*** Qiming has quit IRC | 00:19 | |
*** pratikmallya has quit IRC | 00:22 | |
*** Qiming has joined #senlin | 01:12 | |
Qiming | morning | 01:17 |
---|---|---|
xuhaiwei | morning | 01:20 |
Qiming | just got some complaints about upgrading senlin | 01:21 |
Qiming | the report is from IBM internal | 01:21 |
Qiming | people upgraded senlin to the latest code, they had to recreate database, due to latest revision to db schema | 01:22 |
Qiming | and they have lost references to all objects they have created | 01:22 |
Qiming | so, I'm thinking maybe we should freeze verion 001 database schema starting mitaka-2 | 01:23 |
Qiming | any future revisions to db should go into 002, 003 etc | 01:23 |
xuhaiwei | up to now, we hardly conside backward capability things | 01:28 |
xuhaiwei | that is a problem | 01:28 |
Qiming | yes, this is the right thing to do | 01:28 |
Qiming | the only question was when to start enforcing such a rule/practice | 01:29 |
Qiming | considering that there are users start using it | 01:29 |
Qiming | I'm raising awareness for core reviewers | 01:29 |
Qiming | please be stricter when reviewing DB changes | 01:30 |
xuhaiwei | ok | 01:30 |
xuhaiwei | about the meetup agenda, I want to talk about tempest-external-plugin, it's about function test, currently it is not in the schedule, Qiming | 01:32 |
Qiming | great | 01:32 |
xuhaiwei | 20 mins would be ok I think | 01:33 |
xuhaiwei | can you reschedule it? | 01:33 |
Qiming | Looking | 01:33 |
Qiming | is it an experience sharing? | 01:33 |
xuhaiwei | no | 01:34 |
xuhaiwei | you know in tempest, some projects will make them function tests ran in tempest, but there are two many projects now, tempest only want to support core projects now | 01:35 |
Qiming | so it would be a senlin related topic, which means day 2 is better? | 01:35 |
xuhaiwei | so they bring into this method, make other projects' function test work as a plugin | 01:36 |
xuhaiwei | it's ok I think, not very big topic | 01:36 |
xuhaiwei | after this plugin worked, you can run senlin function with tempest command | 01:37 |
Qiming | check etherpad, editing | 01:38 |
Qiming | xuhaiwei, we have to define our scope | 01:57 |
xuhaiwei | if we create a stack and leave it there, why should we create it? | 01:57 |
Qiming | be very very careful when walking into anoter service's domain | 01:57 |
Qiming | senlin is about a cluster of stacks | 01:57 |
xuhaiwei | I think I am understanding the original design for stack node | 01:57 |
xuhaiwei | I am not | 01:58 |
Qiming | each stack can be a wrapper of a nova server + a cinder volume + a floating IP ... | 01:58 |
Qiming | xuhaiwei, I don't think you understand the design | 01:58 |
Qiming | you are dragging us to a very dangerous zone | 01:58 |
xuhaiwei | yes, I mean I dont understand | 01:58 |
Qiming | stack is an abstraction or a wrapper of many different kinds of resources | 01:59 |
xuhaiwei | so why senlin create a stack node? | 01:59 |
Qiming | in Heat's design, a stack can contain an embeded stack, which contains yet another level of child stacks | 01:59 |
Qiming | xuhaiwei, to avoid reinventing all resource types | 02:00 |
Qiming | think about array and struct metaphore | 02:00 |
Qiming | Heat provides you a struct, Senlin provides you an array | 02:00 |
Qiming | cluster of heat stacks <==> array of structs | 02:00 |
Qiming | we are NEVER touching the internals of a struct | 02:01 |
xuhaiwei | so we only create it? | 02:01 |
Qiming | create, update, delete, scale .... | 02:01 |
Qiming | whatever | 02:01 |
Qiming | we scale a stack as a whole | 02:02 |
Qiming | senlin will never care how many instances, how many volumes inside a stack | 02:02 |
Qiming | if your heat template doesn't have an output of the IPs of the servers created, we will never get a chance to do load-balancing | 02:03 |
xuhaiwei | but if your stack work together with other instances(not in this stack), there will confict | 02:03 |
Qiming | we cannot force users to write heat templates this way and that way | 02:03 |
Qiming | xuhaiwei, that is true | 02:04 |
Qiming | that will be true forever | 02:04 |
Qiming | just like you are running openstack | 02:04 |
Qiming | and you are also launching KVMs using qemu-kvm command line | 02:04 |
xuhaiwei | in that case, the stack auto-scaling will fail in a high opportunity | 02:04 |
Qiming | you are using Heat to do orchestration, while you are also using nova to create VMs | 02:04 |
Qiming | xuhaiwei, I don't agree | 02:05 |
Qiming | if architected properly, you can still get auto-scaling work properly | 02:05 |
Qiming | there are 1000 ways to fail | 02:05 |
Qiming | 999 are not our business | 02:05 |
xuhaiwei | just ask an common question | 02:05 |
Qiming | say you a cluster of heat stacks and you want that cluster to be autoscaled | 02:06 |
xuhaiwei | if we dont care about the instance inside a stack, we auto-scale this stack, which az will the instance go? | 02:06 |
Qiming | tell me the metrics you want to trigger this "auto" scaling | 02:06 |
Qiming | cross-az policy won't deal with heat stack | 02:07 |
Qiming | you cannot put heat stack into an AZ | 02:07 |
Qiming | cinder, if I'm remembering correctly, has its own definition of AZ | 02:08 |
Qiming | there are also questions how to do cross AZ references from one Nova server to a Cinder volume | 02:08 |
Qiming | before we have some basic functional support from nova, cinder and other services, you won't be able to put a Heat stack as a whole into an AZ | 02:09 |
xuhaiwei | I still think stack autoscaleing doesn't make sense, because you have no policy to evaluate a whole stack | 02:10 |
Qiming | well it can be done | 02:10 |
Qiming | pretty easily | 02:10 |
Qiming | you can not speak for the users | 02:10 |
Qiming | cloud monitoring is another domain | 02:11 |
Qiming | a huge one | 02:11 |
Qiming | even today, using ceilometer, you can already do a decent metric collection, aggregation, filtering and then trigger an alarm from that | 02:11 |
xuhaiwei | ok, if you want to scale out a stack, you will create an exactly same stack? | 02:12 |
Qiming | yes | 02:12 |
Qiming | same collection of resources | 02:12 |
Qiming | just like you are calling Heat twice | 02:12 |
*** elynn has joined #senlin | 02:13 | |
xuhaiwei | if the stack has only one node, it seems ok, but it has many nodes, it's dangeous | 02:13 |
Qiming | yes, if you are scaling a stack of many nodes | 02:13 |
Qiming | you will take care of the nodes you created from Heat | 02:13 |
Qiming | one way of doing that is to add metadata to your nova server resource in the stack | 02:14 |
Qiming | so that ceilometer can filter out which resource comes from which stack | 02:14 |
Qiming | then ceilometer will do the correct aggregation and evaluate whether an alarm is needed | 02:14 |
Qiming | if is not Senlin's business | 02:15 |
xuhaiwei | if a cluster contain some nova nodes and some stack nodes, the policy attached to it will still work? | 02:15 |
Qiming | we don't allow a cluster to contain nodes of different profile type | 02:16 |
xuhaiwei | sorry, I forget that | 02:16 |
xuhaiwei | a stack cluster can be attached some policies? | 02:17 |
Qiming | yes | 02:18 |
Qiming | check the policy definition | 02:18 |
Qiming | eacy policy is supposed to tell what profile types it can support | 02:18 |
Qiming | not all policies will be applicable to all profiles | 02:18 |
xuhaiwei | ok | 02:19 |
Qiming | xuhaiwei, I'm wondering if we should take some actions about this: https://review.openstack.org/#/c/156508 | 02:38 |
Qiming | or it should be a patch to SDK | 02:39 |
Qiming | don't have bandwidth to following at the moment | 02:39 |
xuhaiwei | Ok, I will read this patch | 02:40 |
xuhaiwei | we will discuss this in the API topic at the meetup | 02:41 |
Qiming | we don't have an API sesssion | 02:41 |
Qiming | yet | 02:41 |
xuhaiwei | 1-11 project status: API , just share some information? | 02:43 |
xuhaiwei | don't discuss something? | 02:43 |
Qiming | yes, it is a share | 02:43 |
Qiming | day one is about meeting with users | 02:43 |
xuhaiwei | got it | 02:43 |
Qiming | added one on day 2 | 02:44 |
Qiming | putting you as the lead | 02:44 |
xuhaiwei | ok, thanks | 02:45 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Implement node_recover in Profile https://review.openstack.org/262933 | 03:37 |
*** gongysh_ has quit IRC | 03:57 | |
*** elynn has quit IRC | 04:00 | |
*** gongysh_ has joined #senlin | 04:10 | |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Recover into Node Actions and Node Model https://review.openstack.org/264568 | 04:35 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Recover into Node Actions and Node Model https://review.openstack.org/264568 | 04:51 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Recover into Node Actions and Node Model https://review.openstack.org/264568 | 05:09 |
*** elynn has joined #senlin | 05:12 | |
*** dixiaoli has joined #senlin | 05:14 | |
*** gongysh_ has quit IRC | 05:16 | |
*** gongysh_ has joined #senlin | 05:35 | |
*** Yanyanhu has joined #senlin | 05:36 | |
Qiming | Yanyanhu | 06:34 |
Yanyanhu | hi | 06:34 |
Qiming | please jump onto this: https://review.openstack.org/264336 | 06:34 |
Yanyanhu | sure, will check it | 06:35 |
Qiming | it smells dangerous | 06:35 |
Yanyanhu | ok | 06:35 |
Yanyanhu | will confirm it | 06:36 |
Yanyanhu | Qiming, I think this change is safe for us. Actually only loadbalancer has two status related properties: provisioning_status and operating_status | 06:40 |
Yanyanhu | other lb related resources don't have 'status' attr based on Neutron API interface | 06:41 |
Yanyanhu | and we are not using this 'status' attr in our lb driver | 06:41 |
Yanyanhu | so I'm gonna +1 to that patch :) | 06:44 |
Qiming | Yanyanhu, okay | 06:46 |
Qiming | just want to make sure we are not checking 'status' fields | 06:46 |
Yanyanhu | yes, understand | 06:47 |
Qiming | If I'm remembering correctly, we WERE checking status when ever doing some operations | 06:47 |
Yanyanhu | it could influence our lb driver if so | 06:47 |
Yanyanhu | yes, everytime we do lb related operation, we will check loadbalancer's provisioning_status and operating_status | 06:47 |
Qiming | I see | 06:48 |
Yanyanhu | to decide whether the operation is done | 06:48 |
Qiming | those are the two status fields we care about | 06:48 |
Yanyanhu | yes | 06:48 |
Qiming | Yanyanhu, have you verified if Brian's patch to metadata works? | 07:09 |
Yanyanhu | haven't yet | 07:09 |
Yanyanhu | haven't read his latest change | 07:09 |
Yanyanhu | let me have a look | 07:09 |
Qiming | okay, he just proposed a new patchset | 07:09 |
Yanyanhu | yes, the old one looks good I think | 07:10 |
Yanyanhu | not sure image part is also ready. But I noticed he didn't -workflow this time | 07:10 |
Qiming | yep, I think he has tested it | 07:11 |
*** elynn has quit IRC | 07:11 | |
Yanyanhu | yes | 07:11 |
*** elynn has joined #senlin | 07:11 | |
*** elynn has quit IRC | 07:22 | |
*** elynn has joined #senlin | 07:22 | |
*** elynn has quit IRC | 07:27 | |
*** elynn has joined #senlin | 07:27 | |
*** elynn has quit IRC | 07:32 | |
*** elynn has joined #senlin | 07:33 | |
*** dixiaoli_ has joined #senlin | 07:50 | |
*** dixiaoli_ has quit IRC | 07:51 | |
*** dixiaoli has quit IRC | 07:54 | |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Remove references to oslo incubator modules https://review.openstack.org/264664 | 08:30 |
openstackgerrit | Yanyan Hu proposed openstack/senlin: Rework get_details of nova server profile https://review.openstack.org/264669 | 08:54 |
openstackgerrit | Yanyan Hu proposed openstack/senlin: [WIP]Rework policy check for built-in policy types https://review.openstack.org/264670 | 08:54 |
Yanyanhu | hi, Qiming, just propose a patch for policy_check rework. It is still in progress, but any comment is very helpful | 08:54 |
Yanyanhu | I haven't revised test cases for it | 08:55 |
Yanyanhu | BTW, https://review.openstack.org/264670 this problem in sdk is still there | 08:55 |
Yanyanhu | neutron will throw exception when accepting an unexpected argument in request body | 08:56 |
Yanyanhu | this behavior is different from Nova. Nova just ignore this kind of arguments silently | 08:57 |
Qiming | can we revise neutron resource directly? | 08:57 |
Qiming | if it is generic enough, we can propose a revision at resource level | 08:58 |
Qiming | lixinhui, there? | 08:58 |
Yanyanhu | yes, I think Brian want this kind of revision. But he also didn't have very good idea for this issue | 08:58 |
Yanyanhu | I think | 08:58 |
Yanyanhu | revising lb related resource is a possible way to address this issue | 08:59 |
Qiming | if it is only a neutron problem, we can revise itself | 08:59 |
Yanyanhu | without influence other resources | 08:59 |
Qiming | yes | 08:59 |
Yanyanhu | understand, will try it | 08:59 |
Qiming | ok | 08:59 |
Qiming | elynn is pinging me about austin summit proposal | 09:00 |
Qiming | do you have any ideas? about senlin, what can we propose? | 09:00 |
Yanyanhu | oh, right, I received his mail, almost forgot this... | 09:00 |
elynn | yes, Internal review will start from 1/11, if we need to catch up with internal review, need to submit them befoer 1/11 | 09:03 |
Qiming | I'm thinking if the health management thing can be a feature to show case | 09:07 |
xuhaiwei | I think we have not introduced Senlin to openstack community formally yet | 09:08 |
xuhaiwei | except Heat team | 09:09 |
Qiming | yes | 09:12 |
Qiming | we have proposed senlin deep dive to vancouver, no luck, to tokyo, no luck | 09:12 |
Qiming | why don't we try a 3rd time | 09:12 |
Yanyanhu | :) | 09:13 |
Yanyanhu | we should | 09:13 |
*** dixiaoli has joined #senlin | 09:13 | |
elynn | VM HA is a good use case to show in summit, just don't know how tc choose topic... | 09:13 |
Qiming | so we propose two? | 09:14 |
Qiming | one is about a deep dive about senlin | 09:14 |
Qiming | the other is about a real use case | 09:14 |
Qiming | it would be controversial if we propose a topic like "Clustering containers in an OpenStack native way." | 09:15 |
xuhaiwei | we should make a cool topic name anyway, it can add the possibility to be chosen | 09:18 |
xuhaiwei | should we still talk about container? | 09:18 |
Qiming | out of magnum's context | 09:19 |
Qiming | we can do that | 09:19 |
Qiming | :) | 09:19 |
xuhaiwei | we can put that in the future job of Senlin introduction session | 09:19 |
Qiming | yep | 09:21 |
openstackgerrit | Merged openstack/senlin: Fix KeyError caused by getting 'security_groups' from server_data https://review.openstack.org/263558 | 09:35 |
openstackgerrit | Merged openstack/senlin: Updated from global requirements https://review.openstack.org/264467 | 09:36 |
*** lixinhui_ has joined #senlin | 09:42 | |
Qiming | lixinhui_ is back | 09:44 |
Qiming | so ... | 09:44 |
Qiming | opinions | 09:44 |
Qiming | about submitting someting to austin summit | 09:44 |
Qiming | IBM internal deadline is next Monday | 09:44 |
lixinhui_ | cool | 09:46 |
xuhaiwei | Senlin introduction again, about the background, the structure, where are we, future jobs | 09:46 |
lixinhui_ | I like Elynn's mention to VM HA use case | 09:50 |
lixinhui_ | We should china mobile to vote for senlin | 09:50 |
lixinhui_ | we should ask china mobile to vote for senlin | 09:51 |
lixinhui_ | Seems container is very popular in openstack also | 09:52 |
Qiming | yip | 09:53 |
Yanyanhu | new gerrit UI is terrible, sigh... | 09:54 |
Qiming | it has been popular for two years now | 09:54 |
Yanyanhu | anti-human | 09:54 |
Qiming | developers are never grouped into human | 09:55 |
Yanyanhu | sigh... | 09:55 |
Yanyanhu | we are people... | 09:55 |
lixinhui_ | :) | 09:56 |
lixinhui_ | I am thinking if it is proper tp provide a talk/demo like AWS | 10:03 |
Qiming | like AWS ? | 10:04 |
lixinhui_ | as in the blood of senlin, better support to scaling, lb and health management | 10:04 |
Qiming | bragging? | 10:04 |
Qiming | not yet, :) | 10:04 |
Qiming | that is the vision | 10:05 |
Qiming | yes | 10:05 |
lixinhui_ | :) | 10:05 |
Qiming | back in Tokyo, Heat team was asking a question: what should Heat do with the AWS autoscaling API if autoscaling is offloaded to Senlin | 10:06 |
Qiming | I told them this API should be taken over by Senlin as well | 10:06 |
Yanyanhu | so an AWS-compatible API might be needed in future | 10:06 |
Qiming | Angus and Steve Baker was cheering | 10:06 |
Qiming | applauding I think | 10:06 |
Yanyanhu | they sufferred from this kind of thing I think :) | 10:07 |
Yanyanhu | to keep compatible with another API interface | 10:07 |
lixinhui_ | enen | 10:07 |
openstackgerrit | Qiming Teng proposed openstack/senlin: Rework DB sorting functions https://review.openstack.org/264698 | 10:08 |
elynn | So we directly provide an API at senlin side or to organize senlin resources in heat side? | 10:08 |
lixinhui_ | it is a headache before senlin | 10:08 |
Qiming | yes | 10:08 |
Qiming | they were trapped | 10:08 |
Qiming | trapped by something they woud rather stay away from | 10:08 |
xuhaiwei | so can't we drop AWS autoscaling API in senlin? | 10:09 |
xuhaiwei | why should we receive them? | 10:09 |
Qiming | xuhaiwei, it is a community wide requirement | 10:10 |
Qiming | there are requirements for OpenStack to support AWS APIs | 10:10 |
*** elynn has quit IRC | 10:11 | |
xuhaiwei | ec2 api has been kicked out from Nova | 10:11 |
xuhaiwei | Heat can't kick AWS autoscaling API? | 10:11 |
Qiming | I think there is still a work group doing that | 10:11 |
xuhaiwei | manage them somewhere else | 10:12 |
Qiming | nova team ... | 10:12 |
Qiming | no comment | 10:12 |
lixinhui_ | I can not say other users, VMware users really like AWS alike things | 10:12 |
lixinhui_ | I mean the VIO users, VIO is openstack product | 10:12 |
lixinhui_ | running on vSphere | 10:12 |
Qiming | nova cores believe they are Gods | 10:13 |
lixinhui_ | it has AT&T and Nike ... as customers | 10:13 |
Qiming | yes, I do hear a lot requirements on mailinglist, people need AWS APIs sometimes | 10:14 |
Qiming | but anyway, for senlin, I don't think it is a high priority | 10:14 |
Qiming | just put it in FEATURES.rst at the moment | 10:14 |
lixinhui_ | okay | 10:14 |
Qiming | about the proposals | 10:15 |
Qiming | my suggestion | 10:15 |
lixinhui_ | maybe nowdays, we could mentions these advantages as a part of senlin introduction | 10:15 |
Qiming | we propose two | 10:15 |
Qiming | one is about a generic intro to Senlin | 10:15 |
*** Yanyanhu has quit IRC | 10:16 | |
Qiming | another is about a use case, say VM health management | 10:16 |
lixinhui_ | +1 | 10:16 |
xuhaiwei | ok, the chance of being chosen is doubled | 10:17 |
Qiming | hopefully | 10:17 |
lixinhui_ | Just try our best, and collect name list to vote then... | 10:20 |
*** Qiming has quit IRC | 10:22 | |
*** dixiaoli has quit IRC | 11:11 | |
*** Qiming has joined #senlin | 11:14 | |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Description about Overriding Profile https://review.openstack.org/264727 | 11:24 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Description about Recover of Profile https://review.openstack.org/264727 | 12:49 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Description about Recover of Profile https://review.openstack.org/264727 | 12:51 |
openstackgerrit | Cindia-blue proposed openstack/senlin: Add Description about Recover Function in Profile https://review.openstack.org/264727 | 13:10 |
*** lixinhui_ has quit IRC | 13:15 | |
openstackgerrit | Merged openstack/senlin: Add 'wait_for_delete' call for functional tests https://review.openstack.org/263038 | 13:23 |
openstackgerrit | Qiming Teng proposed openstack/senlin: Fix sample placement policies https://review.openstack.org/264773 | 13:26 |
openstackgerrit | Merged openstack/senlin: Rework DB sorting functions https://review.openstack.org/264698 | 13:32 |
openstackgerrit | Qiming Teng proposed openstack/senlin: Add utility function for parsing sort parameter https://review.openstack.org/264796 | 14:31 |
*** Qiming has quit IRC | 16:12 | |
*** rebase has joined #senlin | 17:55 | |
*** bdrich has joined #senlin | 18:18 | |
*** rebase has quit IRC | 19:06 | |
*** junxu has quit IRC | 19:29 | |
*** openstackgerrit has quit IRC | 20:23 | |
*** openstackgerrit has joined #senlin | 20:25 | |
*** dixiaoli has joined #senlin | 22:45 | |
*** dixiaoli_ has joined #senlin | 22:55 | |
*** dixiaoli has quit IRC | 23:01 | |
*** dixiaoli_ has quit IRC | 23:01 | |
*** dixiaoli has joined #senlin | 23:03 | |
*** Qiming has joined #senlin | 23:32 | |
*** bdrich has quit IRC | 23:34 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!