*** Jeffrey4l_ has joined #openstack-kolla | 00:10 | |
*** MarMat_ has joined #openstack-kolla | 00:12 | |
*** weiyu_ has joined #openstack-kolla | 00:14 | |
*** MarMat has quit IRC | 00:16 | |
*** MarMat_ has quit IRC | 00:17 | |
*** rajathagasthya has quit IRC | 00:18 | |
*** mark-casey has quit IRC | 00:18 | |
*** ravig has joined #openstack-kolla | 00:21 | |
*** diogogmt has joined #openstack-kolla | 00:25 | |
*** ccesario___ has quit IRC | 00:26 | |
*** weiyu_ has quit IRC | 00:26 | |
*** jtriley has joined #openstack-kolla | 00:40 | |
*** salv-orlando has joined #openstack-kolla | 00:42 | |
*** ravig has quit IRC | 00:46 | |
*** mbound has joined #openstack-kolla | 00:47 | |
*** daneyon has joined #openstack-kolla | 00:47 | |
*** salv-orlando has quit IRC | 00:48 | |
*** daneyon has quit IRC | 00:52 | |
*** mbound has quit IRC | 00:52 | |
*** unicell1 has quit IRC | 00:55 | |
*** unicell has joined #openstack-kolla | 00:56 | |
*** sdake has joined #openstack-kolla | 01:02 | |
sdake | v1k0d3n still around? | 01:06 |
---|---|---|
sdake | v1k0d3n try to hunt me down tomorrow in the afternoon | 01:11 |
*** weiyu_ has joined #openstack-kolla | 01:17 | |
*** unicell1 has joined #openstack-kolla | 01:17 | |
*** unicell has quit IRC | 01:17 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Do not use AUFS when building centos container images on ubuntu host https://review.openstack.org/314205 | 01:19 |
*** britthouser has quit IRC | 01:21 | |
*** britthouser has joined #openstack-kolla | 01:21 | |
*** mummer has quit IRC | 01:28 | |
v1k0d3n | sdake: sounds good. dude, thanks again! i'm really excited about getting started :) | 01:30 |
sdake | v1k0d3n cool - aare you on win or mac so we can ru nwebex? | 01:31 |
*** kjelly has quit IRC | 01:37 | |
*** sacharya has joined #openstack-kolla | 01:39 | |
*** sacharya has quit IRC | 01:44 | |
*** vhosakot has joined #openstack-kolla | 01:44 | |
*** diogogmt has quit IRC | 01:44 | |
*** salv-orlando has joined #openstack-kolla | 01:44 | |
*** rhallisey has quit IRC | 01:48 | |
*** dolpher has joined #openstack-kolla | 01:48 | |
*** salv-orlando has quit IRC | 01:51 | |
*** kjelly has joined #openstack-kolla | 01:54 | |
*** unicell1 has quit IRC | 01:55 | |
*** ravig has joined #openstack-kolla | 01:56 | |
*** ravig has quit IRC | 02:02 | |
*** sacharya has joined #openstack-kolla | 02:02 | |
*** jtriley has quit IRC | 02:07 | |
*** jtriley has joined #openstack-kolla | 02:15 | |
*** ayoung has quit IRC | 02:17 | |
*** daneyon has joined #openstack-kolla | 02:36 | |
*** sean-k-mooney has joined #openstack-kolla | 02:39 | |
*** daneyon has quit IRC | 02:41 | |
*** coolsvap has joined #openstack-kolla | 02:41 | |
*** allen_gao has quit IRC | 02:44 | |
*** allen_gao has joined #openstack-kolla | 02:48 | |
*** yuanying has quit IRC | 02:51 | |
v1k0d3n | mac | 02:51 |
v1k0d3n | sdake: ^ | 02:52 |
*** salv-orlando has joined #openstack-kolla | 02:53 | |
*** salv-orlando has quit IRC | 02:54 | |
*** klint has joined #openstack-kolla | 02:56 | |
*** allen_gao has quit IRC | 02:57 | |
*** mummer has joined #openstack-kolla | 02:58 | |
*** allen_gao has joined #openstack-kolla | 03:04 | |
*** sdake has quit IRC | 03:07 | |
openstackgerrit | Merged openstack/kolla: Re-order the when condition in the cinder reconfigure https://review.openstack.org/310973 | 03:10 |
*** Jeffrey4l_ has quit IRC | 03:18 | |
*** ravig has joined #openstack-kolla | 03:19 | |
*** Jeffrey4l_ has joined #openstack-kolla | 03:19 | |
*** ravig has quit IRC | 03:21 | |
*** ravig has joined #openstack-kolla | 03:23 | |
*** mummer has quit IRC | 03:32 | |
*** godleon has joined #openstack-kolla | 03:35 | |
*** ayoung has joined #openstack-kolla | 03:42 | |
*** yuanying has joined #openstack-kolla | 03:48 | |
*** salv-orlando has joined #openstack-kolla | 03:58 | |
*** salv-orlando has quit IRC | 04:02 | |
*** sdake has joined #openstack-kolla | 04:13 | |
*** unicell has joined #openstack-kolla | 04:26 | |
*** jtriley has quit IRC | 04:43 | |
*** salv-orlando has joined #openstack-kolla | 04:57 | |
*** fragatina has quit IRC | 04:59 | |
*** sdake has quit IRC | 05:06 | |
*** fragatina has joined #openstack-kolla | 05:17 | |
*** daneyon has joined #openstack-kolla | 05:18 | |
*** fragatina has quit IRC | 05:22 | |
*** daneyon has quit IRC | 05:23 | |
*** coolsvap has quit IRC | 05:25 | |
*** fragatina has joined #openstack-kolla | 05:33 | |
*** fragatina has quit IRC | 05:33 | |
*** fragatina has joined #openstack-kolla | 05:34 | |
*** vhosakot has quit IRC | 05:57 | |
*** rajathagasthya has joined #openstack-kolla | 06:08 | |
*** rajathagasthya has quit IRC | 06:12 | |
*** berendt has joined #openstack-kolla | 06:19 | |
berendt | Hi. Is Ansible 1.9.4 still a valid requirement or is it possible to use Ansible 2.0.x? | 06:22 |
*** ravig has quit IRC | 06:22 | |
*** jmccarthy has joined #openstack-kolla | 06:29 | |
*** jmccarthy has left #openstack-kolla | 06:30 | |
*** gumbyblake has joined #openstack-kolla | 06:35 | |
*** berendt has quit IRC | 06:36 | |
*** berendt has joined #openstack-kolla | 06:36 | |
*** neilus has joined #openstack-kolla | 06:40 | |
*** Serlex has joined #openstack-kolla | 06:42 | |
*** diga has joined #openstack-kolla | 06:46 | |
*** sacharya has quit IRC | 06:53 | |
*** diga has quit IRC | 06:57 | |
*** salv-orlando has quit IRC | 07:05 | |
*** daneyon has joined #openstack-kolla | 07:06 | |
*** Mr_Broken has joined #openstack-kolla | 07:07 | |
*** daneyon has quit IRC | 07:11 | |
*** b_bezak has joined #openstack-kolla | 07:19 | |
*** mikelk has joined #openstack-kolla | 07:34 | |
*** diga has joined #openstack-kolla | 07:35 | |
*** gumbyblake has left #openstack-kolla | 07:35 | |
*** jtriley has joined #openstack-kolla | 07:39 | |
*** dolpher has quit IRC | 07:42 | |
*** diga has quit IRC | 07:43 | |
*** sacharya has joined #openstack-kolla | 07:54 | |
*** shardy has joined #openstack-kolla | 07:56 | |
*** sacharya has quit IRC | 07:58 | |
*** gfidente has joined #openstack-kolla | 08:08 | |
*** sbezverk_ has joined #openstack-kolla | 08:08 | |
*** sbezverk has quit IRC | 08:09 | |
*** athomas has joined #openstack-kolla | 08:14 | |
*** vincent_vdk has quit IRC | 08:28 | |
*** vincent_vdk has joined #openstack-kolla | 08:28 | |
openstackgerrit | Christian Berendt proposed openstack/kolla: Improve logging output of kolla-build command https://review.openstack.org/314460 | 08:34 |
*** vincent_vdk has quit IRC | 08:34 | |
*** vincent_vdk has joined #openstack-kolla | 08:37 | |
*** sacharya has joined #openstack-kolla | 09:00 | |
*** sdake has joined #openstack-kolla | 09:01 | |
*** jtriley has quit IRC | 09:04 | |
*** jtriley has joined #openstack-kolla | 09:04 | |
*** sacharya has quit IRC | 09:04 | |
*** diga has joined #openstack-kolla | 09:11 | |
openstackgerrit | Merged openstack/kolla: Make "cleanup-containers" script works dynamically https://review.openstack.org/313526 | 09:13 |
*** pbourke has quit IRC | 09:24 | |
*** pbourke has joined #openstack-kolla | 09:24 | |
*** salv-orlando has joined #openstack-kolla | 09:27 | |
pbourke | berendt: currently | 09:28 |
pbourke | berendt: yes | 09:28 |
pbourke | berendt: initiall 2.x support being added currently | 09:29 |
berendt | pbourke: thanks for the reply. is it planned to change this? | 09:29 |
berendt | pbourke: ok | 09:29 |
*** sean-k-mooney has quit IRC | 09:29 | |
*** salv-orlando has quit IRC | 09:32 | |
*** ccesario___ has joined #openstack-kolla | 09:36 | |
*** salv-orlando has joined #openstack-kolla | 09:38 | |
*** salv-orlando has joined #openstack-kolla | 09:38 | |
*** gfidente has quit IRC | 09:40 | |
*** diga has quit IRC | 09:40 | |
*** coolsvap has joined #openstack-kolla | 09:58 | |
openstackgerrit | Paul Bourke proposed openstack/kolla: Improve Vagrants bootstrap.sh proxy support https://review.openstack.org/310246 | 10:13 |
*** SiRiuS has joined #openstack-kolla | 10:13 | |
*** ccesario___ has quit IRC | 10:13 | |
openstackgerrit | Paul Bourke proposed openstack/kolla: Improve Vagrants bootstrap.sh proxy support https://review.openstack.org/310246 | 10:14 |
openstackgerrit | Paul Bourke proposed openstack/kolla: Improve Vagrants bootstrap.sh proxy support https://review.openstack.org/310246 | 10:15 |
Mr_Broken | Morning guys | 10:31 |
Mr_Broken | Are there any deployment options available for customizing topology ?for stable/mitaka ? | 10:31 |
Mr_Broken | I need a separate compute node | 10:31 |
Mr_Broken | and want to have cinder | 10:32 |
Mr_Broken | and neutron_server with ovs on this compute node | 10:32 |
Mr_Broken | is there any way , to tell kolla deployer NOT to deploy specific dockers ? | 10:32 |
Mr_Broken | for example deploy cinder_api and cinder_scheduler on controllers, but cinder_volume on Compute node ? | 10:32 |
Mr_Broken | also I would like to split some nova services | 10:40 |
*** sdake has quit IRC | 10:42 | |
*** daneyon has joined #openstack-kolla | 10:43 | |
sbezverk_ | Mr_Broken you can achieve it by placing your hosts in a specific groups in inventory file.. Example host in storage group and networking group will not have deployed controllers' containers | 10:45 |
Mr_Broken | ok , but my config is a bit weird | 10:45 |
Mr_Broken | for example - i would like only cinder_storage service on compute node (instead of Storage node ) | 10:46 |
Mr_Broken | and I need nova_libvirt and neutron_server + OVS installed as standalone on this compute node | 10:46 |
Mr_Broken | (since there is no working LVM driver for stable/mitaka dockers ) | 10:46 |
sbezverk_ | Mr_Broken Hvae you tried to place you compute node into compute group and into storage group at the same time? | 10:48 |
*** Serlex has quit IRC | 10:48 | |
*** daneyon has quit IRC | 10:48 | |
Mr_Broken | yes that would work fine | 10:48 |
Mr_Broken | but what if i dont want a specific container to be deployed at all ? | 10:48 |
Mr_Broken | for example i dont want nova_libvirt to be deployed at all | 10:48 |
sbezverk_ | Mr_Broken Sorry but how then you are going to start VMs? | 10:49 |
Mr_Broken | I will install nova standalone on this Bare Metal compute | 10:49 |
Mr_Broken | and just point it in config files to the Kolla | 10:50 |
Mr_Broken | (since i require LVM Iscsi driver working for example ) | 10:50 |
Mr_Broken | thats why i would like to NOT deploy some containers at all | 10:50 |
Mr_Broken | instead of deploying them - and later on removing | 10:50 |
*** Serlex has joined #openstack-kolla | 10:51 | |
Mr_Broken | so basicaly - half Kolla, half self Openstack :] | 10:51 |
sbezverk_ | Mr_Broken I see at this point I do not believe something like that exist. You can try to use enable_nova: "no" in globals.yaml, but I have never tried it myself.. | 10:51 |
*** ccesario___ has joined #openstack-kolla | 10:52 | |
Mr_Broken | i think there is to many Nova dependent services for that ;] | 10:52 |
Mr_Broken | i still want nova_scheduler and API running on COntrollers ;] | 10:52 |
Mr_Broken | i like this 3 node HA setup ;] | 10:52 |
Mr_Broken | but i guess i just want to much :> | 10:52 |
Mr_Broken | thanks for the clarificatioj | 10:52 |
Mr_Broken | clarification* | 10:53 |
sbezverk_ | Mr_Broken np, is it a production env or test? | 10:53 |
Mr_Broken | PoC - but soon would like to have it prod ;] | 10:54 |
sbezverk_ | Mr_Broken I see, so cherrypick iSCSI support patch is not an option for you then.. | 10:56 |
Mr_Broken | not yet | 10:57 |
Mr_Broken | when it will be merged to stable | 10:57 |
Mr_Broken | than sure | 10:57 |
Mr_Broken | at least that was my architect decision :/ | 10:58 |
sbezverk_ | Mr_Broken it is in master now, but it will be a part of Newton stable | 10:59 |
Mr_Broken | for some stupid reason we are bound to use Mitaka still :/ | 11:00 |
Mr_Broken | i think its because we will need some integration with 3rd party software like ICO | 11:00 |
*** sacharya has joined #openstack-kolla | 11:01 | |
*** b_bezak has quit IRC | 11:01 | |
Mr_Broken | and they are stuck with liberty now, moving to mitaka soon | 11:01 |
*** b_bezak has joined #openstack-kolla | 11:02 | |
*** ccesario___ has quit IRC | 11:02 | |
openstackgerrit | Christian Berendt proposed openstack/kolla: Improve logging output of kolla-build command https://review.openstack.org/314460 | 11:04 |
*** sacharya has quit IRC | 11:05 | |
*** weiyu_ has quit IRC | 11:05 | |
*** rhallisey has joined #openstack-kolla | 11:06 | |
*** sdake has joined #openstack-kolla | 11:07 | |
rhallisey | sdake, morning | 11:08 |
*** godleon has quit IRC | 11:09 | |
sdake | zug | 11:09 |
openstackgerrit | Paul Bourke proposed openstack/kolla: Fix build for oraclelinux base https://review.openstack.org/310774 | 11:13 |
*** dims has quit IRC | 11:20 | |
*** mlima has joined #openstack-kolla | 11:22 | |
*** dims has joined #openstack-kolla | 11:25 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Keystone Service and Pod https://review.openstack.org/312944 | 11:28 |
*** b_bezak has quit IRC | 11:30 | |
*** b_bezak has joined #openstack-kolla | 11:33 | |
*** weiyu has joined #openstack-kolla | 11:37 | |
*** papacz has quit IRC | 11:43 | |
*** allen_gao has quit IRC | 11:44 | |
*** lukl has joined #openstack-kolla | 11:49 | |
openstackgerrit | Ryan Hallisey proposed openstack/kolla: Spec: Deploy Kolla images using Kubernetes https://review.openstack.org/304182 | 11:49 |
*** sdake has quit IRC | 11:56 | |
*** weiyu has quit IRC | 11:57 | |
*** gfidente has joined #openstack-kolla | 11:57 | |
*** dwalsh has joined #openstack-kolla | 11:59 | |
*** sdake has joined #openstack-kolla | 12:02 | |
*** weiyu_ has joined #openstack-kolla | 12:08 | |
*** ozialien10 has quit IRC | 12:08 | |
*** sdake_ has joined #openstack-kolla | 12:08 | |
*** ozialien10 has joined #openstack-kolla | 12:09 | |
*** sdake has quit IRC | 12:11 | |
*** Liuqing has joined #openstack-kolla | 12:11 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/kolla: Spec: Deploy Kolla images using Kubernetes https://review.openstack.org/304182 | 12:15 |
coolsvap | rhallisey, updated some minor typos in kubernetes spec | 12:15 |
rhallisey | coolsvap, thanks | 12:16 |
coolsvap | np just wanted to let you know what i updated | 12:17 |
*** jtriley has quit IRC | 12:17 | |
*** weiyu_ has quit IRC | 12:23 | |
*** Liuqing has quit IRC | 12:25 | |
*** Liuqing has joined #openstack-kolla | 12:27 | |
*** Mr_Broken has quit IRC | 12:28 | |
*** Mr_Broken has joined #openstack-kolla | 12:29 | |
*** weiyu_ has joined #openstack-kolla | 12:30 | |
*** Mr_Broken has quit IRC | 12:34 | |
*** coolsvap has quit IRC | 12:47 | |
*** coolsvap has joined #openstack-kolla | 12:48 | |
*** salv-orlando has quit IRC | 12:49 | |
*** coolsvap has quit IRC | 12:50 | |
*** coolsvap has joined #openstack-kolla | 12:50 | |
*** SiRiuS has quit IRC | 12:55 | |
*** jtriley has joined #openstack-kolla | 12:56 | |
*** mbound has joined #openstack-kolla | 12:57 | |
*** SiRiuS has joined #openstack-kolla | 12:59 | |
*** klint has quit IRC | 13:00 | |
*** sacharya has joined #openstack-kolla | 13:02 | |
*** mbound has quit IRC | 13:02 | |
*** dave-mccowan has joined #openstack-kolla | 13:03 | |
*** salv-orlando has joined #openstack-kolla | 13:03 | |
*** sacharya has quit IRC | 13:06 | |
*** neilus1 has joined #openstack-kolla | 13:07 | |
*** neilus has quit IRC | 13:11 | |
*** Mr_Broken has joined #openstack-kolla | 13:15 | |
ccesario | morning guys! | 13:22 |
*** daneyon has joined #openstack-kolla | 13:25 | |
*** daneyon has quit IRC | 13:30 | |
*** esharao has joined #openstack-kolla | 13:30 | |
sbezverk_ | mlima ping | 13:40 |
*** berendt has quit IRC | 13:40 | |
mlima | sbezverk_, :) | 13:41 |
*** weiyu_ has quit IRC | 13:41 | |
*** signed8bit has joined #openstack-kolla | 13:43 | |
*** maestro1 has joined #openstack-kolla | 13:45 | |
sbezverk_ | mlima I wanted to check with you if you had a chance to see my reply for your review :-) | 13:46 |
mlima | sbezverk_, no, i'm sorry | 13:47 |
mlima | :/ | 13:47 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/kolla-kubernetes: Switch on warnerrors and universal wheel https://review.openstack.org/313137 | 13:47 |
mlima | i'll do it now | 13:47 |
sbezverk_ | mlima thank you, I have not changed anything just provide the logic behind why I think it should be done this way :-) | 13:48 |
*** SiRiuS has quit IRC | 13:48 | |
*** Mr_Broken has quit IRC | 13:49 | |
*** Mr_Broken has joined #openstack-kolla | 13:50 | |
*** sean-k-mooney has joined #openstack-kolla | 13:50 | |
dims | coolsvap : any feedback on https://review.openstack.org/#/c/313958/ ? it would be great to get that in and then iterate as it's mostly just copied code from older repos | 13:51 |
patchbot | dims: patch 313958 - kolla-kubernetes - Implement a command line to start/stop keystone se... | 13:51 |
*** SiRiuS has joined #openstack-kolla | 13:53 | |
openstackgerrit | Paul Bourke proposed openstack/kolla: Debug oraclelinux deploy gates https://review.openstack.org/300988 | 13:53 |
*** coolsvap has quit IRC | 13:54 | |
openstackgerrit | Paul Bourke proposed openstack/kolla: Debug oraclelinux/centos rabbitmq deploy issue https://review.openstack.org/300988 | 13:55 |
*** Liuqing has quit IRC | 13:56 | |
*** MarMat has joined #openstack-kolla | 13:58 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Switch on warnerrors and universal wheel https://review.openstack.org/313137 | 14:02 |
*** SiRiuS has quit IRC | 14:06 | |
sean-k-mooney | o/ how are everyone today | 14:14 |
rhallisey | hey sean-k-mooney | 14:17 |
*** neilus1 has quit IRC | 14:18 | |
*** jtriley has quit IRC | 14:24 | |
rhallisey | dave-mccowan, morning | 14:24 |
dave-mccowan | good morning ryan | 14:24 |
rhallisey | dave-mccowan, can you try to add your custom mariabd, kibana, and elastic search repos into the base directory | 14:24 |
rhallisey | and replace the ones that are there | 14:25 |
rhallisey | dave-mccowan, add those repos to rpm_config_setup is tricky because they are meant to be global | 14:25 |
rhallisey | where as rpm_setup_config is not | 14:25 |
dave-mccowan | what about the gpg signatures? those are hard coded too, instead of being read from the .repos | 14:26 |
rhallisey | ugh this hardcoding is not ideal... | 14:28 |
dave-mccowan | rhallisey the other approach i was considering is to add config variables, that can be global, such as {{ mariadb_repo }}, {{ kibana_rep }}, etc. but the Dockerfile.j2 would still need to be fixed to use all the params in the .repo files. | 14:29 |
*** SiRiuS has joined #openstack-kolla | 14:30 | |
rhallisey | dave-mccowan, I think that is better | 14:35 |
rhallisey | since these repos are exceptions | 14:36 |
dave-mccowan | rhallisey what about line 85 of base/Dockerfile.j2? epel release for rhel is also hard coded to an rpm at https://mirror.centos.org/ | 14:39 |
rhallisey | I think that's also going to need to be abstracted out | 14:40 |
rhallisey | this is kinda of a mess.. oracle linux also has epel + gpg keys | 14:40 |
*** b_bezak has quit IRC | 14:41 | |
*** jamie_h has joined #openstack-kolla | 14:42 | |
dave-mccowan | rhallisey i think we're on the same page now. :-) it seems like each package is installed a little bit differently too. (some with curl, some copy, some yum, some rpm) i haven't been able to figure out the "right" way to attempt to make them consistent, so then it would be easier to make them customizable. | 14:44 |
rhallisey | ya it's not completely consumable | 14:46 |
rhallisey | customizable rather | 14:46 |
rhallisey | may need a blueprint for this one.. :) | 14:46 |
rhallisey | bunch of cleanup to do | 14:46 |
dave-mccowan | rhallisey it's tempting to make an install_metatype == "custom" setting, to branch out of the current template to get it to work for an all-custom build, and leave the current code as technical debt. | 14:47 |
rhallisey | ya I can see how that's tempting | 14:48 |
sbezverk_ | rhallisey have a second for qq? | 14:49 |
rhallisey | dave-mccowan, I don't think I can get to this today, but maybe later this week | 14:50 |
rhallisey | sbezverk_, ya what's up | 14:50 |
rhallisey | dave-mccowan, the patch I came up with last night I don't like :( | 14:50 |
sbezverk_ | rhallisey currently kolla stores config on target host in /etc/kolla/, in kubernetes this will not be the case right? | 14:50 |
sbezverk_ | otherwise we loose "mobility" of pod | 14:51 |
rhallisey | sbezverk_, ya we'll plan on copying the configs to the same place | 14:51 |
dave-mccowan | rhallisey is your patch a good starting point? | 14:51 |
rhallisey | on all nodes | 14:51 |
rhallisey | dave-mccowan, I don't think so. I added a global_config field | 14:52 |
rhallisey | dave-mccowan, it functions like rpm_config_setup | 14:52 |
Lyncos | Yesterday someone talk about Weavescope ... | 14:52 |
rhallisey | it works, but it doesn't fix the gpg key issue | 14:52 |
rhallisey | but I could add that | 14:52 |
rhallisey | dave-mccowan, here I'll just post it | 14:52 |
rhallisey | dave-mccowan, so you can see for yourself | 14:53 |
sbezverk_ | rhallisey why do not we completely stop using host(workers) to store config and use configmap for all confiugration related activities? | 14:56 |
esharao | hi guys.. | 14:57 |
esharao | 14:57 | |
esharao | is it possible to start the kolla deployment on a VM.. or do i need a physical host to start | 14:57 |
*** huikang has joined #openstack-kolla | 14:57 | |
openstackgerrit | Ryan Hallisey proposed openstack/kolla: Allow the base container to be more customizable https://review.openstack.org/314614 | 14:58 |
rhallisey | dave-mccowan, ^ | 14:58 |
rhallisey | esharao, you can use a vm | 14:59 |
esharao | i tried to use the vagrant stuff via a virtual box .. but dont think thats the right way | 14:59 |
rhallisey | I've never used the vagrant stuff unfortunately | 15:00 |
esharao | ok..let me poke around more.. thanks rhallisey | 15:00 |
rhallisey | no problem | 15:00 |
*** huikang has quit IRC | 15:01 | |
rhallisey | sbezverk_, I'm still thinking about the configmap thing | 15:01 |
rhallisey | kfox brought up a good point about using deployments and triggering a rolling upgrade for reconfigure | 15:01 |
*** ntpttr- has joined #openstack-kolla | 15:03 | |
sbezverk_ | rhallisey I see, I got configmap working and it looks pretty good.. it seems they would allow to store all process' config files, so no need to store anything on the worker | 15:04 |
*** Jeffrey4l_ has quit IRC | 15:05 | |
*** ntpttr has quit IRC | 15:06 | |
*** huikang_ has joined #openstack-kolla | 15:07 | |
*** sacharya has joined #openstack-kolla | 15:09 | |
*** rmart04 has joined #openstack-kolla | 15:11 | |
*** ntpttr- is now known as ntpttr | 15:14 | |
*** ntpttr has left #openstack-kolla | 15:16 | |
*** anbehl has quit IRC | 15:16 | |
rhallisey | sbezverk_, where would it get stored then? | 15:17 |
rhallisey | in /var/lib/docker on the host? In other words the container? | 15:18 |
*** huikang_ has quit IRC | 15:18 | |
rhallisey | dave-mccowan, let me know if that patch helps at all | 15:18 |
*** anbehl has joined #openstack-kolla | 15:18 | |
rhallisey | you will still need to change the repos in that directory | 15:19 |
rhallisey | and I'll have to add gpg keys in there too | 15:19 |
rhallisey | dims, I'm trying to run the cli, but I'm not having any success | 15:20 |
*** kbyrne has quit IRC | 15:22 | |
*** jtriley has joined #openstack-kolla | 15:23 | |
*** vhosakot has joined #openstack-kolla | 15:23 | |
dave-mccowan | rhallisey percona is handled differently than the other 3 in this list. we should make them consistent. which way do you think is right? (using yum install or COPY) | 15:23 |
rhallisey | dave-mccowan, well percona is a link | 15:24 |
*** rhallisey has left #openstack-kolla | 15:24 | |
*** rhallisey has joined #openstack-kolla | 15:25 | |
*** ravig has joined #openstack-kolla | 15:25 | |
dims | rhallisey : i bet there's some sensitivity in _load_variables_from_file, i'll add more LOG.debug(s) later today | 15:25 |
dave-mccowan | the others have links too, they are just inside the .repo file. rhallisey do you think we should change percona to also use a .repo file? | 15:26 |
sbezverk_ | rhallisey I think so, what is important k8s should be taking care of config "mobility" in case pod moves | 15:26 |
rhallisey | dims, awesome thanks dims. I'll hopefully have a review up for it soon | 15:26 |
rhallisey | dave-mccowan, it doesn't make a difference because I'm using the same function as rpm_setup_config, which can accept links, rpms, or .repo files | 15:27 |
sbezverk_ | rhallisey I am trying to bring up yuor kesytone pod and service using configmap and test, I have 2 workers.. | 15:27 |
rhallisey | sbezverk_, yes I agree | 15:28 |
*** sacharya_ has joined #openstack-kolla | 15:29 | |
rhallisey | dims, were you running this from /usr/share/kolla-kubernetes? | 15:30 |
*** sacharya has quit IRC | 15:31 | |
dave-mccowan | rhallisey if we support all three methods, then we need to support 3 ways of providing a gpgkey. | 15:31 |
dims | rhallisey : i pip installed kolla. then was running kolla-kubernetes from within a venv (py27) | 15:31 |
rhallisey | dave-mccowan, see the comment https://review.openstack.org/#/c/314614/ | 15:33 |
patchbot | rhallisey: patch 314614 - kolla - Allow the base container to be more customizable | 15:33 |
rhallisey | dave-mccowan, the alternative is to document document | 15:34 |
rhallisey | still does leave tech debt though | 15:34 |
*** kbyrne has joined #openstack-kolla | 15:36 | |
dave-mccowan | rhallisey that's an interesting approach. just make the base/Dockerfile.j2 an operator input file. if so, it should be passed as a parameter to kolla-build, so it doesn't need to be modified in branch. | 15:37 |
dave-mccowan | rhallisey if we expect operators to modify that file, it makes addressing the inconsistencies even more important. | 15:38 |
cinerama | sean-k-mooney, hi are you around? | 15:38 |
*** rhallisey has quit IRC | 15:39 | |
sean-k-mooney | cinerama: hi yes | 15:40 |
*** kbyrne has quit IRC | 15:40 | |
*** rhallisey has joined #openstack-kolla | 15:40 | |
d_code | before I create a new blueprint (I guess this is where this would go), can anyone confirm that today you cannot (directly) deploy kolla as a non-root user via ansible’s --become, right? | 15:40 |
*** kbyrne has joined #openstack-kolla | 15:40 | |
dave-mccowan | rhallisey i think the better approach is to let operators change the .repo files. we just need to get the Dockerfile template to consume them properly/completely. | 15:40 |
d_code | I’m working in an environment where users are not allowed to SSH directly as root | 15:40 |
rhallisey | d_code, I think there is a way sdake_ might know | 15:41 |
d_code | I was able to do it using ansible directly, but it’s dead simple…just pass the become flag | 15:41 |
sean-k-mooney | rhallisey i belive you need to play with your sudoers file but sdake did mention that it should work at the summit | 15:41 |
rhallisey | what's the become flag? | 15:42 |
d_code | ansible-playbook --become | 15:42 |
d_code | it invokes sudo on the remote system | 15:42 |
*** Mr_Broken has quit IRC | 15:43 | |
cinerama | sean-k-mooney: hey i've been following the kolla/bifrost thread & i just started looking at what you've done as a POC | 15:43 |
rhallisey | d_code, ok gotcha | 15:43 |
sean-k-mooney | cinerama: its a horible hack but its partly working. | 15:43 |
rhallisey | d_code, I think there is as sean-k-mooney also mentioned | 15:43 |
rhallisey | I don't know personally though, sdake_ would know when he's back | 15:44 |
rhallisey | d_code, either way, there is no docs around it | 15:44 |
cinerama | wrt venv issues, in our install playbook we do support installation into a venv...you would need to bootstrap the install by creating a venv and installing ansible into it like we do in the env setup script | 15:44 |
rhallisey | d_code, so a blueprint would be an excellent idea | 15:44 |
d_code | right… would something like this go as a bug or blueprint? | 15:44 |
d_code | k | 15:44 |
d_code | I”ll file it | 15:44 |
rhallisey | awesome thx | 15:44 |
*** mummer has joined #openstack-kolla | 15:45 | |
*** mikelk has quit IRC | 15:45 | |
cinerama | sean-k-mooney: also, i think we should be able to split up the installation of services and starting the services into different roles | 15:45 |
rhallisey | dave-mccowan, I think it will be a matter of opinion of whether operators consume by changing the dockerfile or not | 15:45 |
rhallisey | I'm not sure what is right, but what we have now is inconsistent | 15:46 |
rhallisey | since you can only do it for non source centos builds | 15:46 |
sean-k-mooney | cinerama: actully i think the venv support is the problem. in source mode it uses the venv but when ansible runs i dont think its useing the system python and not finding the packages in the venv | 15:46 |
cinerama | sean-k-mooney: oh also just in general i've been hanging out here if folks want to ask stuff about bifrost | 15:46 |
*** kbyrne has quit IRC | 15:47 | |
*** sdake_ has quit IRC | 15:47 | |
sean-k-mooney | cinerama: did you look at how i refactored the bifrost install playbook? https://github.com/SeanMooney/bifrost/commit/e223f4fe73871b76ce87999470a1efc43862671e | 15:48 |
cinerama | huh. so you're running the ansible that lives in the venv and it's not seeing other stuff in that venv? | 15:48 |
sean-k-mooney | cinerama: am ansible is installed in /opt/stack/ansible but kola has a venv in /usr/lib/kolla/venv i think | 15:49 |
*** sdake has joined #openstack-kolla | 15:50 | |
*** maestro1 has quit IRC | 15:50 | |
*** dwalsh has quit IRC | 15:51 | |
cinerama | sean-k-mooney: and who/what is installing ansible? bifrost, or something else? | 15:51 |
*** sdake_ has joined #openstack-kolla | 15:54 | |
*** sdake has quit IRC | 15:57 | |
sean-k-mooney | bifrost kolla and me are all installing ansible. at different points. | 15:58 |
sean-k-mooney | cinerama: ansible is available during the the container build form kolla but seams to be missing after so i explicitly installed it form pip | 16:00 |
cinerama | interesting | 16:00 |
sean-k-mooney | cinerama: bifrost also installs it to /opt/stack/ansible | 16:00 |
*** Avi82 has joined #openstack-kolla | 16:00 | |
cinerama | yup that's right. you can also do export VENV=/path/to/venv and put it in a venv | 16:00 |
cinerama | and the other playbooks should pick up that environment variable, or you can do -e and specify it to ansible when you run the playbooks | 16:01 |
sdake_ | sup | 16:01 |
openstackgerrit | Ryan Hallisey proposed openstack/kolla: Allow the base container to be more customizable https://review.openstack.org/314614 | 16:01 |
sean-k-mooney | cinerama: ah i will try that tonight. | 16:01 |
cinerama | right, -e bifrost_venv_dir | 16:02 |
cinerama | and -e enable_venv=true | 16:02 |
sean-k-mooney | cinerama: but will that install bifrost in a venv? that might work but i was trying to install it system wide but maybe i need to install it into the kolla venv | 16:03 |
cinerama | i had a quick look at the playbooks, just at a high level i think that's about what i would have done | 16:03 |
cinerama | sean-k-mooney: it will install it into a venv. by default that will be in /opt/stack/bifrost but you can change the location & use an existing venv if you wanted to put it in the kolla one | 16:03 |
*** mark-casey has joined #openstack-kolla | 16:05 | |
sean-k-mooney | cinerama: ok well that is something to try. i would like to reuse as much of bifrost as i can with as few changes as possible ideally. | 16:06 |
sean-k-mooney | cinerama: i have been doing this poc with my home server so i will have to take a look at it when i leave the office. | 16:06 |
*** mlima_ has joined #openstack-kolla | 16:09 | |
*** stvnoyes has joined #openstack-kolla | 16:12 | |
*** sacharya_ has quit IRC | 16:12 | |
*** mlima has quit IRC | 16:13 | |
*** berendt has joined #openstack-kolla | 16:15 | |
*** dwalsh has joined #openstack-kolla | 16:16 | |
d_code | sdake_: just created a blueprint proposing ansible —become support | 16:17 |
sean-k-mooney | sdake_: i dont know if you saw my mail eairlier regarding the bifrost poc. im hoping to get it to the point where i can run the deploy_dynamic playbook and have it provision a server. | 16:18 |
sean-k-mooney | sdake_: currently i have got as far as starting the bifrost container enroling the nodes and building the os image with disk image builder. | 16:19 |
sdake_ | that sound spretty far :) | 16:20 |
sdake_ | d_code you can run koll as nonroot currently | 16:20 |
sdake_ | the permissions on the directories have to be perfectly setup | 16:20 |
d_code | hmm…. so the ssh user needs to be in the docker group…and there are config files that go in /etc, right? | 16:21 |
sean-k-mooney | sdake_: the codes a little rough around the edges but when the node is enroled ironic can manage the power state via ipmi but i need to do figure out why the dynamic deploy is broken. ill try some of cinerama's tips and see if it helps tonight | 16:22 |
sdake_ | cool - so you broke up the playbook into 3 parts ? | 16:23 |
sdake_ | just so everyone is aware, I have a stomach virus or flu of some sort | 16:23 |
sdake_ | so today I am not at my best - i'm probably going to be in bed most of the day | 16:24 |
sean-k-mooney | sdake_: yes it can probably be done a better way then how i did it but i have. | 16:24 |
sean-k-mooney | sdake_: that sucks :( | 16:25 |
britthouser | Sorry sdake_ =( | 16:26 |
openstackgerrit | Christian Berendt proposed openstack/kolla: Improve logging output of kolla-build command https://review.openstack.org/314460 | 16:27 |
*** dwalsh has quit IRC | 16:27 | |
*** thomasem has left #openstack-kolla | 16:29 | |
mark-casey | cinerama: sean-k-mooney: would you guys be up for converting the skip_* vars to tags on either the tasks or include statements? http://docs.ansible.com/ansible/playbooks_tags.html <(specifically the section: "And you may also tag basic include statements") | 16:30 |
mark-casey | so... 'ansible-playbook -vvvv -i /bifrost/playbooks/inventory/localhost /bifrost/playbooks/install.yaml --skip-tags: bf_service_start_and_start_dependent' | 16:30 |
*** berendt has quit IRC | 16:31 | |
sean-k-mooney | mark-casey: ansible is not my strongest language so if there is a better(more native) way to do it i would be happy to try it. | 16:31 |
sean-k-mooney | mark-casey: i think cinerama was also suggesting splitting it into seperate roles as another option so you would have 3 roles: install,bootstrap and start | 16:33 |
sean-k-mooney | mark-casey: since bifrost already had a skip_install i just copied the pattern and introduced skip_bootstrap and skip_start | 16:36 |
cinerama | yes we haven't been using tags | 16:36 |
sdake_ | hey cinerama are you a bifrost dev? | 16:38 |
*** sdake_ is now known as sdake | 16:38 | |
sdake | sean-k-mooney 3 roles would be the way to roll :) | 16:38 |
*** stvnoyes has left #openstack-kolla | 16:39 | |
sdake | then you could run one tag at a time vs skip_tag | 16:39 |
sdake | rather skip_install | 16:39 |
sdake | not sure what that is | 16:39 |
cinerama | sdake: i am! | 16:39 |
*** huikang has joined #openstack-kolla | 16:39 | |
sdake | cool have we met in person | 16:40 |
*** berendt has joined #openstack-kolla | 16:40 | |
sdake | happy to have you in channel :) | 16:40 |
mark-casey | cinerama: with the roles option would that be like "- { role: bootstrap, when: "skip_bootstrap" not defined }"? | 16:40 |
sdake | we really want to do what is right by bifrost here | 16:41 |
sdake | rather then ram our agenda down your throats ;) | 16:41 |
sdake | so tell us where we are going wrong, and we can adjust | 16:41 |
sdake | the vm idea - i get that - but that is pretty much not ideal for us | 16:41 |
sdake | i can live with a fat container | 16:41 |
sdake | i think logn term it may or may not make sense to go thin containers | 16:41 |
sdake | but lets take it one step ata time :) | 16:42 |
mark-casey | cinerama: IIRC conditonal roles apply the when ANDed to all tasks in the role. So that would be similar to the current skip_install but easier to read since not written into every single when | 16:42 |
cinerama | mark-casey: so i think what we'd do there is have a boolean to define whether or not we run bootstrap etc | 16:42 |
cinerama | mark-casey: yeah we could put the conditional at the higher level where we call the role - that's why i thought splitting it up would be a good fit | 16:43 |
*** diogogmt has joined #openstack-kolla | 16:43 | |
mark-casey | yea makes sense. | 16:43 |
sean-k-mooney | mark-casey: i placed the skip_* on the include to avoid repeating it an split the main.yml taks into 3 files | 16:44 |
cinerama | we could probably also have some kind of container related boolean and then use supervisord vs systemd etc? | 16:44 |
sean-k-mooney | well systemd does work in a container now though many still dont like the idea | 16:45 |
cinerama | we would have to try it | 16:45 |
sean-k-mooney | just to get the poc working i was using systemd as the init system in the container so i could reuse the service files provided by the OS | 16:46 |
v1k0d3n | sdake: hey man. what time works for you to get together? | 16:46 |
sdake | v1k0d3n vomiting today | 16:47 |
sdake | so not sure how i'll feel this afternoon | 16:47 |
sdake | can we play it by ear? | 16:47 |
*** berendt has quit IRC | 16:47 | |
v1k0d3n | absolutely. sorry to hear :( | 16:47 |
sdake | hazard of having children | 16:47 |
*** sacharya has joined #openstack-kolla | 16:47 | |
cinerama | sean-k-mooney: that makes sense - also in bifrost we try to iterate towards stuff rather than having massive changes where we can | 16:48 |
*** gfidente has quit IRC | 16:48 | |
*** gfidente has joined #openstack-kolla | 16:48 | |
*** sacharya has quit IRC | 16:50 | |
*** sacharya has joined #openstack-kolla | 16:50 | |
sdake | cinerama ack on iteration | 16:51 |
sdake | sean-k-mooney basically make small changes that lead to big changes over time :) | 16:52 |
sdake | but it doesns't have to be big change to start | 16:52 |
cinerama | yep | 16:52 |
sdake | sean-k-mooney if you need help decomposing your patch stream i can provide comments/review, just let me know where the review is | 16:52 |
sean-k-mooney | sdake: yep i agree with that aproch too. | 16:52 |
sdake | sean-k-mooney i'm glad your taking this on, I was getting ready toget started on this proect persoanlly | 16:53 |
sdake | and I'm super overloaded | 16:53 |
cinerama | yes i'd like to review and stuff too. and test | 16:53 |
sdake | does bifrost have gate jobs? | 16:53 |
sdake | cinerama ^^ | 16:54 |
cinerama | yeah we do a test install & deploy | 16:55 |
*** unicell has quit IRC | 16:55 | |
cinerama | you can look at the test scripts in the scripts dir to see what we do | 16:55 |
sean-k-mooney | sdake: well this is currently a gap in the kolla install story and i would personally like to have this so i can clean down my own servers quickly too. | 16:56 |
*** ravig has quit IRC | 16:56 | |
sdake | ya - huge huge gap | 16:56 |
*** huikang has quit IRC | 16:56 | |
sean-k-mooney | cinerama: ye use the agent_ssh driver in the gate right and test with vms? | 16:56 |
sdake | cinerama what some cores have proposed is taking our ironic implementation we have today and making it a standalone since much of the stuff is already containerized | 16:57 |
*** berendt has joined #openstack-kolla | 16:57 | |
sdake | and running it on the undercloud | 16:57 |
sdake | but i'd rather not duplicate all the work of bifrost | 16:58 |
sdake | god i hate undercloud/overcloud terminology | 16:58 |
sdake | so terrible | 16:58 |
cinerama | yes | 16:58 |
cinerama | yes we use agent_ssh & vms | 16:59 |
mark-casey | sdake: yeeeeesss. do not want (over/under) | 17:00 |
sean-k-mooney | trippleo gets a little confusing sometimes... conatinesed trippleo evenmore so | 17:01 |
*** pbourke has quit IRC | 17:01 | |
sean-k-mooney | sdake: we tried(are still trying) to deploy kolla with the ironic driver a few weeks ago and hit alot of problems | 17:01 |
sean-k-mooney | sdake: do you know if anyone has successfully deployed it recently? | 17:02 |
sdake | deployed kolla? | 17:03 |
sdake | or ironic? | 17:03 |
sdake | kolla mitaka works fantasically for me on my 3 node gear | 17:03 |
sdake | liberty is still waiting on a workflow from any of our 12 core reviewers | 17:03 |
sean-k-mooney | ironic or rather kolla with the the iroinc nova driver | 17:03 |
* sdake selling first born | 17:03 | |
sdake | ya kolla ironic implementation is tech preview, it probably doesn't workt he way you would want it to | 17:04 |
sdake | which is as a way to schedule either bare metal or virtual machines via nova | 17:04 |
sdake | that is how I want it to work | 17:04 |
sdake | but jpeeler doesn't think it can be made to do that | 17:04 |
sean-k-mooney | am we were only trying to deploy barmetal | 17:05 |
sdake | yup - not sure if that even wroks | 17:05 |
sdake | i haven't tested it persoanlly | 17:05 |
sdake | its a work in progress IMO | 17:05 |
sean-k-mooney | the issue we had is that pxe infoH | 17:05 |
sdake | i can confirm I have *not* persoanlly tested kolla's implementation of ironic | 17:06 |
sean-k-mooney | the dhcp server was configure with the neutron port uuid instead of the chassi guid so the server never got a reply to the pxe boot form the dhcp server | 17:06 |
sdake | I know its close | 17:06 |
sdake | but devil i ndetails | 17:06 |
*** sacharya has quit IRC | 17:08 | |
*** sacharya has joined #openstack-kolla | 17:08 | |
openstackgerrit | Merged openstack/kolla: Changes needed to deploy liberty with mitaka code https://review.openstack.org/308390 | 17:13 |
*** ccesario___ has joined #openstack-kolla | 17:13 | |
*** jmccarthy1 has joined #openstack-kolla | 17:14 | |
*** rmart04 has quit IRC | 17:15 | |
*** SiRiuS has quit IRC | 17:15 | |
*** athomas has quit IRC | 17:17 | |
openstackgerrit | Ryan Hallisey proposed openstack/kolla: Doc rework for quickstart, multinode, and image-building https://review.openstack.org/311884 | 17:17 |
sdake | rhallisey you up for releasing liberty now | 17:21 |
rhallisey | ya | 17:21 |
sdake | git pull http://github.com/openstack/releases | 17:21 |
sdake | pull/clone | 17:21 |
*** gfidente has quit IRC | 17:22 | |
sdake | find the _independent directory | 17:22 |
*** ccesario___ has quit IRC | 17:23 | |
openstackgerrit | Ryan Hallisey proposed openstack/kolla: Allow the base container to be more customizable https://review.openstack.org/314614 | 17:23 |
*** ccesario___ has joined #openstack-kolla | 17:23 | |
sdake | find the kolla.yml file | 17:23 |
sdake | edit it | 17:23 |
*** unicell has joined #openstack-kolla | 17:23 | |
sdake | and do the right thing with it ;) | 17:23 |
rhallisey | sdake, so is this 1.0.1? | 17:24 |
sdake | NO | 17:24 |
sdake | 1.1.0 | 17:24 |
rhallisey | 1.1 | 17:24 |
rhallisey | right | 17:24 |
rhallisey | meant that | 17:24 |
sdake | dont forge the .0 :) | 17:24 |
sdake | do a git pull of kolla to find the git sha | 17:25 |
sdake | rather git clone | 17:25 |
sdake | and checkout stable/liberty branch | 17:25 |
*** SiRiuS__ has joined #openstack-kolla | 17:27 | |
rhallisey | sdake, does the commit message matter? | 17:28 |
sdake | say somethign like "Release of Kolla Liberty 1.1.0" | 17:28 |
rhallisey | rather does it need to be something specific | 17:28 |
rhallisey | ok | 17:28 |
rhallisey | https://review.openstack.org/314690 | 17:28 |
*** ravig has joined #openstack-kolla | 17:29 | |
*** ccesario___ has quit IRC | 17:36 | |
mark-casey | cinerama: sean-k-mooney: fair warning this is a little unorthodox... but if you're wanting to split the ironic tasks into roles this method prevents hundreds of skipped tasks in ansible-playbook output. Also (unless the 'when:'s are changed to tags, you'd still use -e options as you are now) | 17:36 |
mark-casey | https://github.com/ropsoft/bifrost/commit/148b434f81bc2aac8ac2ee7991ea67d845c7c22e | 17:36 |
mark-casey | essentially, run roles in 3 separate plays which are selected based on localhost being aliased into each of 3 targeted groups | 17:37 |
cinerama | interesting idea | 17:38 |
mark-casey | I admit its a little magic-heavy but I'm pretty sure any other combination of includes or roles with whens or tags will have all tasks print with skipped, vs here it would just be [no hosts matched] | 17:40 |
*** ravig has quit IRC | 17:40 | |
mark-casey | err, maybe tags wouldn't print all skipped; would have to try it | 17:41 |
*** shardy has quit IRC | 17:41 | |
*** ravig has joined #openstack-kolla | 17:41 | |
*** neophy has joined #openstack-kolla | 17:46 | |
*** absubram has joined #openstack-kolla | 17:50 | |
*** ravig has quit IRC | 17:50 | |
*** dwalsh has joined #openstack-kolla | 17:53 | |
*** Serlex has quit IRC | 17:54 | |
*** ravig has joined #openstack-kolla | 17:59 | |
*** rmart04 has joined #openstack-kolla | 17:59 | |
*** SiRiuS__ has quit IRC | 18:01 | |
sean-k-mooney | mark-casey: it took a couple of glances to figure our exactly what your patch does but i get it now. | 18:02 |
sean-k-mooney | mark-casey: with that change you will only get a skip missage for the add to the group if the group is skipped | 18:02 |
dims | rhallisey : around? | 18:07 |
rhallisey | dims, ya | 18:08 |
dims | rhallisey : didn't see a review from you yet, was wondering if you made progress | 18:08 |
rhallisey | I haven't gotten to it yet. I have to handle something internal real quick | 18:09 |
dims | rhallisey : ack thanks | 18:09 |
rhallisey | dims, I haven't gotten it to work yet, still tracing the issue | 18:09 |
rhallisey | but I will come back to it in the next 30 min | 18:10 |
*** ravig has quit IRC | 18:12 | |
*** ravig has joined #openstack-kolla | 18:13 | |
mark-casey | sean-k-mooney: yes. So mostly an output thing since we'll have to run it 2-3 times, but just trying to think of different ways to split it up in hopes of finding the best place to put the complexities | 18:18 |
*** SiRiuS__ has joined #openstack-kolla | 18:20 | |
*** thumpba has joined #openstack-kolla | 18:25 | |
*** haplo37 has joined #openstack-kolla | 18:26 | |
*** sacharya has quit IRC | 18:31 | |
*** haplo37 has quit IRC | 18:40 | |
*** dwalsh has quit IRC | 18:42 | |
*** dwalsh has joined #openstack-kolla | 18:42 | |
*** thumpba has quit IRC | 18:44 | |
*** berendt has quit IRC | 18:45 | |
*** thumpba has joined #openstack-kolla | 18:46 | |
*** berendt has joined #openstack-kolla | 18:47 | |
*** ravig has quit IRC | 18:47 | |
*** ravig has joined #openstack-kolla | 18:49 | |
*** sdake has quit IRC | 18:52 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Implement ceilometer ansible role https://review.openstack.org/300574 | 18:55 |
openstackgerrit | Hui Kang proposed openstack/kolla: Implement ceilometer ansible role https://review.openstack.org/300574 | 18:59 |
*** salv-orl_ has joined #openstack-kolla | 19:02 | |
*** fragatina has quit IRC | 19:02 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Add Kuryr ansible role https://review.openstack.org/298894 | 19:03 |
*** rmart04 has quit IRC | 19:04 | |
*** salv-orlando has quit IRC | 19:05 | |
*** fragatina has joined #openstack-kolla | 19:09 | |
*** mnaser has left #openstack-kolla | 19:10 | |
*** ravig has quit IRC | 19:11 | |
*** Avi82 has quit IRC | 19:13 | |
*** Avi82 has joined #openstack-kolla | 19:17 | |
*** jmccarthy1 has quit IRC | 19:19 | |
*** sacharya has joined #openstack-kolla | 19:22 | |
*** rmart04 has joined #openstack-kolla | 19:24 | |
*** neophy has quit IRC | 19:24 | |
*** neophy has joined #openstack-kolla | 19:30 | |
*** ravig has joined #openstack-kolla | 19:30 | |
*** godleon has joined #openstack-kolla | 19:31 | |
*** rmart04 has quit IRC | 19:32 | |
*** rmart04_ has joined #openstack-kolla | 19:32 | |
*** dwalsh has quit IRC | 19:38 | |
*** rmart04 has joined #openstack-kolla | 19:40 | |
*** rmart04_ has quit IRC | 19:41 | |
*** ravig has quit IRC | 19:45 | |
*** ravig has joined #openstack-kolla | 19:46 | |
*** esharao_ has joined #openstack-kolla | 19:47 | |
*** rmart04_ has joined #openstack-kolla | 19:48 | |
*** rmart04 has quit IRC | 19:49 | |
*** rmart04_ is now known as rmart04 | 19:49 | |
*** erhudy has joined #openstack-kolla | 19:49 | |
*** ravig has quit IRC | 19:50 | |
*** esharao has quit IRC | 19:51 | |
*** mlima_ has quit IRC | 19:54 | |
*** dcwangmit01 has joined #openstack-kolla | 19:54 | |
*** rmart04 has quit IRC | 19:55 | |
*** ravig has joined #openstack-kolla | 19:56 | |
*** ravig has quit IRC | 19:57 | |
*** dcwangmit01 has quit IRC | 19:59 | |
*** dcwangmit01 has joined #openstack-kolla | 19:59 | |
*** rmart04 has joined #openstack-kolla | 19:59 | |
mark-casey | cinerama: sean-k-mooney: this is probably a prereq https://bugs.launchpad.net/bifrost/+bug/1580322 | 19:59 |
openstack | Launchpad bug 1580322 in bifrost "skip_install usage appears inconsistent" [Undecided,New] | 19:59 |
*** unicell has left #openstack-kolla | 20:04 | |
*** rmart04 has quit IRC | 20:04 | |
*** rmart04 has joined #openstack-kolla | 20:05 | |
*** unicell has joined #openstack-kolla | 20:05 | |
*** absubram has quit IRC | 20:06 | |
*** rmart04_ has joined #openstack-kolla | 20:07 | |
*** rmart04 has quit IRC | 20:09 | |
*** rmart04_ is now known as rmart04 | 20:09 | |
*** haplo37 has joined #openstack-kolla | 20:12 | |
*** sacharya_ has joined #openstack-kolla | 20:13 | |
*** rmart04_ has joined #openstack-kolla | 20:14 | |
*** ravig has joined #openstack-kolla | 20:14 | |
*** rmart04 has quit IRC | 20:14 | |
*** rmart04_ is now known as rmart04 | 20:14 | |
cinerama | mark-casey, thanks for filing the bug, i'll have a look soon | 20:15 |
*** ravig has quit IRC | 20:15 | |
*** ravig has joined #openstack-kolla | 20:15 | |
*** sacharya has quit IRC | 20:16 | |
*** ravig has quit IRC | 20:21 | |
*** rmart04 has quit IRC | 20:23 | |
*** ravig has joined #openstack-kolla | 20:25 | |
*** fragatina has quit IRC | 20:27 | |
*** aNupoisc has joined #openstack-kolla | 20:27 | |
aNupoisc | Hi guys I am new to deployment of openstack using kolla and following http://docs.openstack.org/developer/kolla/quickstart.html | 20:29 |
aNupoisc | I am getting "Get https://172.16.2.2:4000/v1/_ping: tls: oversized record received with length 20527" while building container images | 20:29 |
aNupoisc | have anybody come across the same error | 20:30 |
aNupoisc | I am executing ~/kolla/tools/build.py -b ubuntu -t source --registry 172.16.2.2:4000 --no-cache --push for building the container images | 20:30 |
*** absubram has joined #openstack-kolla | 20:33 | |
*** daneyon has joined #openstack-kolla | 20:38 | |
*** jtriley has quit IRC | 20:39 | |
vhosakot | aNupoisc: do you have TLS enabled ? | 20:40 |
vhosakot | aNupoisc: https://github.com/openstack/kolla/blob/master/etc/kolla/globals.yml#L93 | 20:40 |
aNupoisc | Yes TLS is enabled | 20:42 |
aNupoisc | TLSv1.2 | 20:42 |
*** daneyon has quit IRC | 20:43 | |
aNupoisc | vhosakot: is is "no" in globals | 20:44 |
aNupoisc | should it be "yes" | 20:44 |
aNupoisc | but kolla_enable_tls_external: "no" is commented out | 20:44 |
vhosakot | no.. default is "no", which means TLS is disbabled | 20:45 |
aNupoisc | Okay shall i make it "yes"? Will that be a cause | 20:45 |
vhosakot | aNupoisc: TLS is disbaled by default in kolla https://github.com/openstack/kolla/blob/master/ansible/group_vars/all.yml#L220 | 20:46 |
vhosakot | aNupoisc: are you running insecure registry ? | 20:46 |
aNupoisc | vhosakot: can you please tell me where shall i look for insecure registry is running | 20:48 |
aNupoisc | I am not giving option --insecure_registry while running | 20:49 |
aNupoisc | Yes i am giving --insecure_registry in /etc/defualt/docker | 20:50 |
vhosakot | aNupoisc: cool, can you paste the output of "ps -ef | grep insecure" | 20:51 |
vhosakot | aNupoisc: u on Ubuntu ? | 20:52 |
aNupoisc | Yes | 20:53 |
aNupoisc | I got root "1927 1 3 11:32 ? 00:05:23 /usr/bin/docker daemon --dns 8.8.8.8 --dns 8.8.4.4 --insecure-registry 172.16.2.2:5000 --raw-logs | 20:53 |
aNupoisc | root 27998 2506 0 13:52 pts/0 00:00:00 grep --color=auto insecure | 20:53 |
aNupoisc | " | 20:53 |
aNupoisc | oops bad indentation | 20:54 |
aNupoisc | root 1927 1 3 11:32 ? 00:05:23 /usr/bin/docker daemon --dns 8.8.8.8 --dns 8.8.4.4 --insecure-registry 172.16.2.2:5000 --raw-logs | 20:54 |
aNupoisc | root 27998 2506 0 13:52 pts/0 00:00:00 grep --color=auto insecure | 20:54 |
vhosakot | aNupoisc: so, yeah, you running insecure registry.. cool | 20:55 |
vhosakot | aNupoisc: when do you see "Get https://172.16.2.2:4000/v1/_ping: tls: oversized record received with length 20527" ? | 20:55 |
aNupoisc | yes i remember i had made that change in docker file | 20:55 |
aNupoisc | on running build.py with -b ubuntu -t source --registry 172.16.2.2:4000 --no-cache --push | 20:56 |
vhosakot | aNupoisc: what change did you make in which DOckerfile ? can you pastebin it ? | 20:56 |
aNupoisc | is it because here i am specifying port 4000 and in docker file i have given --insecure-registry: "172.16.2.2:5000" | 20:56 |
vhosakot | aNupoisc: which Dockerfile ? I dont think --insecure-registry is needed in any Dockerfile | 20:57 |
aNupoisc | vhosakot: i mean /etc/default/docker | 20:58 |
openstackgerrit | Daneyon Hansen proposed openstack/kolla: WIP: Adds Neutron LBaaS Support https://review.openstack.org/314770 | 20:58 |
vhosakot | aNupoisc: yes, adding --insecure-registry for DOCKER_OPTS in /etc/default/docker is correct | 20:59 |
vhosakot | aNupoisc: can you paste "grep DOCKER_OPTS /etc/default/docker" | 20:59 |
aNupoisc | Yes. | 21:00 |
aNupoisc | vhosakot: DOCKER_OPTS="--dns 8.8.8.8 --dns 8.8.4.4 --insecure-registry 172.16.2.2:5000" | 21:00 |
vhosakot | aNupoisc: cool | 21:00 |
aNupoisc | vhosakot: is it correct? | 21:01 |
vhosakot | aNupoisc: yes.. | 21:01 |
vhosakot | aNupoisc: do you have DOCKER_TLS_VERIFY and DOCKER_CERT_PATH set in your environment any chance ? can you paste the output of "echo $DOCKER_TLS_VERIFY" and "echo $DOCKER_CERT_PATH" | 21:02 |
*** thumpba has quit IRC | 21:02 | |
openstackgerrit | Merged openstack/kolla: Catch all the exception in the push thread https://review.openstack.org/310087 | 21:02 |
aNupoisc | so my doubt is while executing ./tools/build.py -b ubuntu -t source --registry 172.16.2.2:4000, i am giving 4000 as a port | 21:03 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/kolla-kubernetes: Add a memcached service https://review.openstack.org/314772 | 21:03 |
aNupoisc | is that the correct port? | 21:03 |
aNupoisc | vhosakot: yes i will post the paths | 21:03 |
vhosakot | aNupoisc: yes, can you change 5000 to 4000 in DOCKER_OPTS in /etc/default/docker, and restart docker daemon (service restart docker) ? | 21:04 |
*** ravig has quit IRC | 21:04 | |
aNupoisc | vhosakot: no path is given for TLS_VERIFY and DOCKER_CERT_PATH | 21:04 |
vhosakot | aNupoisc: cool, can you change 5000 to 4000 in DOCKER_OPTS in /etc/default/docker, and restart docker daemon (service restart docker) ? | 21:04 |
aNupoisc | vhosakot: Yes of course just a min | 21:05 |
aNupoisc | vhosakot: yes I did restart a service | 21:06 |
aNupoisc | vhosakot: now executing build.py again | 21:07 |
vhosakot | aNupoisc: cool | 21:07 |
aNupoisc | vhosakot: is there a need to specify DOCKER_TLS_VERIFY and DOCKER_CERT_PATH? | 21:07 |
vhosakot | aNupoisc: no | 21:07 |
aNupoisc | also shall i remove all the container and images? | 21:08 |
aNupoisc | vhosakot: or its fine | 21:08 |
vhosakot | aNupoisc: no, they are fine | 21:08 |
*** unicell has quit IRC | 21:14 | |
vhosakot | aNupoisc: did it work ? | 21:16 |
aNupoisc | vhosakot: it is still running | 21:17 |
vhosakot | aNupoisc: cool | 21:17 |
aNupoisc | it takes a bit of time :) | 21:17 |
*** rhallisey has quit IRC | 21:17 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Do not use AUFS when building centos/fedora/oraclelinux container images on ubuntu host https://review.openstack.org/314205 | 21:19 |
*** unicell has joined #openstack-kolla | 21:20 | |
*** unicell has quit IRC | 21:20 | |
*** unicell has joined #openstack-kolla | 21:20 | |
aNupoisc | vhosakot: so generic question, how to decide what port it should use? i mean can it be 6000 instead of 4000? | 21:22 |
openstackgerrit | Hui Kang proposed openstack/kolla: Do not use AUFS for centos/fedora/oraclelinux container images on ubuntu host https://review.openstack.org/314205 | 21:22 |
*** ccesario___ has joined #openstack-kolla | 21:23 | |
aNupoisc | vhosaot: i think so it is running good as no error till now | 21:23 |
aNupoisc | vhosakot: we can specify the port with docker run right? because I had used docker run -d -p 4000:5000 | 21:24 |
aNupoisc | vhosakot: YES it successfully built all the images | 21:24 |
aNupoisc | vhosakot: Thanks a lot | 21:25 |
vhosakot | aNupoisc: cool, so, it was a port mis-match issue.. | 21:25 |
aNupoisc | vhosakot: yeah | 21:25 |
aNupoisc | vhosakot: so how to select the port | 21:25 |
aNupoisc | i mean can i try with 6000? | 21:25 |
vhosakot | aNupoisc: cool.. yes, you can specify the registry port with "docker run" | 21:26 |
vhosakot | aNupoisc: yes, any unused port should be fine | 21:26 |
aNupoisc | vhosakot: okay cool | 21:26 |
aNupoisc | vhosakot: thank you :) | 21:27 |
vhosakot | aNupoisc: cool, np.. glad it worked! :) | 21:29 |
*** mbound has joined #openstack-kolla | 21:30 | |
aNupoisc | vhosakot: yes!! | 21:30 |
*** ravig has joined #openstack-kolla | 21:32 | |
*** neophy has quit IRC | 21:45 | |
*** ccesario___ has quit IRC | 21:48 | |
*** sacharya has joined #openstack-kolla | 21:52 | |
*** sacharya_ has quit IRC | 21:54 | |
*** fragatina has joined #openstack-kolla | 21:57 | |
*** salv-orl_ has quit IRC | 22:07 | |
*** kusr has joined #openstack-kolla | 22:07 | |
kusr | hey, | 22:07 |
*** ccesario___ has joined #openstack-kolla | 22:07 | |
kusr | did you change the kolla documentation page today ? | 22:08 |
kusr | http://docs.openstack.org/developer/kolla/quickstart.html <-- has a few bugs | 22:08 |
*** Avi82 has quit IRC | 22:11 | |
*** Avi82 has joined #openstack-kolla | 22:12 | |
vhosakot | kusr: what bugs ? can you elaborate ? | 22:13 |
kusr | kolla/tools/init-runonce <-- in the end wont run because it needs pip install python-nutreonclient | 22:13 |
*** mark-casey has left #openstack-kolla | 22:15 | |
kusr | https://github.com/openstack/kolla/commit/1c0dcd8ba83aae488e855b371707349451a4e85f <-- this commit fails if you didn't configure the globals.yml properly, but you can't understand why from it | 22:16 |
kusr | the bug report https://bugs.launchpad.net/kolla/+bug/1566571 actually explains in the description what the issue is, but if i didn't read it (while googling the error) i couldn't understand what the error is | 22:16 |
openstack | Launchpad bug 1566571 in kolla mitaka "Add prechecks to validate kolla_internal_vip_address and kolla_external_vip_address" [Medium,Fix committed] - Assigned to Vikram Hosakote (vhosakot) | 22:16 |
kusr | from NoneType is not iterable (output of ansible) | 22:17 |
kusr | nova-compute.conf file referred there in the middle does not exist in latest versions of koola | 22:18 |
vhosakot | yeah, that prechecks validates validate kolla_internal_vip_address and kolla_external_vip_address | 22:18 |
kusr | (and it somehow knew that it should use QEMU automatically) | 22:18 |
kusr | vhosakot: indeed, but it does not give a meaningfull error, and it's easy to misintrepret that step | 22:18 |
kusr | the bug description told me my issue (that i need to put the ip in the range of the netmask) | 22:19 |
kusr | i'm just throwing out stuff i had trying to install today kolla for the first time with that guide | 22:19 |
vhosakot | kusr: the error message is from ansible and not from kolla.. kolla prints whatever ansible returns | 22:20 |
kusr | the how to do other_args for insecure-registry would be nice to have how to do in ubuntu as well (since other steps have both OS) | 22:20 |
vhosakot | kusr: goos point, I will add it | 22:20 |
vhosakot | good* | 22:20 |
*** Avi82 has quit IRC | 22:21 | |
*** esharao_ has quit IRC | 22:21 | |
kusr | vhosakot: dont forget te pip install python-neutreonclient as well | 22:21 |
*** Avi82 has joined #openstack-kolla | 22:21 | |
*** sdake has joined #openstack-kolla | 22:22 | |
kusr | now for the more lame stuff :) | 22:22 |
kusr | the script doesn't explicitly say you should do everything as root | 22:22 |
kusr | if you don't do them as root, it takes quite a few editing / mocking aronud to make it work | 22:22 |
kusr | (or maybe the init-runonce should be moved to use the unified openstack python script) | 22:24 |
*** haplo37 has quit IRC | 22:24 | |
*** berendt has quit IRC | 22:25 | |
kusr | vhosakot: i guess the documentation just lacks in this regard: "Please specify an unused IP address in the network to act as a VIP" isn't specific enough | 22:25 |
kusr | maybe an unused IP address in the same netmask as used by the interface configured in network_interface in the next step | 22:26 |
*** daneyon has joined #openstack-kolla | 22:27 | |
kusr | because as you can see, quite a few people mistake this step | 22:27 |
vhosakot | kusr: yes, good point.. I'll update the docs | 22:27 |
kusr | also... as a first time user | 22:27 |
kusr | i couldn't figure out which user/pass i need to put in the dashboard to use the system (nothing i tried worked) | 22:27 |
kusr | to take it to the next level, i would say maybe do everything as root, or map how to make it work as non root (i think this requires some source modifications) | 22:29 |
*** Avi82 has quit IRC | 22:29 | |
kusr | do you care about those stuff as well or not ? | 22:30 |
kusr | where is the kolla-deploy file / | 22:31 |
sdake | kasr agree on unified openstack python script -that would be great if someone were to do that - although its really a development tool whicfh is why we haven't bothered | 22:31 |
*** daneyon has quit IRC | 22:31 | |
sdake | kusr i agree the VIP thing is confusing, I dont know how to make it more clear :) | 22:31 |
kusr | sdake: i've written a suggestion | 22:32 |
kusr | "unused IP address in the same netmask as used by the interface configured in network_interface in the next step" | 22:32 |
vhosakot | kusr: if adminrc file is soucred, it displays the horizon creds | 22:32 |
kusr | sdake: also maybe get ansible to put a sane error, when it fails the check on it | 22:32 |
sdake | kusr ya I was just reading scrollback | 22:32 |
sdake | I think vhosakot said he would take that into the docs | 22:32 |
vhosakot | yep, I'll update the docs | 22:32 |
kusr | https://github.com/openstack/kolla/blob/cf3e68213edba3e74e74254676085999651a4361/kolla/cmd/build.py#L399 <-- this fails if you do it as non-root | 22:33 |
kusr | since sys.prefix on ubuntu 16.04 is "/usr" but if you did pip install as non root, it goes to "~/.local" | 22:33 |
sdake | build.py doesn't require root on my system | 22:33 |
sdake | but I don't use pip installs | 22:33 |
sdake | sounds like a permissions bug | 22:33 |
kusr | no | 22:33 |
sdake | could you file one? | 22:33 |
kusr | it's a bug in the line i've shown | 22:33 |
kusr | sys.prefix points to the wrong place | 22:34 |
kusr | i had to hardcode it to /home/<user>/.local/ to make it work | 22:34 |
kusr | i dont have a user (unless it's on github) | 22:34 |
vhosakot | kusr: so, the os.path.join failed on your machine as you're not root ? | 22:35 |
kusr | os.path.exists(os.path.join(image_path, 'base')): (a few lines under) because it looked in the wrong place | 22:35 |
kusr | all possible_paths pointed to bad places (on ubuntu 16.04) | 22:35 |
kusr | this is a. non-root and b. not installed pip as root | 22:36 |
kusr | (pip non root installs to ~/.local), but python2 atleast on 16.04 has a bug where os.path.exists doesn't work on ~ as well, so i needed to expand it to full path | 22:37 |
vhosakot | possible_paths is obtained from project's root path or /usr (if pip installed).. so, if it does not work on UBuntu 16, we need to fix it..... can you file a bug ? | 22:37 |
kusr | vhosakot: on github ? | 22:37 |
vhosakot | kusr: no in launchpad | 22:37 |
sdake | launchpad.net/kolla | 22:37 |
kusr | i dont have a launchpad user :| | 22:37 |
vhosakot | https://bugs.launchpad.net/kolla | 22:37 |
kusr | i hate having too many logins i forget abot | 22:38 |
vhosakot | create one.. takes 2 mins :) | 22:38 |
kusr | is it a hassle for you to file it ? :) | 22:38 |
vhosakot | I can file it.. sure.. can you send exact error and distro details ? | 22:38 |
sdake | i dont know what the issue is - everythign works for me as non-root | 22:38 |
kusr | altough i find it hard to belive this isn't an isuse in 14.04 as well | 22:38 |
vhosakot | no, I dont see the issue on 14.04.. I build and deploy everyday | 22:39 |
kusr | sdake: where is your kolla share dir ? | 22:39 |
vhosakot | as non-root | 22:39 |
sdake | kusr not sure, I typically dont pip install | 22:39 |
kusr | do ls ~/.local/share/kolla | 22:39 |
kusr | does that exist ? | 22:39 |
sdake | no that definately doesn't exist | 22:39 |
kusr | so you did pip install root | 22:40 |
kusr | as root | 22:40 |
sdake | no i didn't pip install at all | 22:40 |
kusr | ohh | 22:40 |
kusr | hmm | 22:40 |
sdake | i work from tools directory | 22:40 |
kusr | i followed http://docs.openstack.org/developer/kolla/quickstart.html | 22:40 |
sdake | but if pip install is bustedon ubuntu we should fix it :) | 22:40 |
kusr | where it says "pip install kolla/" in one of the stages | 22:40 |
sdake | kusr ya, I think we need to modify those directions not torecommend pip install for developers | 22:40 |
vhosakot | "pip install" works fine for me Ubuntu 14 | 22:40 |
kusr | vhosakot: pip install as non root ? | 22:41 |
kusr | sdake: could be, i have no idea yet, since i only followed that :0 | 22:41 |
sdake | it may be that sys.perfix on ubuntu 15.z is busted | 22:42 |
kusr | 16.04 LTS is out officaly, yo should try it :0 | 22:42 |
sdake | rather ubuntu 16.z | 22:42 |
sdake | i run centos, others in project run ubuntu | 22:42 |
sdake | provides good balance | 22:42 |
sdake | but nobody is running 16.04 yet | 22:42 |
sdake | but i think should be - if rhel8 were out id be on it from day 1 :) | 22:43 |
kusr | i wanted to try OS quickly, so i installed it on 16.04 | 22:43 |
sbezverk_ | sdake I tried 16.04 but I always use sudo | 22:43 |
kusr | there was another file i needed to edit, le tme remember | 22:43 |
sdake | kusr so it sounds like you worked around that problem | 22:43 |
kusr | sdake: yeah i modified sys.prefix in that script | 22:43 |
kusr | to point to /home/<dir>/.local/ | 22:43 |
kusr | and it worked | 22:43 |
sdake | kusr i see, a local pip install | 22:43 |
kusr | yep | 22:44 |
sdake | pip install -e ? | 22:44 |
kusr | nop, just pip install as non root in 16.04 | 22:44 |
kusr | it auto puts it there | 22:44 |
sdake | wierd, that doesn't work on centos ;/ | 22:44 |
kusr | yeah it's new in ubuntu | 22:44 |
kusr | a few years i think | 22:44 |
kusr | btw, this is from my mind, from what i recall from office :) | 22:45 |
sdake | so after you put in a workaround, did that work for you? | 22:45 |
kusr | http://askubuntu.com/questions/14535/whats-the-local-folder-for-in-my-home-directory | 22:45 |
sdake | its not actually aw orkaround, its a correct solution to the bug | 22:45 |
kusr | sdake: that part yeah | 22:45 |
kusr | also i think /etc/kolla doesn't have correct permissions for writing as non-root (altough i don't have a good solution for that) | 22:46 |
sdake | your right, it doesn't | 22:46 |
sdake | I have been trhinking about handling tyhe non-root case more and more | 22:46 |
sdake | and think an ansible script that sets permissions/owners run as root can solve that | 22:46 |
kusr | https://github.com/openstack/kolla/blob/cf3e68213edba3e74e74254676085999651a4361/tools/kolla-ansible <-- has a bug with non-root as well | 22:46 |
sdake | so atleast deployment isn't run each time as root | 22:46 |
sdake | but only the permission setting | 22:47 |
kusr | BASEDIR doesn't look for ~/.local as wlel | 22:47 |
kusr | so i had to hard cod ethat as wlel | 22:47 |
*** ravig has quit IRC | 22:47 | |
kusr | than everything worked besides the /tc | 22:47 |
sdake | whats /tc? | 22:47 |
kusr | the /etc | 22:47 |
sdake | oh - so you could make /etc group owned by docker | 22:47 |
kusr | but since ansible and everything was installed to ~/.local (cause 16.04 comes with ansible 2 only) | 22:47 |
sdake | and read/writeable | 22:48 |
sdake | you were using ansible 1.9.4 correct? | 22:48 |
kusr | i had to a. hardcode that into BASEDIR in that script, and b. sudo env "path=$PATH:/home/<user>/.local/bin" kolla-ansible to make the deployment work :) | 22:48 |
kusr | ype | 22:48 |
kusr | yep | 22:48 |
kusr | manually installed via pip install to ~/.local | 22:48 |
sdake | cool | 22:49 |
kusr | after all of this stuff, everything ran :0 | 22:49 |
sdake | so sounds like you outsmarted mark shuttleworth :) | 22:49 |
sdake | nice job :) | 22:49 |
kusr | i'm using linux abit longer than him | 22:49 |
kusr | since kernel 0.7 | 22:49 |
vhosakot | nice | 22:49 |
*** ccesario___ has quit IRC | 22:50 | |
sdake | ya - been since 0.99 for me, i'm a bit of a get off my lawn type ;) | 22:50 |
kusr | nice | 22:50 |
kusr | i have a few questions (OS noob) | 22:50 |
sdake | most devs kin kolla are younger however ;) | 22:50 |
kusr | a. does kolla provide a high-availability solution for OS ? | 22:50 |
kusr | b. whats the best way on going to a non-toy AIO installation ? | 22:51 |
sdake | it works using keepalived and haproxy | 22:51 |
sdake | so sort of | 22:51 |
kusr | and c. does it support ironic for bare metal provisinoning | 22:51 |
sdake | in proper HA, you 1. identify fault 2. isolate fault 3. correct fault, 4. ntoify of failure | 22:51 |
sdake | all kolla does is prevent the fault from sort of happening via redundancy | 22:51 |
sdake | in other words its not perfect - that covers A | 22:51 |
kusr | does it do auto replication of the important stuff ? | 22:52 |
sdake | B / multinode - edit the multinode inventory file and you shyould be godo to go - on each node you need docker and docker-py only | 22:52 |
sdake | c. no, however sean-k-mooney is working on that in the community | 22:52 |
sdake | so just t o clarify c, do you mean can kolla bare metal provision your machines? | 22:53 |
kusr | i have stuff i want to run without VM, i would rather it would still live network wise, etc... in the OS eco system | 22:54 |
kusr | so i would want it to PXE boot a machine, and throw an image on top of it | 22:55 |
sdake | ya - we have an implementation does it work, no idea | 22:56 |
kusr | k, might be looking into that | 22:56 |
vhosakot | kusr: didn't know Ubuntu 16 comes with Ansible 2.. so, did you deploy kolla with Ansible 2, or you downgraded to Ansible 1.9.4 ? | 22:57 |
kusr | i manually pip install 1.9.4 as non-root | 22:57 |
vhosakot | cool | 22:57 |
kusr | as the guide said :) | 22:57 |
vhosakot | cool | 22:57 |
kusr | is there any documentation on kolla HA model compard to other ones ? | 22:58 |
kusr | is everything replicated between the state redunded components ? | 22:58 |
sdake | there is no coparison of kolla to other tools that has been prepared by the kolla community | 22:58 |
sdake | I suspect other proprietary software companies have produced an eval doc, but I have not seen it and it is speculatino that they have | 22:58 |
sdake | yes nway active | 22:59 |
kusr | cool | 22:59 |
sdake | where n is your number of nodes | 22:59 |
kusr | so i guess next step is to read about the non AIO mode | 22:59 |
sdake | nubmer of control nodes | 22:59 |
kusr | ohh.. .also please put more info on how to login to the dashboard in the guide | 22:59 |
sdake | ya - our docs need love | 22:59 |
sdake | the problem is people don't file bugs | 23:00 |
sdake | so we lose track of all this ir c conversation | 23:00 |
kusr | can i change 'n' after first deployment and it will adeapt ? | 23:00 |
sdake | and i have stomach flu today - so pretty much i'm barely alive ;) | 23:00 |
kusr | cause people only know how to use githb this days, most projects i contribute to moved their whole systems to github | 23:00 |
sdake | kusr i believe the answer is yes | 23:01 |
kusr | sdake: you can do active-active redundency with vhosakot :) | 23:01 |
kusr | it's nice, took me about 3 hours of messing with this to get the AIO working on 16.04 | 23:01 |
sdake | ya - we find its 1-3 hours for most people from first build to deploy | 23:01 |
sdake | compared to other tools you would still be on the "downloading rpms step" :) | 23:01 |
kusr | devstack was faster, but | 23:02 |
*** ravig has joined #openstack-kolla | 23:02 | |
kusr | i couldn't figure out how to make it non-toy or work after main machine is down (how to heal it) | 23:02 |
vhosakot | haha... we all here in the kolla community try to do as much active-active redundancy as possible to answer/fix kolla questions/bugs :) | 23:02 |
kusr | last question for now | 23:02 |
kusr | how do i get kubernetes after i setup OS ? | 23:02 |
sdake | we haven't implemented any of the k8s underlay at moment | 23:03 |
sdake | but one way is via magnum | 23:03 |
sdake | but our magnum implementation is broken | 23:03 |
sdake | so pretty much, your on your own atm with that | 23:03 |
sdake | but magnum is in progress | 23:03 |
kusr | does kubernetes have a hidden gem like kolla ? | 23:03 |
kusr | (maybe unrelated to OS?) | 23:03 |
vhosakot | kusr: magnum needs neutron-lbaas.. and there is a WIP is kolla review queue to add neutron-lbaas to kolla... you could refer/use it https://review.openstack.org/#/c/314770/ | 23:04 |
patchbot | vhosakot: patch 314770 - kolla (stable/mitaka) - WIP: Adds Neutron LBaaS Support | 23:04 |
kusr | also kolla needs much more visibility imho, i looked for something like this / asked in #openstack and nobody gave me this solution | 23:04 |
kusr | i found it by some dude in a conference who told me it's easy with this | 23:04 |
*** ravig has quit IRC | 23:07 | |
kusr | vhosakot: looks nice, based on your experience what is the timeframe for magnum support ? | 23:08 |
kusr | (officaly) | 23:09 |
vhosakot | kusr: that WIP is not functional yet but is a great first patch set... once neutron-lbaas works, magnum can be added... the magnum config files we have currently in kolla are not completely correct, and hence a magnum bay cannot be created successfully.... I'm planning to fix it if I have time... can't give a date tho :) | 23:10 |
*** ravig has joined #openstack-kolla | 23:11 | |
*** weiyu has joined #openstack-kolla | 23:11 | |
vhosakot | kusr: here is another patch set submitted by another dev to fix some of the magnum configs --> https://review.openstack.org/#/c/300747/ | 23:11 |
patchbot | vhosakot: patch 300747 - kolla - [WIP] Fix the Magnum failed to create trustee | 23:11 |
*** ravig has quit IRC | 23:12 | |
kusr | thanks for the pointers | 23:14 |
vhosakot | kusr: cool | 23:14 |
kusr | gtg to sleep | 23:14 |
kusr | sorry for being scattered with my reports | 23:15 |
vhosakot | kusr: cool, np :) I will update the docs and prechecks... | 23:15 |
*** weiyu has quit IRC | 23:16 | |
sdake | kusr well we are working on visibility, we went from 0 % ro 4% of deployments in the last 6 months | 23:16 |
sdake | so people inside the openstack ecosystem that fill our user surveys and stuff are aware of kolla | 23:16 |
*** salv-orlando has joined #openstack-kolla | 23:16 | |
kusr | sdake: i asked in #openstack for an easy solution, but nobody pointed me a few weeks ago | 23:16 |
sdake | kusr but the general population looking for a cloud solution - hard for us to compete with mirantis scorched earth advertising policies :) | 23:16 |
kusr | i understand, but i think just like the OS model wins | 23:17 |
sdake | #openstack is pretty worthless imo :) | 23:17 |
kusr | so will this | 23:17 |
kusr | but most people inrested in OS will look there | 23:17 |
sdake | I agree, that is why we went from 0% to 4% | 23:17 |
kusr | the only reason i came here, is because i already knew kola existed | 23:17 |
sdake | from nothing to 4% of deplloyments in 6 months is HUGED | 23:17 |
sdake | ya we haven't figured out how to solve the problem of people not knowing how something doesn't exist - specifically in this case | 23:18 |
sdake | kolla | 23:18 |
kusr | vhosakot: if you try to do a non-root installation in 16.04 you'll quickly see all the issues :) | 23:18 |
sdake | I'm not sure nayone has - other then advertising | 23:18 |
kusr | try it on a VM like i did | 23:18 |
sdake | which obviously we aren't goint do so ;) | 23:18 |
vhosakot | kusr: cool, will try it... non-root on 14 is good for me | 23:18 |
kusr | put it in #openstack topic :) | 23:18 |
*** godleon has quit IRC | 23:19 | |
kusr | vhosakot: do it on a clean machine with onthing installed as pip-root | 23:19 |
kusr | that might cover allth eissues | 23:19 |
sdake | there are like 6 deployment tools for openstack | 23:19 |
kusr | sdake: free / easy / non-toy ? :) | 23:19 |
sdake | which one gets to put their stuff in there :) | 23:19 |
sdake | easy | 23:19 |
sdake | hmm | 23:19 |
openstackgerrit | Daneyon Hansen proposed openstack/kolla: WIP: Updates Magnum for Mitaka https://review.openstack.org/314805 | 23:19 |
kusr | do round-robin | 23:19 |
sdake | kolla wins I think on that one :) | 23:19 |
vhosakot | kusr: another magnum PS in queue :) ^^^ | 23:20 |
kusr | exciting ! | 23:20 |
kusr | i have an vacation till next week | 23:20 |
kusr | maybe by than i can test it :) | 23:20 |
*** salv-orlando has quit IRC | 23:21 | |
*** SiRiuS__ has quit IRC | 23:21 | |
kusr | ohh, another Q before i'm killed | 23:21 |
kusr | the SQL DB let's say, is saved in the container | 23:22 |
kusr | or in an outside volume ? | 23:22 |
*** ravig has joined #openstack-kolla | 23:22 | |
*** diogogmt has quit IRC | 23:23 | |
kusr | ohh... another issue i remember.. rabbitmq didn't install until i put the hostname pointing to the interface IP in /etc/hosts (again, the error was kind obscure) | 23:23 |
erhudy | epmd hostname resolution issues can be tricksy | 23:24 |
kusr | yeah, i'm just talking about my experience installing kolla / OS for the first time where i found the error reports to be lacking :) | 23:24 |
*** weiyu has joined #openstack-kolla | 23:24 | |
*** ravig has quit IRC | 23:25 | |
kusr | googling + intuition helped allot, but maybe other people will be more stumpted | 23:25 |
kusr | ok, must go, night | 23:26 |
*** absubram has quit IRC | 23:26 | |
*** jtriley has joined #openstack-kolla | 23:28 | |
aNupoisc | is there some automated script to install docker and few other stuff on target nodes when deploying kolla on multinode | 23:28 |
*** kusr has quit IRC | 23:30 | |
*** sdake has quit IRC | 23:30 | |
*** weiyu has quit IRC | 23:33 | |
*** sacharya has quit IRC | 23:34 | |
*** sacharya has joined #openstack-kolla | 23:35 | |
*** vhosakot has quit IRC | 23:37 | |
*** sdake has joined #openstack-kolla | 23:38 | |
*** signed8bit is now known as signed8bit_Zzz | 23:40 | |
*** mark-casey has joined #openstack-kolla | 23:47 | |
*** vhosakot has joined #openstack-kolla | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!