*** salv-orl_ has joined #kolla | 00:06 | |
*** salv-orlando has quit IRC | 00:09 | |
*** jasonsb has joined #kolla | 00:12 | |
*** asalkeld has joined #kolla | 00:13 | |
*** tzn has joined #kolla | 00:18 | |
*** suro-patz has quit IRC | 00:35 | |
*** tzn has quit IRC | 01:09 | |
*** salv-orl_ has quit IRC | 01:11 | |
*** salv-orlando has joined #kolla | 01:12 | |
*** tzn has joined #kolla | 01:26 | |
*** suro-patz has joined #kolla | 01:39 | |
*** tzn has quit IRC | 01:41 | |
*** gtt116__ has joined #kolla | 01:44 | |
*** gtt116_ has quit IRC | 01:44 | |
*** tzn has joined #kolla | 01:53 | |
*** tzn has quit IRC | 01:58 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Bump versions of os-client-config https://review.openstack.org/263056 | 02:20 |
---|---|---|
openstackgerrit | Sam Yaple proposed openstack/kolla: Bump versions of ansible in kolla-ansible https://review.openstack.org/263086 | 02:20 |
*** Allen_Gao has joined #kolla | 02:22 | |
mandre | happy new year \o/ | 02:29 |
mandre | so many things got merged during end year vacations | 02:32 |
mandre | you guys never rest? | 02:32 |
SamYaple | nihilifer: catching up on the issue | 02:35 |
SamYaple | nihilifer: are you saying kolla-ansible the container? or kolla-ansible the playbooks | 02:35 |
SamYaple | i sounds like you are attempting to reuse the ansible playbooks we have to deploy mesos | 02:36 |
SamYaple | mandre: happy new year! | 02:36 |
SamYaple | mandre: i do not rest, no. | 02:36 |
mandre | SamYaple, indeed | 02:44 |
mandre | so we're not using the ansible docker module anymore? | 02:45 |
*** sdake has joined #kolla | 02:47 | |
SamYaple | correct | 02:48 |
SamYaple | all the patches have merged :D | 02:48 |
SamYaple | docker is unblocked! | 02:48 |
SamYaple | mandre: gate test to prove it https://review.openstack.org/#/c/261601/ | 02:49 |
*** sdake_ has joined #kolla | 02:49 | |
mandre | SamYaple: congrats | 02:50 |
mandre | I'm going to remove the version pinning in vagrant | 02:51 |
SamYaple | cool | 02:51 |
openstackgerrit | Merged openstack/kolla: Unblock Docker pinning in gate https://review.openstack.org/261601 | 02:51 |
SamYaple | docker 1.10 will also allow us to do neutron thin containers (finally) | 02:51 |
SamYaple | thats due feburary | 02:51 |
*** sdake has quit IRC | 02:52 | |
SamYaple | so im going to get the reviews up early and have the team test with the dev version of docker | 02:52 |
mandre | all great news | 02:53 |
sdake_ | evening folks | 02:53 |
SamYaple | hey sdake_ | 02:53 |
SamYaple | long time no see | 02:54 |
sdake_ | ya we have been on a 2 week shutdown | 02:54 |
sdake_ | i'm back in action now htough | 02:54 |
sdake_ | and teeth are all in good health | 02:54 |
SamYaple | awesome! | 02:54 |
sdake_ | we are going to be moving fast now ;) | 02:54 |
SamYaple | just was telling mandre we got kolla_docker all in place and converted | 02:54 |
SamYaple | we are officially unblocked | 02:54 |
sdake_ | nice! | 02:54 |
sdake_ | does it work well? | 02:54 |
sdake_ | I haven' ttried it | 02:54 |
sdake_ | been in midst of remodel of house | 02:54 |
SamYaple | i mean its not worse. i havent had any issues | 02:55 |
SamYaple | its a heck of alot more readable | 02:55 |
sdake_ | cool and it works with any version of docker? | 02:55 |
sdake_ | did itmerge, or does it still need reviews? | 02:55 |
SamYaple | yes and no | 02:55 |
SamYaple | its all merged, kolla is converted | 02:55 |
sdake_ | nice | 02:55 |
SamYaple | it will work with any, but we are requiring docker 1.9 now to use named volumes | 02:55 |
SamYaple | so its only tested with 1.8.2 and 1.9 | 02:56 |
sdake_ | nice | 02:56 |
SamYaple | im not planning on keeping it working with all versions of docker-py and docker, just the ones we 'support' in kolla | 02:56 |
SamYaple | that should actually keep it from getting out of control like the official ansible one | 02:56 |
openstackgerrit | Martin André proposed openstack/kolla: Remove docker version pinning for vagrant envs https://review.openstack.org/263096 | 02:59 |
sdake_ | sounds good | 03:00 |
*** sdake_ is now known as sdake | 03:00 | |
mandre | hey sdake | 03:02 |
mandre | agreed SamYaple, let's keep the module in a manageable state | 03:02 |
SamYaple | yea mandre it will grow with time, adding and removing, just like the rest (probably a bit more though) | 03:03 |
SamYaple | but the removing is key | 03:03 |
SamYaple | the ansible one is trying to work with all versions of everything forever | 03:03 |
SamYaple | esh | 03:03 |
SamYaple | i mean i dont disagree with them, but it makes everything so obtuse | 03:04 |
SamYaple | mandre: sdake https://review.openstack.org/#/c/262614/ | 03:05 |
SamYaple | thats my new project | 03:05 |
*** hparekh has quit IRC | 03:05 | |
mandre | hey cool, what does ekko mean? | 03:11 |
sdake | you spelled scalable incorrectly :) | 03:13 |
sdake | best wishes with your new project sam | 03:14 |
sdake | hope we dont see less of you in kola world ;) | 03:14 |
*** Allen_Gao has quit IRC | 03:14 | |
sdake | mandre Greek Meaning: The name Ekko is a Greek baby name. In Greek the meaning of the name Ekko is: Sound. | 03:15 |
sdake | there is one possible translation for you;) | 03:15 |
SamYaple | mandre: ekko as in echo as in copy | 03:16 |
sdake | word on the street is docke r1.8.2 is sort o lik ea LTS - meant for production deployments | 03:17 |
mandre | google translate autodetects it's Norwegian for echo | 03:17 |
sdake | so we wil. probably want to support that version | 03:17 |
SamYaple | nah it doesnt have volumes support | 03:17 |
SamYaple | plus to go to thin contaienrs we will actually need to be doing docker 1.10 for mitaka | 03:17 |
SamYaple | and thats not released yet | 03:17 |
SamYaple | thin neutron containers* | 03:17 |
*** bmace has quit IRC | 03:18 | |
sdake | i'm pretty happy we jus tremoved a painful depenendenc from our system | 03:18 |
*** bmace has joined #kolla | 03:18 | |
SamYaple | now we can use registry v1 and v2 again! | 03:18 |
* mandre is out for lunch | 03:19 | |
sdake | i was using registry v1 | 03:19 |
sdake | but ithink an older version of docker | 03:19 |
*** hparekh has joined #kolla | 03:19 | |
sdake | i dont recall | 03:19 |
sdake | brain flush | 03:19 |
SamYaple | docker >=1.9 and ansible docker module == no v1 registry support ever again | 03:19 |
sdake | roger | 03:20 |
sdake | tat sucks | 03:20 |
SamYaple | its a single line fix too | 03:20 |
SamYaple | but they said "v1 is deprecated so no change for that" | 03:20 |
SamYaple | i would agree with them if v2 registry wasnt awful | 03:20 |
SamYaple | its getting better, but it still requires suplicate pushing of layers! | 03:21 |
SamYaple | ridiculous | 03:21 |
sdake | it is a hard problem to solve | 03:21 |
sdake | tidiely | 03:21 |
SamYaple | i did it in one line | 03:21 |
SamYaple | https://review.openstack.org/#/c/262842/3/ansible/library/kolla_docker.py | 03:21 |
sdake | no i mean the duplicate pushing thing | 03:22 |
SamYaple | oh yea. i just think its funny they pushed v2 registry when its clearly not ready | 03:22 |
SamYaple | it *might* be ready by docker 1.10 | 03:22 |
sdake | ya that was a big error in judgement | 03:22 |
sdake | that has caused their company all kinds of harm | 03:22 |
SamYaple | i think it was more about scalability | 03:22 |
SamYaple | v1 registry was killing thier backend | 03:22 |
SamYaple | v2 is actually really good for them, bad for devs | 03:23 |
sdake | buy more hardware | 03:23 |
sdake | probleme solved ;) | 03:23 |
SamYaple | v2 also improves pull for clients | 03:23 |
SamYaple | i would say _most_ people are benifiting. just not the devs | 03:23 |
sdake | its the devs that are pushing docker ;) | 03:24 |
SamYaple | indeed | 03:24 |
sdake | so their main audience, is being shat upon | 03:24 |
sdake | anywY nice job, we have a good working solution now ;) | 03:25 |
sdake | I'll have to spin it up tomorrow when i get back to work for the year | 03:25 |
SamYaple | sounds good | 03:25 |
sdake | secure boot i sbusted on my server,s o Ihave to reinstall my os | 03:26 |
* sdake growls | 03:26 | |
SamYaple | setting up a PXE boot server is the way to go | 03:26 |
SamYaple | i rekick weekly | 03:26 |
SamYaple | takes about ~30 min | 03:26 |
*** Allen_Gao has joined #kolla | 03:27 | |
*** unicell has quit IRC | 03:28 | |
*** unicell has joined #kolla | 03:28 | |
*** unicell1 has joined #kolla | 03:31 | |
*** unicell has quit IRC | 03:33 | |
*** coolsvap|away is now known as coolsvap | 03:33 | |
*** suro-patz has quit IRC | 03:37 | |
*** Allen_Gao has quit IRC | 03:37 | |
*** unicell has joined #kolla | 03:39 | |
*** unicell1 has quit IRC | 03:41 | |
*** Jeffrey4l_ has joined #kolla | 03:41 | |
*** klint has joined #kolla | 03:46 | |
*** Allen_Gao has joined #kolla | 03:54 | |
*** suro-patz has joined #kolla | 04:03 | |
*** Allen_Gao has quit IRC | 04:08 | |
*** ssurana has joined #kolla | 04:13 | |
*** ssurana has left #kolla | 04:13 | |
*** sdake has quit IRC | 04:14 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-mesos: Updated from global requirements https://review.openstack.org/262897 | 04:15 |
*** Allen_Gao has joined #kolla | 04:25 | |
mandre | SamYaple here? | 04:31 |
*** Allen_Gao has quit IRC | 04:55 | |
*** suro-patz has quit IRC | 04:56 | |
*** suro-patz has joined #kolla | 04:56 | |
*** suro-patz has quit IRC | 04:58 | |
*** diga has joined #kolla | 05:03 | |
*** salv-orl_ has joined #kolla | 05:07 | |
*** salv-orlando has quit IRC | 05:07 | |
*** Allen_Gao has joined #kolla | 05:09 | |
nihilifer | SamYaple: i meant kolla-ansible as ansible/ dir of kolla, sorry for confusion | 05:36 |
nihilifer | and yea, i'd like to reuse some stuff from it to deploy mesos by ansible | 05:36 |
nihilifer | mostly kolla_docker, of course | 05:37 |
*** Jeffrey4l_ has quit IRC | 05:37 | |
*** Jeffrey4l_ has joined #kolla | 05:50 | |
nihilifer | mandre: hey, if you want to touch vagrant stuff today, can you also disable selinux somehow in bootstrap scripts? | 05:52 |
nihilifer | we recently had some people here having problem with this | 05:53 |
nihilifer | sed -i -r "s,^SELINUX=.+$,SELINUX=permissive," /etc/selinux/config | 05:53 |
nihilifer | should be fine | 05:53 |
nihilifer | or i can do this, but much later, because now i'm digged in the different stuff ;) | 05:54 |
*** suro-patz has joined #kolla | 05:54 | |
nihilifer | 'n | 05:54 |
* nihilifer feels sorry because of making Dan Walsh sad, but our Docker volumes forces us to do this | 05:55 | |
mandre | hey nihilifer, sure I can do it | 05:57 |
mandre | nihilifer, do we have a bug for the selinux issue? | 06:03 |
nihilifer | probably nope | 06:03 |
*** maestro has joined #kolla | 06:04 | |
mandre | i'll fill one quickly | 06:05 |
mandre | what kind of issues with it, do you recall? | 06:06 |
nihilifer | no rights to write on volume | 06:06 |
*** salv-orlando has joined #kolla | 06:06 | |
mandre | thx | 06:06 |
nihilifer | because selinux needs svirt_sandbox_file_t flag on directory mounted as a dir, to not prevent access to it | 06:07 |
*** salv-orl_ has quit IRC | 06:09 | |
nihilifer | well, i registered some blueprint about *enabling* selinux, but it seems to be hard now, almost impossible - to use selinux properly, imo we should figure how to ensure that i.e. ceph container, which should be authorized to r/w to /dev. is really a ceph container and only it's allowed | 06:09 |
nihilifer | probably in some future version of docker there will be a mechanism for hashing container images | 06:10 |
nihilifer | that would be halpful | 06:10 |
nihilifer | helpful* | 06:11 |
mandre | yeah the right thing is to disable selinux in vagrant | 06:12 |
nihilifer | yep, no better options for now | 06:12 |
mandre | i don't even see the point of having it enabled in a dev environment | 06:12 |
nihilifer | if selinux would work for us, maybe for being sure that it really works even in dev environment | 06:13 |
nihilifer | but well, that may happen in some far far future | 06:13 |
nihilifer | probably not in mitaka and even the next release ;) | 06:14 |
nihilifer | hmm... s/probably/surely :P | 06:14 |
mandre | meaning probably never in computer time :) | 06:15 |
openstackgerrit | Martin André proposed openstack/kolla: Set selinux to permissive in Vagrant https://review.openstack.org/263138 | 06:28 |
mandre | nihilifer, ^ | 06:28 |
openstackgerrit | Merged openstack/kolla: Documentation improvements. https://review.openstack.org/263055 | 06:42 |
*** suro-patz has quit IRC | 06:47 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 06:51 |
*** suro-patz has joined #kolla | 06:57 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 07:00 |
openstackgerrit | Merged openstack/kolla: Remove docker version pinning for vagrant envs https://review.openstack.org/263096 | 07:03 |
*** Jeffrey4l_ has quit IRC | 07:11 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 07:11 |
openstackgerrit | Merged openstack/kolla: Support v1 and v2 registry - kolla_docker https://review.openstack.org/262842 | 07:13 |
*** suro-patz has quit IRC | 07:16 | |
*** suro-patz has joined #kolla | 07:17 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 07:18 |
*** reddy has joined #kolla | 07:18 | |
*** suro-patz has quit IRC | 07:21 | |
*** Jeffrey4l_ has joined #kolla | 07:29 | |
*** salv-orlando has quit IRC | 07:34 | |
*** salv-orlando has joined #kolla | 07:35 | |
*** reddy has quit IRC | 07:39 | |
*** salv-orlando has quit IRC | 07:42 | |
*** salv-orlando has joined #kolla | 07:42 | |
openstackgerrit | hardik proposed openstack/kolla: Ansible files for mistral containers. https://review.openstack.org/261876 | 07:48 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the docker may return a empty dict when pull images https://review.openstack.org/263037 | 07:55 |
*** Jeffrey4l_ has quit IRC | 07:59 | |
*** Allen_Gao has quit IRC | 08:00 | |
*** Jeffrey4l_ has joined #kolla | 08:12 | |
*** Allen_Gao has joined #kolla | 08:14 | |
*** chandankumar has joined #kolla | 08:18 | |
*** godleon has joined #kolla | 08:25 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the docker may return a empty dict when pull images https://review.openstack.org/263037 | 08:31 |
*** Jeffrey4l__ has joined #kolla | 08:32 | |
*** Jeffrey4l_ has quit IRC | 08:33 | |
*** salv-orlando has quit IRC | 08:34 | |
*** salv-orlando has joined #kolla | 08:35 | |
*** akwasnie has joined #kolla | 08:39 | |
*** shardy has joined #kolla | 08:44 | |
*** maestro has quit IRC | 08:44 | |
*** godleon has quit IRC | 09:04 | |
*** egonzalez has joined #kolla | 09:04 | |
*** sdake has joined #kolla | 09:08 | |
*** maestro has joined #kolla | 09:08 | |
*** maestro has left #kolla | 09:12 | |
*** gfidente has joined #kolla | 09:38 | |
*** gfidente has joined #kolla | 09:38 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 09:43 |
*** diga_ has joined #kolla | 09:45 | |
*** diga has quit IRC | 09:46 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 09:46 |
*** pbourke has joined #kolla | 09:48 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: [WIP] Vagrant multinode deployment https://review.openstack.org/262242 | 09:51 |
*** jmccarthy has joined #kolla | 09:54 | |
*** Jeffrey4l__ has quit IRC | 09:58 | |
*** openstackgerrit has quit IRC | 10:02 | |
*** openstackgerrit has joined #kolla | 10:02 | |
*** coolsvap is now known as coolsvap|away | 10:04 | |
*** Jeffrey4l__ has joined #kolla | 10:11 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 10:11 |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 10:28 |
*** klint has quit IRC | 10:30 | |
openstackgerrit | Artur Zarzycki proposed openstack/kolla: Horizon ubuntu binary container https://review.openstack.org/263205 | 10:40 |
*** klint has joined #kolla | 10:46 | |
SamYaple | mandre: im here now | 10:48 |
openstackgerrit | Artur Zarzycki proposed openstack/kolla: Horizon ubuntu binary container https://review.openstack.org/263205 | 10:52 |
*** tzn has joined #kolla | 10:54 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Update kolla_docker documentation https://review.openstack.org/262844 | 10:58 |
sdake | is michal rostecki around? | 10:59 |
sdake | morning folks ;) | 11:00 |
sdake | nihilifer ping | 11:00 |
sdake | how close is https://review.openstack.org/#/c/236198/5 to being completed? | 11:01 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the docker may return a empty dict when pull images https://review.openstack.org/263037 | 11:05 |
openstackgerrit | Merged openstack/kolla: Set selinux to permissive in Vagrant https://review.openstack.org/263138 | 11:08 |
nihilifer | sdake: i'll try to get it done till wednesday | 11:08 |
sdake | nihilifer fantasic - would like to get that fixedasap ;) | 11:08 |
nihilifer | ok. btw, we plan to release mitaka-2 on 21th Jan? | 11:09 |
sdake | nihilifer, I think what we should do is just cut mitaka-1 now | 11:09 |
sdake | and yes release mitaka-2 on schedule | 11:09 |
sdake | and probably add a backup to the tagging process in case i fall ill in the future ;) | 11:10 |
sdake | SamYaple up for taking on that responsibiliity? | 11:10 |
nihilifer | sdake: should i move this but to mitaka-2 then? | 11:11 |
nihilifer | bug* | 11:11 |
sdake | nihilifer let me handle the moving please | 11:11 |
sdake | i will do today | 11:11 |
nihilifer | ok | 11:11 |
SamYaple | sdake: sure | 11:16 |
SamYaple | hey nihilifer want to talk about the kolla-ansible in mesos? | 11:17 |
sdake | SamYaple cool - I'll handle tagging, but if I'm late, i'll still hndle tagging, if i need to hand off for some medical reason (I have bad eeth for example) I'll let you know via email. ok? | 11:17 |
nihilifer | SamYaple: yes | 11:17 |
SamYaple | sdake: thats fine | 11:17 |
SamYaple | nihilifer: ok so are you going to be needing everything, or just the library/kolla_docker.py? | 11:17 |
nihilifer | just kolla_docker | 11:18 |
sdake | SamYaple your in the kolla-release group now | 11:18 |
sdake | just make your own signing key | 11:19 |
sdake | i am just going to tag mitaka-1 where it is atm | 11:19 |
SamYaple | sdake: roger | 11:19 |
sdake | and we can make an announcement wednesday saying it may or may not work ;) | 11:19 |
SamYaple | nihilifer: hmm. i hate to say it, but a copy may be the right choice here. | 11:19 |
sdake | along with all of the changes using our standard release announcement etherpad process | 11:19 |
SamYaple | nihilifer: if you do not update the one in the kolla-mesos repo, we should be fine. perhaps set it to read-only? | 11:20 |
sdake | nihilifer can you give me a tldr status report on the mesos repo? | 11:20 |
SamYaple | i don't like the idea of involving submodules/subtrees with git | 11:20 |
sdake | yes no submodules please | 11:20 |
sdake | it just overly complilcates the workflow | 11:21 |
* sdake has finall8y convinced SamYaple ;) | 11:21 | |
SamYaple | sdake: im fine with submodules if the whole repo is needed, not a single file | 11:22 |
nihilifer | sdake: mesos repo currently has 1) vagrant for all-in-one mesos installation 2) yaml config files for each service, similar to those from ansible in kolla 3) command line tool pushing configuration to zookeeper and running the containers on mesos+marathon (or mesos+chronos for bootstrap) | 11:22 |
SamYaple | my new repo is merged! | 11:22 |
SamYaple | join me on #openstack-ekko | 11:23 |
nihilifer | sdake: and 4) extended start script, which creates config files from zk | 11:23 |
sdake | the yaml files are a mesos input? | 11:24 |
nihilifer | sdake: currently supported services are mariadb, rabbit, glance, cinder, keystone and horizon | 11:24 |
nihilifer | sdake: not exactly. yaml files define what's pushed to zookeeper | 11:25 |
SamYaple | fyi sdake, mysql+pymysql is the new default across the openstack. pymysql has python3 support so all connection strings should be that | 11:25 |
sdake | so they are an input to the cli tool? | 11:25 |
SamYaple | we switched kolla over a while ago | 11:25 |
sdake | SamYaple thanks I saw that in a review and was curious what that was all about | 11:26 |
nihilifer | sdake: now i'm working on multinode mesos installations by ansible. then i'll continue on brining support of the other services (probably i'll begin from neutron) | 11:26 |
nihilifer | sdake: yes, they're input for cli | 11:26 |
sdake | cool so sounds like its single node atm then with bringup via vagrand? | 11:27 |
sdake | vagrant? | 11:27 |
nihilifer | tes | 11:27 |
nihilifer | yes* | 11:27 |
sdake | hope you can make our wednesday meeting so we can write the release announcement together | 11:27 |
sdake | or someone else with the appropriate understanding of the changes | 11:28 |
nihilifer | sure | 11:28 |
sdake | cool sounds like i plan | 11:28 |
sdake | i need to reinstall my centos distro today because secre boot is busted for some reason | 11:29 |
* sdake groans | 11:29 | |
*** shardy has quit IRC | 11:29 | |
*** shardy has joined #kolla | 11:30 | |
nihilifer | SamYaple: copying over kolla_docker sounds awfull, but yeah, no better choice for now | 11:30 |
sdake | SamYaple would you like to try to make the tag today so i can walk you through any problems you have (so you can learn how to do it) | 11:30 |
nihilifer | SamYaple: besides cloning kolla in vagrant which i'm doing currently in WIP patches i sent:/ | 11:31 |
SamYaple | sdake: ok. is there special openstack things or just git tag/push? | 11:31 |
sdake | there is some special magic | 11:31 |
sdake | bu8t you have tag permission push becasue of kolla-release | 11:31 |
sdake | the magic is documented | 11:31 |
sdake | first step is to create a signing key | 11:31 |
SamYaple | nihilifer: yea the copy isn't great. but if you set your kolla-mesos copy to read-only and -2 any changes in the kolla-mesos branch it shouldn't be too bad | 11:32 |
SamYaple | nihilifer: but i agree, not amazing. might be the cleanest though. that module won't change often (well it might in the next few days, but not long term) | 11:32 |
sdake | ya copy as unideal ss it is, seems like best solution here, as long as we treat the koll repo s the porper upstream for the module | 11:32 |
SamYaple | i want to start the process of splitting kolla-ansible off before mitaka-2 | 11:33 |
sdake | 3 weeks - get cracking ;) | 11:34 |
SamYaple | i can get it merged into project-config in that time, not problem. we need to get everyone on the same page. i was planning that for this meeting | 11:35 |
SamYaple | upcoming meeting* | 11:35 |
SamYaple | pbourke: hey man! | 11:40 |
pbourke | SamYaple: hey happy new year :) | 11:40 |
*** dims has joined #kolla | 11:40 | |
*** dims has quit IRC | 11:40 | |
SamYaple | you as well | 11:41 |
*** dims has joined #kolla | 11:41 | |
sdake | pbourke happy new year dude | 11:41 |
sdake | we are still all alive and its a new year :) | 11:41 |
sdake | my tooth didn't kill me ;) | 11:42 |
pbourke | same to you, good to be back | 11:42 |
sdake | ya over the last 3-4 weeks i feel a bit what it is like to be unemployed | 11:42 |
sdake | i was bored to death | 11:42 |
*** Jeffrey4l__ has quit IRC | 11:43 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 11:45 |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 11:58 |
sdake | SamYaple pbourke re swift | 12:04 |
sdake | and ceph | 12:04 |
sdake | ceph has a rockin auto disk detect mechnaism | 12:04 |
sdake | can we make swift have the same thing? | 12:04 |
SamYaple | sdake: yes and no. ceph has a rockin add-disk-on-the-fly mechanism. swift does not | 12:06 |
SamYaple | so we would be in the business of autogenerating rings | 12:06 |
*** salv-orl_ has joined #kolla | 12:06 | |
sdake | fundamentally the problem i keep hearing over and over with swift and kolla is -etoohard | 12:07 |
*** salv-orl_ has quit IRC | 12:07 | |
sdake | i dont hear that with ceph | 12:07 |
*** salv-orl_ has joined #kolla | 12:07 | |
sdake | would like to solve that - however is best ;) | 12:08 |
*** salv-orlando has quit IRC | 12:09 | |
*** tzn has quit IRC | 12:10 | |
SamYaple | because swift is fundementally a PITA to setup and manage | 12:10 |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 12:10 |
SamYaple | we can take a more active control in swift, but that means limiting what the end user can do with it | 12:11 |
SamYaple | or its just alot of manual work on the enduers part | 12:11 |
sdake | is both an answer that we could deliver on? | 12:18 |
sdake | "easy mode limited config" + "hard mode full customization"\ | 12:18 |
SamYaple | full customization means reimplementing ring config as ansible code | 12:19 |
sdake | right | 12:19 |
sdake | oh | 12:19 |
sdake | no | 12:19 |
sdake | i mean full customization via the operator | 12:20 |
SamYaple | yup | 12:20 |
sdake | sorry 6am ;) | 12:20 |
SamYaple | thats what i meant to | 12:20 |
SamYaple | either we lock down the ring generation, or we reimplement it | 12:20 |
sdake | i mean the operator is responsible for generating ring configs int he had mode config | 12:20 |
SamYaple | then we cant autoadd disks | 12:20 |
sdake | had/hard | 12:20 |
SamYaple | and the ceph disk code doesnt matter | 12:21 |
sdake | ok, well this is probably8 agood topic for the midcycle | 12:21 |
SamYaple | indeed. swift is annoying | 12:21 |
sdake | its a bear to ssetup | 12:22 |
sdake | but people like it because its dep free | 12:22 |
SamYaple | ok im heading into the office. afk | 12:22 |
sdake | enjoy | 12:28 |
sdake | hey guys zuckerberg is giong to implement AI | 12:28 |
sdake | in his spare time | 12:29 |
sdake | while two generations of maniacal PHD+ brainpower has failed | 12:29 |
openstackgerrit | Merged openstack/kolla: Fix the docker may return a empty dict when pull images https://review.openstack.org/263037 | 12:30 |
pbourke | his last implementation it turned out had humans sitting typing responses | 12:30 |
sdake | haha | 12:31 |
sdake | funny :) | 12:31 |
sdake | fwiw, I dont use facebook ;) | 12:32 |
*** tzn has joined #kolla | 12:32 | |
nihilifer | lucky you | 12:32 |
ajafo | re | 12:36 |
pbourke | are insecure registries supported by the new kolla_docker module? | 12:38 |
*** sdake has quit IRC | 12:39 | |
nihilifer | pbourke: it's probably not the business of kolla_docker module | 12:39 |
nihilifer | afaik, insecure registries can be enabled only on the docker service's config | 12:40 |
*** sdake has joined #kolla | 12:40 | |
nihilifer | in case of redhat based distros - /usr/lib/systemd/system/docker.service | 12:40 |
pbourke | nihilifer: hmm that's true. just trying to figure it out in the context of https://review.openstack.org/#/c/262843/4/etc/kolla/globals.yml | 12:41 |
nihilifer | here you have a solution - https://github.com/openstack/kolla/blob/master/dev/vagrant/bootstrap.sh#L60 | 12:41 |
nihilifer | awww, ok | 12:42 |
nihilifer | hmm... sorry, that insecure_registry option in ansible module was actually for docker-py using http (instead of https), sorry for confusing you | 12:43 |
nihilifer | well, in Sam's patch it seems to be working without anny additional flags | 12:43 |
*** rhallisey has joined #kolla | 12:45 | |
*** reddy has joined #kolla | 12:46 | |
*** diga_ has quit IRC | 12:46 | |
pbourke | I actually can't see how it would work with an insecure reg | 12:46 |
pbourke | but I'll take your word that it does | 12:46 |
rhallisey | morning | 12:49 |
sdake | insecure registry flag happens in the docker.service file imo not the docker.py file | 12:52 |
sdake | rhallisey happy new years! | 12:52 |
rhallisey | hey happy new years! | 12:52 |
rhallisey | sdake, boy you're up early | 12:52 |
*** sdake_ has joined #kolla | 12:53 | |
*** Allen_Gao has quit IRC | 12:56 | |
*** Allen_Gao has joined #kolla | 12:56 | |
*** sdake has quit IRC | 12:57 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 12:59 |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 13:07 |
*** dave-mccowan has joined #kolla | 13:08 | |
*** inc0 has joined #kolla | 13:11 | |
inc0 | good morning everyone | 13:11 |
*** salv-orl_ has quit IRC | 13:13 | |
nihilifer | hey inc0, happy new year | 13:14 |
inc0 | you too | 13:14 |
*** salv-orlando has joined #kolla | 13:14 | |
inc0 | I heard you got hit by winter? | 13:14 |
nihilifer | yes | 13:14 |
nihilifer | -15 degrees | 13:14 |
nihilifer | so after not so cold holidays, is a little bit shocking ;) | 13:16 |
inc0 | yup...well...let's just hope it won't stay this way till April, like it did few years ago | 13:16 |
barthalion | snow in April was awesome | 13:23 |
*** Jeffrey4l__ has joined #kolla | 13:24 | |
inc0 | yeah, and christmas were warm that year, so it was just opposite to what it should be | 13:25 |
SamYaple | pbourke: yes they are | 13:26 |
openstackgerrit | Merged openstack/kolla: Update kolla_docker documentation https://review.openstack.org/262844 | 13:26 |
SamYaple | upstream docker deprecated that option | 13:26 |
SamYaple | hey rhallisey | 13:26 |
SamYaple | pbourke: nihilifer https://github.com/docker/docker-py/blob/master/docker/api/daemon.py#L37-41 | 13:28 |
*** reddy has quit IRC | 13:28 | |
SamYaple | its no longer used on the client side, only set on the daemon to allow which "insecure registries" are ok to connect to | 13:28 |
SamYaple | client cant override anymore hence the deprecation | 13:28 |
SamYaple | morning inc0 | 13:29 |
inc0 | hey SamYaple, so ekko is a thing:) | 13:30 |
SamYaple | it is indeed | 13:30 |
inc0 | cool! At last I get to write some python | 13:32 |
rhallisey | SamYaple, inc0, hey guys happy new year! | 13:35 |
inc0 | you too | 13:35 |
rhallisey | thanks | 13:36 |
openstackgerrit | Merged openstack/kolla: Remove unused params for kolla_docker https://review.openstack.org/262843 | 13:37 |
*** unicell1 has joined #kolla | 13:38 | |
*** reddy has joined #kolla | 13:41 | |
*** unicell has quit IRC | 13:41 | |
*** klint has quit IRC | 13:57 | |
*** gtt116_ has joined #kolla | 14:05 | |
*** dwalsh has joined #kolla | 14:05 | |
*** gtt116__ has quit IRC | 14:08 | |
SamYaple | if I could get some reviewer love on the named volume patch it would be much appreciated (note, must use docker 1.9 and docker-py 1.6) https://review.openstack.org/#/c/261843/ | 14:09 |
inc0 | SamYaple, what part exactly requires sorted? https://review.openstack.org/#/c/248481/9/tools/version_check.py | 14:11 |
*** sdake has joined #kolla | 14:11 | |
SamYaple | inc0: if i remember corerctly (and mandre has changed the logic a bit) but when i did it it used sorted to ensure the version numbers lined up and the latest was successfully choosen | 14:12 |
SamYaple | not the logic itself, but the data coming into the logic | 14:12 |
*** sdake_ has quit IRC | 14:12 | |
SamYaple | beautifulsoup didnt return a deterministic list or something like that | 14:12 |
*** v1k0d3n has joined #kolla | 14:13 | |
*** Kennan has quit IRC | 14:14 | |
SamYaple | either way keeping them sorted is the most readable output. whats your problem with sorted again inc0 ? | 14:14 |
*** Kennan has joined #kolla | 14:15 | |
*** dwalsh has quit IRC | 14:17 | |
inc0 | it's not a problem, just iteritems looks better imho, more pythonic | 14:22 |
inc0 | also there is slight change in sorted logic between 2 and 3 | 14:22 |
inc0 | and that is in 2 it returns list and in 3 it's an iterator | 14:23 |
inc0 | here that won't affect anything, but it might later on | 14:23 |
SamYaple | ah i see. hmm if we go iteritems then well have to have another layer to sort the output | 14:28 |
SamYaple | not ideal | 14:28 |
inc0 | well, anyway it's not bad, it just could be better, I don't think we need to sort output really, but it doesn't hurt | 14:29 |
*** jpeeler has joined #kolla | 14:29 | |
*** jpeeler has joined #kolla | 14:29 | |
SamYaple | since this is a user readable output, i think we do | 14:29 |
SamYaple | when this gets changed to a bot to propose things, we absolutley dont | 14:30 |
inc0 | also, upstream_version is a dict | 14:30 |
inc0 | so it's not sorted | 14:30 |
inc0 | order you put stuff there has completely no effect | 14:31 |
SamYaple | that may have changed. we still want sorted output though. a sorted dict is a great way to ensure that... | 14:31 |
inc0 | you don't use sorted dict there | 14:32 |
nihilifer | you guys mean OrderedDict or what? | 14:33 |
SamYaple | that doesnt affect the output inc0 | 14:33 |
inc0 | yeah, OrderedDict | 14:33 |
SamYaple | nihilifer: we are calling it by sorted(dict()) | 14:33 |
nihilifer | ah, i see it now, sorry | 14:33 |
*** stvnoyes has joined #kolla | 14:34 | |
inc0 | SamYaple, what you do there is sorted(dict()) -> this will produce sorted list of keys of dict, then with that you populate upstream_versions which is just plain dict, and then use this one | 14:34 |
inc0 | and order in which you add stuff to upstream_version has completely no meaning | 14:34 |
*** dwalsh has joined #kolla | 14:34 | |
inc0 | so as it currently looks like, sorted() seems redundant | 14:35 |
inc0 | unless I'm missing somethingt | 14:35 |
SamYaple | inc0: you are missing the fact that this is a for loop and it prints things i think | 14:35 |
SamYaple | this is user output | 14:35 |
nihilifer | i'd propose to change MAJOR_VERSIONS_MAP to OrderedDict | 14:36 |
nihilifer | and then iteritems | 14:36 |
inc0 | yes, it works on upsteam_version and local_versiuon | 14:36 |
inc0 | and again, way you add stuff to it doesn't mean anything | 14:36 |
inc0 | they seem sorted, because that how dicts in python works | 14:36 |
inc0 | but in fact they might not be | 14:36 |
stvnoyes | status | 14:36 |
SamYaple | nihilifer: then we have to get six involoved | 14:38 |
SamYaple | i dont wanna get six involved :( | 14:38 |
inc0 | SamYaple, try to remove sorted and see if it changes anything | 14:38 |
SamYaple | inc0: it does | 14:39 |
SamYaple | because it hashes those keys and then loops through the hashes in order | 14:39 |
nihilifer | SamYaple: you already have six there, so what's the problem? | 14:40 |
SamYaple | oh i didnt realize that nihilifer. like i said mandre took over the patch a while back. I only care about sorted output with the lest complication | 14:40 |
SamYaple | ordereddict seems to be the way to go | 14:41 |
nihilifer | yea | 14:41 |
*** chandankumar has quit IRC | 14:41 | |
inc0 | yeah, make it ordered dict | 14:42 |
*** reddy has quit IRC | 14:48 | |
akwasnie | hi guys, happy new year!:) | 14:49 |
inc0 | hey Ala, hows life? | 14:49 |
SamYaple | happy new year akwasnie! | 14:49 |
akwasnie | Sam, I have one question about kolla_docker in logstash. https://review.openstack.org/#/c/252968/7/ansible/roles/elk/tasks/logstash.yml | 14:50 |
SamYaple | shoot | 14:51 |
akwasnie | Expose and command parameters are required for logstash container but are not supported by kolla_docker. Should we add them to kolla_docker or create Dockerfile for logstash and specify port that should be exposed there? | 14:51 |
SamYaple | akwasnie: i don't think either of those are _required_. The CMD should be in the config.json and the expose is only useful when not using --net=host (and we should be using net=host everywhere) | 14:52 |
*** rmart04 has joined #kolla | 14:53 | |
openstackgerrit | Merged openstack/kolla-mesos: Updated from global requirements https://review.openstack.org/262897 | 14:53 |
*** sdake_ has joined #kolla | 14:55 | |
akwasnie | oh thanks Sam, I'll try it | 14:56 |
*** sdake has quit IRC | 14:58 | |
*** sdake has joined #kolla | 15:00 | |
openstackgerrit | Merged openstack/kolla: Bump versions of os-client-config https://review.openstack.org/263056 | 15:02 |
openstackgerrit | Merged openstack/kolla: Convert to named volumes instead of data container https://review.openstack.org/261843 | 15:02 |
*** sdake_ has quit IRC | 15:03 | |
*** shakamunyi has quit IRC | 15:15 | |
*** barra204 has quit IRC | 15:15 | |
openstackgerrit | Michal Rostecki proposed openstack/kolla: [WIP] Add Dockerfiles for Mesos https://review.openstack.org/261331 | 15:16 |
*** sdake_ has joined #kolla | 15:22 | |
*** sdake has quit IRC | 15:25 | |
*** blahRus has joined #kolla | 15:27 | |
nihilifer | SamYaple: about your latest comments | 15:29 |
SamYaple | the structural thing? | 15:29 |
nihilifer | we already have situation when openstack-base is in some outside directory | 15:29 |
nihilifer | and all the dirs line nova/ etc. are inheriting from that | 15:29 |
SamYaple | ok. thats a fair point | 15:29 |
nihilifer | i think that mesosphere is analogical here | 15:29 |
SamYaple | yea. we did break that rule when doing openstack-base | 15:29 |
*** shakamunyi has joined #kolla | 15:29 | |
SamYaple | youve sold me | 15:29 |
*** egonzalez has quit IRC | 15:32 | |
*** chandankumar has joined #kolla | 15:34 | |
*** sdake has joined #kolla | 15:38 | |
*** sdake_ has quit IRC | 15:39 | |
sdake | nihilifer will you be making it in person to the kolla midcycle? | 15:41 |
*** blahRus1 has joined #kolla | 15:49 | |
*** blahRus has quit IRC | 15:50 | |
*** rahulshr has joined #kolla | 15:50 | |
rahulshr | Hi Guys, I need your suggestion | 15:50 |
SamYaple | rahulshr: ok! | 15:50 |
rahulshr | I have deployed kolla using all in one config | 15:50 |
rahulshr | now I am trying to make it multinode ie to add one compute node | 15:51 |
sdake | nice - what was your impression of the deployment process - just say whatever comes to mind ;) | 15:52 |
rahulshr | as per my understanding I have to deploy the registry and build the images again with that registry | 15:52 |
sdake | rahulshr you need to use a registry when doing multinode | 15:52 |
pbourke | rahulshr: I think you should just need to add the new node to your inventory file and redeploy | 15:52 |
sdake | you can do it without a registry but your life will be much simpler with a registry | 15:52 |
SamYaple | rahulshr: that is the way you should do it, however for testing you can push the images by hand to a second node | 15:52 |
rahulshr | is it possible to do it without building again? | 15:53 |
SamYaple | yes what sdake said | 15:53 |
SamYaple | rahulshr: you can | 15:53 |
sdake | yup you can push to the second node | 15:53 |
rahulshr | ok | 15:53 |
SamYaple | you can `docker import/export` the images | 15:53 |
SamYaple | but again, i don't recommend it. its long. setting up a registry is quick and easy | 15:53 |
rahulshr | how can I push multiple images in one go? | 15:53 |
pbourke | --registry foo:5000 --push | 15:54 |
pbourke | pass above args to your build | 15:54 |
*** cloudnautique has joined #kolla | 15:54 | |
rahulshr | sdake since I was new with kolla/openstack I have to do the deployment couple of time but I hated the kolla-build time :) | 15:55 |
*** cloudnautique has quit IRC | 15:55 | |
sdake | pbourke he means wihtout a registry | 15:55 |
sdake | rahulshr roger on the build time - that is indeed annoying | 15:55 |
sean-k-m1 | you should be able to manually push the currently built images to a regestry also without rebuilding them correct. | 15:55 |
sdake | sean-k-m1 yes you can also do that | 15:55 |
sdake | so docker has an image cache - this is what "docker images" shows you | 15:55 |
*** cloudnautique has joined #kolla | 15:55 | |
sdake | anythign in the image cache can be pushed to a registry | 15:55 |
rahulshr | yes I can ush manually but there are arnd 50 images | 15:56 |
rahulshr | I was thinking if there is some docker cmd to push all in one go | 15:56 |
sdake | rahulshr probably easiest is to cut nad past ethe list and then make a bash script out of it to do the job | 15:56 |
sdake | otherwise you can use awk to do the job | 15:56 |
rahulshr | hmm ok | 15:56 |
sdake | perhaps SamYaple can come up with some simple awk script to do it - for me i'd jsut do docker images and manaully create a script to do it with 50 commands in it | 15:57 |
sdake | but I roll hacky on 1 time jobs ;) | 15:57 |
rahulshr | let me try the script then | 15:58 |
SamYaple | rahulshr: you want to push local images to the regstiry? | 15:59 |
SamYaple | i have a script for that | 15:59 |
SamYaple | well not so much a script. but i do `docker images | awk '/kollaglue/ {print $1}' | xargs -n1 -P1 docker push` | 16:00 |
SamYaple | adjust the value of -P for parallel | 16:00 |
SamYaple | if you have other kollaglue images, you may want to change the match to the registry name like `docker images | awk '/name_of_registry:5000/ {print $1}' | xargs -n1 -P1 docker push` | 16:01 |
rahulshr | Thanks SamYaple | 16:04 |
*** rmart04 has quit IRC | 16:08 | |
*** reddy has joined #kolla | 16:18 | |
*** ajafo has quit IRC | 16:26 | |
*** akwasnie has quit IRC | 16:26 | |
*** coolno1 has joined #kolla | 16:30 | |
coolno1 | hello | 16:30 |
coolno1 | need some help in safely upgrading our open stack setup | 16:30 |
SamYaple | coolno1: what is your question? | 16:31 |
coolno1 | we have got a 2 node setup where each node acts as controller as well as compute node currently | 16:31 |
coolno1 | We need to add another compute node and if this succeeds we will add lot many compute nodes | 16:31 |
coolno1 | how do we go about it without impacting the current setup | 16:32 |
*** ajafo has joined #kolla | 16:32 | |
coolno1 | SamYaple: details as above | 16:32 |
SamYaple | coolno1: is this a kolla set? | 16:32 |
coolno1 | yes corrct | 16:32 |
coolno1 | *correct | 16:32 |
SamYaple | it should be as simple as adding the node to your inventory file and rerunning the playbooks | 16:32 |
SamYaple | if you ahve updated images then its going to try to restart and pull in those changes | 16:33 |
coolno1 | will it bring down the existing setup- by which I mean their are users which are currently using VM's spawned out of openstack setup | 16:33 |
rahulshr | I think you just have to edit a mulitnode file | 16:33 |
SamYaple | coolno1: no it will not, unless you have updated/rebuilt the images in some way | 16:33 |
SamYaple | the playbooks are idempotent, they should only change things if things need to be changed | 16:34 |
coolno1 | SamYaple: Thanks for this info will try it out | 16:34 |
coolno1 | rahulsr: what is multinode file? | 16:34 |
SamYaple | for your info: if this was a fresh vanilla setup, you could just add the compute to the inventory and rerun all the playbooks and only the compute node would have changes | 16:34 |
SamYaple | coolno1: your inventory file | 16:34 |
coolno1 | SamYaple: aah, we do not use the name multinode | 16:35 |
SamYaple | yea its fine | 16:35 |
SamYaple | those inventory files in the repo are examples | 16:35 |
coolno1 | SamYaple: Okay got it, so just add another hostname in our inventory file an rerun kolla deploy | 16:36 |
coolno1 | SamYaple: that should do I guess? | 16:36 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow common role to pull images https://review.openstack.org/262846 | 16:36 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow ceph role to pull images https://review.openstack.org/262847 | 16:36 |
coolno1 | in the [compute] section ofcourse | 16:36 |
SamYaple | coolno1: correct. that is all | 16:38 |
SamYaple | again, if the underlying images have changes it may try to restart your services | 16:38 |
coolno1 | SamYaple: I have a local registry | 16:39 |
SamYaple | but if they havent youll be fine (have you rebuilt the images?) | 16:39 |
coolno1 | SamYaple: No I haven't, I have got a docker registry which host all the images and those are intact(have not been touched) | 16:40 |
coolno1 | SamYaple: I hope this should be safe? | 16:40 |
SamYaple | coolno1: yes it should be fine then | 16:41 |
SamYaple | if it is not, please keep all the info and logs (and ansible output) | 16:41 |
SamYaple | it is also safe to stop ansible while running if you see a problem | 16:41 |
SamYaple | ^C | 16:41 |
coolno1 | stop ansible? | 16:42 |
coolno1 | SamYaple: can you please elaborate? | 16:42 |
SamYaple | if you run the playbooks and see a problem (restarting of your contaienrs for example) you can stop ansible mid-run safely | 16:42 |
SamYaple | and rerun it later | 16:42 |
coolno1 | SamYaple: aah okay, but starting container does not mean my users VM will be impacted? | 16:43 |
coolno1 | as compute node does not use any containers | 16:43 |
SamYaple | correct, its a new compute node | 16:43 |
coolno1 | may be only for sometime the VM's may be inaccessible (for example say neutron containers are restarted) | 16:44 |
coolno1 | and hence networking becomes unavailable | 16:44 |
SamYaple | no you should not be impacted at all because only the new compute node should have anything that has changed on it | 16:44 |
SamYaple | but if containers do get restarted, then yes you will be impacted as you describe | 16:45 |
SamYaple | i am stepping away from lunch. please let me know if you have any issues coolno1 | 16:45 |
coolno1 | SamYaple: Okay so looks like it is safe to add another compute node, I will try this later tonight and will probably get back to you tomorrow | 16:45 |
SamYaple | also let me know if you dont have any issues! | 16:45 |
SamYaple | sounds good | 16:45 |
coolno1 | SamYaple: sure will let you know | 16:46 |
coolno1 | thanks a lot | 16:46 |
*** unicell1 has quit IRC | 16:47 | |
*** unicell has joined #kolla | 16:47 | |
nihilifer | sdake: yes, in 99% i'll be there in person | 16:50 |
*** diga has joined #kolla | 16:50 | |
nihilifer | let me confirm that later this week | 16:50 |
*** barra204 has joined #kolla | 16:51 | |
*** shakamunyi has quit IRC | 16:51 | |
*** sacharya has joined #kolla | 16:51 | |
*** unicell has quit IRC | 16:52 | |
pbourke | nihilifer: what does kolla-mesos-deploy assume to be on the deployment host? | 16:52 |
*** ssurana has joined #kolla | 16:55 | |
nihilifer | pbourke: zookeeper and mesos+marathon | 16:55 |
pbourke | nihilifer: k that's what I thought, thanks | 16:56 |
nihilifer | np | 16:56 |
*** diga has quit IRC | 16:57 | |
*** nihilifer has quit IRC | 17:00 | |
SamYaple | pbourke: you coming to the midcycle? | 17:02 |
pbourke | SamYaple: no | 17:02 |
SamYaple | nooooo | 17:02 |
*** coolno1 has quit IRC | 17:02 | |
SamYaple | but the beers... | 17:02 |
pbourke | we discussed this before :) | 17:03 |
SamYaple | i know i know | 17:03 |
SamYaple | but you should still try | 17:03 |
*** nihilifer has joined #kolla | 17:04 | |
*** diga has joined #kolla | 17:05 | |
pbourke | Id like to but it's too far to travel for 2 days | 17:05 |
*** diga has quit IRC | 17:05 | |
*** diga has joined #kolla | 17:06 | |
SamYaple | make it a week! | 17:06 |
openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Playbook for keystone upgrade https://review.openstack.org/257568 | 17:07 |
pbourke | stop tempting me | 17:08 |
pbourke | mesos is a pita to install | 17:08 |
*** chandankumar has quit IRC | 17:08 | |
openstackgerrit | Sidharth Surana proposed openstack/kolla: Make galeradb bootstraping robust https://review.openstack.org/254961 | 17:10 |
*** diga_ has joined #kolla | 17:10 | |
*** diga has quit IRC | 17:10 | |
*** diogogmt has joined #kolla | 17:11 | |
*** reddy has quit IRC | 17:12 | |
*** diga_ has quit IRC | 17:15 | |
*** nihilifer has quit IRC | 17:17 | |
*** nihilifer has joined #kolla | 17:18 | |
*** sdake_ has joined #kolla | 17:23 | |
*** chandankumar has joined #kolla | 17:24 | |
*** sdake has quit IRC | 17:25 | |
SamYaple | sdake_: food is fine in the building, trust me. ill bring it in if there is any issue (and there wont be) | 17:31 |
SamYaple | pbourke: tempted | 17:31 |
pbourke | SamYaple: your untapped checkins have been doing the trick | 17:32 |
SamYaple | pbourke: thats mostly from the same place | 17:32 |
pbourke | ahem, but its not kolla not the beer right | 17:32 |
SamYaple | right uhh yeaa | 17:32 |
SamYaple | pbourke: want to work on a backup project? | 17:32 |
SamYaple | got a new one in openstack | 17:33 |
SamYaple | ekko | 17:33 |
SamYaple | its official | 17:33 |
pbourke | yup im watching it in gerrit | 17:33 |
pbourke | ill probably lurk for a while and see if it sparks my interest | 17:33 |
SamYaple | join us in #openstack-ekko | 17:33 |
pbourke | already there | 17:33 |
SamYaple | cool :) | 17:34 |
pbourke | you guys seem to have discussed it previously, any tasks for a newbie? | 17:34 |
SamYaple | pbourke: i talked a bit with inc0 the other day, but nihilifer is as new as you to it | 17:37 |
SamYaple | since they are both going to be at the midcycle for kolla, i was going to carve out some ekko time then | 17:38 |
SamYaple | pbourke: i am more than happy to give the run down for you guys | 17:38 |
*** chandankumar has quit IRC | 17:46 | |
*** dwalsh has quit IRC | 17:49 | |
sdake_ | SamYaple ack, sounds good thanks bro | 17:53 |
*** sdake_ is now known as sdake | 17:56 | |
sdake | SamYaple is the venue the actual servosity office or some startup incubator? | 17:57 |
sdake | I htink you told me in the past but i don't recall atm | 17:57 |
SamYaple | startup indubator would probably be most accurate | 17:57 |
SamYaple | but our office is here | 17:57 |
SamYaple | we have shared confrence rooms | 17:57 |
sdake | cool | 17:57 |
sdake | to get in the building, requires a badge or badgless entry? | 17:58 |
sdake | (this stuff goes int eh wiki) | 17:58 |
SamYaple | go through the first floor during business hours (all hours?) and no badge | 17:58 |
*** suro-patz has joined #kolla | 18:00 | |
*** vinkman has joined #kolla | 18:03 | |
*** vinkman has left #kolla | 18:03 | |
*** salv-orl_ has joined #kolla | 18:06 | |
*** v1k0d3n has quit IRC | 18:06 | |
*** salv-orl_ has quit IRC | 18:06 | |
*** salv-orl_ has joined #kolla | 18:07 | |
*** reddy has joined #kolla | 18:08 | |
*** jasonsb has quit IRC | 18:08 | |
*** salv-orlando has quit IRC | 18:08 | |
*** dwalsh has joined #kolla | 18:10 | |
*** reddy has quit IRC | 18:13 | |
sdake | nice cisco is funding midcycle dinner on tuesday - yay :) | 18:20 |
SamYaple | w00t | 18:21 |
sdake | now for breakfast and soda ;) | 18:21 |
*** v1k0d3n has joined #kolla | 18:24 | |
*** tzn has quit IRC | 18:31 | |
*** openstackgerrit has quit IRC | 18:32 | |
*** openstackgerrit has joined #kolla | 18:32 | |
*** akwasnie has joined #kolla | 18:40 | |
*** smartshader has joined #kolla | 18:48 | |
*** Allen_Gao has quit IRC | 18:56 | |
*** Allen_Gao has joined #kolla | 18:57 | |
*** sdake has quit IRC | 19:08 | |
*** gfidente is now known as gfidente|afk | 19:08 | |
*** opennode has joined #kolla | 19:20 | |
*** openstackstatus has quit IRC | 19:20 | |
*** openstackstatus has joined #kolla | 19:20 | |
*** ChanServ sets mode: +v openstackstatus | 19:20 | |
*** sdake has joined #kolla | 19:21 | |
opennode | hi! Im trying to deploy latest master branch (as stable/liberty does not get pass mysql haproxy user creation on centos 7) but deploy fails with “Starting rsyslog container” task with the following error msg: AttributeError("'Client' object has no attribute 'create_host_config’”,) | 19:23 |
sdake | opennode haven't seen that error before | 19:24 |
opennode | can I debug somehow more before failing bugreport? | 19:25 |
sdake | opennode note we just changed to a kolla-developed ansible module for communicating with docker because of untactable upstream problems with the ansible docker module, so there may be a bug or two | 19:25 |
sdake | opennode lets see if SamYaple has any suggestions | 19:25 |
sdake | he wrote the module for the most part iiuc | 19:25 |
opennode | aha ok | 19:25 |
akwasnie | SamYaple: are you there? | 19:25 |
SamYaple | opennode: do you have any logs? how is it failing? what message | 19:25 |
SamYaple | akwasnie: here i am | 19:25 |
opennode | I actually dont wont to build and deploy master branch - just stable/liberty didnt work out for me either... | 19:26 |
opennode | TASK: [common | Starting rsyslog container] *********************************** | 19:26 |
opennode | failed: [localhost] => {"changed": true, "failed": true} | 19:26 |
opennode | msg: AttributeError("'Client' object has no attribute 'create_host_config'",) | 19:26 |
opennode | FATAL: all hosts have already failed -- aborting | 19:26 |
opennode | ansible/library/kolla_docker.py failing perhaps? | 19:27 |
SamYaple | opennode: that is due to a bas os-client-config | 19:27 |
SamYaple | when did you build these images? | 19:27 |
opennode | just now | 19:27 |
SamYaple | hmmm | 19:27 |
opennode | Im on centos 7.1.1503 | 19:28 |
SamYaple | opennode: what version of docker and docker-py? | 19:28 |
opennode | ie the build/deploy host | 19:28 |
SamYaple | on the deploy host | 19:28 |
opennode | Docker version 1.8.2-el7.centos, build a01dc02/1.8.2 | 19:28 |
opennode | python-docker-py.noarch 1.4.0-118.el7 @extras | 19:28 |
SamYaple | ah yes. you need docker 1.9+ and docker-py 1.6. see http://docs.openstack.org/developer/kolla/quickstart.html | 19:28 |
SamYaple | these reqs _just_ changes as of today | 19:29 |
akwasnie | i am working on adapting ELK patchset to kolla_docker and - we discussed earlier today how to handle missing command param in ansible for kolla_docker, and I have a question regarding how to handle the configuration file part | 19:29 |
opennode | ok thnx | 19:29 |
SamYaple | akwasnie: ok | 19:29 |
SamYaple | like the config.json thing? | 19:29 |
akwasnie | we can use config.json to run commands with specific params, but right now ELK containers do not use kolla base image but the oficial ones for elastic, logstash and kibana | 19:30 |
SamYaple | ohh. hmmm | 19:31 |
SamYaple | hmmmmm | 19:31 |
akwasnie | so - this will require to have additonal dockerfiles for E, L and K for each distro | 19:31 |
SamYaple | thats an interesting case | 19:31 |
SamYaple | yea akwasnie to do it the way the other ones work, then yes | 19:31 |
akwasnie | do we want them to be inherited after kolla base and have all the specifics for each service copied into them | 19:31 |
SamYaple | we should discuss this in the meeting on Wednesday | 19:31 |
SamYaple | that would be my vote, to build our own | 19:32 |
SamYaple | but it is additional work | 19:32 |
akwasnie | or, inherit after original items and implement specifics for each distro? | 19:32 |
akwasnie | s/items/images | 19:32 |
SamYaple | akwasnie: my vote would be building our own from scratch, but this is somethign we should talk about in the meeting wednesday with a larger audience | 19:32 |
akwasnie | ok, should I add this to agenda? | 19:33 |
opennode | one thing more - any ides why stable/liberty branch “TASK: [mariadb | Creating haproxy mysql user]” is producing this error: | 19:33 |
opennode | File \"/usr/lib64/python2.7/site-packages/cryptography/hazmat/backends/interfaces.py\", line 12, in <module>", " @six.add_metaclass(abc.ABCMeta)", "AttributeError: 'module' object has no attribute 'add_metaclass'"], "warnings": []} ? | 19:33 |
SamYaple | yes. additionally, you may want to add a mailing list question (if you want) | 19:33 |
akwasnie | ok, thanks for help, SamYaple :) | 19:33 |
SamYaple | opennode: stable liberty only works with docker 1.6.0 to 1.8.2, and docker-py <= 1.4 i believe | 19:34 |
SamYaple | opennode: but that looks like the mysql module in the kolla-ansible contianer. so im not sure | 19:34 |
SamYaple | ive not seen that. i woudl recommend filing a bug | 19:34 |
opennode | it might be variation of https://bugs.launchpad.net/kolla/+bug/1491782 | 19:35 |
openstack | Launchpad bug 1491782 in kolla "Fail to create haproxy mysql user" [Medium,In progress] - Assigned to Sidharth Surana (ssurana) | 19:35 |
SamYaple | opennode: normally that bug you linked is due to the database not running | 19:36 |
opennode | yes - I have different error actually - just the same outcome :) | 19:36 |
opennode | ok I will file a bug | 19:36 |
*** rahulshr has quit IRC | 19:37 | |
*** dims_ has joined #kolla | 19:46 | |
*** dims has quit IRC | 19:47 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow common role to pull images https://review.openstack.org/262846 | 19:48 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow ceph role to pull images https://review.openstack.org/262847 | 19:48 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow common role to pull images https://review.openstack.org/262846 | 19:49 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Allow ceph role to pull images https://review.openstack.org/262847 | 19:49 |
*** tzn has joined #kolla | 19:50 | |
opennode | filed https://bugs.launchpad.net/kolla/+bug/1530962 | 19:51 |
openstack | Launchpad bug 1530962 in kolla "mariadb "Creating haproxy mysql user" task failing with "no no attribute 'add_metaclass'" error" [Undecided,New] | 19:51 |
sdake | akwasnie agree we should discuss on wednesday | 19:53 |
sdake | akwasnie can yo uadd to the the agend aplease | 19:53 |
openstackgerrit | Alicja Kwasniewska proposed openstack/kolla: Added ELK containers for centralised logging https://review.openstack.org/252968 | 19:54 |
*** daneyon has joined #kolla | 20:04 | |
*** akwasnie has quit IRC | 20:04 | |
*** tzn has quit IRC | 21:08 | |
*** tzn has joined #kolla | 21:15 | |
*** tzn has quit IRC | 21:21 | |
*** inc0 has quit IRC | 21:24 | |
*** tzn has joined #kolla | 21:41 | |
*** diogogmt has quit IRC | 21:57 | |
*** sdake has quit IRC | 22:00 | |
*** sdake has joined #kolla | 22:00 | |
*** dwalsh has quit IRC | 22:04 | |
*** rhallisey has quit IRC | 22:11 | |
*** salv-orl_ has quit IRC | 22:15 | |
*** gfidente|afk has quit IRC | 22:16 | |
*** shardy has quit IRC | 22:33 | |
*** Allen_Gao has quit IRC | 22:37 | |
*** Allen_Gao has joined #kolla | 22:38 | |
*** sacharya has quit IRC | 22:48 | |
SamYaple | ok everyone. Here si the pull review. please have a look https://review.openstack.org/#/c/262846/ | 23:02 |
SamYaple | its passing gate and working | 23:02 |
SamYaple | notice the change to the tools/kolla-ansible specifically, its important | 23:03 |
*** blahRus1 has quit IRC | 23:04 | |
*** sacharya has joined #kolla | 23:33 | |
*** sacharya has quit IRC | 23:37 | |
*** tzn has quit IRC | 23:45 | |
*** yuanying has quit IRC | 23:50 | |
*** yuanying has joined #kolla | 23:51 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!