| *** dwalsh has joined #openstack-kolla | 00:01 | |
| *** mdorman has quit IRC | 00:03 | |
| *** zhiwei has joined #openstack-kolla | 00:06 | |
| *** mbound has quit IRC | 00:07 | |
| openstackgerrit | Joshua Harlow proposed openstack/kolla: Allow for externally managed configuration https://review.openstack.org/332590 | 00:11 |
|---|---|---|
| *** zhiwei has quit IRC | 00:11 | |
| *** mdorman has joined #openstack-kolla | 00:14 | |
| *** jtriley has joined #openstack-kolla | 00:19 | |
| *** daneyon has quit IRC | 00:53 | |
| *** phuongnh has joined #openstack-kolla | 00:56 | |
| *** dwalsh has quit IRC | 01:00 | |
| *** fragatina has joined #openstack-kolla | 01:06 | |
| *** mbound has joined #openstack-kolla | 01:08 | |
| *** jpeeler has joined #openstack-kolla | 01:09 | |
| *** v1k0d3n has joined #openstack-kolla | 01:09 | |
| *** mbound has quit IRC | 01:13 | |
| *** huikang has joined #openstack-kolla | 01:18 | |
| *** zhiwei has joined #openstack-kolla | 01:18 | |
| *** v1k0d3n has quit IRC | 01:23 | |
| *** jpeeler has quit IRC | 01:27 | |
| *** sacharya has joined #openstack-kolla | 01:46 | |
| *** sacharya has quit IRC | 01:50 | |
| *** salv-orlando has joined #openstack-kolla | 02:00 | |
| *** salv-orlando has quit IRC | 02:07 | |
| *** jtriley has quit IRC | 02:11 | |
| *** daneyon has joined #openstack-kolla | 02:12 | |
| *** daneyon has quit IRC | 02:17 | |
| *** mewald1 has joined #openstack-kolla | 02:19 | |
| *** coolsvap has joined #openstack-kolla | 02:20 | |
| *** mewald has quit IRC | 02:21 | |
| *** daneyon has joined #openstack-kolla | 02:31 | |
| *** stvnoyes1 has joined #openstack-kolla | 02:37 | |
| *** stvnoyes has quit IRC | 02:40 | |
| *** yuanying has quit IRC | 02:46 | |
| coolsvap | sdake, have a look at https://hub.docker.com/u/kolla/ | 02:51 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/kolla: Remove deprecated package python-tuskarclient https://review.openstack.org/332609 | 02:57 |
| *** caowei has joined #openstack-kolla | 02:57 | |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add Kuryr Docker container https://review.openstack.org/298000 | 02:57 |
| *** daneyon_ has joined #openstack-kolla | 03:08 | |
| *** daneyon has quit IRC | 03:10 | |
| *** salv-orlando has joined #openstack-kolla | 03:14 | |
| *** salv-orlando has quit IRC | 03:21 | |
| coolsvap | centos-binary is busted in master | 03:23 |
| huikang | as well as ubuntu-binary | 03:34 |
| huikang | I meant oraclelinux-binary | 03:35 |
| coolsvap | yes both depend on same packaging source rdo | 03:36 |
| *** Kennan has quit IRC | 03:42 | |
| *** yuanying has joined #openstack-kolla | 03:47 | |
| *** dave-mcc_ has quit IRC | 03:48 | |
| *** sacharya has joined #openstack-kolla | 03:59 | |
| *** williamcaban has joined #openstack-kolla | 04:03 | |
| *** williamcaban has quit IRC | 04:14 | |
| *** williamcaban has joined #openstack-kolla | 04:15 | |
| *** huikang has quit IRC | 04:21 | |
| *** huikang has joined #openstack-kolla | 04:21 | |
| *** Kennan has joined #openstack-kolla | 04:22 | |
| *** huikang has quit IRC | 04:26 | |
| openstackgerrit | Md Nadeem proposed openstack/kolla: Add unit test for stop and restart container https://review.openstack.org/331524 | 04:26 |
| *** salv-orlando has joined #openstack-kolla | 04:28 | |
| *** Kennan2 has joined #openstack-kolla | 04:30 | |
| *** sacharya has quit IRC | 04:31 | |
| *** Kennan has quit IRC | 04:33 | |
| *** salv-orlando has quit IRC | 04:35 | |
| openstackgerrit | Md Nadeem proposed openstack/kolla: Add unit test for stop and restart container https://review.openstack.org/331524 | 04:38 |
| *** williamcaban has quit IRC | 04:39 | |
| *** williamcaban has joined #openstack-kolla | 04:49 | |
| *** huikang has joined #openstack-kolla | 04:50 | |
| *** daneyon has joined #openstack-kolla | 04:55 | |
| *** daneyon has quit IRC | 04:59 | |
| *** caowei has quit IRC | 05:01 | |
| *** Jeffrey4l_ has quit IRC | 05:01 | |
| *** daneyon has joined #openstack-kolla | 05:12 | |
| *** daneyon_ has quit IRC | 05:12 | |
| *** huikang has quit IRC | 05:15 | |
| *** huikang has joined #openstack-kolla | 05:16 | |
| *** shardy has joined #openstack-kolla | 05:20 | |
| *** huikang has quit IRC | 05:21 | |
| *** caowei has joined #openstack-kolla | 05:30 | |
| *** tfukushima has joined #openstack-kolla | 05:32 | |
| *** salv-orlando has joined #openstack-kolla | 05:37 | |
| *** shardy has quit IRC | 05:38 | |
| *** daneyon has quit IRC | 05:40 | |
| *** hanchao has joined #openstack-kolla | 05:41 | |
| hanchao | hello, is there a clear openstack release matched to kolla ? I tried to find that, but it makes me a bit confused... | 05:43 |
| coolsvap | hanchao, can you explain your question? what exactly are you looking for? | 05:49 |
| hanchao | coolsvap yes, I'm trying to figure out the relationship of the release of openstack and kolla. E.g kolla release 3.0.0 indicates to the current developing version of openstack. then which release indicates to Liberty or Mitaka? | 05:52 |
| coolsvap | mitaka is 2.x current mitaka release is 2.0.1 released on June 06 2016 | 05:54 |
| coolsvap | liberty is 1.x current liberty release is 1.1.1 June 06 2016 | 05:54 |
| hanchao | ok, but is there a clear document which mentions about this? | 05:55 |
| *** caowei has quit IRC | 05:56 | |
| hanchao | I can only find something here [http://docs.openstack.org/releasenotes/kolla/] | 05:56 |
| hanchao | but it looks liberty and mitaka are quite close to each other. | 05:56 |
| openstackgerrit | Md Nadeem proposed openstack/kolla: Add test for remove container, get state and get env https://review.openstack.org/331556 | 05:58 |
| *** gfidente has joined #openstack-kolla | 06:00 | |
| *** williamcaban has quit IRC | 06:01 | |
| *** Jeffrey4l_ has joined #openstack-kolla | 06:09 | |
| openstackgerrit | Md Nadeem proposed openstack/kolla: Add test for remove container, get state and get env https://review.openstack.org/331556 | 06:13 |
| coolsvap | hanchao, one of the places i can provide you is https://launchpad.net/kolla | 06:13 |
| hanchao | coolsvap, thanks, here I can see a clear roadmap. :) | 06:15 |
| hanchao | Then, will this milestones be updated/synchronized into the docs of the release note? | 06:18 |
| openstackgerrit | Md Nadeem proposed openstack/kolla: Add test for remove container, get state and get env https://review.openstack.org/331556 | 06:20 |
| *** caowei has joined #openstack-kolla | 06:36 | |
| *** salv-orl_ has joined #openstack-kolla | 06:38 | |
| *** salv-orlando has quit IRC | 06:41 | |
| mewald1 | can someone shed some light on how logging works with Kolla? Are logs stored in the containers for example? | 06:41 |
| *** jmccarthy has joined #openstack-kolla | 06:42 | |
| *** daneyon has joined #openstack-kolla | 06:43 | |
| *** JeroenBo has joined #openstack-kolla | 06:45 | |
| *** david-lyle_ has joined #openstack-kolla | 06:46 | |
| hanchao | @mewald1 good point, I wish to learn that, too. | 06:46 |
| *** daneyon has quit IRC | 06:48 | |
| *** slagle_ has joined #openstack-kolla | 06:50 | |
| *** janmalax_ has joined #openstack-kolla | 06:50 | |
| hanchao | @mewald1 seems a log can be checked inside a running container. | 06:51 |
| *** athomas has joined #openstack-kolla | 06:53 | |
| *** dasm_ has joined #openstack-kolla | 06:53 | |
| hanchao | Then where can we find the log for kolla? I did not see it in the `/var/log/`. Like the `kolla-ansible deploy` can raise exceptions, however the info of the stdout is too less to indicating the root. | 06:54 |
| *** dgonzalez_ has joined #openstack-kolla | 06:55 | |
| *** intr1nsic_ has joined #openstack-kolla | 06:55 | |
| *** Serlex has joined #openstack-kolla | 06:56 | |
| *** shardy has joined #openstack-kolla | 06:56 | |
| *** dmsimard1 has joined #openstack-kolla | 06:57 | |
| *** sacharya has joined #openstack-kolla | 06:59 | |
| *** Mr_Broken has joined #openstack-kolla | 07:00 | |
| *** ccesario_ has joined #openstack-kolla | 07:01 | |
| *** david-lyle has quit IRC | 07:04 | |
| *** dgonzalez has quit IRC | 07:04 | |
| *** ccesario has quit IRC | 07:04 | |
| *** janmalax has quit IRC | 07:04 | |
| *** dmsimard has quit IRC | 07:04 | |
| *** intr1nsic has quit IRC | 07:04 | |
| *** slagle has quit IRC | 07:04 | |
| *** dasm has quit IRC | 07:04 | |
| *** dgonzalez_ is now known as dgonzalez | 07:04 | |
| *** dasm_ is now known as dasm | 07:04 | |
| *** intr1nsic_ is now known as intr1nsic | 07:04 | |
| *** Serlex has quit IRC | 07:06 | |
| *** fragatina has quit IRC | 07:07 | |
| *** fragatina has joined #openstack-kolla | 07:07 | |
| *** Kennan2 has quit IRC | 07:09 | |
| *** sacharya has quit IRC | 07:18 | |
| *** salv-orl_ has quit IRC | 07:20 | |
| mewald1 | hanchao: there is a debug flag in kolla-build.conf | 07:20 |
| mewald1 | maybe that helps | 07:21 |
| hanchao | but still, this is for the stdout, not recorded in a file... | 07:22 |
| mewald1 | true | 07:22 |
| mewald1 | for the build an can live with that though. The runtime logs of services I want to have somewhere safe :) https://github.com/openstack/kolla/blob/master/specs/logging-with-heka.rst This seems interesting in this context | 07:23 |
| *** jmccarthy has left #openstack-kolla | 07:44 | |
| Mech422__ | mewald1: the service logs are available in the 'heka' container on each host | 07:50 |
| Mech422__ | mewald1: in /var/log/kolla/X | 07:50 |
| *** Serlex has joined #openstack-kolla | 07:52 | |
| *** b_bezak has joined #openstack-kolla | 07:56 | |
| mewald1 | Mech422__: yeah I just figured that out :) I found there is "kolla_log" volume where everything is gathered | 07:59 |
| mewald1 | But I haven't found out what the heka_socket volume is there for | 07:59 |
| Mech422__ | I tried the 'kibana' thing - created the index, but all I found in it was apache stuff | 08:00 |
| Mech422__ | Horizon I think | 08:00 |
| mewald1 | What is this heka_socket volume all about? | 08:07 |
| mewald1 | does heka throw the logs into that socket? what's the point? | 08:07 |
| *** salv-orlando has joined #openstack-kolla | 08:14 | |
| *** JianqingJiang has joined #openstack-kolla | 08:16 | |
| JianqingJiang | hi | 08:17 |
| mewald1 | hi | 08:17 |
| JianqingJiang | can password in /etc/kolla/passwords.yml be empty? | 08:17 |
| mewald1 | Mech422__: What did you do to deploy elasticsearch and kibana? How does it work? | 08:18 |
| mewald1 | JianqingJiang: Nope, I dont think so | 08:18 |
| JianqingJiang | I run precheck , and error log appears | 08:18 |
| Mr_Broken | Hey Guys | 08:18 |
| Mr_Broken | are you aware of a simmilar bug opened for glance ? https://launchpad.net/bugs/1571211 | 08:18 |
| openstack | Launchpad bug 1571211 in Cinder "In multinode setup VM fails to launch due to cinder not checking with all glance api servers" [Medium,In progress] - Assigned to John Griffith (john-griffith) | 08:18 |
| JianqingJiang | But I fill the password up and it also show error log | 08:19 |
| Mr_Broken | there is same problem with glance-cache-manage, even if you use different glance/cinder backends (and try to cache the images) | 08:19 |
| JianqingJiang | And I try to run kolla-genpwd command ,but this command is not found in CentOS7 | 08:20 |
| Mech422__ | JianqingJiang: oh - it might be in like ./.tox/..../bin | 08:20 |
| JianqingJiang | Please help me figure it out | 08:20 |
| Mech422__ | JianqingJiang: ./.tox/genconfig/bin/kolla-genpwd | 08:23 |
| *** mbound has joined #openstack-kolla | 08:23 | |
| JianqingJiang | I only find /root/kolla/kolla/cmd/genpwd.py | 08:23 |
| *** dmk0202 has joined #openstack-kolla | 08:23 | |
| Mech422__ | JianqingJiang: Hmm - you ran the kolla-build ? | 08:24 |
| Mech422__ | I think thats what fires off the Tox stuff | 08:24 |
| JianqingJiang | I can't | 08:24 |
| JianqingJiang | I can't find ./.tox/genconfig/bin/kolla-genpwd | 08:25 |
| Mech422__ | JianqingJiang: I think you have to run the kolla-build stuff first, then the preflight checks, then deploy | 08:25 |
| Mech422__ | err.. preflight checks, configure, then deploy | 08:25 |
| JianqingJiang | ./root/kolla/.tox/genconfig/bin/kolla-genpwd | 08:26 |
| JianqingJiang | I find it in this path | 08:26 |
| Mech422__ | yeah - its in the .tox dir of your kolla git checkout | 08:27 |
| *** sacharya has joined #openstack-kolla | 08:28 | |
| coolsvap | JianqingJiang, you need to use kolla-genpwd command which will update the /etc/kolla/passwords.yml | 08:29 |
| coolsvap | all the steps are documented in quickstart | 08:30 |
| JianqingJiang | OK | 08:30 |
| JianqingJiang | But I think the quickstart is a lit confusing | 08:31 |
| *** sacharya has quit IRC | 08:33 | |
| JianqingJiang | TASK [common : Starting heka bootstrap container] ****************************** | 08:36 |
| JianqingJiang | fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "msg": "Unknown error message: Tag 3.0.0 not found in repository docker.io/kollaglue/centos-source-heka"} | 08:36 |
| JianqingJiang | to retry, use: --limit @/root/kolla/ansible/site.retry | 08:36 |
| JianqingJiang | PLAY RECAP ********************************************************************* | 08:36 |
| JianqingJiang | localhost : ok=11 changed=5 unreachable=0 failed=1 | 08:36 |
| JianqingJiang | I get this error... | 08:36 |
| JianqingJiang | My docker image tag is 2.0.2 | 08:37 |
| Mech422__ | JianqingJiang: you have openstack_version 3.0.0 in globals.yml | 08:38 |
| Mech422__ | JianqingJiang: set it to 2.0.2 | 08:38 |
| JianqingJiang | Thanks | 08:40 |
| JianqingJiang | Unknown error message: Tag 2.0.2 not found in repository docker.io/kollaglue/centos-source-heka | 08:43 |
| JianqingJiang | Is my path wrong ? | 08:43 |
| JianqingJiang | I run "docker images" is ok | 08:44 |
| JianqingJiang | docker version 1.11.2 | 08:44 |
| Mr_Broken | is there 2.0.2 already released ? | 08:45 |
| Mr_Broken | you can set the tag to #latest - it should pick one automagicaly :P | 08:45 |
| Mech422__ | Mr_Broken: yeah - but 2.X is for stable/liberty and 3.0 is for stable/mitaka | 08:55 |
| Mech422__ | or 2.x is for stable/mitaka and 3.x is for master...cant remember atm | 08:55 |
| JianqingJiang | I git checkout mitaka... | 08:56 |
| JianqingJiang | And the image tag is 2.0.2 | 08:56 |
| Mech422__ | ahh - 3.X must be for master then | 08:56 |
| JianqingJiang | Mech422__:But I think is the path... I can't find docker.io/kollaglue | 08:58 |
| Mech422__ | JianqingJiang: Oh..dunno - I've only deployed from a local repo | 08:59 |
| JianqingJiang | I don't have a docker repo... | 09:00 |
| Mech422__ | JianqingJiang: yeah - I stand one up in a container before I run kolla... | 09:01 |
| Mech422__ | JianqingJiang: its just a 'regular' docker container, with 1 exposed port | 09:02 |
| JianqingJiang | find -iname "*docker.io*" | 09:06 |
| JianqingJiang | I ran this command and nothing shows | 09:06 |
| JianqingJiang | Is it ok ? | 09:06 |
| Mr_Broken | 2.0 for mitaka (i am using it atm ) | 09:06 |
| Mr_Broken | and 3.0 for Newton | 09:06 |
| JianqingJiang | But the error log is not found in repository docker.io/kollaglue/centos-source-heka | 09:07 |
| Mr_Broken | is there centos from sources build available for Mitaka ? | 09:07 |
| Mech422__ | Mr_Broken: makes sense - I'm tracking master which is newton now | 09:07 |
| JianqingJiang | Can I change the path? | 09:07 |
| Mr_Broken | not if you are using outside registry | 09:08 |
| Mech422__ | JianqingJiang: its trying to use the docker.io registry | 09:08 |
| Mech422__ | JianqingJiang: which is correct if you don't run local registry | 09:08 |
| Mech422__ | JianqingJiang: but I've never used it | 09:08 |
| Mech422__ | JianqingJiang: and I think coolsvap said some images were broken on there atm... | 09:09 |
| JianqingJiang | But the quickstart says it's ok if I don't use local registry | 09:09 |
| Mech422__ | [20:23] <coolsvap> centos-binary is busted in master | 09:10 |
| Mech422__ | [20:34] <huikang> as well as ubuntu-binary | 09:10 |
| Mech422__ | [20:35] <huikang> I meant oraclelinux-binary | 09:10 |
| Mech422__ | [20:36] <coolsvap> yes both depend on same packaging source rdo | 09:10 |
| Mech422__ | JianqingJiang: yeah - it is OK - they were just working on it today, and I think something is busted atm | 09:10 |
| *** JeroenBo has quit IRC | 09:10 | |
| Mr_Broken | so you either go with Ubuntu, or make centos-binary (not sources) | 09:12 |
| Mr_Broken | or even better - try to deploy using local registry | 09:12 |
| Mech422__ | Mr_Broken: I think you have to build centos from sources - the binary is busted atm | 09:12 |
| Mr_Broken | oh, fair play | 09:12 |
| Mr_Broken | you are doing master/newton JianqinqJiang ? | 09:13 |
| Mech422__ | I always build ubuntu source with a local repo - so I don't know anything about centos or docker.io images | 09:13 |
| JianqingJiang | No,I am just want to deploy the openstack | 09:15 |
| JianqingJiang | But I am interested in developing the project | 09:16 |
| *** JeroenBo has joined #openstack-kolla | 09:24 | |
| *** Jeffrey4l__ has joined #openstack-kolla | 09:25 | |
| *** daneyon has joined #openstack-kolla | 09:26 | |
| *** Jeffrey4l_ has quit IRC | 09:26 | |
| *** daneyon has quit IRC | 09:30 | |
| *** Jeffrey4l__ has quit IRC | 09:35 | |
| mewald1 | Does creating custom build profiles work? I just created a profile called "custom" and "profile = custom" in kolla-build.conf. I ran kolla-build and now everything from keystone to mesosphere is being built. | 09:37 |
| *** pbourke has quit IRC | 09:42 | |
| *** pbourke has joined #openstack-kolla | 09:43 | |
| *** fragatina has quit IRC | 10:01 | |
| JianqingJiang | Can I use my local images to deploy ? I run "docker images" is ok, but when I deploy, the error log shows that kolla is using a outside docker registry... | 10:08 |
| *** matrohon has quit IRC | 10:09 | |
| coolsvap | JianqingJiang, yes i have seen some issues with centos-bnary earlier | 10:11 |
| *** fragatina has joined #openstack-kolla | 10:12 | |
| JianqingJiang | How I can do to solve my problem... | 10:13 |
| Mech422__ | JianqingJiang: you need to run a local registry or do a source build .. | 10:18 |
| *** tfukushima has quit IRC | 10:18 | |
| coolsvap | mewald1, what do you want in custom images? | 10:20 |
| *** caowei has quit IRC | 10:24 | |
| Mech422__ | coolsvap: I have to add the novadocker driver to nova image - would that be best as a customization, build profile', or addition to the existing image? | 10:24 |
| coolsvap | I think it is customization | 10:27 |
| coolsvap | because it is not something which can be worked out by changing configurations during deployment | 10:27 |
| *** Jeffrey4l has joined #openstack-kolla | 10:27 | |
| Mech422__ | coolsvap: thanks! | 10:28 |
| *** tfukushima has joined #openstack-kolla | 10:31 | |
| *** fragatina has quit IRC | 10:31 | |
| coolsvap | JianqingJiang, i do not see you deployed the local docker registry | 10:37 |
| coolsvap | JianqingJiang, please show me the content of /etc/systemd/system/docker.service.d/kolla.conf and /etc/sysconfig/docker | 10:38 |
| *** Jeffrey4l has quit IRC | 10:40 | |
| coolsvap | JianqingJiang, also share the output of docker ps -a | 10:41 |
| *** JianqingJiang has quit IRC | 10:50 | |
| *** coolsvap has quit IRC | 10:56 | |
| *** coolsvap has joined #openstack-kolla | 10:57 | |
| mewald1 | coolsvap: I just want a custom profile, but I think I figured it out. | 11:06 |
| *** Mr_Broken has quit IRC | 11:10 | |
| *** Mr_Broken has joined #openstack-kolla | 11:11 | |
| *** Mr_Broken has quit IRC | 11:16 | |
| *** caowei has joined #openstack-kolla | 11:16 | |
| mewald1 | During build I get no error messages and the final output shows only "built" and "unmatched". In "docker images" I see a lot of images with name "<none>". Why? | 11:16 |
| *** phuongnh has quit IRC | 11:17 | |
| coolsvap | mewald1, there was an issue with rdo repos earlier in the day | 11:29 |
| coolsvap | we had same issue in gates causing build failure | 11:30 |
| coolsvap | this seem to have resolved | 11:30 |
| *** mliima has joined #openstack-kolla | 11:36 | |
| openstackgerrit | Merged openstack/kolla: Updated from global requirements https://review.openstack.org/332302 | 11:47 |
| *** sdake has quit IRC | 11:48 | |
| *** tfukushima has quit IRC | 11:51 | |
| *** dwalsh has joined #openstack-kolla | 11:57 | |
| *** Mr_Broken has joined #openstack-kolla | 11:58 | |
| *** dmsimard1 has quit IRC | 12:00 | |
| *** dmsimard has joined #openstack-kolla | 12:00 | |
| *** sdake has joined #openstack-kolla | 12:03 | |
| sdake | coolsvap did you push teh 3.0.0 tags? | 12:04 |
| coolsvap | sdake, yes | 12:07 |
| sdake | i had only wanted 1.1.0 1.1.1 and 2.0.0 and 2.0.1 pushed | 12:07 |
| sdake | 3.0..0 isn't a thing | 12:07 |
| coolsvap | :( | 12:07 |
| coolsvap | i will delete those tonight | 12:07 |
| sdake | they are really hard to delete | 12:08 |
| coolsvap | i know but i will do | 12:08 |
| *** daneyon has joined #openstack-kolla | 12:08 | |
| sdake | there is no api | 12:08 |
| coolsvap | sorry for not checking with you | 12:08 |
| sdake | its ok - 3.0.0 would lmake sense if we could continuously buil | 12:08 |
| sdake | but that tag would be more like latest | 12:08 |
| coolsvap | sdake, that can be done | 12:09 |
| coolsvap | atleast once a day like a cron job | 12:09 |
| *** rhallisey has joined #openstack-kolla | 12:10 | |
| sdake | the problem is i'd like it in ci of openstack | 12:10 |
| *** ccesario_ is now known as ccesario | 12:10 | |
| sdake | and don't knwo how to get the credentials securely in to the gate | 12:10 |
| sdake | i htikn we would better serve our userbae if we provided our stable images first :) | 12:10 |
| coolsvap | sdake, yes i am waitting for some action on the review :) | 12:11 |
| sdake | i really dont understand why there isn't an image delete python api | 12:11 |
| coolsvap | i can start the job for stable tonight and all the images will be available in the morning | 12:11 |
| sdake | inside ci/cd? | 12:12 |
| coolsvap | sdake, no | 12:12 |
| *** daneyon has quit IRC | 12:12 | |
| coolsvap | for ci/cd i dont think we can do it in prescribed limit of a gate job | 12:13 |
| coolsvap | or can we? | 12:13 |
| coolsvap | not sure | 12:13 |
| sdake | we could if we could get the credentials in the gate | 12:13 |
| sdake | but again i ddont want free passwords logged into our ci scripts | 12:13 |
| sdake | git repo is no place for a credential :) | 12:14 |
| coolsvap | yes | 12:14 |
| sdake | coolsvap i just asked again on openstack infra, recommend joining in case anyone has ideas | 12:17 |
| coolsvap | joined | 12:17 |
| *** dave-mccowan has joined #openstack-kolla | 12:19 | |
| mewald1 | I created a new deployment with Kolla this morning. Now I made sure that elasticsearch and kibana images are built by adding them to my profile. I cant figure out how to deploy them now. According to multinode inventory file, the roles are assigned to "control" but neither kolla-build deploy nor kolla-build pull show anything that relates to elasticsearch. Why is it not being rolled out? | 12:23 |
| sbezverk | coolsvap ping | 12:24 |
| coolsvap | sbezverk, pong | 12:25 |
| sbezverk | coolsvap: do you know if kubernetes 1.2.4 rpm exists anywhere? | 12:25 |
| sbezverk | I am rebuilding my multinode cluster and do not want to install 1.2.0 and then update to to 1.2.4 | 12:26 |
| *** Jeffrey4l has joined #openstack-kolla | 12:26 | |
| Mr_Broken | mewald1: did you put enable_kibana: yes in globals.yml ? | 12:27 |
| Mr_Broken | kolla-build - only builds docker images | 12:28 |
| Mr_Broken | kolla-ansible deploy - will deploy those images on the hosts - jsyk | 12:28 |
| mewald1 | Mr_Broken: ha shit, or course not . yeah sorry I meant "kolla-ansible" before :) | 12:28 |
| Mr_Broken | yeah, edit globals.yml ;] | 12:28 |
| Mr_Broken | it has some undocumented "features" like that ;] | 12:28 |
| coolsvap | sbezverk, let me check | 12:29 |
| mewald1 | Mr_Broken: undocumented features - my favorite ;) | 12:29 |
| sbezverk | coolsvap: thank you | 12:31 |
| Mr_Broken | mewald1: yeah was having same issues when tried to customise my topology . you can enable more sevices this way | 12:36 |
| Mr_Broken | I think in newton docs - its described a little bit more | 12:37 |
| mewald1 | Mr_Broken: I managed to get elasticsearch and kibana deployed now, BUT: How should I be able to access them? Should haproxy listen on the VIP? If yes: haproxy should have been redeployed, no? | 12:37 |
| Mr_Broken | yes it should be listening on the VIP if you setup HAProxy | 12:37 |
| mewald1 | So, how do I "change" the haproxy config now? | 12:38 |
| Mr_Broken | http://controler:5601/kibana | 12:38 |
| mewald1 | I just found kolly-ansible reconfigure - does that do it? | 12:38 |
| Mr_Broken | it should be changed during deployment | 12:38 |
| mewald1 | hmm, ahh can see the port | 12:38 |
| mewald1 | thanks! | 12:38 |
| Mr_Broken | np ;] | 12:38 |
| *** salv-orl_ has joined #openstack-kolla | 12:39 | |
| Mr_Broken | whenever you run kolla-ansible it should go thorugh all of the roles/templates and change whats needed ( and skip whats already done ) | 12:40 |
| mewald1 | ok, what's the default user for kibana? :D | 12:40 |
| mewald1 | the kibana default or some other? | 12:40 |
| Mr_Broken | the thing is - when you change your topology a bit (like delete a docker , and install some service outside kolla ) - reconfigure will not work for you :/ | 12:40 |
| Mr_Broken | hmm dont remember now ;] | 12:40 |
| Mr_Broken | i think you have all you need in passwords.yml ? | 12:41 |
| mewald1 | ok, Ill figure it out :) | 12:41 |
| Mr_Broken | there is also a prefix you need to configure when you start kibana for the first time | 12:41 |
| Mr_Broken | you need to check the docs :> | 12:41 |
| kfox1111 | Can you override a template with your own? | 12:41 |
| *** salv-orlando has quit IRC | 12:41 | |
| kfox1111 | for example, if I wanted to use my own local_settings for horizon? | 12:41 |
| Mr_Broken | kfox1111: you can change services configuration | 12:42 |
| Mr_Broken | let me find the link describing that | 12:42 |
| kfox1111 | thx. | 12:43 |
| *** sdake has quit IRC | 12:43 | |
| Mr_Broken | https://github.com/openstack/kolla/blob/master/doc/advanced-configuration.rst#openstack-service-configuration-in-kolla | 12:43 |
| Mr_Broken | basicaly you create /etc/kolla/config/service_name/config.conf - edit this files with lines you want to append | 12:44 |
| Mr_Broken | and run kolla-ansible reconfigure | 12:44 |
| kfox1111 | Ah. perfect. thanks. | 12:45 |
| *** ppowell has joined #openstack-kolla | 12:47 | |
| coolsvap | Jeffrey4l, around? | 12:49 |
| Jeffrey4l | coolsvap, yep | 12:50 |
| coolsvap | Jeffrey4l, i pushed the images today, i will push them for mitaka and liberty | 12:50 |
| Jeffrey4l | coolsvap, cool | 12:50 |
| coolsvap | if you want the access rights ping sdake with your email id for docker hub | 12:50 |
| coolsvap | he pinged most cores yesterday | 12:50 |
| Jeffrey4l | roger that. thanks for the information. | 12:51 |
| *** vhosakot has joined #openstack-kolla | 12:53 | |
| *** zhiwei has quit IRC | 12:53 | |
| *** JeroenBo has left #openstack-kolla | 12:55 | |
| openstackgerrit | Merged openstack/kolla: Notify operators that cinder + LVM doesn't work https://review.openstack.org/332228 | 12:57 |
| *** dwalsh has quit IRC | 12:57 | |
| *** dwalsh has joined #openstack-kolla | 12:59 | |
| *** salv-orl_ has quit IRC | 13:03 | |
| *** zhiwei has joined #openstack-kolla | 13:05 | |
| *** zhiwei has quit IRC | 13:09 | |
| *** vhosakot has quit IRC | 13:09 | |
| openstackgerrit | Merged openstack/kolla: Add Kuryr Docker container https://review.openstack.org/298000 | 13:09 |
| *** Mr_Broken has quit IRC | 13:09 | |
| *** slagle_ is now known as slagle | 13:10 | |
| openstackgerrit | Merged openstack/kolla: Remove deprecated package python-tuskarclient https://review.openstack.org/332609 | 13:10 |
| *** Mr_Broken has joined #openstack-kolla | 13:10 | |
| *** Mr_Broke_ has joined #openstack-kolla | 13:11 | |
| *** salv-orlando has joined #openstack-kolla | 13:12 | |
| *** Mr_Broken has quit IRC | 13:14 | |
| dims | rhallisey : the kolla meeting is in another 2 and half hours...right? | 13:20 |
| rhallisey | dims, noon EST | 13:21 |
| dims | ack thanks | 13:21 |
| rhallisey | no problem | 13:21 |
| dims | rhallisey : the official records needs to reflect weekly - https://review.openstack.org/#/c/327845/1/meetings/kolla-team-meeting.yaml | 13:23 |
| patchbot | dims: patch 327845 - openstack-infra/irc-meetings - Move kolla meeting to 16:00 UTC (MERGED) | 13:23 |
| dims | that shows kolla skipping every week :) | 13:24 |
| dims | every other week | 13:24 |
| rhallisey | dims, good point :) | 13:24 |
| dims | "frequency: biweekly-even" | 13:24 |
| mewald1 | Mr_Broke_: You might remember I just deploy ES and Kibana with your help. I believe heka should show a configuration to send logs to ES, correct? If so: where should I find it? | 13:24 |
| rhallisey | i'll fix that real quick.. | 13:25 |
| dims | thanks rhallisey | 13:25 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fix docs for Ansible 2.1 support, remove some of the notes. https://review.openstack.org/330805 | 13:25 |
| dims | slukjanov got the date wrong and is on a train at the moment... | 13:25 |
| Mr_Broke_ | @mewald1: have you seen this ? http://docs.openstack.org/developer/kolla/kibana-guide.html | 13:27 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Adding debugging documentation for Memcached https://review.openstack.org/330766 | 13:27 |
| Mr_Broke_ | mewald1: it should be done automagicaly, what you need to do is configure kibana after first login - dscribed in the guide | 13:28 |
| *** jtriley has joined #openstack-kolla | 13:30 | |
| mewald1 | Mr_Broke_: yeah I found it, but Kibana gives me "Pattern matches 0% of existing indices and aliases" so I assume messages are not reaching Elasticsearch properly | 13:33 |
| Mr_Broke_ | and you did :Index name or pattern - log-* ? when loggin for the first time ? | 13:35 |
| Mr_Broke_ | and created an index with the button ? | 13:35 |
| mewald1 | yup, but I have a clue: I ran "kolla-ansible reconfigure" and now in the heka container /etc/heka/heka-elasticsearch.toml appeared | 13:35 |
| mewald1 | it wasn't there before | 13:36 |
| mewald1 | now I am able to use log-* as pattern in Kibana, too | 13:36 |
| Mr_Broke_ | weird, it should be there after first deployment | 13:36 |
| Mr_Broke_ | great | 13:36 |
| *** sdake has joined #openstack-kolla | 13:37 | |
| *** inc0 has joined #openstack-kolla | 13:38 | |
| inc0 | good morning | 13:38 |
| dims | rhallisey : we probably need to hold the discussion on the fuel/container-agnostic thing next week :( | 13:38 |
| dims | inc0 : sdake : ^^ | 13:38 |
| rhallisey | kk | 13:39 |
| inc0 | dims, sure, any chances one of you could go for midcycle? | 13:39 |
| inc0 | so we can have also f2f? | 13:39 |
| inc0 | dims, basically the only thing I don't agree with (and community seems to be with me) is repo per container | 13:39 |
| inc0 | but I'm sure we can figure something out.. | 13:40 |
| *** Mr_Broke_ has quit IRC | 13:41 | |
| dims | inc0 : where/when is the midcycle | 13:41 |
| inc0 | 11-12 Jul Durham | 13:41 |
| inc0 | NC | 13:41 |
| *** Mr_Broken has joined #openstack-kolla | 13:42 | |
| sdake | https://wiki.openstack.org/wiki/Sprints/KollaNewtonSprint | 13:42 |
| sdake | 12-13 ctually | 13:42 |
| inc0 | ehh...datetime is hard | 13:42 |
| sdake | dims at ansible HQ | 13:42 |
| mewald1 | Mr_Broken: Have you seen that happening or do we assume I found a bug? | 13:43 |
| rhallisey | did anyone want to share a room? :) | 13:44 |
| sdake | just a heads up - team meeting in 2 hours | 13:45 |
| sdake | normally this would be the afternon (for the us folks) meeting | 13:45 |
| sdake | rhallisey i can if your strapped on budget - but i snore | 13:46 |
| sdake | and it can't be stopped | 13:46 |
| *** Mr_Broken has quit IRC | 13:46 | |
| dims | inc0 : sdake : ah thanks. sergey is on his way to get stuff sorted out to travel back to US. will have to check where he will be | 13:47 |
| rhallisey | sdake, ya I may take you up on that offer | 13:47 |
| dims | we can definitely do a webex/remote during the event inc0 - i will check up on possibilities of me coming, but not sure if i can make too much of a difference organizationally | 13:48 |
| inc0 | dims, we'll make sure that it will be accessible | 13:49 |
| dims | thanks inc0 | 13:49 |
| sdake | its a shame we can't do full on vr yet | 13:54 |
| sbezverk | inc0 do you have a minute? | 13:54 |
| sdake | that would make this midcycle stuf alot cheaper | 13:54 |
| sdake | like in captain america | 13:55 |
| * sdake wants that tech | 13:55 | |
| *** daneyon has joined #openstack-kolla | 13:57 | |
| sdake | dims webex is already setup | 13:57 |
| sdake | dims it only runs on mac and windows tho | 13:57 |
| inc0 | sbezverk, I'm here | 13:58 |
| dims | y no worries sdake : we are used to webex | 13:58 |
| sdake | i used to hate webex at red hat | 13:58 |
| sdake | mainlly becasue it didn't work | 13:58 |
| sbezverk | inc0 do you know if haproxy con work in "on a stick" mode? | 13:58 |
| sdake | and people would ninja webex during meetings | 13:58 |
| sdake | where i'd have to find smoe alternative mechanism to connect | 13:58 |
| inc0 | on a stick? | 13:59 |
| sdake | what is on a stick | 13:59 |
| sbezverk | inc0 when the traffic in and out of the same interface | 13:59 |
| mewald1 | I noticed that Kolla creates Keystone services and endpoints for me. Does it create stuff like cinder volume types etc, too? | 14:00 |
| sdake | mewald1 no only keystone creds | 14:00 |
| sbezverk | with routers it is called on the stick I thought the same term used here as well | 14:00 |
| mewald1 | sdake: ok thx | 14:00 |
| sdake | sbezverk i think haproxy puts data in and out same interface | 14:00 |
| sdake | we do not use it as a firewall for example | 14:01 |
| *** daneyon has quit IRC | 14:01 | |
| inc0 | sbezverk, I have no idea...I thought it just work;) | 14:01 |
| sdake | mewald1 our demo script also creatoes lavors for us | 14:01 |
| sdake | sbezverk the answer is yes - thta is the only way it operates | 14:01 |
| sbezverk | inc0 sdake basically I see request comes in but it does not get proxied to the backend server and I use haproxy vm IP address | 14:01 |
| sdake | sbezverk examples would help :) | 14:02 |
| sbezverk | I am trying to build loadbalancer vm I could use with multinode k8s cluster | 14:02 |
| sdake | do you mean vm or container | 14:03 |
| sbezverk | so kubelet would use a single api server ip | 14:03 |
| sbezverk | sdake vm | 14:03 |
| sbezverk | sdake here is my haproxy config http://paste.openstack.org/show/521176/ | 14:04 |
| kfox1111 | sdake: fyi, I just had to manually tweak some sql migration code to work around a bug. again. :) | 14:07 |
| kfox1111 | so manually being able to go through the upgrade workflow is absolutely nessisary still. :/ | 14:08 |
| mewald1 | Is it possible to configure separate admin, public and internal URLs with Kolla? The services that kind of needs it would be keystone where port 5000 and 35357 represent the public and the admin endpoint of the service. | 14:09 |
| *** williamcaban has joined #openstack-kolla | 14:16 | |
| sdake | kfox1111 coudl yo ugo into more detail | 14:22 |
| *** signed8bit has joined #openstack-kolla | 14:22 | |
| sdake | kfox1111 are you upgrading a stable branch? | 14:22 |
| sdake | kfox1111 fwiw we dont implement upgrades of master branch | 14:22 |
| sdake | it sort of works, but the upstreaams damage their sql code all the time | 14:22 |
| sdake | so its not viable imo | 14:22 |
| sdake | sbezverk i am not a haproxy wizard, but I think liine 33 tells api-server to send and receive off of 10.57.120.254 | 14:23 |
| sdake | sbezverk i would recommend a container though, dealing with vm networking would be a pita | 14:24 |
| sbezverk | sdake nv, I got it working for api and for etcd :-) so the config could be simplified but redundancy stays the same | 14:24 |
| Mech422__ | mewald1: yes, you can configure seperate nets - look at the examples in the inventory file showing how to pass api_interface, ceph_interface, etc | 14:31 |
| Mech422__ | sbezverk: Are you sure haproxy is the problem? Firewalls often dont like stick/hairpin mode ? | 14:32 |
| *** ppowell has quit IRC | 14:34 | |
| sbezverk | Mech422__ that is why I was asking for "on the stick" :-) but it seems haproxy works fine.. | 14:35 |
| *** ppowell has joined #openstack-kolla | 14:36 | |
| Mech422__ | sbezverk: yeah - oh cool - so you routing between like a plain libvirsh VM and the docker bridge? | 14:36 |
| *** williamcaban has quit IRC | 14:46 | |
| *** david-lyle_ has quit IRC | 14:47 | |
| sbezverk | Mech422__ no it is not docker env. I brought up 5 VMs which is 5 k8s cluster nodes and I needed "external" loadbalancer | 14:47 |
| Mech422__ | sbezverk: oh...I always seem to get caught with weird network stuff, because I use OVS for basic 'host' networking, and only pxe runs over native/untagged - all 'real' traffic goes thru ovs trunks | 14:50 |
| Mech422__ | sbezverk: so I usually end up with patches and veths and bridges ! Oh My! :-) | 14:50 |
| *** vhosakot has joined #openstack-kolla | 14:54 | |
| sbezverk | Mech422__ I try as much as I can to use pcipassthrough so I do not need to deal with software bridges | 14:55 |
| Mech422__ | sbezverk: have you tried the sr-iov stuff for pass thru ? | 14:56 |
| sbezverk | Mech422__ yep use it all the time for our virtual router | 15:01 |
| Mech422__ | sbezverk: oh nice - you like it then ? | 15:01 |
| Mech422__ | sbezverk: I haven't tried it yet - or passthru at all. Kinda hoping I get a reason to in the future... | 15:03 |
| *** sacharya has joined #openstack-kolla | 15:08 | |
| sbezverk | Mech422__ well it has its own issues, but when it works, the performance it much much better than any software switching things.. | 15:10 |
| *** sdake has quit IRC | 15:10 | |
| *** coolsvap has quit IRC | 15:10 | |
| *** robcresswell_ has joined #openstack-kolla | 15:10 | |
| *** robcresswell has quit IRC | 15:10 | |
| *** robcresswell_ is now known as robcresswell | 15:10 | |
| *** coolsvap has joined #openstack-kolla | 15:11 | |
| *** huikang has joined #openstack-kolla | 15:11 | |
| *** sdake_ has joined #openstack-kolla | 15:11 | |
| coolsvap | sdake_, sorry i was away when all the discussion took place on infra | 15:15 |
| *** b_bezak has quit IRC | 15:16 | |
| coolsvap | sbezverk, hey the most updated rpm i got is 1.2.0 | 15:17 |
| *** sdake has joined #openstack-kolla | 15:18 | |
| *** sdake_ has quit IRC | 15:19 | |
| sdake | dims you about | 15:21 |
| dims | pong | 15:21 |
| sdake | dims - dont know if you saw my review comment in the multirepo spec for kolla | 15:22 |
| sdake | dims i think there is some confusion around kolla's mission | 15:22 |
| Mech422__ | sbezverk: oh - I wondered how the performance was - so pretty close to native then ? | 15:22 |
| sdake | dims if you haven't had a chance to read it yet, please read it | 15:22 |
| dims | sdake : going to look | 15:22 |
| dims | sdake : looks like you and inc0 have countered the perception sufficiently | 15:25 |
| sdake | dims i think mechnically it would be difficult to manage | 15:25 |
| sdake | rather logistically | 15:26 |
| sdake | dims but before we geet into logistics i'd like to understand the need for serpate repositories - haven't seen that yet | 15:26 |
| inc0 | and I'm concerned about dependency management - we put deps between containers | 15:26 |
| *** williamcaban has joined #openstack-kolla | 15:26 | |
| sdake | inc0 those are those pesky logistics i was tlaking about | 15:26 |
| inc0 | so we might end up in cross-repo dependency hell for multi-container changes | 15:26 |
| *** mbound has quit IRC | 15:27 | |
| inc0 | yeah, I agree sdake, I asked what problem we're solving | 15:27 |
| inc0 | before we jump to conclusions | 15:27 |
| sdake | the elemental spec answered that pretty welll | 15:27 |
| sdake | this spec, less so | 15:27 |
| inc0 | well, elemental spec was about customizations, and we got this one figured out | 15:28 |
| sdake | right | 15:28 |
| inc0 | if there is more issues, I'd like to know them | 15:28 |
| sdake | i would abbandon but i want to incorporate some of that into docs | 15:28 |
| *** v1k0d3n has joined #openstack-kolla | 15:30 | |
| dims | sdake : inc0 : i think the idea is that just like we put out pypi libs with version #'s and then figure out ranges that work togther, eventually we'll want do to the same with containers. | 15:30 |
| inc0 | dims, well, with single repo we don't really need to right? | 15:30 |
| dims | and if a specific repo changes, then run CI for specific things that are implemented in that repo | 15:30 |
| inc0 | there are solutions to this problem, but I'm reluctant in making it a problem at all | 15:31 |
| sdake | dims yes this is per image building - i know exactly what your after | 15:31 |
| sdake | dims this can be done in one repo versus many | 15:31 |
| inc0 | dims, so, we'll have mechanism for that | 15:31 |
| dims | sdake : am trying to voice sergey's concern (not speaking for myself) | 15:31 |
| inc0 | you'll be able to go all batshit crazy about your custom containers | 15:31 |
| sdake | dims yup i know exactly where he is coming from | 15:32 |
| *** caowei has quit IRC | 15:32 | |
| inc0 | including properitary soft inside | 15:32 |
| sdake | inc0 this is different then that | 15:32 |
| sdake | this is an optimization to only build continers tht change | 15:32 |
| inc0 | sdake, we can try to look at build.py to incorporate that | 15:32 |
| sdake | if the rationale is to only build containers that change, there are ways that already exist to solve that problem | 15:32 |
| sdake | although not hiding out in any pubic repos unfortunately | 15:32 |
| dims | right, bump version in repo, put out a new image, run tests against specifically things that changed in that repo | 15:33 |
| inc0 | well, truth be told, you won't be able to do this in docker like that | 15:33 |
| sdake | dims all this is about ci/cd optimization | 15:33 |
| sdake | inc0 we do at cisco ;) | 15:33 |
| inc0 | if base images change, you'll have problem | 15:33 |
| dims | sdake : precisely | 15:33 |
| inc0 | sdake, multiple base images | 15:33 |
| sdake | inc0 in some cases yes | 15:34 |
| dims | a stack of base images, possibly ability to inject additional base images based on things needed at say a customer site | 15:34 |
| dims | s/images/layers | 15:34 |
| *** v1k0d3n has quit IRC | 15:35 | |
| sdake | here is something i wrote: http://paste.openstack.org/show/521194/ | 15:35 |
| dims | sdake : so in the end a customer environment, they should be able to mint what they want and pick and choose from things | 15:35 |
| sdake | that prints out all parent and children containers from a list of changed containers that could be genreated by gerrit | 15:36 |
| dims | then run CI on stuff and deploy | 15:36 |
| dims | separate repos forces you to think that way | 15:36 |
| sdake | this input could be fed right back into build.py | 15:36 |
| * dims still wearing sergey's hat | 15:37 | |
| sdake | dims fwiw Ithink a whole lot of repos for work that is related is an anti-pattern | 15:37 |
| sdake | roger | 15:37 |
| inc0 | dims, to do that well, you can use tags | 15:38 |
| inc0 | you should use tags | 15:38 |
| inc0 | like centos-source-base:my-own-tag | 15:38 |
| sdake | what I am saying is and I'd like you to confirm if you understand, there are ways that what sergey wants (to optimize cic/dd run time) can be done without a bajillion repos | 15:39 |
| inc0 | yeah, what sdake said;) | 15:39 |
| dims | sdake : yes, i confirm that i understand :) | 15:39 |
| sdake | what sergey is saying I think is he wants to force that way of thinking | 15:39 |
| dims | yep | 15:39 |
| sdake | with multiple repos | 15:39 |
| sdake | I understand that part of things, i think there is more to it | 15:40 |
| sdake | but to be entirely blunt, I think separating the repos has the potential to kill the project | 15:40 |
| sdake | and that risk is not worth the gain of forcing dependency modeling in the repos themselves | 15:40 |
| sdake | fwiw, when we did the mission statement, i foguht for universal container idea, but it was rejected | 15:41 |
| sdake | not for multiple repos | 15:41 |
| sdake | but or othres to be able to reuse our containers - and explicitly call it out in our mission statement | 15:41 |
| dims | ack. we can always provide feedback that of the things in that spec, that one is not-negotiable | 15:42 |
| sdake | everything is negotaiable | 15:42 |
| sdake | but the multiple repos is dangerous to me personally | 15:42 |
| wirehead_ | FWIW, I doubt we’re going to get the ‘universal container’ regardless. The best we’re going to do is some refactoring of our existing setup to split the kolla-ansible logic from the feeding of the jinja template itself. | 15:42 |
| sdake | seems dangerously | 15:43 |
| sdake | seems dangerous that is | 15:43 |
| dims | wirehead_ : there's a fuel spec and they will try to do it themselves | 15:43 |
| sdake | we already have an abi and well established precedent of not making it ansible specific | 15:44 |
| sdake | we wnat others to use our containers, so there is one community maintaining them | 15:44 |
| dims | sdake : ++ (speaking for myself) | 15:45 |
| *** daneyon has joined #openstack-kolla | 15:45 | |
| sdake | dims there is another frankly in here too | 15:46 |
| sdake | we have a whole lot of stuff on our plates that could take onwards of two+ years to sort out | 15:46 |
| wirehead_ | I’ve read the spec, dims. | 15:46 |
| sdake | fitting this in as a priority isn't super high in my mind | 15:46 |
| dims | sdake : yep. | 15:46 |
| sdake | the making the containers universal yes - making them separate repos no | 15:46 |
| wirehead_ | That’s less a “I doubt they’ll try it” and more a “I’m not sure you’ll get any farther than some kolla refactoring would get you" | 15:47 |
| *** daneyon_ has joined #openstack-kolla | 15:48 | |
| sdake | we have the abi we have today mostly because of tripleo integration | 15:48 |
| dims | wirehead_ : i dunno how far they would get or if they will succeed, just that fuel-ccp-* repos will be started and there will be folks working on it there | 15:48 |
| sdake | if we had no abi, things woudl be simpler - then we would just deploy via ansible | 15:48 |
| *** daneyon has quit IRC | 15:49 | |
| dims | sdake : as long as there are no ansible specific bits in the images was the concern i think | 15:50 |
| sdake | there are non today | 15:50 |
| sdake | none | 15:50 |
| *** harlowja_at_home has joined #openstack-kolla | 15:50 | |
| inc0 | dims, we were very careful not to put ansible specific stuff | 15:51 |
| *** huikang has quit IRC | 15:51 | |
| dims | so then it becomes an argument for a repo split | 15:52 |
| inc0 | yeah, and I'll argue that any customization or whatever can be done without repo split | 15:52 |
| inc0 | in fact, it will be easier as it won't require lots of merging/rebasing of forked contaniers | 15:52 |
| dims | :) | 15:53 |
| inc0 | ...once we're done implementing it;) | 15:53 |
| Mech422__ | fwiw, I think the real value add in kolla is not the containers themselves, but the knowledge of how to wire them all together, bootstrap, and make them happy.. | 15:53 |
| inc0 | and building only changed containers....well, we can filter out builds in our build.py now | 15:53 |
| inc0 | Mech422__, it's both, creating containers took a lot of time too | 15:53 |
| sdake | Mech422__ figuring that stuff out adds value in the containers themselves | 15:54 |
| inc0 | all the idiosyncrasies between OSes, user and dir management | 15:54 |
| Mech422__ | inc0: I can imagine....but I'd also think it will become more common for various projects to produce containers themselves ? | 15:54 |
| dims | so it looks like the greatest push back is on "One Repository per OpenStack component" and others are yes, we should do it, we are already trying to and we need help to implement those | 15:54 |
| *** ppowell has quit IRC | 15:54 | |
| sdake | dims i dont know if that is totaally accurate | 15:55 |
| inc0 | Mech422__, right now we have well established standards that will make adding new containers easier | 15:55 |
| sdake | dims the problem is my brain sees split repos | 15:55 |
| dims | taking a temperature sdake | 15:55 |
| sdake | and i have a hard time procesing the rest of the spec | 15:55 |
| Mech422__ | inc0: the 'idiosyncrasies' I think are part of the 'wire them together' - thats 'lore' that gets 'lost' quite often | 15:55 |
| inc0 | yeah you can think of it like that | 15:55 |
| inc0 | dims, so I had some issues with bootstrap outside container, but that's mostly tech stuff | 15:56 |
| inc0 | how we actually do it | 15:56 |
| sdake | yes i dont like that idea either | 15:56 |
| sdake | but i dont like our current bootstrapping | 15:56 |
| dims | sdake | 15:56 |
| dims | "Separate containers definitions and deployment tools" << split ansible into separate repo - we have talked about it in the past, we are already not using ansible in images/definitions so do-able. | 15:56 |
| dims | "One Repository per OpenStack component" << Not worth the effort, probably counter productive to health | 15:56 |
| sdake | what I would prefer is one container bootstrap the whole damn thing | 15:56 |
| Mech422__ | dims: personally, 1 repo makes it a bit easier for me to get started with...but to be honest, learning the 'lore' and how everythign fits together is the bigger issue for me | 15:56 |
| *** Serlex has quit IRC | 15:56 | |
| inc0 | I like the idea, I just say we might not be able to without compromising part of security | 15:57 |
| dims | "nified paths and endpoints" << no push back in comments | 15:57 |
| dims | "CI pipeline" << no push back in comments | 15:57 |
| sdake | we have unified paths | 15:57 |
| inc0 | yeah kinda, depends what you mean by that | 15:57 |
| sdake | every container is standardized | 15:57 |
| rhallisey | inc0, I think the split repo argument can change with kube though | 15:57 |
| inc0 | kolla-start.sh is pretty standarized | 15:58 |
| inc0 | rhallisey, what do you mean? | 15:58 |
| inc0 | btw meeting | 15:58 |
| sdake | inc0 he means shaving off ansible into a separate repo | 15:58 |
| sdake | ya meeting tie | 15:58 |
| rhallisey | well you talking about multiple deployment tools in a repo | 15:58 |
| rhallisey | both consuming the containers | 15:58 |
| inc0 | rhallisey, not really, we were taling about multiple repos for containers | 15:58 |
| dims | i gotta run. will rope sergey to the next meeting and give him logs for this one (too bad he is travelling) | 15:58 |
| sdake | later dims | 15:59 |
| inc0 | so heat dockerfiles will be in separate repo than nova | 15:59 |
| inc0 | cya dims | 15:59 |
| rhallisey | dims, bye! | 15:59 |
| rhallisey | inc0, no I mean kolla-anisble split from kolla-kube | 15:59 |
| rhallisey | not for splitting containers | 15:59 |
| inc0 | yeah, we talked about splitting containers | 15:59 |
| inc0 | as for kolla-kube vs kolla-ansible, we agreed to shelve this discussion till we have more concrete evidence;) | 16:00 |
| inc0 | we didn't make a call either way, just to say "we'll examine later" | 16:00 |
| *** britthouser has joined #openstack-kolla | 16:00 | |
| *** number80 has joined #openstack-kolla | 16:01 | |
| number80 | hello, I'm currently checking global-reqs, and it appeared that kolla is the only project having it in its requirements | 16:02 |
| *** dmk0202 has quit IRC | 16:02 | |
| number80 | but it looks like it's not even used, can someone confirm that? | 16:02 |
| number80 | ok found it, it's still used | 16:07 |
| *** harlowja_at_home has quit IRC | 16:12 | |
| *** mbound has joined #openstack-kolla | 16:27 | |
| *** vhosakot_ has joined #openstack-kolla | 16:28 | |
| *** vhosakot has quit IRC | 16:30 | |
| *** JoseMello has joined #openstack-kolla | 16:31 | |
| *** mbound has quit IRC | 16:32 | |
| *** athomas has quit IRC | 16:38 | |
| *** huikang has joined #openstack-kolla | 16:46 | |
| *** sacharya_ has joined #openstack-kolla | 16:47 | |
| *** sacharya has quit IRC | 16:50 | |
| openstackgerrit | Merged openstack/kolla: Spec: Deploy Kolla images using Kubernetes https://review.openstack.org/304182 | 16:50 |
| *** ssurana has joined #openstack-kolla | 16:51 | |
| *** shardy has quit IRC | 16:56 | |
| rhallisey | wirehead_, so what do you have demo able? | 16:59 |
| rhallisey | keystone + mariadb + memcached ? | 17:00 |
| wirehead_ | So, fired up MariaDB, MemcacheD, Keystone, and Horizon.. all of the services that have net=host removed. | 17:00 |
| rhallisey | v nice | 17:00 |
| wirehead_ | Horizon was able to find Keystone and interact with the API to add users, etc. | 17:00 |
| wirehead_ | I can kill Keystone and the RC will restart it and everything’s fine. | 17:00 |
| rhallisey | wirehead_, dude with anisble the deployment experience will be 10x better | 17:00 |
| rhallisey | can't wait | 17:01 |
| wirehead_ | And kfox1111 was able to just deploy Horizon without the rest of the stack. | 17:01 |
| inc0 | wirehead_, did docker networkign stop to suck big time? | 17:01 |
| rhallisey | wirehead_, that's awesome | 17:01 |
| *** mgoddard has quit IRC | 17:01 | |
| wirehead_ | inc0: So, correctly configured kube 1.2+ bypasses the docker proxy. | 17:01 |
| rhallisey | wirehead_, hmmm, that's an interesting idea | 17:01 |
| wirehead_ | inc0: I haven’t run any bench tests yet. | 17:01 |
| inc0 | yay for correctly configured kube | 17:01 |
| rhallisey | we should decouple our plays enough so you can deploy just one managed by kube | 17:01 |
| wirehead_ | Basically, iptables backed by etd. | 17:01 |
| wirehead_ | etcd. | 17:02 |
| * rhallisey gets lunch | 17:02 | |
| wirehead_ | Especially for compute nodes. Compute nodes via kolla-ansible with the rest in kolla-kubernetes. | 17:02 |
| inc0 | wirehead_, that and networking ndoes | 17:03 |
| inc0 | as you don't want l3 agents floating around as well | 17:04 |
| *** huikang has quit IRC | 17:06 | |
| *** Jeffrey4l has quit IRC | 17:08 | |
| *** sdake_ has joined #openstack-kolla | 17:09 | |
| *** sdake has quit IRC | 17:10 | |
| *** ppowell has joined #openstack-kolla | 17:23 | |
| *** huikang has joined #openstack-kolla | 17:24 | |
| *** sdake_ has quit IRC | 17:29 | |
| *** coolsvap has quit IRC | 17:30 | |
| *** david-lyle has joined #openstack-kolla | 17:30 | |
| *** daneyon has joined #openstack-kolla | 17:33 | |
| openstackgerrit | David Wang proposed openstack/kolla-kubernetes: Simplify DNS docs, removing need for user decisions https://review.openstack.org/332937 | 17:35 |
| *** daneyon has quit IRC | 17:38 | |
| *** huikang has quit IRC | 17:42 | |
| *** ppowell has quit IRC | 17:43 | |
| *** zhiwei has joined #openstack-kolla | 17:50 | |
| *** ppowell has joined #openstack-kolla | 17:52 | |
| *** bootsha has joined #openstack-kolla | 17:53 | |
| *** zhiwei has quit IRC | 17:54 | |
| *** salv-orlando has quit IRC | 18:02 | |
| *** salv-orlando has joined #openstack-kolla | 18:03 | |
| harlowja | what'd i miss from meeting :-P | 18:03 |
| *** ssurana has quit IRC | 18:05 | |
| *** ppowell has quit IRC | 18:06 | |
| *** ppowell has joined #openstack-kolla | 18:07 | |
| *** salv-orlando has quit IRC | 18:07 | |
| *** ssurana has joined #openstack-kolla | 18:08 | |
| *** huikang has joined #openstack-kolla | 18:09 | |
| Mech422__ | harlowja: apperently - everyone :-P | 18:12 |
| harlowja | lol | 18:12 |
| Mech422__ | harlowja: umm - there was talk about single vs multiple repos | 18:13 |
| Mech422__ | harlowja: and discussion about tracking gate failures so we can map out what needs to be fixed | 18:13 |
| Mech422__ | harlowja: and stuff about refactoring build.py, and converting dockerfiles to the new customization stuff | 18:14 |
| harlowja | ya, i'd like to know if there is any little docs for how the config stuff is working | 18:14 |
| Mech422__ | harlowja: I think that was the highlights... | 18:14 |
| harlowja | https://review.openstack.org/#/c/332590/ (does this one even make sense) | 18:14 |
| patchbot | harlowja: patch 332590 - kolla - Allow for externally managed configuration | 18:14 |
| harlowja | in that i don't want kolla to do all the config stuff, i'm just gonna bindmount stuff in (that puppet has put in place already) | 18:14 |
| harlowja | but i don't know the details on all that config magic sauce | 18:14 |
| Mech422__ | harlowja: there was a link inc0 sent out on how customization works...but I just realized I closed the meeting tab without copying it | 18:14 |
| harlowja | lol | 18:15 |
| harlowja | inc0 send it again :-P | 18:15 |
| Mech422__ | harlowja: here's the etherpad for tracking gate problems: https://etherpad.openstack.org/p/kolla-gate-failures | 18:15 |
| harlowja | also the whole config stuff, is anyone thinking of refactoring that whole junk | 18:17 |
| harlowja | the oslo.config stuff that is | 18:17 |
| harlowja | its hella confusing | 18:17 |
| harlowja | (unnessasarly confusing imho) | 18:17 |
| harlowja | * unnecessarily | 18:17 |
| Mech422__ | harlowja: have you seen http://augeas.net | 18:18 |
| harlowja | aka, why is it just something like $ kolla-build <config-file> and thats it :-P | 18:18 |
| harlowja | vs u can do some of this via CLI options, u can do some of it via something else .... | 18:18 |
| harlowja | i have seen augeas, i haven't used it much | 18:18 |
| harlowja | i was going to at one time use augeas in cloud-init | 18:18 |
| harlowja | for doing configuration things | 18:19 |
| Mech422__ | harlowja: I'm always surprised there's no pre-existing lenses for openstack for it | 18:19 |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Adding debugging documentation for Keystone https://review.openstack.org/330788 | 18:19 |
| Mech422__ | harlowja: it makes config file manipulation dead easy, and very consistent | 18:19 |
| harlowja | ya, i forget the reason i never used it in cloud-init | 18:19 |
| *** sacharya_ has quit IRC | 18:20 | |
| Mech422__ | harlowja: right now, I 'pre-cook' my globals file with jinja... | 18:21 |
| *** sacharya has joined #openstack-kolla | 18:21 | |
| Mech422__ | harlowja: so I have a copy of globals with all the '{{' changed to '!!' and all the "}}" changed to "@@" | 18:21 |
| harlowja | ya, i need to figure out the X different ways kolla has config, lol | 18:21 |
| Mech422__ | harlowja: so I can add in my custom jinja, run it thru the template module to 'pre-cook' it... | 18:22 |
| harlowja | right, that's for container configuration right? | 18:22 |
| harlowja | or is that for kolla config? | 18:22 |
| Mech422__ | harlowja: then sed the '!!' back to "{{", etc and hand it back to kolla to do ITS jinja processing | 18:22 |
| harlowja | or is it for oslo.config | 18:22 |
| harlowja | or is it for some other config, lol | 18:22 |
| Mech422__ | ah well..time for bed... | 18:23 |
| harlowja | lol | 18:23 |
| Mech422__ | harlowja: take care! | 18:23 |
| harlowja | lata | 18:23 |
| harlowja | let me see what inc0 can find for configs usage in kolla, it'd be real nice to not have 10000 different formats for config into kolla (or into containers or ...) | 18:26 |
| harlowja | in fact i can share how i did this in anvil | 18:27 |
| harlowja | where we sorta settled on 3 different types of config (all yaml) | 18:27 |
| *** daneyon_ has quit IRC | 18:27 | |
| *** daneyon has joined #openstack-kolla | 18:28 | |
| *** bootsha has quit IRC | 18:31 | |
| *** gfidente is now known as gfidente|afk | 18:34 | |
| wirehead_ | see, to the universal container and architecture of openstack thing, I really wish we could inflict upon the rest of the community a single container-friendly config standard. :) | 18:36 |
| *** mbound has joined #openstack-kolla | 18:39 | |
| *** matrohon has joined #openstack-kolla | 18:43 | |
| harlowja | wirehead_ so that's 1 of the configurations for containers | 18:44 |
| harlowja | the other one that i'm also confused about is the kolla config itself, lol | 18:45 |
| harlowja | (for kolla) | 18:45 |
| harlowja | it seems ummm, somewhat all over the place | 18:45 |
| harlowja | like there is globals.yml (which i guess is for ansible?) | 18:45 |
| harlowja | then there is some CLI options | 18:45 |
| harlowja | then there is some config file options | 18:45 |
| harlowja | some of those are dynamic, some aren't | 18:45 |
| harlowja | which is hella-confusing | 18:45 |
| *** sdake has joined #openstack-kolla | 18:47 | |
| wirehead_ | yeah | 18:47 |
| harlowja | so ya, idk maybe people just like the pain of that | 18:47 |
| harlowja | but i sure don't, lol | 18:47 |
| harlowja | <rage pain> | 18:48 |
| harlowja | lol | 18:48 |
| wirehead_ | It’s like the config pain of an expression that requires you to have the C operator prescedence table printed out. | 18:48 |
| harlowja | ya | 18:48 |
| harlowja | in all honesty, i'm not even sure why oslo.config is being used in kolla | 18:48 |
| harlowja | seems sorta useless and just makes shit more confusing | 18:48 |
| harlowja | vs just using the basic configparser and having all things go into that | 18:49 |
| harlowja | then its just kolla-build <kolla-config> | 18:49 |
| harlowja | and thats it | 18:49 |
| harlowja | or damn, don't use cfgparser and just use yaml | 18:49 |
| harlowja | simple ===== good | 18:50 |
| *** sdake_ has joined #openstack-kolla | 18:55 | |
| sdake_ | harlowja i'm not entirely sure why we are using oslo.cfg either | 18:55 |
| harlowja | ya, that whole stuff confuses me, lol | 18:56 |
| sdake_ | the reason we re not using yaml is yaml is a dependency | 18:56 |
| sdake_ | and the target nodes should be dep free minus docker-py and docker | 18:56 |
| harlowja | nah, i meant for kolla itself, not the containers | 18:56 |
| sdake_ | but in this case it would run on the deployment host | 18:56 |
| harlowja | the config in the containers, slightly different imho, i don't quite yet understand that part so far | 18:57 |
| harlowja | (understand the container configuration stuffs) | 18:57 |
| sdake_ | harlowja its a little hard to understand bbecause it was complicated by drop root | 18:57 |
| harlowja | whose drop root? | 18:57 |
| sdake_ | i can wlak you through it if you like | 18:57 |
| sdake_ | the drop root we added to containers | 18:57 |
| harlowja | ah | 18:57 |
| harlowja | i thought drop root was a IRC user, ha | 18:58 |
| harlowja | damn u drop root | 18:58 |
| harlowja | lol | 18:58 |
| *** sdake has quit IRC | 18:58 | |
| harlowja | sdake_ sure at a time, maybe not yet, gotta finish writing up some specs (internally) | 18:58 |
| *** achanda has joined #openstack-kolla | 18:58 | |
| sdake_ | unfortunatley i have a meeting in 2 mins | 18:58 |
| sdake_ | it would tke 5 minutes to go through it | 18:58 |
| harlowja | k | 18:58 |
| sdake_ | maybe 10 or 15 | 18:59 |
| harlowja | u should write that stuff down and put it in a 'kolla config and wtf are the X configs doing/for/why' doc :-P | 18:59 |
| sdake_ | line numbers change in code | 18:59 |
| harlowja | lol | 19:00 |
| harlowja | that's ok, we'll fix those line numbers changing soon enough | 19:00 |
| harlowja | no more code | 19:00 |
| harlowja | lol | 19:00 |
| sdake_ | there has to bbe code to parse the json | 19:00 |
| sdake_ | the json is the abi | 19:00 |
| sdake_ | its super obvious once someone explains it to you | 19:01 |
| harlowja | k | 19:01 |
| sdake_ | but not from looking at the code base randomly | 19:01 |
| harlowja | ya | 19:01 |
| harlowja | i tried that approach, and am like <what the heck> | 19:01 |
| harlowja | lol | 19:01 |
| wirehead_ | Well, might be useful to at least write down the expected protocol for what you would config where. | 19:01 |
| *** achanda has quit IRC | 19:02 | |
| wirehead_ | For example: the discussion related to change 327925 | 19:02 |
| inc0 | wow that's a lot of logs to read through | 19:03 |
| *** bootsha has joined #openstack-kolla | 19:04 | |
| inc0 | harlowja, ok ad configs | 19:05 |
| inc0 | best idea I have right now is to decouple config generation from deployment | 19:05 |
| inc0 | so you generate configs, go to vim, edit whatever the hell you want | 19:05 |
| inc0 | and then actual deploy will use it | 19:05 |
| *** huikang has quit IRC | 19:12 | |
| *** b_bezak has joined #openstack-kolla | 19:14 | |
| *** huikang has joined #openstack-kolla | 19:25 | |
| *** sdake has joined #openstack-kolla | 19:31 | |
| *** sdake_ has quit IRC | 19:33 | |
| harlowja | right, inc0 so thats for the container configs, just want to make sure we talking about which configs, lol | 19:35 |
| harlowja | the stuff that is a mix of yaml, and json and ??? | 19:35 |
| *** fragatina has joined #openstack-kolla | 19:41 | |
| *** fragatina has quit IRC | 19:42 | |
| *** fragatina has joined #openstack-kolla | 19:42 | |
| *** thumpba has joined #openstack-kolla | 19:43 | |
| *** jogam82 has joined #openstack-kolla | 19:55 | |
| jogam82 | hi kolla team anyone out there? | 19:55 |
| inc0 | out where? | 19:55 |
| inc0 | we're here, not sure about "there" | 19:56 |
| *** bootsha has quit IRC | 19:58 | |
| jogam82 | haha... just build/deployed the latest Kolla from github on a 4 node env including cinder via loopback blockdevice, now I can create a volume via 'cinder create 1 --display-name testVolume' which results in status eror | 19:58 |
| inc0 | jogam82, globals.yml plz | 19:59 |
| jogam82 | oh hosts are Ubuntu 16.04 LTS and I build latest | 19:59 |
| inc0 | did you enable cinder and select driver for it? | 20:00 |
| *** sdake_ has joined #openstack-kolla | 20:01 | |
| jogam82 | @inc0: here you go --> http://paste.openstack.org/show/521274/ | 20:02 |
| inc0 | so, could you please paste me logs from cinder? | 20:02 |
| jogam82 | I assume so, followed the kolla openstack cinder intro, which is missing the small part of adding enable_iscsi: "yes" | 20:03 |
| inc0 | you can find all service logs in /var/lib/docker/volumes/log_data | 20:03 |
| inc0 | good catch jogam82 ! wanna make a patch for it?:) | 20:03 |
| inc0 | you'll be a contributor to kolla! | 20:03 |
| *** salv-orlando has joined #openstack-kolla | 20:04 | |
| *** sdake has quit IRC | 20:04 | |
| jogam82 | no such folder on either control or storage node: sudo ls /var/lib/docker/volumes | 20:05 |
| jogam82 | returns cinder heka heka_socket iscsi_info kolla_logs metadata.db | 20:05 |
| inc0 | kolla_logs* | 20:05 |
| inc0 | my mistake | 20:05 |
| jogam82 | whohooo that would be a great contribution I suppose... | 20:05 |
| inc0 | it would be a contribution, and a valuable one | 20:06 |
| jogam82 | is there a definite guide to make contributions or just fork/branch/**FIX**/commit and notify someone? | 20:06 |
| inc0 | jogam82, http://docs.openstack.org/infra/manual/developers.html | 20:07 |
| inc0 | we have proper process for that | 20:07 |
| inc0 | I encourage you to try and contribute, good karma;) | 20:08 |
| inc0 | but take a look at logs, maybe they will give you some insight of what went wrong | 20:08 |
| *** vhosakot_ has quit IRC | 20:11 | |
| *** ppowell has quit IRC | 20:13 | |
| *** vhosakot has joined #openstack-kolla | 20:14 | |
| *** daneyon_ has joined #openstack-kolla | 20:15 | |
| jogam82 | inc0, took a while here is the /var/lib/docker/volumes/kolla_logs/_data/cinder/cinder-volume.log from the storage node: http://paste.openstack.org/show/521276/ | 20:16 |
| *** bootsha has joined #openstack-kolla | 20:19 | |
| inc0 | hmm | 20:19 |
| inc0 | I don't know:S | 20:19 |
| *** daneyon_ has quit IRC | 20:20 | |
| inc0 | sbezverk, around? | 20:20 |
| inc0 | jogam82, sbezverk was person who wrote iscsi support, so he might help you | 20:21 |
| inc0 | if not, please file a bug, we'll try to reproduce | 20:22 |
| *** salv-orlando has quit IRC | 20:27 | |
| *** vhosakot has quit IRC | 20:28 | |
| *** salv-orlando has joined #openstack-kolla | 20:28 | |
| *** matrohon has quit IRC | 20:28 | |
| *** vhosakot has joined #openstack-kolla | 20:31 | |
| *** sdake_ has quit IRC | 20:34 | |
| *** huikang has quit IRC | 20:35 | |
| *** huikang has joined #openstack-kolla | 20:35 | |
| *** jtriley has quit IRC | 20:36 | |
| jogam82 | inc0, seems odd I can create a virtual volume on the storage node itself via lvcreate no problem, but from the cinder_volume instance its failing... | 20:39 |
| *** fragatina has quit IRC | 20:39 | |
| *** huikang has quit IRC | 20:39 | |
| jogam82 | inc0, I was also wondering why iscsi is required for cinder to use LVM volume group? | 20:40 |
| inc0 | jogam82, well, creation of volume is one, but you need to access it somehow right? | 20:41 |
| *** huikang has joined #openstack-kolla | 20:44 | |
| *** b_bezak has quit IRC | 20:46 | |
| jogam82 | inc0, true enough but neither swift nor ceph use the iscsi approach - but who am I to say anything, just wondering and it took me a while to figure it out... | 20:47 |
| inc0 | ceph is different driver | 20:47 |
| inc0 | in fact, I use ceph driver instead of lvm+iscsi | 20:47 |
| inc0 | and true, it's different approach, totally different arch | 20:48 |
| *** matrohon has joined #openstack-kolla | 20:52 | |
| inc0 | jogam82, btw you are aware that we have ceph in kolla right? | 20:52 |
| inc0 | and in fact it's better tested than iscsi..which is counterintuitive, but there are reasons for it | 20:53 |
| *** thumpba has quit IRC | 20:55 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Glance works without host network https://review.openstack.org/333022 | 20:57 |
| *** huikang has quit IRC | 20:58 | |
| jogam82 | inc0, I think the problem is that kolla assumes that build and execution is as root user, is that possible? | 21:00 |
| inc0 | it might | 21:00 |
| inc0 | as rootwrap will require root priviledges | 21:00 |
| jogam82 | inc0, on my storage node I can create volumes via sudo lvcreate ... | 21:00 |
| inc0 | but that's cinder issue, | 21:00 |
| inc0 | well, in container we dont run sudo | 21:01 |
| inc0 | so it will get privs of user who runs docker I guess | 21:01 |
| inc0 | not sure tho | 21:01 |
| jogam82 | but from the cinder_volume instance it executes the rootwrap with sudo, but the command lvcreate itself does not have the sudo prefix | 21:01 |
| jogam82 | I might just 'enhance' the lvm command in the container | 21:02 |
| *** matrohon has quit IRC | 21:06 | |
| *** rhallisey has quit IRC | 21:07 | |
| vhosakot | join #docker | 21:09 |
| vhosakot | oops... sorry :) | 21:09 |
| vhosakot | typo :) | 21:09 |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla: Horizon interface address and memcached override https://review.openstack.org/330828 | 21:09 |
| jogam82 | inc0, is each container pulling openstack released packages into a venv that I can obviously not edit? is there a 'configuration layer' that can be used to add 'minor' changes? | 21:10 |
| Daviey | ayoung: Hey. I read your blog on using kolla for keystone work recently... I wondered, did you set it up to do LDAP/AD (and bonus points for kerberos)? | 21:15 |
| ayoung | Daviey, I did not. Are you interested in help along those lines? | 21:15 |
| Daviey | ayoung: Yeah. :) | 21:16 |
| ayoung | There are some interesting things going on with Kerberos and containers, but I need to get smart on them before I can talk about them. the #freeipa team is a little more current with them due to the need to integrate with openshift | 21:16 |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla: Horizon interface address and memcached override https://review.openstack.org/330828 | 21:16 |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla: Fix interface address for memcached atop kubernetes https://review.openstack.org/327987 | 21:16 |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla: Glance interface address, memcached override, and glance override. https://review.openstack.org/332481 | 21:16 |
| ayoung | Daviey, my guess is that you make each container into its own host/service, so that service tickets will work | 21:17 |
| Daviey | ayoung: Yeah. I guess to start with, just having LDAP/AD integration would be a good start. :) | 21:18 |
| ayoung | Daviey, so that should not be different from any other Keystone LDAP integration | 21:18 |
| Daviey | right | 21:18 |
| ayoung | Daviey, http://adam.younglogic.com/2015/02/adding-an-ldap-backed-domain-to-a-packstack-install/ | 21:19 |
| ayoung | Daviey, the only issue is that, since it is a domain specific config, you need to restart the Keystone server | 21:19 |
| Daviey | ayoung: I doubt i've missed one of your blog posts. Always a good read. | 21:20 |
| ayoung | Thanks | 21:20 |
| ayoung | Daviey, so the guy driving Kerberos Docker integration sits right behind me on the odd day that we both decide to go into the office. | 21:21 |
| ayoung | Nalid D | 21:21 |
| ayoung | Dahyabhai | 21:21 |
| * ayoung checks to see if he spelled it right first try | 21:21 | |
| Daviey | hah. Useful to have people like that nearby | 21:22 |
| ayoung | woot | 21:22 |
| ayoung | he's awesome. Wrote certmonger | 21:22 |
| ayoung | he's on the RH docker team now, but has loads of Crypto based experience | 21:22 |
| Daviey | Looking forward to seeing what comes of that | 21:23 |
| ayoung | Daviey, ask me again in #freeipa | 21:23 |
| *** sdake has joined #openstack-kolla | 21:28 | |
| *** bootsha has quit IRC | 21:34 | |
| *** sdake_ has joined #openstack-kolla | 21:38 | |
| *** sdake has quit IRC | 21:39 | |
| *** dwalsh has quit IRC | 21:43 | |
| *** huikang has joined #openstack-kolla | 21:58 | |
| *** huikang has quit IRC | 21:58 | |
| *** signed8bit is now known as signed8bit_Zzz | 22:03 | |
| *** daneyon_ has joined #openstack-kolla | 22:04 | |
| *** inc0 has quit IRC | 22:05 | |
| *** daneyon_ has quit IRC | 22:08 | |
| *** JoseMello has quit IRC | 22:09 | |
| *** ssurana has quit IRC | 22:13 | |
| *** ssurana1 has joined #openstack-kolla | 22:13 | |
| *** sdake has joined #openstack-kolla | 22:21 | |
| *** sdake_ has quit IRC | 22:24 | |
| *** achanda_ has joined #openstack-kolla | 22:29 | |
| *** sdake_ has joined #openstack-kolla | 22:31 | |
| *** sdake has quit IRC | 22:34 | |
| *** daneyon has quit IRC | 22:41 | |
| *** daneyon has joined #openstack-kolla | 22:41 | |
| *** sdake_ has quit IRC | 22:49 | |
| *** asalkeld has joined #openstack-kolla | 22:49 | |
| wirehead_ | kfox1111: OK, I was really wondering about if it was going to be better to separate out the Glance services. | 22:53 |
| wirehead_ | kfox1111: except that I couldn’t find any recommendations in the docs either way. | 22:54 |
| *** vhosakot has quit IRC | 22:54 | |
| *** salv-orl_ has joined #openstack-kolla | 22:58 | |
| *** salv-orlando has quit IRC | 23:00 | |
| *** salv-orl_ has quit IRC | 23:06 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Glance works without host network https://review.openstack.org/333022 | 23:09 |
| *** williamcaban has quit IRC | 23:13 | |
| *** sdake has joined #openstack-kolla | 23:24 | |
| *** sdake has quit IRC | 23:42 | |
| *** sacharya has quit IRC | 23:51 | |
| *** lmiccini has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!