| *** adrian_otto has joined #openstack-kolla | 00:01 | |
| jascott1 | wrt to testing kk8s and helm upgrades, is that something we will have in the gate? | 00:02 |
|---|---|---|
| kfox1111 | jascott1: yeah. we should. | 00:03 |
| jascott1 | does that need a blueprint or? | 00:04 |
| kfox1111 | oh, probably. | 00:05 |
| *** goldyfruit has quit IRC | 00:07 | |
| kfox1111 | we have 2.0.x images and 3.0.x images in the hub now, so we have material to try an upgrade too. :) | 00:08 |
| *** dave-mccowan has quit IRC | 00:08 | |
| *** hyang has quit IRC | 00:09 | |
| jascott1 | oooh i meant just helm -> helm based. thinking more about making sure the workflow suffices. yeah i could see upgrade path being important irl ;) | 00:11 |
| kfox1111 | oh. like, helm 2.0.0 -> helm 2.0.1 | 00:11 |
| kfox1111 | thats important too I guess. but, they should be testing that. | 00:11 |
| jascott1 | more like I install compute kit with helm and then upgrade varioius pieces, waht does that look like | 00:12 |
| kfox1111 | kind of more in the vain of the: "lets cache external artiacts like helm/kubernetes and ensure they are current/working with our build" | 00:12 |
| sbezverk | kfox1111: I do not follow, if we change helm version in setup_kubernetes and go through the gate, is it not sufficient? | 00:12 |
| *** Pavo has quit IRC | 00:12 | |
| *** JRobinson__ is now known as JRobinson__afk | 00:12 | |
| kfox1111 | sbezverk: for most things, yes. | 00:13 |
| jascott1 | kfox1111 talking about validating the helm upgrade my-release(s) workflow | 00:13 |
| kfox1111 | but if say, 2.0.2 comes out and is buggy and eats template vars but still looks like it works, then that might be a thing to test. | 00:13 |
| kfox1111 | jascott1: ah. ok. | 00:13 |
| sbezverk | jascott1: you were talking helm's "helm upgrade" thing :-) | 00:15 |
| jascott1 | yes! | 00:15 |
| *** Pavo has joined #openstack-kolla | 00:17 | |
| jascott1 | but back to kfox1111's original interpretation of my statement, should we have gate for 2.0->3.0 | 00:17 |
| jascott1 | non-helm to helm, is that what yo were saying kfox1111? | 00:18 |
| *** schwicht has joined #openstack-kolla | 00:20 | |
| kfox1111 | jascott1: no. helm to helm. | 00:21 |
| kfox1111 | major upgrade testing. | 00:21 |
| jascott1 | ah | 00:22 |
| kfox1111 | helm install kolla/neutron-server --version 2.0.2 | 00:22 |
| kfox1111 | #some other part of the process | 00:22 |
| kfox1111 | helm upgrade kolla/neutron-server --version 3.0.1 | 00:22 |
| kfox1111 | where there's some job to do db migration/backup/etc. | 00:23 |
| jascott1 | yup | 00:23 |
| jascott1 | surely it can all be done but how ugly it will have to be idk | 00:23 |
| * kfox1111 nods | 00:25 | |
| kfox1111 | alwasy has been ugly every other time I've done it. :) | 00:25 |
| *** schwicht has quit IRC | 00:26 | |
| jascott1 | sw dev: the relentless pursuit of perfection manifested as omg it almost works most of the time | 00:27 |
| kfox1111 | jascott1: hehe. | 00:27 |
| kfox1111 | with openstack upgrades, its always bit me in one way or another. :/ | 00:27 |
| *** JRobinson__afk is now known as JRobinson__ | 00:29 | |
| *** yingjun has joined #openstack-kolla | 00:29 | |
| sdake_ | jascott1 lol | 00:29 |
| sdake_ | right on dude ;) | 00:29 |
| jascott1 | :) | 00:30 |
| sdake_ | jascott1 we want to make something work all of the time :) | 00:30 |
| sdake_ | you may notice i left out part of the sentence ;) | 00:30 |
| jascott1 | yes i would be happy to see something work all the time | 00:31 |
| sdake_ | any kolla core reviewers about | 00:31 |
| jascott1 | oh hey kfox1111 think I should make a blueprint for upgrades or did someone else want to? | 00:34 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 00:35 |
| sdake_ | jascott1 iirc inc0 said we are not handling upgrades in 1.0.0 | 00:35 |
| sdake_ | jascott1 but nothing stops you from making a blueprint - it just may not be prioritized for 1.0.0 | 00:35 |
| sbezverk | kfox1111: let's forget for now about Release. Name, I will fix it last. please check pv/pvc in the latest ps | 00:36 |
| jascott1 | yeah that makes sense. just want to see it fleshed out | 00:36 |
| jascott1 | sdake_ ^ | 00:36 |
| sdake_ | jascott1 good time to put it in discussion | 00:36 |
| sdake_ | people can discuss it | 00:36 |
| sdake_ | i see the gate promotions are working | 00:37 |
| sdake_ | yay :) | 00:37 |
| *** rstarmer_ has joined #openstack-kolla | 00:37 | |
| * sdake_ always a little concerned when modifying the gate ;) | 00:37 | |
| sdake_ | PUT THE HAMMER DOWN | 00:38 |
| sdake_ | sup rstarmer | 00:38 |
| sdake_ | no more need for check experimental | 00:38 |
| sdake_ | kfox1111 gotta give it to ya man - you figured out the gate, thats some serious hard work ;) | 00:39 |
| jascott1 | no doubt! | 00:39 |
| *** Bico_Fino has quit IRC | 00:39 | |
| *** rstarmer has quit IRC | 00:40 | |
| kfox1111 | sdake_: thanks. :) | 00:44 |
| kfox1111 | sdake_: persistence is key. :) | 00:44 |
| kfox1111 | I'm very stubbern. :) | 00:44 |
| *** sdake_ has quit IRC | 00:45 | |
| sbezverk | kfox1111: wow new word of the day "stubbern" ;-) | 00:45 |
| kfox1111 | sbezverk: unlike you,. I don't have the excuse of not being a native english speaker. I just suck at it. :) | 00:46 |
| *** sdake has joined #openstack-kolla | 00:46 | |
| sbezverk | kfox1111: I thought it was some sort of southern jargon ;-) | 00:47 |
| kfox1111 | :) | 00:47 |
| kfox1111 | could be. :) | 00:47 |
| sdake | what is southern jargon | 00:48 |
| *** Marcellin__ has quit IRC | 00:49 | |
| sbezverk | kfox1111: do you think you will be able to build that small patch to copy stateful thingy? if not I will just copy ceph variables to common for now, I need to test pv/pvc at the gate and wihtout your patch it will not work.. | 00:50 |
| *** hfu has joined #openstack-kolla | 00:52 | |
| jascott1 | kollabot wfa what is "southern jargon"? | 00:53 |
| kollabot1 | No results! | 00:53 |
| jascott1 | hmm you gave up pretty fast there bot | 00:53 |
| jascott1 | im no longer convinced you are connected to the intertubes | 00:54 |
| jascott1 | properly anyway | 00:54 |
| *** tovin07 has joined #openstack-kolla | 00:54 | |
| *** tovin07_ has joined #openstack-kolla | 00:54 | |
| *** rhallisey has quit IRC | 00:55 | |
| *** diogogmt has quit IRC | 00:55 | |
| jascott1 | bots of the hook, wolframalpha has no idea what im asking | 00:56 |
| *** yingjun has quit IRC | 00:57 | |
| *** vilobhmm112 has quit IRC | 00:58 | |
| jascott1 | kollabot wfa MD5 "A kollabot's work is never done." | 00:59 |
| kollabot1 | Input interpretation MD5 | A kollabot's work is never done. | 00:59 |
| *** duonghq has joined #openstack-kolla | 00:59 | |
| kollabot1 | Message digest integer form | 152468366363674955718345317154661134388 | 00:59 |
| kollabot1 | hexadecimal form | 72b4 56d8 1d90 1730 5d10 e1d2 b5c4 f034 | 00:59 |
| duonghq | morning guys | 00:59 |
| duonghq | kollabot1, poke | 00:59 |
| sdake | sup duonghq | 00:59 |
| sdake | kollabot1 feed me dinner | 00:59 |
| sdake | its broken jascott1 | 00:59 |
| jascott1 | lol | 00:59 |
| sbezverk | sdake: you should ask for something more realistic, like write kubernetes version 2 ;-) | 01:02 |
| sdake | the day will come... | 01:02 |
| *** fragatin_ has quit IRC | 01:02 | |
| *** zhubingbing has joined #openstack-kolla | 01:03 | |
| sdake | sup zhangshuai | 01:03 |
| sdake | zhubingbing | 01:03 |
| *** fragatina has joined #openstack-kolla | 01:03 | |
| sdake | sorry lots of z's :) | 01:03 |
| zhubingbing | i am here | 01:03 |
| sdake | just a greeting | 01:03 |
| jascott1 | always wanted to read vonnegut's 'player piano' about future being engineers and non-engineers | 01:03 |
| zhubingbing | hi sdake | 01:03 |
| sdake | greating | 01:03 |
| sdake | meh :) | 01:03 |
| *** rstarmer_ has quit IRC | 01:04 | |
| *** tovin07_ has quit IRC | 01:04 | |
| *** rstarmer has joined #openstack-kolla | 01:04 | |
| *** tovin07_ has joined #openstack-kolla | 01:04 | |
| *** rstarmer has quit IRC | 01:06 | |
| openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: DO NOT MERGE!: Testing gate workflow (so I understand it) https://review.openstack.org/407207 | 01:08 |
| kfox1111 | there's something you don't see everyday... | 01:12 |
| kfox1111 | was seeing something truely bizar. ended up being a router that needed its mac address cache dropped. then everything started working again. :/ | 01:12 |
| *** zhurong has joined #openstack-kolla | 01:13 | |
| kfox1111 | ok. back to sbezverk's ceph thingy | 01:14 |
| sbezverk | kfox1111: could be DoS attack, somebody poisoned router's mac address cache | 01:18 |
| *** schwicht has joined #openstack-kolla | 01:18 | |
| *** tonanhngo has quit IRC | 01:18 | |
| *** sdake_ has joined #openstack-kolla | 01:19 | |
| kfox1111 | oh.... let me check something. | 01:19 |
| *** sdake has quit IRC | 01:20 | |
| kfox1111 | can't find that very quickly. will send it to the network team. thakns for the suggestion. | 01:21 |
| *** sdake has joined #openstack-kolla | 01:22 | |
| *** zhurong has quit IRC | 01:25 | |
| sbezverk | kfox1111: on cisco box you can enable some debugs to see | 01:25 |
| *** sdake_ has quit IRC | 01:25 | |
| *** zhurong has joined #openstack-kolla | 01:25 | |
| *** unicell1 has joined #openstack-kolla | 01:26 | |
| kfox1111 | se cleared the cache already, so probably won't see it again. :/ | 01:26 |
| kfox1111 | we | 01:26 |
| *** unicell has quit IRC | 01:27 | |
| *** adrian_otto has quit IRC | 01:28 | |
| *** adrian_otto has joined #openstack-kolla | 01:28 | |
| kfox1111 | so, values seems to really want to treat things as strings. | 01:29 |
| kfox1111 | so we can't do the same tricks we've been doing with json. | 01:29 |
| kfox1111 | this makes the ceph stuff harder. | 01:29 |
| *** yingjun has joined #openstack-kolla | 01:30 | |
| kfox1111 | grr.... | 01:33 |
| kfox1111 | split doesn't seem to be working. or range isnt. or both. | 01:33 |
| *** sdake_ has joined #openstack-kolla | 01:33 | |
| kfox1111 | or, I'm just dumb. :) | 01:34 |
| kfox1111 | ok. got range working... lets back up and test vals again. | 01:35 |
| *** sdake has quit IRC | 01:35 | |
| *** unicell has joined #openstack-kolla | 01:36 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 01:37 |
| *** tonanhngo has joined #openstack-kolla | 01:37 | |
| *** unicell1 has quit IRC | 01:38 | |
| sbezverk | kfox1111: I got range working got ceph monitors, it is used in pv temaplate | 01:38 |
| kfox1111 | yeah. it looks like it works if you put it in a values file and pas it via the cli. | 01:39 |
| kfox1111 | but it fails if you try to do it via the --set. | 01:39 |
| kfox1111 | or at least, I haven't figured out a way to specify it. | 01:39 |
| sbezverk | kfox1111: for gate I specified the same ip you used for monitor | 01:40 |
| sbezverk | in the gate config | 01:41 |
| kfox1111 | sbezverk: k. | 01:41 |
| *** sayantan_ has quit IRC | 01:41 | |
| kfox1111 | I've got the other review you asked for working. I've got to head home, then I'lll look at committing it. | 01:41 |
| sbezverk | kfox1111: sounds good for now I used common as a workaround | 01:42 |
| *** tonanhngo has quit IRC | 01:42 | |
| kfox1111 | k. | 01:42 |
| kfox1111 | l8r. | 01:42 |
| *** elynn has joined #openstack-kolla | 01:45 | |
| *** elynn_ has joined #openstack-kolla | 01:48 | |
| *** elynn has quit IRC | 01:51 | |
| *** adrian_otto has quit IRC | 01:52 | |
| Jeffrey4l | morning guys | 01:59 |
| duonghq | sup Jeffrey4l | 02:03 |
| *** rstarmer has joined #openstack-kolla | 02:10 | |
| *** Pavo has quit IRC | 02:12 | |
| *** sdake has joined #openstack-kolla | 02:15 | |
| *** hfu has quit IRC | 02:15 | |
| *** sdake_ has quit IRC | 02:17 | |
| *** Pavo has joined #openstack-kolla | 02:17 | |
| *** Pavo has quit IRC | 02:17 | |
| *** vilobhmm11 has joined #openstack-kolla | 02:25 | |
| *** schwicht has quit IRC | 02:29 | |
| *** Pavo has joined #openstack-kolla | 02:30 | |
| *** hfu has joined #openstack-kolla | 02:44 | |
| *** yuanying has quit IRC | 02:46 | |
| *** elynn_ has quit IRC | 02:57 | |
| *** sp_ has joined #openstack-kolla | 02:58 | |
| *** kristian__ has joined #openstack-kolla | 03:02 | |
| *** elynn_ has joined #openstack-kolla | 03:02 | |
| *** elynn_ has quit IRC | 03:06 | |
| *** kristian__ has quit IRC | 03:06 | |
| *** elynn_ has joined #openstack-kolla | 03:07 | |
| *** sdake_ has joined #openstack-kolla | 03:10 | |
| *** mgiles_ has quit IRC | 03:12 | |
| *** sdake has quit IRC | 03:13 | |
| *** lrensing has joined #openstack-kolla | 03:16 | |
| *** adrian_otto has joined #openstack-kolla | 03:17 | |
| *** adrian_otto has quit IRC | 03:23 | |
| *** adrian_otto has joined #openstack-kolla | 03:25 | |
| *** hfu has quit IRC | 03:31 | |
| *** adrian_otto has quit IRC | 03:33 | |
| *** adrian_otto has joined #openstack-kolla | 03:33 | |
| *** hfu has joined #openstack-kolla | 03:36 | |
| *** hfu has quit IRC | 03:40 | |
| *** hfu has joined #openstack-kolla | 03:40 | |
| *** adrian_otto has quit IRC | 03:43 | |
| *** adrian_otto has joined #openstack-kolla | 03:43 | |
| *** hfu has quit IRC | 03:44 | |
| *** hfu has joined #openstack-kolla | 03:45 | |
| *** adrian_otto has quit IRC | 03:46 | |
| *** sp_ has quit IRC | 03:47 | |
| *** adrian_otto has joined #openstack-kolla | 03:47 | |
| *** hfu has quit IRC | 03:47 | |
| *** adrian_otto has joined #openstack-kolla | 03:48 | |
| *** sayantan_ has joined #openstack-kolla | 03:49 | |
| *** hfu has joined #openstack-kolla | 03:50 | |
| *** hfu has quit IRC | 03:54 | |
| *** adrian_otto1 has joined #openstack-kolla | 03:58 | |
| *** adrian_otto1 has quit IRC | 04:00 | |
| *** sp_ has joined #openstack-kolla | 04:01 | |
| *** adrian_otto has quit IRC | 04:01 | |
| sp_ | hello everyone | 04:08 |
| *** hfu has joined #openstack-kolla | 04:08 | |
| *** prameswar has joined #openstack-kolla | 04:12 | |
| *** hfu has quit IRC | 04:15 | |
| *** vilobhmm11 has quit IRC | 04:15 | |
| *** vilobhmm11 has joined #openstack-kolla | 04:17 | |
| *** hfu has joined #openstack-kolla | 04:20 | |
| *** tonanhngo has joined #openstack-kolla | 04:21 | |
| *** mdnadeem has joined #openstack-kolla | 04:21 | |
| *** hfu has quit IRC | 04:27 | |
| *** Pavo has quit IRC | 04:30 | |
| *** elynn_ has quit IRC | 04:30 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add *.egg-info into .gitignore file https://review.openstack.org/407312 | 04:31 |
| *** Pavo has joined #openstack-kolla | 04:35 | |
| *** sdake_ has quit IRC | 04:39 | |
| *** adrian_otto has joined #openstack-kolla | 04:43 | |
| *** sdake has joined #openstack-kolla | 04:44 | |
| sdake | Jeffrey4l about? | 04:48 |
| Jeffrey4l | sdake, yep | 04:49 |
| sdake | https://review.openstack.org/#/c/372133/ | 04:50 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for glance https://review.openstack.org/406978 | 04:50 |
| sdake | Jeffrey4l can you take a look at that review pls | 04:51 |
| Jeffrey4l | sdake, done. cloud u review this https://review.openstack.org/406978 , a new idea about reconfigure. still wip, but almost done. | 04:52 |
| sdake | Jeffrey4l will review in 5-10 mins ok? | 04:52 |
| Jeffrey4l | sdake, np. just need a glance. still have some issue and need fix. | 04:53 |
| *** elynn_ has joined #openstack-kolla | 04:53 | |
| openstackgerrit | Merged openstack/kolla: Follow Docker Inc's trademark guidelines https://review.openstack.org/372133 | 04:54 |
| openstackgerrit | Steven Dake proposed openstack/kolla-ansible: Follow Docker Inc's trademark guidelines https://review.openstack.org/407321 | 04:54 |
| Jeffrey4l | in this way, we can keep copy_once and copy_always. it is easy to implment copy_once now. | 04:55 |
| *** yuanying has joined #openstack-kolla | 04:56 | |
| openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes: Follow Docker Inc's trademark guidelines https://review.openstack.org/407325 | 04:58 |
| *** elynn__ has joined #openstack-kolla | 04:58 | |
| *** elynn_ has quit IRC | 04:58 | |
| sdake | ok looking now | 04:58 |
| sdake | also klindgren_ wanted you to review his spec related to uids/gids | 04:59 |
| Jeffrey4l | ok, np. | 04:59 |
| *** pbourke has quit IRC | 05:01 | |
| *** pbourke has joined #openstack-kolla | 05:01 | |
| sdake | Jeffrey4l not sur ewhtat a flush handler is | 05:02 |
| sdake | so really can't provide a qualified review on that without digging into what that is | 05:02 |
| Jeffrey4l | it will force to run handler define in handlers/main.yml file. | 05:02 |
| sdake | looks pretty good | 05:03 |
| sdake | i made a name suggestion | 05:03 |
| sdake | other then that, I dont see any major problems | 05:03 |
| Jeffrey4l | cool. | 05:03 |
| Jeffrey4l | in this way, we can unify deploy and reconfigure. it is the same right now. | 05:04 |
| Jeffrey4l | and reconfigure is as fast as deploy. | 05:04 |
| Jeffrey4l | and code is less than before :) | 05:04 |
| sdake | yup i picked up on that | 05:05 |
| sdake | lokos pretty good | 05:05 |
| sdake | pretty clever actually ;-) | 05:06 |
| sdake | Jeffrey4l can you +2 those other two reviews i just submitted | 05:06 |
| Jeffrey4l | yes. i will work out two or three role on this. then we can distribute work items to others. | 05:07 |
| Jeffrey4l | np. i will review them. may be later ;) | 05:07 |
| sdake | they are pretty much c&p from othe rrepo | 05:07 |
| sdake | with a depends on for the legal review | 05:07 |
| Jeffrey4l | ya. you are meaning the docker inc trademark PS. done | 05:08 |
| Jeffrey4l | sup pbourke | 05:09 |
| *** sdake_ has joined #openstack-kolla | 05:10 | |
| Jeffrey4l | pbourke, source u review this https://review.openstack.org/406995 , i want change the way gate running. it depends on project-config change in https://review.openstack.org/406997 | 05:10 |
| Jeffrey4l | source/could* | 05:10 |
| *** srwilkers_ has quit IRC | 05:11 | |
| *** sdake has quit IRC | 05:13 | |
| sdake_ | kolla still 3rd busiest project in openstack | 05:19 |
| sdake_ | ok night all | 05:19 |
| Jeffrey4l | night | 05:22 |
| *** yuanying_ has joined #openstack-kolla | 05:26 | |
| *** yuanying has quit IRC | 05:29 | |
| *** sp_ has quit IRC | 05:32 | |
| *** yuanying has joined #openstack-kolla | 05:36 | |
| *** hfu has joined #openstack-kolla | 05:37 | |
| *** yuanying_ has quit IRC | 05:40 | |
| *** hfu has quit IRC | 05:40 | |
| *** JRobinson__ is now known as JRobinson__afk | 05:40 | |
| *** yuanying_ has joined #openstack-kolla | 05:41 | |
| *** hfu has joined #openstack-kolla | 05:42 | |
| *** rstarmer has quit IRC | 05:42 | |
| *** yuanying has quit IRC | 05:44 | |
| *** mdnadeem has quit IRC | 05:44 | |
| *** sp_ has joined #openstack-kolla | 05:45 | |
| *** hfu has quit IRC | 05:47 | |
| *** hfu has joined #openstack-kolla | 05:47 | |
| *** zhubingbing has quit IRC | 05:47 | |
| *** zhubingbing has joined #openstack-kolla | 05:48 | |
| *** sdake_ has quit IRC | 05:48 | |
| *** lrensing has quit IRC | 05:48 | |
| *** mdnadeem has joined #openstack-kolla | 05:52 | |
| *** adrian_otto has quit IRC | 05:56 | |
| *** JRobinson__afk is now known as JRobinson__ | 06:00 | |
| *** kristian__ has joined #openstack-kolla | 06:01 | |
| *** kristian__ has quit IRC | 06:05 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for glance https://review.openstack.org/406978 | 06:05 |
| *** hfu has quit IRC | 06:06 | |
| *** zhubingbing has quit IRC | 06:06 | |
| *** zhubingbing has joined #openstack-kolla | 06:07 | |
| *** adrian_otto has joined #openstack-kolla | 06:08 | |
| *** sdake has joined #openstack-kolla | 06:09 | |
| *** tonanhngo has quit IRC | 06:09 | |
| openstackgerrit | jianyi proposed openstack/kolla-ansible: Enable senlin dashboard when enable_senlin is yes https://review.openstack.org/405078 | 06:10 |
| *** hfu has joined #openstack-kolla | 06:15 | |
| *** adrian_otto has quit IRC | 06:24 | |
| *** adrian_otto has joined #openstack-kolla | 06:26 | |
| *** adrian_otto has quit IRC | 06:27 | |
| *** hfu has quit IRC | 06:27 | |
| *** sdake_ has joined #openstack-kolla | 06:28 | |
| *** Pavo has quit IRC | 06:30 | |
| *** sdake has quit IRC | 06:30 | |
| *** hfu has joined #openstack-kolla | 06:32 | |
| *** Pavo has joined #openstack-kolla | 06:34 | |
| *** hfu has quit IRC | 06:36 | |
| *** JRobinson__ has quit IRC | 06:45 | |
| openstackgerrit | chen.xing proposed openstack/kolla: Fix the format issue https://review.openstack.org/407343 | 06:45 |
| *** hfu has joined #openstack-kolla | 06:46 | |
| openstackgerrit | chen.xing proposed openstack/kolla: Fix the format issue https://review.openstack.org/407343 | 06:46 |
| *** hfu has quit IRC | 06:48 | |
| *** kristian__ has joined #openstack-kolla | 06:49 | |
| *** hfu has joined #openstack-kolla | 06:53 | |
| *** kristian__ has quit IRC | 06:54 | |
| *** hfu has quit IRC | 06:57 | |
| *** hfu has joined #openstack-kolla | 06:57 | |
| *** sp_ has quit IRC | 07:01 | |
| *** sp_ has joined #openstack-kolla | 07:03 | |
| *** prameswar has quit IRC | 07:04 | |
| *** vilobhmm11 has quit IRC | 07:06 | |
| *** vilobhmm11 has joined #openstack-kolla | 07:07 | |
| *** saneax-_-|AFK is now known as saneax | 07:11 | |
| *** yuanying_ has quit IRC | 07:13 | |
| *** kristian__ has joined #openstack-kolla | 07:15 | |
| *** g3ek has quit IRC | 07:20 | |
| *** g3ek has joined #openstack-kolla | 07:21 | |
| *** elynn__ has quit IRC | 07:25 | |
| *** elynn__ has joined #openstack-kolla | 07:26 | |
| *** Guest66666 has quit IRC | 07:27 | |
| *** zhubingbing has quit IRC | 07:28 | |
| *** zhubingbing has joined #openstack-kolla | 07:28 | |
| *** Guest66666 has joined #openstack-kolla | 07:29 | |
| *** kristia__ has joined #openstack-kolla | 07:32 | |
| *** klindgren has joined #openstack-kolla | 07:34 | |
| *** klindgren_ has quit IRC | 07:34 | |
| *** kristi___ has joined #openstack-kolla | 07:34 | |
| *** kristian__ has quit IRC | 07:35 | |
| *** Jeffrey4l has quit IRC | 07:36 | |
| *** kristia__ has quit IRC | 07:38 | |
| *** kristi___ has quit IRC | 07:39 | |
| *** pcaruana has joined #openstack-kolla | 07:42 | |
| *** kristian__ has joined #openstack-kolla | 07:42 | |
| *** klindgren_ has joined #openstack-kolla | 07:45 | |
| *** vilobhmm11 has joined #openstack-kolla | 07:45 | |
| *** klindgren has quit IRC | 07:46 | |
| *** kristian__ has quit IRC | 07:47 | |
| *** hfu has quit IRC | 07:48 | |
| *** hfu has joined #openstack-kolla | 07:49 | |
| *** athomas has joined #openstack-kolla | 07:51 | |
| *** Jeffrey4l has joined #openstack-kolla | 07:52 | |
| *** hfu has quit IRC | 07:53 | |
| *** fragatina has quit IRC | 07:53 | |
| *** hfu has joined #openstack-kolla | 07:53 | |
| *** fragatina has joined #openstack-kolla | 07:54 | |
| openstackgerrit | Merged openstack/kolla-ansible: Follow Docker Inc's trademark guidelines https://review.openstack.org/407321 | 07:55 |
| *** fragatina has quit IRC | 07:59 | |
| *** sdake has joined #openstack-kolla | 07:59 | |
| *** yuanying has joined #openstack-kolla | 08:01 | |
| *** matrohon has joined #openstack-kolla | 08:02 | |
| *** int32bit has joined #openstack-kolla | 08:02 | |
| *** tonanhngo has joined #openstack-kolla | 08:03 | |
| *** sdake_ has quit IRC | 08:03 | |
| *** tonanhngo has quit IRC | 08:04 | |
| *** vilobhmm11 has quit IRC | 08:11 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 08:14 |
| *** Pavo has quit IRC | 08:30 | |
| *** Pavo has joined #openstack-kolla | 08:34 | |
| *** int32bit has quit IRC | 08:35 | |
| *** Serlex has joined #openstack-kolla | 08:35 | |
| *** jascott1 has quit IRC | 08:39 | |
| *** mkoderer has joined #openstack-kolla | 08:46 | |
| *** mdnadeem has quit IRC | 08:52 | |
| openstackgerrit | junbo proposed openstack/kolla-ansible: Integrate ceph-rgw with keystone https://review.openstack.org/406971 | 08:58 |
| openstackgerrit | junbo proposed openstack/kolla-ansible: Integrate ceph-rgw with keystone https://review.openstack.org/406971 | 09:01 |
| openstackgerrit | Christian Berendt proposed openstack/kolla-ansible: Remove store_events parameter from ceilometer configuration https://review.openstack.org/407414 | 09:02 |
| *** mdnadeem has joined #openstack-kolla | 09:04 | |
| *** ram_ has quit IRC | 09:05 | |
| *** egonzalez90 has joined #openstack-kolla | 09:09 | |
| *** shardy has joined #openstack-kolla | 09:11 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for nova https://review.openstack.org/407416 | 09:12 |
| *** gfidente has joined #openstack-kolla | 09:21 | |
| *** rmart04 has joined #openstack-kolla | 09:27 | |
| *** kristian__ has joined #openstack-kolla | 09:27 | |
| *** kristian__ has quit IRC | 09:27 | |
| *** kristian__ has joined #openstack-kolla | 09:27 | |
| *** strigazi_AFK is now known as strigazi | 09:28 | |
| *** Bico_Fino has joined #openstack-kolla | 09:28 | |
| *** Bico_Fino has quit IRC | 09:31 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for glance https://review.openstack.org/406978 | 09:35 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for nova https://review.openstack.org/407416 | 09:36 |
| *** yingjun has quit IRC | 09:40 | |
| *** kristian__ has quit IRC | 09:40 | |
| *** shardy is now known as shardy_mtg | 09:45 | |
| *** elynn__ has quit IRC | 09:47 | |
| *** sdake has quit IRC | 09:49 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 09:51 |
| duonghq | how can we set quota for neutron? in /etc/kolla/config/neutron/neutron.conf? | 09:52 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Install murano dashboard into horizon image https://review.openstack.org/406474 | 09:58 |
| *** huhaoran has joined #openstack-kolla | 10:01 | |
| *** zhurong has quit IRC | 10:03 | |
| *** tovin07_ has quit IRC | 10:05 | |
| *** sayantan_ has quit IRC | 10:05 | |
| *** kristian__ has joined #openstack-kolla | 10:06 | |
| openstackgerrit | Merged openstack/kolla-ansible: Enable senlin dashboard when enable_senlin is yes https://review.openstack.org/405078 | 10:08 |
| egonzalez90 | duonghq: addind the content in the following link to etc/kolla/config/neutron/neutron.conf http://docs.openstack.org/admin-guide/cli-networking-advanced-quotas.html | 10:08 |
| egonzalez90 | then reconfigure | 10:08 |
| *** tovin07 has quit IRC | 10:09 | |
| *** athomas has quit IRC | 10:09 | |
| *** zhangshuai has quit IRC | 10:09 | |
| duonghq | egonzalez90, thank you, I just found the way | 10:10 |
| *** kristian__ has quit IRC | 10:11 | |
| *** zhangshuai has joined #openstack-kolla | 10:11 | |
| openstackgerrit | prameswar proposed openstack/kolla: bifrost guide improved https://review.openstack.org/406177 | 10:15 |
| *** prameswar has joined #openstack-kolla | 10:18 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla-ansible: Remove store_events parameter from ceilometer configuration https://review.openstack.org/407414 | 10:18 |
| prameswar | Hi All, can someone review these two PS :) https://review.openstack.org/#/c/406177/ and https://review.openstack.org/#/c/407061/ | 10:22 |
| *** JoseMello has joined #openstack-kolla | 10:24 | |
| egonzalez90 | berendt: I cannot find any release note about the deprecation of store-events option? Only the PS on which they removed it. Were you able to find it? | 10:26 |
| berendt | egonzalez90: I am not sure if there was a deprecation note, I only found the referenced review request | 10:26 |
| *** Pavo has quit IRC | 10:30 | |
| *** kristian__ has joined #openstack-kolla | 10:30 | |
| *** kristia__ has joined #openstack-kolla | 10:32 | |
| *** kristia__ has quit IRC | 10:33 | |
| *** kristia__ has joined #openstack-kolla | 10:34 | |
| *** kristian__ has quit IRC | 10:34 | |
| *** Pavo has joined #openstack-kolla | 10:35 | |
| *** athomas has joined #openstack-kolla | 10:35 | |
| *** kristia__ has quit IRC | 10:35 | |
| *** kristian__ has joined #openstack-kolla | 10:35 | |
| *** kristia__ has joined #openstack-kolla | 10:40 | |
| *** kristian__ has quit IRC | 10:40 | |
| *** huhaoran has quit IRC | 10:46 | |
| *** huhaoran has joined #openstack-kolla | 10:47 | |
| *** huhaoran has quit IRC | 10:51 | |
| *** huhaoran has joined #openstack-kolla | 10:51 | |
| *** hfu has quit IRC | 10:58 | |
| *** kristian__ has joined #openstack-kolla | 10:59 | |
| *** kristia__ has quit IRC | 11:02 | |
| *** kristia__ has joined #openstack-kolla | 11:04 | |
| *** yingjun has joined #openstack-kolla | 11:05 | |
| openstackgerrit | Li Yingjun proposed openstack/kolla-ansible: Enable searchlight ui when searchlight enabled https://review.openstack.org/405957 | 11:06 |
| *** kristian__ has quit IRC | 11:08 | |
| *** yingjun has quit IRC | 11:13 | |
| openstackgerrit | qiankun.zhang proposed openstack/kolla: Install watcher-dashboard into horizon image https://review.openstack.org/407460 | 11:14 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add enabled_zaqar in group all.yml file https://review.openstack.org/407461 | 11:14 |
| mliima_ | morning guys | 11:20 |
| *** zhangshuai has quit IRC | 11:21 | |
| *** zhangshuai has joined #openstack-kolla | 11:21 | |
| mliima_ | sean-k-mooney, i'm so sorry | 11:22 |
| mliima_ | :/ | 11:22 |
| *** kristia__ has quit IRC | 11:23 | |
| *** kristian__ has joined #openstack-kolla | 11:25 | |
| *** msimonin has joined #openstack-kolla | 11:25 | |
| Jeffrey4l | sup egonzalez90 | 11:25 |
| Jeffrey4l | hi mliima_ | 11:26 |
| *** zhubingbing has quit IRC | 11:26 | |
| mliima_ | hey Jeffrey4l | 11:26 |
| Jeffrey4l | mliima_, i am trying to optimize the reconfiguration. could u take a look https://review.openstack.org/406978 | 11:27 |
| mliima_ | sure | 11:27 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for nova https://review.openstack.org/407416 | 11:27 |
| Jeffrey4l | mliima_, here is the implementation of nova https://review.openstack.org/407416 | 11:28 |
| *** kristia__ has joined #openstack-kolla | 11:28 | |
| mliima_ | okay | 11:28 |
| *** kristian__ has quit IRC | 11:28 | |
| *** kristia__ has quit IRC | 11:29 | |
| *** shardy_mtg has quit IRC | 11:30 | |
| *** yingjun has joined #openstack-kolla | 11:30 | |
| mliima_ | Jeffrey4l, | 11:31 |
| mliima_ | What do you think to change recreate_or_restart_container to reconfigure_container? | 11:31 |
| Jeffrey4l | mliima_, yep? | 11:31 |
| Jeffrey4l | actually it is not just reconfigure. | 11:31 |
| mliima_ | just a suggestion | 11:31 |
| mliima_ | oh | 11:31 |
| mliima_ | okay | 11:31 |
| Jeffrey4l | i removed start.yml and upgrade.yml. | 11:31 |
| mliima_ | np Jeffrey4l | 11:32 |
| Jeffrey4l | so technically, it will start the container for the first time/ in deployment stage. | 11:32 |
| Jeffrey4l | but recreate_or_restart_container is still a bad name. :( | 11:34 |
| *** yingjun has quit IRC | 11:36 | |
| *** yingjun has joined #openstack-kolla | 11:36 | |
| *** mdnadeem has quit IRC | 11:37 | |
| *** yingjun has quit IRC | 11:41 | |
| *** mdnadeem has joined #openstack-kolla | 11:49 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add enable_zaqar in group all.yml file https://review.openstack.org/407461 | 12:00 |
| Jeffrey4l | mliima_, thanks. | 12:01 |
| Jeffrey4l | mliima_, here is another ps related to CI change, could u review it ;) https://review.openstack.org/406995 and related project-config change https://review.openstack.org/406997 | 12:01 |
| egonzalez90 | Jeffrey4l: about https://review.openstack.org/#/c/407461/ | 12:04 |
| Jeffrey4l | yep | 12:04 |
| egonzalez90 | IMHO think make more sense remove zaqar from Horizon container, since we do not support deployment of this service | 12:04 |
| *** tonanhngo has joined #openstack-kolla | 12:04 | |
| mliima_ | egonzalez90, https://github.com/openstack/kolla/tree/master/docker/zaqar | 12:05 |
| egonzalez90 | yup, but vars come from ansible | 12:05 |
| *** tonanhngo has quit IRC | 12:05 | |
| Jeffrey4l | mliima_, docker files exist but the role is not. | 12:05 |
| mliima_ | oh | 12:05 |
| egonzalez90 | I though we have ansible roles since zaqar containers have been here for long time, but not | 12:06 |
| mliima_ | I agree egonzalez90 | 12:06 |
| egonzalez90 | related bp: https://blueprints.launchpad.net/kolla-ansible/+spec/ansible-zaqar | 12:07 |
| Jeffrey4l | yes. may be reverted and wait zaqar role . | 12:07 |
| mliima_ | but it no make any sense | 12:07 |
| mliima_ | why add zaqar dockerfiles and not add role | 12:07 |
| mliima_ | :/ | 12:07 |
| Jeffrey4l | it may raise some issue when enable zaqar without real service. | 12:08 |
| Jeffrey4l | mliima_, role is hard then dockerfiles lol | 12:08 |
| mliima_ | haha | 12:08 |
| egonzalez90 | yup, is in the queue from 2015 | 12:08 |
| Jeffrey4l | or keep this and wait the role to merge? | 12:09 |
| Jeffrey4l | really? is there a PS for this? | 12:09 |
| *** zhubingbing has joined #openstack-kolla | 12:09 | |
| egonzalez90 | not in the review queue, in blueprints waiting for work | 12:09 |
| egonzalez90 | sorry for bad wording | 12:10 |
| Jeffrey4l | roger. | 12:10 |
| mliima_ | well, lets revert it | 12:10 |
| Jeffrey4l | anyway, OK. | 12:10 |
| *** yingjun has joined #openstack-kolla | 12:10 | |
| Jeffrey4l | working on it. | 12:10 |
| *** portdirect__ has joined #openstack-kolla | 12:10 | |
| *** hfu has joined #openstack-kolla | 12:11 | |
| mliima_ | my +2 is waiting for it haha | 12:11 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Revert "Enable zaqar ui when zaqar enabled" https://review.openstack.org/407461 | 12:15 |
| sean-k-mooney | mliima_: hi, dont be sorry by the way if you didnt get to test the conntrack driver its fine. im not going to use manila for a while as i just installed a larger hdd on the server i was going to provide the share too | 12:15 |
| mliima_ | I can test it now sen | 12:16 |
| mliima_ | sean-k-mooney, | 12:16 |
| mliima_ | tell me how | 12:16 |
| *** kristian__ has joined #openstack-kolla | 12:16 | |
| Jeffrey4l | revert is pushed https://review.openstack.org/#/c/407461/ | 12:17 |
| mliima_ | :wq | 12:17 |
| *** kristian__ has quit IRC | 12:17 | |
| *** kristian__ has joined #openstack-kolla | 12:17 | |
| mliima_ | reusing PS Jeffrey4l | 12:17 |
| mliima_ | hahah | 12:17 |
| egonzalez90 | Jeffrey4l: ENABLE_ZAQAR in extend_start don't break anything right? is not needed to be reverted or yes? https://review.openstack.org/#/c/406606/ | 12:18 |
| sean-k-mooney | mliima_: add this to /etc/kolla/config/neutron/ml2_conf.ini and deploy or reconfigure [securitygroup] | 12:18 |
| sean-k-mooney | firewall_driver = openvswitch | 12:18 |
| Jeffrey4l | egonzalez90, i think so. keep it won't hurt anything. | 12:18 |
| mliima_ | yes egonzalez90 | 12:18 |
| Jeffrey4l | egonzalez90, we should not implement kolla's depends kolla-ansible's change now. | 12:18 |
| Jeffrey4l | may kolla-k8s need this ;) | 12:18 |
| mliima_ | sean-k-mooney, | 12:20 |
| mliima_ | within manila.conf | 12:20 |
| Jeffrey4l | egonzalez90, i am optimizing kolla-ansible reconfiguration. could u review the init PS for glance and a following PS for nova https://review.openstack.org/406978 https://review.openstack.org/407416 this change is a little big. so need more eye on the initial PS. | 12:20 |
| mliima_ | when neutron_plugin_agent is openvswitch | 12:20 |
| mliima_ | interface_driver = manila.network.linux.interface.OVSInterfaceDriver | 12:20 |
| mliima_ | have i change something there? | 12:20 |
| egonzalez90 | will do, I'm reviewing PS from yesterday yet | 12:21 |
| Jeffrey4l | thanks. | 12:21 |
| *** kristian__ has quit IRC | 12:22 | |
| *** kristian__ has joined #openstack-kolla | 12:23 | |
| sean-k-mooney | mliima_: no you should not need to change anything there | 12:23 |
| mliima_ | running sean-k-mooney | 12:24 |
| mliima_ | :) | 12:24 |
| sean-k-mooney | mliima_: thanks | 12:25 |
| sean-k-mooney | mliima_: hopefully you can reporduce and it not use a local thing to my setup. that would be annoying to track down | 12:26 |
| *** kristian__ has quit IRC | 12:28 | |
| *** athomas has quit IRC | 12:29 | |
| *** Pavo has quit IRC | 12:30 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for nova https://review.openstack.org/407416 | 12:31 |
| *** schwicht has joined #openstack-kolla | 12:33 | |
| openstackgerrit | Merged openstack/kolla: Add run_gate.sh script for CI https://review.openstack.org/406995 | 12:33 |
| *** schwicht has quit IRC | 12:34 | |
| *** Pavo has joined #openstack-kolla | 12:35 | |
| mliima_ | sean-k-mooney, | 12:56 |
| mliima_ | same error | 12:56 |
| mliima_ | :) | 12:56 |
| openstackgerrit | Merged openstack/kolla-ansible: Tacker NFV Ansible support https://review.openstack.org/399179 | 13:04 |
| *** kristian__ has joined #openstack-kolla | 13:04 | |
| *** rhallisey has joined #openstack-kolla | 13:05 | |
| openstackgerrit | Merged openstack/kolla-ansible: Revert "Enable zaqar ui when zaqar enabled" https://review.openstack.org/407461 | 13:05 |
| *** hfu has quit IRC | 13:07 | |
| *** kristian__ has quit IRC | 13:08 | |
| *** jheroux has joined #openstack-kolla | 13:11 | |
| *** sdake has joined #openstack-kolla | 13:11 | |
| sdake | morning peeps | 13:14 |
| *** fragatina has joined #openstack-kolla | 13:15 | |
| *** fragatin_ has joined #openstack-kolla | 13:15 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add chrony ansible role https://review.openstack.org/399355 | 13:16 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Implement octavia ansible role https://review.openstack.org/404648 | 13:19 |
| *** fragatina has quit IRC | 13:19 | |
| *** dave-mccowan has joined #openstack-kolla | 13:19 | |
| *** lamt has joined #openstack-kolla | 13:19 | |
| *** shardy has joined #openstack-kolla | 13:22 | |
| mliima_ | morning sdake | 13:22 |
| *** hfu has joined #openstack-kolla | 13:22 | |
| sdake | hey mliima_ | 13:23 |
| *** huhaoran has quit IRC | 13:25 | |
| *** yingjun has quit IRC | 13:25 | |
| *** mdnadeem has quit IRC | 13:27 | |
| openstackgerrit | Merged openstack/kolla-ansible: Remove store_events parameter from ceilometer configuration https://review.openstack.org/407414 | 13:28 |
| *** hfu has quit IRC | 13:29 | |
| *** adrian_otto has joined #openstack-kolla | 13:29 | |
| *** hfu has joined #openstack-kolla | 13:29 | |
| *** kristian__ has joined #openstack-kolla | 13:32 | |
| *** Bico_Fino has joined #openstack-kolla | 13:32 | |
| mliima_ | sean-k-mooney, the "main" issue is because manila failt to create a share server | 13:33 |
| *** kristian__ has quit IRC | 13:34 | |
| *** kristian__ has joined #openstack-kolla | 13:35 | |
| *** hfu has quit IRC | 13:38 | |
| zhubingbing | hi guys | 13:39 |
| *** fguillot has joined #openstack-kolla | 13:41 | |
| *** mattmceuen has quit IRC | 13:42 | |
| *** msimonin1 has joined #openstack-kolla | 13:45 | |
| *** msimonin has quit IRC | 13:46 | |
| *** huhaoran has joined #openstack-kolla | 13:47 | |
| *** palexster has quit IRC | 13:51 | |
| *** prameswar has quit IRC | 13:51 | |
| *** fragatin_ has quit IRC | 13:52 | |
| *** prameswar has joined #openstack-kolla | 13:52 | |
| *** fragatina has joined #openstack-kolla | 13:52 | |
| *** hfu has joined #openstack-kolla | 13:53 | |
| *** duonghq_ has joined #openstack-kolla | 13:57 | |
| *** yingjun has joined #openstack-kolla | 13:57 | |
| duonghq_ | evening | 13:57 |
| *** hfu has quit IRC | 13:58 | |
| *** Marcellin__ has joined #openstack-kolla | 13:58 | |
| *** palexster has joined #openstack-kolla | 13:59 | |
| *** CBR09 has joined #openstack-kolla | 14:00 | |
| *** lamt has quit IRC | 14:01 | |
| *** lamt has joined #openstack-kolla | 14:02 | |
| *** Bico_Fino has quit IRC | 14:02 | |
| *** hfu has joined #openstack-kolla | 14:08 | |
| *** goldyfruit has joined #openstack-kolla | 14:09 | |
| *** adrian_otto has quit IRC | 14:11 | |
| v1k0d3n | sdake: hey man | 14:16 |
| v1k0d3n | so i tried docker hub this time...same issue. | 14:16 |
| v1k0d3n | i could use a bit of help if someone is familiar with pushing images to a public registry. | 14:17 |
| v1k0d3n | :( | 14:17 |
| *** sbezverk_ has joined #openstack-kolla | 14:17 | |
| v1k0d3n | or is there a way to debug kolla-build a little bit more? | 14:17 |
| dmsimard | sdake, Jeffrey4l: I'll let you guys discuss some more https://review.openstack.org/#/c/407165/ before doing anything ? I don't mind doing whatever, just need to agree on what | 14:20 |
| dmsimard | berendt: ^ | 14:20 |
| dmsimard | I don't have a strong opinion | 14:21 |
| *** sbezverk has quit IRC | 14:21 | |
| *** schwicht has joined #openstack-kolla | 14:22 | |
| Jeffrey4l | dmsimard, include: {{ ansible_os_family }}.yml is more elegant imo ;) | 14:22 |
| *** hfu has quit IRC | 14:28 | |
| *** hfu has joined #openstack-kolla | 14:29 | |
| *** duonghq_ has quit IRC | 14:29 | |
| *** Pavo has quit IRC | 14:30 | |
| *** Pavo has joined #openstack-kolla | 14:34 | |
| *** hfu has quit IRC | 14:35 | |
| *** hfu has joined #openstack-kolla | 14:38 | |
| *** hfu has quit IRC | 14:40 | |
| dmsimard | Jeffrey4l: well, sure, but then it requires some involved refactoring | 14:40 |
| dmsimard | Jeffrey4l: I didn't plan on refactoring that, or else I would've already changed some other things as well | 14:40 |
| dmsimard | mixing the dict and = notation drives me nuts | 14:40 |
| dmsimard | i.e, http://paste.openstack.org/show/591543/ | 14:41 |
| Jeffrey4l | dmsimard, that's ansible feature. | 14:41 |
| dmsimard | Jeffrey4l: what, mixing syntax ? :) | 14:41 |
| Jeffrey4l | i am OK with the refactor in another PS. just throw a new solution. | 14:42 |
| Jeffrey4l | where this come from? | 14:42 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 14:42 |
| Jeffrey4l | for task, ansible support a=b in one line and a: b | 14:42 |
| *** saneax is now known as saneax-_-|AFK | 14:44 | |
| *** jtriley has joined #openstack-kolla | 14:44 | |
| *** sdake has quit IRC | 14:45 | |
| openstackgerrit | Merged openstack/kolla: add aodh os_aodh_apache_log.lua in heka https://review.openstack.org/384602 | 14:50 |
| *** rstarmer has joined #openstack-kolla | 14:51 | |
| *** rstarmer has quit IRC | 14:53 | |
| dmsimard | Jeffrey4l: I don't know, but when I see playbooks/tasks written using the two different syntaxes simultaneously it makes me sad | 14:54 |
| *** sdake has joined #openstack-kolla | 14:55 | |
| egonzalez90 | dmsimard: agree with you | 14:55 |
| *** rstarmer has joined #openstack-kolla | 14:55 | |
| *** rstarmer has quit IRC | 14:56 | |
| sdake | v1k0d3n - i have meetings most of the day today until 2pm mst | 14:56 |
| Jeffrey4l | dmsimard, egonzalez90 it is a feature of ansible. i am hate at the beginning. but like it now :) | 14:56 |
| *** rstarmer has joined #openstack-kolla | 14:56 | |
| dmsimard | The "dict notation" (the second example in what I pasted above) is proper YAML while the syntax with equal ("=") signs is understood by Ansible | 14:56 |
| sdake | v1k0d3n and have kid duty as well - so not sure I can get to it before then | 14:56 |
| Jeffrey4l | imaging show it looks when using dict format: file: src=a dest=b mode=777 user=aaa group=bbb | 14:56 |
| *** rstarmer has quit IRC | 14:56 | |
| v1k0d3n | anyone else build and push images to a public repository? | 14:57 |
| dmsimard | Jeffrey4l: yeah but then your lines can easily become too long, you know ? | 14:57 |
| *** rstarmer has joined #openstack-kolla | 14:57 | |
| Jeffrey4l | it depends when it too long, i prefer to use dict format :) | 14:57 |
| *** rstarmer has quit IRC | 14:57 | |
| egonzalez90 | in my eyes is harder to read one liner code, also at review if the line is to long have to scroll | 14:57 |
| dmsimard | Jeffrey4l: and then before you know it, you start deciding which notation to use based on the length of your lines .. and that makes everything inconsistent :( | 14:57 |
| v1k0d3n | having issues troubleshooting some of these steps, and i think the docs could use a little love. | 14:57 |
| *** rstarmer has joined #openstack-kolla | 14:58 | |
| v1k0d3n | sdake: can i have custom names on the repos, or do they HAVE to be "keystone-base"? | 14:58 |
| *** rstarmer has quit IRC | 14:58 | |
| v1k0d3n | for instance, can i have clcp-keystone-base? | 14:58 |
| sbezverk_ | v1k0d3n: I have done it recently for ubuntu-binary image of rabbitmq | 14:58 |
| Jeffrey4l | lol i am OK with the two format if it not too long, less 100 char should be OK. | 14:58 |
| v1k0d3n | actually, it's keystone-base-ubuntu-source or something like that :-/ | 14:58 |
| *** rstarmer has joined #openstack-kolla | 14:58 | |
| v1k0d3n | sbezverk_ what am i missing here? | 14:59 |
| *** rstarmer has quit IRC | 14:59 | |
| v1k0d3n | sbezverk_: i am using the command... | 14:59 |
| v1k0d3n | kolla-build keystone --base ubuntu --type source --push --registry hub.docker.com/attcomdev/clcp-keystone --tag 4.0.1 | 14:59 |
| v1k0d3n | i think this is incorrect. | 14:59 |
| *** rstarmer has joined #openstack-kolla | 14:59 | |
| v1k0d3n | becaue i am getting errors for both docker and quay. | 14:59 |
| sbezverk_ | v1k0d3n: First I build image locally, then tag it with correct tag and then push it to docker hub.. | 15:00 |
| sbezverk_ | I do not do it in one step, never tried it | 15:00 |
| openstackgerrit | Merged openstack/kolla: Tacker Docker configuration https://review.openstack.org/396391 | 15:01 |
| *** tonanhngo has joined #openstack-kolla | 15:02 | |
| *** tonanhngo has quit IRC | 15:03 | |
| *** rstarmer has quit IRC | 15:04 | |
| *** hfu has joined #openstack-kolla | 15:04 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 15:06 |
| *** lbeliveau_ is now known as lbeliveau | 15:11 | |
| *** jrich523 has quit IRC | 15:12 | |
| *** jrich523 has joined #openstack-kolla | 15:12 | |
| *** rstarmer has joined #openstack-kolla | 15:14 | |
| *** yingjun has quit IRC | 15:15 | |
| *** yingjun has joined #openstack-kolla | 15:16 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Install watcher dashboard into horizon image https://review.openstack.org/406607 | 15:19 |
| *** srwilkers_ has joined #openstack-kolla | 15:20 | |
| *** yingjun has quit IRC | 15:20 | |
| sbezverk_ | portdirect__: ping | 15:21 |
| openstackgerrit | caoyuan proposed openstack/kolla: Install watcher dashboard into horizon image https://review.openstack.org/406607 | 15:21 |
| portdirect__ | sbezverk_: pong | 15:22 |
| *** rstarmer has quit IRC | 15:23 | |
| v1k0d3n | so sbezverk_ manual works. i'd really like to modify it so the kolla-build command could be more flexible? | 15:24 |
| v1k0d3n | i've gotta be missing something here.... | 15:24 |
| v1k0d3n | gonna keep playing. | 15:24 |
| sbezverk_ | v1k0d3n: hm, it will not be easy as kolla-build feeds images to kolla-ansible as well.. | 15:25 |
| sbezverk_ | portdirect__: do you know in ceph we use for kube, where ceph.conf gets generated? | 15:26 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Stateful services https://review.openstack.org/407566 | 15:26 |
| kfox1111 | morning. | 15:27 |
| v1k0d3n | sbezverk_: not sure i follow... | 15:27 |
| kfox1111 | sbezverk_: we need to document that part. I think we should just have a procedure for loading in the ceph.conf from the ceph cluster, | 15:27 |
| portdirect__ | sbezverk_: https://github.com/openstack/kolla-kubernetes/blob/master/tests/bin/build_test_ceph.sh#L51 | 15:27 |
| kfox1111 | rather then trying to buld it. | 15:27 |
| sbezverk_ | kfox1111: morning, I have qq. I hit a couple of time issue with ceph | 15:27 |
| kfox1111 | sure. | 15:28 |
| kfox1111 | time issues? | 15:28 |
| *** tonanhngo has joined #openstack-kolla | 15:28 | |
| sbezverk_ | it seems to resolve it, I need to add this config statement: rbd default features = 1 | 15:28 |
| sbezverk_ | from time to time I see this error: rbd: sysfs write failed | 15:29 |
| kfox1111 | strange. | 15:29 |
| *** bmace has quit IRC | 15:29 | |
| kfox1111 | which host type are you using? | 15:29 |
| kfox1111 | ubuntu? | 15:29 |
| sbezverk_ | kfox1111: another thing for tiller crashing, it turned out not tiller issue but kubernetes networking.. | 15:30 |
| sbezverk_ | kfox1111: strangely centos, ubuntu works like a charm | 15:30 |
| kfox1111 | sbezverk_: oh, interesting. what did you discover? | 15:30 |
| *** msimonin1 has quit IRC | 15:30 | |
| *** rstarmer has joined #openstack-kolla | 15:30 | |
| *** saneax-_-|AFK is now known as saneax | 15:30 | |
| kfox1111 | sbezverk_: which host is failing then? it may be that the kernel your using is starting to get rbd 2 features then gets confused. I think centos only supports v1 rbd. | 15:31 |
| *** msimonin has joined #openstack-kolla | 15:31 | |
| *** mgiles has joined #openstack-kolla | 15:31 | |
| sbezverk_ | kfox1111: I went through kubelet logs and all pods were complaining about cni networking not being properly installed | 15:31 |
| kfox1111 | sbezverk_: maybe we force the setting in the ceph-rbd pod in an init-container? | 15:31 |
| sbezverk_ | tiller crash was just due to healthcheck they implemented.. | 15:31 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Enable watcher dashboard when enable_watcher is yes https://review.openstack.org/406608 | 15:31 |
| kfox1111 | sbezverk_: oh. the cni stuff does fail for a bit I think. | 15:31 |
| kfox1111 | as we start helm before calico is finished I think? | 15:31 |
| kfox1111 | but it we have a barior after calico, waiting for all pods to it ready. | 15:32 |
| portdirect__ | yeah - its notmal for cni to be unhappy for some time after initialising the cluster | 15:32 |
| kfox1111 | so eventually it starts and settles. | 15:32 |
| portdirect__ | *normal | 15:32 |
| srwilkers_ | morning everyone | 15:32 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Enable watcher dashboard when enable_watcher is yes https://review.openstack.org/406608 | 15:32 |
| portdirect__ | morning srwilkers :) | 15:32 |
| sbezverk_ | kfox1111: we should do it for sure.. I mean wait for calico to settle down.. | 15:32 |
| kfox1111 | hmm.. so maybe we need some more barior after calico to do more detection of if its ready? | 15:33 |
| kfox1111 | if it itself isn't good enough f or readyness checks? | 15:33 |
| kfox1111 | yeah. | 15:33 |
| sbezverk_ | kfox1111: could you try to add the config I asked before? It will take more time for me to figure out where to add it properly, since you know your ceph inside out.. | 15:33 |
| portdirect__ | sbezverk_/kfox1111: the easiest way to to that is check the status of the dns pod - but will will also need to test on a node-by node basis | 15:33 |
| portdirect__ | as dns pod will only tell if the master node and control pane are up | 15:34 |
| kfox1111 | sbezverk_: the rbd default features = 1 ? sure. | 15:34 |
| kfox1111 | portdirect__: I check for all pods ready in the kube-system namespace | 15:35 |
| kfox1111 | portdirect__: so that should have covered it. | 15:35 |
| sbezverk_ | portdirect__: my concern is this check is only for timing issue, but what if dns is ok but other containers have issue with networking? | 15:35 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Install murano dashboard into horizon image https://review.openstack.org/406474 | 15:35 |
| *** lrensing has joined #openstack-kolla | 15:36 | |
| sbezverk_ | kfox1111: here is ceph issue I am seeing: http://logs.openstack.org/22/405422/41/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/21c03f4/logs/pods/kolla-rabbitmq-bootstrap-iudun.txt | 15:36 |
| portdirect__ | sbezverk_: the dns pod uses the cni neworking - so if it's running then by implication CNI is up and running ok on the master node (so in sinlge node testing this is all thats needed) - but for multinode the only real way to get the status of CNI (that I'm aware of) is to try and secedult a pod on it - a there is no universal way to test status accross the various CNI backends that I'm aware of | 15:37 |
| portdirect__ | *secedult -> schedule (sorry my typing is aweful today) | 15:38 |
| sbezverk_ | portdirect__: ok in either case we can start with check for dns but as Kevin has already mentioned it is already done .. | 15:39 |
| *** wwriverrat has quit IRC | 15:41 | |
| sbezverk_ | kfox1111: that was the suggestion from ceph people on similar error hit by other folks.. | 15:41 |
| kfox1111 | sbezverk_: been looking through the logs. | 15:43 |
| kfox1111 | I'm guessing its not that, | 15:43 |
| kfox1111 | as the gate was stable there before the shift from kollakube -> helm for that pv. | 15:43 |
| openstackgerrit | Merged openstack/kolla-ansible: Give plays in the playbooks a name https://review.openstack.org/404822 | 15:43 |
| kfox1111 | probably an issue with the template. can you revert just the pv creation step from the gate and recueck? | 15:43 |
| kfox1111 | I've gota get ready and head in to work. then I'll poke at it some more. | 15:44 |
| *** rstarmer has quit IRC | 15:45 | |
| sbezverk_ | kfox1111: I would agree with you if ubuntu and other centos were failing | 15:46 |
| *** mkoderer has quit IRC | 15:46 | |
| sbezverk_ | now only one out of three deployments show this issue | 15:46 |
| *** mkoderer has joined #openstack-kolla | 15:46 | |
| sbezverk_ | but I will try again with old PV | 15:46 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Install murano dashboard into horizon image https://review.openstack.org/406474 | 15:47 |
| *** saneax is now known as saneax-_-|AFK | 15:47 | |
| egonzalez90 | Jeffrey4l: around? | 15:49 |
| *** rstarmer has joined #openstack-kolla | 15:50 | |
| *** CBR09 has left #openstack-kolla | 15:51 | |
| *** adrian_otto has joined #openstack-kolla | 15:52 | |
| *** fragatina has quit IRC | 15:53 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove the unnecessary "openstack-zaqar-ui" https://review.openstack.org/407576 | 15:54 |
| *** yatin has quit IRC | 15:54 | |
| *** yatin has joined #openstack-kolla | 15:54 | |
| *** mdnadeem has joined #openstack-kolla | 15:54 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Install murano dashboard into horizon image https://review.openstack.org/406474 | 15:55 |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove the unnecessary "openstack-zaqar-ui" https://review.openstack.org/407576 | 15:56 |
| *** rstarmer has quit IRC | 15:58 | |
| *** goldyfruit has quit IRC | 15:58 | |
| Jeffrey4l | egonzalez90, yep | 15:58 |
| *** rstarmer has joined #openstack-kolla | 15:59 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Install murano dashboard into horizon image https://review.openstack.org/406474 | 15:59 |
| *** goldyfruit has joined #openstack-kolla | 16:01 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 16:02 |
| *** ayoung-afk is now known as ayoung | 16:05 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: add panko role https://review.openstack.org/400122 | 16:05 |
| *** sayantani01_ has quit IRC | 16:08 | |
| *** kbyrne has quit IRC | 16:08 | |
| prameswar | https://review.openstack.org/#/c/406177/ and https://review.openstack.org/#/c/407061/ | 16:11 |
| *** kbyrne has joined #openstack-kolla | 16:11 | |
| prameswar | can someone review these two PS ? :) | 16:11 |
| *** wwriverrat has joined #openstack-kolla | 16:12 | |
| *** zhubingbing has quit IRC | 16:12 | |
| prameswar | Jeffrey4l : can you please review ? | 16:13 |
| *** JoseMello has quit IRC | 16:13 | |
| Jeffrey4l | np prameswar | 16:13 |
| sean-k-mooney | mliima_: just saw your respoce so you were able to reproduce the error? | 16:13 |
| sean-k-mooney | mliima_: that is good to know | 16:13 |
| mliima_ | yes | 16:14 |
| sean-k-mooney | mliima_: thanks for taking a look at it | 16:14 |
| *** shardy has quit IRC | 16:14 | |
| mliima_ | can't connect via ssh | 16:14 |
| mliima_ | cant ping | 16:14 |
| sean-k-mooney | mliima_: and can you do it via a vm | 16:14 |
| mliima_ | i asked to manila community guys | 16:14 |
| mliima_ | <mkoderer> mliima_: is there a different security groups api needed to use the feature? | 16:15 |
| sean-k-mooney | i was able to ssh vi a vm just not from the hos | 16:15 |
| mliima_ | <mkoderer> mliima_: I would say it should work.. | 16:15 |
| mliima_ | i tested only from host sean-k-mooney | 16:15 |
| mliima_ | :/ | 16:15 |
| mliima_ | and i can't ping | 16:15 |
| sean-k-mooney | no worries | 16:15 |
| sean-k-mooney | ya i could see the ovs rule that was droping the traffic | 16:15 |
| sean-k-mooney | the manila hare port is part of the service users defualt secuity group | 16:16 |
| sean-k-mooney | i keep missing letters... *the manila share port not hare | 16:16 |
| sean-k-mooney | even opening the default security group it was part of to allow all traffic form any ip did not help | 16:17 |
| mliima_ | np sean-k-mooney | 16:17 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the set config error https://review.openstack.org/371669 | 16:17 |
| *** sayantan_ has joined #openstack-kolla | 16:17 | |
| sean-k-mooney | but in any case the docs in kolla for this seam correct. it would work if the security group driver was not blocking the traffic | 16:17 |
| *** rstarmer has quit IRC | 16:18 | |
| *** rstarmer has joined #openstack-kolla | 16:18 | |
| sean-k-mooney | so i guess this is a manila <-> neutron issue that needs to be resolved | 16:18 |
| *** kbyrne has quit IRC | 16:18 | |
| *** pcaruana has quit IRC | 16:19 | |
| *** rmart04 has quit IRC | 16:19 | |
| mliima_ | perhaps the port-mapping | 16:19 |
| mliima_ | bridge_mappings* | 16:19 |
| *** kbyrne has joined #openstack-kolla | 16:21 | |
| *** rstarmer has quit IRC | 16:21 | |
| sean-k-mooney | perhaps. unfortuetly i wont be able to explore this further for a while. just one last thing, did you open a bug for this with manila? | 16:22 |
| *** rstarmer has joined #openstack-kolla | 16:22 | |
| sean-k-mooney | i can submit a docs change to kolla to indicate that the intergration will fail if the conntrack security group driver is used | 16:22 |
| mliima_ | sorry sean-k-mooney, i can,t, at least not now | 16:23 |
| Jeffrey4l | prameswar, done | 16:23 |
| prameswar | thanks Jeffrey4l | 16:24 |
| sean-k-mooney | mliima_: thats fine i can open the bug just to track this usecase. you have done more then enough to help already. | 16:24 |
| sean-k-mooney | mliima_: thanks | 16:24 |
| *** kollabot1 has quit IRC | 16:24 | |
| mliima_ | sean-k-mooney, when you fill the bug, let me know the bug id | 16:25 |
| mliima_ | ok? | 16:25 |
| *** yee37910 has joined #openstack-kolla | 16:25 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Implement octavia ansible role https://review.openstack.org/404648 | 16:25 |
| *** yee379 has quit IRC | 16:25 | |
| *** kollabot has joined #openstack-kolla | 16:25 | |
| *** kbyrne has quit IRC | 16:25 | |
| *** gfidente has quit IRC | 16:27 | |
| sean-k-mooney | mliima_: sure i can add you to the watchers i think aswell so you get notifed | 16:28 |
| sean-k-mooney | but ill send you the id | 16:28 |
| *** kbyrne has joined #openstack-kolla | 16:28 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add chrony ansible role https://review.openstack.org/399355 | 16:29 |
| kfox1111 | sbezverk_: so it was just the multinode job? Then its even more unlikely that it was the rbd version thing. | 16:29 |
| *** Pavo has quit IRC | 16:30 | |
| *** Satya_ has joined #openstack-kolla | 16:31 | |
| Satya_ | Hi jeffrey4l | 16:31 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add name for tacker playbook https://review.openstack.org/407590 | 16:31 |
| Jeffrey4l | sup Satya_, | 16:31 |
| kfox1111 | sbezverk_: its entirely possible that there's a few very rare race conditions left in the gate. | 16:31 |
| Satya_ | want to check while doing a upgrade of kolla container using kolla-ansible upgrade | 16:31 |
| Satya_ | it always take allinone file by default | 16:31 |
| kfox1111 | maybe we should keep an etherpad of failure types, dates on when they happened, host types, and log links. | 16:32 |
| Satya_ | anyway i can pass multinode file to that? | 16:32 |
| kfox1111 | then we can start correlating issues? | 16:32 |
| Jeffrey4l | Satya_, yes. just do like why you deploy. | 16:32 |
| Jeffrey4l | in default kolla-ansible use allinone inventory file. | 16:32 |
| Jeffrey4l | unless u specify another one | 16:33 |
| openstackgerrit | prameswar proposed openstack/kolla: QSG updated for shared volumes https://review.openstack.org/407061 | 16:33 |
| *** rstarmer has quit IRC | 16:33 | |
| *** fguillot has quit IRC | 16:33 | |
| Satya_ | i tried specifing it as -i | 16:34 |
| Jeffrey4l | then? | 16:35 |
| Satya_ | still it took tha allin one jeffrey4l | 16:35 |
| *** Pavo has joined #openstack-kolla | 16:35 | |
| Jeffrey4l | hrm. wired. could u post the first few lines when u run kolla-ansible? | 16:35 |
| Satya_ | sure | 16:35 |
| *** rstarmer has joined #openstack-kolla | 16:36 | |
| Satya_ | oops sorry | 16:36 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Stateful services https://review.openstack.org/407566 | 16:36 |
| Satya_ | there is another issue | 16:36 |
| Jeffrey4l | what kinda issue? | 16:37 |
| sbezverk_ | kfox1111: I get what you are saying.. I think tracking failure scenarions could be useful. | 16:37 |
| kfox1111 | sbezverk_: know how to setup an etherpad? I've never done it. if not, lets ask sdake. | 16:38 |
| *** unicell has quit IRC | 16:39 | |
| *** vhosakot has joined #openstack-kolla | 16:39 | |
| sbezverk_ | kfox1111: neither have I.. sdake is busy atm, maybe Jeffrey4l can help.. | 16:39 |
| Jeffrey4l | kfox1111, etherpad link? or etherpad server? | 16:40 |
| kfox1111 | Jeffrey4l: new etherpad link off of openstack's server | 16:40 |
| *** rstarmer has quit IRC | 16:40 | |
| Jeffrey4l | kfox1111, hrm this one https://etherpad.openstack.org/ | 16:41 |
| Jeffrey4l | try what u want and click OK | 16:41 |
| kfox1111 | oh. thats super easy. thanks. :) | 16:41 |
| *** rstarmer has joined #openstack-kolla | 16:41 | |
| Jeffrey4l | ;) np | 16:41 |
| kfox1111 | darn.. it doesn't support tables though. :/ | 16:42 |
| Jeffrey4l | https://etherpad.openstack.org/p/you-can-type-what-u-want-here-and-it-will-create-a-new-link-automatically ;) | 16:42 |
| Jeffrey4l | oops, seems too long ^^ | 16:42 |
| kfox1111 | https://etherpad.openstack.org/p/kolla-kubernetes-gate-failure-tracking | 16:42 |
| *** absubram has quit IRC | 16:43 | |
| Satya_ | jeffrey4l can we do a upgrade each component wise like nova then cinder then keystone ... | 16:44 |
| Jeffrey4l | Satya_, kolla-ansible upgrade -t nova | 16:44 |
| Satya_ | ok cool :) | 16:45 |
| Satya_ | Thanks jeffrey4l :) you are my life saver | 16:45 |
| Jeffrey4l | but common role will be upgrade. | 16:45 |
| Jeffrey4l | my pleasure ;) | 16:45 |
| *** goldyfruit has quit IRC | 16:46 | |
| sbezverk_ | kfox1111: I know the problem with PV template.. | 16:46 |
| *** rstarmer has quit IRC | 16:46 | |
| kfox1111 | sbezverk_: oh, you found it? cool. | 16:47 |
| sbezverk_ | in single node case, the monitor ip address is the same | 16:47 |
| *** rstarmer has joined #openstack-kolla | 16:47 | |
| *** goldyfruit has joined #openstack-kolla | 16:47 | |
| kfox1111 | sbezverk_: ah. right. the setup_conf stuff. | 16:47 |
| sbezverk_ | but in multinode it is not the same as configured | 16:47 |
| openstackgerrit | prameswar proposed openstack/kolla: QSG updated for shared volumes https://review.openstack.org/407061 | 16:47 |
| sbezverk_ | how do you get monitor ip for multinode case? | 16:48 |
| kfox1111 | sbezverk_: have a look at https://review.openstack.org/#/c/407566/ | 16:48 |
| kfox1111 | we should merge that, then tweak the setup to lay down an overrides for the monitors. | 16:48 |
| sbezverk_ | kfox1111: sounds good, let's wait for the gate, I will ack as soon as it is all green | 16:48 |
| kfox1111 | kk | 16:50 |
| kfox1111 | gate's running slow today. | 16:52 |
| Marcellin__ | Hey Guys. I'm trying to do an all-in-one kolla deployment following the quick start manual. I'm using Ubuntu 16.04 and trying to deploy with docker. I did every steps without issues and was able to build, but when I try to deploy, the command kolla-ansible is not found | 16:53 |
| Marcellin__ | any thoughts? | 16:53 |
| *** rstarmer has quit IRC | 16:54 | |
| kfox1111 | Marcellin__: if your building from trunk there's a second repo now, kolla-ansible where the deployment tools are. | 16:54 |
| *** rstarmer has joined #openstack-kolla | 16:54 | |
| *** LamT__ has joined #openstack-kolla | 16:55 | |
| Marcellin__ | @kfox1111 Thanks! | 16:55 |
| kfox1111 | Marcellin__: np. :) | 16:55 |
| kfox1111 | oh... its not the gate's slow... | 16:56 |
| kfox1111 | its the promoted gate tests.... | 16:57 |
| kfox1111 | I got use to the fast stuff reporting emediately, and then the slow tests coming in later. | 16:57 |
| kfox1111 | but now its waiting for them all as a set. | 16:57 |
| kfox1111 | that makes sence. | 16:57 |
| sbezverk_ | kfox1111: yep, but on the other hand we do not need to do experiemntal manually | 16:57 |
| kfox1111 | yeah. :) | 16:58 |
| kfox1111 | I'm all for it. just didn't remember it blocked the pep8 stuff. | 16:58 |
| kfox1111 | can simply look at zuul though, as its running to see the status. | 16:58 |
| kfox1111 | looks like the pep8 stuff is all fixed in that review. so once the others pass, should be good to go. | 16:59 |
| sbezverk_ | kfox1111: cool I will need to rebase on this PS too, to pick up stateful stuff | 16:59 |
| kfox1111 | yeah. | 17:00 |
| kfox1111 | I'll work on a ceph-value.yaml builder real quick too. | 17:00 |
| kfox1111 | unfortunately, I couldn't figure out how to make that work with the cli. :/ | 17:00 |
| sbezverk_ | kfox1111: is that for multinode issue? | 17:00 |
| kfox1111 | sbezverk_: yeah. | 17:00 |
| sbezverk_ | kfox1111: cool, thanks | 17:01 |
| Marcellin__ | kfox1111: Just curious, is anyone doing a doc update to reference to that the kolla-ansible repo, because it's a little confusing | 17:01 |
| kfox1111 | Marcellin__: not that I'm aware of. please raise a bug so someone won't forget. | 17:01 |
| kfox1111 | (or a patch to fix it. :) | 17:01 |
| Marcellin__ | Cool! | 17:02 |
| kfox1111 | thanks. | 17:02 |
| portdirect__ | whoop! finally got the gate workflow (with ceph) running in halcyon-kubernetes :) | 17:02 |
| kfox1111 | portdirect__: awesome. :) | 17:03 |
| kfox1111 | thats fantastic news. :) | 17:03 |
| srwilkers_ | portdirect__, thats awesome man | 17:03 |
| portdirect__ | super rough at the mo - but will get it doc'd up and ready over the next few days | 17:03 |
| sbezverk_ | portdirect__: so the gate workflow can be run now on anything, even baremetal? | 17:03 |
| *** rstarmer has quit IRC | 17:04 | |
| *** rstarmer has joined #openstack-kolla | 17:04 | |
| *** imcsk8 has joined #openstack-kolla | 17:04 | |
| portdirect__ | not yet, sbezverk_: it's super hacky at the moment - but now I've got it to run twice on that I can start to make it generic | 17:05 |
| *** egonzalez90 has quit IRC | 17:05 | |
| *** adrian_otto has quit IRC | 17:05 | |
| portdirect__ | (I've hardcoded some IP addresses etc) | 17:06 |
| portdirect__ | runs inside a container outside of the k8s cluster | 17:07 |
| sbezverk_ | portdirect__: cool, let me know when I can run it on bare metal :-) tired of waiting for zuul to test things | 17:07 |
| v1k0d3n | hey guys...have some questions about kolla-build. some concerns actually, and want to know if some changes would be acceptable. | 17:07 |
| *** jascott1 has joined #openstack-kolla | 17:08 | |
| *** rstarmer has quit IRC | 17:08 | |
| *** rstarmer has joined #openstack-kolla | 17:09 | |
| kfox1111 | portdirect__: but, thats the hard part. so that's down. polish is easier/next. :) | 17:09 |
| sbezverk_ | v1k0d3n: you need to sell it kolla core folks :-) | 17:09 |
| portdirect__ | sbezverk_: no probs - I'll loop you in as soon as it's got a change of running on your hardware - getting this a generic tool on the critical path for me so you (and others can use it) | 17:09 |
| kfox1111 | sbezverk_: I've worked around the zuul slowness by just runing multiple reviews at once. | 17:09 |
| kfox1111 | if you get 2 or 3 going, then its not so bad. | 17:10 |
| *** imcsk8 has quit IRC | 17:10 | |
| berendt | v1k0d3n: I am open to bribes... | 17:10 |
| kfox1111 | v1k0d3n: oh, probably. :) | 17:10 |
| v1k0d3n | first is flags, like centos/ubuntu and source/binary is statically driven | 17:10 |
| v1k0d3n | so if i do a docker build and try to push to a repo that is custom...no love. | 17:10 |
| kfox1111 | I'd like to see versioning/tagging done differently too. | 17:11 |
| v1k0d3n | unless i am totally missing something | 17:11 |
| kfox1111 | v1k0d3n: I think there is a repo override too? | 17:11 |
| kfox1111 | so it tags it foo/centos-binary-haproxy:3.0.1 | 17:11 |
| kfox1111 | where foo can be your repo server. | 17:11 |
| v1k0d3n | kfox1111: doesn't really work as expected... | 17:12 |
| v1k0d3n | so if i do this command... | 17:12 |
| v1k0d3n | kolla-build keystone --base ubuntu --type source --push quay.io/attcomdev/clcp-keystone --tag 4.0.1 | 17:12 |
| v1k0d3n | it doesn't work because it wants to push keystone-ubuntu-source:4.0.1 | 17:12 |
| v1k0d3n | which is hard coded i believe. | 17:13 |
| kfox1111 | I think it was designed to not rewrite the container name. | 17:14 |
| *** matrohon has quit IRC | 17:14 | |
| kfox1111 | so --push quay.io/attcomdev ? | 17:14 |
| kfox1111 | or does quay.io allow nested repos? | 17:14 |
| v1k0d3n | kfox1111: so how do i label it "clcp-keystone"? | 17:15 |
| kfox1111 | that, I don't know. I think there may be a way to override the container name, though I'v eonly dealth with the deploy vars, not the build vars. so can't say for sure. | 17:16 |
| portdirect__ | v1k0d3n: think you will need to build the image then tag it manually unless you patch 'kolla-build' as though what your looking for is not possible when building all of the containers without adding a rexgep container renaming function AFAIK | 17:18 |
| sbezverk_ | kfox1111: acked stateful, got to meetings for a couple of hours, looking forward your ceph_values | 17:18 |
| v1k0d3n | portdirect__: yup, that's right. | 17:19 |
| v1k0d3n | so how can i predict and automate builds when i don't know what the image id will be? | 17:19 |
| v1k0d3n | ^^ and that's the problem. | 17:19 |
| v1k0d3n | :( | 17:19 |
| *** msimonin has quit IRC | 17:20 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Stateful services https://review.openstack.org/407566 | 17:21 |
| *** rstarmer_ has joined #openstack-kolla | 17:21 | |
| *** rstarmer has quit IRC | 17:22 | |
| portdirect__ | v1k0d3n: depending on what your aiming for - I previously did something similar by creating a set of images that simply had a FROM: {{OUTPUT_IMAGE}} dockerfile - and used a bash script to popolaute the From value in the dockerfile and then output the desired tagged image - whihc allowed me to decouple the build preocess and deployment for a project | 17:23 |
| portdirect__ | v1k0d3n: but your needing this for your CI pipeline? in which case the above is not likley to be appropriate :( | 17:23 |
| kfox1111 | v1k0d3n: yeah, thats my issue too. I want more ci around versioning. | 17:23 |
| kfox1111 | v1k0d3n: thats a seperate issue from container naming though. | 17:24 |
| kfox1111 | versions can be done with tags. | 17:24 |
| kfox1111 | renaming the containers themselves is a seperate thing. | 17:24 |
| kfox1111 | sbezverk_: still there? | 17:25 |
| sdake | kfox1111 quick q, is neutron done ? | 17:25 |
| kfox1111 | sdake: not yet. close. | 17:25 |
| sdake | kfox1111 or in progress | 17:25 |
| sdake | thanks | 17:26 |
| kfox1111 | helping out with plubming at the moment. | 17:26 |
| kfox1111 | getting the hard stuff out of the way, to make it easier to do stuff in parallel. | 17:26 |
| *** rstarmer_ has quit IRC | 17:26 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 17:27 |
| sbezverk_ | kfox1111: what is up? | 17:27 |
| kfox1111 | sbezverk_: should I amend your review to have the fixed ceph bits? | 17:27 |
| kfox1111 | or do that seperatly. will be harder to test seperately though. | 17:28 |
| sbezverk_ | kfox1111: can you do it in new ps? | 17:28 |
| sbezverk_ | I already rebased mine on stateful | 17:28 |
| kfox1111 | I can, but then you would have to rebase again too. | 17:28 |
| sbezverk_ | sure, that is no issue | 17:28 |
| kfox1111 | k | 17:28 |
| sbezverk_ | kfox1111: for testing you could use rabbitmq template if it gets merged before | 17:30 |
| sbezverk_ | I think it is safe because only multinode is not working and it is not the template issue, but it is up to you | 17:31 |
| kfox1111 | we can just merge the values thing too. it should be pretty obvious if the yaml file isn't correct looking. | 17:31 |
| portdirect__ | kfox1111: quick q - why are we putting haproxy infront of the neutron-server containers? | 17:32 |
| *** diogogmt has joined #openstack-kolla | 17:32 | |
| *** tonanhngo has quit IRC | 17:32 | |
| kfox1111 | portdirect__: it lets us do rolling upgrades safely. | 17:32 |
| kfox1111 | portdirect__: it also gives us a place to plugin ssl termination (pending feature) | 17:32 |
| kfox1111 | some folks have asked for an option to make turning off the haproxy in front optional. I've agreed. the code just hasn't been written yet. | 17:33 |
| openstackgerrit | Merged openstack/kolla-ansible: Add name for tacker playbook https://review.openstack.org/407590 | 17:33 |
| kfox1111 | should just be a couple of strategicallyl placed conditionals in the template. | 17:34 |
| *** mdnadeem has quit IRC | 17:34 | |
| *** rstarmer has joined #openstack-kolla | 17:35 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for glance https://review.openstack.org/406978 | 17:36 |
| portdirect__ | kfox1111: cool - I'd most likley terminate TLS at the container directly if possible (just personal preference - makes PKI client auth easier). how is rolling update improved by haproxy for each pod (I'm not sure - but know you'll have thought it through :D )? | 17:38 |
| kfox1111 | portdirect__: python services don't terminate tls very well. | 17:38 |
| *** rstarmer has quit IRC | 17:38 | |
| kfox1111 | portdirect__: haproxy has a feature to shutdown safely only when all connections are finished. | 17:39 |
| *** rstarmer has joined #openstack-kolla | 17:39 | |
| kfox1111 | so if you look at the deployment, you can see a prestop hook that causes haproxy to stop recieving new connections and stay alive while that happens. | 17:39 |
| v1k0d3n | kfox1111 / portdirect__ true...but i would think others would want CI around this process as well...no? | 17:40 |
| v1k0d3n | as in CI and name to something they want. | 17:40 |
| portdirect__ | kfox1111: ah nice - thats pretty cool | 17:40 |
| kfox1111 | that, in addition to the k8s svc load balancer that removes new connections to terminating pods, allows pods to safely drain out. | 17:40 |
| kfox1111 | zero downtime. :) | 17:40 |
| kfox1111 | v1k0d3n: yeah. I want ci. I'm sure others will too. | 17:41 |
| kfox1111 | v1k0d3n: we need to smooth out that workflow for sure. | 17:41 |
| kfox1111 | be back in a few. meeting time. | 17:41 |
| v1k0d3n | kfox1111: i think it's a big part that should be pluggable for anyone wanting to use it. | 17:41 |
| v1k0d3n | right now it's not easy enough. | 17:42 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Horizon custom policies https://review.openstack.org/399707 | 17:42 |
| portdirect__ | kfox1111: yeah - this is why I'd like us to expore using ingress controllers (with headless services) in future for API endpoints, it's how I put together my old work and worked really well - but thats a discussion for much further down the line | 17:43 |
| *** rstarmer has quit IRC | 17:43 | |
| sbezverk_ | v1k0d3n: since we have haproxy in common deployment template makeing it conditional will be fairly easy | 17:44 |
| *** fguillot has joined #openstack-kolla | 17:46 | |
| *** adrian_otto has joined #openstack-kolla | 17:47 | |
| *** fragatina has joined #openstack-kolla | 17:48 | |
| *** rmart04 has joined #openstack-kolla | 17:48 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimized reconfigure action for nova https://review.openstack.org/407416 | 17:53 |
| *** absubram has joined #openstack-kolla | 17:54 | |
| *** rmart04 has quit IRC | 17:59 | |
| *** rmart04 has joined #openstack-kolla | 18:00 | |
| v1k0d3n | portdirect__ / kfox1111 yup...in support of ingress here too. there are some nifty little workarounds to get them to work outside of 80 :) | 18:02 |
| v1k0d3n | sbezverk_ to be clear...i am only standing up a server to do image builds, pull in a latest commit for say keystone, building and pushing the container (and then tearing down the deployment). no kolla-ansible involved at all. | 18:04 |
| v1k0d3n | in fact i very methodically separate these processes to the point where i don't even want to them of them being in the same project. | 18:04 |
| portdirect__ | v1k0d3n: nice - good to hear you got it working - I've been meaning to play with that, you guys using nginx? | 18:05 |
| v1k0d3n | (in my head only of course :) ) | 18:05 |
| v1k0d3n | well, it's funny... | 18:05 |
| v1k0d3n | alan mentioned that you sent him to that link... | 18:06 |
| v1k0d3n | and i tried it earlier this year...it does in fact work, and i used it with NGINX, yes. | 18:06 |
| *** huhaoran has quit IRC | 18:07 | |
| portdirect__ | sweet - I played around with it months ago and couldn't make it work then (if I recall correctly - though I only spent 30 mins on it) - so ended up sticking with services for rabbit and ovsdb (OVN: northbound/southbound), but if I get time (which seems like a rarer and and rarer commodity these days) I'll test it out with 'em :) | 18:10 |
| *** kristian__ has quit IRC | 18:10 | |
| *** tonanhngo has joined #openstack-kolla | 18:16 | |
| *** tonanhngo_ has joined #openstack-kolla | 18:18 | |
| *** prameswar has quit IRC | 18:19 | |
| *** prameswar has joined #openstack-kolla | 18:20 | |
| *** tonanhngo has quit IRC | 18:20 | |
| harlowja | v1k0d3n how goes it :-P | 18:21 |
| *** tonanhngo_ has quit IRC | 18:23 | |
| srwilkers_ | sup harlowja | 18:24 |
| harlowja | sup sup | 18:24 |
| *** krtaylor has quit IRC | 18:28 | |
| *** Pavo has quit IRC | 18:30 | |
| *** Pavo has joined #openstack-kolla | 18:34 | |
| *** mkoderer has quit IRC | 18:36 | |
| *** prameswar has quit IRC | 18:40 | |
| *** tonanhngo has joined #openstack-kolla | 18:45 | |
| *** v1k0d3n has quit IRC | 18:47 | |
| *** v1k0d3n has joined #openstack-kolla | 18:47 | |
| *** v1k0d3n has quit IRC | 18:52 | |
| *** krtaylor has joined #openstack-kolla | 18:54 | |
| *** rmart04 has quit IRC | 18:58 | |
| *** sdake has quit IRC | 18:59 | |
| *** LamT__ has quit IRC | 19:02 | |
| *** msimonin has joined #openstack-kolla | 19:05 | |
| *** sdake has joined #openstack-kolla | 19:06 | |
| *** huhaoran has joined #openstack-kolla | 19:07 | |
| *** v1k0d3n has joined #openstack-kolla | 19:11 | |
| *** Satya_ has quit IRC | 19:12 | |
| *** huhaoran has quit IRC | 19:12 | |
| *** severion has joined #openstack-kolla | 19:13 | |
| openstackgerrit | James McCarthy proposed openstack/kolla-ansible: Swift data transfer should not be using the API Network https://review.openstack.org/407665 | 19:14 |
| *** severion has quit IRC | 19:15 | |
| *** severion has joined #openstack-kolla | 19:15 | |
| *** severion is now known as v1k0d3n_ | 19:16 | |
| *** v1k0d3n_ has quit IRC | 19:17 | |
| *** v1k0d3n has quit IRC | 19:17 | |
| *** v1k0d3n has joined #openstack-kolla | 19:17 | |
| *** fragatina has quit IRC | 19:21 | |
| openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: WIP: Add Development Setup Script for use with Kubernetes Helm https://review.openstack.org/405720 | 19:23 |
| kfox1111 | back. | 19:25 |
| kfox1111 | too many meetings. :/ | 19:25 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Add ceph values for gate https://review.openstack.org/407672 | 19:28 |
| jascott1 | kfox1111 I created this bp from our conversation yesterday https://blueprints.launchpad.net/kolla-kubernetes/+spec/helm-deployment-upgrade | 19:28 |
| kfox1111 | jascott1: looks good. :) | 19:30 |
| jascott1 | cool was hoping good enough for discussion | 19:30 |
| kfox1111 | jascott1: one of the things I was wanting to do, is come up with an 'upgrade' definition document. | 19:30 |
| kfox1111 | there are lots of different ways to define 'upgrade' and each is different. so we should clearly lay them out. | 19:31 |
| *** yee37910 has quit IRC | 19:31 | |
| *** kollabot has quit IRC | 19:31 | |
| jascott1 | agree | 19:31 |
| *** kollabot has joined #openstack-kolla | 19:31 | |
| *** yee379 has joined #openstack-kolla | 19:31 | |
| kfox1111 | for example. "upgrading from 2.0.2 to 2.0.3 for neutron-server" | 19:31 |
| kfox1111 | zero downtime rolling upgrade | 19:31 |
| jascott1 | exactly | 19:31 |
| kfox1111 | completely different thing then upgrading from 2.0.2 to 3.0.1. that will have downtime. | 19:32 |
| *** zhangshuai has quit IRC | 19:32 | |
| kfox1111 | some services will be zero downtime, or just api downtime depending on how you chose to archetect it. | 19:32 |
| *** zhangshuai has joined #openstack-kolla | 19:32 | |
| kfox1111 | or some even will have data downtime. | 19:32 |
| kfox1111 | like, libvirt upgrade on non ceph backed compute ephemeral nodes. | 19:33 |
| *** jmccarthy has quit IRC | 19:33 | |
| jascott1 | ah | 19:33 |
| kfox1111 | so we shoudl clearly lay out what the various upgrade thingies look like, | 19:33 |
| kfox1111 | so we can talk about them specificly in different situations. | 19:33 |
| kfox1111 | like, in the architecture document, we let users choose different tradeoffs, like, ceph or not ceph backed compute. | 19:33 |
| *** jmccarthy has joined #openstack-kolla | 19:34 | |
| kfox1111 | but then we reference, but then upgradesl ook like X, or like Y depending. | 19:34 |
| *** TxGirlGeek has joined #openstack-kolla | 19:34 | |
| portdirect__ | kfox1111: In working on ceph and getting a gate type enviroment running locally - the biggest issue I found was that it seems that kollakube assumes that the hostname of a node will match the name ascribed to it in k8s - a pretty fair assumption, and a limitation I'm pretty happy to enforce moving foward (due to the large number of things we will have running in the hosts network-namespace) but wondered what your thoughts were? | 19:34 |
| kfox1111 | portdirect__: the config for the gate assumes that. which is a more fair assumption. | 19:35 |
| kfox1111 | something to be figured out if we make the gate scripts more generally useful. which you are trying to do. | 19:35 |
| kfox1111 | (which is good. :) | 19:35 |
| kfox1111 | the ceph it deploys though really was only ever intended for testing. so making it too generic might be problematic. | 19:36 |
| sbezverk_ | kfox1111: any good news? | 19:36 |
| kfox1111 | I'd kind of recommend using ceph-deploy instead. I containerized it for testing just because I was supporting minikube, and it didn't work with it. | 19:36 |
| kfox1111 | sbezverk_: https://review.openstack.org/407672 should do the job I think. | 19:36 |
| kfox1111 | arg... swap hell... | 19:39 |
| kfox1111 | there we go. firefox finally crashed. :) | 19:40 |
| kfox1111 | sbezverk_: with that patch in place, I think you can add a: | 19:40 |
| kfox1111 | --values <(ceph_values) | 19:40 |
| kfox1111 | to the cli of your helm install rabbitmq-pv | 19:40 |
| *** Jeffrey4l has quit IRC | 19:42 | |
| *** Jeffrey4l has joined #openstack-kolla | 19:42 | |
| sbezverk_ | kfox1111: I am not sure how all this will work with already existing ceph: section in vlaues.yaml | 19:42 |
| kfox1111 | arg... more meeting.s got one at noon, and one at 1. :/ | 19:42 |
| kfox1111 | sbezverk_: should be overridden by the one on the cli. | 19:43 |
| kfox1111 | end user always trumps package values. | 19:43 |
| *** kristian__ has joined #openstack-kolla | 19:43 | |
| kfox1111 | the package values are defaults. | 19:43 |
| sbezverk_ | kfox1111: I will copy this code and insert in in rabbitmq ps and test it.. if it works I remove it and ack this ps | 19:43 |
| kfox1111 | sbezverk_: perfect. thanks. :) | 19:43 |
| portdirect__ | kfox1111: yeah I ended up fighting for a lot longer than I expected with it :) - should have shouted for help at the weekend - but have my head round it pretty well now - I'm gonna get keystone packaged up before getting back on with it, but my plan is to implement your ceph within helm (providing a 1-1 upgrade path from the existing, so rabbitmq/mariadb/nova/cinder can coniunie interacting with it as they currently do) and then we can review | 19:44 |
| portdirect__ | the config generation and see how best to proceed long term. | 19:44 |
| kfox1111 | portdirect__: ok. sounds good. yeah, someone other then me knowing it is a great thing too. :) | 19:44 |
| kfox1111 | so I'm glad you stuck with it. :) | 19:44 |
| kfox1111 | I'd kind of like to see a ceph multinode variant in the gate with ceph-deploy added though. | 19:45 |
| kfox1111 | which will test what is probably going to be most common case for kolla-kubernetes. | 19:45 |
| portdirect__ | kfox1111: it's also worth checking out the work upstream - their containers have moved on a lot - and they publish k8s templates for them now. | 19:46 |
| kfox1111 | portdirect__: +1. they still need a helm / k8s stuff though, to be really useful I think? | 19:46 |
| kfox1111 | thats what I think ceph-deploy really brings to the table at the moment. | 19:47 |
| kfox1111 | it makes it much easier to acutally manage the thing past initial deploy, as well as help with the initial deploy. | 19:47 |
| sbezverk_ | kfox1111: I see this new function, but which file will call it though? | 19:47 |
| kfox1111 | helm/k8s should help with that once that part is done, but I think k8s may need to grow a fair bit first. :/ | 19:47 |
| kfox1111 | sbezverk_: the helm install. | 19:48 |
| kfox1111 | so, something like: | 19:48 |
| kfox1111 | helm install kolla/rabbit-pv --set foo=bar,... --values <(ceph_values) --name rabbit-pv ... | 19:48 |
| *** ccesario has quit IRC | 19:49 | |
| kfox1111 | <( ) in bash is a construct that runs the stuff in the () and passes it through the cli as a named unix socket at the argument, with the contents of stdout of the command in () | 19:49 |
| kfox1111 | so its basically building the values file as a temporary file. but doing it all in memory. | 19:50 |
| sbezverk_ | kfox1111: interesting, never seen it | 19:50 |
| kfox1111 | sbezverk_: yeah. picked up that trick only a couple years ago. :) | 19:50 |
| kfox1111 | been realy useful. :) | 19:50 |
| kfox1111 | you can pipe stuff around then to commands that only expect filenames. | 19:51 |
| kfox1111 | I use it with stuff like: | 19:51 |
| kfox1111 | diff -u <(cat file1 | sort) <(cat file2 | sort) | 19:51 |
| srwilkers_ | oh thats pretty neat kfox1111 | 19:52 |
| kfox1111 | diff -u <(cobbler system report --name=foo) <(cobbler system report --name=bar) | 19:52 |
| sbezverk_ | kfox1111: wow, nice trick! | 19:52 |
| kfox1111 | :) | 19:52 |
| kfox1111 | gota head out for a few hours. | 19:53 |
| jascott1 | sbezverk_ kfox1111 seeing Values.name instead of Release.Name in some charts.. that needs to be changed or I am misunderstanding the naming? | 19:53 |
| kfox1111 | bash is awesome. :) | 19:53 |
| kfox1111 | jascott1: yeah. should be .Release.Name. | 19:53 |
| kfox1111 | bbl | 19:53 |
| jascott1 | thanks | 19:53 |
| jascott1 | sbezverk_ were you having problems getting .Release.Name to work? | 19:53 |
| jascott1 | helm 2.0.2 released, upped the gRPC limit to 10mb https://github.com/kubernetes/helm/releases/tag/v2.0.2 | 19:55 |
| sbezverk_ | jascott1: I will use your ps as an example when you finish it ;-) | 19:55 |
| jascott1 | i dont know what else it needs to be honest, was looking around the other charts to see if I had missed anything | 19:56 |
| sean-k-mooney | v1k0d3n: hi are your around? | 19:56 |
| sean-k-mooney | v1k0d3n: was just looking at https://github.com/att-comdev/halcyon-vagrant-kubernetes | 19:57 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 19:57 |
| sbezverk_ | jascott1: how is your gate look like when you use your ps to deploy? | 19:58 |
| jascott1 | i need to test it with config mps | 19:59 |
| kfox1111 | yay. meeting cancled. so hour repreve. | 19:59 |
| jascott1 | nice | 19:59 |
| sbezverk_ | kfox1111: sent for review hoefully multi node will be green ;-) | 20:00 |
| kfox1111 | sbezverk_: cool. will keep an eye on it. | 20:00 |
| *** kristian__ has quit IRC | 20:04 | |
| *** portdirect__ is now known as portdirect_away | 20:04 | |
| jascott1 | should I be worried about this failing? gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv | 20:05 |
| *** Marx314 has joined #openstack-kolla | 20:05 | |
| kfox1111 | yes | 20:06 |
| jascott1 | ok | 20:06 |
| kfox1111 | got a link? | 20:06 |
| jascott1 | http://logs.openstack.org/77/405677/6/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/e15dd02/console.html | 20:06 |
| kfox1111 | yeah. seeing this every once in a while. | 20:07 |
| kfox1111 | go ahead and recheck. | 20:07 |
| kfox1111 | we're goign to try and gather some statistics for these sorts of things here: | 20:08 |
| kfox1111 | https://etherpad.openstack.org/p/kolla-kubernetes-gate-failure-tracking | 20:08 |
| sbezverk_ | that is networking issue I was talking about | 20:08 |
| sbezverk_ | Error syncing pod, skipping: failed to "SetupNetwork" | 20:08 |
| *** kristian__ has joined #openstack-kolla | 20:08 | |
| sbezverk_ | that is what tiller's pod reports | 20:09 |
| kfox1111 | this is the interesting bit: | 20:09 |
| kfox1111 | http://logs.openstack.org/77/405677/6/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/e15dd02/console.html#_2016-12-06_00_11_44_894124 | 20:09 |
| kfox1111 | kube-dns is not in a ready state. | 20:09 |
| sbezverk_ | also indicator that network is not in a good state | 20:09 |
| *** rhallisey has quit IRC | 20:10 | |
| kfox1111 | canal nodes are all up though. really weird. | 20:10 |
| portdirect_away | yeah - I found the wait_for_pods.sh caused issues when testing locally | 20:10 |
| *** portdirect_away is now known as portdirect | 20:10 | |
| kfox1111 | portdirect: how so? | 20:11 |
| portdirect | i think the timeout should be increased - to perhaps 8 mins? | 20:11 |
| sbezverk_ | kfox1111: I have seen it a few times but since I moved to 1.4.6 and the latest cni script/binaries do not see it anymore.. | 20:11 |
| portdirect | the kube sceduler backs off it's restart attempts - and so sometimes it can take 6 mins or so for th dns pod to be fully up | 20:12 |
| kfox1111 | seems like an awefully long time for the system to become stable. | 20:12 |
| portdirect | ^^ | 20:12 |
| portdirect | :( | 20:12 |
| sbezverk_ | kfox1111: another change I have done was to move to etcd 3.X for all components | 20:12 |
| *** sdake has quit IRC | 20:12 | |
| *** hfu has quit IRC | 20:12 | |
| kfox1111 | sbezverk_: interesting. | 20:12 |
| kfox1111 | so it might just resolve itself once kubeadm upgrades again? | 20:13 |
| portdirect | also rememeber that sometimes it may take longer to pull images from the docker hub than you would like - and we are not pre-pulling the calico/flannel images | 20:13 |
| kfox1111 | portdirect: the wait script does not include the time to pull. so its like 3 min I think to actually start and settle. | 20:13 |
| *** sdake has joined #openstack-kolla | 20:14 | |
| kfox1111 | we could up it further, but then it will also take longer to debug reviews where the things start but don't become ready. | 20:14 |
| sbezverk_ | gents, I am still struggling with using release.name not techincally but conceptually. Let me please explain where my problem is.. | 20:15 |
| kfox1111 | sbezverk_: k. yeah. lets talk thorugh it. | 20:15 |
| portdirect | kfox1111: sorry (I should have phraded it better) it's beacuse the wait script doesnt account for pulling that this issue is more likley to occur setting up networking | 20:16 |
| *** sdake_ has joined #openstack-kolla | 20:16 | |
| sbezverk_ | so when we use helm to deploy any components the release name will be unique right? | 20:16 |
| sbezverk_ | for each | 20:16 |
| kfox1111 | portdirect: it does acount for pulling though. it prepulls. | 20:16 |
| kfox1111 | ensuring the container is on the host before going through the 3 min wait for start loop. | 20:17 |
| portdirect | kfox1111: not for cni components though? for everything else it does | 20:17 |
| kfox1111 | sbezverk_: correct. | 20:17 |
| kfox1111 | portdirect: the cni plugin has a wait too, on the kube-system namespace. so will pull there too and wait. | 20:17 |
| kfox1111 | see setup_calico.sh | 20:17 |
| sbezverk_ | now when we create service with unique name and pod with unique name how we tie them together, before they had common name | 20:17 |
| kfox1111 | sbezverk_: ah... I see the problem. hmm... | 20:18 |
| sbezverk_ | so we would have rabbitmq for svc pod | 20:18 |
| sbezverk_ | etc | 20:18 |
| sbezverk_ | but now they are unique | 20:18 |
| portdirect | here: https://github.com/openstack/kolla-kubernetes/blob/master/tests/bin/setup_canal.sh#L13? | 20:18 |
| jascott1 | svc = release name, pod selector = release name right? | 20:18 |
| *** sdake has quit IRC | 20:19 | |
| sbezverk_ | I do not see how it can be connected logically | 20:19 |
| kfox1111 | portdirect: two lines down. | 20:19 |
| sbezverk_ | jascott1: it works in your case only becuase you do not create svc | 20:19 |
| kfox1111 | sbezverk_: yeah. I think your correct. we need a name var seperate from release in that case. :/ | 20:20 |
| sbezverk_ | otherwise your memcache pod would not be created with the same name | 20:20 |
| sbezverk_ | as already created service | 20:20 |
| sbezverk_ | kfox1111: yeah it was bugging me until today when I finally managed to formulate the problem | 20:20 |
| jascott1 | ok should we just go back to using .Values.name? | 20:21 |
| portdirect | kfox1111: sorry - I'm confused - it's that time out thats failing isn't it? the wait for pods script does not account for pulling the images and kubectl is simply launching a config which will put the pods straigt into "pending" state - hence the timeout? | 20:21 |
| sbezverk_ | jascott1: I think so at least unless somebody comes up with a bright idea, but then why we need to move out from using name still mistery to me | 20:22 |
| kfox1111 | jascott1: yeah. probably should. :/ | 20:22 |
| portdirect | kfox111: all other tests run a pull_containers.sh pror to testing: https://github.com/openstack/kolla-kubernetes/blob/master/tests/bin/ceph_workflow.sh#L40 | 20:23 |
| kfox1111 | portdirect: it does though | 20:23 |
| kfox1111 | I thought.... | 20:23 |
| jascott1 | i think the idea was to make 'helm install <chart> --name myname' work | 20:23 |
| kfox1111 | oh... | 20:24 |
| sbezverk_ | jascott1: why it is better than --set name=blah? | 20:24 |
| kfox1111 | its in the pull_containers.sh script. | 20:24 |
| kfox1111 | and thats not executed there... | 20:24 |
| kfox1111 | there should be a pull_containers.hs call before the wait_containers | 20:25 |
| jascott1 | sbezverk_ apparently its not better :) | 20:25 |
| kfox1111 | jascott1: we still need --name too. :/ | 20:26 |
| kfox1111 | so: | 20:26 |
| portdirect | kfox1111: all good :) I had to remove all the pull conatiners stuff from my local testing (as it only works on a single node) | 20:26 |
| kfox1111 | helm install kolla/foo-deployment --name foo-deployment --set name=foo | 20:26 |
| *** portdirect is now known as portdirect_away | 20:26 | |
| sbezverk_ | kfox1111: ceph thing did not work :( | 20:27 |
| *** portdirect_away has quit IRC | 20:27 | |
| kfox1111 | portdirect_away: yeah, we should add multinode support to it probably. | 20:27 |
| kfox1111 | sbezverk_: yeah, digging through the logs. its a little strange. | 20:27 |
| sbezverk_ | kfox1111: it definetly helped with multinode but on single not broke it | 20:28 |
| kfox1111 | I think the ceph stuff did work, as a lot of things that relies on rabbit worked. | 20:28 |
| kfox1111 | oh... | 20:28 |
| kfox1111 | hmm... | 20:28 |
| kfox1111 | let me parse through setup_config again. | 20:28 |
| kfox1111 | I thought it was always overriding. but maybe it was only with the multinode case.... | 20:29 |
| sbezverk_ | kfox1111: please I think we are very close :) | 20:29 |
| jascott1 | sbezverk_ i guess the real question is, will using both --name nad --set name together be confusing ;) | 20:30 |
| *** Pavo has quit IRC | 20:30 | |
| kfox1111 | jascott1: I'm not sure if its avoidable. | 20:30 |
| kfox1111 | jascott1: as if you don't set '--name', hyou will have a very hard time finding the release again to manage it. | 20:30 |
| sbezverk_ | jascott1: do we need --name? or we could change name for something like "label" | 20:31 |
| jascott1 | true | 20:31 |
| jascott1 | can we make --set name required? | 20:31 |
| sbezverk_ | .Values.label | 20:31 |
| kfox1111 | so helm delete / helm upgrade would suck, | 20:31 |
| sbezverk_ | instead | 20:31 |
| jascott1 | should/can | 20:31 |
| *** kollabot2 has joined #openstack-kolla | 20:32 | |
| kfox1111 | and helm ls would be gibberish. | 20:32 |
| kfox1111 | label has a connotation to me that this is a subset of. | 20:32 |
| kfox1111 | labels are abitrary key/value parirs a user can tag on it for query. | 20:32 |
| kfox1111 | this is just a single moniker | 20:32 |
| sbezverk_ | so you would have: --name rabit-pod --set label=rabbitmq | 20:32 |
| kfox1111 | I'd prefer to leave label alone. maybe some day we could use it to let the user pass through whole label keypairs through and laabel all the pods that way. | 20:33 |
| kfox1111 | label: | 20:33 |
| *** kollabot has quit IRC | 20:33 | |
| kfox1111 | mystuff: myvalue | 20:33 |
| jascott1 | id? | 20:33 |
| kfox1111 | then kubectl get pods -L mystuff=myvalue would work. :) | 20:33 |
| jascott1 | since its more than just a name or label | 20:33 |
| kfox1111 | id's/names are sometimes the same things... sometimes not... not sure on this one... | 20:34 |
| jascott1 | hmmm | 20:34 |
| kfox1111 | lets think more about what its being use dfor... | 20:34 |
| kfox1111 | we have some microservices.... | 20:34 |
| sbezverk_ | serviceName | 20:34 |
| kfox1111 | rabbitmq-pod rabbitmq-service. | 20:35 |
| *** Pavo has joined #openstack-kolla | 20:35 | |
| kfox1111 | they are in a way sub instances of the same thing. | 20:35 |
| kfox1111 | the instance the user is trying to instantiate. | 20:35 |
| kfox1111 | servicename might work... | 20:35 |
| kfox1111 | the instance is something like: | 20:36 |
| jascott1 | metadata.name, spec.selector.matchlabels.service, template.metadata.labels.service, are the ones I have been dealing wih | 20:36 |
| kfox1111 | nova-rabbit | 20:36 |
| kfox1111 | neutron-rabbit | 20:36 |
| sbezverk_ | :-) make sence as it tie them together | 20:36 |
| sbezverk_ | part of common serviceName | 20:36 |
| kfox1111 | where nova-rabbit has a pod, and a service. | 20:36 |
| sbezverk_ | but serviceName will nova-rabbit | 20:37 |
| kfox1111 | ok. I guess the issue I have with servicename, is it overloads the term service a bit. | 20:37 |
| sbezverk_ | same as service xxxxx-rabbit | 20:37 |
| kfox1111 | so, we're currently defining service to be the templates/packages/charts. | 20:37 |
| kfox1111 | microservices being just fignergrained those things. | 20:38 |
| kfox1111 | with this, we're talking about the instantiation of those things having a name. | 20:38 |
| kfox1111 | where one service can be instantiated multiple times. | 20:38 |
| v1k0d3n | sean-k-mooney hey man | 20:38 |
| v1k0d3n | whats up? | 20:38 |
| kfox1111 | each needing a unique name. | 20:38 |
| sean-k-mooney | v1k0d3n: i just figured it out i think | 20:39 |
| cliles | where is the reopen bug button in launchpad?! | 20:39 |
| sean-k-mooney | im trying out your vagrent file with openstack as the provider | 20:39 |
| kfox1111 | service_instance_name might be one way to word that, but is quite verbose. | 20:39 |
| kfox1111 | service_name is shorter, but is giveing a name to an instance more then a service. | 20:39 |
| kfox1111 | so instance_name might be the better short choice? | 20:40 |
| sean-k-mooney | v1k0d3n: i had to modify the vagrent file alittle to work with keystone v3 its running now | 20:40 |
| jascott1 | well it is name of an instance | 20:40 |
| sbezverk_ | kfox1111: not sure it could get confused with openstack instance | 20:41 |
| kfox1111 | sbezverk_: true. but less less overloaded still then service | 20:42 |
| kfox1111 | maybe some synonym then? | 20:42 |
| jascott1 | deployment wouldve been good :/ | 20:42 |
| kfox1111 | jascott1: yeah. that one's really heavily overloaded already too. :/ | 20:43 |
| jascott1 | they already took all the words :) | 20:43 |
| sbezverk_ | :) | 20:43 |
| sbezverk_ | srvName | 20:43 |
| kfox1111 | hehe: "thing_name" :) | 20:44 |
| kfox1111 | what we want to try and capture, is the logic that sbezverk had to think through, and get it into a name that intuitively makes that understandign come across. :/ | 20:45 |
| kfox1111 | not an easy thing. :/ | 20:45 |
| kfox1111 | heh. instance really is the closes definition I can come up with. :/ maybe the fact nova is trying to get rid of the usage of that term themselves leaves it up for use. :/ | 20:46 |
| jascott1 | component_name | 20:46 |
| *** fguillot has quit IRC | 20:47 | |
| jascott1 | since its not part of k8s vocab its prob fine right? | 20:47 |
| kfox1111 | is component used by openstack? | 20:47 |
| kfox1111 | cann't offhand say it is... | 20:47 |
| *** diogogmt has quit IRC | 20:48 | |
| sbezverk_ | kfox1111: I do not know, I get what you are saying about a serviceName, but man it matches exactly what we do.. | 20:48 |
| *** matrohon has joined #openstack-kolla | 20:49 | |
| kfox1111 | yeah.. I could be convinced to do service_name.... just looking for something else first. | 20:49 |
| kfox1111 | I just don't want to have to answer questions that have one person talking about helm --set service_name=neutron and another looking at helm install kolla/neutron | 20:50 |
| kfox1111 | the package name being neutron, but its a service package with name neutron. | 20:50 |
| sbezverk_ | servicePackageName? | 20:51 |
| kfox1111 | heh. which may have an "instance_name" that is something other then neutron. :/ | 20:51 |
| jascott1 | install_name? | 20:51 |
| jascott1 | maybe too close to release name | 20:51 |
| kfox1111 | though instance name could be used to describe the helm instance name, --name. :/ | 20:51 |
| kfox1111 | bleh. | 20:51 |
| jascott1 | werds | 20:52 |
| kfox1111 | yeah. release name. that. | 20:52 |
| kfox1111 | arg. | 20:52 |
| kfox1111 | naming stuff is hard. :/ | 20:52 |
| kfox1111 | yeah. I forgot --name is relase name. | 20:52 |
| kfox1111 | they avoided that issue themselves. | 20:52 |
| kfox1111 | so instance_name is still available there. | 20:52 |
| kfox1111 | so, there's another way to solve it.... | 20:52 |
| kfox1111 | call it more what it is. service_instance_name. | 20:53 |
| kfox1111 | and when too long, shorten it via acronym. | 20:53 |
| jascott1 | sin! | 20:53 |
| kfox1111 | helm install neutron-l3-agent --name neutron-l3-agent --set sin=neutron or whatever | 20:53 |
| kfox1111 | or, instance_name_of_service | 20:54 |
| jascott1 | crap now im back to id cause what the hell is a sin? :) | 20:54 |
| kfox1111 | inos :) | 20:54 |
| sbezverk_ | svcInstName | 20:54 |
| kfox1111 | sbezverk_: that would work too. | 20:54 |
| kfox1111 | though I hate camlecasing on the cli. | 20:54 |
| *** david-lyle has quit IRC | 20:55 | |
| kfox1111 | and svc has a bit of a connotation in k8s. | 20:55 |
| sbezverk_ | kfox1111: man you are hard to please ;-) | 20:55 |
| kfox1111 | as they shorten their term for service that way. | 20:55 |
| kfox1111 | sbezverk_: hehe. | 20:55 |
| kfox1111 | the problem is, I've had to please users way to long. :/ | 20:55 |
| kfox1111 | and a little bit of picking good monikers goes a long way towards me not having to explain things over and over and over. :/ | 20:56 |
| jascott1 | no doubt | 20:56 |
| kfox1111 | speaking of which, meeting next. :/ | 20:56 |
| kfox1111 | be back in an hour or so. | 20:56 |
| kfox1111 | I'd prefer to make a user go look up an acronym, | 20:56 |
| kfox1111 | then them assume they know what a term means, and get the wrong idea. | 20:56 |
| kfox1111 | often saves me a lot of time. :) | 20:57 |
| kfox1111 | that really kills me in the keystone project vs tenant thing. | 20:57 |
| sbezverk_ | I checked and work is synonym to service | 20:57 |
| kfox1111 | they think they know what projet means (wrongly). they dont know what a tenant is, so they ask and get an easy answer. | 20:57 |
| sbezverk_ | how about work_name | 20:57 |
| kfox1111 | gota go. will think about that one. | 20:58 |
| *** vilobhmm11 has joined #openstack-kolla | 20:58 | |
| *** vilobhmm111 has joined #openstack-kolla | 21:01 | |
| *** vilobhmm11 has quit IRC | 21:04 | |
| openstackgerrit | sayantani proposed openstack/kolla: Add new guide for single-node deployment https://review.openstack.org/404993 | 21:06 |
| openstackgerrit | James McCarthy proposed openstack/kolla-ansible: Swift data transfer should not be using the API Network https://review.openstack.org/407665 | 21:06 |
| *** schwicht has quit IRC | 21:07 | |
| *** fragatina has joined #openstack-kolla | 21:09 | |
| *** huhaoran has joined #openstack-kolla | 21:09 | |
| *** dave-mccowan has quit IRC | 21:10 | |
| sbezverk_ | meeting time, will be back in 1 hour.. | 21:11 |
| *** huhaoran has quit IRC | 21:14 | |
| *** msimonin1 has joined #openstack-kolla | 21:16 | |
| *** msimonin has quit IRC | 21:18 | |
| *** kristia__ has joined #openstack-kolla | 21:22 | |
| *** kristian__ has quit IRC | 21:22 | |
| *** schwicht has joined #openstack-kolla | 21:22 | |
| *** schwicht has quit IRC | 21:26 | |
| *** schwicht has joined #openstack-kolla | 21:27 | |
| *** Jeffrey4l has quit IRC | 21:32 | |
| *** Jeffrey4l has joined #openstack-kolla | 21:34 | |
| *** schwicht has quit IRC | 21:36 | |
| *** goldyfruit has quit IRC | 21:37 | |
| *** schwicht has joined #openstack-kolla | 21:39 | |
| *** kbyrne has quit IRC | 21:41 | |
| sean-k-mooney | v1k0d3n: out of interest how long should the initialize the kubernetes master task take with halcyon? | 21:45 |
| *** schwicht has quit IRC | 21:46 | |
| *** kbyrne has joined #openstack-kolla | 21:47 | |
| *** schwicht has joined #openstack-kolla | 21:48 | |
| *** goldyfruit has joined #openstack-kolla | 21:50 | |
| dmsimard | why | 21:52 |
| dmsimard | whyyyyyyyyyyyyyyy https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/baremetal/defaults/main.yml#L21 | 21:52 |
| dmsimard | why do you need epel-release :) | 21:52 |
| dmsimard | epel is evil | 21:52 |
| dmsimard | mandre, sdake_ ^ | 21:53 |
| sean-k-mooney | dmsimard: it was stated as a requirement in the quickstart | 21:55 |
| sean-k-mooney | dmsimard: the barmetal roles just automates that | 21:56 |
| dmsimard | sean-k-mooney: afaict that's just to retrieve ansible | 21:56 |
| dmsimard | ansible is either not in CentOS base OS or simply too old | 21:57 |
| dmsimard | http://docs.openstack.org/developer/kolla/quickstart.html ctrl+f epel | 21:57 |
| sean-k-mooney | really? i taught i installed ansible from pip... | 21:57 |
| dmsimard | I didn't want to invade bootstrap-servers too much but it looks like I'll have to take a stab at it | 21:58 |
| dmsimard | :D | 21:58 |
| dmsimard | It's sort of the expectation that ansible is installed because bootstrap-servers RUNS ansible and it's installed as part of installing kolla-ansible in the first place :P | 21:58 |
| *** schwicht has quit IRC | 21:58 | |
| sean-k-mooney | actully it does not install ansible you are correct | 21:58 |
| sean-k-mooney | i havent written boostrap-buildhost yet. | 21:59 |
| *** JRobinson__ has joined #openstack-kolla | 21:59 | |
| dmsimard | Yeah, but I mean is there any other requirement on EPEL ? You mentioned quickstart but all I see there is about Ansible itself | 21:59 |
| sean-k-mooney | that was the one that was going to prep the first node withch is the only one that needs ansible installed. | 21:59 |
| sean-k-mooney | im not sure | 22:00 |
| dmsimard | Is that role self-tested in the gate ? If I do stupid things, is that going to fail the gate ? | 22:00 |
| sean-k-mooney | i normally deploy based on ubuntu | 22:00 |
| dmsimard | Like, do you guys leverage it to prep the nodepool vm ? | 22:01 |
| sean-k-mooney | no | 22:01 |
| sean-k-mooney | its not currently tested | 22:01 |
| dmsimard | :( | 22:01 |
| sean-k-mooney | this is something i would like to test next cycle but have not had time to look at that yet | 22:01 |
| *** matrohon has quit IRC | 22:02 | |
| sean-k-mooney | dmsimard: if you spin up a new centos vm and override redhat_pkg_install: and remove epel-release then you can see do the role complete without it | 22:03 |
| sean-k-mooney | dmsimard: if it still passes the prechecks then you should not have broke anything | 22:03 |
| sean-k-mooney | dmsimard: that is effectivly the test. if you start with a vanila cloud image do all the prechecks pass after bootstrap-servers. | 22:05 |
| *** lrensing has quit IRC | 22:05 | |
| *** krtaylor has quit IRC | 22:07 | |
| *** srwilkers_ has quit IRC | 22:10 | |
| kfox1111 | back. | 22:10 |
| kfox1111 | sbezverk_: how about 'element' ? | 22:11 |
| jascott1 | i trust your meeting was of exceptional value as they usually are ;) | 22:12 |
| jascott1 | what about kolla_name? | 22:12 |
| kfox1111 | jascott1: I have been told all meetings are "work" by definition. | 22:12 |
| jascott1 | boy i agree to that | 22:12 |
| kfox1111 | jascott1: as if they arn't "getting work done", then there are folks who's whole lives are about not getting work done. ;) | 22:12 |
| kfox1111 | kolla_name is too overloaded I think. | 22:13 |
| jascott1 | its def "work" being in them sometimes | 22:13 |
| jascott1 | oh ok wasnt sure | 22:13 |
| kfox1111 | "an element is an instance of one or more k8s objects working together to acomplish one common goal" | 22:14 |
| kfox1111 | "for example, a rabbitmq instance along with its corresponding service is an element." | 22:15 |
| kfox1111 | each element is givin a unique name with the helm element_name value. | 22:15 |
| *** wwriverrat has quit IRC | 22:16 | |
| *** bmace has joined #openstack-kolla | 22:19 | |
| jascott1 | kfox1111 works for me | 22:21 |
| *** kristia__ has quit IRC | 22:21 | |
| *** kristian__ has joined #openstack-kolla | 22:21 | |
| kfox1111 | we probably should start a definition document. | 22:23 |
| *** wwriverrat has joined #openstack-kolla | 22:23 | |
| *** david-lyle has joined #openstack-kolla | 22:24 | |
| *** jtriley_ has joined #openstack-kolla | 22:24 | |
| *** schwicht has joined #openstack-kolla | 22:25 | |
| kfox1111 | oh. here we go... | 22:25 |
| kfox1111 | I got the ip wrong... | 22:26 |
| *** kristian__ has quit IRC | 22:26 | |
| *** jtriley has quit IRC | 22:27 | |
| *** msimonin1 has quit IRC | 22:29 | |
| *** jtriley_ has quit IRC | 22:29 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Add ceph values for gate https://review.openstack.org/407672 | 22:30 |
| *** jheroux has quit IRC | 22:30 | |
| *** Pavo has quit IRC | 22:30 | |
| kfox1111 | sbezverk_: when you get back, try that updated function. will have to pass "$1" to it too. | 22:30 |
| kfox1111 | I reread through setup_config.sh and relized I was looking at the wrong piece of code before. | 22:30 |
| *** dave-mccowan has joined #openstack-kolla | 22:32 | |
| jascott1 | it would be great if we declared the vars instead of sprinkling $1 and $2 | 22:32 |
| kfox1111 | ture. | 22:32 |
| kfox1111 | true | 22:32 |
| jascott1 | the gates have been very informative but had to figure those out | 22:33 |
| kfox1111 | probably should make a blueprint for it and go patch them all. | 22:33 |
| *** dave-mcc_ has joined #openstack-kolla | 22:34 | |
| kfox1111 | that part of the code has grown fairly organically. I'll admit, not my best work. :/ | 22:34 |
| jascott1 | all good man | 22:35 |
| jascott1 | it works | 22:35 |
| *** Pavo has joined #openstack-kolla | 22:35 | |
| *** mgiles has quit IRC | 22:35 | |
| jascott1 | bb in an hour | 22:35 |
| sean-k-mooney | has anyone here used halcyon to deploy kubernetes? | 22:36 |
| *** dave-mccowan has quit IRC | 22:37 | |
| *** vilobhmm111 has quit IRC | 22:40 | |
| *** vilobhmm11 has joined #openstack-kolla | 22:40 | |
| *** david-lyle has quit IRC | 22:41 | |
| *** david-lyle has joined #openstack-kolla | 22:41 | |
| *** harlowja has quit IRC | 22:43 | |
| *** harlowja has joined #openstack-kolla | 22:43 | |
| *** vilobhmm11 has quit IRC | 22:45 | |
| *** david-lyle has quit IRC | 22:46 | |
| *** adrian_otto has quit IRC | 22:50 | |
| *** saneax-_-|AFK is now known as saneax | 23:03 | |
| *** dave-mcc_ has quit IRC | 23:08 | |
| *** TxGirlGeek has quit IRC | 23:11 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: WIP: fix zaqar container https://review.openstack.org/407757 | 23:16 |
| *** vilobhmm11 has joined #openstack-kolla | 23:17 | |
| *** srwilkers_ has joined #openstack-kolla | 23:20 | |
| kfox1111 | random crazy question. has anyone ever tried rabbit or mariadb on top of cephfs? | 23:20 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: WIP: zaqar ansible https://review.openstack.org/407760 | 23:21 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: WIP: zaqar ansible https://review.openstack.org/407760 | 23:22 |
| sayantan_ | Hi everyone. I had opened a patch for minor documentation fixes for the docs in kolla. https://review.openstack.org/#/c/406534/ . tox -e docs fails without these being corrected. | 23:24 |
| sayantan_ | Will be great if someone can review it | 23:24 |
| sdake_ | yo peeps | 23:26 |
| srwilkers_ | ill take a look sayantan_ | 23:26 |
| srwilkers_ | sup sdake_ | 23:27 |
| sdake_ | just woke up | 23:27 |
| sdake_ | wiped out | 23:27 |
| *** Serlex has quit IRC | 23:27 | |
| sdake_ | how about yourself ;) | 23:27 |
| srwilkers_ | running on fumes myself | 23:27 |
| sayantan_ | srwilkers_ thanks :) | 23:27 |
| srwilkers_ | np sayantan_ :) | 23:27 |
| sbezverk_ | kfox1111: so you decided on the name? | 23:30 |
| *** krtaylor has joined #openstack-kolla | 23:30 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 23:31 |
| sayantan_ | Hi sdake_: I uploaded a new patch and incorporated the changes you suggested. I have separated the instructions for deployment using pip workflow and git clone into two different documents. I think this might serve to reduce confusions. | 23:32 |
| sayantan_ | https://review.openstack.org/#/c/404993/ | 23:32 |
| sbezverk_ | kfox1111: what about $1, where do you want me to pass? | 23:32 |
| sdake_ | satcool | 23:32 |
| sdake_ | sati'd review, but i dont review right when i wake up - and also going to y folks for dinner | 23:33 |
| sdake_ | sayantan_ i'll get to it after dinner hopefully - or possibly other core reviewers will get tot it | 23:33 |
| sayantan_ | Thanks :) | 23:34 |
| sayantan_ | No rush | 23:34 |
| srwilkers_ | ill take a look at that as well sayantan_. i cant approve anything, but can at least offer another set of eyes | 23:35 |
| sayantan_ | that would be great. I appreciate it. :) Thanks | 23:36 |
| sdake_ | srwilkers_ ya especially docs on kolla-ansible are crucal to not break | 23:37 |
| sdake_ | srwilkers_ it was very difficult getting them correct in the first place ;) | 23:37 |
| kfox1111 | sbezverk_: see the conversation above about 'element'. hows that one sound? | 23:37 |
| srwilkers_ | sdake_, most definitely. thats one of my concerns, so happy to look this over thoroughly | 23:38 |
| sbezverk_ | kfox1111: just element or element_name? | 23:39 |
| sdake_ | sayantan_ just one quick note, i didn't look throughly at your ocs but had a glance, - I have seen in other docs for other projects the distro choice can be selected and then the docs render the right thing | 23:39 |
| kfox1111 | hmm... | 23:39 |
| sdake_ | sayantan_ i think if we could do that that would be huge | 23:39 |
| kfox1111 | I guess element by itself would work. | 23:39 |
| sdake_ | sayantan_ i'm not sure how to do that- but if you can figure it out that would rock ;) | 23:40 |
| sdake_ | ok gotta get rolling to dinner bbiaf | 23:40 |
| sdake_ | so... tired... | 23:40 |
| sayantan_ | sdake: i'll figure it out :) | 23:40 |
| sbezverk_ | kfox1111: I would prefer element_name as it tells what to be expected, but I will not fight over it.. | 23:40 |
| kfox1111 | element_name works for me too. | 23:41 |
| sbezverk_ | jascott1: what about you? | 23:41 |
| *** vhosakot has quit IRC | 23:42 | |
| *** absubram has quit IRC | 23:42 | |
| sbezverk_ | kfox1111: if you are ok, great.. let's use it then | 23:43 |
| kfox1111 | works for me. :) | 23:43 |
| sbezverk_ | kfox1111: I will use it in the next revesion of rabbitmq ps | 23:44 |
| kfox1111 | k | 23:44 |
| sbezverk_ | right now I am testing your new version of ceph_values | 23:44 |
| kfox1111 | ok. cool. | 23:45 |
| kfox1111 | yeah. I thought I was reading the setup_config file correctly last time. but then after rereading it realizedI was looking at the wrong place. sry about that. | 23:45 |
| *** sdake_ has quit IRC | 23:51 | |
| *** v1k0d3n has quit IRC | 23:53 | |
| sean-k-mooney | has anyone tested http://docs.openstack.org/developer/kolla-kubernetes/development-environment.html with ubuntu? | 23:58 |
| sean-k-mooney | i keep getting stuck at TASK [kube-init : initialize the kubernetes master] | 23:59 |
| *** Marcellin__ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!