| *** yingjun has joined #openstack-kolla | 00:03 | |
| *** ducttape_ has joined #openstack-kolla | 00:05 | |
| *** kbaegis has joined #openstack-kolla | 00:08 | |
| *** ducttape_ has quit IRC | 00:10 | |
| *** aagate has quit IRC | 00:10 | |
| *** itlinux_ has quit IRC | 00:12 | |
| *** itlinux has quit IRC | 00:12 | |
| *** itlinux_ has joined #openstack-kolla | 00:15 | |
| *** itlinux has joined #openstack-kolla | 00:15 | |
| *** jascott1 has joined #openstack-kolla | 00:15 | |
| *** feisky has joined #openstack-kolla | 00:17 | |
| *** itlinux_ has quit IRC | 00:18 | |
| *** itlinux has quit IRC | 00:18 | |
| *** feisky has quit IRC | 00:19 | |
| *** feisky has joined #openstack-kolla | 00:19 | |
| *** feisky has quit IRC | 00:24 | |
| *** feisky has joined #openstack-kolla | 00:24 | |
| *** feisky has quit IRC | 00:26 | |
| *** rama_y_ has quit IRC | 00:28 | |
| *** masber has joined #openstack-kolla | 00:35 | |
| *** jascott1 has quit IRC | 00:37 | |
| *** duonghq has joined #openstack-kolla | 00:43 | |
| duonghq | morning guys | 00:43 |
|---|---|---|
| *** tovin07_ has joined #openstack-kolla | 00:50 | |
| *** caowei has joined #openstack-kolla | 00:54 | |
| *** masber has quit IRC | 00:54 | |
| *** awidders_ has quit IRC | 00:55 | |
| *** awiddersheim has joined #openstack-kolla | 00:56 | |
| *** renmak_ has quit IRC | 00:57 | |
| *** masber has joined #openstack-kolla | 01:08 | |
| *** harlowja has quit IRC | 01:10 | |
| vhosakot | morning duonghq | 01:13 |
| duonghq | morning vhosakot | 01:14 |
| *** kbaegis has quit IRC | 01:35 | |
| *** zhangfei has joined #openstack-kolla | 01:39 | |
| *** ducttape_ has joined #openstack-kolla | 01:45 | |
| *** ducttape_ has quit IRC | 01:50 | |
| *** masber has quit IRC | 01:50 | |
| *** masber has joined #openstack-kolla | 01:52 | |
| *** itlinux_ has joined #openstack-kolla | 02:01 | |
| *** itlinux has joined #openstack-kolla | 02:01 | |
| *** itlinux has quit IRC | 02:03 | |
| *** itlinux_ has quit IRC | 02:03 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Stop and start container rather than use restart directly https://review.openstack.org/488259 | 02:04 |
| *** rama_y has joined #openstack-kolla | 02:12 | |
| *** unicell has joined #openstack-kolla | 02:16 | |
| *** jamesbenson has joined #openstack-kolla | 02:17 | |
| *** unicell has quit IRC | 02:21 | |
| *** ducttape_ has joined #openstack-kolla | 02:21 | |
| *** jamesbenson has quit IRC | 02:21 | |
| masber | Hi sambetts|afk are you around? | 02:22 |
| *** ducttape_ has quit IRC | 02:25 | |
| *** itlinux has joined #openstack-kolla | 02:26 | |
| *** itlinux_ has joined #openstack-kolla | 02:26 | |
| *** unicell has joined #openstack-kolla | 02:29 | |
| *** zhurong has joined #openstack-kolla | 02:30 | |
| *** unicell has quit IRC | 02:33 | |
| openstackgerrit | Yotaro Konishi proposed openstack/kolla-ansible master: Support custom_config for trove-taskmanager/conductor https://review.openstack.org/487764 | 02:33 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-ansible master: Specify 'become' to necessary tasks (general roles) https://review.openstack.org/398682 | 02:38 |
| *** vhosakot has quit IRC | 02:40 | |
| *** unicell has joined #openstack-kolla | 02:41 | |
| *** unicell has quit IRC | 02:45 | |
| *** rama_y has quit IRC | 02:46 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-ansible master: Specify 'become' for only necessary tasks (default roles) https://review.openstack.org/398684 | 02:49 |
| duonghq | Jeffrey4l, I fixed ansible-become as your comment | 02:49 |
| *** unicell has joined #openstack-kolla | 02:53 | |
| *** unicell has quit IRC | 02:57 | |
| spsurya | morning guys | 03:07 |
| duonghq | morning spsurya | 03:07 |
| spsurya | duonghq: hi.... | 03:08 |
| *** masber has quit IRC | 03:13 | |
| *** unicell has joined #openstack-kolla | 03:17 | |
| *** unicell has quit IRC | 03:22 | |
| *** emccormick has quit IRC | 03:24 | |
| sbezverk | Jeffrey4l ping | 03:24 |
| *** emccormick has joined #openstack-kolla | 03:29 | |
| *** unicell has joined #openstack-kolla | 03:30 | |
| *** zhurong has quit IRC | 03:31 | |
| *** unicell has quit IRC | 03:34 | |
| *** ramishra has joined #openstack-kolla | 03:34 | |
| *** unicell has joined #openstack-kolla | 03:42 | |
| *** rama_y has joined #openstack-kolla | 03:42 | |
| *** unicell has quit IRC | 03:46 | |
| Jeffrey4l | sbezverk, pong | 03:47 |
| Jeffrey4l | duonghq, roger. will check it. | 03:47 |
| *** masber has joined #openstack-kolla | 03:48 | |
| duonghq | Jeffrey4l, thanks | 03:52 |
| *** skramaja has joined #openstack-kolla | 03:57 | |
| spsurya | duonghq: are we testing ks rolling upgrade in this week ? | 04:02 |
| duonghq | spsurya, there is one major issue left | 04:04 |
| duonghq | we must ensure anytime there are some container still running | 04:04 |
| duonghq | with current implementation, maybe there is a period all container restart | 04:05 |
| duonghq | which in turn introduce a short downtime | 04:05 |
| duonghq | my approach is develop ansible strategy but got license issue with GPLv3 | 04:06 |
| duonghq | I'm looking for improve egonzalez's suggestion | 04:06 |
| *** itlinux_ has quit IRC | 04:07 | |
| *** itlinux has quit IRC | 04:07 | |
| spsurya | duonghq: yes then it is performance impact | 04:08 |
| *** unicell has joined #openstack-kolla | 04:09 | |
| spsurya | duonghq: but we can land it for PIke at a level | 04:09 |
| spsurya | ? | 04:09 |
| duonghq | spsurya, I'm not sure | 04:17 |
| duonghq | we should ask inc0 or Jeffrey | 04:18 |
| duonghq | can you try egonzalez's suggestion? | 04:18 |
| *** yuanying has quit IRC | 04:28 | |
| *** harlowja has joined #openstack-kolla | 04:32 | |
| *** skramaja has quit IRC | 04:32 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-kubernetes master: Updated from global requirements https://review.openstack.org/485905 | 04:33 |
| spsurya | duonghq: I will check, but major work has been done by you and egonzalez, so keeping release date int consideration , it would more helpful if you guys give more hand to it. Meanwhile i will be testing it and let you know if i get some issue related to this. | 04:34 |
| duonghq | spsurya, sure | 04:34 |
| duonghq | this afternoon I'll try egonzalez suggestion, it's quite easy and straightforward | 04:34 |
| spsurya | duonghq: thanks | 04:35 |
| spsurya | duonghq: nice | 04:35 |
| duonghq | but I cannot test the logic on large scale | 04:35 |
| spsurya | you mean for n-node s ? | 04:35 |
| duonghq | yes | 04:36 |
| duonghq | 2-3 nodes are not enough | 04:36 |
| duonghq | I think this solution will make large performance impact | 04:36 |
| spsurya | i also got two node of 32GB | 04:36 |
| spsurya | only | 04:36 |
| duonghq | anw, it's a working option | 04:36 |
| duonghq | without headache of gplv3 | 04:36 |
| duonghq | I'm reading mail thread about license | 04:37 |
| spsurya | duonghq i am also checking of gplv3 licence | 04:41 |
| duonghq | spsurya, nice | 04:41 |
| *** yuanying has joined #openstack-kolla | 04:42 | |
| *** jamesbenson has joined #openstack-kolla | 04:45 | |
| *** caowei has quit IRC | 04:58 | |
| *** jamesbenson has quit IRC | 05:02 | |
| *** jamesbenson has joined #openstack-kolla | 05:03 | |
| *** emccormick has quit IRC | 05:04 | |
| *** jamesbenson has quit IRC | 05:07 | |
| *** janki has joined #openstack-kolla | 05:16 | |
| *** harlowja has quit IRC | 05:28 | |
| *** zhurong has joined #openstack-kolla | 05:30 | |
| *** caowei has joined #openstack-kolla | 05:36 | |
| *** renmak_ has joined #openstack-kolla | 05:36 | |
| *** harlowja has joined #openstack-kolla | 05:41 | |
| *** kevzha01 has joined #openstack-kolla | 05:47 | |
| *** omenv has joined #openstack-kolla | 05:51 | |
| *** ducttape_ has joined #openstack-kolla | 05:52 | |
| *** ducttape_ has quit IRC | 05:56 | |
| *** skramaja has joined #openstack-kolla | 06:00 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/kolla-ansible master: use internalURL to connect to cinder in magnum https://review.openstack.org/488300 | 06:03 |
| *** yuanying has quit IRC | 06:05 | |
| *** yuanying has joined #openstack-kolla | 06:06 | |
| *** lan has joined #openstack-kolla | 06:06 | |
| *** yuanying_ has joined #openstack-kolla | 06:07 | |
| *** zhubingbing_ has quit IRC | 06:08 | |
| *** yuanying has quit IRC | 06:10 | |
| *** magicboiz has joined #openstack-kolla | 06:15 | |
| *** magicboiz has quit IRC | 06:15 | |
| *** magicboiz has joined #openstack-kolla | 06:15 | |
| *** lan has quit IRC | 06:16 | |
| *** lan has joined #openstack-kolla | 06:17 | |
| *** yeyon has joined #openstack-kolla | 06:22 | |
| *** magicboiz has quit IRC | 06:24 | |
| *** magicboiz has joined #openstack-kolla | 06:27 | |
| duonghq | spsurya, can review this: https://review.openstack.org/#/c/398682/ | 06:30 |
| duonghq | this is 2-cycle old bp | 06:31 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: Stop container before container is removed https://review.openstack.org/488306 | 06:31 |
| spsurya | duonghq: sure | 06:31 |
| *** magicboiz has quit IRC | 06:31 | |
| *** magicboiz has joined #openstack-kolla | 06:32 | |
| *** zhubingbing has joined #openstack-kolla | 06:41 | |
| *** kevzha01 has quit IRC | 06:41 | |
| *** harlowja has quit IRC | 06:48 | |
| *** csuttles_ has quit IRC | 06:50 | |
| yeyon | i use kolla-ansible prechecks ,but not check successful, error message:TASK[rabbitmq:check if all rabbit hostnames are resolvable] failed:[localhost] (item=localhost)=>{.....,"failed":true,"item":"localhost"} | 06:50 |
| *** ducttape_ has joined #openstack-kolla | 06:52 | |
| yeyon | is my hosts file error? | 06:53 |
| mewald1 | Is kolla able to merge json just like ini? | 06:54 |
| *** ducttape_ has quit IRC | 06:54 | |
| *** ducttape_ has joined #openstack-kolla | 06:55 | |
| *** jamesbenson has joined #openstack-kolla | 06:56 | |
| duonghq | mewald1, what do you mean? | 06:58 |
| *** ducttape_ has quit IRC | 06:59 | |
| *** ducttape_ has joined #openstack-kolla | 07:00 | |
| *** jamesbenson has quit IRC | 07:00 | |
| mewald1 | duonghq: Kolla can merge INI files to generate config files for example for OpenStack. Does that work with json already? | 07:00 |
| duonghq | mewald1, not yet | 07:01 |
| *** ducttape_ has quit IRC | 07:04 | |
| *** masber has quit IRC | 07:13 | |
| *** masber has joined #openstack-kolla | 07:20 | |
| *** masber has quit IRC | 07:38 | |
| *** paramite has joined #openstack-kolla | 07:39 | |
| *** cristicalin has joined #openstack-kolla | 07:41 | |
| *** mgoddard has joined #openstack-kolla | 07:42 | |
| *** egonzalez has joined #openstack-kolla | 07:47 | |
| *** cristicalin has quit IRC | 08:00 | |
| *** ducttape_ has joined #openstack-kolla | 08:01 | |
| *** ducttape_ has quit IRC | 08:03 | |
| *** ducttape_ has joined #openstack-kolla | 08:03 | |
| *** strigazi_ is now known as strigazi | 08:04 | |
| *** ducttape_ has quit IRC | 08:07 | |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 08:10 |
| *** ducttape_ has joined #openstack-kolla | 08:10 | |
| *** ducttap__ has joined #openstack-kolla | 08:12 | |
| *** ducttape_ has quit IRC | 08:12 | |
| kolla-slack1 | <adrianreza> hello guys, I tried deploy 3 nodes of kolla-k8s, 1 node as control plane, and 2 nodes as compute node. when I deployed mariadb, it always CrashLoopBackOff until now. here is the logs http://paste.ubuntu.com/25189188/ I’m unable to find what’s wrong with my installation. | 08:14 |
| *** ducttap__ has quit IRC | 08:16 | |
| *** openstackgerrit has quit IRC | 08:18 | |
| *** sambetts|afk is now known as sambetts | 08:23 | |
| *** rmart04 has joined #openstack-kolla | 08:25 | |
| *** yuanying_ has quit IRC | 08:35 | |
| *** yuanying has joined #openstack-kolla | 08:35 | |
| *** yuanying has quit IRC | 08:35 | |
| *** yuanying has joined #openstack-kolla | 08:36 | |
| *** openstackgerrit has joined #openstack-kolla | 08:39 | |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 08:39 |
| *** serlex has joined #openstack-kolla | 08:40 | |
| *** caowei_ has joined #openstack-kolla | 08:41 | |
| *** caowei has quit IRC | 08:41 | |
| *** caowei_ is now known as caowei | 08:41 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: Assume the container is removed if it is not show in docker ps https://review.openstack.org/488306 | 08:46 |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 08:49 |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 08:51 |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 08:52 |
| yeyon | hi! I use kolla-ansible prechecks ,but not check successful, error message:TASK[rabbitmq:check if all rabbit hostnames are resolvable] failed:[localhost] (item=localhost)=>{.....,"failed":true,"item":"localhost"},i do not know where is error?is my hosts file error? | 08:55 |
| *** mewald has joined #openstack-kolla | 08:57 | |
| *** mewald1 has quit IRC | 09:00 | |
| *** ducttape_ has joined #openstack-kolla | 09:00 | |
| *** ducttap__ has joined #openstack-kolla | 09:03 | |
| *** ducttape_ has quit IRC | 09:03 | |
| *** ducttap__ has quit IRC | 09:07 | |
| *** rbergero1 is now known as rbergeron | 09:17 | |
| *** rbergeron has joined #openstack-kolla | 09:17 | |
| egonzalez | yeyon, check if the host can resolve its own hostname | 09:19 |
| egonzalez | *by api interface IP | 09:19 |
| duonghq | egonzalez, can you review this one: https://review.openstack.org/#/c/398682/ | 09:20 |
| egonzalez | duonghq, can the last ensure dir permissions task be merged with the dir creation? | 09:25 |
| egonzalez | should not be a problem if not using recursive in the task | 09:26 |
| duonghq | egonzalez, this is just for ensure permission, it's not necessary, indeed | 09:26 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Stop and start container rather than use restart directly https://review.openstack.org/488259 | 09:26 |
| *** cristicalin has joined #openstack-kolla | 09:26 | |
| duonghq | due to every file is set proper permission and owner | 09:26 |
| *** srnbckr has joined #openstack-kolla | 09:26 | |
| egonzalez | duonghq, is looking good, just a missing mode in one file and the dir comment | 09:27 |
| duonghq | egonzalez, thanks, so last task can be safety removed? | 09:28 |
| egonzalez | duonghq, I think so, but move the permission to the dir creation | 09:28 |
| duonghq | sure | 09:28 |
| duonghq | thank you | 09:29 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Assume the container is removed if it is not show in docker ps https://review.openstack.org/488346 | 09:30 |
| *** duonghq has quit IRC | 09:30 | |
| openstackgerrit | Chiawei Xie proposed openstack/kolla-kubernetes master: Fix Deployment Guide for kolla-kubernetes https://review.openstack.org/488349 | 09:33 |
| *** cristicalin has quit IRC | 09:37 | |
| *** wahnsinn_ has joined #openstack-kolla | 09:38 | |
| *** srnbckr has quit IRC | 09:39 | |
| *** prateek has joined #openstack-kolla | 10:01 | |
| *** ducttape_ has joined #openstack-kolla | 10:03 | |
| *** ducttap__ has joined #openstack-kolla | 10:06 | |
| jbadiapa | egonzalez, have you got a minute? | 10:07 |
| *** tovin07_ has quit IRC | 10:07 | |
| *** ducttape_ has quit IRC | 10:08 | |
| *** mnasiadka has joined #openstack-kolla | 10:09 | |
| *** yingjun has quit IRC | 10:10 | |
| *** ducttap__ has quit IRC | 10:10 | |
| egonzalez | jbadiapa, yep | 10:12 |
| jbadiapa | regarding the issue with the gates on the fluentd patch | 10:13 |
| jbadiapa | egonzalez^ | 10:13 |
| jbadiapa | egonzalez, the problem is that the kolla-ansible gate wont work due to the problem with the td-agent path, is taht right? | 10:15 |
| egonzalez | jbadiapa, yep, if the images are merged, ansible cannot find fluentd executable | 10:15 |
| jbadiapa | egonzalez, the main thing is that the -e genconfig does not look at the containers as far I could find out | 10:16 |
| egonzalez | jbadiapa, genconfig only generates config files based on ansible variables and templates | 10:17 |
| egonzalez | thats not the problem, ansible change fix the executable path, problem is if image change is merged will broke ansible gates until ansible change is merged | 10:18 |
| egonzalez | so, ansible change depends-on kolla change, need to be reviewed and +w (wont not merged until images are merged) | 10:19 |
| jbadiapa | egonzalez, I know, the two approach would be 1.- change the dockerfile to make work on the current gate or 2.- change the kolla-ansible first to make work on both cases | 10:24 |
| egonzalez | jbadiapa, current approach is 2 | 10:25 |
| jbadiapa | egonzalez, perfect, that's what I thought as well | 10:25 |
| egonzalez | we need reviews on ansible change too to allow merge images | 10:26 |
| egonzalez | https://review.openstack.org/#/c/482528/ | 10:27 |
| *** daidv has quit IRC | 10:31 | |
| jbadiapa | ack, I reviewed a couple of times | 10:32 |
| *** jamesbenson has joined #openstack-kolla | 10:32 | |
| *** caowei has quit IRC | 10:33 | |
| *** jamesbenson has quit IRC | 10:36 | |
| openstackgerrit | Mathias Ewald proposed openstack/kolla master: Add sensu images https://review.openstack.org/486379 | 10:37 |
| mewald | I still got this redis role that needs some review :) https://review.openstack.org/#/c/486304/ | 10:38 |
| sbezverk | Jeffrey4l have you played with selinux labeling for containers? | 10:39 |
| openstackgerrit | Mathias Ewald proposed openstack/kolla-ansible master: Add sensu role https://review.openstack.org/488367 | 10:39 |
| Jeffrey4l | sbezverk, sorry. no. i know a little about selinux. | 10:41 |
| sbezverk | Jeffrey4l thanks will keep looking then.. | 10:41 |
| *** kornicameister has quit IRC | 10:45 | |
| *** kornicameister has joined #openstack-kolla | 10:46 | |
| *** lan has quit IRC | 10:52 | |
| *** zhurong has quit IRC | 10:53 | |
| openstackgerrit | Mathias Ewald proposed openstack/kolla master: Add sensu images https://review.openstack.org/486379 | 10:56 |
| *** aolwas has quit IRC | 11:02 | |
| *** kornicameister has quit IRC | 11:03 | |
| *** aolwas has joined #openstack-kolla | 11:03 | |
| *** masber has joined #openstack-kolla | 11:04 | |
| *** renmak_ has quit IRC | 11:06 | |
| openstackgerrit | Javier Castillo proposed openstack/kolla-ansible master: Fix designate interface selection https://review.openstack.org/488329 | 11:06 |
| *** ducttape_ has joined #openstack-kolla | 11:07 | |
| masber | mgoddard, sorry I had to go yesterday, are you available to talk? | 11:09 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Fixes horizon reconfigure for adding custom *_policy.json files https://review.openstack.org/488377 | 11:09 |
| mgoddard | hi masber: yes I'm available | 11:09 |
| *** ducttape_ has quit IRC | 11:11 | |
| *** cristicalin has joined #openstack-kolla | 11:15 | |
| *** rama_y has quit IRC | 11:18 | |
| *** cristicalin has quit IRC | 11:20 | |
| *** emccormick has joined #openstack-kolla | 11:27 | |
| *** skramaja has quit IRC | 11:37 | |
| mewald | can someone help me figure out what failed here https://review.openstack.org/#/c/488367/ ? | 11:37 |
| *** prateek_ has joined #openstack-kolla | 11:37 | |
| mewald | I cannot see it from the console.html | 11:38 |
| *** prateek has quit IRC | 11:40 | |
| *** zhangfei has quit IRC | 11:40 | |
| *** zhangfei has joined #openstack-kolla | 11:41 | |
| *** zhangfei has quit IRC | 11:45 | |
| *** markhargrove28 has joined #openstack-kolla | 11:46 | |
| egonzalez | mewald, http://logs.openstack.org/67/488367/1/check/gate-kolla-ansible-pep8-ubuntu-xenial/6327d12/console.html#_2017-07-28_10_43_31_208103 | 11:49 |
| openstackgerrit | Mathias Ewald proposed openstack/kolla-ansible master: Add sensu role https://review.openstack.org/488367 | 11:49 |
| mewald | egonzalez: yeah I found that but it gives no hint (at least for me) what the real problem is. "No JSON object could be decoded" ok, but the rolls out and sensu takes it and runs … Also it's not even a json but a jinja :D | 11:51 |
| mewald | json lint says my json is valid | 11:52 |
| *** yeyon has quit IRC | 11:54 | |
| *** yingjun has joined #openstack-kolla | 11:57 | |
| *** yingjun has quit IRC | 11:58 | |
| *** ducttape_ has joined #openstack-kolla | 12:05 | |
| *** rwallner has joined #openstack-kolla | 12:08 | |
| *** ducttape_ has quit IRC | 12:11 | |
| *** rhallisey has joined #openstack-kolla | 12:12 | |
| *** awiddersheim has quit IRC | 12:14 | |
| *** dosaboy has quit IRC | 12:14 | |
| *** awiddersheim has joined #openstack-kolla | 12:15 | |
| *** strigazi is now known as strigazi_AFK | 12:20 | |
| *** magicboiz has quit IRC | 12:27 | |
| *** bmace has quit IRC | 12:29 | |
| *** bmace has joined #openstack-kolla | 12:30 | |
| *** goldyfruit has joined #openstack-kolla | 12:35 | |
| *** ducttape_ has joined #openstack-kolla | 12:38 | |
| *** ducttape_ has quit IRC | 12:39 | |
| *** ducttape_ has joined #openstack-kolla | 12:41 | |
| *** cristicalin has joined #openstack-kolla | 12:42 | |
| *** prateek_ has quit IRC | 12:43 | |
| rwellum | ping sbezverk | 12:48 |
| sbezverk | rwellum pong | 12:49 |
| *** cristicalin has quit IRC | 12:53 | |
| *** ducttape_ has quit IRC | 12:54 | |
| *** magicboiz has joined #openstack-kolla | 12:55 | |
| rwellum | Morning sbezverk, regarding your review comment - there is no 'orchestration/scripts' location - so you want me to create this? And maybe move the AIO shell script here too? | 12:55 |
| *** magicboiz has quit IRC | 12:55 | |
| sbezverk | rwellum you need to add (mkdir) it in your PS | 12:56 |
| rwellum | Yeah the process I am ok with - | 12:56 |
| *** ducttape_ has joined #openstack-kolla | 12:56 | |
| rwellum | I thought the location existed somewhere and I couldn't see it. | 12:57 |
| *** dosaboy has joined #openstack-kolla | 12:57 | |
| *** markhargrove281 has joined #openstack-kolla | 12:57 | |
| *** rwallner has quit IRC | 12:58 | |
| *** markhargrove28 has quit IRC | 12:58 | |
| *** rwallner has joined #openstack-kolla | 12:58 | |
| *** ducttape_ has quit IRC | 13:01 | |
| *** jaosorior has quit IRC | 13:04 | |
| *** schwicht has joined #openstack-kolla | 13:07 | |
| masber | mgoddard, sorry I was rebuilding my openstack test environment | 13:08 |
| mgoddard | no problem. how can I help? | 13:08 |
| masber | mgoddard, I have a kolla-ansible environment and I also put Ironic, I enabled automatic clean and setup the clean network on the ironic.conf file | 13:09 |
| masber | then I created a node | 13:09 |
| masber | the node start the clean state | 13:10 |
| masber | and reboots | 13:10 |
| masber | but I don't know why it boots to local disk, my feeling is that the ramdisk is not deployed and I don't if the issue is the ironic configuration with the randisk or pxe network | 13:11 |
| masber | deploy logs are missing | 13:11 |
| masber | so I don't know how to troubleshoot | 13:12 |
| masber | or what to do next, I was wondering if you have any advise or idea? | 13:13 |
| *** prateek has joined #openstack-kolla | 13:16 | |
| mgoddard | well first of all I'd try to get console access to the node and watch it during cleaning | 13:19 |
| mgoddard | is it a baremetal or 'fake' vm? | 13:20 |
| masber | mgoddard, it is a baremetal | 13:21 |
| *** jamesbenson has joined #openstack-kolla | 13:21 | |
| masber | I took a screenshot during pxe boot | 13:21 |
| masber | let me show to you | 13:21 |
| masber | https://ibin.co/3UvkyRGbsjnq.png | 13:22 |
| mgoddard | that's a little odd - it looks like it's found the pxe config via IP address but it should find it via MAC | 13:24 |
| mgoddard | anyway, looks like PXE has failed | 13:24 |
| *** jtriley has joined #openstack-kolla | 13:25 | |
| *** jamesbenson has quit IRC | 13:25 | |
| mgoddard | what does your network look like? flat or VLAN? | 13:25 |
| masber | public network is flat | 13:25 |
| mgoddard | and the cleaning network? | 13:26 |
| masber | http://paste.openstack.org/show/616859/ | 13:27 |
| *** rama_y has joined #openstack-kolla | 13:27 | |
| mgoddard | I thought you said you had configured the cleaning network in ironic.conf? | 13:27 |
| masber | I put the public network as cleaning so it is flat | 13:28 |
| masber | yes | 13:28 |
| masber | this is my ironic.conf file http://paste.openstack.org/show/616861/ | 13:29 |
| masber | and my network configuration http://paste.openstack.org/show/616862/ | 13:29 |
| openstackgerrit | Bradley Jones proposed openstack/kolla-kubernetes master: Deploy playbook using microservices https://review.openstack.org/473588 | 13:31 |
| mgoddard | ok. Do you normally boot bare metal instances on the public network? | 13:31 |
| openstackgerrit | Bradley Jones proposed openstack/kolla-kubernetes master: Deploy playbook using microservices https://review.openstack.org/473588 | 13:32 |
| openstackgerrit | Rich Wellum proposed openstack/kolla-kubernetes master: Moved ko.py to orchestration/tools https://review.openstack.org/487972 | 13:33 |
| masber | mgoddard, I apologize, I don't understand the question | 13:33 |
| mgoddard | when you do nova boot for bare metal instances, which network do they use? | 13:34 |
| *** markhargrove28 has joined #openstack-kolla | 13:37 | |
| *** prateek has quit IRC | 13:37 | |
| rwellum | sbezverk: this is in kolla-ansible quickstart.rst - clearly says to bring up the second interface. Even though I've been told that will break neutron? https://www.irccloud.com/pastebin/xBj3ka1z/ | 13:37 |
| *** markhargrove281 has quit IRC | 13:39 | |
| sbezverk | rwellum not sure what to say.. | 13:39 |
| rwellum | sbezverk: you think this is wrong though? I'll ping inc0 or egonzalez and see what they think. | 13:40 |
| masber | mgoddard, no idea, how can I know? | 13:41 |
| sbezverk | rwellum from what I remember network interface should not have any ip, it is controlled by neutron | 13:41 |
| sbezverk | check with them please | 13:41 |
| egonzalez | rwellum, network_interface is for API traffic, should be UP and with IP | 13:42 |
| egonzalez | rwellum, neutron_external_interface is for openstack public network, interface UP with no IP assigned | 13:43 |
| rwellum | egonzalez: you're suggesting you need 3 NIC's/Interfaces? | 13:43 |
| sbezverk | rwellum I was talking about latter | 13:43 |
| rwellum | Yes sbezverk | 13:43 |
| rwellum | If the doc says 2 interfaces with IP addresses then you need a 3rd that is not up for neutron_external_interface? | 13:44 |
| egonzalez | rwallner, need minimun 2, one for all services/tunnel,data interfaces (network_interface) and other for neutron_external_interface | 13:44 |
| egonzalez | rwellum, sorry rwallner again | 13:45 |
| rwellum | egonzalez: :) | 13:45 |
| *** jistr is now known as jistr|mtg | 13:45 | |
| rwellum | egonzalez: in the text above, what is that second interface - neutron_external_interface? | 13:46 |
| egonzalez | rwellum, yep, check https://docs.openstack.org/kolla-ansible/latest/production-architecture-guide.html | 13:47 |
| *** ducttape_ has joined #openstack-kolla | 13:47 | |
| egonzalez | there is a brief explanation of the interfaces used | 13:48 |
| egonzalez | as sort resume, every interface is pointing to network_interface except neutron_external_interface that __should__ be a different NIC | 13:48 |
| *** mgoddard has quit IRC | 13:48 | |
| *** ducttape_ has quit IRC | 13:49 | |
| *** ducttape_ has joined #openstack-kolla | 13:49 | |
| rwellum | egonzalez: thanks - but then that readme (the first one I pointed out is wrong) - if that is neutron external - it should NOT be up. That's the piece I think is wrong/confusing. | 13:50 |
| sbezverk | rwellum neutron external must be up | 13:50 |
| sbezverk | I collapse all other interfaces onto the management one | 13:51 |
| rwellum | sbezverk: egonzalez said quote: "rwellum, neutron_external_interface is for openstack public network, interface UP with no IP assigned" | 13:52 |
| rwellum | Am I going crazy here.... | 13:52 |
| sbezverk | rwellum that is it what I told you too, UP and no IP | 13:53 |
| *** prateek has joined #openstack-kolla | 13:55 | |
| serlex | ha! easily done | 13:55 |
| rwellum | sbezverk: Ok sorry. Then why does the doc say: "Verify if the second interface has an IP address" | 13:56 |
| rwellum | Maybe it should be "Verify if the second interface is UP and does not have an IP address" | 13:56 |
| sbezverk | rwellum agree, it should be more explicit which interface it is talking about | 13:57 |
| *** strigazi_AFK is now known as strigazi_OFF | 13:58 | |
| egonzalez | rwellum, my config in a working environment http://paste.openstack.org/show/616864/ | 13:59 |
| *** mnasiadka has quit IRC | 14:04 | |
| *** ramishra has quit IRC | 14:04 | |
| egonzalez | rwellum, opened a bug for the lack of network configuration docs https://bugs.launchpad.net/kolla-ansible/+bug/1707211 | 14:07 |
| openstack | Launchpad bug 1707211 in kolla-ansible "lack of network documentation" [Undecided,New] | 14:07 |
| *** prateek has quit IRC | 14:08 | |
| *** rmart04 has quit IRC | 14:10 | |
| *** jamesbenson has joined #openstack-kolla | 14:13 | |
| *** kevzha01 has joined #openstack-kolla | 14:13 | |
| *** kevzha01 has quit IRC | 14:13 | |
| *** kevzha01 has joined #openstack-kolla | 14:14 | |
| *** jtriley_ has joined #openstack-kolla | 14:16 | |
| *** jamesbenson has quit IRC | 14:17 | |
| *** jtriley has quit IRC | 14:18 | |
| *** jamesbenson has joined #openstack-kolla | 14:20 | |
| jamesbenson | morning all :-) and happy friday! | 14:22 |
| *** sbezverk has quit IRC | 14:22 | |
| *** itlinux has joined #openstack-kolla | 14:24 | |
| *** itlinux has quit IRC | 14:24 | |
| *** kevzha01_ has joined #openstack-kolla | 14:24 | |
| *** paramite has quit IRC | 14:26 | |
| *** kevzha01 has quit IRC | 14:27 | |
| *** eaguilar has joined #openstack-kolla | 14:36 | |
| *** hongbin has joined #openstack-kolla | 14:38 | |
| *** eaguilar is now known as eaguilar|afk | 14:45 | |
| *** hrw has quit IRC | 14:45 | |
| *** janki has quit IRC | 14:45 | |
| *** jistr|mtg is now known as jistr | 14:46 | |
| *** masber has quit IRC | 14:46 | |
| *** hrw has joined #openstack-kolla | 14:46 | |
| *** eaguilar|afk is now known as eaguilar | 14:46 | |
| *** sbezverk has joined #openstack-kolla | 14:50 | |
| *** schwicht has quit IRC | 14:51 | |
| *** harlowja has joined #openstack-kolla | 14:51 | |
| inc0 | good morning | 14:53 |
| *** mewald has quit IRC | 14:55 | |
| *** eaguilar is now known as eaguilar|afk | 14:57 | |
| *** eaguilar|afk is now known as eaguilar | 14:58 | |
| *** eaguilar is now known as eaguilar|afk | 14:58 | |
| *** kbaegis has joined #openstack-kolla | 14:59 | |
| kbaegis | Hi all. Is there a defect with the inventory handling for openvswitch? | 14:59 |
| kbaegis | I need kolla-ansible not to deploy ovs or the db | 15:00 |
| inc0 | kbaegis: with master it's easy | 15:01 |
| kbaegis | inc0: and by master you mean k8s? | 15:01 |
| inc0 | with ocata you need to disable it in globals | 15:01 |
| inc0 | no, I mean master kolla-ansible | 15:01 |
| kbaegis | inc0: It can't be done in inventory? | 15:01 |
| inc0 | as in not released | 15:01 |
| inc0 | no | 15:02 |
| kbaegis | thanks. That answers that | 15:02 |
| inc0 | https://github.com/openstack/kolla-ansible/blob/stable/ocata/ansible/roles/neutron/defaults/main.yml#L28 | 15:02 |
| inc0 | you need to override this var | 15:02 |
| kbaegis | I just want it to use the host managed ovs, not within docker, and it obviously breaks networking when both are set up | 15:02 |
| inc0 | just make sure that volume monunts are compatible | 15:03 |
| kbaegis | Wait, it will still deploy the ovs-l2 agent right? | 15:03 |
| inc0 | yeah | 15:03 |
| inc0 | you need sth like this | 15:04 |
| inc0 | neutron_services: openvswitch-vswitchd: enabled: false | 15:04 |
| inc0 | (with indents) | 15:04 |
| *** egonzalez has quit IRC | 15:10 | |
| *** kiennt has joined #openstack-kolla | 15:15 | |
| *** eaguilar|afk is now known as eaguilar | 15:16 | |
| openstackgerrit | Bradley Jones proposed openstack/kolla-kubernetes master: Deploy playbook using microservices https://review.openstack.org/473588 | 15:28 |
| *** s-dean has joined #openstack-kolla | 15:29 | |
| inc0 | kfox1111_ sbezverk quick question...you are aware that we're using secrets for passwords and whatnot but they're also written in plain configmaps right? | 15:31 |
| inc0 | genconfig puts passwords where passwords should be put | 15:32 |
| sbezverk | inc0: password are encoded | 15:32 |
| inc0 | by what? | 15:33 |
| sbezverk | by kube | 15:33 |
| inc0 | not in configs | 15:33 |
| sbezverk | when you create a secret | 15:33 |
| inc0 | what I'm saying is | 15:33 |
| inc0 | when you run ansible genconfig | 15:33 |
| inc0 | it will put password in config | 15:33 |
| inc0 | so while secrets are secure, passwords themselves are visible in configmap | 15:34 |
| inc0 | just so you're aware | 15:34 |
| sbezverk | ah ah ok, you mean in opensatck config | 15:34 |
| inc0 | yes | 15:34 |
| sbezverk | but how it is different from kolla-ansible? | 15:34 |
| inc0 | but all of the passwords (maybe besides db admin passoword) are visible | 15:34 |
| inc0 | it is not | 15:34 |
| inc0 | just saying that we might want to use k8s secret management there and make it little more secure | 15:35 |
| inc0 | not | 15:35 |
| inc0 | not sure how yet, but something to keep in mind | 15:35 |
| sbezverk | inc0 that would be great, but it would then require preprocessing | 15:35 |
| sbezverk | we could add decrypt init container | 15:35 |
| inc0 | alternatively we could remove passwords from config | 15:36 |
| sbezverk | to every pod to decrypt config password | 15:36 |
| inc0 | and create secrets | 15:36 |
| sbezverk | yes, but we still need to add them before the service starts | 15:36 |
| inc0 | and run nova-compute --config-file /etc/nova/nova.conf --config-file=/etc/nova/secrets.conf | 15:37 |
| sbezverk | inc0: I do not follow how this will help | 15:37 |
| inc0 | secret will effectively be another config file | 15:38 |
| inc0 | with just secret-related configs in it | 15:38 |
| inc0 | but not just a password, .conf file with group header and whatnot | 15:38 |
| *** rama_y has quit IRC | 15:38 | |
| inc0 | https://paste.fedoraproject.org/paste/hIexHaepmkLrtKJAEb8bgA like this | 15:39 |
| *** itlinux has joined #openstack-kolla | 15:40 | |
| inc0 | oh btw sbezverk your registry chart is awesome | 15:40 |
| *** itlinux_ has joined #openstack-kolla | 15:40 | |
| inc0 | gave me few ideas | 15:40 |
| inc0 | I'll publish something today, if I'm right we can make prep before kolla-k8s deploy super easy | 15:40 |
| inc0 | configmaps, secrets and chart building that is | 15:40 |
| rwellum | inc0: +1 | 15:41 |
| kfox1111_ | morning. | 15:46 |
| sbezverk | inc0: great, let's see | 15:47 |
| sbezverk | kfox1111: morning | 15:47 |
| sbezverk | kfox1111: I found the source of that bug | 15:47 |
| kfox1111_ | sbezverk: oh cool. sorry I hadn't gotten to it yet. | 15:47 |
| kfox1111_ | what was it? | 15:47 |
| *** itlinux_ has quit IRC | 15:48 | |
| sbezverk | kfox1111_ : helm/kolla-common/templates/_common_dependency_container.yaml | 15:48 |
| *** itlinux has quit IRC | 15:48 | |
| sbezverk | kfox1111_ {{- define "joinListWithColon" -}} | 15:48 |
| sbezverk | there is an inconsistency between when dependency specified in yaml file and when it is overridden from command line with set | 15:49 |
| *** eaguilar is now known as eaguilar|afk | 15:49 | |
| kfox1111_ | ah. maybe different data types? | 15:50 |
| kfox1111_ | I know it tries to automatically figure out data types if done via --set and I have ran into problems with that. | 15:50 |
| sbezverk | here is how data gets represented for both cases: | 15:50 |
| sbezverk | https://paste.fedoraproject.org/paste/fsSW0C71NlahUH1lzyc36w | 15:50 |
| kfox1111_ | I see why that may fail for the use case. | 15:51 |
| *** harlowja has quit IRC | 15:51 | |
| kfox1111_ | it was written before we had 'set' I think. | 15:51 |
| kfox1111_ | I have a different way of doing it now... | 15:51 |
| sbezverk | kfox1111_ oh great, if you have the fix, let's fix it | 15:51 |
| kfox1111_ | oh... | 15:52 |
| kfox1111_ | though that looks weird. | 15:52 |
| *** kevzha01_ has quit IRC | 15:53 | |
| kfox1111_ | try changing the macro contents to be: | 15:54 |
| kfox1111_ | {{- $local = dict "first" true }} | 15:54 |
| kfox1111_ | {{ range $k, $v := . }}{{ if $local.first }},{{ end }}{{ $v }}{{ $_ = set $local "first" false }}{{ end }} | 15:54 |
| sbezverk | kfox1111_ no kidding, if you specify in yam, it looks like map[string]string | 15:54 |
| kfox1111_ | map[string]string is the desired result. | 15:54 |
| kfox1111_ | --set may be converting the "0" to be int. | 15:55 |
| kfox1111_ | and then the range isn't sure what to do with it. | 15:55 |
| sbezverk | but second looks like map[[]int]string | 15:55 |
| *** mewald has joined #openstack-kolla | 15:56 | |
| kfox1111_ | what was the --set you used? | 15:56 |
| kbaegis | inc0: :/ didn't work | 15:57 |
| inc0 | really, what happened? | 15:57 |
| sbezverk | --set global.kolla.mariadb.statefulset.dependencies.jobs.[0]="mariadb-storage-init-element" --set global.kolla.mariadb.statefulset.dependencies.jobs.[1]="mariadb-storage-init-element-2" | 15:57 |
| inc0 | also show me how did you define it in globals | 15:57 |
| kbaegis | I went directly to the main.yml and set to false | 15:58 |
| jamesbenson | inc0: I'm going to open a bug about prechecks passing when external ceph is still missing files in the config dir. | 15:58 |
| kfox1111_ | sbezverk: not sure thats the right syntax for that. but not sure what it would be otherwise. | 15:59 |
| kfox1111_ | maybe: dependencies.jobs[1] = ? | 15:59 |
| inc0 | damn | 15:59 |
| sbezverk | kfox1111_ that is the only way to override lists | 15:59 |
| kfox1111_ | not sure --set is even capabale of defining this. | 15:59 |
| sbezverk | kfox1111_ technosophos told me that it was right syntax | 16:00 |
| kfox1111_ | ok. | 16:00 |
| jamesbenson | inc0: Also, at least one of the files that's required could be generated automatically, nothing deployment specific about it, do you want ansible to generate the file automatically? | 16:00 |
| kfox1111_ | looks really weird. but if he said thats the way it should work, I believe him. | 16:00 |
| inc0 | kbaegis: well, worst case scenario you can just remove ovs starting from here: https://github.com/openstack/kolla-ansible/blob/stable/ocata/ansible/roles/neutron/handlers/main.yml#L2 | 16:00 |
| kfox1111_ | try the above different macro and see what happens. | 16:01 |
| inc0 | but this needs more debug, disabling service should work | 16:01 |
| sbezverk | kfox1111_ trying, there were a couple of typos ;) | 16:01 |
| kfox1111_ | oh. but I think it should be 'if not $local.first' | 16:01 |
| kfox1111_ | yeah. probably. its untested. :) | 16:01 |
| inc0 | jamesbenson: that'd be good change imo | 16:01 |
| jamesbenson | inc0: the auto generate? | 16:01 |
| inc0 | if it can be autogenerated, sure why not | 16:02 |
| sbezverk | kfox1111_ you almost nailed it "value": "\n,mariadb-storage-init-elementmariadb-storage-init-element-2" | 16:02 |
| jamesbenson | cool, will do. | 16:02 |
| inc0 | less steps to work through | 16:02 |
| jamesbenson | that would happen in the deploy correct? | 16:02 |
| sbezverk | first comma is not needed and missing comma between them | 16:02 |
| jamesbenson | ino0: I'll update the docs with a note well stating that this particular file is autogenerated but this is what it looks like.... | 16:03 |
| kfox1111_ | sbezverk: yeah. add the not into the if. | 16:04 |
| kfox1111_ | should fix it then I think. | 16:04 |
| kbaegis | inc0: Yeah, "ansible/roles/neutron/defaults/main.yml" is set to false. Ansible scripts still deploy it | 16:04 |
| jamesbenson | ...sorry I've been out the past few days, I was setting up irc on slack and a really nasty stomach bug around the house.... | 16:04 |
| jamesbenson | inc0: I'll also try to do upgrade with external ceph to 5.0.0 | 16:04 |
| jamesbenson | :-) | 16:04 |
| kfox1111_ | looks like hostpath may go in finally. | 16:05 |
| inc0 | kbaegis: I'll test it out later, but it's strange | 16:05 |
| inc0 | in Pike ovs is in separate role | 16:06 |
| inc0 | so it will be as easy as "enable_openvswitch: false" | 16:06 |
| sbezverk | kfox1111_ yep excellent works for all three cases | 16:06 |
| kfox1111_ | cool. | 16:06 |
| sbezverk | in yaml, 1 parameters in command line and 2 parameters in command line | 16:06 |
| sbezverk | could you please push a fix ? | 16:07 |
| *** serlex has quit IRC | 16:07 | |
| kfox1111_ | sbezverk: can you do it? I think you mgiht be able to quicker then I can. | 16:07 |
| kfox1111_ | otherwise, I can try and do it later. | 16:07 |
| sbezverk | kfox1111_ ok, sure | 16:08 |
| kfox1111_ | thanks. :) | 16:08 |
| kbaegis | inc0: What would be the syntax in globals.yml? | 16:09 |
| *** eaguilar|afk is now known as eaguilar | 16:09 | |
| kbaegis | enable_openvswitch: "no" ;neutron_plugin_agent: "openvswitch"; and then I set the default/main.yml to enabled: false | 16:10 |
| kbaegis | *shrug* I'll just delete the handler then | 16:10 |
| inc0 | kbaegis: something like that http://paste.openstack.org/show/616883/ | 16:11 |
| inc0 | dear k8s, y u no allow to port forward to service?:( | 16:12 |
| kbaegis | k8s networking is crazy gross. | 16:12 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fix rendering problem when yaml value gets overridden via cli https://review.openstack.org/488513 | 16:13 |
| inc0 | kbaegis: well, matter of taste | 16:13 |
| inc0 | openstack does crazy things with bridges | 16:13 |
| inc0 | they do crazy things with iptables | 16:14 |
| kbaegis | inc0: That's true. | 16:14 |
| inc0 | openstack does crazy things with iptables too so I guess all and all k8s is a bit simpler | 16:14 |
| kbaegis | inc0: Yeah, my main objection to k8s right now are the inflexible networking parts | 16:14 |
| kfox1111_ | inc0: +1. please file a feature request. and ping me on it. I'll upvote it too. :) | 16:14 |
| kbaegis | or assumptions that you can't change | 16:14 |
| kbaegis | But I'm not an expert :) | 16:14 |
| *** david-lyle has quit IRC | 16:14 | |
| inc0 | kfox1111_: we're not alone https://github.com/kubernetes/kubernetes/issues/15180 | 16:15 |
| kbaegis | dns comes to mind | 16:15 |
| kolla-slack1 | <jmjava> i am also see this: No valid host was found. There are not enough hosts available in downstream openstack/helm - was there a fix for this? | 16:15 |
| kfox1111_ | inc0: thanks. upvoted and following. :) | 16:15 |
| inc0 | they have oneliner there | 16:16 |
| inc0 | to do it | 16:16 |
| inc0 | long and ugly but I guess it's something | 16:16 |
| *** david-lyle has joined #openstack-kolla | 16:16 | |
| inc0 | jmjava: I'll need some context | 16:17 |
| inc0 | what did you do?;) | 16:17 |
| *** rama_y has joined #openstack-kolla | 16:17 | |
| *** eaguilar_ has joined #openstack-kolla | 16:18 | |
| kolla-slack1 | <jmjava> sorry yesterday there was a post about nova throwing error. I got this message when trying to start an instance of cirros via Horizon UI | 16:18 |
| inc0 | ah right | 16:18 |
| inc0 | nova scheduler and compute logs shows anything? | 16:18 |
| kolla-slack1 | <jmjava> checking | 16:19 |
| *** eaguilar has quit IRC | 16:19 | |
| *** mewald has quit IRC | 16:21 | |
| *** david-lyle has quit IRC | 16:22 | |
| kbaegis | inc0: I was editing the wrong file. My venv had confused me | 16:23 |
| inc0 | heh | 16:23 |
| inc0 | that happens;) | 16:23 |
| kolla-slack1 | <jmjava> 2017-07-27 20:38:44.054 90228 ERROR nova.compute.manager [req-a2e8e209-3f23-4df4-9c4e-2935f035f4d1 - - - - -] No compute node record for host MyHostName | 16:23 |
| inc0 | check nova-compute logs on host MyHostName;) | 16:25 |
| *** jascott1 has joined #openstack-kolla | 16:27 | |
| *** jascott1 has quit IRC | 16:27 | |
| *** jascott1 has joined #openstack-kolla | 16:28 | |
| *** eaguilar_ has quit IRC | 16:32 | |
| *** unicell has quit IRC | 16:37 | |
| *** openstackstatus has quit IRC | 16:41 | |
| *** openstack has joined #openstack-kolla | 16:45 | |
| *** unicell has joined #openstack-kolla | 16:47 | |
| *** tonanhngo has joined #openstack-kolla | 16:48 | |
| *** tonanhngo_ has joined #openstack-kolla | 16:52 | |
| *** tonanhngo_ has quit IRC | 16:52 | |
| *** emccormick has quit IRC | 16:52 | |
| *** tonanhngo has quit IRC | 16:52 | |
| *** unicell has quit IRC | 16:54 | |
| kolla-slack1 | <jmjava> Ok i see https://review.openstack.org/#/c/487413/ - maybe that is the fix? | 16:54 |
| *** tonanhngo has joined #openstack-kolla | 16:54 | |
| *** harlowja has joined #openstack-kolla | 16:54 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Add OpenDaylight role https://review.openstack.org/416367 | 16:55 |
| *** renmak_ has joined #openstack-kolla | 16:55 | |
| *** harlowja_ has joined #openstack-kolla | 16:55 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Add OpenDaylight role https://review.openstack.org/416367 | 16:57 |
| *** harlowja has quit IRC | 16:59 | |
| *** sambetts is now known as sambetts|afk | 17:03 | |
| *** omenv has quit IRC | 17:08 | |
| -openstackstatus- NOTICE: The Gerrit service on review.openstack.org is being taken offline for roughly 5 minutes to perform a database backup and reconfiguration | 17:12 | |
| *** itlinux has joined #openstack-kolla | 17:14 | |
| *** junbo has quit IRC | 17:15 | |
| *** eaguilar has joined #openstack-kolla | 17:28 | |
| *** kiennt has quit IRC | 17:30 | |
| *** openstackgerrit has quit IRC | 17:33 | |
| *** renmak_ has quit IRC | 17:40 | |
| *** openstackgerrit has joined #openstack-kolla | 17:41 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: WIP Dockerfile to run ansible orchiestration https://review.openstack.org/474770 | 17:41 |
| inc0 | kfox1111_ sbezverk ^ | 17:41 |
| inc0 | so idea goes like that: do docker build ~/kolla-kubernetes --tag localhost:30400/kolla-kubernetes-orchestration | 17:42 |
| inc0 | helm install registry | 17:42 |
| inc0 | kubectl port-forward <<registry-pod>> 4000:30400 | 17:42 |
| *** eaguilar_ has joined #openstack-kolla | 17:42 | |
| inc0 | docker push localhost:30400/kolla-kubernetes-orchestration | 17:42 |
| inc0 | helm install helm/orchesration/kolla-configs | 17:42 |
| inc0 | on release we can put kolla-kubernetes-orchestration image on dockerhub and you can just use helm install | 17:43 |
| *** eaguilar has quit IRC | 17:45 | |
| sbezverk | inc0: where do you generate password.yaml? | 17:47 |
| inc0 | sbezverk: kubectl create secret --from-file=/etc/kolla/passwords.yaml | 17:47 |
| inc0 | yeah you need to do this and globals | 17:47 |
| *** unicell has joined #openstack-kolla | 17:47 | |
| sbezverk | inc0: that was secrets, but the passwords file which is used for secrets, it needs to be generated too | 17:48 |
| inc0 | it's generated in same way as in kolla-ansible | 17:48 |
| inc0 | or kolla-k8s for that matter | 17:48 |
| *** eaguilar_ is now known as eaguilar|afk | 17:48 | |
| inc0 | current secret-generator uses /etc/kolla/passwords.yml as input too | 17:49 |
| sbezverk | inc0: right it is just I did not spot the place where you generate new /etc/kolla/passwords.yml | 17:49 |
| inc0 | yeah, I don't | 17:50 |
| inc0 | I'm assuming it's there | 17:50 |
| sbezverk | inc0: that was my pount | 17:50 |
| sbezverk | if you put everything in container | 17:50 |
| sbezverk | then you need to build it too I think | 17:50 |
| inc0 | nah | 17:50 |
| inc0 | I mount it from secret | 17:51 |
| inc0 | I don't build passwords.yaml into container | 17:51 |
| *** eaguilar|afk is now known as eaguilar_ | 17:55 | |
| *** eaguilar_ is now known as eaguilar|afk | 17:55 | |
| rwellum | sbezverk: where does the cinder.conf file live? In the container? | 17:55 |
| sbezverk | inc0: Ok I see, I guess since there is no yet a part that generate a secret with all passwords | 17:55 |
| inc0 | sbezverk: it's just the same as today with kolla-k8s and ansible | 17:56 |
| inc0 | and I think that's good | 17:56 |
| inc0 | getting these 2 configs up is for users to do | 17:56 |
| inc0 | rwellum: what do you mena? | 17:56 |
| inc0 | mean | 17:57 |
| sbezverk | rwellum : the one cinder uses, yes in the container, the one you use to build configmap, is on local host | 17:57 |
| rwellum | ok | 17:57 |
| sbezverk | inc0: you have reference in your chart to secrets which supposedly contains all password, if you do not provide a way to create that k8s secret object | 17:58 |
| sbezverk | then the job will get stuck in Pending | 17:58 |
| inc0 | yeah, you need to run kubectl create secret passwords --from-file=/etc/kolla/passwords.yml | 17:59 |
| sbezverk | inc0: yep | 17:59 |
| inc0 | also create configmap globals --from-file=/rtc/kolla/globals.yml | 17:59 |
| inc0 | I'm taking same config input as kolla-k8s now | 18:00 |
| *** vhosakot has joined #openstack-kolla | 18:02 | |
| inc0 | do you like this approach? | 18:03 |
| inc0 | it's going to be easy to run chart kolla-prep | 18:03 |
| inc0 | that will run registry and all that | 18:03 |
| rwellum | sbezverk: so if i want to change the cinder.conf file that the container comes up with? | 18:03 |
| sbezverk | rwellum A bit tricky but doable | 18:06 |
| sbezverk | you need to change configmap providing cinder.conf and then delete cinder pod NOT deployment or daemonset | 18:07 |
| sbezverk | when new pod gets scheduled it will pick up updated config map | 18:07 |
| sbezverk | there is a probably better way to do it with helm upgrade but I have not used it much | 18:08 |
| sbezverk | inc0: I am personally ok with any approach which does not force us to use only this or that. | 18:09 |
| jascott1 | inc0 interesting https://www.meetup.com/Cloud-Native-PDX/events/240116992/?rv=me1 | 18:11 |
| *** openstackstatus has quit IRC | 18:11 | |
| *** openstack has joined #openstack-kolla | 18:15 | |
| sbezverk | jascott1 http://paste.openstack.org/show/616894/ | 18:19 |
| sbezverk | jascott1 there is \n right before the first value | 18:19 |
| jascott1 | sbezverk oh ok toom me a sec, thought i was looking for a newline itself not a literal \n :) | 18:21 |
| sbezverk | jascott1 :) | 18:22 |
| jascott1 | thats "weird" | 18:22 |
| sbezverk | yeah and it is consistent | 18:22 |
| sbezverk | only happens when values provided from the command line | 18:22 |
| sbezverk | I think there is a bug in command line parser | 18:23 |
| sbezverk | jascott1 hm, even if taken from yaml, it still adds \n | 18:32 |
| jascott1 | interesting sounds like a bug maybe but wonder it its in the templating lib rather than helm | 18:37 |
| jascott1 | I will check into it after I grab a bite | 18:37 |
| sbezverk | jascott1 sure | 18:37 |
| *** dave-mccowan has joined #openstack-kolla | 18:38 | |
| *** kbaegis has quit IRC | 18:39 | |
| *** openstackstatus has quit IRC | 18:42 | |
| *** openstack has joined #openstack-kolla | 18:45 | |
| *** kbaegis has joined #openstack-kolla | 19:00 | |
| *** dave-mccowan has quit IRC | 19:05 | |
| *** jamesbenson has quit IRC | 19:06 | |
| *** jamesbenson has joined #openstack-kolla | 19:08 | |
| *** dciabrin has quit IRC | 19:12 | |
| *** jamesbenson has quit IRC | 19:12 | |
| *** jamesbenson has joined #openstack-kolla | 19:16 | |
| *** jamesbenson has quit IRC | 19:16 | |
| *** kbaegis has quit IRC | 19:17 | |
| *** jamesbenson has joined #openstack-kolla | 19:20 | |
| *** ducttape_ has quit IRC | 19:24 | |
| *** ducttape_ has joined #openstack-kolla | 19:25 | |
| *** rama_y has quit IRC | 19:38 | |
| inc0 | sbezverk: registry will be available in all the gates or just master? | 19:43 |
| sbezverk | inc0: in theroy it could be used with any version as long as there is tar file for that version | 19:44 |
| *** jascott1 has quit IRC | 19:44 | |
| inc0 | well, there is tar for both ocata and master | 19:44 |
| inc0 | but what I'm asking is if it's even run in every gate | 19:44 |
| *** jascott1 has joined #openstack-kolla | 19:44 | |
| inc0 | because if it is, I can move gates to container-based genconfig | 19:44 |
| *** rama_y has joined #openstack-kolla | 19:45 | |
| *** rhallisey has quit IRC | 19:58 | |
| sbezverk | inc0: for now it is implemented only with master 2 jobs | 19:58 |
| *** mewald has joined #openstack-kolla | 20:09 | |
| *** mewald has quit IRC | 20:21 | |
| *** mewald has joined #openstack-kolla | 20:22 | |
| openstackgerrit | Mathias Ewald proposed openstack/kolla-ansible master: Add sensu role https://review.openstack.org/488367 | 20:28 |
| jascott1 | sbezverk sorry its taking me so long but I have at least reproduced in isolated test chart | 20:33 |
| jascott1 | i tihnk its the template function | 20:33 |
| jascott1 | one sec | 20:33 |
| jascott1 | sbezverk I replied at line 31 http://paste.openstack.org/show/616900/ | 20:40 |
| *** jtriley_ has quit IRC | 20:43 | |
| jascott1 | also do we not have strings.Join from sprig? by this template I assume not | 20:44 |
| *** jamesbenson has quit IRC | 20:54 | |
| *** jamesbenson has joined #openstack-kolla | 20:56 | |
| *** markhargrove28 has quit IRC | 20:58 | |
| jascott1 | sbezverk interestingly I see a join in the codebase | 20:58 |
| *** jamesbenson has quit IRC | 21:01 | |
| *** jamesbenson_ has joined #openstack-kolla | 21:02 | |
| *** jamesbenson_ has quit IRC | 21:02 | |
| *** jamesbenson has joined #openstack-kolla | 21:02 | |
| *** ducttap__ has joined #openstack-kolla | 21:11 | |
| *** ducttape_ has quit IRC | 21:14 | |
| *** ducttape_ has joined #openstack-kolla | 21:17 | |
| *** ducttap__ has quit IRC | 21:20 | |
| inc0 | hmm, can we move all 4.0 gates to registry too? | 21:27 |
| *** david-lyle has joined #openstack-kolla | 21:30 | |
| kfox1111_ | inc0: I was working on caching in the gate. if we had registries for all versions that would be ok by me. | 21:31 |
| inc0 | I dont think that's doable...Mitaka is EOL | 21:31 |
| kfox1111_ | heh. people give up too quickly. :/ | 21:32 |
| inc0 | Newton will be EOL soon | 21:32 |
| kfox1111_ | we just gotta put a tarball in a place. why is that hard? | 21:32 |
| inc0 | let me put it this way, I'm not super happy with having most of our testing done on EOLed repo | 21:33 |
| inc0 | version* | 21:33 |
| kfox1111_ | we already have a gate job taring up/posting all the containers we have. | 21:33 |
| kfox1111_ | that I agree with. | 21:33 |
| kfox1111_ | someone's just gotta put in some work. | 21:33 |
| inc0 | if we cut that out, that leaves us with ocata and pike | 21:33 |
| inc0 | both available in tarballs | 21:34 |
| kfox1111_ | you keep saying that is a lot of work to carry. | 21:34 |
| kfox1111_ | I still don't agree. | 21:34 |
| inc0 | well, we're multiplying things we carry | 21:34 |
| inc0 | now we have tarballs and no tarballs | 21:34 |
| kfox1111_ | agreed. but a forloop is a forloop either way. | 21:34 |
| inc0 | also gates for ansible-orch and no orch | 21:34 |
| kfox1111_ | if it has 2 or 5 things in it. | 21:34 |
| kfox1111_ | orch is something seperate. | 21:35 |
| kfox1111_ | test what you want there. | 21:35 |
| kfox1111_ | go ahead and only support pike. should be fine. | 21:35 |
| inc0 | well, I know | 21:35 |
| inc0 | I don't want to break our env tho | 21:35 |
| inc0 | also I'm not opposed to have big coverage | 21:35 |
| kfox1111_ | we agreed only to add for orch. it wont break anything there. | 21:35 |
| inc0 | but complexity is growing:/ | 21:36 |
| kfox1111_ | not sure it really is. | 21:36 |
| inc0 | ok, reason I'm complaining is now I'm trying to put configmaps generation in container | 21:36 |
| kfox1111_ | keep orch only to one version. I'm cool with that. | 21:36 |
| kfox1111_ | configmap generation works on all versions currently. | 21:36 |
| kfox1111_ | not sure what the issue there is. | 21:37 |
| inc0 | it won't work if we'd deprecate kollakube | 21:37 |
| kfox1111_ | why not? | 21:37 |
| kfox1111_ | I dont see anything in it that doesn't work accross all versions? | 21:37 |
| inc0 | it's not code I'm talking about | 21:37 |
| inc0 | it's sphagetthi in our gate shell scritps | 21:38 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/488008 | 21:38 |
| kfox1111_ | theres a little bit of mess there. a lot more of the mess simply is not sharing code we probably could share. :/ | 21:38 |
| inc0 | right | 21:38 |
| kfox1111_ | ceph vs iscsi. | 21:38 |
| inc0 | right | 21:38 |
| kfox1111_ | not 2 vs 4. | 21:38 |
| inc0 | well 2 vs 4 would is slightly different image distribution | 21:39 |
| kfox1111_ | how so? | 21:39 |
| inc0 | and don't forget how many times you guys have hit issue that sth wasn't available in images | 21:39 |
| inc0 | pushed | 21:39 |
| inc0 | even tho it was merged | 21:39 |
| kfox1111_ | ight now we have two different distrib versions. trunk and everything else. | 21:39 |
| kfox1111_ | gates are working now though. so the images are good. not sure the issue. if things broke, then we shoudl evaluate if we carry forward or abandon. | 21:40 |
| inc0 | 4 can be moved to tarballs too | 21:40 |
| kfox1111_ | so can 2 and 3 with the same effort as 4. | 21:41 |
| kfox1111_ | then we're down to 1 distrib mechansim. | 21:41 |
| kfox1111_ | but not throwing out konwn working stuff. | 21:41 |
| inc0 | no, I'm not suggesting that | 21:42 |
| kfox1111_ | still not sure why dropping 2 and 3 reduces our workload a significant amount. | 21:42 |
| inc0 | well for now it works so it's fine | 21:43 |
| inc0 | but orch will have to be 4+ or better yet just single version | 21:43 |
| kfox1111_ | yeah. I'm ok reevaluating it once it breaks, or we have to do major surgery to support the next version. | 21:44 |
| kfox1111_ | works for me. | 21:44 |
| kfox1111_ | since p is iminant, I dont really think only supoprting P is too bad. | 21:44 |
| kfox1111_ | by the time orch is really done, so will pike. | 21:44 |
| inc0 | right | 21:44 |
| inc0 | ok, that works | 21:44 |
| kfox1111_ | I wana keep the low level building bloks as widely usable as possible. | 21:45 |
| kfox1111_ | the orchestration stuff can be more opinionated. | 21:45 |
| kfox1111_ | it really has to be. | 21:45 |
| inc0 | what I'm saying is "shell is kinda orchestration too" | 21:46 |
| kfox1111_ | there is no such thing an unopinionated orchestration tool Ithink. | 21:46 |
| inc0 | and it grew to sphagetthi;) | 21:46 |
| kfox1111_ | it kind of is, but its just intended for gate testing. | 21:46 |
| kfox1111_ | whole different level of requirements. | 21:47 |
| kfox1111_ | I don't disagree on the sphagetthiish. | 21:47 |
| kfox1111_ | version one of gate tests for anything complicated rarely is pretty. :) | 21:47 |
| *** david-lyle has quit IRC | 21:48 | |
| *** renmak_ has joined #openstack-kolla | 21:48 | |
| inc0 | yeah, I don't think it's bad thing, I'm just thinking how to solve this mess without destabilizing our dev | 21:48 |
| kfox1111_ | I think that task is much lower priority for kolla as a whole then getting orchestration going. | 21:49 |
| kfox1111_ | its like kolla-containers vs kolla-ansible. | 21:50 |
| kfox1111_ | the kolla-containers tests are a little less critical then getting kolla-ansible working with them. | 21:50 |
| kfox1111_ | anyway... | 21:51 |
| inc0 | but deploy gates in kolla-containers aren't based of shell;) anyway, we agree that we need to clean the spaghetti | 21:51 |
| kfox1111_ | I think between your ps, and the microchart ps, your close. | 21:51 |
| inc0 | right, they're meant to be complementary | 21:52 |
| kfox1111_ | inc0: whatever. we coudl go around and around on gates. ;) like, having multinode, etc. :) | 21:52 |
| kfox1111_ | I think once those to ps's stablalize/merge, then we can start to talk about getting rid of the older gates. | 21:53 |
| kfox1111_ | until then, what we got works and is the best we got. | 21:53 |
| kfox1111_ | the other thing I think the shell buys us is a good seperation from bugs in the orchestration system. | 21:54 |
| kfox1111_ | having the shell gate will help let you know that the helm charts are good and stable, and any bugs found are likly in the new orchestration code. | 21:54 |
| inc0 | yeah no argument there | 21:54 |
| kfox1111_ | and visa versa. | 21:55 |
| kfox1111_ | should help bootstrap the orchestration code much quicker. | 21:55 |
| inc0 | but let's agree that if there is significant backwards incompatibility, we're not adding too many more ifs please.. | 21:55 |
| kfox1111_ | +1 | 21:55 |
| kfox1111_ | I'm also thinking right after 1.0, whenever that is, | 21:55 |
| kfox1111_ | we drop all support but for current release'd version. | 21:56 |
| kfox1111_ | whichever comes first. | 21:56 |
| kfox1111_ | either major incompatability or 1.0. | 21:56 |
| *** eaguilar|afk has quit IRC | 21:56 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fix rendering problem when yaml value gets overridden via cli https://review.openstack.org/488513 | 22:00 |
| *** jamesbenson has quit IRC | 22:01 | |
| *** itlinux has quit IRC | 22:05 | |
| *** mewald has quit IRC | 22:06 | |
| *** renmak_ has quit IRC | 22:12 | |
| *** ducttap__ has joined #openstack-kolla | 22:30 | |
| *** ducttape_ has quit IRC | 22:33 | |
| *** vhosakot has quit IRC | 22:35 | |
| *** ducttap__ has quit IRC | 22:42 | |
| *** ducttape_ has joined #openstack-kolla | 22:45 | |
| *** pbourke has quit IRC | 22:46 | |
| *** pbourke has joined #openstack-kolla | 22:48 | |
| *** goldyfruit has quit IRC | 22:49 | |
| *** ducttape_ has quit IRC | 22:50 | |
| jascott1 | sbezverk is this still a version dependent problem? https://github.com/kubernetes/helm/issues/2739 | 22:51 |
| *** ducttape_ has joined #openstack-kolla | 22:52 | |
| *** ducttape_ has quit IRC | 22:56 | |
| *** jamesbenson has joined #openstack-kolla | 22:57 | |
| *** inc0 has quit IRC | 23:00 | |
| *** inc0 has joined #openstack-kolla | 23:00 | |
| *** jamesbenson has quit IRC | 23:02 | |
| *** tonanhngo has quit IRC | 23:22 | |
| *** tonanhngo has joined #openstack-kolla | 23:28 | |
| *** hongbin has quit IRC | 23:30 | |
| *** tonanhngo has quit IRC | 23:33 | |
| *** ddyer has quit IRC | 23:35 | |
| *** tonanhngo has joined #openstack-kolla | 23:36 | |
| *** tonanhngo has quit IRC | 23:41 | |
| *** renmak_ has joined #openstack-kolla | 23:48 | |
| sbezverk | jascott1 for extra \n what you suggested fixed the issue, but version related issue is still there | 23:52 |
| sbezverk | jascott1 thanks a lot | 23:52 |
| *** renmak_ has quit IRC | 23:58 | |
| *** ducttape_ has joined #openstack-kolla | 23:59 | |
| jascott1 | ah interesting. I think new `set` went in between 2.4.2 and 2.5 | 23:59 |
| *** bmace has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
