| *** malini has quit IRC | 00:01 | |
| *** kgriffs is now known as kgriffs|afk | 00:03 | |
| *** amitgandhinz has quit IRC | 00:03 | |
| *** flwang1 has quit IRC | 00:10 | |
| *** jeffrey4l__ has joined #openstack-zaqar | 00:12 | |
| *** openstackgerrit has quit IRC | 00:31 | |
| *** openstackgerrit has joined #openstack-zaqar | 00:32 | |
| *** echevemaster has quit IRC | 00:45 | |
| *** cpallares has joined #openstack-zaqar | 00:45 | |
| *** kgriffs|afk is now known as kgriffs | 00:54 | |
| *** kgriffs is now known as kgriffs|afk | 01:04 | |
| *** echevemaster has joined #openstack-zaqar | 01:16 | |
| *** jeffrey4l__ has quit IRC | 01:21 | |
| *** jeffrey4l__ has joined #openstack-zaqar | 01:37 | |
| *** echevemaster has quit IRC | 01:48 | |
| *** vipul has quit IRC | 01:48 | |
| *** dmitryme has quit IRC | 01:48 | |
| *** echevemaster has joined #openstack-zaqar | 01:54 | |
| *** vipul has joined #openstack-zaqar | 01:54 | |
| *** dmitryme has joined #openstack-zaqar | 01:54 | |
| *** kgriffs|afk is now known as kgriffs | 01:55 | |
| *** sebasmagri has quit IRC | 01:57 | |
| *** sebasmagri has joined #openstack-zaqar | 01:58 | |
| *** sebasmagri is now known as Guest57163 | 01:59 | |
| *** kgriffs is now known as kgriffs|afk | 02:05 | |
| *** cpallares has quit IRC | 02:22 | |
| *** alcabrera is now known as alcabrera|afk | 02:46 | |
| *** kgriffs|afk is now known as kgriffs | 02:55 | |
| *** jeffrey4l__ has quit IRC | 03:04 | |
| *** kgriffs is now known as kgriffs|afk | 03:05 | |
| *** jeffrey4l has joined #openstack-zaqar | 03:17 | |
| *** vkmc has quit IRC | 03:33 | |
| *** prashanthr_ has quit IRC | 03:45 | |
| *** prashanthr_ has joined #openstack-zaqar | 03:47 | |
| *** prashanthr_ has quit IRC | 03:54 | |
| *** prashanthr_ has joined #openstack-zaqar | 03:55 | |
| *** kgriffs|afk is now known as kgriffs | 03:56 | |
| *** echevemaster has quit IRC | 04:00 | |
| *** Guest57163 is now known as sebasmagri | 04:02 | |
| *** sebasmagri is now known as Guest80276 | 04:03 | |
| *** prashanthr_1 has joined #openstack-zaqar | 04:05 | |
| *** kgriffs is now known as kgriffs|afk | 04:06 | |
| *** Guest80276 is now known as sebasmagri | 04:06 | |
| *** sebasmagri has joined #openstack-zaqar | 04:06 | |
| *** sebasmagri has joined #openstack-zaqar | 04:06 | |
| *** prashanthr_ has quit IRC | 04:07 | |
| *** jeffrey4l has quit IRC | 04:11 | |
| *** prashanthr_ has joined #openstack-zaqar | 04:13 | |
| *** prashanthr_1 has quit IRC | 04:14 | |
| *** prashanthr_ has quit IRC | 04:22 | |
| *** jeffrey4l has joined #openstack-zaqar | 04:30 | |
| *** prashanthr_ has joined #openstack-zaqar | 04:45 | |
| *** kgriffs|afk is now known as kgriffs | 04:57 | |
| *** kgriffs is now known as kgriffs|afk | 05:07 | |
| *** sgotliv has joined #openstack-zaqar | 05:09 | |
| *** prashanthr_1 has joined #openstack-zaqar | 05:30 | |
| *** prashanthr_ has quit IRC | 05:31 | |
| *** prashanthr_1 has quit IRC | 05:34 | |
| *** prashanthr_ has joined #openstack-zaqar | 05:47 | |
| *** kgriffs|afk is now known as kgriffs | 05:58 | |
| *** kgriffs is now known as kgriffs|afk | 06:08 | |
| *** jeffrey4l has quit IRC | 06:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/zaqar: Imported Translations from Transifex https://review.openstack.org/122070 | 06:12 |
|---|---|---|
| *** flaper87|afk is now known as flaper87 | 06:20 | |
| *** jeffrey4l has joined #openstack-zaqar | 06:24 | |
| *** kgriffs|afk is now known as kgriffs | 06:59 | |
| *** kgriffs is now known as kgriffs|afk | 07:09 | |
| *** kgriffs|afk is now known as kgriffs | 08:00 | |
| *** sgotliv has quit IRC | 08:07 | |
| *** kgriffs is now known as kgriffs|afk | 08:09 | |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Cleanup technical debt https://review.openstack.org/121656 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Rename `shard` to `pool` https://review.openstack.org/121494 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Use queue_creation to send queue's metadata https://review.openstack.org/121564 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Add support for `pop` implemented in v1.1 https://review.openstack.org/120155 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Respect queue's lazy creation https://review.openstack.org/121565 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Always pass a project_id https://review.openstack.org/121596 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Add support for flavor to the client library https://review.openstack.org/121534 | 08:14 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Fix functional tests https://review.openstack.org/121563 | 08:14 |
| *** flwang1 has joined #openstack-zaqar | 08:19 | |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/zaqar: Remove config sample and generate it using code https://review.openstack.org/122089 | 08:26 |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/zaqar: Fix the redis opts https://review.openstack.org/122093 | 08:45 |
| *** kgriffs|afk is now known as kgriffs | 09:00 | |
| *** kgriffs is now known as kgriffs|afk | 09:10 | |
| *** sgotliv has joined #openstack-zaqar | 09:17 | |
| *** sgotliv has quit IRC | 09:57 | |
| *** malini has joined #openstack-zaqar | 10:00 | |
| *** jeffrey4l has quit IRC | 10:01 | |
| *** kgriffs|afk is now known as kgriffs | 10:01 | |
| *** jeffrey4l has joined #openstack-zaqar | 10:04 | |
| *** kgriffs is now known as kgriffs|afk | 10:11 | |
| *** tmu_ is now known as tmu | 10:34 | |
| *** sgotliv has joined #openstack-zaqar | 10:37 | |
| *** myra has quit IRC | 10:37 | |
| *** kgriffs|afk is now known as kgriffs | 11:02 | |
| *** prashanthr_ has quit IRC | 11:05 | |
| *** kgriffs is now known as kgriffs|afk | 11:12 | |
| *** malini has quit IRC | 11:27 | |
| *** vkmc has joined #openstack-zaqar | 11:32 | |
| *** vkmc has quit IRC | 11:32 | |
| *** vkmc has joined #openstack-zaqar | 11:32 | |
| vkmc | good morning o/ | 11:33 |
| vkmc | flaper87, hey Fla, how are you? | 11:50 |
| vkmc | I'm checking https://review.openstack.org/#/c/120155/4/zaqarclient/queues/v1/api.py | 11:50 |
| vkmc | according to the API spec, shouldn't it be delete? | 11:50 |
| vkmc | https://wiki.openstack.org/wiki/Zaqar/specs/api/v1.1#Delete_Multiple_Messages | 11:51 |
| vkmc | I mean, there is no POP method, we implemented it with DELETE | 11:52 |
| vkmc | https://wiki.openstack.org/wiki/Zaqar/specs/api/v1.1#Endpoints_Synopsis | 11:52 |
| vkmc | sorry I misread your fix -.- | 11:53 |
| vkmc | rm -rf ^ | 11:54 |
| *** sgotliv has quit IRC | 12:01 | |
| *** kgriffs|afk is now known as kgriffs | 12:03 | |
| flaper87 | goooooooooooooo morning | 12:03 |
| flaper87 | vkmc: and I wrote my comment in the wrong line | 12:04 |
| flaper87 | that's my bad :( | 12:04 |
| *** sgotliv has joined #openstack-zaqar | 12:04 | |
| vkmc | flaper87, hey :) | 12:05 |
| vkmc | it seemed odd to me when I reviewed TBH, I misunderstood https://wiki.openstack.org/wiki/Zaqar/specs/api/v1.1#Delete_Multiple_Messages | 12:06 |
| flaper87 | vkmc: hehe, indeed :D | 12:06 |
| vkmc | flaper87, how are you doing? | 12:07 |
| flaper87 | vkmc: I'm good, you? | 12:07 |
| flaper87 | the wait is kinda killing me | 12:08 |
| flaper87 | we're at 7 -1 and 6 +1 | 12:08 |
| vkmc | I'm fine too | 12:08 |
| flaper87 | if ttx changes his mind, we're in | 12:08 |
| vkmc | yes... it's truly a bummer | 12:08 |
| flaper87 | but he doesn't seem to be around | 12:08 |
| vkmc | I was going to send you a tweet or something when I saw jpipes vote haha | 12:08 |
| flaper87 | haahahahaha | 12:08 |
| vkmc | and his feedback | 12:08 |
| vkmc | (probably you were f5 in dreams) | 12:09 |
| vkmc | s/f5/hitting f5 | 12:09 |
| flaper87 | kinda, | 12:09 |
| vkmc | I dunno... keep away from gratuation for one vote | 12:10 |
| vkmc | one 'hesitant' vote | 12:10 |
| vkmc | that shouldn't be allowed haha | 12:10 |
| flaper87 | what really worries me right now and I think this was not discussed enough is that holding back zaqar will affect its adaption really bad | 12:11 |
| flaper87 | holding it back for 1 more cycle means that it'll be released with other openstack projects in 1 year | 12:11 |
| flaper87 | that's way too much time | 12:11 |
| vkmc | what I don't understand | 12:12 |
| vkmc | well... not understand... what I cannot conceive | 12:12 |
| *** kgriffs is now known as kgriffs|afk | 12:12 | |
| vkmc | is the fact that there are no plans for changing anything in the core in the next cycle | 12:13 |
| vkmc | we may be discussing and clearing up the whole messaging or queue dillema | 12:13 |
| vkmc | but nothing else | 12:13 |
| vkmc | do we really need a cycle to do that? | 12:14 |
| flaper87 | right | 12:14 |
| vkmc | shouldn't it be more fair to get graduate and make sure that situation is defined by K-1? | 12:14 |
| flaper87 | yeah, that's what I'm still trying to explain to ppl | 12:14 |
| flaper87 | we know what we are and where we are headed | 12:14 |
| vkmc | we have a summit in between, if needed, we could have some sort of session to sit together and discuss what is Zaqar trying to be and how we are planning to do it | 12:15 |
| vkmc | sit together with those who doesn't understand Zaqar position I meant | 12:15 |
| flaper87 | TBH, I'm tired of doing that | 12:15 |
| vkmc | yeah I understand that | 12:15 |
| flaper87 | I prefer dedicating time to those who already understand it and want to use it | 12:15 |
| vkmc | but integration also means having to explain others what we are up to | 12:15 |
| flaper87 | instead of using that time to explain it to others and start discussing what queuing and messaging is | 12:16 |
| flaper87 | right but we've done that for 2 cycles | 12:16 |
| vkmc | yeah, but it seems that it's not enough :( | 12:16 |
| vkmc | don't get me wrong | 12:16 |
| flaper87 | I think there's people that need to do their homework | 12:16 |
| vkmc | I don't want to waste time explaining what Zaqar is | 12:16 |
| flaper87 | that's why I don't think we should dedicate more time to this "explaining zaqar' thing | 12:16 |
| flaper87 | We failed 1 time and some people learned from it | 12:17 |
| flaper87 | they dug into zaqar and tried to udnerstand | 12:17 |
| flaper87 | we made mistakes and we worked on those things | 12:17 |
| flaper87 | that's what really frustrates me | 12:18 |
| vkmc | yeah, it's obvious that there is a good predisposition from Zaqar side | 12:18 |
| flaper87 | We've put so much time reaching out to people | 12:18 |
| flaper87 | explaining what zaqar is and collecting use cases | 12:18 |
| vkmc | right now, as ttx admitted, is TC's fault | 12:18 |
| vkmc | maybe we should poke TC so they keep an eye on us | 12:18 |
| vkmc | but that is expected from their side, not from ours | 12:19 |
| flaper87 | I'm not expecting the TC to sya: "Hey we screwed so lets approve regardless of what the project does". What I'd like is to make people understand that we know where we are headed and regardless of the result of the queuing/messaging discussion there are still use cases for it and people *want* it | 12:19 |
| flaper87 | the API is stable and we've several different deployments covered | 12:19 |
| vkmc | yeah | 12:19 |
| vkmc | that is what I wanted to say in last TC meeting | 12:20 |
| vkmc | it doesn't matter if we are a messaging system, a queue, or a jar of nutella | 12:20 |
| flaper87 | LOL @ jar of nutella | 12:20 |
| flaper87 | FWIW, I said it and zaneb said that too | 12:20 |
| flaper87 | "I don' | 12:20 |
| flaper87 | ops | 12:20 |
| flaper87 | he said something like: "I don't care how you call it" | 12:21 |
| vkmc | it has been demostrated that Zaqar has several use cases for integrated projects and for other projects as well | 12:21 |
| vkmc | yea I think I remember now | 12:21 |
| flaper87 | I'm not willing to make 3 guys happy at the cost of sacrificing use-case | 12:21 |
| flaper87 | use-cases* | 12:22 |
| vkmc | that is way more polite that my way of expressing it haha | 12:22 |
| vkmc | yeah | 12:22 |
| vkmc | btw, why they say that removing get by id we won't qualify as a messaging system? | 12:22 |
| flaper87 | that's why I said: "Regardless of the result of this discussion, it ain't going to change" | 12:22 |
| vkmc | AFAIK there is no contraint about message system features | 12:22 |
| flaper87 | right... At this point, I think it's just a matter of perception, really | 12:23 |
| vkmc | they have to start thinking out of the box | 12:23 |
| flaper87 | vkmc: PM ;) | 12:24 |
| flaper87 | anyway | 12:24 |
| flaper87 | I still have hope that ttx will reconsider his vote | 12:24 |
| flaper87 | but that's the only hope I have left | 12:24 |
| vkmc | yeah, I was thinking to leave a gentle review with this thoughts | 12:25 |
| vkmc | I'm not sure | 12:25 |
| vkmc | if there is a slightly chance to get some of the -1's flip to our side | 12:26 |
| vkmc | we should take it | 12:27 |
| vkmc | flaper87, ^ | 12:28 |
| flaper87 | yeah, feel free to do it. I added a couple of comments last night. One of them saying what we discussed here | 12:29 |
| vkmc | cool | 12:30 |
| *** malini has joined #openstack-zaqar | 12:35 | |
| *** malini has quit IRC | 12:48 | |
| *** jchai has joined #openstack-zaqar | 12:54 | |
| *** kgriffs|afk is now known as kgriffs | 13:04 | |
| *** sriram has joined #openstack-zaqar | 13:08 | |
| *** mpanetta has joined #openstack-zaqar | 13:12 | |
| *** kgriffs is now known as kgriffs|afk | 13:13 | |
| *** mpanetta has quit IRC | 13:16 | |
| *** mpanetta has joined #openstack-zaqar | 13:16 | |
| *** echevemaster has joined #openstack-zaqar | 13:17 | |
| *** malini has joined #openstack-zaqar | 13:47 | |
| *** malini has left #openstack-zaqar | 13:47 | |
| *** malini has joined #openstack-zaqar | 13:47 | |
| malini | hello!! | 13:49 |
| vkmc | hi malini! | 13:50 |
| malini | how are you? | 13:50 |
| *** sgotliv has quit IRC | 13:51 | |
| vkmc | good good and you? | 13:51 |
| vkmc | last news on the Zaqar graduation http://ttx.re/next-steps-for-zaqar.html | 13:52 |
| vkmc | :/ | 13:52 |
| flaper87 | and you can tell by the face | 13:53 |
| flaper87 | :D | 13:53 |
| vkmc | yeah... | 13:53 |
| vkmc | spoiler alert | 13:53 |
| mpanetta | blah | 13:53 |
| malini | I saw the votes y'day evening :( | 13:53 |
| malini | But I am determined not to let some folks lack of understanding ruin my day | 13:54 |
| flaper87 | vkmc: malini girls, would you have time to go through the client reviews today? | 13:54 |
| malini | flaper87: you have pop-tarts? | 13:54 |
| vkmc | flaper87, I'm doing that, check out https://review.openstack.org/#/c/121534/4 | 13:54 |
| vkmc | +1 malini | 13:54 |
| flaper87 | malini: I have gummy bears that I'm willing to share | 13:55 |
| malini | flaper87: I can live with tht | 13:55 |
| flaper87 | actually, these ones are called: "funny puzzle" | 13:56 |
| flaper87 | WTF ? | 13:56 |
| vkmc | there is something I'm not getting... hmm in L338 here https://review.openstack.org/#/c/121563/4/zaqarclient/tests/queues/queues.py why you assert that the length of the messages list is 1 instead of 3? | 13:56 |
| vkmc | and in L354 0 instead of 1 | 13:56 |
| flaper87 | vkmc: that's a falcon bug | 13:56 |
| flaper87 | well, not really a bug | 13:56 |
| flaper87 | lemme explain | 13:56 |
| vkmc | sure | 13:56 |
| flaper87 | requests serialices ids as ?ids=1&ids=2&ids=3 | 13:57 |
| flaper87 | however, falcon doesn't know how to parse that | 13:57 |
| flaper87 | it knows how to parse ids=1,2,3 | 13:57 |
| vkmc | aha... | 13:57 |
| flaper87 | I talked with kgriffs|afk and that's comming in 0.2 | 13:57 |
| flaper87 | we need to have a fix for that in the client, though | 13:57 |
| flaper87 | but the fix is in the CLI patch | 13:57 |
| flaper87 | once all that lands, I'll fix the test | 13:57 |
| vkmc | doesn't the endpoint expects ?ids=1,2,3? | 13:58 |
| flaper87 | there's no standard on that, TBH | 13:58 |
| vkmc | it seems odd to have ?ids=1&ids=2&ids=3 | 13:58 |
| flaper87 | there's no "URL-rfc" explaining how multiple parameters must be specified | 13:58 |
| flaper87 | AFAIK | 13:58 |
| flaper87 | vkmc: yup but it's quite common | 13:59 |
| vkmc | now that I'm reading... yeah | 13:59 |
| vkmc | it's quite common for different parameters though | 13:59 |
| vkmc | anyhow | 14:02 |
| vkmc | that is on the request | 14:03 |
| vkmc | but when you get the set of messages | 14:03 |
| vkmc | you get an iterator with the messages | 14:03 |
| flaper87 | right | 14:05 |
| *** kgriffs|afk is now known as kgriffs | 14:05 | |
| flaper87 | but since falcon is not parsing the `ids` thing correctly, it's just getting 1 id and sending 1 message back | 14:05 |
| vkmc | oic | 14:06 |
| flaper87 | 1 is wrong for sure, I just can't fix it in that specific patch | 14:06 |
| flaper87 | :P | 14:06 |
| flaper87 | once the CLI patch lands with the fix for it, I'll update the test | 14:06 |
| vkmc | don't we require Falcon update for that? | 14:07 |
| *** sgotliv has joined #openstack-zaqar | 14:07 | |
| flaper87 | vkmc: we can either wait for falcon to support `ids=&ids=&ids` or just serialize our ids in a way that falcon knows how to parse `ids=1,2,3` | 14:07 |
| vkmc | flaper87, fair enough | 14:08 |
| flaper87 | vkmc: this is the fix I'm referring to https://review.openstack.org/#/c/120966/4/zaqarclient/queues/v1/core.py,cm | 14:08 |
| *** jchai is now known as jchai_afk | 14:08 | |
| flaper87 | read my comments there and let me know if you agree | 14:08 |
| vkmc | kewl | 14:09 |
| vkmc | I agree | 14:09 |
| *** kgriffs is now known as kgriffs|afk | 14:14 | |
| *** amitgandhinz has joined #openstack-zaqar | 14:23 | |
| vkmc | flaper87, L73 https://review.openstack.org/#/c/121564/4/zaqarclient/queues/v1/queues.py what happens if the queue already exist? just updates the metadata? | 14:27 |
| vkmc | it's not in the wiki, we should add it | 14:27 |
| vkmc | https://wiki.openstack.org/wiki/Zaqar/specs/api/v1.1#Create_Queue | 14:27 |
| *** jeffrey4l has quit IRC | 14:28 | |
| flaper87 | vkmc: yup | 14:33 |
| flaper87 | AFAIK | 14:33 |
| flaper87 | it just overwrites it | 14:33 |
| flaper87 | we need to work on the support for `patch` | 14:33 |
| *** alcabrera|afk is now known as alcabrera | 14:34 | |
| vkmc | k | 14:34 |
| *** sebasmagri has quit IRC | 14:36 | |
| *** jeffrey4l has joined #openstack-zaqar | 14:36 | |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Cleanup technical debt https://review.openstack.org/121656 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Rename `shard` to `pool` https://review.openstack.org/121494 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Use queue_creation to send queue's metadata https://review.openstack.org/121564 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Respect queue's lazy creation https://review.openstack.org/121565 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Always pass a project_id https://review.openstack.org/121596 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Add support for flavor to the client library https://review.openstack.org/121534 | 14:41 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Fix functional tests https://review.openstack.org/121563 | 14:41 |
| flaper87 | vkmc: malini ^^^ | 14:42 |
| vkmc | thanks flaper87 | 14:42 |
| malini | it is weird why it carries over my prev +2 https://review.openstack.org/#/c/121656/ | 14:43 |
| flaper87 | malini: it carries it if I didn't change the code | 14:44 |
| flaper87 | and I didn't touch that patch | 14:44 |
| vkmc | flaper87, you added the warning | 14:45 |
| malini | so why did you have a new patchset? coz of dependency? | 14:45 |
| flaper87 | malini: yes | 14:45 |
| flaper87 | vkmc: the warning is another patch | 14:45 |
| flaper87 | AFAIK | 14:45 |
| vkmc | oh | 14:45 |
| vkmc | makes sense now | 14:45 |
| vkmc | magic git | 14:45 |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/python-zaqarclient: CLI support for message CRUD operations https://review.openstack.org/120966 | 14:45 |
| jeffrey4l | flaper87, Added the FIXME comments | 14:47 |
| flaper87 | jeffrey4l: will take a look now | 14:47 |
| jeffrey4l | thx | 14:48 |
| vkmc | flaper87, https://review.openstack.org/#/c/121534/ | 14:50 |
| flaper87 | vkmc: ? :) | 14:51 |
| vkmc | flaper87, can you check the schema? | 14:51 |
| flaper87 | vkmc: mmh, I swear I responded | 14:51 |
| *** sebasmagri has joined #openstack-zaqar | 14:51 | |
| flaper87 | vkmc: it's `flavors_name` it has to match the ref | 14:51 |
| vkmc | flaper87, yeah, I meant there is a required param missing | 14:52 |
| flaper87 | and the pool_name and capabilities are in the flavor_data | 14:52 |
| flaper87 | ah, mmh. | 14:52 |
| flaper87 | wah yeah so here's the thing | 14:52 |
| flaper87 | vkmc: it's a required param here: https://review.openstack.org/#/c/121534/4/zaqarclient/queues/v1/flavor.py,cm | 14:52 |
| vkmc | yes | 14:53 |
| flaper87 | the reason I didn't make it required in the core module is because people should know what they're doing if they're using the core module | 14:53 |
| flaper87 | :) | 14:53 |
| flaper87 | I had a 10mins discussion with myself on that | 14:53 |
| vkmc | and what happens if no pool_name is passed there? | 14:53 |
| flaper87 | "Should I add a bunch of required arguments in the core module? or should I tell the user to read the spec?" | 14:54 |
| flaper87 | vkmc: zaqar will raise 400 | 14:54 |
| flaper87 | core module is a lower-level module | 14:54 |
| vkmc | ok.. | 14:54 |
| flaper87 | if the users use it, they must know what they're doing | 14:54 |
| *** kgriffs|afk is now known as kgriffs | 14:54 | |
| flaper87 | if they don't then zaqar will make sure to let them know | 14:55 |
| vkmc | flaper87, and in here? https://review.openstack.org/#/c/121534/4/zaqarclient/queues/v1/core.py | 14:55 |
| vkmc | btw, why your gerrit is fancier that mine? | 14:55 |
| flaper87 | vkmc: go to settings and enable the "new ui" | 14:57 |
| flaper87 | ;) | 14:57 |
| flaper87 | vkmc: same thing, the pool should be in the flavor_data dict | 14:58 |
| vkmc | ok | 14:59 |
| vkmc | well, no more comments on that patchset | 15:01 |
| vkmc | I added this as a response to the comments I made before | 15:01 |
| * vkmc enables new ui | 15:01 | |
| vkmc | http://alltheragefaces.com/img/usercreated/4f71bfdabd626.png | 15:04 |
| *** jchai_afk is now known as jchai | 15:04 | |
| alcabrera | good morning. :) | 15:05 |
| jeffrey4l | good evening ;p | 15:06 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Cleanup technical debt https://review.openstack.org/121656 | 15:08 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Use queue_creation to send queue's metadata https://review.openstack.org/121564 | 15:08 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Respect queue's lazy creation https://review.openstack.org/121565 | 15:08 |
| openstackgerrit | Flavio Percoco proposed a change to openstack/python-zaqarclient: Always pass a project_id https://review.openstack.org/121596 | 15:08 |
| vkmc | ok, I'm lost with the new ui | 15:08 |
| * vkmc goes back to the old ui | 15:08 | |
| flaper87 | vkmc: replied to all your comments :D | 15:09 |
| flaper87 | and addressed them | 15:09 |
| vkmc | flaper87, yup I saw them | 15:10 |
| vkmc | hi alcabrera! :) | 15:13 |
| alcabrera | heya vkmc. :) | 15:14 |
| vkmc | flaper87, thaaaaaanks for being so patient :p | 15:14 |
| flaper87 | vkmc: thank YOU for being patient and reviewing all those patches | 15:14 |
| vkmc | oh we deserve tons of poptarts <3 | 15:15 |
| alcabrera | ^ | 15:16 |
| * alcabrera hands out poptarts, tons of them | 15:16 | |
| *** jeffrey4l_ has joined #openstack-zaqar | 15:17 | |
| vkmc | yaaaay | 15:17 |
| vkmc | flaper87, can you rebase this? https://review.openstack.org/#/c/121478/1 | 15:17 |
| flaper87 | vkmc: yup, I wil | 15:18 |
| *** jeffrey4l has quit IRC | 15:20 | |
| vkmc | :) | 15:23 |
| vkmc | alcabrera, if you have a moment later, can you review this one? https://review.openstack.org/#/c/121424/ | 15:23 |
| alcabrera | sure thing, vkmc! | 15:24 |
| vkmc | thx! | 15:24 |
| flaper87 | kgriffs: if you've a chance, could you go through the client patches again? | 15:24 |
| flaper87 | malini: ^ | 15:24 |
| alcabrera | flaper87, vkmc: all word-smithed on #121424 (-1) | 15:28 |
| vkmc | thx alcabrera | 15:31 |
| alcabrera | np vkmc. :) | 15:31 |
| kgriffs | flaper87: I'll see what I can do. But I won't lie. I am behind on some other work that I put aside to do benchmarking for Zaqar over the last couple days. | 15:34 |
| * flaper87 is back | 15:34 | |
| flaper87 | kgriffs: sure, no worries! | 15:34 |
| flaper87 | I think I'll go ahead and ninja-approve some patches | 15:35 |
| flaper87 | unless malini has time to go through them now | 15:35 |
| malini | flaper87: all the client patches? | 15:35 |
| malini | on it now | 15:36 |
| flaper87 | malini: awesome, thanks | 15:36 |
| flaper87 | malini: if there are concerns, feel free to ping me directly | 15:36 |
| flaper87 | sorry for the rush, I just don't want to miss the 18th deadline | 15:36 |
| malini | ok | 15:36 |
| flaper87 | :D | 15:36 |
| *** jchai is now known as jchai_afk | 15:39 | |
| malini | vkmc, flwang, kgriffs: Can you take a look at this one https://review.openstack.org/#/c/121596/ if you have a minute? | 15:42 |
| vkmc | malini, sure thing | 15:44 |
| malini | & this https://review.openstack.org/#/c/121596/ :) | 15:44 |
| malini | they are the same :D | 15:44 |
| kgriffs | flaper87: re rally | 15:44 |
| kgriffs | there are some sorts of tests/graphs that I don't see it reporting right now | 15:44 |
| kgriffs | I think we can contribute to make those happen | 15:44 |
| kgriffs | malini: will you be able to at the rally hangout on thur? | 15:45 |
| kgriffs | I have daddy duty @ home turns out | 15:45 |
| flaper87 | boris-42: there? | 15:46 |
| malini | kgriffs: sure..I just need some background on what/why & also when | 15:46 |
| kgriffs | ok | 15:46 |
| flaper87 | Did you read my messages w.r.t the Rally hangout ? | 15:46 |
| flaper87 | Is tomorrow 21 UTC a good time for you ? | 15:46 |
| flaper87 | kgriffs: what graphs ? | 15:46 |
| *** ametts has joined #openstack-zaqar | 15:46 | |
| kgriffs | so here is an example of what is currently reported | 15:47 |
| kgriffs | http://logs.openstack.org/76/110876/2/check/gate-rally-dsvm-glance/76dd628/rally-plot/results.html.gz | 15:47 |
| kgriffs | maybe I'm just not understanding the charts | 15:47 |
| kgriffs | what does "iteration" mean? | 15:48 |
| flaper87 | number of calls ? | 15:48 |
| kgriffs | I don't understand what "order number of method's call" is | 15:48 |
| flaper87 | mmh | 15:48 |
| flaper87 | boris-42: ^ | 15:48 |
| flaper87 | :D | 15:48 |
| kgriffs | maybe rally can already do this, but I would like the ability to say | 15:49 |
| kgriffs | we want to track X against increasing load | 15:50 |
| malini | flaper87: I think I am done with all the client patches | 15:50 |
| malini | let me know if I missed anything | 15:50 |
| kgriffs | where X is (latency, throughput, stdev, min, max, etc.) | 15:50 |
| kgriffs | and then I would like to say | 15:50 |
| kgriffs | let's now run all that again and increase the cluster size, then compare side-by-side to see how things scale | 15:51 |
| flaper87 | malini: thanks a lot | 15:51 |
| malini | thank you! | 15:51 |
| flaper87 | kgriffs: oh mmh | 15:52 |
| kgriffs | finally, you might show a histogram of per-request latency for all clients | 15:52 |
| flaper87 | I don't think it can do it now | 15:52 |
| flaper87 | AFAIU all that is in the works | 15:52 |
| flaper87 | but we definitely need i | 15:52 |
| flaper87 | it | 15:52 |
| kgriffs | ok | 15:52 |
| kgriffs | oh, also it would be great to graph that against CPU and disk I/O on the cluster nodes | 15:53 |
| flaper87 | vkmc: you're missing this one: https://review.openstack.org/#/c/121596/ | 15:53 |
| flaper87 | :D | 15:53 |
| * kgriffs wants the moon | 15:53 | |
| *** prashanthr_ has joined #openstack-zaqar | 15:53 | |
| kgriffs | final thought | 15:53 |
| kgriffs | idk what the gates run on now | 15:53 |
| flaper87 | those are important metrics for us. Does rally report that already? | 15:53 |
| *** jeffrey4l__ has joined #openstack-zaqar | 15:53 | |
| flaper87 | i mean, even if not graphed, does it have info about CPU usage? | 15:53 |
| kgriffs | but perhaps we can get Rackspace to donate some onmetal boxes for a weekly large-scale test. this would happen out-of-band (not gated on each patch) | 15:54 |
| vkmc | flaper87, no I don't :P | 15:54 |
| kgriffs | flaper87: idk about cpu usage. you'd have to use munin or something | 15:54 |
| kgriffs | statsd | 15:54 |
| kgriffs | whatever you like | 15:54 |
| flaper87 | vkmc: gerrit says you reviewed it at " | 15:54 |
| flaper87 | 5:54 PM" | 15:54 |
| flaper87 | :P | 15:54 |
| vkmc | flaper87, gerrit is lying | 15:54 |
| vkmc | flaper87, I reviewed it at 12.55PM | 15:54 |
| flaper87 | oh ok, me shuts gerrit down | 15:55 |
| flaper87 | kgriffs: yeah, I'll write this down and talk to boris | 15:55 |
| kgriffs | maybe the swift folks could lend us a hand as well, since I have to think they care about performance at least as much as we do. :) | 15:55 |
| * vkmc plays the timezone card | 15:55 | |
| kgriffs | flaper87: so maybe you can talk to notmyname as well | 15:55 |
| flaper87 | I'm pretty sure you just woke him up | 15:56 |
| flaper87 | :D | 15:56 |
| kgriffs | lol | 15:56 |
| flaper87 | you mentioned his nick and his baby in less than 10s | 15:56 |
| *** jeffrey4l_ has quit IRC | 15:56 | |
| flaper87 | ok, I gtg for a bit guys | 15:56 |
| flaper87 | brb | 15:56 |
| kgriffs | kk | 15:56 |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/zaqar: Add genconfig to tox.ini https://review.openstack.org/122190 | 15:56 |
| *** ametts has quit IRC | 15:57 | |
| *** jeffrey4l__ has quit IRC | 16:02 | |
| *** ametts has joined #openstack-zaqar | 16:08 | |
| *** prashanthr_1 has joined #openstack-zaqar | 16:20 | |
| prashanthr_1 | Gooood morning :) | 16:20 |
| *** jchai_afk is now known as jchai | 16:21 | |
| *** jeffrey4l__ has joined #openstack-zaqar | 16:22 | |
| jeffrey4l__ | flaper87, my patch https://review.openstack.org/#/c/122093/ does three thing. 1) fix the redis opt typo in setup.cfg 2) add the redis to oslo.config.generator.rc 3) fix the reconnect_sleep default value type. So the better way is split this commit into three or just describe it in commit summary? | 16:22 |
| prashanthr_1 | kgriffs: Good morning :) I just had a small question about https://bugs.launchpad.net/zaqar/+bug/1363811 | 16:22 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Add support for `pop` implemented in v1.1 https://review.openstack.org/120155 | 16:22 |
| *** jeffrey4l__ is now known as jeffrey4l | 16:24 | |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Rename `shard` to `pool` https://review.openstack.org/121494 | 16:24 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Add support for flavor to the client library https://review.openstack.org/121534 | 16:26 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Fix functional tests https://review.openstack.org/121563 | 16:27 |
| kgriffs | prashanthr_1: sorry, gotta run - be back in an hour or so (lunch group) | 16:30 |
| prashanthr_1 | Sure kgriffs :) | 16:31 |
| *** kgriffs is now known as kgriffs|afk | 16:31 | |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Use queue_creation to send queue's metadata https://review.openstack.org/121564 | 16:32 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Respect queue's lazy creation https://review.openstack.org/121565 | 16:33 |
| *** jeffrey4l_ has joined #openstack-zaqar | 16:33 | |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/zaqar: Add the redis opts to the generate sample shell script https://review.openstack.org/122093 | 16:34 |
| openstackgerrit | Jeffrey Zhang proposed a change to openstack/zaqar: Add genconfig to tox.ini https://review.openstack.org/122190 | 16:34 |
| *** sgotliv has quit IRC | 16:37 | |
| *** jeffrey4l has quit IRC | 16:37 | |
| boris-42 | flaper87 hi there | 16:38 |
| boris-42 | kgriffs|afk hi there | 16:38 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Always pass a project_id https://review.openstack.org/121596 | 16:40 |
| openstackgerrit | A change was merged to openstack/python-zaqarclient: Cleanup technical debt https://review.openstack.org/121656 | 16:41 |
| *** jeffrey4l_ has quit IRC | 16:46 | |
| notmyname | kgriffs|afk: flaper87: yo. I wasn't online at all yesterday and just got through my 401 unread emails. now going through IRC buffers | 16:56 |
| notmyname | kgriffs|afk: flaper87: sorry the vote went the way it did yesterday (from what I heard) | 16:57 |
| notmyname | kgriffs|afk: flaper87: I see (recently in my buffer) that you were talking about me wrt perf in zaqar. let me know what you need | 16:57 |
| *** prashanthr_1 has quit IRC | 17:00 | |
| *** ametts has quit IRC | 17:00 | |
| *** amitgandhinz has quit IRC | 17:00 | |
| *** fifieldt_ has quit IRC | 17:00 | |
| *** prashanthr_1 has joined #openstack-zaqar | 17:00 | |
| *** ametts has joined #openstack-zaqar | 17:00 | |
| *** amitgandhinz has joined #openstack-zaqar | 17:00 | |
| *** fifieldt_ has joined #openstack-zaqar | 17:00 | |
| *** echevemaster has quit IRC | 17:00 | |
| *** kgriffs|afk is now known as kgriffs | 17:01 | |
| boris-42 | flaper87 kgriffs so guys | 17:01 |
| *** echevemaster has joined #openstack-zaqar | 17:01 | |
| boris-42 | flaper87 kgriffs "iterations" means order number of call of method that is benchmark | 17:02 |
| boris-42 | flaper87 kgriffs we are creating load by calling multiple times the same method (that is called benchmark scenarios) | 17:02 |
| boris-42 | flaper87 kgriffs in graphs we have duration / iteration | 17:03 |
| *** kgriffs is now known as kgriffs|afk | 17:11 | |
| *** sgotliv has joined #openstack-zaqar | 17:46 | |
| *** sgotliv has quit IRC | 17:51 | |
| *** sgotliv has joined #openstack-zaqar | 17:53 | |
| *** sgotliv has quit IRC | 17:53 | |
| *** sgotliv has joined #openstack-zaqar | 17:53 | |
| *** kgriffs|afk is now known as kgriffs | 18:02 | |
| *** kgriffs is now known as kgriffs|afk | 18:12 | |
| *** kgriffs|afk is now known as kgriffs | 18:24 | |
| *** AAzza is now known as AAzza_afk | 18:28 | |
| *** AAzza_afk is now known as AAzza | 18:46 | |
| *** kgriffs is now known as kgriffs|afk | 18:59 | |
| *** prashanthr_ has quit IRC | 19:09 | |
| *** jchai is now known as jchai_afk | 19:11 | |
| *** jchai_afk is now known as jchai | 19:18 | |
| *** AAzza has left #openstack-zaqar | 19:28 | |
| *** malini has quit IRC | 19:31 | |
| *** jogo has joined #openstack-zaqar | 19:32 | |
| jogo | I have a random question | 19:36 |
| vkmc | jogo, shoot :) | 19:37 |
| jogo | how does zaqar compare to the properties that SQS has | 19:38 |
| jogo | http://docs.aws.amazon.com/AWSSimpleQueueService/latest/SQSDeveloperGuide/DistributedQueues.html | 19:38 |
| jogo | no guarantee of in order delivery | 19:39 |
| jogo | at least once delivery | 19:39 |
| jogo | and message sample | 19:40 |
| jogo | vkmc: ^ | 19:40 |
| vkmc | jogo, as in SQS message order is not guaranteed | 19:41 |
| vkmc | sequence information is required | 19:41 |
| jogo | vkmc: is that true just for your API specs or do any existing backends have that behavior? | 19:42 |
| vkmc | jogo, for the API specs | 19:42 |
| jogo | vkmc: so no backends have that behavior? | 19:43 |
| vkmc | jogo, AFAIK, no | 19:43 |
| jogo | ok, what about the other two properties | 19:43 |
| vkmc | I might be wrong though | 19:43 |
| jogo | vkmc: anyone else who may know? | 19:44 |
| vkmc | kgriffs|afk, malini|afk, alcabrera around? ^ | 19:45 |
| alcabrera | I'm here. | 19:45 |
| alcabrera | I remember there was a time when zaqar guaranteed FIFO for the mongodb backend | 19:46 |
| alcabrera | I'm not sure if this still holds | 19:46 |
| alcabrera | FIFO itself guaranteeing things like: order is preserved, once-and-only-once, at-most-once | 19:46 |
| alcabrera | jogo: ^ | 19:46 |
| vkmc | I'm not sure if it also true for the Redis driver | 19:47 |
| alcabrera | I'm not sure, either | 19:47 |
| vkmc | flaper87, ^ | 19:47 |
| jogo | what about the message sample property? | 19:48 |
| jogo | vkmc alcabrera: so I ask because as far as I can tell, those properties are what helps amazon SQS scale really well | 19:48 |
| jogo | and if zaqar is trying to cover the SQS use case then I would assume it would hold the same properties as well | 19:49 |
| alcabrera | message sample... let's see... | 19:49 |
| alcabrera | interesting | 19:50 |
| alcabrera | jogo: by design, zaqar gives you all messages avaialble (paginated) | 19:50 |
| alcabrera | so it is different here | 19:50 |
| vkmc | jogo, of course, makes sense :) | 19:51 |
| alcabrera | it seems like SQS shards messages across storage differently | 19:51 |
| jogo | alcabrera: yeah, how does zaqar shard? | 19:51 |
| alcabrera | whereas zaqar instead shards queues | 19:51 |
| alcabrera | different queues may live on different storage nodes | 19:51 |
| jogo | alcabrera: that sounds like individual queues are limited in size etc | 19:51 |
| jogo | and can be hurt by noisy neighbor syndrom etc | 19:52 |
| alcabrera | possibly | 19:52 |
| alcabrera | yeah, there's definitely capacity problems if one queue gets too big. :) | 19:52 |
| jogo | alcabrera: yup | 19:53 |
| flwang1 | flaper87: I'm reading your mail | 19:53 |
| jogo | alcabrera: so where/how can I get further information on this | 19:53 |
| alcabrera | jogo: check out the wiki: https://wiki.openstack.org/wiki/Zaqar, and feel free to ask more questions in here | 19:54 |
| jogo | alcabrera: hmm, I am thinking of posting this to the ML so flaper87 and others can chime in | 19:57 |
| *** flwang1 has quit IRC | 19:57 | |
| jogo | alcabrera: as these sound like questions that the wiki should answer | 19:58 |
| alcabrera | go for it, jogo. it's a good way to clarify these sorts of things, and I imagine many others are also curious about these properties of zaqar. | 19:58 |
| alcabrera | with regards to asking on the ML | 19:58 |
| jogo | alcabrera: yeah, and I wanted to get a head start and getting answers to these questions well before the end of next cycle | 19:58 |
| alcabrera | awesome! :) | 19:59 |
| vkmc | the long polling property was planned for K | 20:00 |
| vkmc | https://blueprints.launchpad.net/zaqar/+spec/long-poll | 20:00 |
| alcabrera | thanks vkmc! | 20:01 |
| vkmc | now wrt shared queues, I remember there was some plans too | 20:02 |
| *** amitgandhinz has quit IRC | 20:16 | |
| *** amitgandhinz has joined #openstack-zaqar | 20:16 | |
| *** amitgandhinz has quit IRC | 20:21 | |
| jogo | alcabrera vkmc: thanks, I started a ML thread | 20:36 |
| alcabrera | thanks, jogo. :) | 20:36 |
| jogo | http://lists.openstack.org/pipermail/openstack-dev/2014-September/046271.html | 20:36 |
| vkmc | jogo, thanks! | 20:37 |
| *** ametts has quit IRC | 20:40 | |
| *** ametts has joined #openstack-zaqar | 20:59 | |
| *** malini has joined #openstack-zaqar | 21:01 | |
| flwang | jogo: thanks for the fantastic summary, but seems I can't see the image in you mail | 21:13 |
| jogo | flwang: I cut and pasted the middle part from http://docs.aws.amazon.com/AWSSimpleQueueService/latest/SQSDeveloperGuide/DistributedQueues.html (link at the bottom) | 21:14 |
| flwang | awesome | 21:15 |
| * flaper87 back | 21:21 | |
| flaper87 | jogo: just read the backlog | 21:21 |
| flaper87 | jogo: did you read this? https://wiki.openstack.org/wiki/Zaqar/Frequently_asked_questions#How_does_Zaqar_compare_to_AWS_.28SQS.2FSNS.29.3F | 21:21 |
| flaper87 | is not very informative | 21:21 |
| flaper87 | but it has something | 21:21 |
| flaper87 | I'm sure we have a more detailed page | 21:21 |
| flaper87 | but I can't find it | 21:21 |
| flaper87 | fucking wiki | 21:21 |
| flaper87 | jogo: just read your email, I'll likely reply tomorrow but thanks for raising those questions | 21:22 |
| *** mpanetta has quit IRC | 21:23 | |
| *** sriram has quit IRC | 21:26 | |
| openstackgerrit | Victoria MartÃnez de la Cruz proposed a change to openstack/zaqar: Added checks enforcements rule to the developers manual https://review.openstack.org/121424 | 21:28 |
| flaper87 | boris-42: there? | 21:29 |
| flaper87 | vkmc: flwang malini around? | 21:29 |
| vkmc | flaper87, here | 21:29 |
| flaper87 | and kgriffs|afk | 21:29 |
| flaper87 | just a quick and hopefully not long request | 21:29 |
| flwang | flaper87: ~~~ | 21:30 |
| flwang | listening | 21:30 |
| flaper87 | I'd like us to start working on the integration with other projects | 21:30 |
| flwang | awesome | 21:30 |
| vkmc | +1000 | 21:30 |
| flaper87 | 2 projects that I'd like us to target first are: Heat and Horizon | 21:30 |
| flaper87 | We've collected some use cases here: https://etherpad.openstack.org/p/zaqar-integrated-projects-use-cases | 21:30 |
| flaper87 | I'll send an email tomorrow to start the discussion with people in those teams | 21:30 |
| flaper87 | however, before sending that email, I'd like us to think a bit on how we can help them | 21:31 |
| flwang | I have mentioned that with kgriffs|afk about the integration | 21:31 |
| flaper87 | and what we need to do so | 21:31 |
| flwang | and IMHO, it can prompt Zaqar a lot | 21:31 |
| flaper87 | so, it'd be really nice if you could take 15 mins and add comments to those etherpads | 21:31 |
| flaper87 | for example: Would you use prod/con or pub/sub for each use-case? | 21:32 |
| flaper87 | Would you recommend separate queues or just 1 ? | 21:32 |
| flaper87 | etc | 21:32 |
| vkmc | I can take a look to the Horizon ones | 21:32 |
| vkmc | it has been a while since I don't contribute to Horizon, but I'd truly like to go back to it | 21:32 |
| flaper87 | there's no much we can add without interacting with each team but lets at least add something that can help starting the discussion | 21:32 |
| flaper87 | vkmc: awesome | 21:32 |
| vkmc | and I know some Horizon folks, so I can ask them | 21:33 |
| vkmc | of course that the core questions would go to the ml | 21:33 |
| flwang | I can help on the heat+zaqar+ceilometer | 21:33 |
| flaper87 | flwang: lets focus on heat and horizon for now | 21:34 |
| flwang | I think the heat's painpoint is how the ceilometer trigger the auto scale | 21:34 |
| flaper87 | it'd be unrealistic to try to help them all at once | 21:34 |
| flwang | fine | 21:34 |
| *** malini has quit IRC | 21:34 | |
| vkmc | true that | 21:34 |
| flaper87 | we don't need to cover all the use cases, we can work on those in a per-milestone basis | 21:35 |
| *** jchai has quit IRC | 21:35 | |
| vkmc | :) | 21:35 |
| vkmc | flaper87, I'd also like to chat about the features jogo mentioned SQS has | 21:36 |
| vkmc | and add documentation for them if we are providing it, or discuss about the possibility of adding them, if we are not | 21:37 |
| vkmc | bbl | 21:39 |
| flaper87 | we're providing it or well, we provide something better | 21:40 |
| flaper87 | :D | 21:40 |
| flaper87 | ok, we can talk about it later | 21:40 |
| flaper87 | when you're back | 21:40 |
| *** alcabrera is now known as alcabrera|afk | 21:52 | |
| vkmc | back | 21:54 |
| * vkmc wanna know | 21:55 | |
| vkmc | :D | 21:55 |
| vkmc | its too late for you though, we can chat tomorrow | 21:55 |
| flaper87 | vkmc: so | 21:59 |
| flaper87 | Zaqar guarantees FIFO which is something SQS, as stated in that link, doesn't do | 22:00 |
| flaper87 | FIFO, however, depends on the backend driver | 22:00 |
| flaper87 | if the backend storage can't guarantee FIFO then there's nothing Zaqar can do there | 22:00 |
| flaper87 | Message distribution happens at a storage level, the storage itself will replicate the data and distribute it | 22:01 |
| flaper87 | Just as for FIFO, the smarter the backend is with data distribution and replication, the more benefits the user will have | 22:01 |
| flaper87 | With mongodb it'd be easy to have a per-region distribution | 22:01 |
| flaper87 | with redis that's not really possible | 22:01 |
| jogo | flaper87: sounds good | 22:02 |
| flaper87 | Messages are not distributed across multiple storage nodes, zaqar talks to 1 pool and the pool is responsible for doing the remaining job | 22:02 |
| vkmc | great | 22:02 |
| flaper87 | jogo: you've a preview :D | 22:02 |
| * flaper87 tries to stay away from work when is midnight but vkmc keeps forcing him to get back to it | 22:03 | |
| flaper87 | :P | 22:03 |
| *** flwang1 has joined #openstack-zaqar | 22:03 | |
| flaper87 | (jooooooooooooooooooke) | 22:03 |
| vkmc | oh I told you we can chat tomorrow | 22:03 |
| vkmc | although I appreciate you are letting me know now | 22:03 |
| flaper87 | vkmc: i know I know | 22:03 |
| jogo | flaper87: I am concerned about the implications of FIFO on scaling etc | 22:03 |
| vkmc | :D | 22:03 |
| jogo | flaper87: and how sharding is by queue and not messages | 22:04 |
| jogo | flaper87: anyway I look forward to seeing your email | 22:04 |
| vkmc | we should document this in the how we compare to AWS section, I wasn't aware of mongo FIFO guarantee | 22:04 |
| flaper87 | jogo: re FIFO, that's a fair concern. Depending on the storage it may have different costs | 22:04 |
| flaper87 | jogo: and re sharding, we have thought about making the sharding key configurable. It really depends on the consumption pattern | 22:05 |
| flaper87 | depending on whether the queue is write heavy or read-heavy it may be better to keep messages together in 1 shard or distributed in several | 22:06 |
| flaper87 | I'll explain this thoughts better in the email | 22:06 |
| jogo | flaper87: kk | 22:06 |
| flaper87 | these* | 22:06 |
| vkmc | makes sense | 22:06 |
| jogo | flaper87: w.r.t. sharding model IMHO zaqar should be able to automatically select which one makes the most sense | 22:08 |
| jogo | and not add yet another config option | 22:08 |
| flaper87 | jogo: now that we have flavors, we could do that in a more reliable way. there are some other related issues. For example, we're not using 1 db per queue because there's a chance we could hit mongodb's namespaces limit. This means that the sharding key selected will apply to all queues stored in the database `N` | 22:11 |
| flaper87 | jogo: TBH, there are still some experiments left to do in sharding environments | 22:12 |
| flaper87 | right now we prefer to recomend using replica sets | 22:12 |
| flaper87 | which are safer and also easier to scale at the cost of requiring more space on disk | 22:12 |
| flaper87 | (due to replication) | 22:13 |
| flaper87 | kk, gtg now :) | 22:15 |
| flaper87 | have a great rest of the day y'all | 22:15 |
| *** flaper87 is now known as flaper87|afk | 22:16 | |
| vkmc | thanks flaper87|afk, enjoy the rest of your evening o/ | 22:16 |
| *** sgotliv has quit IRC | 22:21 | |
| *** ametts has quit IRC | 23:08 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
