| *** ametts has quit IRC | 00:03 | |
| *** echevemaster has quit IRC | 00:16 | |
| *** amitgandhinz has joined #openstack-zaqar | 00:43 | |
| *** amitgandhinz has quit IRC | 00:48 | |
| *** achanda has joined #openstack-zaqar | 00:50 | |
| *** achanda has quit IRC | 00:55 | |
| *** achanda has joined #openstack-zaqar | 00:56 | |
| *** mpanetta has joined #openstack-zaqar | 00:57 | |
| *** mpanetta has joined #openstack-zaqar | 00:58 | |
| *** achanda has quit IRC | 01:01 | |
| *** kgriffs is now known as kgriffs|afk | 01:04 | |
| *** kgriffs|afk is now known as kgriffs | 01:05 | |
| *** JAHoagie has quit IRC | 01:07 | |
| *** kgriffs is now known as kgriffs|afk | 01:15 | |
| *** amalagon has quit IRC | 01:21 | |
| *** amalagon has joined #openstack-zaqar | 01:22 | |
| *** amalagon has quit IRC | 01:26 | |
| *** cpallares has quit IRC | 01:53 | |
| *** amitgandhinz has joined #openstack-zaqar | 01:53 | |
| *** amitgandhinz has quit IRC | 01:54 | |
| *** amitgandhinz has joined #openstack-zaqar | 01:54 | |
| *** exploreshaifali has quit IRC | 02:06 | |
| *** amitgandhinz has quit IRC | 02:13 | |
| *** mpanetta has quit IRC | 02:28 | |
| *** amitgandhinz has joined #openstack-zaqar | 02:31 | |
| *** vkmc has quit IRC | 03:04 | |
| *** amitgandhinz has quit IRC | 03:07 | |
| *** amalagon has joined #openstack-zaqar | 04:27 | |
| *** flwang1 has quit IRC | 04:35 | |
| *** JAHoagie has joined #openstack-zaqar | 04:58 | |
| *** JAHoagie has quit IRC | 05:51 | |
| *** reed has quit IRC | 06:54 | |
| openstackgerrit | Zhi Yan Liu proposed openstack/zaqar: Integrate OSprofiler with Zaqar https://review.openstack.org/141356 | 08:34 |
|---|---|---|
| *** bradjones has quit IRC | 11:28 | |
| *** bradjones_ has joined #openstack-zaqar | 11:29 | |
| *** bradjones_ is now known as bradjones | 11:29 | |
| *** bradjones has joined #openstack-zaqar | 11:29 | |
| *** exploreshaifali has joined #openstack-zaqar | 11:29 | |
| *** vkmc has joined #openstack-zaqar | 11:46 | |
| *** vkmc has quit IRC | 11:46 | |
| *** vkmc has joined #openstack-zaqar | 11:46 | |
| vkmc | morning o/ | 11:48 |
| exploreshaifali | vkmc: good morning \o | 11:51 |
| exploreshaifali | whats going on? | 11:51 |
| vkmc | exploreshaifali, hey there! | 11:52 |
| vkmc | all good and you? | 11:52 |
| exploreshaifali | yeah good me too :) | 11:53 |
| vkmc | could you make the server run? | 11:54 |
| exploreshaifali | no.... tbh I don't know what should be done next | 11:54 |
| exploreshaifali | I tried few guesses but all were worthless | 11:55 |
| exploreshaifali | vkmc: should I try to change the method name and replace everywhere | 11:55 |
| exploreshaifali | and then try to run server | 11:56 |
| exploreshaifali | method name--something else than queue_controller | 11:56 |
| exploreshaifali | but I think there should be some other problem, as getting error QueueController object is not callable | 11:57 |
| exploreshaifali | flaper87: around? | 12:12 |
| vkmc | is always a good idea to do something else when you are stucked with something | 12:20 |
| vkmc | also, we have to fix the debugger | 12:21 |
| vkmc | so you have an extra tool | 12:21 |
| vkmc | right now I have to do some stuff for work, otherwise I'd help you to debug | 12:22 |
| exploreshaifali | yeah okay!! | 12:22 |
| exploreshaifali | no worries..... I will keep trying :) | 12:23 |
| vkmc | would you like to tackle the problem with testtools? | 12:23 |
| vkmc | that requires researching a bit | 12:23 |
| exploreshaifali | yea sure | 12:23 |
| *** shibanis has joined #openstack-zaqar | 13:19 | |
| *** shibanis has left #openstack-zaqar | 13:19 | |
| vkmc | pcaruana, o/ | 13:21 |
| vkmc | exploreshaifali, btw, did you moved *all* appearances of queue_controller in DataDriver to ControlDriver? | 13:35 |
| vkmc | exploreshaifali, for instance, in pooling.py? | 13:35 |
| *** flwang1 has joined #openstack-zaqar | 14:00 | |
| flwang1 | flaper87: ping | 14:10 |
| *** sriram has joined #openstack-zaqar | 14:14 | |
| *** kragniz has quit IRC | 14:22 | |
| *** kragniz has joined #openstack-zaqar | 14:22 | |
| *** flwang1 has quit IRC | 14:31 | |
| *** JAHoagie has joined #openstack-zaqar | 14:33 | |
| *** shaifali_ has joined #openstack-zaqar | 14:41 | |
| *** mpanetta has joined #openstack-zaqar | 14:42 | |
| *** shaifali_ has quit IRC | 14:44 | |
| *** shaifali_ has joined #openstack-zaqar | 14:45 | |
| *** shaifali_ has quit IRC | 14:45 | |
| *** mpanetta has quit IRC | 14:51 | |
| *** mpanetta has joined #openstack-zaqar | 14:51 | |
| *** vipul has quit IRC | 14:52 | |
| *** vipul has joined #openstack-zaqar | 14:52 | |
| *** kgriffs|afk has quit IRC | 14:58 | |
| *** kgriffs|afk has joined #openstack-zaqar | 14:58 | |
| exploreshaifali | vkmc: sorry I was out for some time | 14:58 |
| vkmc | np | 14:58 |
| *** kgriffs|afk is now known as kgriffs | 14:58 | |
| exploreshaifali | I have actually not moved queue_controller from DataDrive to ControlDrive | 14:59 |
| exploreshaifali | have added queue_controller to ControllerDrive | 14:59 |
| exploreshaifali | and modified the defination of queue_controller in DataDrive | 14:59 |
| exploreshaifali | such that now it dose not use QueueController directly | 14:59 |
| vkmc | move it directly | 15:00 |
| exploreshaifali | now it use QueueController instance(or create and delete methods) trough ControllDriver | 15:00 |
| exploreshaifali | vkmc: you want me to move whole queue_controller to ControlDriver? | 15:01 |
| vkmc | exploreshaifali, isn't that the desired change? | 15:02 |
| *** cpallares has joined #openstack-zaqar | 15:02 | |
| exploreshaifali | though main task is to move queue_controller from data to control plane...but if we look in deep main thing is that we should not use use QueueController directly in data driver | 15:03 |
| exploreshaifali | as we want data and control planes to be separated | 15:04 |
| vkmc | so... remove it | 15:04 |
| vkmc | for v1_1 queue creation is lazy | 15:04 |
| vkmc | (a queue is created when a message is sent) | 15:04 |
| vkmc | I really need to see a detailed spec of thigs change :| | 15:05 |
| vkmc | cpallares, wasaaaaaaaaaaaaap | 15:05 |
| exploreshaifali | vkmc: yea but does new queue is created for each new message we send? | 15:05 |
| cpallares | hey vkmc! | 15:05 |
| cpallares | hi exploreshaifali! | 15:05 |
| exploreshaifali | cpallares: heeeeeeeeeeelllllllllloooooooo :D | 15:05 |
| vkmc | exploreshaifali, suppose you send a message to queue 'fizbit' | 15:05 |
| vkmc | if the queue doesn't exist, 'fizbit' is created | 15:06 |
| vkmc | if it exists, then nothing happens | 15:06 |
| vkmc | although that is not the case for v1_0 | 15:06 |
| vkmc | so we might have to make a control for that | 15:06 |
| *** ametts has joined #openstack-zaqar | 15:07 | |
| exploreshaifali | vkmc: yeah that is correct | 15:07 |
| exploreshaifali | but I think method name(queue_controller) dosen't matter | 15:07 |
| exploreshaifali | what does that method do is what we need to change | 15:08 |
| exploreshaifali | though I may be wrong | 15:08 |
| vkmc | come again? | 15:09 |
| exploreshaifali | vkmc: see https://github.com/openstack/zaqar/blob/master/zaqar/storage/base.py#L212 | 15:09 |
| exploreshaifali | that is abstract method | 15:09 |
| exploreshaifali | used at line 154 | 15:09 |
| vkmc | yep | 15:09 |
| exploreshaifali | and one more time below it | 15:10 |
| vkmc | that is what I say | 15:10 |
| vkmc | remove those lines | 15:10 |
| exploreshaifali | ohhhh....... you are asking to remove those line? | 15:10 |
| vkmc | yeap | 15:10 |
| vkmc | because queue creation is lazy | 15:10 |
| vkmc | buuuuuuuuuut | 15:11 |
| vkmc | we might need to make an special case for v1_0 | 15:11 |
| vkmc | but for now, just to make your server up and running | 15:11 |
| vkmc | try that | 15:11 |
| exploreshaifali | okay | 15:11 |
| exploreshaifali | but let me explain what I was trying to do | 15:11 |
| exploreshaifali | I didn't though to remove those line | 15:12 |
| exploreshaifali | rather change the defination of overried method | 15:12 |
| vkmc | I know | 15:12 |
| exploreshaifali | hahaha | 15:13 |
| exploreshaifali | okay | 15:13 |
| exploreshaifali | so if I will remove those line for creating and deleting queues | 15:13 |
| exploreshaifali | than either I should provide some method for the same task to be done htere | 15:13 |
| exploreshaifali | *there | 15:13 |
| exploreshaifali | and those methods will be called on ControlDriver object | 15:14 |
| *** JAHoagie has quit IRC | 15:15 | |
| vkmc | nope | 15:15 |
| vkmc | just remove them | 15:15 |
| vkmc | and leave the health method where it is now | 15:16 |
| exploreshaifali | vkmc: but in that case we will not create and delete queue whenever needed | 15:17 |
| exploreshaifali | right? | 15:17 |
| vkmc | hhmm | 15:17 |
| vkmc | what do you mean by that? | 15:17 |
| exploreshaifali | if I will remove those lines and will provide no replacement for them | 15:18 |
| vkmc | then, as I mentioned, the queue required will be lazily created | 15:18 |
| vkmc | or that is what I expect | 15:18 |
| exploreshaifali | okay.... let see :) | 15:19 |
| exploreshaifali | vkmc: thanks!! | 15:19 |
| vkmc | exploreshaifali, np | 15:20 |
| vkmc | try that and let me kno | 15:20 |
| vkmc | w | 15:20 |
| exploreshaifali | yeah.....sure!! | 15:20 |
| *** amitgandhinz has joined #openstack-zaqar | 15:23 | |
| *** kgriffs has quit IRC | 15:25 | |
| *** kgriffs|afk has joined #openstack-zaqar | 15:25 | |
| *** kgriffs|afk is now known as kgriffs | 15:26 | |
| * cpallares waves at kgriffs | 15:33 | |
| * vkmc waves at * | 15:34 | |
| kgriffs | o/ | 15:37 |
| vkmc | \o | 15:37 |
| *** mpanetta has quit IRC | 15:37 | |
| *** mpanetta has joined #openstack-zaqar | 15:38 | |
| vkmc | mpanetta, yooo | 15:39 |
| *** miqui_ has joined #openstack-zaqar | 15:49 | |
| *** vkmc_ has joined #openstack-zaqar | 15:53 | |
| *** reed has joined #openstack-zaqar | 15:59 | |
| *** vkmc has quit IRC | 16:03 | |
| *** vkmc_ is now known as vkmc | 16:03 | |
| *** vkmc has quit IRC | 16:03 | |
| *** vkmc has joined #openstack-zaqar | 16:03 | |
| *** vkmc_ has joined #openstack-zaqar | 16:03 | |
| sriram | \o/ | 16:04 |
| *** JAHoagie has joined #openstack-zaqar | 16:04 | |
| * sriram frantically waves at everyone | 16:04 | |
| kragniz | ~o~ | 16:05 |
| *** vkmc_ has quit IRC | 16:05 | |
| cpallares | kragniz: o~o | 16:10 |
| vkmc | ~o/ | 16:10 |
| * vkmc waves back to sriram | 16:11 | |
| cpallares | sriram: /o\ | 16:11 |
| * cpallares salutes vkmc | 16:12 | |
| * vkmc gives cpallares some crackers | 16:13 | |
| vkmc | its lunchtime here :o | 16:13 |
| kragniz | vkmc: eat some lunch | 16:14 |
| vkmc | kragniz, oh I will | 16:15 |
| vkmc | :D | 16:15 |
| kragniz | :( ||| ) <- vkmc eating a cracker | 16:15 |
| vkmc | *crunch* *crunch* | 16:15 |
| vkmc | I'm that annoying colleague that makes weird noises when eating | 16:16 |
| kragniz | I've got a mechanical keyboard, which I think may annoy people | 16:16 |
| vkmc | lol | 16:18 |
| *** kgriffs has quit IRC | 16:26 | |
| *** kgriffs has joined #openstack-zaqar | 16:27 | |
| pcaruana | vkmc hi there, sorry forgot to change my afk status :) | 16:28 |
| vkmc | pcaruana, no problem :) | 16:29 |
| vkmc | pcaruana, saw you in #sysarmy and #pyar... so I had to say hi :D | 16:29 |
| vkmc | glad to have you here! | 16:29 |
| pcaruana | vkmc, yes keeping my .ar network ;) | 16:30 |
| vkmc | pcaruana, that's great! | 16:30 |
| vkmc | pcaruana, how long have you been working with the stack? | 16:32 |
| vkmc | I don't know many ar people involved unfortunately | 16:32 |
| mpanetta | hiya vkmc :) | 16:38 |
| vkmc | heeeeeeeey mpanetta :) | 16:38 |
| mpanetta | How goes? | 16:39 |
| vkmc | all good... trying to debug something weird with pip :| | 16:40 |
| vkmc | you? | 16:40 |
| pcaruana | vkmc, played a little between blosson and essex, rejoined interest between grizzly and havanna ,ON the other hand I am an Argie expat living in Europe for at least 5 seasons :). | 16:41 |
| mpanetta | Pip eh? Is it broken again? heh | 16:42 |
| vkmc | pcaruana, that's super cool! it has been a while then :) | 16:43 |
| vkmc | mpanetta, yeaaaah, not in Zaqar though... I'm with TripleO's DIB | 16:44 |
| vkmc | but its super weird http://paste.openstack.org/show/155762/ | 16:44 |
| * vkmc sighs | 16:44 | |
| *** exploreshaifali is now known as exploreshai|afk | 16:46 | |
| vkmc | off to lunch, brb | 16:49 |
| *** amalagon has quit IRC | 16:52 | |
| *** amalagon has joined #openstack-zaqar | 16:52 | |
| *** amalagon has quit IRC | 16:57 | |
| mpanetta | Hmm lunch sounds good | 16:59 |
| kragniz | a second lunch also sounds good | 17:00 |
| cpallares | a second lunch does sound good | 17:02 |
| cpallares | I have to eat breakfast first though | 17:02 |
| mpanetta | cpallares: Brunch ftw? | 17:03 |
| kragniz | an n+1 lunch also sounds good, where n is the current lunch number | 17:03 |
| cpallares | mpanetta: Double brunch :P | 17:03 |
| cpallares | haha | 17:03 |
| mpanetta | brunch lunch | 17:03 |
| kragniz | brulunchfast | 17:04 |
| *** nakul_cpani has joined #openstack-zaqar | 17:04 | |
| cpallares | breakbrunlunch | 17:06 |
| *** amalagon has joined #openstack-zaqar | 17:11 | |
| *** amalagon has quit IRC | 17:57 | |
| *** amalagon has joined #openstack-zaqar | 17:57 | |
| *** amitgandhinz has quit IRC | 18:11 | |
| *** amitgandhinz has joined #openstack-zaqar | 18:21 | |
| *** nakul_cpani has quit IRC | 18:46 | |
| *** reed has quit IRC | 18:49 | |
| * vkmc lurks | 18:50 | |
| *** exploreshai|afk has quit IRC | 18:52 | |
| *** nakul_cpani has joined #openstack-zaqar | 19:13 | |
| *** reed has joined #openstack-zaqar | 19:17 | |
| vkmc | nakul_cpani, how are you doing? | 19:19 |
| *** nakul_cpani has quit IRC | 19:53 | |
| *** kgriffs has quit IRC | 20:01 | |
| *** flwang1 has joined #openstack-zaqar | 20:01 | |
| *** kgriffs has joined #openstack-zaqar | 20:01 | |
| *** nakul_cpani has joined #openstack-zaqar | 20:05 | |
| *** flwang1 has quit IRC | 20:06 | |
| *** flwang1 has joined #openstack-zaqar | 20:08 | |
| flwang1 | flaper87: piiiiiiiiiiiiiiiiiiiiiing | 20:16 |
| vkmc | flwang1, hey! | 20:17 |
| flwang1 | vkmc: hi | 20:18 |
| vkmc | flwang1, how are you? I saw you uploaded some patches, I'll review those | 20:24 |
| flwang1 | vkmc: not bad | 20:31 |
| flwang1 | yep, I uploaded 2 patch set for the notification service | 20:32 |
| flwang1 | but I still run into some design issues for the pool support for notification | 20:34 |
| flwang1 | not sure if you saw the discussion between kgriffs and me | 20:35 |
| flwang1 | vkmc: our current pool implement is really depending on 'queue' | 20:35 |
| vkmc | flwang1, exactly yes | 20:36 |
| vkmc | I saw your discussion with kgriffs yesterday | 20:36 |
| vkmc | so... in the dropping queue in favor of topics spec | 20:38 |
| flwang1 | vkmc: right, so for messages, claims, queue, you have to pass in 'queue' as the parameter | 20:38 |
| vkmc | one of the reviewers pointed out that it would be a good idea to keep it in the url | 20:38 |
| flwang1 | to get the correct pool | 20:38 |
| vkmc | yeah | 20:38 |
| vkmc | that won't respect rest though | 20:38 |
| flwang1 | vkmc: yep, that's one option | 20:39 |
| flwang1 | and no matter if we want to drop queue in the future, that's another problem | 20:39 |
| flwang1 | i'd like to get feedback from core reviewers to move on | 20:40 |
| flwang1 | since it will impact the design and many other stuff | 20:40 |
| vkmc | yeah, its important that we move on with notifications | 20:40 |
| vkmc | I'm also in a similar situation with the persistent transport one | 20:40 |
| flwang1 | bad :( | 20:41 |
| vkmc | let's have a meeting tomorrow, flaper87 should be back then | 20:42 |
| flwang1 | vkmc: ok, sounds good | 20:42 |
| vkmc | in the meantime, I see that you are mentioning that if we drop the queue concept then all the operations are more complex | 20:42 |
| vkmc | so... we should discuss if dropping the queue concept is actually a good idea | 20:42 |
| kgriffs | seems like we have a couple topics here ^^^ | 20:44 |
| kgriffs | first is simply the design of the api - where/when to specify a queue name when doing CRUD on subscriptions | 20:45 |
| *** pcaruana is now known as pcaruana|afk| | 20:45 | |
| kgriffs | the second is how to map a subscription to a queue, and by extension a pool, right? | 20:45 |
| vkmc | kgriffs, exactly | 20:46 |
| kgriffs | if you know the queue name and the project ID, then you can look up the pool, right? | 20:46 |
| vkmc | right | 20:47 |
| kgriffs | ok, so next question: during which operations do you need to look up the pool when dealing with subscriptions? | 20:48 |
| vkmc | so that happens in a pooled deployment when the number of subscriptions are bigger than certain value | 20:51 |
| vkmc | flwang1, ^ | 20:51 |
| vkmc | that is what I understood at least | 20:51 |
| kgriffs | are you saying that we are sharding the subscriptions list across pools? | 20:51 |
| kgriffs | (the list itself) | 20:52 |
| flwang1 | kgriffs: yes | 20:52 |
| flwang1 | for now, one subscription will have one record in database | 20:53 |
| flwang1 | that means for a mobile app notification service, the subscription number maybe very big | 20:53 |
| flwang1 | so it's useful to sharding the subscriptions across pools, thoughts? | 20:54 |
| * kgriffs is considering | 20:54 | |
| flwang1 | based on that, for the CRUD actions, there is the 'source' field for creation, so it's ok | 20:54 |
| flwang1 | but for get, delete, update, we only have the subscription id and project id, so it would be hard to get the correct pool given we're depending 'queue' for sharding | 20:56 |
| flwang1 | maybe it's time to consider to shard based on 'project' instead of 'queue' | 20:56 |
| vkmc | I remember talking about that with flaper87 | 20:57 |
| vkmc | he redirected me to his blog :p | 20:57 |
| kgriffs | so, a few things | 20:57 |
| kgriffs | first, I want to verify our assumption that subscriptions have to be sharded | 20:57 |
| kgriffs | second, I think we should not confuse how we use pools for data plane with how we (may) want to shard control plane data | 20:58 |
| kgriffs | so, #1 | 20:58 |
| kgriffs | the mobile device scenario i suppose would mean subscriptions that use APN, etc. for the delivery | 20:59 |
| flwang1 | so you mean there is a proxy point to delivery the notifications, right? | 21:00 |
| flwang1 | that means not too much subscriptions we should concern? | 21:00 |
| kgriffs | right, you would need a notification worker that could deliver the notification using mobile push | 21:00 |
| kgriffs | but that worker (or pool of workers) would need to know all the mobile devices it was supposed to push a given message to | 21:01 |
| kgriffs | i guess we need to think about some hypothetical services/apps that would want to do this | 21:02 |
| *** JAHoagie has quit IRC | 21:03 | |
| kgriffs | what we need to figure out is if this use case is something that people would want to use Zaqar for directly. could be, but I want to be sure. | 21:03 |
| kgriffs | for the moment, however, let's assume that we do need to support this | 21:04 |
| flwang1 | kgriffs: so you believe that's the key if there will be huge subscriptions in zaqar db, right? | 21:04 |
| kgriffs | how many subscriptions would that be? 10 million? more? | 21:04 |
| flwang1 | kgriffs: i have no idea, but let's assume it's a number we should concern about the performance | 21:05 |
| *** nakul_cpani has quit IRC | 21:06 | |
| *** nakul_cpani has joined #openstack-zaqar | 21:06 | |
| kgriffs | so, I would recommend testing mongo, mysql, maybe also redis - just seed it with, say, 10 million rows containing the kind of data we need for each subscription record. Add an index to lookup based on queue name and project ID. Then do some quick and dirty query perf tests, and see how much storage space the set of records consumes. Then do the same for 100 million records. | 21:12 |
| kgriffs | That will let us know whether a single DB cluster is up to the task. | 21:13 |
| flwang1 | so you prefer to keep one db cluster and put it on management plane, is it? | 21:14 |
| flwang1 | sorry | 21:15 |
| flwang1 | i mean keep the table on mangement plance, is it | 21:15 |
| kgriffs | well, I think regardless of what we do, we should keep the subscription data separate from the messages. they will have different scaling and performance characteristics. | 21:16 |
| kgriffs | also, consider that subscriptions are mostly read-only | 21:19 |
| kgriffs | you will read them way more often than write them | 21:19 |
| flwang1 | kgriffs: good points | 21:19 |
| kgriffs | this changes how you might want to deploy them | 21:19 |
| kgriffs | I suppose the pool catalog is similar. | 21:20 |
| kgriffs | so to me, it makes sense to have a DB cluster dedicated to the control data (pool catalog, subscriptions) and then a separate N clusters AKA pools for messages | 21:20 |
| flwang1 | kgriffs: make sense for me, but it depends on the performance | 21:21 |
| kgriffs | I suspect that 100 million records wouldn't be a big deal for a single control plane cluster. You couldn't run it on a Raspberry Pi, but you wouldn't need an IBM Power box either | 21:22 |
| kgriffs | but yeah, we should check it out | 21:22 |
| flwang1 | kgriffs: yep | 21:22 |
| flwang1 | kgriffs: we can revisit this topic after flaper87 back | 21:22 |
| kgriffs | scaling a read-heavy workload can be done with mysql by reading from a pool of slaves behind HAProxy or similar | 21:23 |
| kgriffs | but it is still one overall cluster, so need for sharding | 21:23 |
| kgriffs | on mongo you could easily use built-in sharding on a single cluster | 21:23 |
| flwang1 | kgriffs: right, TBH, that's my initial design | 21:23 |
| kgriffs | and also with redis, redis-cluster would probably work well (TBD) | 21:23 |
| flwang1 | but then maybe i just think too much | 21:23 |
| kgriffs | anyway, like I said, we should actually test it out a little before jumping to conclusions | 21:24 |
| flwang1 | agree | 21:24 |
| flwang1 | I will paste above discussion into the etherpad | 21:24 |
| kgriffs | flwang1: heh, it is always good to think about handling huge numbers of records/traffic/etc. But you have to balance that with YAGNI | 21:24 |
| kgriffs | kewl | 21:24 |
| * vkmc was in read only mode | 21:25 | |
| flwang1 | https://etherpad.openstack.org/p/zaqar-notification | 21:26 |
| *** JAHoagie has joined #openstack-zaqar | 21:26 | |
| *** nakul_cpani has quit IRC | 21:47 | |
| *** sriram has quit IRC | 22:16 | |
| *** mpanetta has quit IRC | 22:19 | |
| *** kgriffs has quit IRC | 23:20 | |
| *** miqui has quit IRC | 23:20 | |
| *** notmyname has quit IRC | 23:21 | |
| *** flwang1 has quit IRC | 23:21 | |
| *** reed has quit IRC | 23:21 | |
| *** ekarlso- has quit IRC | 23:21 | |
| *** notmyname has joined #openstack-zaqar | 23:22 | |
| *** flwang1 has joined #openstack-zaqar | 23:22 | |
| *** reed has joined #openstack-zaqar | 23:22 | |
| *** ekarlso- has joined #openstack-zaqar | 23:22 | |
| *** boris-42 has quit IRC | 23:24 | |
| *** miqui has joined #openstack-zaqar | 23:25 | |
| *** boris-42 has joined #openstack-zaqar | 23:26 | |
| *** amitgandhinz has quit IRC | 23:37 | |
| *** kgriffs has joined #openstack-zaqar | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!