() function when generating the view as this one takes only 2 seconds | 07:48 |
| makowals | Any ideas what can I look for in order to make it faster? | 07:49 |
| *** itlinux has joined #openstack-horizon | 07:52 | |
| *** itlinux has quit IRC | 08:04 | |
| makowals | I have already measured the whole horizon/tables/views.py and this delay is coming somewhere from the outside | 08:13 |
| *** e0ne has joined #openstack-horizon | 08:23 | |
| *** yamamoto has joined #openstack-horizon | 08:23 | |
| *** pcaruana has quit IRC | 08:27 | |
| *** pcaruana has joined #openstack-horizon | 08:31 | |
| *** openstackgerrit has quit IRC | 08:33 | |
| vladiskuz | makowals: hi! What do you mean when you said "outside"? | 08:36 |
| makowals | Outside of the "views.py" file | 08:36 |
| vladiskuz | makowals: we are trying to improve performance of the instances tab. There are several patches on review. You can use profile to understand what's going wrong | 08:40 |
| *** spectr has quit IRC | 08:40 | |
| vladiskuz | makowals: if you want run a profiler I can help you with this | 08:41 |
| makowals | vladiskuz: I would appreciate. I have already made some patches upstream for this, but it seems like apparently for some projects I still have a lot of issues | 08:41 |
| makowals | Also at this moment I'm not really able to tell what really makes "the good" and "the bad" projects different | 08:42 |
| *** ianychoi has quit IRC | 08:42 | |
| *** jeremy_moffitt_ has joined #openstack-horizon | 08:46 | |
| *** gary-smith_ has joined #openstack-horizon | 08:46 | |
| *** jgravel has quit IRC | 08:47 | |
| *** jeremy_moffitt has quit IRC | 08:47 | |
| *** gary-smith has quit IRC | 08:48 | |
| *** spectr has joined #openstack-horizon | 08:50 | |
| *** Sudheer has joined #openstack-horizon | 08:52 | |
| Sudheer | Hello all i have question regarding support for displaying gnoochi/adoh information in horizon | 08:56 |
| Sudheer | I came across a bule print which basically address this issue: https://blueprints.launchpad.net/horizon/+spec/horizon-gnocchi-graphs | 08:57 |
| Sudheer | but i have n't found any useful links from the same to check the progress of UI support | 08:58 |
| *** yamamoto has quit IRC | 08:58 | |
| *** belmoreira has quit IRC | 08:58 | |
| Sudheer | Can any one suggest whether the support for gnoochi/adoh is already exists or in progress | 08:59 |
| *** yamamoto has joined #openstack-horizon | 09:01 | |
| *** yamamoto has quit IRC | 09:06 | |
| *** ianychoi has joined #openstack-horizon | 09:06 | |
| *** belmoreira has joined #openstack-horizon | 09:08 | |
| *** belmoreira has quit IRC | 09:08 | |
| *** belmoreira has joined #openstack-horizon | 09:10 | |
| *** ratailor has joined #openstack-horizon | 09:14 | |
| *** priteau has joined #openstack-horizon | 09:20 | |
| *** tosky has joined #openstack-horizon | 09:21 | |
| *** rmart04 has joined #openstack-horizon | 09:40 | |
| *** mvk has quit IRC | 09:40 | |
| *** belmoreira has quit IRC | 09:53 | |
| *** jamesbenson has joined #openstack-horizon | 09:54 | |
| *** udesale has quit IRC | 09:58 | |
| *** jamesbenson has quit IRC | 09:58 | |
| *** gokhan has quit IRC | 10:03 | |
| -openstackstatus- NOTICE: The CI system will be offline starting at 11:00 UTC (in just under an hour) for Zuul v3 rollout: http://lists.openstack.org/pipermail/openstack-dev/2017-October/123337.html | 10:08 | |
| *** wolverineav has joined #openstack-horizon | 10:08 | |
| *** ratailor has quit IRC | 10:21 | |
| *** ratailor has joined #openstack-horizon | 10:22 | |
| *** itxaka has joined #openstack-horizon | 10:25 | |
| *** early has quit IRC | 10:29 | |
| *** early has joined #openstack-horizon | 10:32 | |
| *** openstackgerrit has joined #openstack-horizon | 10:35 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/django_openstack_auth master: Imported Translations from Zanata https://review.openstack.org/510054 | 10:35 |
| *** ratailor_ has joined #openstack-horizon | 10:41 | |
| *** gokhan has joined #openstack-horizon | 10:41 | |
| *** jgravel has joined #openstack-horizon | 10:44 | |
| *** ratailor has quit IRC | 10:45 | |
| *** itxaka has quit IRC | 10:50 | |
| *** erlon has joined #openstack-horizon | 10:57 | |
| *** itlinux has joined #openstack-horizon | 11:05 | |
| *** mvk has joined #openstack-horizon | 11:12 | |
| *** belmoreira has joined #openstack-horizon | 11:16 | |
| *** hoonetorg has quit IRC | 11:21 | |
| *** udesale has joined #openstack-horizon | 11:21 | |
| *** hoonetorg has joined #openstack-horizon | 11:35 | |
| *** ratailor__ has joined #openstack-horizon | 11:38 | |
| *** rmart04 has quit IRC | 11:41 | |
| *** ratailor_ has quit IRC | 11:41 | |
| *** spectr has quit IRC | 11:44 | |
| *** spectr has joined #openstack-horizon | 11:46 | |
| *** ratailor__ has quit IRC | 11:49 | |
| *** spectr-RH has joined #openstack-horizon | 11:54 | |
| *** yamamoto has joined #openstack-horizon | 11:55 | |
| *** spectr has quit IRC | 11:57 | |
| *** yamamoto has quit IRC | 11:59 | |
| *** dave-mccowan has joined #openstack-horizon | 12:01 | |
| *** jlabarre has joined #openstack-horizon | 12:04 | |
| *** yamamoto has joined #openstack-horizon | 12:08 | |
| *** itlinux has quit IRC | 12:10 | |
| *** spectr-RH has quit IRC | 12:13 | |
| *** spectr has joined #openstack-horizon | 12:13 | |
| *** itlinux has joined #openstack-horizon | 12:14 | |
| *** spectr-RH has joined #openstack-horizon | 12:14 | |
| *** spectr has quit IRC | 12:18 | |
| *** openstackgerrit has quit IRC | 12:18 | |
| *** priteau has quit IRC | 12:30 | |
| *** yamamoto has quit IRC | 12:31 | |
| *** itlinux has quit IRC | 12:35 | |
| *** wolverineav has quit IRC | 12:35 | |
| *** itlinux has joined #openstack-horizon | 12:35 | |
| *** belmoreira has quit IRC | 12:41 | |
| *** catintheroof has joined #openstack-horizon | 12:43 | |
| -openstackstatus- NOTICE: Due to unrelated emergencies, the Zuul v3 rollout has not started yet; stay tuned for further updates | 13:05 | |
| *** zhurong has joined #openstack-horizon | 13:20 | |
| *** wolverineav has joined #openstack-horizon | 13:27 | |
| *** priteau has joined #openstack-horizon | 13:35 | |
| *** tosky_ has joined #openstack-horizon | 13:36 | |
| *** itlinux has quit IRC | 13:36 | |
| *** tosky has quit IRC | 13:38 | |
| *** lblanchard has joined #openstack-horizon | 13:43 | |
| *** udesale has quit IRC | 13:43 | |
| *** jtriley has joined #openstack-horizon | 13:49 | |
| *** udesale has joined #openstack-horizon | 13:55 | |
| *** Tengu has joined #openstack-horizon | 13:57 | |
| Tengu | hello! | 13:57 |
| Tengu | small question: on three load-balanced controllers with openstack HA, we have disconnections from horizon way before the configured session_timeout. How are sessions managed? is it stateless? | 13:58 |
| robcresswell | Tengu: We just hold the token in a cookie, but thats about it | 14:01 |
| Tengu | robcresswell: hmmm ok. and any backend understands that cookie and can match the keytone ticket right? | 14:01 |
| robcresswell | Tengu: It's the keystone auth token that we store, so yes | 14:02 |
| Tengu | hmm ok. so maybe it's a timeout on keystone side | 14:02 |
| *** spectr-RH has quit IRC | 14:04 | |
| robcresswell | iirc we have a session timeout too, and so it'll time out at the lower of either our setting or keystones | 14:04 |
| *** spectr has joined #openstack-horizon | 14:04 | |
| *** chlong has joined #openstack-horizon | 14:05 | |
| Tengu | ok. so might be something like keystone_authtoken/memcache_pool_unused_timeout | 14:05 |
| cshen | Hi all, following the question from yesterday. I enabled the debug log in horizon. and try to make a network_topology call to it directly. it timed out after 5 mins. | 14:05 |
| *** lblanchard1 has joined #openstack-horizon | 14:05 | |
| cshen | the debug only showed [Wed Oct 11 13:38:03.491088 2017] [wsgi:error] [pid 255:tid 140246016505600] [client 10.97.160.27:45328] Timeout when reading response headers from daemon process 'horizon': /openstack/venvs/horizon-14.1.0/lib/python2.7/dist-packages/openstack_dashboard/wsgi/django.wsgi | 14:05 |
| cshen | nothing more. how can I make horizon to give me more verbosed logs? | 14:06 |
| *** lblanchard has quit IRC | 14:06 | |
| cshen | I want to figure where it took most time. | 14:07 |
| *** tosky_ is now known as tosky | 14:10 | |
| robcresswell | cshen: That doesnt look like debug level logging to me :) | 14:10 |
| robcresswell | you'll need to update your settings, then restart horizon | 14:11 |
| robcresswell | May also need some changes to your wsgi server | 14:11 |
| cshen | I did this change in settings. | 14:12 |
| cshen | DEBUG = True | 14:12 |
| cshen | and restarted horizon. | 14:13 |
| *** rmart04 has joined #openstack-horizon | 14:14 | |
| cshen | and I believe horizon did produce debug logs. because for other requets. I saw these logs. | 14:16 |
| cshen | [Wed Oct 11 13:48:32.385371 2017] [wsgi:error] [pid 251:tid 140246174328576] DEBUG:keystoneauth.session:REQ: curl -g -i --insecure -X GET http://xxx:5000/v3/users/ad262a99ab9d49a18e1d0e63c2912d2a/projects -H "User-Agent: python-keystoneclient" -H "Accept: application/json" -H "X-Auth-Token: {SHA1}63f5fc9dccaf2eac804035be9200df1007dfbc67" | 14:16 |
| cshen | [Wed Oct 11 13:48:32.614877 2017] [wsgi:error] [pid 251:tid 140246174328576] DEBUG:keystoneauth.session:RESP: [200] Date: Wed, 11 Oct 2017 13:48:32 GMT Server: Apache Vary: X-Auth-Token x-openstack-request-id: req-b1298a21-523a-484b-84c1-139557eed92e Content-Length: 4918 Content-Type: application/json | 14:16 |
| *** rmart04 has left #openstack-horizon | 14:21 | |
| *** zhurong has quit IRC | 14:29 | |
| *** ramineni_ has joined #openstack-horizon | 14:32 | |
| *** lajoskatona has quit IRC | 14:32 | |
| *** udesale has quit IRC | 14:34 | |
| *** felipemonteiro has joined #openstack-horizon | 14:38 | |
| *** mvk has quit IRC | 14:39 | |
| *** jamesbenson has joined #openstack-horizon | 14:40 | |
| *** felipemonteiro_ has joined #openstack-horizon | 14:40 | |
| Sudheer | Hello all i have question regarding support for displaying gnoochi/adoh information in horizon | 14:43 |
| Sudheer | I came across a bule print which basically address this issue: https://blueprints.launchpad.net/horizon/+spec/horizon-gnocchi-graphs | 14:43 |
| Sudheer | but i have n't found any useful links from the same to check the progress of UI support | 14:43 |
| *** spectr has quit IRC | 14:43 | |
| Sudheer | Can some one point me in the right direction whether there is new panel/plugin creation is in progress for displaying Gnoochi/Adoh services data | 14:43 |
| *** gary-smith__ has joined #openstack-horizon | 14:44 | |
| *** felipemonteiro has quit IRC | 14:44 | |
| *** openstackgerrit has joined #openstack-horizon | 14:47 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/horizon master: Switch Cinder-related tests to mock https://review.openstack.org/510118 | 14:47 |
| *** spectr has joined #openstack-horizon | 14:58 | |
| david-lyle | Sudheer, try #openstack-telemetry it should be owned by the telemetry team | 15:01 |
| robcresswell | Sudheer: There's no plugin that I'm aware of to display telemetry data within Horizon | 15:02 |
| david-lyle | There was work on Aodh alarms, and I vaguely remember a mailing post about a gnocchi panel, but I didn't track it | 15:03 |
| *** chlong has quit IRC | 15:06 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/horizon master: Switch Cinder-related tests to mock https://review.openstack.org/510118 | 15:30 |
| *** gyee has joined #openstack-horizon | 15:33 | |
| *** jamesbenson has quit IRC | 15:33 | |
| *** jamesbenson has joined #openstack-horizon | 15:34 | |
| *** jamesbenson has quit IRC | 15:37 | |
| *** chlong has joined #openstack-horizon | 15:38 | |
| *** jamesbenson has joined #openstack-horizon | 15:43 | |
| *** jamesbenson has quit IRC | 15:44 | |
| *** jamesbenson has joined #openstack-horizon | 15:45 | |
| *** chlong has quit IRC | 15:51 | |
| *** gary-smith__ has quit IRC | 15:53 | |
| *** ratailor has joined #openstack-horizon | 15:54 | |
| *** pcaruana has quit IRC | 15:59 | |
| *** rcernin has quit IRC | 15:59 | |
| *** gary-smith__ has joined #openstack-horizon | 16:01 | |
| *** ratailor has quit IRC | 16:03 | |
| *** jpich has quit IRC | 16:08 | |
| *** wolverineav has quit IRC | 16:09 | |
| *** itlinux has joined #openstack-horizon | 16:11 | |
| *** chlong has joined #openstack-horizon | 16:11 | |
| *** wolverineav has joined #openstack-horizon | 16:13 | |
| *** jamesbenson has quit IRC | 16:21 | |
| *** jamesbenson has joined #openstack-horizon | 16:23 | |
| *** gary-smith__ has quit IRC | 16:27 | |
| *** e0ne has quit IRC | 16:38 | |
| *** felipemonteiro_ has quit IRC | 16:40 | |
| *** ramineni_ has quit IRC | 16:41 | |
| *** chlong has quit IRC | 16:48 | |
| *** felipemonteiro has joined #openstack-horizon | 17:00 | |
| *** felipemonteiro_ has joined #openstack-horizon | 17:01 | |
| *** felipemonteiro has quit IRC | 17:06 | |
| *** nikhil has joined #openstack-horizon | 17:14 | |
| *** tesseract has quit IRC | 17:15 | |
| *** tosky has quit IRC | 17:34 | |
| *** itlinux has quit IRC | 17:40 | |
| *** e0ne has joined #openstack-horizon | 17:41 | |
| *** cshen_ has joined #openstack-horizon | 17:51 | |
| *** tosky has joined #openstack-horizon | 17:58 | |
| *** e0ne has quit IRC | 18:06 | |
| *** flwang1 has joined #openstack-horizon | 18:16 | |
| flwang1 | david-lyle: ying_zuo: robcresswell: two topics I'd like to discuss if you're around | 18:17 |
| flwang1 | 1. https://bugs.launchpad.net/horizon/+bug/1722704 bug about image name filter | 18:18 |
| openstack | Launchpad bug 1722704 in OpenStack Dashboard (Horizon) "Image name filter on instance panel doesn't work" [Undecided,New] - Assigned to Feilong Wang (flwang) | 18:18 |
| flwang1 | 2. https://bugs.launchpad.net/horizon/+bug/1722701 adding a white list for nova extension | 18:19 |
| openstack | Launchpad bug 1722701 in OpenStack Dashboard (Horizon) "[Performance] Add a white list for nova extension to avoid API calls" [Undecided,In progress] - Assigned to Feilong Wang (flwang) | 18:19 |
| ying_zuo | flwang1: I am here | 18:23 |
| flwang1 | ying_zuo: sorry i'm making noisy here again :) | 18:23 |
| flwang1 | ying_zuo: yesterday, i had a look at the code about image name filter and I think there are some design faults, pls see bug https://bugs.launchpad.net/horizon/+bug/1722704 | 18:24 |
| openstack | Launchpad bug 1722704 in OpenStack Dashboard (Horizon) "Image name filter on instance panel doesn't work" [Undecided,New] - Assigned to Feilong Wang (flwang) | 18:24 |
| ying_zuo | no worries :) | 18:24 |
| robcresswell | flwang1: isn't there already a Nova extensions blacklist setting? | 18:25 |
| flwang1 | robcresswell: yep, hence why i want to add a white list to reduce unnecessary calls | 18:26 |
| flwang1 | in our env, those function calls take 0.2-0.3 second | 18:26 |
| flwang1 | and with a white list in settings.py, the number will be 0.01 second | 18:27 |
| *** gyee has quit IRC | 18:27 | |
| *** gyee has joined #openstack-horizon | 18:29 | |
| *** jlabarre has quit IRC | 18:33 | |
| ying_zuo | flwang1: where is the blacklist being set? | 18:34 |
| *** gyee has quit IRC | 18:37 | |
| *** gyee has joined #openstack-horizon | 18:38 | |
| flwang1 | ying_zuo: wait a sec | 18:38 |
| flwang1 | https://github.com/openstack/horizon/blob/master/openstack_dashboard/api/nova.py#L989 | 18:39 |
| flwang1 | it's set by ops in settings.py | 18:39 |
| *** gyee has quit IRC | 18:39 | |
| flwang1 | ying_zuo: robcresswell: so for the image filter issue, i can't see there is good way to fix them, they're limitations | 18:42 |
| ying_zuo | hmm. should probably have a default value for it set on either local_settings or settings | 18:42 |
| flwang1 | ying_zuo: you mean the extension list? yep, it would be nice | 18:42 |
| flwang1 | and the list extension API will be called when loading horizon first time | 18:43 |
| *** gyee has joined #openstack-horizon | 18:45 | |
| *** cshen_ has quit IRC | 18:47 | |
| ying_zuo | having a whitelist is not a very robust solution | 18:49 |
| flwang1 | ying_zuo: i understand. so i just want to get some feedback from you guys, otherwise, we may keep it in our private repo | 18:49 |
| ying_zuo | it can be a long list | 18:50 |
| flwang1 | ying_zuo: no, not really | 18:50 |
| flwang1 | you can only put the names you cared | 18:50 |
| flwang1 | for instance list panel, only two names | 18:50 |
| flwang1 | Shelve, AdminActions | 18:51 |
| ying_zuo | I think there are more | 18:52 |
| ying_zuo | oh, just for extensions | 18:53 |
| ying_zuo | if we already have a blacklist maybe it's okay to have a whitelist too | 18:55 |
| flwang1 | ying_zuo: it would be nice if you can leave your comments on the bug/patch, so that we can work out a final decision | 18:57 |
| flwang1 | and I do need your comments on the image name filter bug | 18:57 |
| ying_zuo | are you going to be around in an hour? | 18:58 |
| flwang1 | yep, i'm going to office now and will be online after 1.5 hours | 18:59 |
| ying_zuo | okay. I just added the image name filter bug to the list for the weekly meeting https://etherpad.openstack.org/p/horizon-bug-review | 19:00 |
| flwang1 | don't think the traffic at Wellington is bad, i'm going to grab breakfast and then send my kids to kindergarten ;) | 19:00 |
| flwang1 | ying_zuo: cool, thanks a lot | 19:00 |
| *** cshen_ has joined #openstack-horizon | 19:09 | |
| *** mvk has joined #openstack-horizon | 19:18 | |
| *** flwang1 has quit IRC | 19:22 | |
| ying_zuo | Weekly Horizon meeting will start in half hour. #openstack-meeting-3 | 19:30 |
| *** gary-smith has joined #openstack-horizon | 19:30 | |
| *** gyee has quit IRC | 19:31 | |
| *** makowals_ has joined #openstack-horizon | 19:39 | |
| *** e0ne has joined #openstack-horizon | 19:40 | |
| *** gyee has joined #openstack-horizon | 19:51 | |
| *** ppiela_ has quit IRC | 19:52 | |
| robcresswell | flwang, ying_zuo: why a blacklist *and* a whitelist? | 19:52 |
| robcresswell | That's just more logic to muddle settings with : | 19:53 |
| robcresswell | :/ * | 19:53 |
| robcresswell | flwang: If there's already a blacklist... can't you just put all the extensions in there, and you're done? | 19:53 |
| *** gyee has quit IRC | 19:57 | |
| *** jamesbenson has quit IRC | 19:58 | |
| *** harlowja has quit IRC | 20:00 | |
| *** itlinux has joined #openstack-horizon | 20:03 | |
| *** cshen_ has quit IRC | 20:03 | |
| *** cshen_ has joined #openstack-horizon | 20:03 | |
| *** jamesbenson has joined #openstack-horizon | 20:07 | |
| *** jamesbenson has quit IRC | 20:09 | |
| *** jamesbenson has joined #openstack-horizon | 20:13 | |
| *** tsmith has joined #openstack-horizon | 20:16 | |
| flwang | robcresswell: that's a good point, though I don't think the blacklist should be used for my purpose | 20:22 |
| robcresswell | flwang: Could you elaborate? | 20:23 |
| *** flwang1 has joined #openstack-horizon | 20:23 | |
| flwang | my purpose is checking if the extension is supported or not, to save the calls to nova api and even the functions in api/nova.py | 20:24 |
| flwang | so do you think i should just check if the extention name is in the blacklist, if it's not, then I think it's supported? | 20:25 |
| *** cshen_ has quit IRC | 20:25 | |
| flwang | robcresswell: currently, Horizon will do things like https://github.com/openstack/horizon/blob/master/openstack_dashboard/dashboards/project/instances/tables.py#L280 | 20:27 |
| flwang | then call https://github.com/openstack/horizon/blob/master/openstack_dashboard/api/nova.py#L999 | 20:28 |
| flwang | as I mentioned in the bug, now there are 5 actions are doing the check like this, and there are 20 instances per pages | 20:28 |
| robcresswell | Right, but if you just blacklist that extension... | 20:29 |
| flwang | so 100 calls, with cache, it still take about 0.2 seconds | 20:29 |
| *** marlinc has quit IRC | 20:29 | |
| flwang | robcresswell: in otherwords, not in blacklist doesn't mean the nova server supports that | 20:30 |
| flwang | but if it's in a whitelist, that means the extension name is added explicitly by ops, because ops know what they're doing | 20:31 |
| flwang | robcresswell: does that make any sense? | 20:31 |
| *** gary-smith__ has joined #openstack-horizon | 20:31 | |
| robcresswell | flwang: Not really. If 0.2 seconds is truly unbearable, just put every extension in the blacklist and explicitly remove the ones your support | 20:33 |
| robcresswell | thats identical to a whitelist anyway | 20:33 |
| robcresswell | Except without Horizon having to carry logic for both | 20:34 |
| *** gary-smith has quit IRC | 20:34 | |
| *** priteau has quit IRC | 20:38 | |
| *** gary-smith has joined #openstack-horizon | 20:50 | |
| *** gary-smith__ has quit IRC | 20:51 | |
| *** gary-smith has quit IRC | 20:55 | |
| *** lblanchard1 has quit IRC | 20:59 | |
| *** makowals_ has quit IRC | 21:00 | |
| robcresswell | flwang: I'd prefer if we kept the image name there. Just make the ID a popover. problem solved. | 21:01 |
| robcresswell | Then remove the image name filter, because its a magic Horizon invention (we have got to stop doing those) | 21:01 |
| flwang | robcresswell: but we also need to remove the image name filter? | 21:01 |
| robcresswell | Yup | 21:01 |
| *** catintheroof has quit IRC | 21:02 | |
| e0ne | bugs discussions are very useful and important, but we have no time for other topics:( | 21:03 |
| ying_zuo | oh sorry e0ne | 21:04 |
| e0ne | robcresswell, flwang: glance v2 api supports image filtering by name. can we use it? | 21:04 |
| david-lyle | robcresswell, the API supports filtering by name | 21:04 |
| robcresswell | david-lyle: The Nova API? | 21:04 |
| ying_zuo | I see you added a topic on the meeting note | 21:04 |
| david-lyle | we just have to pass the query parameter | 21:04 |
| e0ne | ying_zuo: np, it's not something that blocks me | 21:04 |
| david-lyle | no image | 21:04 |
| flwang | e0ne: as i mentioned, it still returns a list | 21:04 |
| flwang | so that said, there are two images with same name | 21:04 |
| e0ne | flwang: yes, we have to map the correct image to instance | 21:04 |
| flwang | you got two images and which one is the one you care about? | 21:05 |
| david-lyle | flwang, yes I don't think names are exclusive | 21:05 |
| robcresswell | I guess the other alternative is just to assume people don't give their images the same name. | 21:05 |
| e0ne | flwang: I'm more concerned about performance as pointed above | 21:05 |
| david-lyle | flwang, it doesn't matter, the API will return all the match | 21:05 |
| david-lyle | it's up to the user beyond that | 21:05 |
| flwang | e0ne: yep, the performance is the one i really care about | 21:05 |
| e0ne | flwang: +1 on performance | 21:06 |
| flwang | and you can see I have proposed several patches about the instance panel performance | 21:06 |
| robcresswell | david-lyle: I think the issue is that its not obvious which image you'll actually be filtering by if you just take the first in a list | 21:06 |
| robcresswell | Unless you go through each and concat them all | 21:06 |
| david-lyle | robcresswell, I'm saying it doesn't matter | 21:06 |
| david-lyle | all matches based on name are valid | 21:06 |
| e0ne | flwang: thanks for them. I appreciate your help | 21:06 |
| david-lyle | the user can filter beyond that | 21:06 |
| flwang | david-lyle: no | 21:06 |
| robcresswell | david-lyle: How? | 21:07 |
| flwang | user can't do that | 21:07 |
| robcresswell | It's a single keyword filter | 21:07 |
| flwang | because horizon always pick the first one | 21:07 |
| david-lyle | they have a brain | 21:07 |
| david-lyle | :) | 21:07 |
| david-lyle | but artificially filtering for them is wrong | 21:07 |
| flwang | for that case, why don't let them filter by id | 21:07 |
| *** harlowja has joined #openstack-horizon | 21:07 | |
| ying_zuo | e0ne: I made a note to talk about the mox to mock migration before reviewing the bug reports in the next meeting | 21:08 |
| david-lyle | that is also allowed | 21:08 |
| david-lyle | image_id and image_name are acceptable filters | 21:08 |
| robcresswell | david-lyle: So, get every image that matches the name, then get the instances that match each ID and return that entire list? Because atm, we just return whatever is first in the list glance returns | 21:08 |
| david-lyle | The API call is different | 21:08 |
| e0ne | ying_zuo: thanks | 21:08 |
| robcresswell | Which is problematic, because the user has no control over that | 21:08 |
| *** jamesbenson has quit IRC | 21:09 | |
| david-lyle | I'm not claiming the current implementation is correct, but it could be made correct by making the correct API query | 21:09 |
| robcresswell | Sure | 21:09 |
| e0ne | ying_zuo: probably, we should start meetings with proposed topics if any and then continue bugs discussion | 21:09 |
| ying_zuo | e0ne: yes | 21:10 |
| e0ne | ying_zuo: but it's up to you and rest of the community to decide how to deal with meetings :) | 21:10 |
| david-lyle | gary-smith_, I triaged the bug you mentioned me on from the meeting, thanks, it should be a rather simple fix | 21:11 |
| flwang | robcresswell: david-lyle: so what's the conclusion? the main purpose I'd like to do is improve the performance actually ;) | 21:11 |
| e0ne | flwang, robcresswell, david-lyle: I have to go, it's 12am already in my TZ. I'll read the conversation tomorrow and will be ready to help if needed | 21:11 |
| robcresswell | e0ne: \o | 21:11 |
| flwang | and fix it bug by the way if there is | 21:11 |
| flwang | e0ne: thanks, have a good night | 21:12 |
| e0ne | flwang: IMO, we should support only that kind of filters, which are implemented by APIs | 21:13 |
| e0ne | flwang: thanks | 21:13 |
| david-lyle | flwang, I'm for the performance improvement of providing setting to not get image name. | 21:13 |
| david-lyle | I think the filter is debatable, it's doable, but is it an overreach for the instances panel to be filtering on a different API result | 21:13 |
| *** jtomasek has quit IRC | 21:14 | |
| *** jtriley has quit IRC | 21:15 | |
| flwang | david-lyle: ok, I will focus on the image name configurable first | 21:15 |
| david-lyle | flwang, you had a bug for the filter problem, what was the number? | 21:18 |
| flwang | https://bugs.launchpad.net/horizon/+bug/1722704 | 21:18 |
| openstack | Launchpad bug 1722704 in OpenStack Dashboard (Horizon) "Image name filter on instance panel doesn't work" [Undecided,New] - Assigned to Feilong Wang (flwang) | 21:18 |
| david-lyle | flwang, do you want me to take that? | 21:19 |
| david-lyle | I imagine the problem is larger than just that | 21:19 |
| flwang | david-lyle: feel free to take it | 21:20 |
| flwang | david-lyle: but i have a condition | 21:20 |
| flwang | david-lyle: please review my performance patches ;) | 21:20 |
| david-lyle | I reviewed the image setting one, I need to look at it again | 21:21 |
| flwang | david-lyle: for example, https://review.openstack.org/#/c/510718/ | 21:21 |
| david-lyle | that was just a quick tour | 21:21 |
| flwang | and https://review.openstack.org/511095 | 21:21 |
| flwang | instance panel is calling too many api calls | 21:21 |
| david-lyle | the floating IP sync was added because of operator demand :) | 21:22 |
| david-lyle | now we have to have an operator grudge match | 21:22 |
| david-lyle | the delay on fIP sync can be quite large | 21:23 |
| flwang | david-lyle: ok, why don't just get the FIP instead of everything? | 21:23 |
| flwang | currently, those neutron API calls take 1.3-1.5 seconds in our env | 21:24 |
| david-lyle | flwang, I agree it makes too many API calls | 21:24 |
| david-lyle | unfortunately it's the panel that everyone want to tie everything together on | 21:24 |
| david-lyle | and most of the APIs are not intended to be user consumable alone | 21:25 |
| david-lyle | so we have to balance usability and performance | 21:25 |
| david-lyle | I would prefer getting an async load of the data going rather than removing the usability elements | 21:26 |
| david-lyle | because values like UUIDs are fairly useless alone | 21:27 |
| david-lyle | hence image name lookup | 21:27 |
| david-lyle | project name lookup | 21:27 |
| david-lyle | flwang, I completely understand your frustration | 21:28 |
| david-lyle | this is why we tried moving toward angular and async loading, we just have been reeeaaaalllly slow at it | 21:29 |
| flwang | david-lyle: thank you, and i totally understand the situation | 21:29 |
| david-lyle | it won't fix everything, but perceived page load speed will be much better | 21:29 |
| flwang | yep, TBH, currently, the page is taking 9-10 seconds if there are 20 instances on the page | 21:30 |
| flwang | that's very bad experience | 21:30 |
| david-lyle | flwang, indeed | 21:30 |
| *** e0ne has quit IRC | 21:31 | |
| david-lyle | maybe your image name lookup setting should be expanded to instance panel performance in general | 21:31 |
| *** wolverineav has quit IRC | 21:31 | |
| flwang | david-lyle: that's what i was going to do | 21:32 |
| david-lyle | I fear the code will become a mess of if/else | 21:32 |
| flwang | before we totally angulaize the page | 21:32 |
| david-lyle | but maybe that's what we have to do | 21:32 |
| flwang | could we have a dict in settings.py to turn off something, like the image name, to speed up the page from different angels? | 21:32 |
| david-lyle | sure | 21:33 |
| flwang | would you mind giving a good name? as I'm not a English native speaker | 21:33 |
| robcresswell | I wonder how much work this is, over just adopting the angular instances panel | 21:34 |
| robcresswell | Richard worked on that for a few weeks | 21:34 |
| flwang | so that we have xxxx = {"ENABLE_INSTANCE_IMAGENAME": True, "ENABLE_NEUTEON_REALTIME_FIP": True} | 21:35 |
| flwang | robcresswell: could we have it in Queens? | 21:35 |
| david-lyle | maybe something around "performance_flags" | 21:35 |
| david-lyle | and change ENABLE to DISABLE | 21:35 |
| robcresswell | flwang: Probably, if someone wants to adopt it | 21:35 |
| robcresswell | Then it would be much easier to load the panel and async all the other data | 21:36 |
| david-lyle | it just needs an owner | 21:36 |
| david-lyle | robcresswell, I agree | 21:36 |
| flwang | robcresswell: but ying_zuo told me now there is nobody working on the angularize stuff for instance panel | 21:36 |
| flwang | it sounds like we have everything but a developer | 21:36 |
| david-lyle | flwang, robcresswell is saying if you're going to put this much time into the current panel, you should just finish the angular panel | 21:36 |
| david-lyle | I think | 21:36 |
| robcresswell | david-lyle: +1 | 21:37 |
| flwang | david-lyle: ok, i see. but i don't have the skill of angularJS | 21:37 |
| david-lyle | no one does, we'll never know | 21:37 |
| flwang | I mostly work on the backend, you know | 21:37 |
| david-lyle | :) | 21:37 |
| robcresswell | My point is that all this time making the Python panel "less bad" could just be invested into making a good panel :p | 21:37 |
| flwang | zaqar, nova, ceilometer, balablabla | 21:37 |
| flwang | but not Js and css | 21:37 |
| robcresswell | flwang: *shrug* I dont really know anything, but I manage to contribute :p | 21:38 |
| david-lyle | Js is just like python with some exceptions | 21:38 |
| flwang | david-lyle: and another thing is we're using stable pike | 21:38 |
| flwang | and I don't have much time to get this done (in angularJS) | 21:39 |
| flwang | I'm sorry | 21:39 |
| david-lyle | flwang, well your fixes likely won't qualify for a backport | 21:39 |
| david-lyle | but a new panel can be turned on fairly easily | 21:39 |
| david-lyle | I htink most of the REST API support is already there | 21:39 |
| * david-lyle would have to check on the last part | 21:39 | |
| david-lyle | flwang, fair enough | 21:39 |
| david-lyle | but what we're proposing for the current panel is going to make a mess of it | 21:40 |
| david-lyle | and it's going to be very hard to maintain | 21:40 |
| david-lyle | IMO | 21:41 |
| flwang | david-lyle: ok, but at least that's the thing i'd like to figure out, if community is not happy with the work I'm proposing, we may have to keep them in private repo | 21:41 |
| robcresswell | Hence our reluctance on some of the changes :) | 21:41 |
| robcresswell | Some sound okay. Happy to keep evaluating on a case-by-case basis. | 21:41 |
| flwang | david-lyle: yep, i totally understand | 21:41 |
| robcresswell | I'm gonna head off; will catch up with rest of conversation tomorrow | 21:42 |
| robcresswell | night all | 21:42 |
| flwang | robcresswell: good night, thanks | 21:42 |
| flwang | david-lyle: so basically the image name configurable should be ok? | 21:42 |
| flwang | and the neutron one is not? | 21:42 |
| flwang | david-lyle: and do you think this one is backportable https://review.openstack.org/511072 ? | 21:43 |
| david-lyle | flwang, that one's tough, I think amotoki's interpretation is accurate | 21:48 |
| flwang | david-lyle: ok, fair enough | 21:50 |
| *** ppiela has joined #openstack-horizon | 21:51 | |
| *** gary-smith has joined #openstack-horizon | 22:27 | |
| *** wolverineav has joined #openstack-horizon | 22:27 | |
| *** felipemonteiro_ has quit IRC | 22:34 | |
| *** felipemonteiro_ has joined #openstack-horizon | 22:34 | |
| *** marlinc has joined #openstack-horizon | 22:35 | |
| *** priteau has joined #openstack-horizon | 22:39 | |
| *** priteau has quit IRC | 22:44 | |
| *** wolverineav has quit IRC | 22:54 | |
| *** wolverineav has joined #openstack-horizon | 22:55 | |
| *** wolverineav has quit IRC | 22:59 | |
| *** priteau has joined #openstack-horizon | 23:01 | |
| *** priteau has quit IRC | 23:06 | |
| *** felipemonteiro_ has quit IRC | 23:12 | |
| *** yamamoto has joined #openstack-horizon | 23:28 | |
| *** jtriley has joined #openstack-horizon | 23:38 | |
| *** tosky has quit IRC | 23:43 | |
| *** ppiela has quit IRC | 23:54 | |
| *** itlinux has quit IRC | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!