| *** eaguilar has joined #openstack-kolla | 00:04 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 00:05 |
|---|---|---|
| *** eaguilar_ has quit IRC | 00:06 | |
| *** yingjun has joined #openstack-kolla | 00:11 | |
| *** eaguilar has quit IRC | 00:13 | |
| *** sbezverk has quit IRC | 00:18 | |
| *** sbezverk has joined #openstack-kolla | 00:20 | |
| *** sbezverk has quit IRC | 00:24 | |
| *** Jeffrey4l has quit IRC | 00:25 | |
| *** gema has quit IRC | 00:27 | |
| *** gema has joined #openstack-kolla | 00:27 | |
| *** jrobinson has quit IRC | 00:27 | |
| *** jaesang has joined #openstack-kolla | 00:39 | |
| *** duonghq has joined #openstack-kolla | 00:47 | |
| duonghq | morning guys | 00:47 |
| *** deadnull has quit IRC | 00:48 | |
| *** sbezverk has joined #openstack-kolla | 00:53 | |
| kfox1111 | duonghq: hi. :) | 00:53 |
| duonghq | evening kfox1111 | 00:54 |
| *** cuongnv has joined #openstack-kolla | 01:00 | |
| *** MasterOfBugs has quit IRC | 01:01 | |
| hawi | hi. i do have question. ubuntu 16.04.2, Images that failed to build list is really long. what's wrong? | 01:01 |
| *** jrobinson has joined #openstack-kolla | 01:03 | |
| *** tonanhngo has quit IRC | 01:04 | |
| *** sayantani01 has quit IRC | 01:07 | |
| *** daidv has joined #openstack-kolla | 01:09 | |
| duonghq | hawi, can you post the log? | 01:11 |
| duonghq | normally, one of base image failed | 01:11 |
| duonghq | lead to all other fail | 01:11 |
| *** sbezverk has quit IRC | 01:12 | |
| *** sbezverk has joined #openstack-kolla | 01:12 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 01:13 |
| *** tonanhngo has joined #openstack-kolla | 01:15 | |
| hawi | duonghq: http://pastebin.com/eSkxAsSx | 01:16 |
| duonghq | hawi, your registry is not proper setup | 01:19 |
| *** unicell has quit IRC | 01:19 | |
| *** tonanhngo has quit IRC | 01:19 | |
| hawi | duonghq: thanks, will try to find out what's wrong with it. | 01:21 |
| duonghq | hawi, how do you setup this? | 01:21 |
| *** unicell has joined #openstack-kolla | 01:21 | |
| *** schwicht has joined #openstack-kolla | 01:21 | |
| hawi | duonghq: docker run -d -p 4000:5000 --restart=always --name registry registry:2 and i modified /etc/default/docker as well | 01:25 |
| hawi | duonghq: DOCKER_OPTS="--insecure-registry http://10.79.3.254:4000" | 01:26 |
| duonghq | how about kolla-build.conf? | 01:27 |
| *** vhosakot has joined #openstack-kolla | 01:28 | |
| hawi | duonghq: it seems that i dont have it where i expected it to be | 01:28 |
| *** zhubingbing has joined #openstack-kolla | 01:29 | |
| duonghq | hawi, sorry, do not understand | 01:29 |
| duonghq | I see that your settings is applied | 01:29 |
| duonghq | but it cannot connect to the registry | 01:29 |
| duonghq | you can try push some images to the registry manually | 01:30 |
| vhosakot | sbezverk: thanks for your reply, checking IRC logs now. so, did you run the steps in these scripts manually to install kolla-k8s? https://github.com/openstack/kolla-kubernetes/blob/master/tools/build_dev_image.sh and https://github.com/openstack/kolla-kubernetes/blob/master/tools/run_dev_image.sh ? | 01:31 |
| *** unicell has quit IRC | 01:31 | |
| hawi | duonghq: thanks, will try to sort it out | 01:33 |
| vhosakot | sbezverk: again, I'll read your reply later in the IRC logs if I get disconnected when I'm eating dinner now :) | 01:38 |
| *** schwicht has quit IRC | 01:44 | |
| *** schwicht_at_work has joined #openstack-kolla | 01:48 | |
| sbezverk | vhosakot: yes but I modified them to match my environment | 01:48 |
| vhosakot | sbezverk: cool, thanks for the info, will run them now. | 01:49 |
| *** Jeffrey4l has joined #openstack-kolla | 01:53 | |
| *** zioproto has quit IRC | 02:01 | |
| *** sayantani01 has joined #openstack-kolla | 02:06 | |
| *** schwicht_at_work has quit IRC | 02:12 | |
| *** Jeffrey4l has quit IRC | 02:14 | |
| *** Jeffrey4l has joined #openstack-kolla | 02:14 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 02:15 |
| sbezverk | kfox1111: another pain :( nova scheduler failing to contact placement.. but curling placement service returns correct response | 02:19 |
| *** masber has joined #openstack-kolla | 02:26 | |
| *** cuongnv has quit IRC | 02:29 | |
| *** cuongnv has joined #openstack-kolla | 02:32 | |
| *** vhosakot has quit IRC | 02:32 | |
| *** vhosakot has joined #openstack-kolla | 02:32 | |
| sdake | hey peeps | 02:34 |
| duonghq | hi sdake | 02:35 |
| sdake | sup peeps | 02:38 |
| sdake | supdu | 02:38 |
| sdake | sup duonghq | 02:39 |
| *** Jeffrey4l has quit IRC | 02:41 | |
| *** Jeffrey4l has joined #openstack-kolla | 02:41 | |
| *** yuanying has quit IRC | 03:03 | |
| *** jrobinson has quit IRC | 03:08 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible master: Fix can't find /usr/lib/libCryptoki2_64.so in barbican https://review.openstack.org/444561 | 03:14 |
| zhubingbing | https://review.openstack.org/#/c/444561/ | 03:14 |
| zhubingbing | who please help me review it | 03:14 |
| zhubingbing | ;) | 03:14 |
| zhubingbing | duonghq | 03:14 |
| vhosakot | hi sdake.. since I ran into many vagrant issues, I reverse-engineered the Vagrfile's contents, setup the networking myself, booted four bare-metal VMs, and deployed a four-node (1 master, 3 workers) k8s cluster manually using kubeadm using https://kubernetes.io/docs/getting-started-guides/kubeadm/. Here is my cluster --> http://paste.openstack.org/show/603032/. | 03:16 |
| vhosakot | I'll now refer https://github.com/openstack/kolla-kubernetes/blob/master/tools/build_dev_image.sh and https://github.com/openstack/kolla-kubernetes/blob/master/tools/run_dev_image.sh, and deploy kolla-k8s on my k8s cluster. | 03:16 |
| vhosakot | then, I'll help sbezverk and kfox1111 improve the docs so docs have all the steps for new contributors. | 03:17 |
| *** rwsu has joined #openstack-kolla | 03:23 | |
| *** Margin has joined #openstack-kolla | 03:27 | |
| Margin | hi, I met an issue ,https://bugs.launchpad.net/kolla/+bug/1673657 | 03:27 |
| openstack | Launchpad bug 1673657 in kolla "port 6000 which is used by swift object-server isn't listening after deploy openstack with multi-node " [Undecided,New] | 03:27 |
| *** lazyPwr has quit IRC | 03:29 | |
| *** zhugaoxiao has quit IRC | 03:29 | |
| *** zhugaoxiao has joined #openstack-kolla | 03:30 | |
| *** lazyPower has joined #openstack-kolla | 03:30 | |
| duonghq | vhosakot, thank for you doc, does the ps is merged? | 03:37 |
| vhosakot | duonghq: no, I'll add the docs soon, I just got my cluster up :) | 03:38 |
| duonghq | vhosakot, roger | 03:38 |
| sdake | sbezverk i pushed 4.0.0 for centos source | 03:41 |
| sdake | vhosakot there is a doc available to shows how to deploy kolla-kubernetes using kubeadm | 03:44 |
| *** vhosakot_ has joined #openstack-kolla | 03:44 | |
| sdake | vhosakot vhosakot_ https://etherpad.openstack.org/p/kolla-kubernetes-deploy-guide-BP | 03:44 |
| *** vhosakot has quit IRC | 03:46 | |
| vhosakot_ | sdake: I started with that, a lot of the steps have strikethrough. hence, did not know the exact steps/flow. I was able to use the upstream docs for kubeadm and deploy 4-node cluster successfully - https://kubernetes.io/docs/getting-started-guides/kubeadm/. | 03:47 |
| sdake | vhosakot_ the non-strikethrough are the correct docs more or less | 03:47 |
| sdake | they have been tested by 5 people+ | 03:47 |
| vhosakot_ | sdake: can I refer them to deploy kolla-k8s on my empty cluster? | 03:48 |
| sdake | i'd recommend testing them | 03:49 |
| sdake | are you doing it today or tomorrow? | 03:49 |
| sdake | if today - i have some time to do with you | 03:49 |
| vhosakot_ | sdake: yes, I have a working empty cluster on which I'm deploying my test apps to learn k8s... I now need OpenStack (kolla-k8s) on my empty cluster. | 03:50 |
| vhosakot_ | sdake: here is my empty cluster on which sbezverk says I can now start installing OpenStack (kolla-k8s) http://paste.openstack.org/show/603032/ | 03:51 |
| sdake | vhosakot_ if your oding it now - i can help you via webex | 03:52 |
| vhosakot_ | sdake: how long does it take? does building take a lot of time.. I can spend few hours now, sure.. I wast kolla-k8s up to be able to contribute more to kolla-k8s project :) | 03:53 |
| vhosakot_ | wast ==want* | 03:53 |
| vhosakot_ | sdake: do you think my empty cluster at http://paste.openstack.org/show/603032/ is a good one to start with, or should I deploy k8s form scratch? | 03:54 |
| sdake | thats what i'd use | 03:54 |
| sdake | the one you ahve deployed | 03:54 |
| sdake | few hours max | 03:54 |
| sdake | http://cisco.webex.com/join/stdake | 03:55 |
| vhosakot_ | cool, let me jump in then :) 2 mins :) | 03:55 |
| vhosakot_ | I'm in :) can you hear me | 03:58 |
| duonghq | sdake, I'm not sure why we softlink kolla-ansible to kolla | 04:01 |
| duonghq | It's confuse from very beginning | 04:01 |
| sdake | duonghq we shouldn't | 04:04 |
| *** Margin has quit IRC | 04:04 | |
| sdake | its nanoying | 04:04 |
| sdake | vhosakot_ i can't hear you | 04:04 |
| sdake | vhosakot_ i think my network imploded - let me reocnnect | 04:04 |
| vhosakot_ | I'm back | 04:04 |
| *** goldyfruit has quit IRC | 04:08 | |
| *** skramaja has joined #openstack-kolla | 04:20 | |
| duonghq | sdake, I'm not sure why we clone into kolla-ansible, then cd kolla/ansible? | 04:33 |
| duonghq | there is not setup.py here | 04:33 |
| *** lamt has joined #openstack-kolla | 04:36 | |
| *** unicell has joined #openstack-kolla | 04:41 | |
| *** sayantani01 has quit IRC | 04:52 | |
| *** sayantani01 has joined #openstack-kolla | 04:54 | |
| *** yuanying has joined #openstack-kolla | 04:56 | |
| *** unicell1 has joined #openstack-kolla | 05:11 | |
| *** unicell has quit IRC | 05:11 | |
| *** caowei has joined #openstack-kolla | 05:11 | |
| *** Margin has joined #openstack-kolla | 05:29 | |
| *** erlon has quit IRC | 05:35 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 05:42 |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla master: Fix oslo_debug_helper not running https://review.openstack.org/444398 | 05:52 |
| masber | hi, I am testing kolla and i got this error --> http://paste.openstack.org/show/603070/ | 06:02 |
| masber | was wondering if someone could help me out finding out the issue | 06:02 |
| *** Jeffrey4l has quit IRC | 06:02 | |
| masber | this paste here has more information --> http://paste.openstack.org/show/603071/ | 06:04 |
| *** jascott1- is now known as jascott1 | 06:06 | |
| *** sayantan_ has joined #openstack-kolla | 06:11 | |
| *** sayantani01 has quit IRC | 06:14 | |
| Margin | masber, I met your issue | 06:15 |
| masber | how did you fixed it, I am checking on stackoverflow and looks like configuration parameters are missing? | 06:16 |
| Margin | https://bugs.launchpad.net/kolla/+bug/1626339 | 06:16 |
| openstack | Launchpad bug 1626339 in kolla-ansible "Kolla mongodb error: can't find self in the replset config" [High,In progress] - Assigned to Jeffrey Zhang (jeffrey4l) | 06:16 |
| Margin | please refer https://review.openstack.org/#/c/422377/ | 06:16 |
| masber | ok, will do | 06:16 |
| *** Jeffrey4l has joined #openstack-kolla | 06:17 | |
| *** zhurong has joined #openstack-kolla | 06:21 | |
| sbezverk | Jeffrey4l: ping | 06:21 |
| Jeffrey4l | sbezverk, pong | 06:21 |
| masber | this patch has been commited right? | 06:22 |
| sbezverk | Jeffrey4l: to be able to start an instance with 4.0.0 images is it mandatory to have cell confiugred? | 06:22 |
| Jeffrey4l | sbezverk, yep. | 06:22 |
| sbezverk | Jeffrey4l: :( | 06:22 |
| Jeffrey4l | otherwise, sync db will complain, iirc. | 06:22 |
| masber | sbezverk, ocata required cell configuration as far as I am aware of | 06:23 |
| sbezverk | Jeffrey4l: what do I need for it? | 06:23 |
| sbezverk | Jeffrey4l: is it just a sequence of commands? or it requires more? | 06:23 |
| Jeffrey4l | sbezverk, check https://docs.openstack.org/developer/nova/cells.html#fresh-install | 06:23 |
| sbezverk | Jeffrey4l: thank you.. | 06:24 |
| Jeffrey4l | sbezverk, simple_cell_setup is easier. | 06:24 |
| sbezverk | Jeffrey4l: man, it is like anohter bootstrap process :( | 06:25 |
| Jeffrey4l | kolla-ansible is using simple_cell_setup command. | 06:26 |
| Jeffrey4l | during deploy | 06:26 |
| Jeffrey4l | 1. create nova_cell0 database and proper credential. 2. sync db as normal 3. run nova-manage cell_v2 simple_cell_setup | 06:26 |
| sbezverk | Jeffrey4l: which image is used to run this command nova-api? | 06:26 |
| Jeffrey4l | yep. | 06:26 |
| Jeffrey4l | nova_api with nova.conf file. | 06:26 |
| masber | Margin, has this patch been applied yet? | 06:27 |
| *** mpansky has joined #openstack-kolla | 06:27 | |
| *** saneax-_-|AFK is now known as saneax | 06:27 | |
| sbezverk | Jeffrey4l: I use nova-api for bootstrapping nova-api DB | 06:27 |
| Margin | oh, I don't know | 06:28 |
| sbezverk | curious why it is not executeds as a part of normal bootstrap process? | 06:28 |
| Jeffrey4l | sbezverk, which? | 06:28 |
| sbezverk | Jeffrey4l: we run 2 bootstrap jobs, 1 for db create/manage of nova db adn one for create/manage nova-api db | 06:29 |
| Jeffrey4l | we need third database: nova_cell0 | 06:29 |
| sbezverk | Jeffrey4l: so extended start script does not automatically execute simple_cell_setup | 06:30 |
| sbezverk | Jeffrey4l: OMG another db :( | 06:30 |
| Jeffrey4l | simple cell setup should run after all nova compute is up. | 06:30 |
| *** satyar has joined #openstack-kolla | 06:31 | |
| sbezverk | Jeffrey4l: how do you tell nova-api image to run simple_cell_setup | 06:31 |
| sbezverk | if the same image is used for everything | 06:31 |
| sbezverk | ? | 06:31 |
| Jeffrey4l | just "docker exec nova_api nova-manage cell_v2 simple_cell_setup" | 06:31 |
| sbezverk | Jeffrey4l: ok got it, thank you, will build a chart for it tomorrow.. | 06:32 |
| Jeffrey4l | np. | 06:32 |
| *** mnasiadka has joined #openstack-kolla | 06:41 | |
| *** unicell1 has quit IRC | 07:04 | |
| *** unicell has joined #openstack-kolla | 07:04 | |
| *** sayantan_ has quit IRC | 07:15 | |
| *** mrhillsman has quit IRC | 07:20 | |
| *** caowei has quit IRC | 07:23 | |
| *** caowei has joined #openstack-kolla | 07:23 | |
| *** vhosakot_ has quit IRC | 07:25 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Update infra components list in README.rst https://review.openstack.org/446873 | 07:33 |
| *** mrhillsman has joined #openstack-kolla | 07:34 | |
| *** manheim has joined #openstack-kolla | 07:35 | |
| *** manheim has quit IRC | 07:37 | |
| *** manheim has joined #openstack-kolla | 07:38 | |
| *** manheim has quit IRC | 07:38 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible master: Fix can't find /usr/lib/libCryptoki2_64.so in barbican https://review.openstack.org/444561 | 07:38 |
| *** l4yerffeJ_ has joined #openstack-kolla | 07:38 | |
| *** manheim has joined #openstack-kolla | 07:38 | |
| *** pcaruana has joined #openstack-kolla | 07:43 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Install kolla-ansible in develop mode in gate https://review.openstack.org/446877 | 07:49 |
| *** matrohon has joined #openstack-kolla | 07:54 | |
| *** Jezekus has joined #openstack-kolla | 07:57 | |
| *** mpansky has quit IRC | 08:09 | |
| *** gfidente has joined #openstack-kolla | 08:11 | |
| *** zioproto has joined #openstack-kolla | 08:18 | |
| *** mgoddard has joined #openstack-kolla | 08:27 | |
| *** zhurong has quit IRC | 08:27 | |
| masber | do I need to run kolla-ansible deploy after kolla-ansible reconfigure? | 08:29 |
| mgoddard | thomas_oneill: I've got some patches for Bifrost that I've not yet pushed upstream including sorting out the logging. Let me know if you would like any. | 08:29 |
| openstackgerrit | Dou Rui Yuan proposed openstack/kolla-ansible master: update expressions for more clear and unified https://review.openstack.org/446882 | 08:31 |
| openstackgerrit | pangliye proposed openstack/kolla-ansible master: Fixes Bug1673691 There is no files in directory "/etc/kolla/telegraf/config/",so the operation of copy should be deleted Fixes Bug1673691 https://review.openstack.org/446888 | 08:36 |
| *** duke__ has joined #openstack-kolla | 08:36 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Register Cinder v3 endpoint in keystone https://review.openstack.org/446889 | 08:37 |
| *** skramaja has quit IRC | 08:37 | |
| *** mgoddard has quit IRC | 08:43 | |
| Jeffrey4l | pbourke, could u review https://review.openstack.org/446877 which fix the gate ( pbr version issue ) | 08:48 |
| *** l4yerffeJ_ has quit IRC | 08:51 | |
| *** mpansky has joined #openstack-kolla | 08:53 | |
| *** shardy has joined #openstack-kolla | 09:01 | |
| *** Serlex has joined #openstack-kolla | 09:06 | |
| *** blallau has joined #openstack-kolla | 09:12 | |
| *** skramaja has joined #openstack-kolla | 09:14 | |
| *** david-lyle_ has joined #openstack-kolla | 09:18 | |
| *** david-lyle has quit IRC | 09:18 | |
| *** cmceniry has quit IRC | 09:20 | |
| *** mgoddard has joined #openstack-kolla | 09:20 | |
| *** cmceniry has joined #openstack-kolla | 09:20 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/kolla master: RHEL: Add mod_ssl for services running over httpd https://review.openstack.org/446911 | 09:27 |
| thomas_oneill | mgoddard: morning!! yes please!! :) | 09:27 |
| thomas_oneill | I was really zonked at the end of yesterday so I could hardly think, but once I'd gotten bifrost actually deploying in a container and the ironic service (at least the API) running, it seemed that the ironic conductor was not running and no drivers were being listed as available | 09:28 |
| thomas_oneill | need to take another look again now, just going to run through it manually and try to see what might be going wrong | 09:28 |
| thomas_oneill | but yes will be useful to see at any patches that might make things easier :D | 09:29 |
| thomas_oneill | mgoddard: could I also ask your opinion on the python mysqldb dep screw-up issue? There are two sets of bugfixes for this problem, sayantani posted this at the end of january https://bugs.launchpad.net/kolla-ansible/+bug/1658840 https://review.openstack.org/#/c/424391/ and then I posted this just yesterday I think https://bugs.launchpad.net/kolla/+bug/1673420 and | 09:32 |
| openstack | Launchpad bug 1658840 in kolla-ansible "Bootstrapping bifrost services in Bifrost container fails with error in creating MySQL DB" [Undecided,In progress] - Assigned to sayantani (sayantani.goswami) | 09:32 |
| thomas_oneill | https://review.openstack.org/#/c/446609/ | 09:32 |
| openstack | Launchpad bug 1673420 in kolla "Bifrost image's mysqldb module is broken" [Undecided,Confirmed] | 09:32 |
| *** qiliang27 has quit IRC | 09:33 | |
| *** athomas has joined #openstack-kolla | 09:38 | |
| *** zhubingbing has quit IRC | 09:43 | |
| mgoddard | thomas_oneill: I'm running centos so haven't hit the mysql issue | 09:46 |
| mgoddard | thomas_oneill: here's bifrost: https://github.com/stackhpc/bifrost/commits/stackhpc-3.0.0 | 09:49 |
| thomas_oneill | cheers :) | 09:50 |
| mgoddard | thomas_oneill: this one pins bifrost dependencies in the kolla image: https://github.com/stackhpc/kolla/commit/4bf7d181f60a2195a7de14060dea4f6d2215fe72 | 09:50 |
| mgoddard | thomas_oneill: and some for kolla-ansible: https://github.com/stackhpc/kolla-ansible/commits/stackhpc-4.0.0.0rc1 | 09:51 |
| mgoddard | I think they're all rebased onto a recent stable/ocata, so stop looking when the commits are no longer by me or 'stack' | 09:52 |
| *** egonzalez has joined #openstack-kolla | 09:52 | |
| thomas_oneill | damn there's quite a lot here eh! are you going to submit some PS's soon? :p | 09:53 |
| mgoddard | thomas_oneill: some are already submitted, the rest I will submit soon | 09:54 |
| mgoddard | thomas_oneill: tried to prioritise stuff that was actually broken over improvements | 09:55 |
| openstackgerrit | caoyuan proposed openstack/kolla master: Update the Fluent link format in README.rst https://review.openstack.org/446928 | 09:56 |
| *** pbourke has quit IRC | 10:05 | |
| *** pbourke has joined #openstack-kolla | 10:05 | |
| *** Margin has quit IRC | 10:06 | |
| *** Jezekus has quit IRC | 10:06 | |
| *** jaesang has quit IRC | 10:08 | |
| *** daidv has quit IRC | 10:08 | |
| *** mgoddard has quit IRC | 10:09 | |
| *** cuongnv has quit IRC | 10:15 | |
| *** yingjun has quit IRC | 10:15 | |
| *** satyar has quit IRC | 10:20 | |
| openstackgerrit | Merged openstack/kolla master: Update the Fluent link format in README.rst https://review.openstack.org/446928 | 10:25 |
| *** duonghq has quit IRC | 10:26 | |
| *** Jezekus has joined #openstack-kolla | 10:28 | |
| *** deadnull has joined #openstack-kolla | 10:30 | |
| *** deadnull has quit IRC | 10:30 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Register Cinder v3 endpoint in keystone https://review.openstack.org/446889 | 10:40 |
| *** egonzalez has quit IRC | 10:43 | |
| *** yuanying has quit IRC | 10:45 | |
| thomas_oneill | https://review.openstack.org/#/c/443747/ I think this needs more reviews for workflow? Or does a core have to tick the workflow separately? | 10:46 |
| *** athomas has quit IRC | 10:54 | |
| *** jaosorior has joined #openstack-kolla | 10:55 | |
| *** jaosorior has joined #openstack-kolla | 10:56 | |
| *** mgoddard has joined #openstack-kolla | 10:57 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Update infra components list in README.rst https://review.openstack.org/446873 | 10:59 |
| *** caowei has quit IRC | 11:00 | |
| *** rwallner has joined #openstack-kolla | 11:04 | |
| *** zhurong has joined #openstack-kolla | 11:14 | |
| *** athomas has joined #openstack-kolla | 11:17 | |
| *** papacz has joined #openstack-kolla | 11:21 | |
| jaosorior | hey folks, any help debugging this failure? https://review.openstack.org/#/c/446911/ not sure if it's related to the patch. | 11:24 |
| thomas_oneill | jaosorior: one or more of the images failed to build :p | 11:31 |
| thomas_oneill | ubuntu source | 11:31 |
| thomas_oneill | bifrost-base failed, dunno if that's excluded | 11:31 |
| thomas_oneill | http://logs.openstack.org/11/446911/1/check/gate-kolla-dsvm-build-ubuntu-source-ubuntu-xenial/294a687/console.html find which image(s) failed and aren't excluded, try to build manually locally to verify that they can be built or not, I guess? | 11:32 |
| jaosorior | thomas_oneill: I only changed stuff for RHEL O_o | 11:32 |
| thomas_oneill | I don't know I've not debugged failed gates before | 11:32 |
| thomas_oneill | oh | 11:32 |
| *** duke__ has quit IRC | 11:32 | |
| *** athomas has quit IRC | 11:36 | |
| thomas_oneill | it tests against the stable/ocata releases of the openstack projects codes right? | 11:40 |
| *** yuanying has joined #openstack-kolla | 11:40 | |
| *** mpansky has quit IRC | 11:40 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/446771 | 11:41 |
| mgoddard | thomas_oneill, jaosorior: bifrost is an odd one though - kolla installs ocata release of bifrost, but bifrost by default pulls down latest versions of its dependencies | 11:42 |
| mgoddard | hence https://github.com/stackhpc/kolla/commit/4bf7d181f60a2195a7de14060dea4f6d2215fe72 | 11:43 |
| jaosorior | mgoddard: so... what do I do there? do I just recheck the gate? | 11:44 |
| mgoddard | jaosorior: let me check the logs :) | 11:45 |
| mgoddard | jaosorior: hmm. looks like bifrost is failing to install shade | 11:47 |
| jaosorior | mgoddard: seems quite unrelated to my change | 11:47 |
| mgoddard | jaosorior: error message is TypeError: __init__() got an unexpected keyword argument 'encoding' | 11:48 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Remove global privilege escalation for deploy_aio.sh https://review.openstack.org/446645 | 11:49 |
| thomas_oneill | mgoddard: oh yeah that's why it always ends up being 3.0.1 which is newton right? | 11:49 |
| jaosorior | mgoddard: so, all I did was add mod_ssl to RHEL based systems https://review.openstack.org/#/c/446911/ | 11:49 |
| mgoddard | jaosorior: yeh I don't think it's caused by your change | 11:50 |
| mgoddard | jaosorior: try a recheck, just to see, but I doubt it will work :) | 11:53 |
| hawi | hi. i have question regarding prechecks. When checking for docker-py, it gives me failed, invalid syntax. so, my question is - what version of python and python-py it expects actually? on ubuntu 16.04.2 i do have only python3 right now | 12:00 |
| mnaser | morning everyone | 12:01 |
| *** jaosorior is now known as jaosorior_brb | 12:01 | |
| mnaser | jaosorior_brb im checking your fails | 12:02 |
| *** mpansky has joined #openstack-kolla | 12:02 | |
| mnaser | jaosorior_brb found your problem | 12:03 |
| *** rhallisey has joined #openstack-kolla | 12:03 | |
| mnaser | (98)Address already in use: AH00072: make_sock: could not bind to address 0.0.0.0:443 | 12:03 |
| mnaser | there is another service taking 443 | 12:03 |
| mnaser | most likely its haproxy | 12:03 |
| *** rhallisey has quit IRC | 12:04 | |
| jaosorior_brb | mnaser: interesting, it shouldn't even be trying to bind to 443; it should just add the mod_ssl package :/ | 12:04 |
| mnaser | jaosorior_brb im pretty sure installing mod_ssl drops a file in conf.d with a Listen 443 in rhel-based systems | 12:04 |
| *** shardy is now known as shardy_lunch | 12:04 | |
| mnaser | is there a reason why you dont want haproxy to do ssl instead? | 12:05 |
| mnaser | (just out of curiosity) | 12:05 |
| jaosorior_brb | mnaser: cause we're trying to do SSL everywhere | 12:06 |
| mnaser | ok so even ssl between haproxy and horizon | 12:06 |
| jaosorior_brb | mnaser: so haproxy is doing TLS termination, and then it will go to the services, also encrypted. | 12:06 |
| jaosorior_brb | mnaser: that's correct. | 12:06 |
| *** satyar has joined #openstack-kolla | 12:07 | |
| mnaser | i see | 12:07 |
| mnaser | btw, i got this setup which can maybe help us gauge our overall stability - http://grafana.openstack.org/dashboard/db/kolla-failure-rate | 12:08 |
| *** caowei has joined #openstack-kolla | 12:09 | |
| mnaser | looks like ubuntu xenial source builds are a bit of am ess | 12:09 |
| *** zhugaoxiao has quit IRC | 12:13 | |
| *** zhugaoxiao has joined #openstack-kolla | 12:14 | |
| *** caowei has quit IRC | 12:16 | |
| *** Serlex has quit IRC | 12:17 | |
| satyar | Hi everyone | 12:18 |
| satyar | one question is freezer implemented fully with kolla? | 12:18 |
| *** athomas has joined #openstack-kolla | 12:23 | |
| *** rwellum has joined #openstack-kolla | 12:27 | |
| hawi | installed python-minimal (2.7.11-1) and python-docker (1.8.0-0ubuntu1), now i can move forward with installation, but with python3 docker-py fails on ubuntu 16.04.2, when doing prechecks | 12:27 |
| *** saneax is now known as saneax-_-|AFK | 12:27 | |
| hawi | and got another weird thing from, for some hosts precheck rabbitmq says that Hostname has to resolve to IP address of api_interface | 12:28 |
| *** goldyfruit has joined #openstack-kolla | 12:29 | |
| *** lamt has quit IRC | 12:30 | |
| dims | SamYaple : could use your input https://etherpad.openstack.org/p/single-approach-to-containers-for-projects | 12:31 |
| dims | inc0 : anyone else interested ^ | 12:31 |
| *** erlon has joined #openstack-kolla | 12:33 | |
| hawi | rabbitmq precheck fail "Hostname has to resolve to IP address of api_interface" doesnt appear for other machines, only for three machines (all controllers) | 12:44 |
| hawi | i see | 12:44 |
| mnaser | can every node resolve the hostname of other controllers hawi ? | 12:46 |
| hawi | yes. i think this precheck (havent looked inside yet) assumes that interfaces from API network for these controller hosts resolv to hostname | 12:47 |
| mnaser | it has to hawi | 12:48 |
| mnaser | or the rabbitmq cluster wont form quorum because it uses dns resolution | 12:48 |
| hawi | mnaser: http://pastebin.com/vNWbKjGE | 12:50 |
| hawi | something like this | 12:50 |
| mnaser | do they resolve? | 12:51 |
| hawi | yes, they do | 12:53 |
| hawi | from each node | 12:53 |
| hawi | but yes, API ip is different, hostname is related to PXE network IP, what i have | 12:53 |
| hawi | API interface ip | 12:53 |
| mnaser | the hostname has to resolve to the api interface ip | 12:53 |
| mnaser | it doesnt sound like it is | 12:54 |
| *** schwicht has joined #openstack-kolla | 12:54 | |
| hawi | i used MAAS for deploying machines, so PXE network has hostname. aghhh. anyway, i can fix it | 12:55 |
| *** manheim has quit IRC | 12:56 | |
| *** manheim has joined #openstack-kolla | 12:56 | |
| *** mpansky has quit IRC | 12:58 | |
| *** mpansky has joined #openstack-kolla | 12:58 | |
| *** rhallisey has joined #openstack-kolla | 13:01 | |
| *** schwicht has quit IRC | 13:03 | |
| nea1 | how long will it take until ocata is officially released? | 13:03 |
| *** schwicht has joined #openstack-kolla | 13:05 | |
| mnaser | nea1 ocata is out i believe | 13:05 |
| mnaser | stable/ocata is tagged already | 13:05 |
| *** Jezekus has quit IRC | 13:06 | |
| mnaser | stuff tagged 4.0.0 i think is stable/ocata | 13:06 |
| nea1 | mnaser: ok then other question when is it uploaded to https://hub.docker.com/r/kolla/centos-binary-kolla-toolbox/tags/ or a other docker_registry? | 13:18 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 13:19 |
| *** blallau_ has joined #openstack-kolla | 13:21 | |
| *** schwicht has quit IRC | 13:21 | |
| *** Jezekus has joined #openstack-kolla | 13:22 | |
| *** duonghq has joined #openstack-kolla | 13:23 | |
| duonghq | evening guys | 13:23 |
| nea1 | TASK [haproxy : Waiting for virtual IP to appear] ****************************** | 13:24 |
| nea1 | fatal: [localhost]: FAILED! => {"changed": false, "elapsed": 301, "failed": true, "msg": "Timeout when waiting for vip.kolla.[...]:3306"} | 13:24 |
| nea1 | seams like keepalived is failing to add the IP to the interface | 13:25 |
| sdake | morning peeps | 13:27 |
| sdake | you off duonghq ? | 13:27 |
| sdake | duonghq you had asked a quesiton earlier before i PTFO'ed for the night | 13:28 |
| sdake | 21:33:03] <duonghq>sdake, I'm not sure why we clone into kolla-ansible, then cd kolla/ansible? | 13:28 |
| sdake | [21:33:13] <duonghq>there is not setup.py here | 13:28 |
| sdake | duonghq would you mind expanding on the question - I'm not sure what your referencing there | 13:28 |
| duonghq | sdake, minutes, I need finding the link to kolla-k8s deploy bp etherpad | 13:29 |
| duonghq | bookmark at work and home are not synced | 13:29 |
| sdake | moment i've got it | 13:29 |
| *** dave-mccowan has joined #openstack-kolla | 13:30 | |
| sdake | https://etherpad.openstack.org/p/kolla-kubernetes-deploy-guide-BP | 13:31 |
| sdake | sup dave-mccowan | 13:31 |
| sdake | duonghq ^^ | 13:31 |
| duonghq | thanks | 13:31 |
| *** mpansky has quit IRC | 13:31 | |
| dave-mccowan | hi sdake | 13:31 |
| *** srwilkers has joined #openstack-kolla | 13:31 | |
| sdake | sup srwilkers | 13:31 |
| duonghq | in line 78, we clone kolla-ansible repository, but in like 80, we cd kolla/ansible | 13:32 |
| duonghq | hi srwilkers | 13:32 |
| *** jtriley has joined #openstack-kolla | 13:32 | |
| Jeffrey4l | srwilkers, my irc client said: " srwilkers aka. srwilkers_away aka. srwilkers_aaway aka. srwilkers_coffee aka. srwilkers_brb aka. srwilkers_lunch " | 13:33 |
| Jeffrey4l | hi sdake | 13:34 |
| sdake | sup Jeffrey4l | 13:34 |
| Jeffrey4l | could u review https://review.openstack.org/446877 , which related to the pbr issue. | 13:34 |
| Jeffrey4l | need install kolla-ansible in env to make pbr work as expected. | 13:34 |
| sdake | why -e? | 13:35 |
| sdake | and not just pip install kolla-ansible? | 13:35 |
| Jeffrey4l | just like "usedevelop=True" in tox.ini | 13:35 |
| Jeffrey4l | i can not explain why. | 13:36 |
| *** zhugaoxiao has quit IRC | 13:36 | |
| Jeffrey4l | but by using "pip install kolla-ansible" it return "4.0.0" on ocata branch | 13:36 |
| Jeffrey4l | when using "pip install -e kolla-ansible" it return "4.0.1" on ocata branch. | 13:36 |
| Jeffrey4l | the latter one is expected. | 13:36 |
| Jeffrey4l | and tox -e works like this. | 13:36 |
| *** zhugaoxiao has joined #openstack-kolla | 13:37 | |
| * sdake groans | 13:37 | |
| Jeffrey4l | so currently, kolla stable/ocata generation 4.0.1 image, but kolla-ansible try to use 4.0.0 image. | 13:37 |
| sdake | ok - are we expecting operators to pip install -e then? | 13:37 |
| Jeffrey4l | operator should use "pip install kolla-ansible" ( from pypi ) | 13:37 |
| sdake | cool | 13:37 |
| Jeffrey4l | we are using " pip install -e kolla-ansible ( from git repo ) " | 13:38 |
| Jeffrey4l | in gate. | 13:38 |
| sdake | https://etherpad.openstack.org/p/single-approach-to-containers-for-projects | 13:43 |
| *** schwicht has joined #openstack-kolla | 13:45 | |
| duonghq | sdake, can you take a look at my question above?, | 13:47 |
| duonghq | in kolla-kubernetes deploy guide in line 78, we clone kolla-ansible repository, but in like 80, we cd kolla/ansible | 13:47 |
| sdake | duonghq yes - I fixed it in the etherpad | 13:47 |
| sdake | can you reload? | 13:47 |
| duonghq | thanks, | 13:48 |
| sdake | I answered that question on line 87 | 13:48 |
| duonghq | network laggy, I expected it update in realtime | 13:48 |
| sdake | ya etherpad a little flakey sometimes | 13:48 |
| duonghq | sorry | 13:48 |
| sdake | no reason to apologize | 13:48 |
| sdake | its a toosl problem not a person problem :) | 13:48 |
| duonghq | :) | 13:48 |
| sdake | no reason to apologize for tools problems | 13:49 |
| duonghq | roger | 13:49 |
| sdake | apologize when you actually hurt someone - or alternatively ask for forgiveness ;) | 13:49 |
| *** mpansky has joined #openstack-kolla | 13:50 | |
| *** jaosorior_brb is now known as jaosorior | 13:50 | |
| *** blallau_ has quit IRC | 13:51 | |
| *** Jezekus has quit IRC | 13:52 | |
| *** yingjun has joined #openstack-kolla | 13:52 | |
| *** shardy_lunch is now known as shardy | 13:56 | |
| *** zhurong has quit IRC | 13:56 | |
| *** mpansky has quit IRC | 14:02 | |
| *** mpansky has joined #openstack-kolla | 14:03 | |
| *** lamt has joined #openstack-kolla | 14:06 | |
| sbezverk | Jeffrey4l: tried to create cell, it is failing.. but it does not produce much debugging info. I was wondering if you might have log generated by simple_cell create command so I could use it as a baseline | 14:06 |
| Jeffrey4l | sbezverk, when running simple_cell_setup, the nova-compute must up and shown in 'nova service-list'. did you di this? | 14:07 |
| sdake | sbezverk morning fine sir | 14:08 |
| Jeffrey4l | simple_cell_setup generate few lines of logs. | 14:08 |
| sbezverk | sdake: good morning | 14:08 |
| sdake | sbezverk quick q, are you back in the office? | 14:08 |
| Jeffrey4l | i do not think it is useful. | 14:08 |
| sbezverk | sdake: nope still in SJ | 14:08 |
| sdake | sbezverk cool - you are back monday? | 14:08 |
| sbezverk | Jeffrey4l: :( I will make sure all services are up before running it.. | 14:09 |
| sdake | sbezverk reason I ask is - I want to start cracking on external config maps and ejecting the kolla-ansible dependency | 14:09 |
| sbezverk | sdake: yeah should be back on monday | 14:09 |
| sdake | I think this along with docs are our number one and one priorities | 14:09 |
| Jeffrey4l | not up. need keep "nova-compute" is shown in "nova service-list" | 14:09 |
| sdake | I am not quite sure where to start | 14:09 |
| sbezverk | sdake: makes two of us ;) | 14:09 |
| sdake | ok - well the way to answer that question is to get started :) | 14:10 |
| sdake | let me look at what hcoas my calendar entails | 14:10 |
| sbezverk | sdake: I need to get vm running with ocata images and it is not yet happening, so it is a priority for me.. I will join you after | 14:11 |
| sdake | sbezverk perhaps we can have a livve brainstorming session for those folks interested in the community | 14:11 |
| sdake | sbezverk is ocata in the gating atm? | 14:11 |
| sdake | sbezverk (not sure if it was added yet) | 14:11 |
| sdake | (is that what your working on?) | 14:11 |
| sbezverk | sdake: my ironic jobs are using 4.0.0 images | 14:11 |
| sdake | nice | 14:11 |
| sdake | i pushed centos src yesterday | 14:11 |
| sbezverk | sdake: but it is not wokring | 14:11 |
| sdake | use_neutron = true | 14:12 |
| sdake | could be one thing | 14:12 |
| sdake | ok - so your working on gating 4.0.0 images then? | 14:12 |
| sbezverk | sdake: no, it is cell | 14:12 |
| sbezverk | sdake: no cell, no vm :) | 14:12 |
| sdake | right | 14:13 |
| * sdake groans | 14:13 | |
| sdake | well gating 4.0.0 images is probably t he top of the stack of things we need to do :) | 14:13 |
| sdake | next on the stack is docs + external config maps | 14:13 |
| sbezverk | sdake: that is what I think too | 14:13 |
| sbezverk | Jeffrey4l: could you take a look I got from a container creating cell to see if it makes sense to you, since I have never seen any working scenario, it is hard for m to tell if it is ok or not, I do not see any obvious errors | 14:16 |
| Jeffrey4l | ok. logs or? | 14:16 |
| sbezverk | Jeffrey4l: nvm log size is 0 bytes :( | 14:18 |
| Jeffrey4l | check the return code. | 14:18 |
| sbezverk | Jeffrey4l: cool thanks for the idea | 14:19 |
| Jeffrey4l | sbezverk, check the nova code here https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L1083 | 14:19 |
| Jeffrey4l | i guess you may exit in here https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L1114 | 14:19 |
| Jeffrey4l | nova-compute isn't up and shown in "nova service-list" | 14:19 |
| sbezverk | Jeffrey4l: I have a function in the gate code that can wait for nova to be fully up before creating cell.. will try it and also get ret code.. | 14:22 |
| Jeffrey4l | got. | 14:22 |
| *** Serlex has joined #openstack-kolla | 14:25 | |
| *** mnasiadka has quit IRC | 14:27 | |
| *** sayantani01 has joined #openstack-kolla | 14:31 | |
| thomas_oneill | any of you cores fancy final review/workflow for this? https://review.openstack.org/#/c/443747/ | 14:33 |
| *** lamt has quit IRC | 14:34 | |
| matrohon | hi kolla experts | 14:39 |
| matrohon | I just registered a new BP in kolla | 14:39 |
| matrohon | do I need the corresponding one in kolla-ansible? | 14:39 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 14:40 |
| *** rmart04 has joined #openstack-kolla | 14:42 | |
| *** zhubingbing_ has joined #openstack-kolla | 14:44 | |
| *** manheim has quit IRC | 14:57 | |
| mnaser | urgh | 14:59 |
| mnaser | more gate breakge | 14:59 |
| mnaser | looks like master source seems broken? | 15:00 |
| mnaser | btw i pasted this earlier but i figured ill repost, this will hopefully help us track down the state of things and see where we can improve things - http://grafana.openstack.org/dashboard/db/kolla-failure-rate | 15:00 |
| *** mpansky has quit IRC | 15:01 | |
| *** sbezverk has quit IRC | 15:03 | |
| thomas_oneill | sayantani01: did you see updates on my bug/fix that's the same as yours but also completely different? :P | 15:04 |
| satyar | Hi all | 15:04 |
| satyar | any one tried freezer? | 15:04 |
| thomas_oneill | I'm pretty new to contributing so my ability is limited, but a decision needs to be made on how best to resolve the issue, I don't think either of our PS's are great, both pretty hackish/workaroundish | 15:04 |
| sayantani01 | thomas_oneill I didn't get a chance to take a look yesterday but I will definitely get on it today | 15:05 |
| thomas_oneill | cool | 15:05 |
| *** mpansky has joined #openstack-kolla | 15:05 | |
| *** Serlex has quit IRC | 15:06 | |
| sayantani01 | Can you also look at the comments on my patch thomas_oneill ? | 15:06 |
| thomas_oneill | yeah, sec | 15:07 |
| dmsimard | mnaser: I wonder if you can drill down those graphs further, (with distros) | 15:09 |
| dmsimard | mnaser: oh wait they're already there | 15:09 |
| mnaser | :D | 15:09 |
| mnaser | ubuntu source seems to have been pooping out so im working on that | 15:10 |
| mnaser | im trying to replicate right now | 15:10 |
| *** mpansky has quit IRC | 15:10 | |
| mnaser | unfortunately source build failures => things cant go through :( | 15:10 |
| dmsimard | mnaser: if kolla would run tempest, you could also get data from openstack-health | 15:12 |
| mnaser | dmsimard thats the goal eventually | 15:13 |
| mnaser | im trying to figure out these ubuntu failures.. random pip install problem about a library | 15:13 |
| thomas_oneill | how do I actually make a comment on a PS? I've added a comment and it's saying it's a draft | 15:14 |
| dmsimard | thomas_oneill: gotta click "post" after clicking on the "reply" button at the top | 15:14 |
| *** manheim has joined #openstack-kolla | 15:15 | |
| *** dave-mccowan has quit IRC | 15:15 | |
| thomas_oneill | dmsimard: I don't see this, it's not a reply it's an original comment. I've done replies before with no problem. | 15:16 |
| mnaser | thomas_oneill once you have drafts in, going back to the main gerrit review page and click on "reply" | 15:16 |
| mnaser | it'll show all the comments and click post | 15:16 |
| thomas_oneill | oh I see | 15:16 |
| goldyfruit | Do you guys had this issue with Keystone: ContextualVersionConflict: (pbr 2.0.0 (/var/lib/kolla/venv/lib/python2.7/site-packages), Requirement.parse('pbr<2.0,>=1.3'), set(['sqlalchemy-migrate'])) | 15:21 |
| goldyfruit | I'm trying to redeploy Mitaka but Keystone complains every time that I try to reach his endpoint | 15:22 |
| *** manheim has quit IRC | 15:22 | |
| sdake | portdirect pingola | 15:23 |
| goldyfruit | If I downgrade pbr to 1.10 I have complains about Olso: ContextualVersionConflict: (pbr 1.10.0 (/var/lib/kolla/venv/lib/python2.7/site-packages), Requirement.parse('pbr>=2.0.0'), set(['oslo.i18n', 'oslo.utils', 'oslo.middleware', 'oslo.context', 'oslo.log', 'stevedore', 'oslo.messaging', 'debtcollector', 'oslo.concurrency'])) | 15:26 |
| *** ipsecguy_ is now known as ipsecguy | 15:27 | |
| *** dave-mccowan has joined #openstack-kolla | 15:29 | |
| *** dciabrin is now known as dciabrin|away | 15:29 | |
| portdirect | hey sdake o/ | 15:30 |
| portdirect | just heading into a meeting but will be back in 1:30 (I hope :/ ) | 15:30 |
| sdake | portdirect enjoy | 15:31 |
| *** yingjun has quit IRC | 15:31 | |
| sdake | portdirect ping me when you return | 15:31 |
| *** hrw has quit IRC | 15:35 | |
| *** hrw has joined #openstack-kolla | 15:36 | |
| *** Serlex has joined #openstack-kolla | 15:40 | |
| *** Jezekus has joined #openstack-kolla | 15:50 | |
| *** rmart04 has quit IRC | 15:52 | |
| *** sbezverk has joined #openstack-kolla | 15:52 | |
| *** rmart04 has joined #openstack-kolla | 15:54 | |
| *** dciabrin|away is now known as dciabrin | 15:57 | |
| *** kbaegis has joined #openstack-kolla | 15:58 | |
| *** rmart04 has quit IRC | 16:00 | |
| kfox1111 | morning. | 16:03 |
| goldyfruit | Morning | 16:03 |
| mnaser | o/ | 16:03 |
| *** duonghq has quit IRC | 16:05 | |
| kfox1111 | heh; https://etherpad.openstack.org/p/single-approach-to-containers-for-projects | 16:12 |
| kfox1111 | getting called out for multiple docker container implementations. :/ | 16:12 |
| kfox1111 | ah. I see others have already found it. :) | 16:13 |
| *** kbaegis has quit IRC | 16:13 | |
| *** MasterOfBugs has joined #openstack-kolla | 16:17 | |
| *** kbaegis has joined #openstack-kolla | 16:18 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 16:19 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Switch to Kolla virtual environment for Bifrost https://review.openstack.org/447053 | 16:20 |
| mnaser | another gate fix ^ | 16:20 |
| mnaser | this should clear out ubuntu-source build fails | 16:20 |
| mnaser | which has been at around 50% | 16:21 |
| mnaser | at at the moment 35% failure rate compared to centos/oracle's 6% | 16:21 |
| *** shananigans has quit IRC | 16:23 | |
| mnaser | mgoddard thanks for the catch! | 16:38 |
| mgoddard | mnaster: no problem. odd how often these 'ships in the night' moments happen :) | 16:39 |
| mnaser | glad we caught the other one, better off to just keep working with our venv rather and avoid getting bit by weird system python issues | 16:40 |
| *** goldyfruit has quit IRC | 16:42 | |
| *** kbaegis has quit IRC | 16:44 | |
| mgoddard | mnaster: agreed | 16:46 |
| *** bmace has quit IRC | 16:48 | |
| *** bmace has joined #openstack-kolla | 16:48 | |
| *** eaguilar has joined #openstack-kolla | 16:49 | |
| *** zhubingbing_ has quit IRC | 16:50 | |
| *** kbaegis has joined #openstack-kolla | 16:52 | |
| *** rabel has joined #openstack-kolla | 16:55 | |
| *** rabel has left #openstack-kolla | 16:55 | |
| *** jaosorior has quit IRC | 16:55 | |
| *** skramaja has quit IRC | 16:57 | |
| *** goldyfruit has joined #openstack-kolla | 16:58 | |
| sdake | kfox1111 any progress on external config maps? | 17:04 |
| *** athomas has quit IRC | 17:07 | |
| sdake | jascott1 you alive | 17:09 |
| jascott1 | idk maybe | 17:10 |
| jascott1 | feeling like crap today | 17:10 |
| jascott1 | hows things sdake? | 17:10 |
| sdake | jascott1 i have really tried to take your advice and not drop knowledge or wisdom | 17:10 |
| sdake | jascott1 yet everyone wants it :) | 17:11 |
| jascott1 | LOL | 17:11 |
| sdake | jascott1 you got a plan B:) | 17:11 |
| sdake | jascott1 here was my plan A: https://etherpad.openstack.org/p/single-approach-to-containers-for-projects | 17:12 |
| *** matrohon has quit IRC | 17:13 | |
| *** haplo37 has quit IRC | 17:15 | |
| *** kbaegis has quit IRC | 17:16 | |
| jascott1 | gee | 17:18 |
| *** pcaruana has quit IRC | 17:18 | |
| jascott1 | lotsa stuff in there oh ya | 17:18 |
| goldyfruit | About my previous issue: (11:21:22 AM) goldyfruit: Do you guys had this issue with Keystone: ContextualVersionConflict: (pbr 2.0.0 (/var/lib/kolla/venv/lib/python2.7/site-packages), Requirement.parse('pbr<2.0,>=1.3'), set(['sqlalchemy-migrate'])) | 17:18 |
| goldyfruit | Solution: In keystone container: sed -i 's/Requires-Dist: pbr (>=1.3,<2.0)/Requires-Dist: pbr (>=1.3,<=2.0)/g' /var/lib/kolla/venv/lib/python2.7/site-packages/sqlalchemy_migrate-0.10.0.dist-info/METADATA | 17:18 |
| *** Jezekus has quit IRC | 17:24 | |
| *** haplo37 has joined #openstack-kolla | 17:24 | |
| jascott1 | sdake are you going for blueprint outta that doc? | 17:28 |
| sdake | jascott1 nope | 17:30 |
| sdake | jascott1 this is dims rodeo :) | 17:30 |
| sdake | jascott1 dims is interested in seeking a collaboration so OpenStack doesn't end up with two dueling container implementations when one would suffice | 17:30 |
| sdake | atleast that is my speculation, I actually dont know what he is after ;-) | 17:31 |
| jascott1 | i see | 17:31 |
| dims | "We have openstack-deb* openstack-ansible* openstack-charm* openstack-puppet* in our OpenStack git repo, can we have openstack-docker-* for folks to collaborate together?" | 17:31 |
| dims | right on top of the etherpad | 17:31 |
| dims | jascott1 : ^ | 17:35 |
| mnaser | isnt kolla what openstack-docker-* is? | 17:38 |
| jascott1 | yeah but community differs on opinions about what should be in images | 17:39 |
| sdake | mnaser kolla pioneered containers - in openstack we don't stop people from competing on the periphery (such as deployment tools) | 17:40 |
| mnaser | sdake correct, but kolla right now is a project with dockerfiles only | 17:40 |
| mnaser | kolla-ansible is the deployment tool that uses kolla container | 17:40 |
| sdake | mnaser as a result, there are a couple of different teams that have tried different approaches | 17:40 |
| mnaser | i mean in all terms, we can rename kolla to openstack-docker and keep kolla-ansible to keep using it | 17:41 |
| SamYaple | mnaser: kolla is a project with a bunch of custom tooling and is very non-standard docker project | 17:41 |
| SamYaple | mnaser: so youd get alot of pushback there | 17:41 |
| SamYaple | rightfully so | 17:41 |
| sdake | mnaser kolla could do that - and take that namespace easily | 17:41 |
| mnaser | i think the problem is there is so many ways of deploying with docker | 17:41 |
| sdake | mnaser we can rename our repos however we like | 17:42 |
| mnaser | and that's a very hard thing to solve | 17:42 |
| mnaser | ex: how do you configure services in containers, kolla took the volume approach | 17:42 |
| mnaser | *maybe* some people might want to go and run it through something like etcd for example | 17:42 |
| sdake | mnaser however it is confusing to have one global namespace such as openstack-docker or openstack-helm for a project that should be neutral to all parties | 17:42 |
| SamYaple | mnaser: no thats deploy tooling problems | 17:42 |
| mnaser | you're right, at the end of the day $service expects a config in /etc/$service/$service.conf | 17:43 |
| SamYaple | mnaser: if you look at the official docker projects youll find the kolla build process is the issue here. a process that i at one point pushed hard (stuff like adding openstack-base) | 17:43 |
| SamYaple | but docker as a whole has gone another direction | 17:43 |
| mnaser | what do you mean by the build process being the issue | 17:43 |
| SamYaple | we produce shared common layers in kolla.... that have no other purpose | 17:43 |
| SamYaple | no one does that anywhere else | 17:44 |
| sdake | SamYaple docker has gone the same direction it was headed when the kolla images were created | 17:44 |
| SamYaple | and it kills CI/CD | 17:44 |
| SamYaple | sdake: this is incorrect info | 17:44 |
| *** blallau has quit IRC | 17:44 | |
| mnaser | i think having a Dockerfile that sits inside say, openstack/nova would be very nice | 17:44 |
| sdake | common shared layers *were* best practice according to you | 17:44 |
| sdake | however they are *NOT* best practices today | 17:44 |
| mnaser | but there would be a lot of duplicating layers | 17:44 |
| SamYaple | sdake: this is correct | 17:45 |
| sdake | mnaser yup indeed there is duplication | 17:45 |
| lazyPower | duplicated layers only matter if you *really* care about image size being transferrred. | 17:45 |
| mnaser | but i think the duplication is very minimal and a common base shouldnt be enforced | 17:45 |
| SamYaple | lazyPower: actually funny you say that, images can be prouduced a third of the size of kolla.... and not share layers | 17:45 |
| mnaser | honestly the kolla images are so layer heavy right now | 17:45 |
| lazyPower | SamYaple: sure, bu ti said transferred, meaning fetching that shared common base once and not again for each subsidiary image | 17:46 |
| lazyPower | SamYaple: the savings are marginal at best imho. | 17:46 |
| SamYaple | lazyPower: no im saying how ever you want to slice it, total size transfered and stored _can_ be less than kolla right now | 17:46 |
| sdake | SamYaple most people don't really care about image size | 17:46 |
| mnaser | also what sdake said. i cant imagine it being that much heavier than it is now | 17:46 |
| sdake | SamYaple most peple care about two thigns -> CURL all over the internet | 17:47 |
| SamYaple | mnaser: right. and thats what is being discussed. Kolla being layer heavy. what that adds and what it prevents (like good CICD) | 17:47 |
| sdake | and CI/CD | 17:47 |
| lazyPower | ^ | 17:47 |
| SamYaple | sdake: kolla sucks at cicd and we all know it | 17:47 |
| lazyPower | that | 17:47 |
| mnaser | ive been doing an effort to fix the ci/cd stuff :( | 17:47 |
| lazyPower | curl + ci process. | 17:47 |
| lazyPower | oh but i wasn't trying to jump on a bash wagon though | 17:47 |
| sdake | SamYaple i agree curl and ci/cd are a problem | 17:47 |
| mnaser | there is just so many things going everywhere and the fact that its not voting seems to have kinda let it stale a bit | 17:47 |
| sdake | i never said they weren't | 17:47 |
| sdake | in fact i wrote down in the etherpad curl shouldn't be used in the image at all | 17:48 |
| SamYaple | sdake: but you are fighting for a structure that inhibits CI/CD | 17:48 |
| mnaser | sdake but then you have another interesting issue | 17:48 |
| mnaser | where do you get deps from | 17:48 |
| lazyPower | so, dumb question | 17:48 |
| sdake | SamYaple i have not fought for any structure that inhibits ci/cd | 17:48 |
| lazyPower | why aren't we running a mirror or a minio bucket that warehouses all the deps | 17:48 |
| lazyPower | or a swift bucket because #openstack | 17:48 |
| sdake | SamYaple i haev fought for a structure that preserves the API | 17:48 |
| sdake | SamYaple i dont care if the images are one or many | 17:48 |
| sdake | rather have one layer or many | 17:49 |
| mnaser | how would you handle source/binary types in the Dockerfile | 17:49 |
| lazyPower | thats typically how this is resolved in enterprise scenarios, they setup a mirror and the mirror becomes teh canonicalized source to prevent random curls. | 17:49 |
| sdake | lazyPower good question - we can't get it done | 17:49 |
| SamYaple | sdake: which is great. preserve that api. you should. but thats not what me (or others) are looking for these days | 17:49 |
| sdake | SamYaple not sure what your argument has to do with layering at all? | 17:50 |
| lazyPower | oh ok. | 17:50 |
| mnaser | on another point, while i love the enthusiasm behind reducing layers, is that really the reason behind the CI/CD problems? :x | 17:50 |
| sdake | mnaser the answer is yes | 17:50 |
| mnaser | how so? i would feel things would become so much scarier if things were all over the place (in-repo Dockerfile's for example) | 17:50 |
| sdake | mnaser here is how today's CI/CD systems work | 17:51 |
| sdake | take ZUUl as an example | 17:51 |
| sdake | ZUUL checks out a clean fresh repo on a clean fresh OS | 17:51 |
| sdake | it builds a nova-novncproxy | 17:51 |
| sdake | its done | 17:51 |
| jascott1 | does new squash make the layering conversation moot as user can just run that at end of their process? | 17:51 |
| SamYaple | sdake: the sharedlayering is whats preventing ci/cd | 17:51 |
| sdake | to do that, it has to build atleast 3 layers | 17:51 |
| sdake | SamYaple yes i agreed above - "The answer is yes" | 17:52 |
| mnaser | openstack-base, nova-base and nova-novncproxy -- so what's the problem with it having to build all 3 layers? | 17:52 |
| sdake | the 3 layers have to be rebuilt every time each container is built | 17:52 |
| sdake | this takes 4-5 hours in a real datacenter | 17:52 |
| lazyPower | i think its just the level of complexity being introduced to have those 3 common layers. | 17:52 |
| sdake | with a fresh checkout of each container | 17:52 |
| SamYaple | mnaser: with cicd, if those layers are shared, they have to build all the layers at the same time | 17:52 |
| lazyPower | or am i off base here? | 17:52 |
| sdake | lazyPower no the layers reduce complexity | 17:53 |
| SamYaple | lazyPower: youre right, but there is more to the picuter as well | 17:53 |
| sdake | lazyPower compressing the layers into one dockerfile increeases complexity | 17:53 |
| mnaser | but at the end of the day, the series of commands which are done to build nova-novncproxy will still be the same in a single Dockerfile | 17:53 |
| *** m1dev has joined #openstack-kolla | 17:53 | |
| SamYaple | lazyPower: they completely increase complexity, despite what others say. | 17:53 |
| mnaser | you'll still have to add the repos, you'll still have to install base packages, you'll still have to install nova and then install nova-novncproxy | 17:53 |
| lazyPower | well this is fun, 2 differing opinions being voiced quite loudly. | 17:53 |
| SamYaple | lazyPower: asking anyone who works with docker and big projects like this, and they will agree | 17:53 |
| mnaser | now if i have to make a change to the base, 7 dockerfiles have to be edited | 17:54 |
| lazyPower | i work with docker every day, and i avoid shared layers like the plague if i can. I only use the base layer as my shared layer, and from there each component gets its own dockerfile and goes. | 17:54 |
| sdake | lazyPower if you want to really break out the salt and butter check out https://etherpad.openstack.org/p/single-approach-to-containers-for-projects | 17:54 |
| SamYaple | lazyPower: to be fair, most people who agree with me have left this project. im just here to try to sway the project for the future into a better path | 17:54 |
| lazyPower | and when i say base layer, i mean FROM: ubuntu or FROM: alpine | 17:55 |
| sdake | SamYaple i disagree with your reasoning on why people have left the project - i think you are making a blanket assertion without stating facts | 17:55 |
| sdake | SamYaple and poisining relationships in the process | 17:55 |
| sdake | enjoy :) | 17:55 |
| lazyPower | but at the end of the day, this is the complexity problem with the golden image paradigm yes? | 17:55 |
| lazyPower | you make a change in a base layer and it domino's all teh way out to each edge component | 17:55 |
| SamYaple | lazyPower: yea. no image can meet 100% | 17:55 |
| SamYaple | thats just how it works | 17:55 |
| mnaser | lazyPower while id love to have a very small thin layer, but editing something in nova base could result in 7 changes | 17:56 |
| mnaser | instead of one change in a base layer | 17:56 |
| SamYaple | so if you notice some patchsets in the kolla queue, youll see where we have discussed making these images flatter | 17:56 |
| SamYaple | the issue comes down to sort of a lock-in affect | 17:56 |
| *** mgoddard has quit IRC | 17:56 | |
| mnaser | my question goes back to whats the difference between 3 images to get a final one | 17:56 |
| SamYaple | we have to maintain compatibility with the apis that have been put im place. that unfortunately means arbitrary jinja2 injects like, everywhere | 17:56 |
| mnaser | or 1 single thick one | 17:57 |
| SamYaple | mnaser: so the issue comes down to building | 17:57 |
| mnaser | i mean if you want to template the dockerfiles to include a base | 17:57 |
| SamYaple | mnaser: if you want keystone you build base->openstack-base->keystone-* | 17:57 |
| SamYaple | mnaser: if you want nova you build base->openstack-base->nova-* | 17:57 |
| SamYaple | youve build base and openstack-base twice | 17:57 |
| SamYaple | and they are bit-for-bit different images | 17:57 |
| SamYaple | they do not share layers | 17:57 |
| mnaser | i thought kolla builds in dependency order? | 17:58 |
| SamYaple | right, but we are talking cicd | 17:58 |
| sdake | mnaser indeed images are built in order | 17:58 |
| SamYaple | now do w emake every commit build _all_ kolla images? | 17:58 |
| mnaser | why not | 17:58 |
| SamYaple | thats the only way to do it in the current system | 17:58 |
| sdake | mnaser what SamYaple is speaking of is that docker's CI/Cd system doesn't provide ordering | 17:58 |
| SamYaple | sdake: dont speak for me. thats not what im saying at all | 17:58 |
| openstackgerrit | Alicja Kwasniewska proposed openstack/kolla master: Added snap-telemetry container https://review.openstack.org/446164 | 17:58 |
| mnaser | the same reason we run devstack/tempest for a nova change | 17:58 |
| sdake | SamYaple you speak for me why shouldn't i speak for you? | 17:59 |
| mnaser | the same reason tripleo deploys an entire cloud | 17:59 |
| lazyPower | but you're running this through jenkins right? CloudBee's has released an addon to do this for you, where you mark jobs as dependent, and it will notify downstream jobs to rebuild when a base-layer image change is merged.... | 17:59 |
| mnaser | as part of their CI | 17:59 |
| SamYaple | sdake: you are an openstack board member. stop please. | 17:59 |
| sdake | SamYaple i always represent kolla in this channel | 17:59 |
| lazyPower | not that i'm necessarily + or - on this convo, i just know that this conversation is getting quite heated and it doesn't seem like all the tech is being represented. | 17:59 |
| sdake | and nobody else | 17:59 |
| mnaser | lazyPower i agree :x | 17:59 |
| sdake | SamYaple DONT FORGET IT | 17:59 |
| lazyPower | ad-hominim? | 18:00 |
| mnaser | lazyPower imho, i think building $world is the way to continue going | 18:00 |
| mnaser | to say that we should test a single component alone if that single component changes is risky | 18:00 |
| lazyPower | have you guys looked at how kubernetes does its CI pipeline? | 18:00 |
| lazyPower | honest question ^ | 18:00 |
| sdake | lazyPower indeed kuberntes uses jenkins at present | 18:01 |
| sdake | and in the future they plan to use zuul theoretically | 18:01 |
| mnaser | i havent, any interesting docs lazyPower ? | 18:01 |
| lazyPower | mnaser: so this is to build off the "single component change single component check" statement | 18:01 |
| *** zioproto has quit IRC | 18:01 | |
| mnaser | but then how do you make sure that these components didnt break something else | 18:01 |
| lazyPower | google has like proably 18 different jobs, just realted to how it verifies a k8s change. And its layered. and it always ENDS with an e2e test with that changeset | 18:01 |
| mnaser | i dont know if we can do depenencies in zuul? | 18:02 |
| mnaser | ex: how i imagine ci ideally: build all images, deploy cloud with kolla, run tempest against it | 18:02 |
| lazyPower | meaning it can stgart by testing 1 component in a quick test, but before its fully accepted, its making it into an e2e test suite, that flexes the full stack of components (images w/ components in this case) in an integration test suite. | 18:02 |
| lazyPower | take a look at the gubernator dashboard. and for docs - http://github.com/kubernetes/test-infra/ | 18:02 |
| mnaser | right, i guess that helps in detecting early failures | 18:02 |
| lazyPower | thats the only way they can feesibly do it, and we joined up to contribue results daily from the CDK stack of things. So we can pull in the prior days commits and both a) determine if we are broken because of a change, b) indicate to the project they've broken the build in a configuration they didn't account for in their own e2e suite. | 18:03 |
| lazyPower | not that we break very often, but the 1.6 cycle has been in no-go state since freeze, which is a telling indicator of gaps in e2e. | 18:03 |
| lazyPower | but just because there are issues there, doesn't mean the approach is inferior. | 18:04 |
| mnaser | lazyPower i see what you're onto, but i think that's an optimization of the current state of CI | 18:04 |
| lazyPower | mnaser: maybe thats where we start instead of making changes to the baseline plubming | 18:04 |
| lazyPower | *plumbing | 18:04 |
| mnaser | and i agree with you lazyPower. that's why i've been pushing patches to fix CI failures | 18:05 |
| mnaser | i also added this http://grafana.openstack.org/dashboard/db/kolla-failure-rate | 18:05 |
| lazyPower | thats an excellent direction to be in, in my opinion | 18:05 |
| mnaser | so that we can drive that % down as much as possible | 18:05 |
| mnaser | (the ubuntu source failures i managed to find here - https://review.openstack.org/#/c/447053/) | 18:05 |
| lazyPower | inc0: i provided some modicum of value ;) ^ | 18:05 |
| lazyPower | mnaser: oh nice, even better if its fixing builds on ubuntu | 18:06 |
| lazyPower | i'm clearly biased here :O | 18:06 |
| mnaser | my ideal world: get the fail % as low as possible | 18:06 |
| mnaser | add tempest runs to the check/gate | 18:06 |
| lazyPower | right, once you have confidence in your built artifacts you can start mucking about in the plubming and making bigger changes | 18:06 |
| mnaser | and get the deploy jobs to become voting | 18:06 |
| mnaser | 100% agreed | 18:06 |
| sdake | mnaser kolla is almost there wrt voting gates | 18:07 |
| sdake | in fact the images vote | 18:07 |
| mnaser | yep ive noticed which is a great step | 18:07 |
| sdake | the deploy jobs depend on infra to become voting - and that is in progress | 18:07 |
| mnaser | if we cant deploy everything in the simplest and cleanest way ... then trying to fix the plumbing is looking the wrong place (imho) | 18:07 |
| sdake | the deploy jobs for kolla-kubernetes could be made voting today - I believe but can't confirm that is the case kfox1111 would have more detail | 18:07 |
| mnaser | that's great to hear so we can have both ansible and k8s deploy jobs | 18:08 |
| mnaser | anyways, i've also personally mentioned that if anyone finds themself "recheck"-ing too often, feel free to reach out and ill do my best to help troubleshoot and solve | 18:09 |
| * inc0 reading back | 18:12 | |
| inc0 | sorry I had cf of meetings this morning | 18:12 |
| *** eglute has joined #openstack-kolla | 18:12 | |
| inc0 | mnaser: sdake so ad voting gates, I plan to rework our gates when zuul v3 lands | 18:13 |
| mnaser | inc0 do you know when that is because it's been a while i think :p | 18:14 |
| inc0 | 1. multinode gates are going to be much easier and clean so multinode will be first class citizen | 18:14 |
| sdake | zuul v3 lands in june | 18:14 |
| mnaser | oh cool thanks sdake | 18:14 |
| sdake | not sure when inc0 will land his rework of the gates for zuulv3 however | 18:14 |
| mnaser | i was hoping to start doing some work to getting tempest to run now (that can hopefully be reused when zuulv3 comes) | 18:14 |
| mnaser | instead of the simple sanity check | 18:15 |
| sdake | mnaser i dont think the project wants to block on voting gates for zuulv3 to release | 18:15 |
| inc0 | sdake: it's going to be ansible based so I'm hopeful we can write cool ansible code to turon one ansible to another | 18:15 |
| sdake | mnaser Jeffrey4l is doing gating work today | 18:15 |
| sdake | mnaser along with the rest of our emea crew | 18:15 |
| inc0 | not only him | 18:15 |
| sdake | inc0 right | 18:15 |
| mnaser | okay cool, i can help with some of that progress (if need be) | 18:15 |
| inc0 | tempest voting is one thing we need, true | 18:15 |
| sdake | mnaser contributions much appreciated :) | 18:16 |
| inc0 | I also want full upgrade gates | 18:16 |
| inc0 | but that really matters when we can do multinode gating | 18:16 |
| inc0 | and that depends on zuulv3 | 18:16 |
| mnaser | in the meantime, clearing out all the small issues that keep coming up is what i've been trying to do | 18:16 |
| inc0 | sure | 18:17 |
| inc0 | with zuul v3 gates will be built with ansible rather than bash | 18:17 |
| *** satyar has quit IRC | 18:17 | |
| mnaser | hint hint https://review.openstack.org/#/c/447053/ fixes ubuntu source builds which are blocking gate | 18:17 |
| inc0 | so I'd be reluctant to spend too much time hacking bashscripts when we're only few weeks before we could do something much better | 18:17 |
| sdake | if by few weeks you mean 8-12 weeks :) | 18:18 |
| inc0 | sdake: when did you hear about june tho? | 18:18 |
| inc0 | last time I checked they didn't have final date yet | 18:18 |
| sdake | inc0 moment | 18:18 |
| sdake | inc0 can't find the reference - saw it on twitter iirc from monty - can't find it now - who knows why :) | 18:24 |
| mnaser | sdake i remember in the ML they mentioned that there is something using zuulv3 right now | 18:24 |
| sdake | mnaser possible - I am not on the zuulv3 team so I woulnd't know | 18:25 |
| mnaser | http://lists.openstack.org/pipermail/openstack-dev/2017-February/113065.html | 18:25 |
| mnaser | this | 18:25 |
| mnaser | buuut i dont think there's a date | 18:25 |
| sdake | mnaser right - perhaps i inferred one from this: | 18:26 |
| sdake | "There are far too many new and exciting things in Zuul v3 to cover in a | 18:26 |
| sdake | single post, and many of the topics (such as Ansible Jobs, or Job | 18:26 |
| sdake | Inheritance and Reuse) are deep topics we can dive in to over time. The | 18:26 |
| sdake | long and short of it is that Zuul v3 is coming soon to an OpenStack | 18:26 |
| sdake | Infra near you, so expect more and more communication about what that | 18:26 |
| sdake | means over the next few months." | 18:26 |
| sdake | two months + middle of march = middle of june | 18:26 |
| mnaser | i guess that's a fair way of putting it | 18:26 |
| sdake | hmm off by 1 error, may at the earliest :) | 18:26 |
| sdake | in any regard, few months of communication + some finalizatino time - inc0 still dont think we want to b lock | 18:28 |
| inc0 | depends on what you mean by block | 18:28 |
| inc0 | I'd block all major refactors on it, yes | 18:28 |
| inc0 | not fixes of what's broken | 18:28 |
| inc0 | also if someone actually wants to make big refactor before that - I'll be happy to review and help | 18:29 |
| inc0 | and won't -2 | 18:29 |
| inc0 | maybe I phrased it incorrectly - I'll, myself, will work on multinode gates, upgade gates and all-out gate refactoring | 18:30 |
| inc0 | but after zuulv3 | 18:30 |
| sdake | inc0 right sounds good | 18:30 |
| inc0 | any major gate changes disrupts development of other features | 18:31 |
| inc0 | especially if we're talking voting | 18:31 |
| inc0 | I want to limit these disruptions by ensuring that we'll work efficiently | 18:31 |
| inc0 | and zuul v3 will allow us to be much more efficient | 18:31 |
| portdirect | word up - back o/ | 18:33 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 18:35 |
| *** gfidente has quit IRC | 18:37 | |
| *** mgoddard has joined #openstack-kolla | 18:37 | |
| *** rhallisey has quit IRC | 18:37 | |
| *** rhallisey has joined #openstack-kolla | 18:38 | |
| *** rwellum has quit IRC | 18:40 | |
| v1k0d3n | hey all good afternoon | 18:43 |
| *** david-lyle_ is now known as david-lyle | 18:43 | |
| *** alop has joined #openstack-kolla | 18:48 | |
| alop | has anyone seen this error deploying compute-kit: | 18:48 |
| alop | l3-agent-daemonset.yaml" at <eq $restype "compute...>: error calling eq: invalid type for comparison | 18:49 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 18:51 |
| *** unicell has joined #openstack-kolla | 18:56 | |
| *** unicell has quit IRC | 18:57 | |
| v1k0d3n | sdake, i'm just reading back...and i really wish i didn't have to comment on this (trust me) | 18:59 |
| v1k0d3n | but size actually came up from several folks at the fishbowl in barcelona. inc0 was there. i know he know this to be true. | 18:59 |
| v1k0d3n | portdirect: was actually one of those as an individual contributor. | 19:00 |
| v1k0d3n | SAP pleaded for a roadmap to reduce size. | 19:00 |
| v1k0d3n | i know it has to be done the way that kolla is today, but to say that it's not a concern isn't the whole picture/story. | 19:00 |
| v1k0d3n | and that's my only .02 out of that whole coversation earlier with SamYaple | 19:01 |
| *** unicell has joined #openstack-kolla | 19:01 | |
| *** Jezekus has joined #openstack-kolla | 19:01 | |
| sdake | the most common arguments i have heard for size reduction are tied up in CI/CD, not the actual fact that the images take 30gig on disk | 19:02 |
| sdake | image size reduction is about increasing performance of CI/CD mostly - not deployment performance | 19:03 |
| sdake | solving CI/CD performance naturally alleviates people's size concerns | 19:04 |
| sdake | deployment performance of kolla is 20 minutes on 123 node cluster | 19:05 |
| v1k0d3n | but if there is a vulnerability common to base, that's not exactly ideal. | 19:05 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Add mechanism to store keys locally to avoid going out to network https://review.openstack.org/447101 | 19:05 |
| sdake | building 170 images individually in CI/CD takes hours/days | 19:05 |
| sdake | build wold works perfectly well and is very performant - many people want a faster CI/CD pipeline that doesn't "build world" but instead "builds indivdiually" | 19:06 |
| jascott1 | doesnt `squash` help with size? or at least when its not beta feature anymore | 19:06 |
| sdake | jascott1 it does help with size | 19:06 |
| sdake | jascott1 it doesn't help with ci/cd build performance | 19:06 |
| sdake | jascott1 as most ppls that are deploying kolla want to put a ci/cd pipeline in front of the docker image creation | 19:06 |
| sdake | and they do that by using jenkins or zuul to build each container indepndently | 19:07 |
| sdake | (or the DBS) | 19:07 |
| kfox1111 | hi | 19:07 |
| jascott1 | hey kfox1111 | 19:07 |
| sdake | hey kfox1111 | 19:07 |
| jascott1 | whats up on this fine Friday? | 19:07 |
| sdake | alop had a question above - which I am unable to answer | 19:07 |
| sdake | jascott1 ITS FRIDAY | 19:07 |
| jascott1 | \o/ | 19:07 |
| alop | sdake: I think something is wrong with my test setup | 19:08 |
| sdake | alop i see - helm version work for you? | 19:08 |
| jascott1 | kfox1111 was that issue with child->parent imports limitation I just posted a problem? | 19:09 |
| kfox1111 | jascott1: I responded just now. | 19:09 |
| kfox1111 | I don't think so? | 19:09 |
| jascott1 | oh ok great | 19:09 |
| kfox1111 | but maybe you have a use case in mind that I don't? | 19:09 |
| alop | yeah, I think my installer picked a much newer version than what is expected | 19:09 |
| jascott1 | im not on VPN atm so not getting the mails | 19:09 |
| sdake | alop what is the output of helm version? | 19:10 |
| alop | 2.2.2 | 19:10 |
| alop | I think that's my problem | 19:10 |
| sdake | let me run 2.2.2 throug hthe gate | 19:10 |
| alop | I notice that the Dockerfile in kolla-kubernetes specifies 2.1.3 | 19:10 |
| kfox1111 | should be 2.2.2 everywhere, unless we missed a place. | 19:11 |
| kfox1111 | we should be requiring 2.2.x now. | 19:11 |
| alop | I can update the dockerfile | 19:11 |
| *** rwallner has quit IRC | 19:12 | |
| kfox1111 | ah. yeah. its wrong in the docker file. :/ | 19:12 |
| kfox1111 | we really need to gate the dev docs. :/ | 19:12 |
| kfox1111 | so easy to break them. :/ | 19:12 |
| mnaser | sdake just so i can keep this on my mind and give it thought over the next few days. so the issue here is that a change in dockerfile for nova-novncproxy will result in $world being rebuilt, when in reality, we only need to rebuild nova-novncproxy because the rest did not change | 19:12 |
| sdake | mnaser if you run in jenkins, that is the case as jenkins doesn't keep the existing docker cache in place | 19:13 |
| sdake | mnaser you can build each image indiviudallly (with layers) | 19:13 |
| mnaser | so ideally CI would just rebuild nova-novncproxy, tack it into the rest of the prebuilt and run CI against it | 19:13 |
| sdake | mnaser and then each image takes less time to build (rather then buidl world) | 19:13 |
| sdake | mnaser yup | 19:14 |
| sdake | jenkins provides a fresh environment each time | 19:14 |
| sdake | docker cache is removed per ci/cd run | 19:14 |
| kfox1111 | yeah. with some kind of container fingerprinting, | 19:14 |
| kfox1111 | ci/cd can then deterimine which doesn't need rebuilding all the time. | 19:14 |
| sdake | kfox1111 ya like that docker hash that containers have ;) | 19:14 |
| kfox1111 | sdake: no, thats not sufficient. | 19:15 |
| sayantani01 | Can someone please point me to the document which talks about various interfaces? network interfaces, api interfaces etc? | 19:15 |
| sdake | kfox1111 i see | 19:15 |
| kfox1111 | sdake: cause the hash changes all the time for things that don't matter. | 19:15 |
| mnaser | and to build the hash, you'd have to build the images | 19:15 |
| mnaser | which brings you back to square one | 19:15 |
| mnaser | so how long can a docker image tag be | 19:15 |
| mnaser | aka what about tagging images in a common registry with service:<md5ofdockerfile> | 19:16 |
| kfox1111 | mnaser: true. I guess I'm thinking, you don't have to ci/cd the artifact if it hashes to an already tested/deployed thing. | 19:16 |
| kfox1111 | so no expensive deployment testing, or notifying admins they need to roll it into production, etc. | 19:16 |
| mnaser | so you host an internal registry where instead you tag things based on the md5 of the dockerfile | 19:16 |
| sdake | kfox1111 - quick question - https://github.com/openstack/kolla-kubernetes/blob/master/tools/test.sh#L10-L11 | 19:16 |
| sdake | i see 2.1.3 templates used | 19:17 |
| kfox1111 | mnaser: doesn't work either. repo deps change. | 19:17 |
| sdake | and 2.2.0 is deployed in gate | 19:17 |
| sdake | we ideally want to be using 2.2.2 | 19:17 |
| sdake | which templates should be used? | 19:17 |
| mnaser | this issue seems so straightforward yet so complicated, heh. | 19:17 |
| kfox1111 | sdake: yeah, probablhy should update that one too. I think only 2.2.0 was out at the time. | 19:17 |
| sdake | kfox1111 i can update 2.2.0 to 2.2.2 what should I update the templates to? | 19:18 |
| kfox1111 | sdake: we probably should try and figure out how to make ONE and only one variable that defines which version to use, to simplify tweaking it. | 19:18 |
| kfox1111 | sdake: not sure if there is a newer version of templates. | 19:18 |
| sdake | kfox1111 for the moment i only want to run a test, alop indicates he ran 2.2.2 and got a helm error | 19:18 |
| alop | FWIW - I'm not doing *exactly* the same thing | 19:18 |
| kfox1111 | mnaser: yeah. if it was easy, it would be fixed in the docker community a long time ago. :/ | 19:19 |
| kfox1111 | mnaser: I think its something like, (rpm -qa | sort; pip list ...| sort ) | md5sum | 19:19 |
| mnaser | i mean if you look at how devstack gate runs | 19:19 |
| mnaser | how would that differ between what kolla does | 19:20 |
| mnaser | build and install everything then run tests | 19:20 |
| mnaser | and that's a model that has worked for quite sometime | 19:20 |
| kfox1111 | devstack isn't for production though. | 19:20 |
| openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Update to Helm 2.2.2 https://review.openstack.org/447102 | 19:20 |
| mnaser | right | 19:20 |
| kfox1111 | its for building dev stuff. | 19:20 |
| mnaser | i may be biased because i'm thinking this in the scope of openstack ci only | 19:20 |
| kfox1111 | I want a much more clear pipeline for prod. | 19:20 |
| mnaser | and not some on site instance at $company that is doing CI before deploying | 19:20 |
| kfox1111 | 1. artefacts are produced. | 19:21 |
| kfox1111 | 2. artefacts are tested. | 19:21 |
| kfox1111 | 3. artefacts are used to stand up a production system. | 19:21 |
| mnaser | TIL you can spell it as artefact | 19:21 |
| kfox1111 | 4. step 1 is repeated and checked to see if actually different. | 19:21 |
| sdake | mnaser ;-) | 19:21 |
| kfox1111 | 5. step 4 things tested. | 19:22 |
| kfox1111 | 6. prodcution system updated using stuff from step 4 | 19:22 |
| sdake | alop right - are you installing helm templates per chance? | 19:22 |
| sdake | https://github.com/technosophos/helm-template/releases/download/2.2.2%2B1/helm-template-linux-2.2.2.1.tgz | 19:22 |
| sdake | alop if so, you may need this helm template to go along with helm 2.2.2 | 19:22 |
| sdake | alop i just submitted a review to test 2.2.2 against the gate | 19:22 |
| sdake | the gate was previously on 2.2.0 | 19:23 |
| alop | I'll try that out, thanks | 19:23 |
| sdake | alop rather currently on 2.2.0 of helm and 2.1.3 of the templtews | 19:23 |
| *** kbaegis has joined #openstack-kolla | 19:23 | |
| kfox1111 | I guess those two versions are compatible and only used for testing the template rendering, not talking with tiller, so never broke the gate. | 19:24 |
| sdake | kfox1111 the templates do what then? | 19:28 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Add mechanism to store keys locally to avoid going out to network https://review.openstack.org/447101 | 19:28 |
| sdake | alop rather - the gate is currently on 2.2.0 - not previously - if your using a recent git checkout - it will default to 2.2.0 | 19:29 |
| sdake | alop although if your not using the dev_env.sh script it may do somethign different | 19:29 |
| hogepodge | inc0: fwiw, I can't get the kolla-ansible bifrost container to work. One problem is the binaries in the virtualenv aren't being put into the exec path, so ironic can't find them. When I fix that iscsi falls over when ironic tries to provision machines with it | 19:29 |
| sdake | we are documeing a non-dev_env deployment guide in etherpad | 19:29 |
| alop | right, I think I'm on the right path now, thanks sdake | 19:29 |
| sdake | alop all godo thanks :) | 19:30 |
| kfox1111 | sdake: theres a py27 check that does some sanity checks on the rendered templates. | 19:30 |
| *** Pavo has joined #openstack-kolla | 19:30 | |
| kfox1111 | the helm template plugin allows you to render templates without having to install k8s/tiller just to spit out some yaml. :) | 19:30 |
| sdake | kfox1111 i see - so linting more or less | 19:30 |
| kfox1111 | yeah. | 19:30 |
| sdake | kfox1111 i attempted to customize the image_name_full btw | 19:31 |
| sdake | kfox1111 and found that doesn't work for the nova bootstrap init job | 19:31 |
| sdake | kfox1111 i looked at the helm chart - it looks like it reference a toolbox container | 19:31 |
| sdake | kfox1111 any idea how to decipher that? | 19:32 |
| kfox1111 | yeah, we need to come up with a document that describes how each package can be configured. or a tool or something. | 19:32 |
| sdake | kfox1111 well i did it with cloud.yaml | 19:32 |
| sdake | kfox1111 i was able to laod the custom image name for nova | 19:32 |
| sdake | however nova's db create job didn't seem to use the same image | 19:32 |
| kfox1111 | sdake: see: helm/kolla-common/templates/_common_create_db_job.yaml | 19:33 |
| sdake | i saw that - it looks like it pulls in toolbox | 19:33 |
| kfox1111 | it does use kolla_toolbox to do the db creation. | 19:33 |
| sdake | kolla_toolbox contains every db bootstrap then? | 19:34 |
| sdake | the container was in a crash loopbackoff | 19:34 |
| sdake | (the db create job) | 19:34 |
| sdake | nova itself was in an init state iirc | 19:34 |
| kfox1111 | I think the create db is basically just 'mysql create db; mysql create user; mysql perm user x to db y' | 19:34 |
| kfox1111 | the manage-db is service specific, which loads the schema. | 19:35 |
| sdake | i see | 19:35 |
| *** tonanhngo has joined #openstack-kolla | 19:35 | |
| *** tonanhngo has quit IRC | 19:35 | |
| kfox1111 | firefox can't get a feature to allow you to figure out what tab is eating up 100% cpu fast enough IMO. :/ | 19:36 |
| sdake | kfox1111 how does one go about customizing the image name in kolla-kubernetes? | 19:41 |
| sdake | kfox1111 have image tag customization sorted out | 19:41 |
| *** manheim has joined #openstack-kolla | 19:41 | |
| kfox1111 | we support either tweaking just the tag, or setting image_full. | 19:41 |
| sdake | OH | 19:41 |
| sdake | image_full = centos-source-nova-api:4.0.0.A? | 19:42 |
| kfox1111 | if you don't follow the naming convention for image name, then you might not follow it for tags too. | 19:42 |
| kfox1111 | yeah. | 19:42 |
| sdake | i didn't have the tag in there | 19:42 |
| kfox1111 | and namespace too. | 19:42 |
| kfox1111 | and hub if you got one. | 19:42 |
| sdake | ya i did have docker.io/kolla/centos-source-nova-api | 19:42 |
| sdake | and that didn't work | 19:42 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Speed up builds by increasing threads for builds https://review.openstack.org/446701 | 19:42 |
| *** Pavo has quit IRC | 19:42 | |
| kfox1111 | the issue is, image name is built out of components normally. | 19:43 |
| sdake | i would think if the tag overrides it overrides always | 19:43 |
| kfox1111 | $distro-$[binary/source]-nova-api | 19:43 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Remove global privilege escalation for deploy_aio.sh https://review.openstack.org/446645 | 19:43 |
| kfox1111 | well, | 19:43 |
| *** shardy has quit IRC | 19:44 | |
| kfox1111 | $reposerver/$repo/$distro-$[binary/source]-nova-api | 19:44 |
| kfox1111 | if you follow the pattern, you can override any of those things. | 19:44 |
| kfox1111 | if you deviate from the pattern, you must specify it all as part of _full | 19:45 |
| sdake | and _full must include a tag? | 19:45 |
| kfox1111 | yeah. | 19:45 |
| kfox1111 | _full is handed directly to image: | 19:45 |
| sdake | image_short should be a thing ;-) | 19:45 |
| kfox1111 | people will be shocked to hear it from me, but, I think thats too much configurability. ;) | 19:45 |
| sdake | or image_name | 19:45 |
| *** tonanhngo has joined #openstack-kolla | 19:46 | |
| sdake | perhaps image_full_and_tag would help clarify it | 19:46 |
| kfox1111 | is image_name = nova-api, and the centos-source override stuff tacked on? | 19:46 |
| sdake | image_name = nova-api | 19:46 |
| kfox1111 | image_full is litterally the full thingy. | 19:46 |
| *** vhosakot has joined #openstack-kolla | 19:46 | |
| sdake | kfox1111 right got it | 19:47 |
| kfox1111 | thats why its full. :) | 19:47 |
| inc0 | kfox1111: can you make it to next weeks meeting? | 19:47 |
| kfox1111 | otherwise, we get image_full_with_type_and_distro_and_repo_and.... for clarity. :) | 19:47 |
| inc0 | we have one thing to talk about which needs your presence | 19:47 |
| kfox1111 | inc0: I'm going to try | 19:47 |
| kfox1111 | k. Ill try hard. :) | 19:47 |
| inc0 | thank you good sir;) | 19:47 |
| kfox1111 | np. :) | 19:47 |
| inc0 | we can move it around | 19:47 |
| inc0 | just let me know | 19:48 |
| inc0 | but quick context - canonical has it's own "proper" way to deploy k8s | 19:48 |
| inc0 | and offered help to work with us on CI and stuff to streamline it | 19:48 |
| kfox1111 | cool. | 19:49 |
| kfox1111 | which 'it' are we talking about? | 19:49 |
| kfox1111 | kolla-kubernetes? | 19:49 |
| inc0 | yup | 19:49 |
| kfox1111 | cool. :) | 19:50 |
| inc0 | gates and/or dev guide | 19:50 |
| * srwilkers sees charms in the future | 19:51 | |
| sdake | alop looks like 2.2.2 with the templates works | 19:51 |
| *** Jezekus has quit IRC | 19:51 | |
| sdake | alop https://review.openstack.org/#/c/447102/ | 19:52 |
| kfox1111 | srwilkers: thats some of that alternative orchestration options I've been saying someone will want to code at some point or another. :) | 19:52 |
| kfox1111 | sdake: cool. | 19:52 |
| srwilkers | kfox1111, :) | 19:52 |
| sdake | kfox1111 it appears atlaeest to pass the gates - which are extensively annoying :) | 19:53 |
| sdake | kfox1111 are we waiting on something to make deploy gates voting? | 19:53 |
| *** kbaegis has quit IRC | 19:53 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 19:54 |
| kfox1111 | sdake: no, I think we can make them voting at any time? | 19:54 |
| sdake | kfox1111 i will make it so | 19:54 |
| kfox1111 | sdake: we've been treating them as voting for a long time. | 19:54 |
| kfox1111 | k. | 19:54 |
| sdake | kfox1111 after i get back from a tour of a gaming office today ;) | 19:54 |
| sbezverk | kfox1111: I am not sure it is a good idea to make voting gates version 2.0.0 jobs | 19:56 |
| sbezverk | kfox1111: it seems you can run just bare basic things with this version and it should not be our target | 19:57 |
| kfox1111 | sbezverk: they test a lot of stuff 3.x currently doesnt. | 19:57 |
| sbezverk | kfox1111: 2.0.0 cannot test lots of things too | 19:57 |
| sdake | are the 4.0.0 helm charts going to be compatible with 2.0.0 charts? | 19:57 |
| kfox1111 | yeah. but that doens't meen they shouldn't be voting? | 19:57 |
| sdake | given the placement api thing | 19:57 |
| kfox1111 | sdake: yeah. placement is purely additive. | 19:58 |
| sdake | there was something else to 4.0.0 -forget which | 19:58 |
| sbezverk | sdake: kfox1111: running old images adds complexity | 19:58 |
| sbezverk | as you need to check which features are supported which not | 19:58 |
| kfox1111 | not testing old images doesn't allow easy migration to kolla-kubernetes. | 19:58 |
| sdake | sbezverk yup we had this discussion in kolla-ansible a long long time ago | 19:58 |
| sdake | and pretty much decided to only implement master plus latest release | 19:58 |
| kfox1111 | once we're closer to a 1.0, I'm totally for dropping all older versions and only doing one version at a time per release. | 19:59 |
| sbezverk | sdake: right | 19:59 |
| kfox1111 | with the quick turnover in functionality in the repo though, cutting off a version from trunk basically seals its fate as not an option for a migration path. | 19:59 |
| kfox1111 | if it was a large burdon to carry, I'd say we drop it. but I don't think it currently is. | 20:00 |
| kfox1111 | and a migration path for folks with older clouds has some big benifits. | 20:00 |
| *** dave-mccowan has quit IRC | 20:07 | |
| sbezverk | so many version makes life more and more difficult, now this damn cells :( | 20:08 |
| kfox1111 | new versions are making things difficult. | 20:09 |
| kfox1111 | not sure the old ones are. | 20:09 |
| kfox1111 | but ops have known that for a long time. :/ | 20:10 |
| kfox1111 | devs like to break stuff all the time moving forward. :/ | 20:10 |
| sbezverk | kfox1111: but it is very hard without moving forward | 20:11 |
| *** rwallner has joined #openstack-kolla | 20:12 | |
| *** rwallner has quit IRC | 20:17 | |
| kfox1111 | what is? | 20:17 |
| kfox1111 | not breaking things? | 20:17 |
| sbezverk | kfox1111: no :) developmnt.. | 20:18 |
| kfox1111 | you can move forward and develop stuff without breaking things. :) | 20:18 |
| kfox1111 | or you can break things and pass the buck to the ops to deal with. | 20:18 |
| kfox1111 | sometimes breaking things really is the right solution. | 20:18 |
| kfox1111 | but often its just becuase of lack of developer attention. | 20:18 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Add mechanism to store keys locally to avoid going out to network https://review.openstack.org/447101 | 20:22 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Speed up builds by increasing threads for builds https://review.openstack.org/446701 | 20:23 |
| sbezverk | kfox1111: could you check this, http://paste.openstack.org/show/603184/ ny head is about to explode from training :( by some reason it does not run all caoomnds, do you see something wrong? | 20:24 |
| *** manheim has quit IRC | 20:26 | |
| mnaser | i think your pipe is breaking things sbezverk | 20:26 |
| mnaser | && (yes | cp -aLf {{ $containerConfigDirectory }}/..data/nova.conf /etc/nova/nova.conf) \ | 20:27 |
| sbezverk | mnaser: have noi choice since nova.conf already exists | 20:27 |
| sbezverk | needs to overwrtie it wihtout prompt | 20:27 |
| mnaser | yes i know | 20:27 |
| mnaser | but if you add the parantheses(spelling?) around it | 20:28 |
| mnaser | it will run that command as a unit | 20:28 |
| mnaser | instead it's piping the whole thing and creating a mess | 20:28 |
| sbezverk | mnaser: ah, true, let me try it, thanks for the idea | 20:28 |
| mnaser | also sbezverk its probably the bash alias breaking things | 20:28 |
| mnaser | instead of that | 20:28 |
| mnaser | just do /bin/cp | 20:28 |
| mnaser | cp is aliased to cp -i usually | 20:28 |
| sbezverk | mnaser: cool will do both ;) just in case | 20:29 |
| *** jmontleon has quit IRC | 20:30 | |
| kfox1111 | sbezverk: looking | 20:32 |
| *** jmontleon has joined #openstack-kolla | 20:32 | |
| *** jtriley has quit IRC | 20:32 | |
| kfox1111 | sbezverk: looks ok to me. | 20:33 |
| kfox1111 | the pipe only shoudl go from the yes to the cp, but not sure the follow on &&'s error behavior. | 20:34 |
| kfox1111 | parans would help there I think. | 20:34 |
| sbezverk | kfox1111" thanks | 20:35 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 20:35 |
| *** srwilkers has quit IRC | 20:40 | |
| *** kbaegis has joined #openstack-kolla | 20:41 | |
| *** Serlex has quit IRC | 20:50 | |
| *** eaguilar has quit IRC | 20:52 | |
| *** eaguilar has joined #openstack-kolla | 20:52 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 21:10 |
| *** ipsecguy has quit IRC | 21:21 | |
| *** ipsecguy has joined #openstack-kolla | 21:22 | |
| *** sbezverk has quit IRC | 21:23 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** schwicht has quit IRC | 21:36 | |
| *** Jeffrey4l has joined #openstack-kolla | 21:36 | |
| *** kbaegis has quit IRC | 21:37 | |
| *** srwilkers has joined #openstack-kolla | 21:40 | |
| *** lamt has joined #openstack-kolla | 21:43 | |
| *** srwilkers has quit IRC | 21:53 | |
| *** lamt has quit IRC | 21:53 | |
| *** lamt has joined #openstack-kolla | 22:01 | |
| *** eaguilar has quit IRC | 22:10 | |
| *** sbezverk has joined #openstack-kolla | 22:12 | |
| *** m1dev has quit IRC | 22:12 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding virtualbmc to the gate https://review.openstack.org/441550 | 22:16 |
| *** saneax-_-|AFK is now known as saneax | 22:19 | |
| *** kbaegis has joined #openstack-kolla | 22:19 | |
| *** schwicht has joined #openstack-kolla | 22:22 | |
| *** kbaegis has quit IRC | 22:27 | |
| *** kbaegis has joined #openstack-kolla | 22:27 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-ansible master: WIP: Fix for Kolla-Kubernetes Fernet Tokens https://review.openstack.org/447159 | 22:31 |
| *** alop has quit IRC | 22:31 | |
| *** kbaegis has quit IRC | 22:32 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Fernet Token Support https://review.openstack.org/446274 | 22:32 |
| sdake | portdirect kfox1111 sbezverk how about some reviews on https://review.openstack.org/#/c/447102/ | 22:34 |
| sdake | or inc0 if your about | 22:34 |
| kfox1111 | looking | 22:34 |
| kfox1111 | +2 | 22:35 |
| sdake | tx | 22:35 |
| kfox1111 | we need to split that those values out somewhere though, so we have a shared shell script with those versions in it. | 22:35 |
| kfox1111 | so we don't keep forgetting where all the places are. :/ | 22:35 |
| sdake | you mean like a versions.sh or something? | 22:35 |
| kfox1111 | yeah. | 22:35 |
| sdake | i'll see if i can sort that out | 22:35 |
| kfox1111 | cool. thanks. :) | 22:36 |
| *** rwallner has joined #openstack-kolla | 22:37 | |
| *** rwallner has quit IRC | 22:42 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Update to Helm 2.2.2 https://review.openstack.org/447102 | 22:47 |
| *** kbaegis has joined #openstack-kolla | 22:49 | |
| *** kbaegis has quit IRC | 22:50 | |
| *** schwicht has quit IRC | 22:51 | |
| *** kbaegis has joined #openstack-kolla | 22:52 | |
| *** kbaegis has quit IRC | 22:56 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Fernet Token Support https://review.openstack.org/446274 | 22:57 |
| *** lamt has quit IRC | 22:59 | |
| *** jbadiapa has quit IRC | 23:01 | |
| openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Remove EXPORT and use search and replace instead https://review.openstack.org/447164 | 23:06 |
| sdake | portdirect kfox1111 ^^ | 23:06 |
| *** mgoddard has quit IRC | 23:10 | |
| openstackgerrit | Sayantani Goswami proposed openstack/kolla-ansible master: Added playbooks for snap deployment https://review.openstack.org/446165 | 23:11 |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Update Searchlight resource_plugin configuration https://review.openstack.org/442628 | 23:11 |
| *** erlon has quit IRC | 23:15 | |
| *** schwicht has joined #openstack-kolla | 23:17 | |
| openstackgerrit | Sayantani Goswami proposed openstack/kolla-ansible master: [WIP] Added playbooks for snap deployment https://review.openstack.org/446165 | 23:18 |
| *** eaguilar has joined #openstack-kolla | 23:21 | |
| *** kbaegis has joined #openstack-kolla | 23:24 | |
| *** kbaegis has quit IRC | 23:24 | |
| *** eaguilar has quit IRC | 23:27 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Fernet Token Support https://review.openstack.org/446274 | 23:29 |
| *** unicell has quit IRC | 23:31 | |
| sdake | kfox1111 around? | 23:51 |
| sdake | kfox1111 next week is mostly clear for me | 23:51 |
| sdake | kfox1111 plan to make external config helm charts | 23:51 |
| sdake | kfox1111 do we need one per microchart - or one per service? | 23:51 |
| sdake | sbezverk did the ironic service chart land? | 23:53 |
| *** kbaegis has joined #openstack-kolla | 23:54 | |
| spsurya | morning all | 23:58 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Fernet Token Support https://review.openstack.org/446274 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!