| sdake | the zstreams i mean | 00:00 |
|---|---|---|
| sdake | as for tagging a "special" image that you can use | 00:00 |
| kfox1111 | yeah. tried to raise it with before I left on vacation. but schedules just didn't match up. | 00:00 |
| sdake | we could probably do that | 00:00 |
| portdirect | mgiles: what distro/type are you using? | 00:00 |
| mgiles | centos | 00:00 |
| sdake | like we could tag 3.0.0.0 | 00:00 |
| sdake | which we could later delete | 00:00 |
| kfox1111 | yeah. so, what about a centos-binary-kolla-toolbox:3.0.0-2 ? or something | 00:01 |
| sdake | or 3.0.1.0 | 00:01 |
| sdake | or whatever | 00:01 |
| sdake | we don't have a -1 syntax atm in the images so kolla-ansible wouldn't work | 00:01 |
| portdirect | mgiles: would it unblock you if i threw up a set of images to the docker hub which you could use untill the main kolla repo get the bugfix? | 00:01 |
| kfox1111 | sdake: revisions break kolla-ansible? | 00:01 |
| mgiles | portdirect I think so. Obviously would need to test it to be sure. | 00:02 |
| sdake | kfox1111 nah - kolla-ansible expects 3.0.1 | 00:02 |
| sdake | not 3.0.1-1 | 00:02 |
| kfox1111 | sdake: lame. :/ | 00:02 |
| kfox1111 | we should discuss how to fix that. :/ | 00:02 |
| sdake | its a one line change | 00:02 |
| sdake | in all.yml or globals.yml | 00:02 |
| kfox1111 | cause this is going to be a continuing problem. :/ | 00:02 |
| kfox1111 | well, kolla-ansible's ok anyway though. | 00:03 |
| kfox1111 | iths the openstack cli that's shipped in kolla-toolbox that's broken. | 00:03 |
| kfox1111 | which I don't htink kolla-ansible's using. but kolla-kubernetes is. | 00:03 |
| SamYaple | kfox1111: that naming convention has been _long_ solved and agreed on | 00:03 |
| sbezverk | kfox1111: I still do not get why we cannot just replace existing toolbox with new client | 00:03 |
| kfox1111 | so if we uploaded 3.0.1-2 and updated kolla-kubernetes to point to it, we'd be good. | 00:03 |
| *** sdake has quit IRC | 00:04 | |
| SamYaple | the 3.0.1 tag is not permenant, but a symlink if you will to a 3.0.1-x tag | 00:04 |
| kfox1111 | SamYaple: long implemented, yes. solved, no. | 00:04 |
| SamYaple | kfox1111: nope, not implemetned at all | 00:04 |
| SamYaple | long solved, never implemented | 00:04 |
| kfox1111 | SamYaple: I think its broken. :/ | 00:04 |
| kfox1111 | no way to upgrade the depenedencies in the container without overwriting the tag or bumping the version. | 00:04 |
| SamYaple | whats broken? the system we agreed on but arent using because we dont have a good way to push images? | 00:04 |
| Pavo | nice I just finally got done with my heat template that will deploy ansible tower | 00:05 |
| SamYaple | correct. overwriting the 3.0.1 tag was always the plan. you can tag an image multiple times | 00:05 |
| kfox1111 | then you get version skews in your prodution custer. | 00:05 |
| kfox1111 | bad solution. | 00:05 |
| SamYaple | its the docker solution and it works well when used correctly | 00:05 |
| kfox1111 | I want atomimc tags on my systems. | 00:05 |
| SamYaple | 3.0.1 and 3.0.1-1 point to the same image. then 3.0.1-2 is pushed and 3.0.1 is retagged | 00:06 |
| SamYaple | you are welcome to use the atomic 3.0.1-1 tag | 00:06 |
| kfox1111 | 3.0.1-1 3.0.1-2, etc. then when I upgrade a container, I know it actually applies to places. | 00:06 |
| kfox1111 | ah. | 00:06 |
| kfox1111 | maybe we're misspeaking then. | 00:06 |
| SamYaple | at least, that was what was agreed on originally. like i said, this is not in place | 00:06 |
| kfox1111 | I didn't see -X revisions on the hub | 00:06 |
| SamYaple | right. its not implemented | 00:06 |
| kfox1111 | ah | 00:06 |
| kfox1111 | gotcha. | 00:06 |
| kfox1111 | ok. then we're totally agreeing. :) | 00:07 |
| SamYaple | until we can push images regularly to docker then this can't really be done well | 00:07 |
| SamYaple | kfox1111: let me find you the old ether pad where this was all worked out | 00:07 |
| SamYaple | i spent alot of time on it | 00:07 |
| kfox1111 | cool. I'd appreciate that. :) | 00:07 |
| mgiles | ok. so I'll abandon the ps as long as we plan to find a more permanent solution. | 00:08 |
| mgiles | thanks all | 00:08 |
| kfox1111 | mgiles: sorry to turn your contribution away. no offence. it was good. | 00:09 |
| kfox1111 | just really wana get rid of the hack thats already there. | 00:09 |
| mgiles | kfox1111 nah - I din't really like the fix myself. I agree that fixing the root problem is better | 00:10 |
| portdirect | yeah - it's the correct solution to the wrong problem (as in a problem that shouldn't exist) :) | 00:10 |
| mgiles | on second thought, should I just wf -1 the review and leave it open until we have a real solution - just to serve as a placeholder/reminder | 00:11 |
| portdirect | that would be great | 00:11 |
| kfox1111 | mgiles: yeah. thats a good idea. then others can easly get to it to work around the issue if they need to too. | 00:12 |
| kfox1111 | hmm... theres another temporary option... | 00:14 |
| *** salv-orlando has quit IRC | 00:14 | |
| SamYaple | kfox1111: welp. the link that i did have is expired | 00:15 |
| openstackgerrit | Marcus Williams proposed openstack/kolla: [WIP] Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 00:16 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: [WIP] Add OpenDaylight to kolla-ansible https://review.openstack.org/416367 | 00:16 |
| kfox1111 | SamYaple: bummer. :/ | 00:16 |
| kfox1111 | SamYaple: interested in working on it in the near future though? I am. | 00:17 |
| SamYaple | kfox1111: yea i remember the structure. trying to see if i have the full page | 00:18 |
| SamYaple | archived* | 00:18 |
| portdirect | Just so I stop asking I take it my offer of providing a nightly master from my own hardware is a non-starter? | 00:19 |
| SamYaple | portdirect: of course not, but i dont think you know what that entails.... though you might! | 00:19 |
| SamYaple | its a pita to push to dockerhub, and cant be done in parallel | 00:19 |
| SamYaple | if done in parallel, it wont share layers (even on download sometimes) | 00:20 |
| kfox1111 | portdirect: sounds like a good idea to me. | 00:20 |
| SamYaple | i would suggest trying to push a full set up to your own dockerhub account first, and measure the time and bandwidth before you agree to that | 00:20 |
| kfox1111 | mgiles: I just pushed an updated image here. kfox1111/centos-binary-kolla-toolbox:3.0.1 | 00:20 |
| SamYaple | if its reasonable for you, by all means do it. up until this point (even with gigabit internet) it hasnt been reasonable to do this | 00:20 |
| kfox1111 | should be able to use that with the pip install removed from all the places. | 00:20 |
| mgiles | kfox1111 thanks. I'll give it a try. | 00:21 |
| kfox1111 | SamYaple: have you tried any of the alternates? | 00:21 |
| kfox1111 | mgiles: cool. let me know if it works for you. or it doesn't. I'll see what Ican do. :) | 00:22 |
| SamYaple | kfox1111: local repo (hub.openstack.org) would be a good alternative! | 00:22 |
| SamYaple | kfox1111: there is also quay.io | 00:22 |
| SamYaple | neither of these are great | 00:22 |
| kfox1111 | hub.openstack.org would be cool. yeah. | 00:22 |
| SamYaple | kfox1111: eventually dockerhub will be fixed, heck it might already been. havent checked in a few months | 00:22 |
| kfox1111 | quay have the same performance issue? | 00:23 |
| SamYaple | kfox1111: the conversation has been had with openstack-infra | 00:23 |
| inc0 | kfox1111, ad pushing images | 00:23 |
| kfox1111 | SamYaple: did you try after they added the hashing? | 00:23 |
| SamYaple | its greenlit if we can show an acceptable way | 00:23 |
| SamYaple | nope. | 00:23 |
| inc0 | I'll ping Jeffrey later today to see what's the issue with tagging z stream | 00:23 |
| *** Pavo has quit IRC | 00:23 | |
| SamYaple | might all be ready, i was suggesting portdirect try it before promising anything | 00:23 |
| inc0 | also kfox1111 if it's about gates, I have better way to do it now:) | 00:23 |
| kfox1111 | cool. yeah. so if portdirect tries, and it works, | 00:24 |
| kfox1111 | we can try in the gate again and see if thats fast. it might be now. | 00:24 |
| portdirect | gonna ssh into the office now and set one up | 00:24 |
| inc0 | kfox1111, http://tarballs.openstack.org/kolla/images/ | 00:24 |
| *** Pavo has joined #openstack-kolla | 00:24 | |
| kfox1111 | inc0: sweet. was wondering about that. | 00:24 |
| inc0 | this will hold master-built images | 00:24 |
| inc0 | it's tar.gz-ed registry | 00:25 |
| portdirect | I think that it would be hard for me to do all - but could almost cratianly do a nightly for centos-source | 00:25 |
| portdirect | *certainly | 00:25 |
| kfox1111 | how often are they built? the times are odd. | 00:25 |
| inc0 | still wip, but pretty solid now | 00:25 |
| inc0 | kfox1111, https://review.openstack.org/#/c/404435/ | 00:25 |
| kfox1111 | registry or docker save? | 00:25 |
| inc0 | when you do recheck experimental | 00:25 |
| inc0 | registry | 00:25 |
| kfox1111 | registry... interesting.... | 00:26 |
| inc0 | we do docker run registry -v /var/lib/registry:/somedirectory | 00:26 |
| kfox1111 | ok, so very strange question.... | 00:26 |
| kfox1111 | can you upload that container then to the hub? ;) | 00:26 |
| inc0 | well, it's not a container | 00:26 |
| inc0 | it's a tar.gz directory now;) | 00:26 |
| kfox1111 | oh. a tarball with the registry's backing store? | 00:26 |
| inc0 | yeah | 00:26 |
| *** prameswar has joined #openstack-kolla | 00:26 | |
| inc0 | it's a hack, but it works for now and will give us gate functionality we need | 00:27 |
| kfox1111 | well, similar question then.... couldn't you make a registry with that embeded ready to go, then push that to the hub? :) | 00:27 |
| inc0 | registry ran by infra is wip | 00:27 |
| kfox1111 | wouldn't have the performance issue in question. | 00:27 |
| inc0 | kfox1111, we could, just why? | 00:27 |
| kfox1111 | so folks could get more up to date images. | 00:27 |
| inc0 | it still would, that's couple gigs of uplink transfer | 00:28 |
| inc0 | so no-no for gates | 00:28 |
| kfox1111 | are the gates that resource constrained? | 00:28 |
| inc0 | and for normal people I'd rather them build their images if they want master | 00:28 |
| kfox1111 | like, once a night job? | 00:28 |
| inc0 | well, this is a lot of data for every gate test | 00:28 |
| kfox1111 | well, less thinking master, | 00:28 |
| inc0 | it's downloading, not pushing that's the issue | 00:28 |
| kfox1111 | as latest stable.... | 00:28 |
| inc0 | pushing would even be doable outside infra | 00:29 |
| inc0 | stable as in not-released? | 00:29 |
| inc0 | well, we might do master too | 00:29 |
| kfox1111 | like, keeping 3.0.1 up to date. | 00:29 |
| inc0 | it's really someone doing cronjob... | 00:29 |
| inc0 | running somewhere | 00:29 |
| kfox1111 | like, zuul nightly job, build container,s run tests, push kollalocalregistry container to hub on success. | 00:30 |
| inc0 | we can have images with 3.0.* tag | 00:30 |
| inc0 | doing it in infra will be a lot of work | 00:30 |
| inc0 | doing it by anyone on any vm, much less | 00:30 |
| kfox1111 | but infra's common. | 00:30 |
| inc0 | lots of work == people in infra have tons on their plate | 00:31 |
| kfox1111 | I think I'd put in the work to infra if it was considered a valuable thing if no one else would. | 00:31 |
| inc0 | and this would be less trivial and new kind of functionality | 00:31 |
| kfox1111 | I've done a lot of gate jobs so far. | 00:31 |
| kfox1111 | is a periodic one worse? | 00:31 |
| inc0 | I dunno if they have any stuff for periodic tasks | 00:31 |
| kfox1111 | they do. | 00:31 |
| kfox1111 | trove has a periodic job to build thier own version of http://tarballs.openstack.org/kolla/images/ | 00:32 |
| inc0 | well something we could talk to them | 00:32 |
| kfox1111 | inc0: http://tarballs.openstack.org/trove/images/ :) | 00:32 |
| kfox1111 | though it hasn't ran in a while. | 00:32 |
| inc0 | well it's periodic or after merge? | 00:32 |
| kfox1111 | infra told me about it. | 00:32 |
| inc0 | we will have functionality to push new images after merge | 00:32 |
| inc0 | to tarballs.o.o | 00:32 |
| inc0 | that's the thing we aim for | 00:33 |
| SamYaple | kfox1111: https://etherpad.openstack.org/p/JSuWFXF5lh | 00:33 |
| SamYaple | kfox1111: thats the best i can remember, but i think im missing something. so feel free to update | 00:33 |
| SamYaple | inc0: you know hub.openstack.org is still a thing that infra offers, why are you not trying to pursue that? | 00:33 |
| SamYaple | inc0: why do the tarball thing since that wastes _so_ much space? | 00:34 |
| inc0 | SamYaple, issue we had about this is deploying actual registry | 00:34 |
| kfox1111 | SamYaple: yeah. thats exactly what I was looking for. | 00:34 |
| inc0 | in a way to mirror it across regions | 00:34 |
| kfox1111 | the ability for immutable x.y.z-xxx | 00:34 |
| inc0 | write puppet to do it | 00:34 |
| inc0 | it's just a lot of work | 00:34 |
| inc0 | while with tarballs we could have something right away | 00:34 |
| inc0 | we agreed tarballs is temporary | 00:34 |
| inc0 | it just uses mechanisms infra already has in place as opposed to create new one | 00:35 |
| inc0 | s | 00:35 |
| SamYaple | kfox1111: yea i think these tags caught all the use-cases | 00:35 |
| SamYaple | kfox1111: since tags cost us nothing, we can throw on as many as we would like | 00:35 |
| *** jrobinson has quit IRC | 00:36 | |
| inc0 | well, I gtg for now guys, happy to return to this conversation | 00:36 |
| kfox1111 | SamYaple: +1 :) | 00:36 |
| inc0 | in the meantime kfox, for gates we can use tarballs | 00:36 |
| inc0 | I just need this change to merge | 00:36 |
| kfox1111 | inc0: sweet. will work on a trunk version of the gates. :) | 00:36 |
| kfox1111 | but probably after 0.4.0. | 00:36 |
| inc0 | if you need to re-upload trunk - run check experimnetal on this review | 00:37 |
| SamYaple | inc0: gotcha. if that works well and infra is happy, then i guess its cool. would still like to see a dockerhub.openstack.org | 00:37 |
| SamYaple | i think infra wont like it so much though | 00:37 |
| inc0 | after that we'll do it automatic on merge | 00:37 |
| inc0 | SamYaple, issue is - for them it's yet another service to support | 00:37 |
| inc0 | can't blame them to try to avoid it:) | 00:37 |
| SamYaple | they are not trying to avoid it at all | 00:38 |
| SamYaple | they are waiting on you | 00:38 |
| inc0 | yeah, what I'm saying is | 00:38 |
| SamYaple | infra is 100% down for hub.openstack.org | 00:38 |
| inc0 | we agreed to work with tarballs | 00:38 |
| inc0 | and move to hub afterwards | 00:38 |
| *** Pavo has quit IRC | 00:38 | |
| SamYaple | you dont need to explain anythign to me | 00:38 |
| inc0 | as tarballs are quick and dirty and we need this for gates after repo split | 00:38 |
| *** jrobinson has joined #openstack-kolla | 00:38 | |
| kfox1111 | SamYaple: doesn't the registry have a swift backend? | 00:38 |
| inc0 | kfox1111, infra doesn't ahve swift anywhere | 00:39 |
| inc0 | that we could use | 00:39 |
| SamYaple | kfox1111: yep. i wasnt going to bring it up | 00:39 |
| kfox1111 | hah. funny. :) | 00:39 |
| *** Pavo has joined #openstack-kolla | 00:39 | |
| SamYaple | they do have a globally sync'd AFS | 00:39 |
| kfox1111 | wow. | 00:39 |
| inc0 | thats why tarballs are easy | 00:39 |
| inc0 | it's just a file;) | 00:40 |
| SamYaple | so is the registry backend | 00:40 |
| kfox1111 | does the registry have a shared fs backend? :) | 00:40 |
| SamYaple | yep | 00:40 |
| kfox1111 | afs should be posix compliant. :) | 00:40 |
| inc0 | well, not really shared as much as master-slave | 00:40 |
| kfox1111 | ah. ok. | 00:40 |
| kfox1111 | :) | 00:40 |
| SamYaple | kfox1111: pretty sure it is | 00:40 |
| SamYaple | for out usage | 00:40 |
| SamYaple | our* | 00:40 |
| kfox1111 | well, even master->slaves, should be good enough. | 00:41 |
| SamYaple | but enough of that, i think youre going to get infra to turn off this linent tarballs thing | 00:41 |
| inc0 | anyway, for future | 00:41 |
| SamYaple | they are going to hate kolla | 00:41 |
| kfox1111 | yeah. | 00:41 |
| kfox1111 | heh. | 00:41 |
| kfox1111 | or learn to love us. ;) | 00:41 |
| SamYaple | you should really detail what the bandwidth usage is going to be for them | 00:41 |
| inc0 | SamYaple, I worked on tarballs with them;) | 00:41 |
| kfox1111 | they are good people though. | 00:41 |
| inc0 | it's not like I did it behind their backs | 00:41 |
| inc0 | in fact we came to this idea with pabelanger | 00:41 |
| *** saneax is now known as saneax-_-|AFK | 00:42 | |
| SamYaple | a wee defensive are we? | 00:42 |
| SamYaple | :) | 00:42 |
| kfox1111 | yeah. they are aware of what I had to do for the kolla-kubernetes gates too. | 00:42 |
| inc0 | -.- | 00:42 |
| SamYaple | :P | 00:42 |
| inc0 | anyway, thats quick and dirty registry | 00:42 |
| inc0 | if someone is willing to write proper puppet module for proper registry | 00:42 |
| inc0 | project is up there already | 00:42 |
| SamYaple | yea this really desnt affect me, the hub.openstack.org thing was jsut so it could sync with dockerhub | 00:43 |
| SamYaple | that way no _person_ pushes to dockerhub | 00:43 |
| inc0 | https://review.openstack.org/#/c/399181/ | 00:43 |
| kfox1111 | inc0: this is a big step tough. very cool. :) | 00:43 |
| kfox1111 | https://review.openstack.org/#/c/416077/ is ready for some reviews. | 00:43 |
| inc0 | lol | 00:44 |
| inc0 | progress meter | 00:45 |
| inc0 | do it in ncurses! be a man! | 00:45 |
| kfox1111 | yupers. :) | 00:45 |
| kfox1111 | heh. | 00:45 |
| inc0 | anyway, I'm off | 00:45 |
| kfox1111 | was trying not to. :) | 00:45 |
| inc0 | have a great evening | 00:45 |
| SamYaple | inc0: have fun in not-texas | 00:45 |
| kfox1111 | you too | 00:45 |
| inc0 | haha | 00:45 |
| openstackgerrit | Joshua Harlow proposed openstack/kolla: Not much need to cache good, bad, unmatched https://review.openstack.org/406391 | 00:45 |
| *** inc0 has quit IRC | 00:46 | |
| *** tovin07 has joined #openstack-kolla | 00:51 | |
| *** jmccarthy has quit IRC | 00:52 | |
| *** jmccarthy has joined #openstack-kolla | 00:53 | |
| *** jrobinson has quit IRC | 00:53 | |
| *** jrobinson has joined #openstack-kolla | 00:54 | |
| *** duonghq has joined #openstack-kolla | 00:57 | |
| portdirect | kfox1111: that progress meter is a joke right? | 00:58 |
| *** jtriley has joined #openstack-kolla | 00:59 | |
| kfox1111 | portdirect: shrug. it has come in handy a few times on some of my slower/heavily loaded systems. | 00:59 |
| portdirect | dude - you need some new hardware :/ | 01:00 |
| kfox1111 | :) | 01:00 |
| portdirect | I'll write a letter to your boss if it helps :) | 01:00 |
| srwilkers | ill cosign it | 01:01 |
| kfox1111 | waste not, want not. :) | 01:02 |
| *** adrian_otto has joined #openstack-kolla | 01:02 | |
| kfox1111 | still using spinny disks on my laptop as its still a bit hard to run openstacks on a laptop with ssd's. | 01:03 |
| portdirect | ? | 01:03 |
| kfox1111 | I run openstack and kubernetes on my laptop. :) | 01:04 |
| portdirect | now your just messing with me - please say you are | 01:04 |
| kfox1111 | nope. works great, other then battery life... which really sucks. :) | 01:04 |
| portdirect | Re the ssd/hdd - I'm running the same but would go mad with an hdd | 01:05 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Visual feedback https://review.openstack.org/416077 | 01:05 |
| kfox1111 | yeah. still got a couple of sizable windows vm's around too. for dev purposes. :/ | 01:05 |
| portdirect | fedora 25, k8s 1.5, freeipa server and openstack on 2011 mpb - no probs but running ssd's in raid | 01:05 |
| kfox1111 | their dev environments are not light. :/ | 01:05 |
| *** mattmceuen has quit IRC | 01:05 | |
| kfox1111 | I think I might be able to get off disk soon though, if Ican ditch those. | 01:06 |
| *** caowei_ has joined #openstack-kolla | 01:06 | |
| *** yingjun has joined #openstack-kolla | 01:06 | |
| *** caowei has quit IRC | 01:07 | |
| *** caowei_ is now known as caowei | 01:07 | |
| *** caowei has quit IRC | 01:09 | |
| portdirect | yeah a 512gb ssd aint that expensive these days (my 240's in raid 0 is a bit scary - and dual bays is rare as these days) but running windows vm's of fa 2.5inch hdd is impressive - I suppose it means you get to take plenty of coffee breaks :) | 01:13 |
| kfox1111 | I usually build stuff in clouds with faster resource. | 01:14 |
| kfox1111 | laptops mostly for thin terminal/airplanes. | 01:14 |
| kfox1111 | and I don't travel much. :) | 01:14 |
| *** salv-orlando has joined #openstack-kolla | 01:15 | |
| portdirect | nice - thats about 2/3 of the images pushed for both centos-source and ubuntu-source, the rest should only take about 25 mins unless docker hub starts throttling me again :D | 01:17 |
| *** jtriley has quit IRC | 01:19 | |
| kfox1111 | how long did the first 2/3'rds take? | 01:20 |
| *** salv-orlando has quit IRC | 01:20 | |
| portdirect | 53 mins | 01:21 |
| *** goldyfruit has quit IRC | 01:21 | |
| kfox1111 | nice. | 01:22 |
| kfox1111 | so, yeah. | 01:22 |
| portdirect | forgot to run time on centos, but will get proper results from ubuntu: 'time kolla-build --base ubuntu --type source --push --push-threads=1 --namespace kollakube' | 01:22 |
| kfox1111 | if we did like, one job per type, we might get it through faster yet. | 01:22 |
| kfox1111 | was that with build too? | 01:23 |
| portdirect | yup | 01:23 |
| kfox1111 | very nice. | 01:23 |
| kfox1111 | yeah, I think we should try and get that into a zuul job. :) | 01:23 |
| *** jtriley has joined #openstack-kolla | 01:23 | |
| *** unicell has joined #openstack-kolla | 01:24 | |
| kfox1111 | would need a way to compare for changes too if at all possible. | 01:24 |
| kfox1111 | so, like, something like: rpm -qa | sort | md5sum | 01:24 |
| kfox1111 | and some logic to then retag/upload only if a change is seen. | 01:25 |
| SamYaple | kfox1111: how would that work for source distros? | 01:25 |
| SamYaple | or, for that matter, when we occastionally have to pull things from source in binary distros? | 01:26 |
| kfox1111 | SamYaple: would need some way for that.. (rpm -qa; pip list) or something? | 01:26 |
| SamYaple | or for outside things like rabbitmq containers that pull clusterer | 01:26 |
| *** hfu has joined #openstack-kolla | 01:27 | |
| openstackgerrit | wangwei proposed openstack/kolla: Add the function of docker login before pushing https://review.openstack.org/399426 | 01:27 |
| portdirect | yeah - I think that in practice that would be an entire project in its own right :/ | 01:27 |
| kfox1111 | yeah. there would be a fue edge cases for sure. but should be a manageble set. | 01:27 |
| SamYaple | portdirect: yea. better to get the tiny images _first_ | 01:27 |
| kfox1111 | I've done it with other projects ok. | 01:27 |
| SamYaple | then we can do a md5/sha hash accross all the files in the tiny image | 01:28 |
| portdirect | SamYaple: +1 | 01:28 |
| kfox1111 | hmm.. that would work too. | 01:28 |
| SamYaple | with some "we dont care about those files" ignores, it should work well | 01:28 |
| kfox1111 | how big are the images now though? would it be all that bad now actually? | 01:28 |
| portdirect | SamYaple: do you have a debbootstrap image or simlar you're working from? | 01:28 |
| portdirect | or just cut out thr crud from the ubuntu base and then squash? | 01:29 |
| portdirect | used to build centos images from a chroot - but havent done that in a while | 01:30 |
| SamYaple | portdirect: its more of a bigass command at the end of the image building process, for each image. then flatten those into a layer. _then_ recompute common layers and at the end you have shared layers | 01:31 |
| SamYaple | so we can acctually remove all the && crap | 01:31 |
| SamYaple | kfox1111: give it a shot | 01:32 |
| SamYaple | kfox1111: you can hash per layer too | 01:32 |
| kfox1111 | SamYaple: I guess there's another option too. | 01:32 |
| portdirect | SamYaple: has per laywer wont work akaik - to many things add timestamps :/ | 01:32 |
| portdirect | *hash per layer | 01:32 |
| kfox1111 | rpm -qa and exclude all files managed by rpm. and then md5sum the rest. | 01:33 |
| *** sdake has joined #openstack-kolla | 01:33 | |
| SamYaple | hash per layer with ignore list? | 01:33 |
| portdirect | also SamYaple the '_then_ ' in your sentence is the bit that gets me interested :) | 01:33 |
| SamYaple | i think youd be suprised at how many timestamps there _arent_ | 01:33 |
| SamYaple | portdirect: the docker image is acatually well documented... in code | 01:34 |
| SamYaple | so its just a matter of assemling layers there. which is actually quite easy, even in python | 01:34 |
| SamYaple | they are just tarballs | 01:34 |
| portdirect | yeah - I've been there :) | 01:34 |
| portdirect | and know you have too - it was more the way of determining what need to be in each layer - a simple conceptual problem - but interested in how you've implemented it | 01:35 |
| SamYaple | portdirect: oh gotcha. its not dynamic as i would like yet. its once you have the flattened image, reproduce the major layers with a subset of files that existed in those layers (if hashes match) | 01:36 |
| SamYaple | so in the end, you still have base->openstack-base->nova-base->nova-api | 01:37 |
| SamYaple | but only 4 layers, not 40 | 01:37 |
| portdirect | ah ok - good 1st step though | 01:38 |
| SamYaple | portdirect: what I would _like_ is a priority list of some kind. but i dont know what those rules would be. 'include file if it exists in >70% of images or is >XMB in size' | 01:39 |
| SamYaple | dont know if that would be good or bad though | 01:39 |
| kfox1111 | heh. | 01:40 |
| kfox1111 | don't know how computationally expensive it would be, | 01:40 |
| portdirect | yeah - thats the hardest one - I'm more moving toward only have it in if you need it there - but then tomorrow I'll be thinkting along the same lines as you are today... | 01:40 |
| kfox1111 | but the ideal solution would be to search the set of desired end containers and dynamically discover how many layers you need to optomize for total space consumed. | 01:40 |
| *** MasterOfBugs has quit IRC | 01:41 | |
| kfox1111 | not sure how large a set that would be in the end though. | 01:41 |
| kfox1111 | might be too many permutations though... | 01:41 |
| portdirect | kfox1111: thats kinda the beauty of these things though - never have any preconceptions and let the maths determine the optimal solution :) | 01:42 |
| * kfox1111 nods. | 01:43 | |
| sdake | sup peeps | 01:43 |
| kfox1111 | that woudl be the ideal solution. | 01:43 |
| kfox1111 | (same with chess) | 01:43 |
| portdirect | blue sky thinking sdake :) | 01:43 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move trove precheck into its own role https://review.openstack.org/416032 | 01:43 |
| sdake | the game of chess can be solved computationally ;) | 01:43 |
| sdake | go on the othe rhand cannot | 01:44 |
| SamYaple | kfox1111: portdirect as far as expensive, its very! dont recommend it. its why i havent purused it until i build a repo/cache contaienr | 01:44 |
| sdake | kings gambit loses by force if played perfectly - every position calculated | 01:44 |
| SamYaple | so the idea is to built a pip wheel and package cache container so everything is cached on the local build host | 01:44 |
| portdirect | SamYaple: | 01:45 |
| SamYaple | portdirect: | 01:45 |
| sdake | SamYaple: | 01:46 |
| SamYaple | sdake: | 01:46 |
| sdake | sdake: | 01:46 |
| portdirect | SamYaple: yeah - every time I've started to experiment with size optiisation its ended up costing more time than its saved | 01:46 |
| portdirect | SamYaple: | 01:46 |
| * portdirect is a irc clutz | 01:46 | |
| SamYaple | portdirect: indeed. but with a repo container, gcc will never be in the images! | 01:46 |
| *** jrobinson has quit IRC | 01:46 | |
| SamYaple | portdirect: its doesnt help so much with binary, other than caching the package downloads | 01:47 |
| portdirect | yeah - thats real ugly - I'm fine with it persoanlly but suprised no ones kicked off about that | 01:47 |
| portdirect | (the gcc/dev libs thing) | 01:48 |
| *** jrobinson has joined #openstack-kolla | 01:48 | |
| SamYaple | yea so far theres been no way around it. but I can build a way around it, optionally i think | 01:48 |
| portdirect | thats centos-source just finished | 01:49 |
| portdirect | https://hub.docker.com/r/kollakube/ | 01:50 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move ironic precheck into its own role https://review.openstack.org/415481 | 01:50 |
| *** sdake_ has joined #openstack-kolla | 01:50 | |
| sdake_ | kfox1111 nick change | 01:52 |
| *** dave-mccowan has joined #openstack-kolla | 01:52 | |
| sdake_ | sup dave-mccowan | 01:52 |
| *** sdake has quit IRC | 01:52 | |
| portdirect | and ubuntu-source: 76m14.013s | 01:53 |
| portdirect | think I can semi officially commit to a nightly for the rest of this dev cycle | 01:53 |
| *** jrobinson has quit IRC | 01:54 | |
| sdake_ | portdirect re CFP for kolla-kubernetes | 01:54 |
| sdake_ | i believe the deadline is 2/6 iirc for the may summit in boston | 01:54 |
| portdirect | k - so who's going? | 01:55 |
| portdirect | (I'm up for it obviously) And sbezverk? kfox1111? you? | 01:56 |
| portdirect | (cant rememeber) | 01:56 |
| sdake_ | portdirect a talk can only have 3 members | 01:56 |
| sdake_ | iirc we decided on you me and sbezverk | 01:57 |
| portdirect | sweet - when you free to knock heads on working out the abstract? | 01:58 |
| sdake_ | monday-ish | 01:58 |
| sdake_ | you are making the travel right? | 01:58 |
| sdake_ | I am unclear if kfox1111 was ogin gto make it or not | 01:58 |
| portdirect | sounds good-ish - yeah for that one I'll be there | 01:58 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add tarballs-base parameter for kolla-build https://review.openstack.org/396830 | 01:59 |
| portdirect | SamYaple: looking at the push code - we should be able to get that time down a lot - once the parents for an image are pushed then we can up the threads i think | 02:00 |
| sbezverk | sdake_: I am going to boston for sure | 02:02 |
| sbezverk | it is 5 hours drive for me | 02:03 |
| portdirect | sbezverk: didn't want to wake you - but forgot it's not stupid o-clock over there :) you free on monday? | 02:03 |
| *** caowei has joined #openstack-kolla | 02:05 | |
| *** dave-mcc_ has joined #openstack-kolla | 02:06 | |
| *** dave-mccowan has quit IRC | 02:08 | |
| sbezverk | portdirect: depending on time, what do you have in mind? | 02:10 |
| openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 02:10 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: TEST MASTER BRANCH https://review.openstack.org/398095 | 02:12 |
| *** Pavo has quit IRC | 02:13 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 02:14 |
| *** salv-orlando has joined #openstack-kolla | 02:16 | |
| *** Pavo has joined #openstack-kolla | 02:17 | |
| *** unicell has quit IRC | 02:19 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 02:19 |
| Jeffrey4l | morning | 02:20 |
| *** salv-orlando has quit IRC | 02:21 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move tacker precheck into its own role https://review.openstack.org/416052 | 02:22 |
| kfox1111 | sdake_: around? | 02:23 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move senlin precheck into its own role https://review.openstack.org/415486 | 02:26 |
| SamYaple | portdirect: well if thats the case, thats how the current kolla-build --push code works :P | 02:26 |
| SamYaple | portdirect: --push --push-threads 4 is a thing that works just like that | 02:27 |
| SamYaple | thats what was broken on the dockerhub | 02:27 |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 02:28 |
| *** sdake_ has quit IRC | 02:29 | |
| openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 02:30 |
| *** sayanta__ has quit IRC | 02:31 | |
| portdirect | SamYaple: oh missed the bottom - nice :) 4 thread push a go-go | 02:33 |
| *** dave-mcc_ has quit IRC | 02:33 | |
| *** dave-mccowan has joined #openstack-kolla | 02:39 | |
| *** goldyfruit has joined #openstack-kolla | 02:46 | |
| *** mattmceuen has joined #openstack-kolla | 02:46 | |
| *** jrich523 has quit IRC | 02:47 | |
| *** sdake has joined #openstack-kolla | 02:49 | |
| sdake | sup Jeffrey4l | 02:49 |
| Jeffrey4l | hi sdake | 02:49 |
| *** Pavo has quit IRC | 02:51 | |
| *** srwilkers has quit IRC | 02:54 | |
| *** jrobinson has joined #openstack-kolla | 02:55 | |
| *** jrobinson is now known as jrobinson-afk | 02:58 | |
| *** dave-mccowan has quit IRC | 03:01 | |
| Jeffrey4l | kfox1111, around? | 03:08 |
| *** mgiles has quit IRC | 03:12 | |
| *** sp_ has joined #openstack-kolla | 03:13 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla-ansible: Removing unused directory "etc/oslo-config-generator" from kolla-ansible https://review.openstack.org/415399 | 03:17 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Fix gate in iax-ord node when using libvirt 2.0 https://review.openstack.org/416417 | 03:22 |
| *** jrobinson-afk is now known as jrobinson- | 03:28 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add horizon into gate profile https://review.openstack.org/416425 | 03:45 |
| openstackgerrit | zhongshengping proposed openstack/kolla: Add support for installing binary packages of watcher https://review.openstack.org/416426 | 03:58 |
| *** sayantan_ has joined #openstack-kolla | 04:01 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes: Helm Ceph Initial Mon https://review.openstack.org/410641 | 04:03 |
| *** yingjun has quit IRC | 04:10 | |
| *** bmace has quit IRC | 04:14 | |
| *** bmace has joined #openstack-kolla | 04:14 | |
| *** salv-orlando has joined #openstack-kolla | 04:17 | |
| *** sayantan_ has quit IRC | 04:20 | |
| *** salv-orlando has quit IRC | 04:21 | |
| *** Dishwasha has joined #openstack-kolla | 04:46 | |
| *** goldyfruit has quit IRC | 04:47 | |
| Dishwasha | howdy. After figuring a few things out, I finally got kolla running on 3 compute and 1 controller node. I've switched from openvswitch to linuxbridge and I still can't figure out how the kolla external interface works since it assigns a /32 to that interface | 04:47 |
| Dishwasha | I can ping out from the controller node, but I can't ping to it from another host | 04:48 |
| Dishwasha | I've added some blanket allow statements in iptables and even cleared the ebtables configuration with no effect | 04:50 |
| *** saneax-_-|AFK is now known as saneax | 04:53 | |
| *** unicell has joined #openstack-kolla | 05:01 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add tarballs-base parameter for kolla-build https://review.openstack.org/396830 | 05:06 |
| *** mdnadeem has joined #openstack-kolla | 05:10 | |
| Dishwasha | bleh, I'm silly. rtl8139 doesn't support 802.1q in kvm, only virtio does | 05:12 |
| Dishwasha | k, re-deployed and I can reach the external, but only after I change from /32 to /24 | 05:32 |
| *** g3ek has quit IRC | 05:33 | |
| *** bachp has quit IRC | 05:34 | |
| *** g3ek has joined #openstack-kolla | 05:42 | |
| *** kristian__ has joined #openstack-kolla | 05:47 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add "Virt_type" options into globals.yml https://review.openstack.org/416437 | 05:50 |
| *** yingjun has joined #openstack-kolla | 05:50 | |
| *** kristian__ has quit IRC | 05:52 | |
| *** bachp has joined #openstack-kolla | 05:54 | |
| *** kristian__ has joined #openstack-kolla | 06:00 | |
| *** kristian__ has quit IRC | 06:00 | |
| *** jrobinson- is now known as jrobinson | 06:06 | |
| *** Dishwasha has quit IRC | 06:12 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Consistent home directory creation for all the services https://review.openstack.org/390179 | 06:14 |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 06:28 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add "virt_type" options into globals.yml https://review.openstack.org/416437 | 06:31 |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 06:35 |
| *** satyar has joined #openstack-kolla | 06:41 | |
| *** salv-orlando has joined #openstack-kolla | 06:43 | |
| openstackgerrit | Cao Xuan Hoang proposed openstack/kolla: Remove trailing backtick https://review.openstack.org/416447 | 06:47 |
| openstackgerrit | Cao Xuan Hoang proposed openstack/kolla-ansible: Remove trailing backtick https://review.openstack.org/416449 | 06:51 |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 06:51 |
| *** pcaruana has joined #openstack-kolla | 06:51 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes: Helm Ceph OSD Bootstrap https://review.openstack.org/410696 | 06:52 |
| openstackgerrit | Li Yingjun proposed openstack/kolla: Improve kolla-build cmd help message https://review.openstack.org/416450 | 06:54 |
| *** msimonin has joined #openstack-kolla | 06:55 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes: Helm Ceph Mon https://review.openstack.org/410660 | 06:56 |
| *** jrobinson has quit IRC | 06:59 | |
| *** adrian_otto has quit IRC | 07:00 | |
| openstackgerrit | Li Yingjun proposed openstack/kolla: Improve kolla-build cmd help message https://review.openstack.org/416450 | 07:08 |
| openstackgerrit | Merged openstack/kolla-ansible: Update the horizon port into all.yml https://review.openstack.org/415495 | 07:19 |
| *** msimonin has quit IRC | 07:21 | |
| satyar | Hi jeffrey4l | 07:24 |
| Jeffrey4l | satyar, poing | 07:24 |
| satyar | :) | 07:24 |
| satyar | jeffrey4l have you seen the high CPU usage at cinder volume container? | 07:27 |
| satyar | almost 200% | 07:27 |
| Jeffrey4l | satyar, no, how did u get that? | 07:28 |
| Jeffrey4l | during boot from volume? | 07:28 |
| satyar | nope | 07:28 |
| satyar | i see the spikes | 07:28 |
| *** mnasiadka has joined #openstack-kolla | 07:29 | |
| satyar | may be because of socket connections getting dropped and established | 07:29 |
| satyar | like mysql and amqp | 07:29 |
| Jeffrey4l | satyar, from htop? or monitor tools? | 07:29 |
| satyar | i used top | 07:30 |
| satyar | and then used lsof -np <child pid of cinder volume> | 07:30 |
| satyar | 50 cinder 20 0 7817368 170920 18440 R 173.3 0.1 697:18.13 cinder-volume | 07:31 |
| Jeffrey4l | so it see 200% cpu with any operation in openstack? | 07:31 |
| satyar | yes | 07:31 |
| satyar | and it drops the mysql and amqp connections almost every 20 mins | 07:32 |
| Jeffrey4l | does new deployment has such issue? | 07:32 |
| satyar | yes | 07:33 |
| satyar | i am seeing the issue with new deployment | 07:33 |
| Jeffrey4l | satyar, how about cinder-volume process exist in your one node? | 07:33 |
| satyar | i have not seen the issue with 21/11/2016 code | 07:33 |
| satyar | tried that seen the same | 07:34 |
| satyar | i tried moving the cinder volume to a node with only cinder volume container | 07:35 |
| Jeffrey4l | in my env, i do not see such issue ( almost the latest code ) | 07:35 |
| satyar | is your backend ceph? | 07:35 |
| satyar | or lvm | 07:35 |
| Jeffrey4l | yep. | 07:35 |
| Jeffrey4l | ceph | 07:35 |
| satyar | strange | 07:35 |
| Jeffrey4l | try to get the cinder-volume process number " ps aux | grep cinder-volume | wc -l " | 07:35 |
| satyar | 3 | 07:36 |
| satyar | there are 2 cinder volume process running | 07:36 |
| *** kristian__ has joined #openstack-kolla | 07:37 | |
| Jeffrey4l | weird ;( | 07:38 |
| satyar | hmm | 07:39 |
| *** awiddersheim has quit IRC | 07:39 | |
| *** awiddersheim has joined #openstack-kolla | 07:39 | |
| *** salv-orlando has quit IRC | 07:39 | |
| *** kristian__ has quit IRC | 07:39 | |
| *** kristian__ has joined #openstack-kolla | 07:39 | |
| *** kristian__ has quit IRC | 07:42 | |
| *** bmace has quit IRC | 07:43 | |
| *** yingjun has quit IRC | 07:49 | |
| *** yingjun has joined #openstack-kolla | 07:50 | |
| *** haplo37 has quit IRC | 07:57 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 07:58 |
| *** matrohon has joined #openstack-kolla | 07:59 | |
| *** haplo37 has joined #openstack-kolla | 08:06 | |
| *** caowei has quit IRC | 08:14 | |
| *** caowei has joined #openstack-kolla | 08:20 | |
| *** rmart04 has joined #openstack-kolla | 08:21 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add "virt_type" options into globals.yml https://review.openstack.org/416437 | 08:23 |
| *** caowei has quit IRC | 08:24 | |
| *** caowei has joined #openstack-kolla | 08:24 | |
| *** caowei_ has joined #openstack-kolla | 08:28 | |
| *** caowei has quit IRC | 08:29 | |
| *** caowei_ is now known as caowei | 08:29 | |
| *** hfu has quit IRC | 08:29 | |
| *** egonzalez90 has joined #openstack-kolla | 08:33 | |
| egonzalez90 | morning o/ | 08:33 |
| *** shardy has joined #openstack-kolla | 08:34 | |
| *** eaguilar has joined #openstack-kolla | 08:34 | |
| *** hfu has joined #openstack-kolla | 08:35 | |
| *** eaguilar has quit IRC | 08:37 | |
| *** Serlex has joined #openstack-kolla | 08:38 | |
| *** zhubingbing has joined #openstack-kolla | 08:38 | |
| zhubingbing | hello guys | 08:39 |
| *** caowei has quit IRC | 08:40 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 08:42 |
| *** jimmygc has joined #openstack-kolla | 08:43 | |
| *** saneax is now known as saneax-_-|AFK | 08:44 | |
| *** caowei has joined #openstack-kolla | 08:44 | |
| *** athomas has joined #openstack-kolla | 08:44 | |
| *** salv-orlando has joined #openstack-kolla | 08:49 | |
| *** kristian__ has joined #openstack-kolla | 08:50 | |
| *** msimonin has joined #openstack-kolla | 08:59 | |
| *** caowei_ has joined #openstack-kolla | 08:59 | |
| *** caowei has quit IRC | 09:01 | |
| *** caowei_ is now known as caowei | 09:01 | |
| *** saneax-_-|AFK is now known as saneax | 09:03 | |
| *** prameswar has quit IRC | 09:07 | |
| jistr | hi, is there a way to persuade kolla-build to push to a custom registry using http instead of https, so that i don't have to go through the exercise of setting up self-signed certs every time for dev environment purposes? | 09:28 |
| *** tis has joined #openstack-kolla | 09:29 | |
| jistr | i tried explicitly adding the protocol to --registry, like `kolla-build --registry http://some-custom-registry --push`, but it's apparently not a valid format for the --registry option | 09:29 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Ansible-ize OpenStack Designate https://review.openstack.org/408465 | 09:36 |
| *** hfu has quit IRC | 09:37 | |
| *** hfu has joined #openstack-kolla | 09:40 | |
| *** bjolo has joined #openstack-kolla | 09:40 | |
| bjolo | morning | 09:41 |
| bjolo | happy new year everyone :) | 09:41 |
| openstackgerrit | caoyuan proposed openstack/kolla: Move Vmtp down to the "Infrastructure components" section https://review.openstack.org/416501 | 09:41 |
| egonzalez90 | jistr: have you tried adding EXTRA_ARGS="--insecure-registry http://192.168.2.5:5000? | 09:42 |
| egonzalez90 | happy new year bjolo | 09:42 |
| *** yingjun has quit IRC | 09:42 | |
| jistr | egonzalez90: i haven't will try :) | 09:42 |
| egonzalez90 | seems like a docker bug | 09:42 |
| *** hfu has quit IRC | 09:47 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Unify haproxy prechecks condition order https://review.openstack.org/415510 | 09:48 |
| *** pbourke has quit IRC | 09:52 | |
| *** pbourke has joined #openstack-kolla | 09:53 | |
| *** salv-orl_ has joined #openstack-kolla | 09:53 | |
| *** kristian__ has quit IRC | 09:54 | |
| *** apuimed|away is now known as apuimedo | 09:56 | |
| *** salv-orlando has quit IRC | 09:56 | |
| *** kristian__ has joined #openstack-kolla | 09:56 | |
| *** salkin has joined #openstack-kolla | 09:59 | |
| *** Mech422 has joined #openstack-kolla | 09:59 | |
| Mech422 | hey guys - dumb question - what was the final decision for the logging stuff ? heka + fluentd ? | 09:59 |
| *** l4yerffeJ has quit IRC | 10:03 | |
| *** l4yerffeJ_ is now known as l4yerffeJ | 10:03 | |
| *** tovin07 has quit IRC | 10:05 | |
| *** jimmygc has quit IRC | 10:05 | |
| egonzalez90 | Mech422: fluentd https://review.openstack.org/#/c/407392/ | 10:08 |
| Mech422 | ahh - fluentd + kibana? oh thanks! | 10:08 |
| egonzalez90 | not sure about the kibana part, zhubingbing is working on fluentd for login recolection | 10:09 |
| *** prameswar has joined #openstack-kolla | 10:11 | |
| Mech422 | thanks ! | 10:18 |
| egonzalez90 | someone have the link for BCN summit etherpad? | 10:18 |
| jistr | egonzalez90: oh you meant EXTRA_ARGS for the docker service right? I actually tried that before, but messed up a bit. It works now :) | 10:22 |
| jistr | I focused on getting rid of the http/https error, while i should have focused on making the insecure registry known to docker. So first i made the registry talk https, i still got errors that the cert is untrusted (self-signed), so only then i found out that i have to edit the systemd unit file for docker to add `--insecure-registry blabla`, which made the cert error disappear. It looks that if i had | 10:23 |
| jistr | done only the --insecure-registry part, it would have worked fine as well. So as it seems to me now, the --insecure-registry parameter for docker deamon can be used two ways -- either to allow http://, or allow https:// with an untrusted cert. Which one of those it is, is probably autodetected based on whether the registry is listening for TLS on the specified port. | 10:23 |
| egonzalez90 | Recent docker versions need to put http in insecure registry url, before it worked without the protocol | 10:23 |
| jistr | egonzalez90: ^^ thanks for the help! | 10:23 |
| * jistr biab | 10:24 | |
| *** gfidente has joined #openstack-kolla | 10:26 | |
| *** zhurong has joined #openstack-kolla | 10:26 | |
| *** satyar has quit IRC | 10:28 | |
| *** pbourke has quit IRC | 10:41 | |
| *** pbourke has joined #openstack-kolla | 10:41 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Ansible-ize OpenStack Designate https://review.openstack.org/408465 | 10:48 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Ansible-ize OpenStack Designate https://review.openstack.org/408465 | 10:50 |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla-ansible: Remove genconfig from tox.ini https://review.openstack.org/416509 | 10:50 |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla-ansible: Remove genconfig from tox.ini https://review.openstack.org/416509 | 10:50 |
| pbourke | can someone please +a https://review.openstack.org/#/c/384496/ it's been open since october | 10:58 |
| pbourke | same for https://review.openstack.org/#/c/398373/ https://review.openstack.org/#/c/401979/ https://review.openstack.org/#/c/401996/ https://review.openstack.org/#/c/402399/ https://review.openstack.org/#/c/408599/ https://review.openstack.org/#/c/409832/ | 11:00 |
| pbourke | Jeffrey4l egonzalez90 Daviey sdake ^ | 11:01 |
| Daviey | looking | 11:02 |
| *** duonghq has quit IRC | 11:03 | |
| *** salv-orl_ has quit IRC | 11:05 | |
| sp_ | Jeffrey4l: ping.... https://review.openstack.org/#/c/390179/8 | 11:06 |
| *** matrohon has quit IRC | 11:09 | |
| *** zhurong has quit IRC | 11:19 | |
| *** tovin07 has joined #openstack-kolla | 11:19 | |
| *** saneax is now known as saneax-_-|AFK | 11:23 | |
| egonzalez90 | zhubingbing: ping | 11:28 |
| *** salv-orlando has joined #openstack-kolla | 11:28 | |
| *** Bico_Fino has joined #openstack-kolla | 11:33 | |
| *** msimonin has quit IRC | 11:40 | |
| mliima | morning guys | 11:49 |
| *** sdake has quit IRC | 11:49 | |
| *** hfu has joined #openstack-kolla | 11:49 | |
| *** hfu has quit IRC | 11:50 | |
| egonzalez90 | morning mliima , how was your holidays? | 11:54 |
| mliima | pretty nice egonzalez90 | 11:55 |
| mliima | :) | 11:55 |
| openstackgerrit | Merged openstack/kolla-ansible: Removing unused directory "etc/oslo-config-generator" from kolla-ansible https://review.openstack.org/415399 | 11:55 |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla: Remove kolla-genpwd from setup.cfg https://review.openstack.org/416532 | 11:58 |
| openstackgerrit | Merged openstack/kolla-ansible: Fix gate in iax-ord node when using libvirt 2.0 https://review.openstack.org/416417 | 12:03 |
| *** shardy is now known as shardy_lunch | 12:07 | |
| *** caowei has quit IRC | 12:15 | |
| *** kristian__ has quit IRC | 12:26 | |
| *** gfidente has quit IRC | 12:30 | |
| *** gfidente has joined #openstack-kolla | 12:31 | |
| *** openstackgerrit has quit IRC | 12:33 | |
| *** jtriley has quit IRC | 12:36 | |
| *** openstackgerrit has joined #openstack-kolla | 12:38 | |
| openstackgerrit | Merged openstack/kolla: Add horizon into gate profile https://review.openstack.org/416425 | 12:38 |
| openstackgerrit | Merged openstack/kolla: Readme Steps fail for Vagrant https://review.openstack.org/414698 | 12:39 |
| *** skramaja has quit IRC | 12:40 | |
| *** msimonin has joined #openstack-kolla | 12:43 | |
| *** shardy_lunch is now known as shardy | 12:45 | |
| *** jtriley has joined #openstack-kolla | 12:46 | |
| *** tovin07 has quit IRC | 12:55 | |
| *** salv-orlando has quit IRC | 12:56 | |
| *** zhurong has joined #openstack-kolla | 12:57 | |
| *** matrohon has joined #openstack-kolla | 12:59 | |
| *** jtriley has quit IRC | 13:01 | |
| *** msimonin has quit IRC | 13:02 | |
| *** rmart04_ has joined #openstack-kolla | 13:05 | |
| *** rmart04 has quit IRC | 13:06 | |
| *** rmart04_ is now known as rmart04 | 13:06 | |
| *** salv-orlando has joined #openstack-kolla | 13:07 | |
| *** dave-mccowan has joined #openstack-kolla | 13:08 | |
| *** goldyfruit has joined #openstack-kolla | 13:11 | |
| *** goldyfruit has quit IRC | 13:11 | |
| *** zhubingbing has quit IRC | 13:13 | |
| *** mnasiadka has quit IRC | 13:14 | |
| *** rmart04_ has joined #openstack-kolla | 13:14 | |
| *** rmart04 has quit IRC | 13:15 | |
| *** rmart04_ is now known as rmart04 | 13:15 | |
| *** sdake has joined #openstack-kolla | 13:16 | |
| sdake | morning peeps | 13:16 |
| portdirect | hey sdake | 13:17 |
| sdake | hey portdirect | 13:18 |
| sdake | did that time work out for you on monday | 13:18 |
| sdake | i think it was a bit late for the uk | 13:18 |
| portdirect | all good :) | 13:18 |
| portdirect | ( I just wake up later these days...) | 13:19 |
| sdake | i went to bed at 7pm last night | 13:19 |
| sdake | its 5am | 13:19 |
| sdake | must have been tired :) | 13:20 |
| sdake | brb reconnecting to vpn | 13:20 |
| sdake | wow these passages malibu people have built an empire | 13:24 |
| sdake | i remember 4-5 years ago they had a "book" on tv | 13:24 |
| sdake | and now they have a big location | 13:24 |
| sdake | and now they accept insurance | 13:24 |
| sdake | (if you coudln't tell I rarely watch cable:) | 13:25 |
| sdake | althoguh I watch roku alot | 13:25 |
| portdirect | lol - I've not had a TV in over ten years | 13:25 |
| portdirect | though spend too much time watching netflix and other streaming services... | 13:26 |
| *** jheroux has joined #openstack-kolla | 13:27 | |
| sdake | wow tv ads are annoying | 13:27 |
| sdake | proactiv | 13:27 |
| sdake | passages malibu | 13:27 |
| sdake | insurance companies | 13:27 |
| sdake | that is all that advertises on discovery channel | 13:28 |
| sdake | i had a recollection about that - but its still the same advertisements | 13:28 |
| portdirect | tuned up build/pushes at my end, so think I can pretty much commit to a nightly build from master for all images - and a weekly for n-1 until we get infra to provide a registry | 13:30 |
| portdirect | takes under 45 mins for each to build/push to docker hub | 13:31 |
| *** lamt has joined #openstack-kolla | 13:34 | |
| *** skramaja has joined #openstack-kolla | 13:34 | |
| *** rafaln has joined #openstack-kolla | 13:36 | |
| *** mattmceuen has quit IRC | 13:37 | |
| *** zhubingbing has joined #openstack-kolla | 13:42 | |
| rafaln | Hi, I have a question regarding Designate component for openstack in Kolla, I would like to test it, but I can't find any information how can I turn it on. I can't simply enable it in globals.yml file like other components, because I can't find it there. Any advice? | 13:43 |
| *** msimonin has joined #openstack-kolla | 13:43 | |
| *** gfidente has quit IRC | 13:44 | |
| *** msimonin has quit IRC | 13:46 | |
| openstackgerrit | Edmund Rhudy proposed openstack/kolla: Enable building without Internet access for most Ubuntu binary containers https://review.openstack.org/414639 | 13:46 |
| *** gfidente has joined #openstack-kolla | 13:46 | |
| openstackgerrit | Edmund Rhudy proposed openstack/kolla: Enable no-Internet build for most Ubuntu binary containers https://review.openstack.org/414639 | 13:46 |
| *** mdnadeem has quit IRC | 13:50 | |
| openstackgerrit | Edmund Rhudy proposed openstack/kolla: Enable no-Internet build for most Ubuntu binary containers https://review.openstack.org/414639 | 13:54 |
| *** mgiles has joined #openstack-kolla | 14:01 | |
| egonzalez90 | rafaln: currently is under review, if you want to test it use this PS https://review.openstack.org/#/c/408465/ | 14:05 |
| egonzalez90 | rafaln: also will need to use this fix to designate images https://review.openstack.org/#/c/413186/ | 14:06 |
| kfox1111 | morning | 14:06 |
| kfox1111 | Jeffrey4l: ping | 14:07 |
| egonzalez90 | rafaln: any input/issue you got in designate let me know, as i'm working on it | 14:08 |
| rafaln | egonzalez90: thank you for the info | 14:13 |
| *** msimonin has joined #openstack-kolla | 14:16 | |
| *** salv-orlando has quit IRC | 14:17 | |
| *** salv-orlando has joined #openstack-kolla | 14:17 | |
| *** aric49 has joined #openstack-kolla | 14:21 | |
| *** rhallisey has joined #openstack-kolla | 14:22 | |
| bjolo | sorry for being lazy, whats the target release date for kolla-ansible ocata? | 14:24 |
| mliima | 2017-02-22 bjolo | 14:25 |
| mliima | *planned* | 14:25 |
| bjolo | but thats for ocata right, will kolla-ansible ocata be ready same day? | 14:25 |
| pbourke | kolla is trailing release | 14:26 |
| pbourke | so some time within 2 weeks after the above date | 14:26 |
| bjolo | ok | 14:26 |
| bjolo | tnx | 14:26 |
| *** salv-orlando has quit IRC | 14:27 | |
| *** fguillot has joined #openstack-kolla | 14:27 | |
| *** mattmceuen has joined #openstack-kolla | 14:28 | |
| bjolo | how many are going to PTG in atlanta? | 14:28 |
| *** skramaja is now known as skramaja|afk | 14:30 | |
| egonzalez90 | bjolo: i'm going | 14:31 |
| bjolo | nice | 14:32 |
| bjolo | im nagging my manager daily :) | 14:32 |
| bjolo | he will soon cave in | 14:33 |
| *** satyar has joined #openstack-kolla | 14:36 | |
| *** rafaln has quit IRC | 14:38 | |
| sdake | bjolo the schedule is here: https://releases.openstack.org/ocata/schedule.html | 14:42 |
| sdake | bjolo the release-trailing date is March 8th | 14:42 |
| bjolo | tnx sdake | 14:43 |
| sdake | np bjolo | 14:43 |
| sbezverk | good morning | 14:44 |
| kfox1111 | morning | 14:44 |
| satyar | Hi jeffrey4l | 14:46 |
| *** gfidente has quit IRC | 14:46 | |
| *** gfidente has joined #openstack-kolla | 14:48 | |
| *** athomas has quit IRC | 14:49 | |
| kfox1111 | sbezverk: I think I'm going to try and go through all the templates and add searchPath definitions to each one. shouldn't take too long this morning. | 14:50 |
| *** msimonin has left #openstack-kolla | 14:50 | |
| kfox1111 | then it would be easier to add searchPath support for the val's more independently? | 14:53 |
| *** rhallisey has quit IRC | 14:53 | |
| *** athomas has joined #openstack-kolla | 14:55 | |
| *** mattmceuen has quit IRC | 14:55 | |
| sbezverk | kfox1111: still not convinced it is right approach, by distributing this work to community it should help to learn better and faster new api ideas. but it is opensource, you do what you think is right ;-) | 14:56 |
| kfox1111 | yeah. I'm not saying I'm going to do it all. | 14:58 |
| kfox1111 | just add one line to the top of each file. | 14:58 |
| sdake | bjolo lol | 14:58 |
| kfox1111 | but, yeah. whoever wants to work on it, can. I don't want to step on anyone's toes. | 14:58 |
| kfox1111 | but also want to get it done if at al possible by next wed, which is really quickly coming up. :/ | 14:59 |
| kfox1111 | sbezverk: so, I kind of asked yesterday. how would you like to break up the work? | 15:00 |
| sdake | amex only allows 1.5 your average 6 month monhtly spend to be outstanding | 15:00 |
| sdake | surprised they gave that data up :) | 15:00 |
| sbezverk | kfox1111: do not worry about me :-) I found one interesting topic which I wan to explore, could be helpfull for future.. | 15:01 |
| sbezverk | I was just tlaking in general | 15:01 |
| kfox1111 | cool. whatcha looking at? | 15:01 |
| sbezverk | kfox1111: custom kube controllers | 15:02 |
| kfox1111 | the 3rd party thing? | 15:03 |
| kfox1111 | (operators) | 15:03 |
| *** rmart04 has quit IRC | 15:03 | |
| kfox1111 | or a different use case? | 15:04 |
| *** bjolo has quit IRC | 15:05 | |
| sbezverk | kfox1111: for now I just want to understand how it works inside, because it seems there are multiple approaches. | 15:05 |
| *** rmart04 has joined #openstack-kolla | 15:06 | |
| sdake | sbezverk there are a few reviews in the queue for operators | 15:06 |
| *** adrian_otto has joined #openstack-kolla | 15:06 | |
| *** zhurong has quit IRC | 15:06 | |
| sdake | sbezverk look for owner ryan | 15:07 |
| sbezverk | sdake: I am not sure if these reviews are now relevant.. | 15:07 |
| kfox1111 | yeah. sounds cool. :) | 15:08 |
| kfox1111 | sbezverk: I think they probably still are. | 15:08 |
| kfox1111 | the helm/entrypoint deployment is one option. operator based deployment is another. | 15:08 |
| *** vhosakot has joined #openstack-kolla | 15:09 | |
| *** mdnadeem has joined #openstack-kolla | 15:10 | |
| *** jtriley has joined #openstack-kolla | 15:10 | |
| *** zhubingbing has quit IRC | 15:12 | |
| *** zhubingbing has joined #openstack-kolla | 15:13 | |
| sdake | sbezverk why wouldn't they be | 15:13 |
| sdake | sbezverk they are just prototype work | 15:13 |
| sdake | sbezverk they need mor eattention | 15:13 |
| sdake | sbezverk we always originally planned to develop operators in python | 15:13 |
| sdake | just to speed up dev | 15:13 |
| *** haplo37 has quit IRC | 15:13 | |
| *** g3ek has quit IRC | 15:13 | |
| *** skramaja|afk has quit IRC | 15:13 | |
| *** duonghq has joined #openstack-kolla | 15:13 | |
| *** skramaja|afk has joined #openstack-kolla | 15:13 | |
| *** sp__ has joined #openstack-kolla | 15:13 | |
| *** g3ek has joined #openstack-kolla | 15:14 | |
| *** adrian_otto1 has joined #openstack-kolla | 15:14 | |
| *** haplo37 has joined #openstack-kolla | 15:14 | |
| sbezverk | sdake: :-) it is like you put together what you want to carry, but have not designed a car to be able to carry it | 15:14 |
| *** zhubingbing has quit IRC | 15:15 | |
| sdake | sbezverk ya - well now that we have the microservices layer | 15:15 |
| sdake | we could potentiallly write an operator to operate it | 15:15 |
| sdake | or some part of it | 15:15 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Fix code style in start_compute.yml https://review.openstack.org/416583 | 15:16 |
| kfox1111 | yeah. all the building blocks are there. now just need someone with interest/time to work on it. | 15:16 |
| *** adrian_otto has quit IRC | 15:17 | |
| sbezverk | sdake: I suggest you check the code coreos put together for their etcd operator, it will give you and idea of the complexity behind it.. | 15:17 |
| sdake | sbezverk i have seen their code | 15:17 |
| sdake | sbezverk i've studied it in great detail | 15:17 |
| sbezverk | we do not have anything even 100 miles close to that state | 15:17 |
| sdake | alot of the complexity has to do with language choice | 15:17 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Fix code style in start_compute.yml https://review.openstack.org/416583 | 15:17 |
| sdake | python shelling out to kubectl should be simple enough | 15:18 |
| *** Pavo has joined #openstack-kolla | 15:18 | |
| kfox1111 | sbezverk: I think the complexity in the etcd is around day2 stuff. hot adding additional etcd's, etc. | 15:19 |
| kfox1111 | openstack's going to have the same issue. with upgrades. | 15:19 |
| sbezverk | kfox1111: that part yes, but even integrating properly with kube is not a trivial thing.. | 15:19 |
| kfox1111 | true. | 15:20 |
| sbezverk | sdake: have you seen any other implementation of operator written on other languges? | 15:20 |
| kfox1111 | the first one's hard. | 15:20 |
| *** srwilkers has joined #openstack-kolla | 15:21 | |
| srwilkers | morning | 15:21 |
| sbezverk | srwilkers: morning | 15:21 |
| kfox1111 | morning | 15:21 |
| srwilkers | the nostalgia is kicking in. internet at work today is very similar to how i felt when i was playing starcraft over dialup back in the day :( | 15:22 |
| openstackgerrit | Merged openstack/kolla: Remove trailing backtick https://review.openstack.org/416447 | 15:23 |
| sbezverk | srwilkers: I bet it was still faster than 1200 baud we used for out first interbanking communication system ;-) | 15:24 |
| *** lrensing has joined #openstack-kolla | 15:24 | |
| srwilkers | sbezverk, you are correct ;) | 15:24 |
| *** vhosakot_ has joined #openstack-kolla | 15:25 | |
| *** vhosakot has quit IRC | 15:26 | |
| *** g3ek has quit IRC | 15:27 | |
| *** Guest66666 has quit IRC | 15:29 | |
| *** absubram has joined #openstack-kolla | 15:30 | |
| *** Guest66666 has joined #openstack-kolla | 15:30 | |
| *** rmart04 has quit IRC | 15:31 | |
| *** goldyfruit has joined #openstack-kolla | 15:36 | |
| *** g3ek has joined #openstack-kolla | 15:37 | |
| *** Pavo has quit IRC | 15:37 | |
| *** Pavo has joined #openstack-kolla | 15:38 | |
| *** inc0 has joined #openstack-kolla | 15:42 | |
| jemcevoy | Pavo: Did you get cobbler to boot UEFI | 15:43 |
| inc0 | good morning | 15:43 |
| *** salv-orlando has joined #openstack-kolla | 15:43 | |
| Pavo | not yet, haven't really had time to work on it | 15:43 |
| jemcevoy | inc0 Morning... Do you work for Intel? | 15:43 |
| inc0 | indeed I do | 15:44 |
| inc0 | last time I checked that is | 15:44 |
| inc0 | haven't seen my emails today | 15:44 |
| Pavo | nope you fired | 15:44 |
| Pavo | lol | 15:44 |
| Pavo | jk | 15:44 |
| inc0 | Pavo, well, let's hope not;) | 15:45 |
| jemcevoy | inc0: In the cluster I am setting up this week and next with Kolla we have the user's home directories on a Luster filesystem accessable through the new Intel Infiband called OPA. | 15:46 |
| inc0 | woho, cool tech | 15:46 |
| jemcevoy | How should I setup that connection? Can I bridge to ib0 on the compute hosts? | 15:46 |
| inc0 | I have no idea in the slightest;) | 15:47 |
| inc0 | I held infinibad gear in my hands once | 15:47 |
| inc0 | thats about it when it comes to my esperience with this tech | 15:47 |
| *** salv-orlando has quit IRC | 15:48 | |
| Pavo | inc0 I was told you was the person to ask when the latest murano and magma fixes would be pushed to the pip package of kolla | 15:48 |
| inc0 | meeting in 10 | 15:48 |
| *** gfidente has quit IRC | 15:48 | |
| inc0 | Pavo, pip gets code when it's released | 15:48 |
| inc0 | Jeffrey4l, are you around? | 15:48 |
| sdake | mgiles did you get unblocked? | 15:49 |
| aric49 | Hello, has anyone experienced issues with the RabbitMQ container in the 3.0.1 kolla release? | 15:49 |
| Pavo | any idea when thats going to be inc0 | 15:49 |
| mgiles | sdake still trying... I've had a few interruptions this morning so going a bit slower than I'd like | 15:49 |
| inc0 | I know we have some gate issues now, I'll have to come back to you. Jeffrey4l is our release liason | 15:49 |
| inc0 | he manages that part | 15:49 |
| aric49 | I am seeing issues where the rabbitmq container is showing "Exited(0)" on a fresh multinode 3.0.1 deployment | 15:50 |
| *** adrian_otto has joined #openstack-kolla | 15:50 | |
| *** adrian_otto1 has quit IRC | 15:51 | |
| *** gfidente has joined #openstack-kolla | 15:51 | |
| *** gfidente has quit IRC | 15:51 | |
| *** gfidente has joined #openstack-kolla | 15:51 | |
| jemcevoy | inc0 Is there anyone in Hillsboro I could ask about connecting IB to openstack? | 15:51 |
| inc0 | uhh, half of Hillsboro is intel | 15:51 |
| sdake | mgiles i am in team meetings all week | 15:51 |
| inc0 | I'm pretty sure there is | 15:51 |
| sdake | mgiles tonight i will be around at 6pm pst and do whatever you need to get unblocked | 15:51 |
| inc0 | but I don't know who | 15:51 |
| inc0 | I'll try to find out | 15:52 |
| mgiles | sdake thanks. I hope to be up and running by then, but will keep you posted | 15:52 |
| sdake | mgiles don't know your schedule or your tz, so not sure if that will work for you or not | 15:52 |
| *** Pavo has quit IRC | 15:52 | |
| *** vhosakot_ has quit IRC | 15:52 | |
| sdake | mgiles i may be away from irc today as well so i can focus on this team meeting | 15:52 |
| Jeffrey4l | sup kfox1111 | 15:52 |
| egonzalez90 | aric49: what kind of issues with rabbit, paste docker logs rabbitmq to paste.openstack.org and logs from rabbit in kolla_logs volume | 15:52 |
| Jeffrey4l | hey inc0, | 15:52 |
| mgiles | sdake I'm EST, so that's a little late, but if I need help I can try to get back on-line then. | 15:53 |
| sdake | mgiles if you could let sbezverk know or Jeffrey4l knwo that would be great befoer you log out | 15:53 |
| mgiles | sdake will do | 15:53 |
| aric49 | engonzalez90, i have the following logs: | 15:53 |
| *** vhosakot has joined #openstack-kolla | 15:53 | |
| sdake | mgiles they are usully aroudn in my rotating 24 hr a day timezone :) | 15:53 |
| inc0 | Jeffrey4l, I got asked several times about us releasing 3.0.2, can you give me brief heads up about that? | 15:53 |
| aric49 | when the container first starts: https://gist.github.com/aric49/8c75adac82449ae26388d4e2ee901381 | 15:53 |
| aric49 | when i try to restart the container: https://gist.github.com/aric49/36c795d4f0df8a7ac8a90467672fd7e2 | 15:54 |
| Jeffrey4l | i think 3.0.2 is ready to go now. | 15:54 |
| Jeffrey4l | but first of all, need sync selinux issue with you. | 15:54 |
| Jeffrey4l | sdake told me you want to add announcement that kolla support selinux. but mitaka branch gate is bust duo to selinux ( disable selinux works ). then can we just release next z stream mitaka now? | 15:55 |
| Jeffrey4l | inc0, ^^ | 15:55 |
| *** Pavo has joined #openstack-kolla | 15:56 | |
| Jeffrey4l | kfox1111, around? | 15:56 |
| egonzalez90 | aric49: no error on that logs, can you paste logs from rabbitmq in kolla_logs volume, should be in /var/lib/docker/volumes/ | 15:57 |
| Jeffrey4l | sup satyar | 15:57 |
| *** Pavo has quit IRC | 15:58 | |
| satyar | jeffrey4l any clue you got for boot from vol | 15:58 |
| Jeffrey4l | i got different error message. i am working that that. | 15:59 |
| inc0 | Jeffrey4l, well nobody created selinux profiles yet | 15:59 |
| satyar | hmmm | 15:59 |
| inc0 | and i'm pretty sure I haven't been releasing any announcements | 15:59 |
| inc0 | especially about that | 15:59 |
| Jeffrey4l | inc0, but enable selinux in gate ( newton branch and master ) works. | 15:59 |
| srwilkers | morning inc0 sdake | 15:59 |
| inc0 | sdake, what did you mean by me announcing anything on that? | 15:59 |
| *** sp__ has quit IRC | 15:59 | |
| sdake | inc0 huh? | 16:00 |
| *** TxGirlGeek has joined #openstack-kolla | 16:00 | |
| sdake | inc0 reading scrollback minute | 16:00 |
| sdake | inc0 you wanted me to change the docs to indicate selinux works (as it doess) | 16:00 |
| Jeffrey4l | sdake, selinux support in kolla. you told me that inc0 wanna support selinux | 16:00 |
| sdake | inc0 i never said anything about announcing support though | 16:00 |
| *** Marcellin__ has joined #openstack-kolla | 16:01 | |
| sdake | inc0 i told jeffrey4l specifically you wanted me to change the docs (which I haven't got to yet) | 16:01 |
| Jeffrey4l | ok. i used a bad word. ;( | 16:01 |
| sdake | Jeffrey4l not a bad word, just a misunderstanding | 16:01 |
| jemcevoy | What is fixed in 3.0.2? Should I be deploying 3.0.2 into production ? | 16:01 |
| Jeffrey4l | yep. | 16:01 |
| inc0 | meeting time | 16:01 |
| sdake | oh is that now? | 16:01 |
| sdake | I thought its in 1 hr | 16:01 |
| Jeffrey4l | jemcevoy, it is hard to say. there are some critial bug is fixed. | 16:01 |
| sdake | i am in a different state, my tz's are afu :) | 16:02 |
| sdake | anyway i wont be able to make it - have to take shower and get to office | 16:02 |
| egonzalez90 | what kind of selinux profiles are needed? we rely on docker, ansible/k8s that have their own selinux policies | 16:02 |
| sdake | see ya all this afternoon | 16:02 |
| sdake | egonzalez90 nothign is needed afaik - it just works in the gate | 16:02 |
| sdake | morning srwilkers | 16:02 |
| Jeffrey4l | jemcevoy, check this page https://launchpad.net/kolla/+milestone/3.0.2 | 16:02 |
| jemcevoy | Thanks Jeffery4l | 16:03 |
| Jeffrey4l | np ;) | 16:03 |
| pbourke | meeting now all | 16:03 |
| pbourke | in you can make it :) | 16:03 |
| pbourke | *if | 16:03 |
| Jeffrey4l | dmsimard, sdake fyi, it is cpu_mode who cause the master gate broken. | 16:04 |
| dmsimard | Jeffrey4l: well there you go | 16:05 |
| dmsimard | cpu_mode fixed it ? | 16:05 |
| dmsimard | er, cpu_mode=none | 16:05 |
| Jeffrey4l | dmsimard, the weird things is that we didn't get any error message like other reporter except that libvirt segment fault. | 16:06 |
| Jeffrey4l | dmsimard, yep. i am pretty sure. | 16:06 |
| Jeffrey4l | which is merged already. | 16:06 |
| dmsimard | It's a complicated problem, too low level for me but there are smarter people working on one or two bugzillas related to the issue | 16:07 |
| Jeffrey4l | dmsimard, https://review.openstack.org/416417 the patch set. | 16:07 |
| Jeffrey4l | i made several recheck to confirm. | 16:07 |
| Jeffrey4l | cool. | 16:07 |
| Jeffrey4l | i also got other error message today in another env. ;( | 16:08 |
| dmsimard | Jeffrey4l: there's https://bugzilla.redhat.com/show_bug.cgi?id=1404627 and https://bugzilla.redhat.com/show_bug.cgi?id=1371617 you can follow if you'd like | 16:09 |
| openstack | bugzilla.redhat.com bug 1404627 in libvirt "OpenStack instances fail to boot with qemu-kvm-rhev 2.6.0 when virt_mode=qemu and cpu_mode=host-model in nova.conf" [Unspecified,Assigned] - Assigned to jdenemar | 16:09 |
| openstack | bugzilla.redhat.com bug 1371617 in libvirt "Libvirt passes unsupported "arat" flag to QEMU when using host-model" [High,Assigned] - Assigned to jdenemar | 16:09 |
| jemcevoy | Jeffrey4l: Is 3.0.2 released? It looks like I hit a few of the fixed issues like fernet and html5 spice | 16:09 |
| dmsimard | in the first bug, they clearly mention "Using "host-model" with TCG is very likely to cause problems. QEMU is being started with -cpu SandyBridge,+osxsave,+hypervisor,+xsaveopt but TCG doesn't support most of the SandyBridge features." | 16:09 |
| Jeffrey4l | jemcevoy, no. but soon. | 16:09 |
| *** sayantan_ has joined #openstack-kolla | 16:09 | |
| Jeffrey4l | dmsimard, i got TCG relate error message today. | 16:10 |
| dmsimard | There's also https://bugs.launchpad.net/qemu/+bug/1652333 | 16:10 |
| openstack | Launchpad bug 1652333 in QEMU "TCG mode fails to boot Linux kernel with qemu 2.6.0 and libvirt 2.0.0" [Undecided,Incomplete] | 16:10 |
| Jeffrey4l | cool. | 16:10 |
| jemcevoy | Is there an announcement list for the release? | 16:10 |
| *** jrich523 has joined #openstack-kolla | 16:11 | |
| Jeffrey4l | jemcevoy, iirc, the announcement mail will be sent automatically. ( not sure ) | 16:11 |
| *** vhosakot_ has joined #openstack-kolla | 16:12 | |
| *** msimonin has joined #openstack-kolla | 16:13 | |
| *** vhosakot has quit IRC | 16:14 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove kolla-genpwd from setup.cfg https://review.openstack.org/416532 | 16:16 |
| *** msimonin1 has joined #openstack-kolla | 16:17 | |
| *** msimonin has quit IRC | 16:17 | |
| *** sp__ has joined #openstack-kolla | 16:18 | |
| *** tis has quit IRC | 16:19 | |
| *** msimonin1 has quit IRC | 16:22 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Remove the "stop-containers" from setup.cfg https://review.openstack.org/416609 | 16:23 |
| *** msimonin has joined #openstack-kolla | 16:24 | |
| *** unicell has quit IRC | 16:29 | |
| *** athomas has quit IRC | 16:32 | |
| *** mattmceuen has joined #openstack-kolla | 16:34 | |
| *** athomas has joined #openstack-kolla | 16:40 | |
| kfox1111 | back | 16:41 |
| *** salkin has quit IRC | 16:41 | |
| kfox1111 | Jeffrey4l: ping | 16:41 |
| inc0 | kfox1111, we have meeting now | 16:42 |
| *** haplo37_ has quit IRC | 16:42 | |
| *** sdake has quit IRC | 16:44 | |
| *** haplo37_ has joined #openstack-kolla | 16:45 | |
| *** jtriley has quit IRC | 16:48 | |
| *** gfidente has quit IRC | 16:48 | |
| *** mdnadeem has quit IRC | 16:48 | |
| *** Pavo has joined #openstack-kolla | 16:49 | |
| *** Pavo has quit IRC | 16:50 | |
| *** gfidente has joined #openstack-kolla | 16:50 | |
| *** sayanta__ has joined #openstack-kolla | 16:57 | |
| *** sayantan_ has quit IRC | 17:00 | |
| *** sayanta__ has quit IRC | 17:00 | |
| *** sayantan_ has joined #openstack-kolla | 17:00 | |
| inc0 | bbiaf guys, moving to office | 17:01 |
| Jeffrey4l | SamYaple, so could u remove your -1 now ? | 17:01 |
| *** inc0 has quit IRC | 17:01 | |
| Jeffrey4l | cores please review https://review.openstack.org/406978 :) | 17:01 |
| duonghq | night guys | 17:01 |
| Jeffrey4l | duonghq, night | 17:01 |
| SamYaple | Jeffrey4l: removed, just throw up a link to that discussion if yo uwould please | 17:02 |
| *** duonghq has quit IRC | 17:02 | |
| Jeffrey4l | SamYaple, will do. thanks. | 17:02 |
| Jeffrey4l | egonzalez90, pbourke please review this ;) https://review.openstack.org/406978 | 17:03 |
| *** salv-orlando has joined #openstack-kolla | 17:04 | |
| *** mdnadeem has joined #openstack-kolla | 17:04 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Support multi key in merge_config module https://review.openstack.org/412101 | 17:06 |
| *** zhugaoxiao has quit IRC | 17:07 | |
| *** zhugaoxiao has joined #openstack-kolla | 17:07 | |
| *** sdake has joined #openstack-kolla | 17:08 | |
| *** vhosakot_ has quit IRC | 17:09 | |
| *** sdake has quit IRC | 17:13 | |
| *** JoseMello has joined #openstack-kolla | 17:14 | |
| *** msimonin has quit IRC | 17:17 | |
| Jeffrey4l | satyar, still around? | 17:17 |
| openstackgerrit | Mark Giles proposed openstack/kolla-kubernetes: Fix typo in variable name https://review.openstack.org/416673 | 17:17 |
| *** msimonin has joined #openstack-kolla | 17:17 | |
| *** msimonin has quit IRC | 17:17 | |
| *** msimonin has joined #openstack-kolla | 17:18 | |
| *** msimonin has quit IRC | 17:18 | |
| *** msimonin has joined #openstack-kolla | 17:18 | |
| *** msimonin has quit IRC | 17:18 | |
| mgiles | kfox1111 The image you pushed allowed me to work around the problem (once I noticed the typo fixed by the ps I just put up). Thanks | 17:18 |
| *** bmace has joined #openstack-kolla | 17:18 | |
| *** msimonin has joined #openstack-kolla | 17:18 | |
| mgiles | Jeffrey4l sbezverk - If sdake asks you later you can tell him I'm unblocked - he asked me to let you guys know. | 17:20 |
| Jeffrey4l | roger. but i am going offline soon ;) | 17:20 |
| kfox1111 | mgiles: great. good catch on the typo. | 17:21 |
| kfox1111 | mgiles: Jeffrey4l said 3.0.2 should be released by the end of the week, so should be able to drop the workaround very soon. | 17:22 |
| *** msimonin has quit IRC | 17:22 | |
| Jeffrey4l | kfox1111, why you are depends on 3.0.2 tag image rather then master images? | 17:23 |
| mgiles | kfox1111 saw that during the meeting. that'll be good | 17:23 |
| kfox1111 | Jeffrey4l: was easier to get the gate up and running with released packages. | 17:23 |
| kfox1111 | Jeffrey4l: will add support for the master gate packages that are getting built now, soon. | 17:23 |
| Jeffrey4l | roger. cool. | 17:24 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Updated from global requirements https://review.openstack.org/416199 | 17:24 |
| *** msimonin has joined #openstack-kolla | 17:25 | |
| *** matrohon has quit IRC | 17:31 | |
| *** Pavo has joined #openstack-kolla | 17:31 | |
| *** mdnadeem has quit IRC | 17:37 | |
| *** mdnadeem has joined #openstack-kolla | 17:41 | |
| *** rhallisey has joined #openstack-kolla | 17:42 | |
| *** msimonin has quit IRC | 17:46 | |
| *** sean-k-mooney has joined #openstack-kolla | 17:51 | |
| *** sp__ has quit IRC | 17:51 | |
| *** mdnadeem has quit IRC | 17:53 | |
| *** vhosakot has joined #openstack-kolla | 18:05 | |
| jemcevoy | sean-k-mooney: Do you have any ideas on how configure Neutron for Intel OPA infiband to the instances to mount Luster home directories? Each VM host has ib0 with IPoIB | 18:09 |
| erhudy | sean-k-mooney: since you own https://bugs.launchpad.net/kolla/+bug/1633187, would you be so kind as to look at https://bugs.launchpad.net/kolla/+bug/1653740, which I ran into when trying to fix the former | 18:19 |
| openstack | Launchpad bug 1633187 in kolla "kolla build and dockerfile filenames do not match for sources.list" [Low,Confirmed] - Assigned to sean mooney (sean-k-mooney) | 18:19 |
| openstack | Launchpad bug 1653740 in kolla "Kolla Ubuntu gates rely on bug in Kolla to work" [Undecided,New] | 18:19 |
| sean-k-mooney | jemcevoy: no sorry. im not sure if that is supported | 18:23 |
| sean-k-mooney | erhudy: im taking a look at it now | 18:24 |
| erhudy | thanks | 18:24 |
| jemcevoy | sean-k-mooney: Do you I could route to the local IB port? | 18:24 |
| sean-k-mooney | erhudy: that does not mean i know how to fix the bug you linked :) | 18:25 |
| erhudy | yeah, i've just been trying to get somebody's attention on it because i'm pretty sure it's a problem in the way the gates are configured | 18:25 |
| erhudy | and OSIC's ubuntu mirror being unsigned | 18:25 |
| sean-k-mooney | jemcevoy: what neutron backand are you using | 18:25 |
| jemcevoy | ovs | 18:26 |
| erhudy | unsigned or not being signed with a key that is added to the base image | 18:26 |
| jemcevoy | and vxlan | 18:26 |
| sean-k-mooney | jemcevoy: and how do you intend to use infiniband in that setup? | 18:26 |
| sean-k-mooney | vxlan over infinaband or infiniband interface passed through to the guest? | 18:26 |
| sean-k-mooney | erhudy: ya if the osic mirror has an ssl cert then we sould need to add the key. | 18:27 |
| erhudy | i think up to this point the gate has worked mostly by accident because the sources.list override isn't actually used, so the base image defaults to using canonical's mirrors | 18:29 |
| sean-k-mooney | jemcevoy: it sounds like you want to do a pci pasthrough of the infiniband interface to the guest which is done via nova not neutorn | 18:29 |
| jemcevoy | home directories are on on Luster on the infiniband network the which is used on the 4032 core HPC cluster... So the data to be processed in the beowulf cluster is in those home directories | 18:29 |
| jemcevoy | sean-k-mooney: All the login node VMs need access to Luster | 18:30 |
| sean-k-mooney | the issue is i dont think ovs supports infiniband. | 18:30 |
| sean-k-mooney | if it was a normal network interface i would say add the infinaband port to hte bridge and configre a flat network provider network and then add a second interface to the vm to connect over the infinaband network | 18:31 |
| *** eaguilar has joined #openstack-kolla | 18:31 | |
| erhudy | sean-k-mooney: should i try to talk to people in openstack-infra about the gate setups? i don't know who's actually responsible | 18:33 |
| *** adrian_otto has quit IRC | 18:33 | |
| sean-k-mooney | erhudy: the reason i have not done much on https://bugs.launchpad.net/kolla/+bug/1633187 bar https://review.openstack.org/#/c/386592/ was i was waiting for the gate to be fixed. i taught Jeffrey4l might be fixing it but i guess not | 18:34 |
| openstack | Launchpad bug 1633187 in kolla "kolla build and dockerfile filenames do not match for sources.list" [Low,Confirmed] - Assigned to sean mooney (sean-k-mooney) | 18:34 |
| jemcevoy | I'll be doing cpi passthrough for the K2 Grid cards used with our remote viewer https://pod.penguincomputing.com/hpc_on_demand#rviz | 18:34 |
| jemcevoy | cpi -> pci | 18:34 |
| sean-k-mooney | jemcevoy: well what you are really trying to do is expose an interface to the guest so that they can mount a remote file system over infiniband. | 18:36 |
| sean-k-mooney | jemcevoy: normally you would use cinder to do that | 18:36 |
| *** sdake has joined #openstack-kolla | 18:36 | |
| *** unicell has joined #openstack-kolla | 18:38 | |
| sean-k-mooney | jemcevoy: i dont know if there is an intel equivalent but this would appear to be one option on the mellanox side but i dont know how it works https://wiki.openstack.org/wiki/Mellanox-Neutron-Newton-InfiniBand | 18:41 |
| *** Pavo has quit IRC | 18:41 | |
| sean-k-mooney | jemcevoy: ah they support infiniband as a transport for sriov nics | 18:42 |
| *** sayantan_ has quit IRC | 18:42 | |
| jemcevoy | I have not used sriov where can I read up on it? | 18:44 |
| sean-k-mooney | jemcevoy: so this is not what you want correct. if you have an intel infiniband adapter that you can allocate virtual functions from you may be able to use the sriov nic agent to enable the same functionality but i dont thinkt that has ever been tested | 18:44 |
| jemcevoy | is it different from pci passthrough? | 18:44 |
| *** inc0 has joined #openstack-kolla | 18:44 | |
| inc0 | back | 18:45 |
| sean-k-mooney | sriov (single route io virutalisation) allows a singel phyical interface to be shared by multiple guests | 18:45 |
| sean-k-mooney | it is similar to pci passthrough | 18:45 |
| *** vhosakot has quit IRC | 18:45 | |
| *** vhosakot has joined #openstack-kolla | 18:46 | |
| *** sayantan_ has joined #openstack-kolla | 18:46 | |
| sean-k-mooney | jemcevoy: sriov has very similar performance and latency chariteristics of pci passthough as it is a hardware feature so there is no software overhead | 18:46 |
| *** sayantan_ has quit IRC | 18:47 | |
| *** egonzalez90 has quit IRC | 18:47 | |
| inc0 | sean-k-mooney, long time no see:) | 18:47 |
| sean-k-mooney | inc0: yes just back form vacation today | 18:48 |
| sean-k-mooney | inc0: how about you? did you have a good break | 18:48 |
| inc0 | busy | 18:48 |
| inc0 | but very nice | 18:48 |
| inc0 | 'murica is beautiful | 18:48 |
| sean-k-mooney | thats good too hear. hope you were busy with non kolla stuff | 18:49 |
| sean-k-mooney | ill be hopefully rebasing all the dpdk pathes this week. the repo split makes things a little more confuing | 18:51 |
| sean-k-mooney | comeing back to coding and trying to remember where you left everthing is hard after 2 weeks | 18:52 |
| *** pcaruana has quit IRC | 18:52 | |
| *** sdake has quit IRC | 18:54 | |
| *** adrian_otto has joined #openstack-kolla | 18:55 | |
| srwilkers | oh inc0 youre back | 18:56 |
| inc0 | that I am | 18:56 |
| srwilkers | at the end of the meeting, you mentioned someone donating effort/hardware/whatever for building images routinely. portdirect has volunteered more than a few times to do this, and his remarks have gone largely unanswered | 18:57 |
| kfox1111 | +1 for portdirect building for now. | 19:00 |
| kfox1111 | his results seemed to indicate we might be able to move it into zuul too. | 19:00 |
| kfox1111 | just either need someone to do the legwork, or I can take a stab at it later. | 19:01 |
| inc0 | if we have hub.o.o that would work too | 19:01 |
| inc0 | anyway, for now we need cronjob to do kolla-build and docker push | 19:01 |
| kfox1111 | we would still need a way to build/refresh/push to hub.o.oo. | 19:02 |
| inc0 | I can give push privileges to any core willing | 19:02 |
| kfox1111 | inc0: please give portdirect push privileges. | 19:02 |
| inc0 | kfox1111, if we can write puppet script to setup mirrored registry | 19:02 |
| inc0 | that's what we need | 19:02 |
| inc0 | yeah, I just need his dockerhub username | 19:02 |
| kfox1111 | inc0: we're talking about more then just gate jobs. | 19:02 |
| kfox1111 | inc0: but keeping up to date images on the hub. | 19:03 |
| kfox1111 | (or wherever) | 19:03 |
| kfox1111 | portdirect: ping :) | 19:03 |
| portdirect | pong - 2 mins | 19:03 |
| kfox1111 | wow. you figured out how to bend time. you gotta tell me how to do that. :) | 19:03 |
| srwilkers | portdirect, http://giphy.com/gifs/filmeditor-the-legend-of-sleepy-hollow-l2YWl26uPafdrZVC0 | 19:04 |
| kfox1111 | the only way I know of involves being far to close to comfort to a black hole.... I'd rather not.. :) | 19:04 |
| inc0 | portdirect, do you have dickerhub account? | 19:05 |
| inc0 | oh, I made a typo... | 19:05 |
| inc0 | o and i are so close to each other | 19:05 |
| portdirect | wow | 19:05 |
| portdirect | srwilkers's finished stuffing my face :) | 19:05 |
| kfox1111 | inc0: heh. | 19:06 |
| *** sp__ has joined #openstack-kolla | 19:06 | |
| kfox1111 | my dentist is angus square dental. | 19:07 |
| kfox1111 | the 'g' is important. :) | 19:07 |
| kfox1111 | made a similar mistake once. :) | 19:07 |
| mliima | https://review.openstack.org/#/c/411794/ need act here | 19:07 |
| *** Bico_Fino has quit IRC | 19:08 | |
| srwilkers | inc0, the amount of 'docker' commands ive mistyped is too high to be proud of :( | 19:08 |
| portdirect | inc0: so what should I be pushing? | 19:09 |
| portdirect | a build from master for each of centos/ubuntu * binary/source | 19:09 |
| lrensing | o.o | 19:09 |
| portdirect | and tagged with master or latest? | 19:09 |
| *** vhosakot_ has joined #openstack-kolla | 19:10 | |
| *** prameswar has quit IRC | 19:10 | |
| *** vhosakot has quit IRC | 19:10 | |
| kfox1111 | portdirect: what I really would like to see, is builds from current release, | 19:10 |
| kfox1111 | with tags updated when changed. | 19:10 |
| kfox1111 | so, 3.0.1-1, 3.0.1-2, 3.0.1-3, etc. | 19:10 |
| inc0 | that's good idea | 19:11 |
| inc0 | I like this structure | 19:11 |
| kfox1111 | trunk builds can be pulled from tarballs.o.o | 19:11 |
| srwilkers | i agree. thats a good approach | 19:11 |
| portdirect | could one of you provide me a kolla-build command example and I'll get on it :) | 19:12 |
| kfox1111 | there's a bit of an etherpad here: https://etherpad.openstack.org/p/JSuWFXF5lh | 19:12 |
| kfox1111 | portdirect: the sticky part is, there isn't a machanism in place to detect if anyh changes were made from the last container. | 19:12 |
| kfox1111 | so we need some kind of diff-ing mechanism. | 19:13 |
| portdirect | yeah - i was imagining somthing simpler, eg this is what ive been running: https://gist.githubusercontent.com/intlabs/0a9f976496dbfb83c7bd5a4e5df0de42/raw/3b7f6fd7d800f7b32207c416b10041ad3807afee/gistfile1.txt | 19:13 |
| kfox1111 | I think rpm -qa | md5sum or (rpm -qa; pip list ) | md5sum would do something like it though. | 19:13 |
| srwilkers | kfox1111, thats what i was thinking. that should work | 19:14 |
| portdirect | (I have another job that cleans up all docker images that runs prior to that) | 19:14 |
| kfox1111 | portdirect: yeah. that looks like a good start. :) | 19:14 |
| kfox1111 | would just need to add some logic in between to make it skip the push if things didn't change, | 19:15 |
| kfox1111 | and a way to keep track of which revision it was up to. | 19:15 |
| *** sp__ has quit IRC | 19:15 | |
| sean-k-mooney | kfox1111: out of interest why 3.0.1-1 and not 3.0.1.1? | 19:15 |
| *** sdake has joined #openstack-kolla | 19:15 | |
| kfox1111 | sean-k-mooney: convention in most pckage managers is to have 2 seperate versions. | 19:16 |
| *** sdake has quit IRC | 19:16 | |
| kfox1111 | use the version of the software , followed by a dash, and then the version of the build/packaging. | 19:16 |
| portdirect | yeah - i was kinda volentering to have a rolling master (ie 4.0.0 nightly) for us to use | 19:16 |
| sean-k-mooney | kfox1111: ah ok that makes sense | 19:17 |
| kfox1111 | since the build/packaging versioning is usually simple, its most of the time just an integer. | 19:17 |
| kfox1111 | though it can get pretty crazy. :) | 19:17 |
| kfox1111 | see redhat's kernel package as an example. :) | 19:17 |
| kfox1111 | 3.10.0-229.14.1.el7 | 19:18 |
| kfox1111 | so, 3.10.0 of the kernel, 229.14.1.el7 of the packaging. :) | 19:18 |
| sean-k-mooney | inc0: are the gates now caching the images now to tarballs.openstack.org. could you set up a nightly cron job to pull the registry and push form a vm that can update docker hub. | 19:18 |
| sean-k-mooney | inc0: that is assuming that portdirect does not build them :) | 19:19 |
| kfox1111 | the job thats pushing them into tarballs.openstack.org could potentially push them to the hub too. | 19:19 |
| inc0 | sean-k-mooney, someday we'll do it;) | 19:19 |
| inc0 | kfox1111, that's a not really - issue is uplink consumption | 19:20 |
| inc0 | might be | 19:20 |
| inc0 | nightly would be ok, or weekly | 19:20 |
| inc0 | but every commit might be a tad too much | 19:20 |
| portdirect | so just to clarify - I can build and push a set a single set of all images on a daily basis - and for that I was envisioning a master. I could set up a build for 3.0.x-fu.bar on a weekly if thats ok? | 19:20 |
| kfox1111 | inc0: I think its building only when trunk gets updated. | 19:21 |
| kfox1111 | inc0: which is not very frequent for the containers? | 19:21 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla: Fix wrong register in ceph reconfigure task https://review.openstack.org/416713 | 19:21 |
| *** aric49 has quit IRC | 19:21 | |
| sean-k-mooney | portdirect: were you asking for permission or makeing a statement of what you were willing/able to do? | 19:23 |
| *** aric49 has joined #openstack-kolla | 19:23 | |
| portdirect | sean-k-mooney: little bit of a little bit of b :) | 19:23 |
| portdirect | would need help setting up the 3.0.0x script | 19:23 |
| portdirect | and also stating where Im willing to put my bandwidth | 19:24 |
| *** matrohon has joined #openstack-kolla | 19:24 | |
| *** vhosakot_ has quit IRC | 19:24 | |
| *** msimonin has joined #openstack-kolla | 19:24 | |
| inc0 | once we get out tarballs ready | 19:25 |
| *** vhosakot has joined #openstack-kolla | 19:25 | |
| portdirect | as from my personal persperctive having trun images is most important to me - and I hope that the released images are not required at such a cadence | 19:25 |
| inc0 | let's discuss it more with infra to make it proper | 19:25 |
| sean-k-mooney | well i dont see anything wrong with haveing a different cdence of master vs stable tags (weekly sounds resonable) as for pushing them ill leave that to someone more qualified | 19:25 |
| *** msimonin has quit IRC | 19:26 | |
| *** msimonin has joined #openstack-kolla | 19:26 | |
| *** msimonin has quit IRC | 19:27 | |
| kfox1111 | portdirect: the upload cadence isn't the important thing to me, its that updates happen soon after updated packages land. | 19:27 |
| sean-k-mooney | actully on the bandwidt side when you push to the docker hub is it pushhing the layerd deltas or does it push a flat tar? i would hope deltas? | 19:27 |
| kfox1111 | portdirect: so, the next openssl volunerability shows up, redhat releases a fix, kolla containers bump to 3.0.1-N+1. | 19:27 |
| portdirect | so putting it out there - I'm gonna push master lagged 'master' and if someone sends me a script to build updated release images I'm happy to run that untill at least the end of this dev cycle | 19:27 |
| kfox1111 | k. | 19:28 |
| kfox1111 | fair enough. :) | 19:28 |
| *** adrian_otto has quit IRC | 19:28 | |
| kfox1111 | my interest is more from an operator perspective, not a trunk one. | 19:28 |
| *** adrian_otto has joined #openstack-kolla | 19:28 | |
| *** Serlex has quit IRC | 19:29 | |
| *** vhosakot has quit IRC | 19:29 | |
| kfox1111 | each use case has its benifits. | 19:29 |
| sean-k-mooney | kfox1111: do you think operators will use vanilla images build upstream in production or are they more likely to have customistatiions | 19:29 |
| portdirect | kfox1111: I cannot take that on :( if you need that I can happily rent you bandwith :) | 19:29 |
| *** sdake has joined #openstack-kolla | 19:29 | |
| kfox1111 | sean-k-mooney: I would, if I knew they were kept up to date. | 19:30 |
| portdirect | kfox1111: this is for the developer community | 19:30 |
| kfox1111 | sean-k-mooney: I've only had to patch a few packages before. and those are easier to maintain just one or two rather then the whole set. | 19:30 |
| kfox1111 | portdirect: I'm going to try and take it on in the gate at some point. trying to get 0.4.0 out th edoor first. :) | 19:31 |
| sean-k-mooney | kfox1111: i guess if you did need some custimiations perhaps a horizon plugin you could buld just that one image and use the rest upstream | 19:31 |
| kfox1111 | portdirect: but I do thank you for the offer. :) | 19:31 |
| *** vhosakot has joined #openstack-kolla | 19:31 | |
| sean-k-mooney | kfox1111: ya | 19:31 |
| kfox1111 | sean-k-mooney: yeah. most of the time, what one op needs ,the rest do too. | 19:31 |
| *** vhosakot has quit IRC | 19:32 | |
| kfox1111 | contributing back as much as possible's a good thing. :) | 19:32 |
| sdake | hey peeps | 19:32 |
| sean-k-mooney | i think the neutron-base image and horizon are the two container that would be changed most often | 19:32 |
| sean-k-mooney | sdake: hi | 19:32 |
| sdake | hey sean | 19:32 |
| kfox1111 | yeah. we've had to customize horizon a few times. nothing major though. | 19:32 |
| kfox1111 | just inhereted from kolla's and tweaked it slightly. | 19:32 |
| kfox1111 | I guess I can see neutron needing it too, depending on what plugins you use. | 19:33 |
| kfox1111 | though I think in either case, it would be good to add support upstream to kolla when possible. | 19:33 |
| sean-k-mooney | sdake: did you have a good holiday season to be pc. | 19:33 |
| sdake | sean-k-mooney i survived it :) | 19:34 |
| sdake | sean-k-mooney 2016 was an onerous year | 19:34 |
| sean-k-mooney | sdake: well that is always a good first step. | 19:34 |
| sean-k-mooney | well im sure 2017 being prime will make it a much better year | 19:34 |
| *** dmsimard has quit IRC | 19:35 | |
| sdake | sean-k-mooney god you would hope so :) | 19:35 |
| *** dmsimard has joined #openstack-kolla | 19:35 | |
| sdake | prime number years cecoming harder and harder to find :) | 19:35 |
| *** negronjl has quit IRC | 19:36 | |
| *** Mech422 has quit IRC | 19:36 | |
| *** manjeets has quit IRC | 19:36 | |
| sean-k-mooney | kfox1111: well as you said its easy to have a container that inherits form the upstream on on the docker hub and jsut addes the plugin if you wanted to miniuse what you build locally | 19:36 |
| *** Mech422 has joined #openstack-kolla | 19:36 | |
| *** negronjl has joined #openstack-kolla | 19:38 | |
| sean-k-mooney | ok i better head home and cook dinner. see everyone tomorow. | 19:38 |
| sean-k-mooney | oh by the way do people know if the ptg has a cutoff for registration. i should know shortly if travel is approved. | 19:39 |
| kfox1111 | sean-k-mooney: I don't think so, but I think there are limited tickets. | 19:40 |
| *** aric49 has quit IRC | 19:40 | |
| sean-k-mooney | kfox1111: well they are still availabel to book but since the refund date is February 16, 2017 i would guess they wil close the registration then also but i could be wrong | 19:40 |
| *** vhosakot has joined #openstack-kolla | 19:41 | |
| kfox1111 | yeah. that sounds reasonable. | 19:42 |
| *** msimonin has joined #openstack-kolla | 19:42 | |
| *** vhosakot_ has joined #openstack-kolla | 19:43 | |
| *** manjeets has joined #openstack-kolla | 19:44 | |
| SamYaple | inc0: where do we stand on kolla-salt stuff. the repo is now in full POC for infra services to deploying keystone | 19:44 |
| *** vhosakot has quit IRC | 19:46 | |
| *** msimonin has quit IRC | 19:52 | |
| *** msimonin1 has joined #openstack-kolla | 19:52 | |
| *** aric49 has joined #openstack-kolla | 19:52 | |
| inc0 | SamYaple, I'll ask TC how to deal with iot | 19:55 |
| kfox1111 | SamYaple: is there anyone other then you interested in supporting it? | 19:55 |
| inc0 | :D | 19:55 |
| inc0 | that's a good question really | 19:56 |
| inc0 | who will give you +2s? | 19:56 |
| *** adrian_otto has quit IRC | 19:56 | |
| kfox1111 | or, are you the only one, and your hoping more will come once contributed? | 19:56 |
| inc0 | when you get to kolla, you'll need to follow the process, which might not be possible without any community | 19:56 |
| *** adrian_otto has joined #openstack-kolla | 19:56 | |
| kfox1111 | inc0: yeah. probably need at least 2 folks minimum to bootstrap? | 19:57 |
| inc0 | well, technicaly it should be 3 | 19:57 |
| inc0 | one proposes patch, 2 approves | 19:57 |
| kfox1111 | it took 2 at least to do kolla-kubernetes. | 19:57 |
| inc0 | no, it was much more then | 19:57 |
| kfox1111 | inc0: yeah. we bent the rules sligtly to get over the hump in kolla-kubernetes. | 19:58 |
| kfox1111 | we all agreed to a tempoarary 1 + 1 until enough cores were premoted. | 19:58 |
| kfox1111 | once portdirect and srwilkers got promoted, we went back to 1 + 2. | 19:58 |
| inc0 | yeah | 19:59 |
| kfox1111 | otherwise, we were waiting multiple days per ps. | 19:59 |
| kfox1111 | and it was backing up very badly. | 19:59 |
| inc0 | so think twice SamYaple if you want to have gerrit:) | 19:59 |
| SamYaple | kfox1111: ive heard several people experss intrest | 20:01 |
| kfox1111 | SamYaple: cool. maybe see if any of them want to be cores? | 20:01 |
| *** mattmceuen has quit IRC | 20:02 | |
| *** athomas has quit IRC | 20:02 | |
| SamYaple | as far as merging, self-merging is definetely a thing | 20:02 |
| SamYaple | for smaller projects | 20:02 |
| sean-k-mooney | self-merging is a thing but once there are 2 or 3 cores its better to avoid for anything but a trival fix | 20:03 |
| kfox1111 | inc0: so the question is, whats the minimum number of cores kolla wants to see to see a new deliverable in kolla? | 20:04 |
| SamYaple | sean-k-mooney: completely agree | 20:05 |
| kfox1111 | though I guess that may be a meteric that doesn't matter in reality. kolla-kubernetes had a bunch of cores, but only 2 committers/alive reviewers for a while. | 20:05 |
| sean-k-mooney | kfox1111: i think a project can bootstrap with 1 but ideally 3 so that there isnt a singel point of failure if one of the reviews is unavialble for a while | 20:06 |
| SamYaple | yea starting projects always gain followers (and lose them) as they grow | 20:07 |
| SamYaple | hell, there are some openstack projects that go entire cycles _without_ committers | 20:07 |
| * portdirect wasn't promoted, he just had more work dumped on his plate... | 20:07 | |
| *** Slower has joined #openstack-kolla | 20:07 | |
| SamYaple | portdirect: sounds like a promotion to me | 20:08 |
| kfox1111 | portdirect: hehe. your on to us. ;) | 20:08 |
| Slower | SamYaple!! | 20:08 |
| Slower | how goes it sam? | 20:08 |
| *** salv-orlando has quit IRC | 20:08 | |
| SamYaple | Slower! my man! | 20:08 |
| SamYaple | it goes ok | 20:08 |
| SamYaple | you? | 20:08 |
| Slower | yeah good.. can't complain :) | 20:08 |
| SamYaple | well you can, but you dont. and thats why i like you | 20:08 |
| Slower | hehe true :) | 20:08 |
| *** eaguilar has quit IRC | 20:09 | |
| portdirect | SamYaple: You've worked for the wrong companies dude - I expect cigars and yachts.... not a four extra radio buttons... | 20:09 |
| Slower | hehe | 20:09 |
| sbezverk | kfox1111: how did you like the comment we got today from helm guy about microservices ;-) ? | 20:10 |
| Slower | so I have a potentially dumb question.. I see these customizable() things in the jinja templates but it's not clear how I set them? | 20:10 |
| SamYaple | Slower: i have the same question, ive just been ignoring it | 20:11 |
| Slower | is there an example someplace? | 20:11 |
| sbezverk | kfox1111: I think it is indication that we are on the right track | 20:11 |
| kfox1111 | sbezverk: +1 :) | 20:11 |
| Slower | SamYaple: ah, hmmm :) | 20:11 |
| *** gfidente is now known as gfidente|afk | 20:23 | |
| *** jtriley has joined #openstack-kolla | 20:26 | |
| *** jtriley has quit IRC | 20:31 | |
| kfox1111 | sbezverk: incoming. :) | 20:31 |
| *** sayantan_ has joined #openstack-kolla | 20:32 | |
| srwilkers | i missed the call today. what was the comment re: microservices? | 20:32 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Add default searchPath to almost all templates. https://review.openstack.org/416729 | 20:33 |
| kfox1111 | well, missed a few it looks like... hmmm... | 20:35 |
| sayantan_ | Hi guys | 20:36 |
| SamYaple | hi sayantan_ | 20:38 |
| sayantan_ | I'm trying to get bifrost running. While trying to bootstrap and start services in the bifrost container, I get the error http://paste.openstack.org/raw/593907/ | 20:38 |
| sayantan_ | Hi SamYaple | 20:38 |
| *** nwik has joined #openstack-kolla | 20:39 | |
| *** lamt has quit IRC | 20:39 | |
| sayantan_ | Do you have an idea why the bifrost container might fail with file system being locked errors? | 20:39 |
| *** sdake has quit IRC | 20:40 | |
| srwilkers | sbezverk, ping | 20:40 |
| *** jtriley has joined #openstack-kolla | 20:43 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Add default searchPath to almost all templates. https://review.openstack.org/416729 | 20:43 |
| kfox1111 | hmm... still missed one file.... | 20:44 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Add default searchPath to almost all templates. https://review.openstack.org/416729 | 20:44 |
| kfox1111 | there we go. | 20:45 |
| *** gfidente|afk has quit IRC | 20:45 | |
| *** eaguilar has joined #openstack-kolla | 20:47 | |
| kfox1111 | heh.... http://opensource.googleblog.com/2017/01/grumpy-go-running-python.html | 20:48 |
| inc0 | Slower, :) now we're almost neighbors:) | 20:50 |
| inc0 | kfox1111, so I have no idea if there is any openstack rule for minimum number of cores | 20:53 |
| kfox1111 | k. was just curious. maybe there isnt any. | 20:54 |
| *** JoseMello has quit IRC | 20:55 | |
| *** sayantan_ has quit IRC | 20:55 | |
| *** sayantan_ has joined #openstack-kolla | 20:56 | |
| SamYaple | inc0: there is not | 20:56 |
| Slower | inc0: oh ya!? where'd you go again? oregon? | 21:02 |
| inc0 | yeah | 21:03 |
| inc0 | near Portland | 21:03 |
| Slower | yeah nice area | 21:04 |
| inc0 | it is very nice | 21:04 |
| inc0 | my kind of place:) | 21:04 |
| Slower | awesome | 21:04 |
| Slower | still doing the paragliding? | 21:04 |
| inc0 | trying to find people around | 21:05 |
| *** aric49 has quit IRC | 21:05 | |
| inc0 | here I could | 21:05 |
| inc0 | in TX not so much | 21:05 |
| inc0 | so if I find local pilots, I'm going back to it on spring | 21:06 |
| Slower | cool | 21:06 |
| *** sayanta__ has joined #openstack-kolla | 21:09 | |
| *** sayantan_ has quit IRC | 21:12 | |
| *** srwilkers has quit IRC | 21:14 | |
| *** salv-orlando has joined #openstack-kolla | 21:23 | |
| openstackgerrit | Merged openstack/kolla-ansible: Support multi key in merge_config module https://review.openstack.org/412101 | 21:23 |
| *** shardy has quit IRC | 21:26 | |
| *** shardy has joined #openstack-kolla | 21:28 | |
| *** sdake has joined #openstack-kolla | 21:28 | |
| openstackgerrit | Merged openstack/kolla-ansible: Optimize reconfigure action for glance https://review.openstack.org/406978 | 21:31 |
| sdake | sup peeps | 21:31 |
| *** l4yerffeJ_ has joined #openstack-kolla | 21:32 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** l4yerffeJ has quit IRC | 21:35 | |
| *** lrensing has quit IRC | 21:36 | |
| openstackgerrit | Merged openstack/kolla-ansible: Optimize reconfigure action for memcached https://review.openstack.org/412225 | 21:39 |
| *** adrian_otto1 has joined #openstack-kolla | 21:42 | |
| *** adrian_otto has quit IRC | 21:45 | |
| *** adrian_otto1 has quit IRC | 21:47 | |
| *** Jeffrey4l has joined #openstack-kolla | 21:49 | |
| *** lamt has joined #openstack-kolla | 21:50 | |
| *** satyar has quit IRC | 21:52 | |
| *** salv-orl_ has joined #openstack-kolla | 21:54 | |
| kfox1111 | sbezverk: portdirect: https://review.openstack.org/#/c/416729/ thoughts? | 21:54 |
| *** TxGirlGeek has quit IRC | 21:55 | |
| *** TxGirlGeek has joined #openstack-kolla | 21:56 | |
| *** salv-orlando has quit IRC | 21:56 | |
| *** matrohon has quit IRC | 21:58 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fix the mariadb/rabbitmq elements and add searchPaths. https://review.openstack.org/416744 | 22:00 |
| *** TxGirlGeek has quit IRC | 22:01 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:01 | |
| *** srwilkers has joined #openstack-kolla | 22:02 | |
| *** adrian_otto has joined #openstack-kolla | 22:03 | |
| *** fguillot has quit IRC | 22:03 | |
| *** TxGirlGeek has quit IRC | 22:03 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:04 | |
| sbezverk | kfox1111: you are kidding right ;-) ?? PS of this size to review it is crazy :-) | 22:04 |
| kfox1111 | sbezverk: I can break it up into any number of chunks. | 22:06 |
| *** TxGirlGeek has quit IRC | 22:06 | |
| sbezverk | if experimental comes back green, I will ack it | 22:06 |
| kfox1111 | its basically add one search path line, and add it to $env for each file. | 22:06 |
| kfox1111 | so, really one line per file to review. | 22:06 |
| *** TxGirlGeek has joined #openstack-kolla | 22:07 | |
| sbezverk | centos with lvm was stable everytime I tired | 22:07 |
| *** MasterOfBugs has joined #openstack-kolla | 22:07 | |
| kfox1111 | sounds good. | 22:07 |
| sbezverk | kfox1111: any chance you could add some non default cases, some no default ports just to see that it is working? | 22:08 |
| sbezverk | I mean not permanently | 22:08 |
| sbezverk | but for a sake of testing | 22:08 |
| *** lbeliveau has joined #openstack-kolla | 22:08 | |
| *** TxGirlGeek has quit IRC | 22:09 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:09 | |
| sbezverk | kfox1111: scratch that, it seems you do not call new macro | 22:09 |
| *** TxGirlGeek has quit IRC | 22:11 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:11 | |
| sbezverk | kfox1111: ansible is not happy, was there any ansible commits in kolla-ansible today which could have broken things? | 22:12 |
| portdirect | kfox1111: just before i look a it why: 'almost all templates?' | 22:12 |
| portdirect | (and not 'all templates') | 22:13 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fix typo in variable name https://review.openstack.org/416673 | 22:15 |
| srwilkers | sbezverk, looks like memcached is bombing out. raising an issue with docker not being found when checking the memcached container | 22:15 |
| *** TxGirlGeek has quit IRC | 22:16 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:16 | |
| sbezverk | srwilkers: it is failing at ansible playbook | 22:16 |
| sbezverk | before anything we touched | 22:17 |
| srwilkers | right. the ansible task thats failing is the task to check the memcached container | 22:17 |
| sbezverk | srwilkers: it is failing at kolla config stage | 22:18 |
| sbezverk | when kolla generates config for us for later usage.. | 22:18 |
| sbezverk | at this point even kube cluster is not up | 22:18 |
| sbezverk | check this out | 22:21 |
| sbezverk | commit 3513ab825d061a8b30dfe0e5fac480b059c5c623 | 22:21 |
| sbezverk | Merge: 7798da7 6e6327e | 22:21 |
| sbezverk | Author: Jenkins <jenkins@review.openstack.org> | 22:21 |
| sbezverk | Date: Wed Jan 4 21:39:02 2017 +0000 | 22:21 |
| sbezverk | Merge "Optimize reconfigure action for memcached" | 22:21 |
| portdirect | hmm so this may be related? https://review.openstack.org/#/c/412225/? | 22:21 |
| portdirect | lol | 22:21 |
| portdirect | sbezverk: thinking about it why do we work from kolla-ansible master? | 22:22 |
| *** TxGirlGeek has quit IRC | 22:23 | |
| srwilkers | thats exactly what im looking at portdirect | 22:23 |
| sbezverk | well guys, I think we should expose it and let kolla-ansible fix it.. | 22:24 |
| srwilkers | im aware this is happening at the config stage. this specifically is what i was pointing out: http://logs.openstack.org/29/416729/3/experimental/gate-kolla-kubernetes-deploy-ubuntu-binary-iscsi-nv/a105184/console.html#_2017-01-04_22_10_10_523914 | 22:24 |
| inc0 | what's up? | 22:24 |
| portdirect | and sbezverk - though i think we took different routes to get to the same place (and he beat me to the punch :) | 22:24 |
| srwilkers | which is why im curious as to if its associated with the review portdirect posted | 22:24 |
| sbezverk | inc0: out gate is broken | 22:25 |
| openstackgerrit | Merged openstack/kolla: Fix docs to require docker config on all nodes https://review.openstack.org/380851 | 22:25 |
| sbezverk | due to Optimize reconfigure action for memcached | 22:25 |
| sbezverk | commit | 22:25 |
| *** jrobinson has joined #openstack-kolla | 22:25 | |
| inc0 | uhh | 22:25 |
| inc0 | can you show me example? | 22:25 |
| sbezverk | inc0: yes one sec | 22:26 |
| sbezverk | inc0: http://logs.openstack.org/44/416744/1/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/7aa3016/console.html | 22:26 |
| sbezverk | it is failing on ansible config tast for memcached | 22:26 |
| *** jheroux has quit IRC | 22:27 | |
| sbezverk | inc0: somebody again decided to optimize things by remving kolla-kube needed code ;-) | 22:27 |
| inc0 | hmm, I see | 22:27 |
| inc0 | sbezverk, well, we need to loose dependency | 22:27 |
| inc0 | between these two | 22:27 |
| srwilkers | i agree entirely | 22:28 |
| *** dave-mccowan has quit IRC | 22:28 | |
| sbezverk | inc0: sonner we get config gen for kube to be independent the better | 22:29 |
| inc0 | well my dear kolla-k8s cores | 22:29 |
| inc0 | there is actually a community that could do it | 22:29 |
| inc0 | << staring at you all >> | 22:29 |
| inc0 | but lets fix this one | 22:30 |
| *** TxGirlGeek has joined #openstack-kolla | 22:30 | |
| srwilkers | i think this is a fantastic candidate for a thorough discussion at the PTG | 22:32 |
| SamYaple | it was a mistake to reuse kolla-ansible configs and inject k8s code into ansible configs in the first place | 22:33 |
| SamYaple | but moving forward, i would love to talk about options. i happen to like config management | 22:34 |
| SamYaple | and templating | 22:34 |
| kfox1111 | back. | 22:34 |
| sbezverk | SamYaple: we did not have a choice at that time | 22:34 |
| SamYaple | sbezverk: why not? | 22:34 |
| inc0 | we had to have configs generated somewhere | 22:34 |
| sbezverk | SamYaple: we needed to produce running code and second there was an idea to use ansible for deployment | 22:35 |
| kfox1111 | portdirect: there's a followon patch that completes the work: https://review.openstack.org/#/c/416744/ | 22:35 |
| SamYaple | ahh. ok i was not aware of the ansible-for-deploemnt piece | 22:35 |
| sbezverk | it was kind of logical | 22:35 |
| SamYaple | that makes this situation infinitly more reasonable to me | 22:35 |
| sbezverk | but since ansible is off the table | 22:35 |
| srwilkers | re config: http://giphy.com/gifs/technology-IF7M8IEKbHby | 22:35 |
| kfox1111 | portdirect: its much harder to review though as it had to fix some missing element_name stuff too. | 22:35 |
| SamYaple | yea agree sbezverk | 22:35 |
| kfox1111 | so I broke it out. | 22:35 |
| sbezverk | kfox1111: nope | 22:36 |
| portdirect | srwilkers: +1 | 22:36 |
| sbezverk | srwilkers: lol | 22:36 |
| kfox1111 | sbezverk: nope what? | 22:36 |
| portdirect | thats the kinds think I imagine SamYaple does in his spare time | 22:36 |
| portdirect | *kind of thing - man my typing is getting worse and worse... | 22:37 |
| sbezverk | kfox1111: you did not brreak anything, it was a change in kolla-ansible that broke our gate | 22:37 |
| kfox1111 | sbezverk: ah. ok. | 22:37 |
| SamYaple | portdirect: everything ive ever done for kolla is in my spare time, fyi | 22:37 |
| kfox1111 | yeah. I defined the value but don't do anything with it yet. just trying to get the major changes that touch all the files done and out of the way. | 22:37 |
| kfox1111 | then we can go through the files a few at a time and update things as needed. | 22:37 |
| sbezverk | kfox1111: make sense.. after realizing that, your ps did not look too too scary ;-) | 22:38 |
| kfox1111 | cool. :) | 22:39 |
| kfox1111 | so, kolla-ansible broke.... | 22:39 |
| kfox1111 | is that being fixed? | 22:39 |
| inc0 | dont be afraid my dear friends! help is on the way | 22:40 |
| *** dave-mccowan has joined #openstack-kolla | 22:40 | |
| kfox1111 | inc0: cool. thx. :) | 22:40 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible: Limit check containers to reconfigure action https://review.openstack.org/416754 | 22:42 |
| inc0 | there you go | 22:42 |
| kfox1111 | thx. :) | 22:42 |
| srwilkers | inc0: http://www.quickmeme.com/meme/3pcju6 | 22:44 |
| sbezverk | man where do we find kolla cores to push it through ?!?!? | 22:45 |
| inc0 | beats me | 22:45 |
| inc0 | have you tried on #openstack-kolla? | 22:45 |
| SamYaple | /join #kolla | 22:45 |
| kfox1111 | :) | 22:47 |
| kfox1111 | 2017-01-04 22:46:36.976726 | An exception occurred during task execution. To see the full traceback, use -vvv. The error was: ImportError: No module named docker | 22:48 |
| kfox1111 | uh, weird. | 22:48 |
| kfox1111 | is it still broken? | 22:49 |
| SamYaple | kfox1111: you know docker-py isnt the name anymore right? | 22:49 |
| SamYaple | pip install docker | 22:49 |
| sbezverk | kfox1111: yes we need to get inc0 fix in | 22:49 |
| kfox1111 | doh. I took that as merged... didn't notice it was a ps. :/ | 22:49 |
| inc0 | well, I could merge my own change | 22:50 |
| inc0 | but I won't | 22:50 |
| inc0 | I don't want to end up in hell for that | 22:50 |
| kfox1111 | +1 for not sending inc0 to hell. :) | 22:50 |
| kfox1111 | more kolla cores though please? | 22:50 |
| kfox1111 | trunks broken. :/ | 22:50 |
| sbezverk | inc0 if we get 10 +1m would it be == 2 +2 ;-) ?? | 22:51 |
| inc0 | yeah, that works, but no matter how many +0 workflow you do, it's never going to be +1 | 22:51 |
| *** openstack has joined #openstack-kolla | 22:56 | |
| kfox1111 | more reviews please https://review.openstack.org/#/c/416754/ | 22:57 |
| kfox1111 | this is really holding me up. :/ | 22:57 |
| *** rhallisey has quit IRC | 22:58 | |
| portdirect | kfox1111: https://media.giphy.com/media/D3OdaKTGlpTBC/giphy.gif | 22:58 |
| sbezverk | kfox1111: I pinged sdake internally hopefully he replies | 22:58 |
| portdirect | kfox1111/sbezverk why not just base our gate on b2? | 22:59 |
| sbezverk | portdirect: lol | 22:59 |
| portdirect | as for config it should be no different? | 22:59 |
| *** TxGirlGeek has quit IRC | 22:59 | |
| *** shardy has quit IRC | 22:59 | |
| kfox1111 | portdirect: b2? | 23:00 |
| sbezverk | portdirect: I am not sure if with cloning we do it will work right away without other problems | 23:00 |
| portdirect | kfox1111: https://github.com/openstack/kolla-ansible/tree/4.0.0.0b2 | 23:00 |
| kfox1111 | portdirect: oh. | 23:00 |
| portdirect | sbezverk: is it worth trying? | 23:00 |
| *** dave-mccowan has quit IRC | 23:00 | |
| kfox1111 | if it isn't resolved fast, I'd say thats a good option. | 23:01 |
| kfox1111 | inc0: there's another option. | 23:01 |
| kfox1111 | reverting a bad patch shouldn't requre a full review? | 23:01 |
| kfox1111 | can we just revert the offending patch until its fixed / rereviewed? | 23:02 |
| SamYaple | kfox1111: reverting still requires 2 cores | 23:02 |
| sbezverk | portdirect: for sure as a plan B it is great | 23:02 |
| inc0 | kfox1111, revert is same thing - 2* +2 | 23:03 |
| kfox1111 | seems bad that cores can push a bad patch in, then can't get it reverted if enough disapear for a while. :/ | 23:03 |
| inc0 | only these are 2 patches, so even harder | 23:03 |
| inc0 | well, that's how bugs happen | 23:03 |
| SamYaple | kfox1111: better than one core not understanding something and then reverting it because _they_ are mistaken | 23:03 |
| sdake | sbezverk i got the msg | 23:03 |
| inc0 | and since we don't have CI in kolla-ansible to check kolla-k8s | 23:03 |
| SamYaple | rule of 2 | 23:04 |
| sdake | but it has not passed the gate | 23:04 |
| inc0 | bad stuff happends | 23:04 |
| * kfox1111 shrugs | 23:04 | |
| kfox1111 | I guess. | 23:04 |
| sdake | inc0 we kolla is cross-repo gating kolla | 23:04 |
| sbezverk | sdake: huray | 23:04 |
| sdake | rather kolla-kubernetes | 23:04 |
| sdake | rather kolla-ansible | 23:04 |
| sdake | too many things :) | 23:04 |
| inc0 | sdake, no, kolla-ansible gates won't check kolla-k8s | 23:04 |
| sbezverk | sdake: I was trying to get vhosakot internally but he is not available :-( | 23:04 |
| inc0 | sdake, https://review.openstack.org/#/c/416754/ | 23:05 |
| sdake | inc0 yes i know | 23:05 |
| inc0 | ack this plz | 23:05 |
| portdirect | sbezverk: there is no plan B... just plan sdake. | 23:05 |
| * srwilkers blinks | 23:05 | |
| sdake | inc0 moment | 23:06 |
| inc0 | it's 7am in China so our Asian friends should be up soon | 23:06 |
| inc0 | ironically, kolla doesn't have that many cores in US any more | 23:06 |
| sdake | inc0 sadly | 23:07 |
| sdake | inc0 alot of apac and emea | 23:07 |
| inc0 | I'm actually pretty happy because of that | 23:07 |
| sdake | we are a world wide project | 23:07 |
| inc0 | exactly | 23:07 |
| inc0 | we might be only non-us centric project | 23:07 |
| sdake | ok - gotta jet | 23:07 |
| inc0 | which kinda makes our PTG even less sensible in Atlanta;) | 23:07 |
| SamYaple | its not our PTG, its all PTG | 23:08 |
| inc0 | yeah | 23:08 |
| SamYaple | as you said, most projects are US central | 23:08 |
| inc0 | what I'm saying is, if I'd organize midcycle it wouldn't be in US any more | 23:08 |
| inc0 | I'd say...Poland | 23:09 |
| *** sdake has quit IRC | 23:09 | |
| SamYaple | im sure you would! | 23:09 |
| kfox1111 | :) | 23:09 |
| inc0 | hey, it makes sense | 23:09 |
| inc0 | it's in Europe, but still close to Asia | 23:09 |
| sbezverk | inc0: man it is too cold in feb to go to Poland | 23:09 |
| inc0 | relative to rest of europe;) | 23:09 |
| sbezverk | think Brazil | 23:09 |
| sbezverk | :-) | 23:09 |
| SamYaple | no no no, think about a place you _want_ to go | 23:09 |
| SamYaple | Norway | 23:09 |
| inc0 | SamYaple, you hate sun? | 23:10 |
| inc0 | like...you are willing to not see it at all for a week? | 23:10 |
| portdirect | ^^ truth | 23:10 |
| kfox1111 | bah. evil day star. good riddence. :) | 23:10 |
| SamYaple | inc0: that sounds lovely | 23:11 |
| inc0 | yeah I knew you might like it | 23:11 |
| inc0 | anyway, Norway is super expensive | 23:11 |
| SamYaple | dont forget, i worked thirdshift for most of the time youve known me | 23:11 |
| inc0 | Poland on the other hand... | 23:12 |
| SamYaple | where is amsterdam? | 23:12 |
| SamYaple | lets go there | 23:12 |
| inc0 | yeah, actually | 23:13 |
| inc0 | Amsterdam | 23:13 |
| inc0 | I mean weed is legal where I am now | 23:13 |
| inc0 | but Amsterdam... | 23:13 |
| inc0 | or Berlin | 23:14 |
| inc0 | I miss Europe..:( | 23:14 |
| SamYaple | its not legal inc0 | 23:14 |
| sbezverk | inc0: good thing about europe is everything is close, but at the same time it is bad thing.. | 23:14 |
| SamYaple | federally it is still illegal | 23:15 |
| inc0 | SamYaple, but in Oregon it is legal | 23:15 |
| inc0 | not sure how it works, but I can buy it in store | 23:15 |
| SamYaple | no its not, federal superceed state | 23:15 |
| inc0 | so meh | 23:15 |
| SamYaple | its just not enforced | 23:15 |
| SamYaple | but its not legal either | 23:15 |
| SamYaple | the dea can (and has) come in and arrest people | 23:15 |
| inc0 | funny country we live in right? | 23:15 |
| SamYaple | inorite | 23:16 |
| srwilkers | funny is certainly one way of saying it | 23:20 |
| kfox1111 | inc0: its a werid thing. federal technically trumps state. | 23:20 |
| *** salv-orl_ has quit IRC | 23:20 | |
| kfox1111 | so it is elegal. but federal's not enforcing in legalized states. | 23:20 |
| srwilkers | which could change soon | 23:21 |
| kfox1111 | so, defacto legal. | 23:21 |
| kfox1111 | but could change at any time for any reason they want to get you on. | 23:21 |
| inc0 | yeah...it's like....how you 'muricans call it? FREEEEDOM! | 23:21 |
| inc0 | Give me a person, we'll find a paragraph. | 23:22 |
| * srwilkers groans | 23:22 | |
| inc0 | I think it was Stalin who said that? | 23:22 |
| *** sdake has joined #openstack-kolla | 23:24 | |
| SamYaple | kfox1111: nah like i said, DEA still drops in and enforces | 23:24 |
| kfox1111 | I'm a federal employe technically. so Its still elegal in my book. | 23:24 |
| kfox1111 | even though its legal in washington state. | 23:25 |
| SamYaple | i dont think things are ever legal, they are just not illegal | 23:25 |
| SamYaple | its illegal, but unenfored? | 23:25 |
| SamYaple | idk. anyway. NO ONE IS SAFE | 23:25 |
| srwilkers | haha | 23:25 |
| *** sdake has quit IRC | 23:26 | |
| inc0 | funny how THC is schedule I but vicodin is schedule II | 23:26 |
| inc0 | or LSD - schedule I | 23:26 |
| srwilkers | its a lot like lane splitting in the US. up until recently, it was only legal in some states because it wasnt explicitly illegal | 23:26 |
| SamYaple | things with no medcinal use are schedule I | 23:26 |
| SamYaple | the medicinal thing is whats been challenged | 23:26 |
| portdirect | inc0 berlin? careful now | 23:27 |
| inc0 | what about it? cool city | 23:27 |
| inc0 | not that I like Germany | 23:27 |
| srwilkers | ;) | 23:27 |
| inc0 | sorry berendt | 23:27 |
| inc0 | ;) | 23:27 |
| inc0 | anyway, Poland, we should have midcycle in Poland | 23:28 |
| kfox1111 | https://review.openstack.org/#/c/416754/ more please | 23:28 |
| inc0 | Jeffrey4l, ^ when you're awake please | 23:28 |
| SamYaple | inc0: no such thing as midcycles anymore! | 23:29 |
| inc0 | now you are mistaken | 23:29 |
| inc0 | ops still have midcycles | 23:29 |
| *** sdake has joined #openstack-kolla | 23:30 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: [WIP] Add OpenDaylight to kolla-ansible https://review.openstack.org/416367 | 23:30 |
| SamYaple | i just assume im always mistaken | 23:30 |
| inc0 | no, that would make it easy | 23:30 |
| inc0 | you're mistaken most of the time | 23:30 |
| SamYaple | i can still assume im always mistaken | 23:31 |
| inc0 | then you'd be mistaken | 23:31 |
| inc0 | to do so | 23:31 |
| SamYaple | be doign so is a subset of your statment! | 23:31 |
| inc0 | on the other hand, if you'd be always mistakend and assume that, you wouldn't be always mistaken any more | 23:31 |
| inc0 | so assumption would be a mistake | 23:31 |
| SamYaple | though i assume i could be wrong here | 23:31 |
| *** vhosakot_ has quit IRC | 23:35 | |
| * portdirect wonders about life choices | 23:36 | |
| inc0 | SamYaple, I've sent mail to ML regarding kolla-salt and other deliverables | 23:39 |
| inc0 | and how we deal with them | 23:39 |
| inc0 | feel free to chime in | 23:39 |
| *** cliles has quit IRC | 23:41 | |
| *** cliles has joined #openstack-kolla | 23:43 | |
| *** lamt has quit IRC | 23:48 | |
| inc0 | kfox1111, pm plz | 23:53 |
| *** lbeliveau has quit IRC | 23:54 | |
| *** goldyfruit has quit IRC | 23:54 | |
| SamYaple | inc0: chimed in | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
