| *** Qiming has quit IRC | 00:00 | |
| flwang | https://github.com/openstack/horizon/blob/master/horizon/tables/base.py#L472 | 00:00 |
|---|---|---|
| flwang | https://github.com/openstack/horizon/blob/master/horizon/tables/base.py#L663 | 00:01 |
| * vkmc clicks | 00:07 | |
| vkmc | I see | 00:10 |
| vkmc | those updates the status of things in the tables | 00:10 |
| vkmc | which will be the perf issue we will focus for the poc | 00:11 |
| vkmc | (I was looking for the balloon notifications) | 00:11 |
| *** davideagnello has quit IRC | 00:12 | |
| *** davideagnello has joined #openstack-zaqar | 00:12 | |
| flwang | vkmc: the balloon notification will be easier than this one | 00:13 |
| vkmc | flwang, so maybe we want to do those | 00:14 |
| flwang | for example, we can easily get the notifications user maybe interested in, like instance delete, image delete, etc | 00:14 |
| vkmc | perhaps the perf impact is higher on the status change for the resources... given that horizon has to do frequent ajax polls to have things updated | 00:15 |
| flwang | vkmc: so how about let's focus on the workflow and figure out how to get the detailed status meanwhile? | 00:15 |
| vkmc | but for the balloons notifications the improvement is visual | 00:15 |
| flwang | vkmc: yep, i believe so | 00:15 |
| vkmc | which may be slightly better for a demo | 00:15 |
| vkmc | sounds good to me | 00:16 |
| flwang | but i don't want the status collecting to be a blocker for us, thoughts? | 00:16 |
| vkmc | why would it be a blocker? | 00:16 |
| vkmc | at least at this stage | 00:16 |
| flwang | i mean if we only focus on the status update case, then the problem is now it's difficult to get the status from different services | 00:17 |
| vkmc | well yes | 00:17 |
| flwang | so we can work on the flow/process and then/meanwhile figure out the status collecting | 00:18 |
| vkmc | sure | 00:18 |
| vkmc | https://github.com/openstack/horizon/blob/master/horizon/middleware.py#L189 | 00:20 |
| flwang | so the process should be: service(glance, nova, cinder, etc ) send notifications to ceilometer -> ceilometer republish the notifications/msg to zaqar -> zaqar notify horizon | 00:20 |
| vkmc | there is a TODO there | 00:20 |
| vkmc | left by Gabriel lol... it has been a while | 00:20 |
| vkmc | (PTL in 2012) | 00:20 |
| vkmc | something I'm missing there | 00:21 |
| flwang | haha, cool | 00:21 |
| vkmc | service sends notification to ceilometer | 00:21 |
| vkmc | the kind of information ceilometer can collect is metering information | 00:21 |
| vkmc | IIRC | 00:22 |
| flwang | if you would like a service sending notifications to zaqar directly, it works either, but it needs more changes | 00:22 |
| flwang | brb, grab lunch | 00:22 |
| flwang | will you be around after 40 mins? | 00:22 |
| vkmc | yes | 00:23 |
| vkmc | :) | 00:23 |
| flwang | cool, catch you later ;) | 00:23 |
| vkmc | sure, enjoy your lunch! | 00:24 |
| vkmc | another con I see about using Ceilo | 00:30 |
| vkmc | is that Ceilo does polling itself | 00:30 |
| vkmc | which means we are just moving the problem somewhere else | 00:31 |
| vkmc | (correct me if I am wrong) | 00:31 |
| *** Qiming has joined #openstack-zaqar | 00:54 | |
| *** achanda has quit IRC | 00:55 | |
| *** davideagnello has quit IRC | 00:58 | |
| *** flwang1 has quit IRC | 00:58 | |
| flwang | vkmc: back | 01:02 |
| flwang | not really, ceilometer support both polling and notification | 01:03 |
| flwang | depends on the metrics | 01:03 |
| *** flwang1 has joined #openstack-zaqar | 01:04 | |
| vkmc | cool | 01:08 |
| flwang | but TBH, there are several parts I haven't figure out clearly | 01:10 |
| flwang | 1. Should we hacking services directly or just consume ceilometer? | 01:10 |
| flwang | 2. Who should be responsible for the notification's filter? in other words, what kind of notification should be sent to the queue | 01:11 |
| flwang | 3. How horizon consume the messages? polling? NO. notifications? it needs an endpoint in horizon i think | 01:12 |
| vkmc | 1. seems fuzzy to me... consuming Ceilometer is easier but more constraining | 01:16 |
| vkmc | but consuming from the services directly would be quite painful to maintain | 01:17 |
| vkmc | 2. haven't thought about that, I assumed the producer would feed something like service name + message and that's all | 01:17 |
| vkmc | 3. I'd say we use notifications, but certainly we will need a websocket endpoint in Horizon | 01:18 |
| flwang | vkmc: #1 if we go for ceilometer, we need a publisher at here https://github.com/openstack/ceilometer/tree/master/ceilometer/publisher | 01:21 |
| flwang | if we go without ceilometer, we need to change the notifier code for different service like this https://github.com/openstack/glance/blob/master/glance/notifier.py | 01:23 |
| flwang | personally, i prefer to use ceilometer, since we will only add a new driver instead of changing/breaking any existing logic | 01:26 |
| *** flwang1 has quit IRC | 01:31 | |
| *** flwang1 has joined #openstack-zaqar | 01:32 | |
| vkmc | sounds good to me | 01:43 |
| flwang | vkmc: i can help for the publisher if you need a hand | 01:47 |
| vkmc | flwang, would it make sense we fork Ceilometer/Horizon and start adding the code there? | 01:48 |
| vkmc | I'll try to start with that tomorrow | 01:48 |
| flwang | vkmc: works for me | 01:48 |
| vkmc | cool | 01:48 |
| vkmc | thanks :d | 01:48 |
| flwang | i have experience with horizon either | 01:48 |
| vkmc | awesome! | 01:49 |
| flwang | so just let me know if you need any help | 01:49 |
| flwang | otherwise, i will take a look at the ceilometer side | 01:49 |
| vkmc | will do | 01:50 |
| vkmc | I'll check out the zaqar publisher creation for the Ceilometer side | 01:51 |
| flwang | so do you need any help for the horizon side or cpallares will do that? | 01:52 |
| flwang | i assume we don't need any change in zaqar | 01:53 |
| vkmc | I'd assume not | 01:57 |
| *** csoukup has joined #openstack-zaqar | 01:57 | |
| vkmc | yeah please, if you can check in the Horizon side | 01:57 |
| vkmc | it would be very helpful :) | 01:57 |
| vkmc | precisely, were we should put the websocket interface | 01:57 |
| vkmc | cpallares will probably join us later | 01:58 |
| *** c_soukup has joined #openstack-zaqar | 02:01 | |
| *** csoukup has quit IRC | 02:04 | |
| flwang | vkmc: because we don't support long poll, and currently our notification only supports webhook | 02:10 |
| flwang | so seems we have to create a new endpoint in horizon, though i'm not sure if it's too 'hacking' | 02:10 |
| *** csoukup_ has joined #openstack-zaqar | 02:19 | |
| *** c_soukup has quit IRC | 02:22 | |
| vkmc | flwang, wouldn't therve's patch do the trick if we can move it forward? | 02:39 |
| flwang | vkmc: it just supports the CRUD of subscriptions via websocket | 02:45 |
| flwang | vkmc: we only support the webhook and email as the task type | 02:46 |
| vkmc | flwang, oh yes, that's right | 02:46 |
| flwang | TBH, that's my main concern from the horizon PoV | 02:46 |
| flwang | from the horizon perspective, i would like to see a lightweight way to resolve the problem instead of introducing more complex | 02:47 |
| flwang | so we should keep that in mind to avoid scaring horizon team :) | 02:48 |
| vkmc | agree | 02:50 |
| vkmc | mhmh | 02:50 |
| *** davideagnello has joined #openstack-zaqar | 02:59 | |
| *** davideagnello has quit IRC | 03:04 | |
| *** diga has joined #openstack-zaqar | 03:28 | |
| *** stanchan has joined #openstack-zaqar | 03:58 | |
| *** davideagnello has joined #openstack-zaqar | 04:12 | |
| *** achanda has joined #openstack-zaqar | 04:34 | |
| *** flwang1 has quit IRC | 05:03 | |
| *** davideagnello has quit IRC | 05:05 | |
| *** csoukup_ has quit IRC | 05:36 | |
| therve | flwang, My patch supports subscription over websocket transport | 05:39 |
| therve | Not just CRUD operations | 05:39 |
| flaper87 | o/ | 07:00 |
| *** davideagnello has joined #openstack-zaqar | 07:06 | |
| *** davideagnello has quit IRC | 07:11 | |
| openstackgerrit | Flavio Percoco proposed openstack/zaqar: Allow running zaqarclient for gate tests https://review.openstack.org/221068 | 07:19 |
| *** dynarro has joined #openstack-zaqar | 07:26 | |
| openstackgerrit | Claudiu Belu proposed openstack/zaqar: Remove unused zaqar.openstack.common modules https://review.openstack.org/220250 | 07:34 |
| *** achanda has quit IRC | 07:38 | |
| *** achanda has joined #openstack-zaqar | 07:39 | |
| openstackgerrit | Merged openstack/zaqar: Allow running zaqarclient for gate tests https://review.openstack.org/221068 | 08:04 |
| *** achanda has quit IRC | 08:20 | |
| *** boris-42 has quit IRC | 08:53 | |
| *** sebasmagri has quit IRC | 08:53 | |
| *** penchal has joined #openstack-zaqar | 09:03 | |
| *** davideagnello has joined #openstack-zaqar | 09:07 | |
| *** sebasmagri has joined #openstack-zaqar | 09:11 | |
| *** davideagnello has quit IRC | 09:11 | |
| zigo | May I have an answer about falcon 3.0? Does Zaqar support it? | 09:19 |
| zigo | flaper87: therve: ^ | 09:19 |
| zigo | The Mailman maintainers in Debian need version 3.0... | 09:19 |
| therve | zigo, https://bugs.launchpad.net/zaqar/+bug/1495402 | 09:20 |
| openstack | Launchpad bug 1495402 in zaqar "Update falcon requirement" [Undecided,New] | 09:20 |
| zigo | therve: Do you have a working patch available somewhere? | 09:23 |
| zigo | Zaqar seems to be the only user of Falcon these days. | 09:23 |
| zigo | Everyone else moved to another framework. | 09:23 |
| flaper87 | zigo: in openstack we've always been the only users of falcon | 09:24 |
| *** boris-42 has joined #openstack-zaqar | 09:24 | |
| *** dynarro has quit IRC | 09:27 | |
| zigo | Barbican seems to also use it. | 09:29 |
| flaper87 | oh really? | 09:30 |
| * flaper87 had no idea | 09:30 | |
| *** diga has quit IRC | 09:43 | |
| therve | zigo, The patch is mentioned in the bug: https://review.openstack.org/225269 | 09:51 |
| jasondotstar | zigo: hi | 10:10 |
| zigo | jasondotstar: Hi ! | 10:10 |
| jasondotstar | how art thou? :-) | 10:11 |
| jasondotstar | hope you had a nice weekend | 10:11 |
| jasondotstar | hey qq | 10:11 |
| jasondotstar | not sure if you use an IRC logger but, flwang and I were talking about the puppet module | 10:12 |
| jasondotstar | need to catch up w you on the packaging | 10:12 |
| jasondotstar | how's that looking these days? | 10:12 |
| jasondotstar | zigo: ^ | 10:15 |
| zigo | jasondotstar: Well, I did all the packaging, and uploaded to Debian Experimental, so it's waiting in the FTP master NEW queue for approval. | 10:16 |
| zigo | jasondotstar: But I still have 2 pending issues. | 10:16 |
| zigo | jasondotstar: 1/ If I do export ZAQAR_TEST_MONGODB=yes, then there's unit tests failure all over the place 2/ I need zaqar to support Falcon 3.0 | 10:16 |
| zigo | For 2/, it's because the Mailman package maintainer would like to package Mailman 3, which uses Falcon 3.0.0. | 10:17 |
| zigo | Both of these issues are *very* important to fix. | 10:17 |
| jasondotstar | ack | 10:17 |
| zigo | As for puppet, I know nothing about it, but you could talk with Emillien about it, I believe. | 10:17 |
| zigo | jasondotstar: If you want to try the packaging, you can git clone git://anonscm.debian.org/openstack/zaqar.git | 10:18 |
| jasondotstar | yeah, EmilienM is aware of what we're working on. | 10:18 |
| zigo | Then try to build taking dependencies from Debian Experimental. | 10:18 |
| zigo | It's to be tried in Sid. | 10:18 |
| zigo | (all Debian packaging work is always done in Debian Sid). | 10:18 |
| zigo | So, I'm basically almost done, pending these 2 ... | 10:18 |
| jasondotstar | thing is... we have to detect linux distro so we can determine if we need to install debs or rpms | 10:19 |
| jasondotstar | the rdo guys are helping with the rpm stuff | 10:19 |
| jasondotstar | so i just needed to figure out the deb pkgs too | 10:19 |
| jasondotstar | so...this sounds great | 10:20 |
| jasondotstar | the debs end up in Debian Experimental, you say? | 10:20 |
| jasondotstar | do you have have the link to that repo? | 10:21 |
| *** itisha has joined #openstack-zaqar | 10:26 | |
| *** Qiming has quit IRC | 10:48 | |
| *** penchal has quit IRC | 11:02 | |
| *** bradjones has joined #openstack-zaqar | 11:05 | |
| *** bradjones has joined #openstack-zaqar | 11:05 | |
| *** davideagnello has joined #openstack-zaqar | 11:08 | |
| *** bradjones has quit IRC | 11:08 | |
| *** bradjones has joined #openstack-zaqar | 11:08 | |
| *** bradjones has joined #openstack-zaqar | 11:08 | |
| *** davideagnello has quit IRC | 11:13 | |
| *** bradjones has quit IRC | 11:14 | |
| *** bradjones has joined #openstack-zaqar | 11:14 | |
| *** bradjones has joined #openstack-zaqar | 11:14 | |
| *** Nesze has joined #openstack-zaqar | 11:24 | |
| EmilienM | zigo, jasondotstar: yeah, we are in touch :) | 11:35 |
| zigo | EmilienM: I'll let you know when zaqar gets out of the NEW queue, ok? | 11:36 |
| EmilienM | zigo: cool - though we don't test on debian, only uca + rdo | 11:37 |
| *** Qiming has joined #openstack-zaqar | 11:37 | |
| zigo | Sure! :) | 11:37 |
| *** flwang1 has joined #openstack-zaqar | 11:37 | |
| flwang1 | morning | 11:38 |
| flwang1 | vkmc: around? | 11:42 |
| flwang1 | flaper87: ping | 11:43 |
| flaper87 | flwang1: pong | 11:51 |
| flwang1 | flaper87: we need to release a new client version asap https://bugs.launchpad.net/python-zaqarclient/+bug/1492866 | 11:51 |
| openstack | Launchpad bug 1492866 in Python client library for Zaqar "can't install zaqarclient as pbr version is too low" [Critical,Confirmed] | 11:51 |
| flaper87 | flwang1: how are you doing? Mr. PTL | 11:51 |
| flaper87 | flwang1: why asap? | 11:51 |
| flaper87 | flwang1: oh, mmh, ok | 11:52 |
| flwang1 | flaper87: it's breaking heat, i think | 11:52 |
| flaper87 | flwang1: so, wait a bit | 11:52 |
| flaper87 | flwang1: here's what we can do | 11:52 |
| * flaper87 checks something | 11:53 | |
| flwang1 | ok | 11:53 |
| flaper87 | flwang1: so, stable/liberty hasn't been cut yet | 11:54 |
| flaper87 | BUT, it'll likely be cut off of 0.1.1 | 11:54 |
| flaper87 | since that's the latest tag we have for zaqarclient | 11:54 |
| flaper87 | we need this as part of liberty but we need the branch created first | 11:55 |
| flaper87 | releasing a client w/ this fix requires us to release a minor version | 11:55 |
| flaper87 | that means, 0.2.0 | 11:55 |
| flwang1 | why not 0.1.2? | 11:56 |
| flaper87 | I was hoping to get v2 support in before the client was released | 11:56 |
| flwang1 | what's the difference for this case? | 11:56 |
| flaper87 | flwang1: because that'd be a patch release | 11:56 |
| flaper87 | and that's just for bug fixes | 11:56 |
| flwang1 | flaper87: https://review.openstack.org/#/c/224683/1 | 11:56 |
| flaper87 | we should follow semver | 11:56 |
| flaper87 | ih, I thought you had modified that patch yourself | 11:58 |
| flwang1 | which one> | 11:58 |
| flaper87 | nvm, I commented but approved | 11:59 |
| flaper87 | flwang1: try to reach out to dhellman and ask what would be the best way to sort this out | 11:59 |
| flaper87 | flwang1: client libraries releases are now managed by the release team | 12:00 |
| flwang1 | flaper87: ah, not by the project team anymore? | 12:00 |
| flwang1 | flaper87: i saw your comment, and i have the similar question, but after checked nova and other project, they're doing the same thing, that's why i +2 | 12:01 |
| flaper87 | flwang1: lol, I did the same. Next time, comment mentioning you checked that to avoid doing these things twice | 12:02 |
| flaper87 | :P | 12:02 |
| flaper87 | flwang1: yeah, libraries are all released by the release team. There's a process for that | 12:03 |
| flwang1 | haha :) | 12:03 |
| flaper87 | check the project called: openstack/releases | 12:03 |
| flwang1 | flaper87: cool | 12:03 |
| flaper87 | flwang1: also, I recommend you to join: #openstack-relmgr-office | 12:03 |
| flwang1 | flaper87: nice | 12:04 |
| openstackgerrit | Merged openstack/python-zaqarclient: Updated from global requirements https://review.openstack.org/224683 | 12:05 |
| flwang1 | it looks like an upper class club :D | 12:06 |
| flaper87 | :P | 12:06 |
| flwang1 | flaper87: so anything else i can help for this? | 12:07 |
| flaper87 | flwang1: release ? | 12:08 |
| flaper87 | flwang1: yeah, you need to ping Doug and ask him what the best way to handle this is | 12:09 |
| flaper87 | he asked me, before the elections, whether we had a client release planned | 12:09 |
| flaper87 | and I said no because well, we didnt have | 12:09 |
| flaper87 | this seems like a pressing issue that may require some exceptions to be sorte dout | 12:09 |
| flwang1 | flaper87: and we may need to release another one soon to support v2, right? | 12:10 |
| flwang1 | flaper87: can you help ping dhellmann? it's late for me and i have to drop | 12:35 |
| flaper87 | flwang1: sure, I'll ping him | 12:38 |
| flaper87 | flwang1: I can't lead the meeting today | 12:38 |
| flaper87 | :( | 12:38 |
| flwang1 | flaper87: oh, you still around | 12:38 |
| flaper87 | but I don't think there's anything in the agenda | 12:39 |
| flwang1 | flaper87: i can do that | 12:39 |
| flaper87 | other than collecting things for the summit | 12:39 |
| flaper87 | flwang1: isn't it at 15 UTC ? | 12:39 |
| flwang1 | does the channel need any permission to start meeting? | 12:39 |
| flwang1 | flaper87: oh, 15 UTC? | 12:39 |
| flwang1 | i can't join, sorry | 12:39 |
| flwang1 | flaper87: i have some questions as for the horizon demo | 12:40 |
| flwang1 | are you available for a 10 mins talk? | 12:41 |
| flaper87 | flwang1: I am | 12:44 |
| flaper87 | flwang1: no permissions needed at all | 12:44 |
| flwang1 | flaper87: we had a discussion with horizon guys | 12:45 |
| flwang1 | and the two pain points now for horizon are 1. the status update for instance, image and volumes 2. critical/important notifications | 12:45 |
| flwang1 | for #1, horizon has to poll per 2.5 sec to get the latest resource status | 12:46 |
| flwang1 | for example, horizon has to call glance client to get the latest image status | 12:47 |
| flaper87 | Sure, the issue with these scenarios is that it requires us to update other projects as well | 12:47 |
| flaper87 | Not just horizon | 12:47 |
| flwang1 | yes | 12:47 |
| flaper87 | The best way to get this sorted out sooner rather than later is by creating an oslo.messaging driver for Zaqar | 12:48 |
| flwang1 | and for example, glance/nova/cinder doesn't support sending the status change notifications | 12:48 |
| vkmc | helloooooooooooooooooo | 12:48 |
| flaper87 | I've been wrapping my head around this issue for a while and I'm still not 100% convinced such driver is what we want | 12:48 |
| flaper87 | I've mentioned that we could experiment by having that driver living in zaqarclient | 12:48 |
| flaper87 | Which means, whenever Zaqarclient is present, the driver will be present as well | 12:49 |
| flwang1 | flaper87: but what's the drawback? | 12:49 |
| flwang1 | flaper87: it's a nice idea | 12:49 |
| flaper87 | we need to give that a try | 12:49 |
| flaper87 | I'm not against that, FWIW | 12:49 |
| flwang1 | flaper87: yep, but as i mentioned above | 12:49 |
| flaper87 | And, the demo, could be around that | 12:49 |
| vkmc | what, where, why? | 12:50 |
| flwang1 | but for #1, another problem is there is no status change notification | 12:50 |
| flwang1 | vkmc: hey | 12:50 |
| flwang1 | what: horizon demo | 12:51 |
| flaper87 | flwang1: sure, but we could start with something simpler. Like, there are notifications sent for other things | 12:51 |
| flaper87 | image creation, image deletion, etc | 12:51 |
| flwang1 | where: requirement | 12:51 |
| flaper87 | Use that as an example | 12:51 |
| flwang1 | why: i don't know why :D | 12:51 |
| vkmc | haha | 12:51 |
| flaper87 | vkmc: helloooooooooooo | 12:51 |
| flwang1 | flaper87: exactly, that's the #2 | 12:51 |
| flaper87 | flwang1: OR, if we really want to demo this with status changes, then we'll have to hack nova. I'd rather go with vanilla code, though. | 12:52 |
| flaper87 | that is, #2 | 12:52 |
| flwang1 | flaper87: instead of have a oslo.messaging driver, personally i prefer to add a new publish driver in ceilometer | 12:52 |
| flwang1 | i mean for this demo | 12:52 |
| flwang1 | we can figure out the better way later | 12:53 |
| flaper87 | sure, but that cuts off ppl that are not using ceilo | 12:53 |
| flaper87 | ok | 12:53 |
| flwang1 | and another issue is how to notify horizon | 12:53 |
| flwang1 | given we don't support long poll, so technically, notification is a good way to go | 12:54 |
| flwang1 | but that means we may need a new endpoint in horizon to achieve that, does that make sense for u? | 12:54 |
| flwang1 | the new endpoint is the webhook zaqar can send msg to | 12:55 |
| therve | flwang1, Have you seen my comment about that earlier on? | 12:59 |
| flaper87 | flwang1: mmh, what about using websocket ? | 12:59 |
| flwang1 | therve: no, would you mind repeating? :) | 12:59 |
| flaper87 | that requires therve's patch to be merged | 12:59 |
| flaper87 | and, obviously, some javascript magic in horizon | 13:00 |
| therve | flwang1, My patch doesn't just support CRUD, it supports notifications over websocket | 13:00 |
| flaper87 | imho, either work. A webhook in horizon or websocket | 13:00 |
| flwang1 | therve: it would be really cool, i haven't tried that, TBH | 13:01 |
| therve | I haven't tested the whole stack for a while to be honest :) | 13:01 |
| flwang1 | therve: hah | 13:02 |
| flwang1 | therve: i will test it and may bug you for some questions | 13:03 |
| flwang1 | flaper87: btw, i just talked with dhellmann about the zaqar client release | 13:04 |
| flwang1 | i will submit a patch to openstack/release repo and you need +1 for that | 13:04 |
| therve | We may need to do something about that falcon cap too | 13:05 |
| therve | I wonder if the requirement freeze can be lifted for that | 13:05 |
| flaper87 | flwang1: saw it | 13:05 |
| flaper87 | :) | 13:05 |
| flwang1 | flaper87: technically, you're still the PTL? or me? | 13:06 |
| flaper87 | flwang1: you are | 13:06 |
| flaper87 | There's no election for zaqar | 13:06 |
| flwang1 | i'm running for Mikata and now we're still in Liberty :) | 13:06 |
| flwang1 | TBH, i haven't figure out the policy | 13:07 |
| flaper87 | flwang1: you start being a PTL from the moment the election is over | 13:07 |
| *** exploreshaifali has joined #openstack-zaqar | 13:07 | |
| flaper87 | since there was no election for Zaqar, you're the PTL | 13:07 |
| flwang1 | therve: falcon cap is for server side, so you're talking about the zaqar release not zaqar client, right? | 13:07 |
| therve | Yeah | 13:08 |
| flwang1 | flaper87: ok, sir | 13:08 |
| flwang1 | therve: ok, i see. i will keep an eye on that | 13:08 |
| *** davideagnello has joined #openstack-zaqar | 13:09 | |
| therve | It's meeting day too! | 13:09 |
| therve | We need to talk about test environments now that we actually run tests in the gate :) | 13:09 |
| flwang1 | therve: yep, but i can't join due to the timezone :( | 13:09 |
| therve | Ah too bad | 13:10 |
| flwang1 | now it's 1:10 AM | 13:10 |
| *** davideagnello has quit IRC | 13:13 | |
| flwang1 | flaper87: can you reminder me again why we use 0.2.0 instead of 0.1.2 for the next client version? | 13:14 |
| *** penchal has joined #openstack-zaqar | 13:14 | |
| *** dynarro has joined #openstack-zaqar | 13:23 | |
| flaper87 | flwang1: because an update to the requirements file requires a minor release | 13:23 |
| flwang1 | flaper87: got it | 13:23 |
| flwang1 | anything else we need to highlight for the release? | 13:23 |
| flaper87 | flwang1: apparently, you still need my +1 on the release for the liberty thingy | 13:24 |
| *** amitgandhinz has joined #openstack-zaqar | 13:24 | |
| flaper87 | flwang1: but just for this case, the rest I said above stands true | 13:24 |
| flwang1 | flaper87: yep, sure, no problem | 13:24 |
| flwang1 | you're still the PTL for L for sure ;) | 13:25 |
| flwang1 | flaper87: https://review.openstack.org/#/c/225671/1 | 13:29 |
| flwang1 | i have to drop | 13:30 |
| vkmc | so, if we move forward therve's patch and we implement a websocket iface in Horizon side | 13:30 |
| flwang1 | flaper87: pls review it and feel free to submit a patch set | 13:30 |
| flwang1 | vkmc: i'm ok to answer your question before i offline | 13:31 |
| vkmc | we could implement the demo by simply getting data from Ceilometer and relying in Zaqar to notify Horizon on updates? | 13:31 |
| therve | vkmc, For a demo, listening directly to notifications on the bus might be easier | 13:32 |
| flwang1 | vkmc: i think so, and as flaper87 mentioned, you can also create an oslo.messaging driver for zaqar to collect the messages into zaqar | 13:32 |
| flwang1 | therve: define the bus? message bus of rabbitmq? | 13:33 |
| therve | flwang1, Yeah | 13:33 |
| vkmc | and how we filter the messages on the bus that actually mean something to Horizon? | 13:33 |
| flwang1 | therve: but we're trying to promote zaqar :) | 13:33 |
| therve | flwang1, Right. Consume the message from the bus, and put them in zaqar | 13:34 |
| therve | flwang1, vkmc: https://github.com/therve/nabu FWIW. Very basic try. | 13:34 |
| flwang1 | therve: cool, that's one option as well | 13:35 |
| flwang1 | vkmc: do you have a few mins to sync with mrunge? | 13:36 |
| vkmc | flwang1, I do | 13:36 |
| flwang1 | vkmc: let's rock in horizon channel | 13:36 |
| vkmc | :D | 13:37 |
| vkmc | seems mrunge is not around | 13:40 |
| vkmc | and its probably pretty late for you | 13:40 |
| flwang1 | vkmc: yep | 13:41 |
| vkmc | I could ask him what he thinks about our current approach | 13:41 |
| vkmc | and update you later today/tomorrow (for me/for you) | 13:41 |
| flwang1 | vkmc: cool, and for now, we will try to focus on the msg pop up | 13:42 |
| vkmc | agree | 13:42 |
| flwang1 | or call it message center | 13:42 |
| flwang1 | a link display how many message there are | 13:42 |
| flwang1 | and after click the link, show details | 13:43 |
| flwang1 | or something like that | 13:43 |
| flwang1 | and we can figure out the status change scenario later | 13:43 |
| flwang1 | vkmc: i have to drop for sure ;) | 13:43 |
| flwang1 | ttyl | 13:43 |
| vkmc | flwang1, ttyt, get some rest! | 13:44 |
| openstackgerrit | Merged openstack/zaqar: Close backend connections in tests https://review.openstack.org/224466 | 13:46 |
| openstackgerrit | Merged openstack/zaqar: Fix few typos in headings https://review.openstack.org/225412 | 13:47 |
| *** flwang1 has quit IRC | 13:48 | |
| openstackgerrit | Flavio Percoco proposed openstack/python-zaqarclient: Release notes for 0.2.0 https://review.openstack.org/225686 | 13:51 |
| flaper87 | ryansb: vkmc sanity check ^ | 13:51 |
| *** Qiming has quit IRC | 13:55 | |
| *** ametts has joined #openstack-zaqar | 13:55 | |
| *** Qiming has joined #openstack-zaqar | 13:57 | |
| *** sriram has joined #openstack-zaqar | 13:59 | |
| *** mpanetta_ has joined #openstack-zaqar | 14:00 | |
| ryansb | gripe: there are 3 "updated from global requirements", I know they were 3 commits but I don't think users care | 14:00 |
| ryansb | otherwise +1 | 14:00 |
| *** david-ly_ is now known as david-lyle | 14:01 | |
| ryansb | flaper87: ^ | 14:07 |
| openstackgerrit | Merged openstack/python-zaqarclient: Using messaging instead of queueing https://review.openstack.org/224983 | 14:13 |
| flaper87 | ryansb: I agree but there were 3 different upgrades in 3 different times for different requirements :/ | 14:16 |
| flaper87 | I'd rather have those | 14:16 |
| ryansb | yeah, but I feel like users just care about "did they update deps?" not "how many individual times have they updated deps since the last release?" | 14:17 |
| ryansb | and if they have the latter question, they can use the git log | 14:17 |
| *** csoukup has joined #openstack-zaqar | 14:17 | |
| ryansb | I'm fine with merging it either way, since it's an aesthetic choice anyway | 14:17 |
| Nesze | Hey, I’m fairly new to Zaqar, trying to experiment with it: currently I’m trying to split up zaqar in two docker containers, one with zaqar itself and another with a mongo backend, however when starting zaqar http://paste.openstack.org/show/472798/ I get this. Have you guys any tip? Here is my config file: http://paste.openstack.org/show/472804/ | 14:20 |
| flaper87 | ryansb: tbh, I'm good with either too but I guess, I'd lean towards an extended changelog since those happened in 3 different commits | 14:22 |
| vkmc | Nesze, welcome! | 14:22 |
| *** david-lyle has quit IRC | 14:22 | |
| vkmc | Nesze, add "unreliable=True" in your conf file and restart Zaqar | 14:23 |
| vkmc | that should work | 14:23 |
| *** penchal has quit IRC | 14:23 | |
| *** david-lyle has joined #openstack-zaqar | 14:26 | |
| *** Qiming has quit IRC | 14:34 | |
| *** kgriffs|afk is now known as kgriffs | 14:34 | |
| *** Qiming has joined #openstack-zaqar | 14:35 | |
| *** achanda has joined #openstack-zaqar | 14:51 | |
| *** diga has joined #openstack-zaqar | 14:54 | |
| Nesze | thanks vkmc, unfortunately it still fails with the same message, new config: https://pastee.org/bqdx5 | 14:55 |
| Nesze | my mongo is v3.0.6, the pymongo client is 3.0.3, should that be the cause? mongo logs https://pastee.org/4c9cc at each try. (This might not even be a zaqar thing, so sry if it’s off-topic) | 14:58 |
| *** achanda has quit IRC | 15:01 | |
| Nesze | cool, it works, I moved the “unreliable=True” under [DEFAULT], thanks again | 15:01 |
| vkmc | Nesze, sorry, I got distracted | 15:02 |
| vkmc | glad its working for you :) | 15:03 |
| ryansb | ah, that'd do it | 15:03 |
| *** exploreshaifali has quit IRC | 15:41 | |
| *** Qiming has quit IRC | 15:46 | |
| *** diga has quit IRC | 15:51 | |
| ryansb | argh. Missed the meeting by an hour. Calendar fail :( | 16:05 |
| *** Nesze has quit IRC | 16:06 | |
| vkmc | ryansb, seems we all kinda missed it | 16:07 |
| vkmc | some of us should run the meeting at 3pmUTC | 16:08 |
| ryansb | yeah... fixed the event so I will actually go next week | 16:10 |
| ryansb | err, in two weeks since we alternate | 16:10 |
| *** stanchan has quit IRC | 16:18 | |
| *** sriram1 has joined #openstack-zaqar | 16:49 | |
| *** sriram has quit IRC | 16:49 | |
| *** dynarro has quit IRC | 16:55 | |
| *** kgriffs is now known as kgriffs|afk | 16:57 | |
| *** kgriffs|afk is now known as kgriffs | 16:59 | |
| *** kgriffs is now known as kgriffs|afk | 17:15 | |
| *** achanda has joined #openstack-zaqar | 17:20 | |
| *** davideagnello has joined #openstack-zaqar | 17:23 | |
| *** amitgandhinz has quit IRC | 17:40 | |
| *** amitgandhinz has joined #openstack-zaqar | 17:41 | |
| *** stanchan has joined #openstack-zaqar | 17:42 | |
| *** stanchan has quit IRC | 17:43 | |
| openstackgerrit | Flavio Percoco proposed openstack/zaqar: DO NOT MERGE: testing falcon 0.3.x https://review.openstack.org/225269 | 17:58 |
| *** stanchan has joined #openstack-zaqar | 18:06 | |
| *** kgriffs|afk is now known as kgriffs | 18:14 | |
| EmilienM | flaper87: I wish I could hire you, but I realize we already are in the same company | 18:53 |
| EmilienM | hopefully | 18:53 |
| *** ametts has quit IRC | 19:01 | |
| *** bradjones has quit IRC | 19:04 | |
| *** amitgandhinz has quit IRC | 19:05 | |
| jasondotstar | zaqarians | 19:25 |
| jasondotstar | 19:25 | |
| jasondotstar | is python3 required? | 19:26 |
| jasondotstar | for zaqar? | 19:26 |
| jasondotstar | the server apparently requires the enum module, and as I've read.... that module was only added in Python v3.4? | 19:26 |
| ryansb | no, zaqar is compatible with python2.7 and python 3.3/3.4 | 19:27 |
| jasondotstar | ryansb: ok | 19:27 |
| jasondotstar | guess i need to find the python2.7 version of that module.... | 19:28 |
| jasondotstar | thx | 19:28 |
| ryansb | where does it require `enum`? | 19:28 |
| ryansb | and have you installed the requirements? | 19:29 |
| ryansb | because "enum34" is a required package for python 2.6 and 2.7 | 19:29 |
| jasondotstar | zaqar/storage/base.py | 19:29 |
| jasondotstar | yeah | 19:29 |
| jasondotstar | python-enum34.noarch : Backport of Python 3.4 Enum | 19:29 |
| ryansb | if you haven't, please "pip install -r requirements.txt" | 19:29 |
| jasondotstar | just found it. | 19:29 |
| jasondotstar | well, I'm installing this from the RPM pkg | 19:29 |
| jasondotstar | but... might be a good idea | 19:30 |
| jasondotstar | for me to look at that file | 19:30 |
| jasondotstar | and see what i need to install as reqs | 19:30 |
| ryansb | yeah, often (for RHEL-derivatives) you can just toss "python-" in front of every line of requirements.txt and the package will be there | 19:30 |
| jasondotstar | ryansb: i literally *juuust* noticed that | 19:31 |
| jasondotstar | :-) | 19:31 |
| vkmc | boris-42, for some reason this is falling https://review.openstack.org/#/c/225192/ | 19:53 |
| vkmc | boris-42, any hint? | 19:53 |
| vkmc | boris-42, seems like it cannot parse it | 19:54 |
| vkmc | boris-42, nvm, I know where is the problem | 19:54 |
| * vkmc should RTFM | 19:55 | |
| openstackgerrit | Victoria Martinez de la Cruz proposed openstack/zaqar: Adds max failure rate limit https://review.openstack.org/225192 | 19:59 |
| *** ametts has joined #openstack-zaqar | 20:15 | |
| *** amitgandhinz has joined #openstack-zaqar | 20:16 | |
| *** csoukup has quit IRC | 20:18 | |
| *** Nesze has joined #openstack-zaqar | 20:23 | |
| Nesze | I’m running zaqar in a Docker container, and having difficulties accessing the server with ‘curl’ from the host machine. The port forwarding is set correctly, the iptables containes the mapping, also, I have no problem to curl the server from inside the container. This might be a docker issue, but could be that by default the python webserver listens only on the loopback interface, 127.0.0.1? Is there a way to bind it to all the interfaces: 0.0. | 20:38 |
| vkmc | Nesze, yeah, that's another configuration | 20:39 |
| vkmc | under [drivers:transport:wsgi] | 20:40 |
| vkmc | bind = 0.0.0.0 | 20:40 |
| vkmc | port = 8888 | 20:41 |
| vkmc | that should work | 20:41 |
| Nesze | thanks vkmc, I’m going to try it out | 20:41 |
| vkmc | np | 20:42 |
| * EmilienM | 20:50 | |
| Nesze | cool, it works | 20:53 |
| * vkmc | 20:53 | |
| vkmc | Nesze, :D | 20:54 |
| *** Nesze has quit IRC | 21:01 | |
| *** itisha has quit IRC | 21:03 | |
| *** sriram1 has quit IRC | 21:12 | |
| flwang | vkmc: around? | 21:18 |
| vkmc | flwang, here | 21:22 |
| flwang | vkmc: anything from the chat your and mrunge? | 21:22 |
| vkmc | flwang, he didn't reply | 21:22 |
| flwang | vkmc: ok, no worries | 21:22 |
| vkmc | seems like he wasn't around today | 21:22 |
| flwang | vkmc: are you going to use therve's tool to collect the msg? | 21:24 |
| vkmc | I'm trying to wrap my head around things | 21:24 |
| *** achanda has quit IRC | 21:24 | |
| flwang | ok, cool | 21:25 |
| vkmc | I didn't follow what flaper87 suggested | 21:25 |
| vkmc | that thing about having the driver living in zaqarclient | 21:25 |
| vkmc | seems like something that should live in the server side | 21:26 |
| *** amitgandhinz has quit IRC | 21:28 | |
| flwang | vkmc: it's talking about adding a olso.messaging driver, and given zaqar is not a true 'message bus' like rabbitmq and qpid, so it may be a little strange to place the code in oslo.messaging repo(they may not accept it either :) ), so i think that's why flaper87 said keep the code in zaqarclient so that, when the zaqar client installed, the driver is there as well | 21:29 |
| *** amitgandhinz has joined #openstack-zaqar | 21:29 | |
| vkmc | oh yes, that I understand | 21:30 |
| vkmc | still, seems odd to have a driver for oslo in the client | 21:30 |
| flwang | yep | 21:30 |
| flwang | so we don't care about that for now | 21:31 |
| flwang | we either use ceilometer to collect the msg, or leverage a tool like therve's to grab the msg from bus and post it into zaqar | 21:31 |
| flwang | all work | 21:31 |
| flwang | today i'm going to review therve's patch to see how horizon can be benefited by the websocket/notifications | 21:32 |
| *** mpanetta_ has quit IRC | 21:33 | |
| *** flwang1 has joined #openstack-zaqar | 21:36 | |
| fungi | zigo: noticing the falcon 3.x discussion in scrollback. barbican seems to have switched from falcon to pecan over a year ago in https://review.openstack.org/89746 so shouldn't be a blocker | 21:36 |
| fungi | i think as long as we get zaqar working with latest falcon asap, you should be all set when you package liberty | 21:37 |
| zigo | fungi: Thanks. | 21:38 |
| zigo | fungi: So, do you confirm Zaqar is now the only user of Falcon? | 21:39 |
| openstackgerrit | Merged openstack/zaqar: Adds max failure rate limit https://review.openstack.org/225192 | 21:39 |
| fungi | zigo: it's the only one i can find, yes | 21:39 |
| fungi | zigo: well, i mean there are a couple other new/low-profile projects too but they're also not relevant in this situation because they don't receive global requirements sync proposals anyway (kiloeyes, freezer-api, monasca-api, stacktach-quincy) | 21:42 |
| zigo | Ok, they are all out of my scope ! :) | 21:43 |
| *** achanda has joined #openstack-zaqar | 21:44 | |
| *** achanda has quit IRC | 21:44 | |
| *** achanda has joined #openstack-zaqar | 21:44 | |
| *** kgriffs is now known as kgriffs|afk | 21:45 | |
| *** amitgandhinz has quit IRC | 22:07 | |
| *** kgriffs|afk is now known as kgriffs | 22:13 | |
| *** JAHoagie has joined #openstack-zaqar | 22:48 | |
| *** ametts has quit IRC | 22:53 | |
| *** stanchan has quit IRC | 22:54 | |
| *** JAHoagie has quit IRC | 23:03 | |
| *** JAHoagie has joined #openstack-zaqar | 23:29 | |
| *** Qiming has joined #openstack-zaqar | 23:36 | |
| *** kgriffs is now known as kgriffs|afk | 23:39 | |
| *** JAHoagie has quit IRC | 23:40 | |
| *** achanda_ has joined #openstack-zaqar | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!