| *** kklimonda has quit IRC | 00:00 | |
| *** kklimonda has joined #openstack-kolla | 00:01 | |
| *** lmiccini has joined #openstack-kolla | 00:07 | |
| kfox1111 | wirehead_: yeah. I only knew because I talked with some of the devs. :/ | 00:20 |
|---|---|---|
| *** mbound has quit IRC | 00:20 | |
| kfox1111 | wirehead_: since I think we'll be able to switch to multinode for at least the api services in pretty short order, should we start a spec or something for how we will make termination safe? | 00:21 |
| kfox1111 | wrap it in wsgi/nginx or something? | 00:21 |
| kfox1111 | or maybe this is a general mailing list question. | 00:24 |
| kfox1111 | it would be good to know if any other operators have wrapped the services or know of any other way of solving it. | 00:24 |
| kfox1111 | wirehead_: also, one other thought I had on a smilar line of thinking to the glance api/registry split. | 00:25 |
| kfox1111 | neutron server (and maybe nova-api, I think) have web processing threads and rpc processing threads. | 00:25 |
| kfox1111 | I think the api side of things will scale nicely without net=host. | 00:26 |
| kfox1111 | but the rpc things register in nova service-list and neutron agent-list respectively and should probably not change names frequently. | 00:26 |
| kfox1111 | so we might want to run the same container, twice, with slightly different configs. (api_workers = 0, rpc_wokers=4 & api_wokers = 4, rpc_workers = 0) or something like that. | 00:27 |
| *** achanda_ has quit IRC | 00:27 | |
| kfox1111 | in fact, we may want to do the api ones with deployments and the rpc ones with daemonsets. | 00:28 |
| *** zhiwei has joined #openstack-kolla | 00:38 | |
| *** zhiwei has quit IRC | 00:43 | |
| *** Jeffrey4l has joined #openstack-kolla | 00:44 | |
| *** daneyon_ has joined #openstack-kolla | 00:46 | |
| *** ssurana1 has quit IRC | 00:50 | |
| *** daneyon_ has quit IRC | 00:50 | |
| *** phuongnh has joined #openstack-kolla | 00:57 | |
| *** fragatina has joined #openstack-kolla | 00:59 | |
| *** zhiwei has joined #openstack-kolla | 01:04 | |
| openstackgerrit | Jack Ning proposed openstack/kolla: Replace horizon default config with custom config https://review.openstack.org/306928 | 01:07 |
| *** fragatina has quit IRC | 01:07 | |
| *** fragatina has joined #openstack-kolla | 01:08 | |
| *** salv-orlando has joined #openstack-kolla | 01:18 | |
| *** salv-orlando has quit IRC | 01:25 | |
| *** bootsha has joined #openstack-kolla | 01:38 | |
| *** daneyon has quit IRC | 01:42 | |
| *** bootsha has quit IRC | 01:43 | |
| *** williamcaban has joined #openstack-kolla | 01:46 | |
| *** haplo37 has quit IRC | 01:55 | |
| *** haplo37 has joined #openstack-kolla | 01:56 | |
| *** signed8bit_Zzz is now known as signed8bit | 01:57 | |
| *** sacharya has joined #openstack-kolla | 01:59 | |
| *** sacharya has quit IRC | 02:03 | |
| *** sacharya has joined #openstack-kolla | 02:04 | |
| *** asalkeld has quit IRC | 02:09 | |
| *** g3ek has quit IRC | 02:13 | |
| *** mewald has joined #openstack-kolla | 02:18 | |
| *** g3ek has joined #openstack-kolla | 02:19 | |
| *** mewald1 has quit IRC | 02:21 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Consolidating path resolution and searching for program files https://review.openstack.org/329676 | 02:26 |
| *** daneyon has joined #openstack-kolla | 02:34 | |
| *** stvnoyes1 has quit IRC | 02:37 | |
| *** stvnoyes has joined #openstack-kolla | 02:38 | |
| *** daneyon has quit IRC | 02:39 | |
| *** inc0 has joined #openstack-kolla | 02:39 | |
| *** bootsha has joined #openstack-kolla | 02:39 | |
| *** bootsha has quit IRC | 02:44 | |
| *** inc0 has quit IRC | 02:46 | |
| *** jtriley has joined #openstack-kolla | 02:46 | |
| *** salv-orlando has joined #openstack-kolla | 02:49 | |
| *** JianqingJiang has joined #openstack-kolla | 02:53 | |
| *** salv-orlando has quit IRC | 02:54 | |
| JianqingJiang | I've install a local docker registry,but How can I move my local images show in command "docker images" to my local docker regsitry? | 02:55 |
| britthouser | You'll have to docker tag them and give them a name that starts off with the local registry IP:PORT | 02:58 |
| britthouser | and then docker push them into the local registry | 02:58 |
| *** g3ek has quit IRC | 02:59 | |
| *** haplo37 has quit IRC | 02:59 | |
| *** achanda has joined #openstack-kolla | 03:03 | |
| JianqingJiang | Thanks | 03:06 |
| *** g3ek has joined #openstack-kolla | 03:08 | |
| *** achanda has quit IRC | 03:08 | |
| *** haplo37 has joined #openstack-kolla | 03:08 | |
| *** daneyon has joined #openstack-kolla | 03:18 | |
| *** daneyon_ has joined #openstack-kolla | 03:21 | |
| *** daneyon has quit IRC | 03:24 | |
| *** williamcaban has quit IRC | 03:27 | |
| *** williamcaban has joined #openstack-kolla | 03:34 | |
| *** bootsha has joined #openstack-kolla | 03:40 | |
| *** signed8bit has quit IRC | 03:43 | |
| *** bootsha has quit IRC | 03:44 | |
| *** williamcaban has quit IRC | 03:47 | |
| *** williamcaban has joined #openstack-kolla | 03:50 | |
| *** salv-orlando has joined #openstack-kolla | 04:00 | |
| *** huikang has joined #openstack-kolla | 04:03 | |
| *** williamcaban has quit IRC | 04:06 | |
| *** JianqingJiang has quit IRC | 04:06 | |
| *** salv-orlando has quit IRC | 04:14 | |
| *** achanda has joined #openstack-kolla | 04:18 | |
| *** daneyon has joined #openstack-kolla | 04:22 | |
| *** daneyon has quit IRC | 04:27 | |
| *** sacharya has quit IRC | 04:30 | |
| *** huikang has quit IRC | 04:36 | |
| *** huikang has joined #openstack-kolla | 04:36 | |
| *** dave-mccowan has quit IRC | 04:41 | |
| *** huikang has quit IRC | 04:41 | |
| *** Liuqing has joined #openstack-kolla | 04:46 | |
| *** coolsvap has joined #openstack-kolla | 04:49 | |
| *** salv-orlando has joined #openstack-kolla | 04:50 | |
| *** salv-orl_ has joined #openstack-kolla | 04:59 | |
| *** salv-orlando has quit IRC | 05:02 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/kolla: Update kollaglue repo on docker to kolla https://review.openstack.org/333110 | 05:03 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/kolla: Update kollaglue repo on docker to kolla https://review.openstack.org/333111 | 05:07 |
| *** daneyon has joined #openstack-kolla | 05:13 | |
| *** daneyon__ has joined #openstack-kolla | 05:14 | |
| *** tfukushima has joined #openstack-kolla | 05:16 | |
| *** JianqingJiang has joined #openstack-kolla | 05:16 | |
| *** daneyon_ has quit IRC | 05:17 | |
| *** daneyon has quit IRC | 05:18 | |
| *** JianqingJiang has quit IRC | 05:21 | |
| *** JianqingJiang has joined #openstack-kolla | 05:21 | |
| *** Liuqing has quit IRC | 05:33 | |
| *** shardy has joined #openstack-kolla | 05:36 | |
| *** shardy has quit IRC | 05:38 | |
| *** daneyon__ has quit IRC | 05:58 | |
| *** JeroenBo has joined #openstack-kolla | 06:24 | |
| *** zhiwei has quit IRC | 06:28 | |
| *** zhiwei has joined #openstack-kolla | 06:28 | |
| *** zhiwei has quit IRC | 06:29 | |
| *** zhiwei has joined #openstack-kolla | 06:29 | |
| *** sdake has joined #openstack-kolla | 06:34 | |
| *** gfidente|afk has quit IRC | 06:35 | |
| mewald | I noticed that in addition to serveral software requirements (ansible, python, docker, etc …) I need proper name resolution between some nodes. For example, for rabbitmq services to join into a cluster. Is there any solution or plan to solve this name resolution thing in general? My suggested solution would be to set up Consul (http://www.consul.io) on all nodes. It would be easy to package that into a container and deploy it everywh | 06:35 |
| sdake | evening | 06:41 |
| sdake | someone ninja added me to a panel | 06:41 |
| sdake | mewald yes that is right you do need name resolution | 06:42 |
| sdake | mewald hui kang has written consul but i dont see the value in it and neithr do other core revieewers so it hasnt merged | 06:43 |
| *** jmccarthy has joined #openstack-kolla | 06:44 | |
| mewald | sdake: so how does the suggested way look like to make Kolla deployment work? Is it just seen as a prerequisite that must be matched but nothing Kolla takes care of? | 06:46 |
| sdake | typically people make an /etc/hosts file that contains their hostnams | 06:46 |
| sdake | and copy it to all nodes | 06:46 |
| sdake | we are working on bare metal deployment | 06:46 |
| mewald | yeah, shortcomings I see with this is: add node X, touch X-1 nodes to fix /etc/hosts | 06:47 |
| sdake | that bare meal work will include work to do the base os setup like install docker engine and docker-py | 06:47 |
| sdake | add all nodes to 1 file | 06:47 |
| sdake | then you can use the ansible all module to copy to all nodes | 06:47 |
| *** zhiwei has quit IRC | 06:47 | |
| sdake | (I think:) | 06:47 |
| mewald | have to check, thx | 06:47 |
| sdake | this is what we will do with bare metal setup (automated) | 06:48 |
| sdake | but we aren't there yet | 06:48 |
| sdake | there is one component which does require dns resolution - specifically rabbitmq | 06:48 |
| sdake | nothing else should i think | 06:48 |
| *** belmoreira has joined #openstack-kolla | 06:49 | |
| mewald | yeah think you're right. the rabbitmq guys should fix this asap … really annoying | 06:50 |
| *** darkmike has joined #openstack-kolla | 06:54 | |
| mewald | any monitoring plans other than elastic search? I would love to see sensu in the stack | 06:54 |
| *** jmccarthy has left #openstack-kolla | 06:54 | |
| *** achanda has quit IRC | 06:56 | |
| *** fragatina has quit IRC | 06:57 | |
| *** fragatina has joined #openstack-kolla | 06:58 | |
| *** Mr_Broken has joined #openstack-kolla | 06:59 | |
| *** achanda has joined #openstack-kolla | 07:01 | |
| *** shardy has joined #openstack-kolla | 07:02 | |
| *** bootsha has joined #openstack-kolla | 07:02 | |
| *** b_bezak has joined #openstack-kolla | 07:02 | |
| *** daneyon has joined #openstack-kolla | 07:05 | |
| *** achanda has quit IRC | 07:06 | |
| darkmike | Hi, I want some help please. I want to know wish version of Kolla is mostly suggested to quickstart openstack. For docker image, do you know where is the best namespace like kollaglue or coolsvap and binary or source ? I just want recommendation for 1rst time success | 07:09 |
| *** daneyon has quit IRC | 07:09 | |
| darkmike | And, I need hints to choose vs version 2 or 3 of kolla | 07:10 |
| coolsvap | darkmike, you can go with mitaka | 07:10 |
| darkmike | thank out all | 07:10 |
| coolsvap | version 2.0.1 | 07:10 |
| darkmike | ok great | 07:10 |
| coolsvap | kollaglue is not updated i suppose | 07:10 |
| darkmike | true | 07:10 |
| coolsvap | coolsvap is something i am pushing to docker hub | 07:10 |
| darkmike | then, you recommend me yours? binary or source ? | 07:10 |
| coolsvap | the new dockerhub is kolla | 07:11 |
| darkmike | ok | 07:11 |
| coolsvap | but its not updated as well | 07:11 |
| coolsvap | we will have it updated in couple of days | 07:11 |
| darkmike | ok | 07:11 |
| darkmike | cool | 07:11 |
| coolsvap | i think you can go with either binary or source | 07:11 |
| sdake | darkmike if you want to get started right now you can use mitaka 2.0.0 | 07:11 |
| sdake | binary or source or ubuntu or centos work properly | 07:11 |
| coolsvap | you can build images locally as well | 07:11 |
| coolsvap | if you wish | 07:11 |
| coolsvap | hey sdake :) | 07:11 |
| sdake | darkmike to do so, add a line "openstack_release: 2.0.0" to your globals.yml file | 07:12 |
| sdake | it should be commented out | 07:12 |
| sdake | definately dont use master unless you plan to develop for kolla | 07:12 |
| sdake | 2.0.0 is availablei nt he kollaglue namespace | 07:12 |
| darkmike | Ok, I just test version 3 using coolsvap-source and it launch with few bugs | 07:12 |
| sdake | we arei n the process of movingthings over to kolla namespace | 07:12 |
| darkmike | cool | 07:13 |
| sdake | darkmike woud you prefer source or binary to start - i'll push 2.0.1 to kollaglue | 07:14 |
| sdake | it needs to be there anyway | 07:14 |
| darkmike | I don't know seriously wish is the best / safetest | 07:14 |
| darkmike | I just want to give a quick course on Openstack | 07:15 |
| darkmike | the next week | 07:15 |
| darkmike | Maybe, I want the version 2.0.1 if version 2.0.0 has some edge. Thanks sdake. | 07:16 |
| darkmike | And many thanks coolsvap | 07:17 |
| *** Serlex has joined #openstack-kolla | 07:17 | |
| sdake | what base os are you planning to deploy? | 07:17 |
| darkmike | I don't have preference. I'm ok with Centos and Ubuntu. | 07:17 |
| *** salv-orl_ has quit IRC | 07:18 | |
| darkmike | oh base os | 07:18 |
| sdake | ok lets go with centos since that is what i have available to build with :) | 07:18 |
| darkmike | ok, allright | 07:18 |
| sdake | ya container and baseos should match | 07:18 |
| darkmike | I use centos actually | 07:18 |
| sdake | sounds good i'll push in a few minutes | 07:18 |
| darkmike | Coool | 07:18 |
| darkmike | Maybe I can't test it tonight since it's 3am | 07:19 |
| sdake | ya it will take me atelast 1 hour to push | 07:19 |
| darkmike | I ll go deeper next days | 07:19 |
| darkmike | thank you again. I will do bug report or fix problems if needed | 07:20 |
| *** sdake_ has joined #openstack-kolla | 07:21 | |
| *** sdake has quit IRC | 07:24 | |
| *** matrohon has joined #openstack-kolla | 07:48 | |
| *** mbound has joined #openstack-kolla | 07:53 | |
| *** neilus has joined #openstack-kolla | 07:58 | |
| *** Mr_Broken has quit IRC | 07:58 | |
| *** gfidente has joined #openstack-kolla | 08:04 | |
| *** salv-orlando has joined #openstack-kolla | 08:05 | |
| *** mbound has quit IRC | 08:19 | |
| *** b_bezak has quit IRC | 08:28 | |
| *** bootsha has quit IRC | 08:31 | |
| *** b_bezak has joined #openstack-kolla | 08:33 | |
| *** zhiwei has joined #openstack-kolla | 08:44 | |
| *** zhiwei has quit IRC | 08:44 | |
| *** zhiwei has joined #openstack-kolla | 08:45 | |
| *** Jeffrey4l has quit IRC | 08:47 | |
| *** salv-orlando has quit IRC | 08:49 | |
| *** salv-orlando has joined #openstack-kolla | 08:50 | |
| *** Mr_Broken has joined #openstack-kolla | 08:53 | |
| *** daneyon has joined #openstack-kolla | 08:53 | |
| *** salv-orlando has quit IRC | 08:55 | |
| *** daneyon has quit IRC | 08:58 | |
| *** Mr_Broken has quit IRC | 08:59 | |
| *** Jeffrey4l has joined #openstack-kolla | 09:00 | |
| *** achanda has joined #openstack-kolla | 09:03 | |
| *** jtriley has quit IRC | 09:06 | |
| *** b_bezak has quit IRC | 09:06 | |
| *** achanda has quit IRC | 09:08 | |
| *** dmk0202 has joined #openstack-kolla | 09:09 | |
| *** zhiwei has quit IRC | 09:11 | |
| *** zhiwei has joined #openstack-kolla | 09:15 | |
| sdake_ | coolsvap ping | 09:22 |
| coolsvap | sdake_, hey | 09:22 |
| sdake_ | can you delete the repos in kolla please | 09:23 |
| coolsvap | sdake_, yes will do | 09:23 |
| pbourke | sdake_: hey, wondering do you have a handy filter you use for kolla bugs on launchpad. going to do some triaging | 09:23 |
| sdake_ | pbourke i do not | 09:23 |
| pbourke | how did you pull that number you mentioned a while back | 09:23 |
| sdake_ | pbourke i eyeball em | 09:23 |
| pbourke | ok | 09:23 |
| sdake_ | https://bugs.launchpad.net/kolla | 09:24 |
| sdake_ | 45 bugs | 09:24 |
| pbourke | if i filter for only un-triaged, excluding incomplete, in progress, etc. I get 42 | 09:24 |
| pbourke | which is not that many | 09:24 |
| sdake_ | in new state | 09:24 |
| sdake_ | https://bugs.launchpad.net/kolla/+bugs?search=Search&field.status=New | 09:25 |
| *** diga has joined #openstack-kolla | 09:25 | |
| pbourke | thanks | 09:25 |
| sdake_ | it was at 60 previously :) | 09:26 |
| sdake_ | we have a few hundred bugs in b1 that I have yet to move to b2 | 09:26 |
| sdake_ | since moving bugs sucks | 09:26 |
| sdake_ | so hot - 119 118 118 days | 09:27 |
| sdake_ | hardly can deal with it | 09:27 |
| coolsvap | sdake_, isnt there a mutli-select kinda option | 09:28 |
| coolsvap | for moving bugs | 09:28 |
| *** Jeffrey4l has quit IRC | 09:28 | |
| *** athomas has joined #openstack-kolla | 09:30 | |
| *** Jeffrey4l has joined #openstack-kolla | 09:42 | |
| *** salv-orlando has joined #openstack-kolla | 09:42 | |
| *** sdake_ has quit IRC | 09:42 | |
| *** pbourke has quit IRC | 09:43 | |
| *** pbourke has joined #openstack-kolla | 09:43 | |
| *** JianqingJiang has quit IRC | 09:46 | |
| *** b_bezak has joined #openstack-kolla | 09:49 | |
| *** bootsha has joined #openstack-kolla | 09:49 | |
| *** Mr_Broken has joined #openstack-kolla | 09:50 | |
| *** fragatina has quit IRC | 09:56 | |
| *** shardy has quit IRC | 10:00 | |
| *** sdake has joined #openstack-kolla | 10:01 | |
| sdake | coolsvap not that i know of | 10:02 |
| *** coolsvap has quit IRC | 10:02 | |
| sdake | pbourke you notice how some of thosse bugs have priorities assigned but haven't entered the triagef state | 10:02 |
| sdake | that shouldn't happen | 10:02 |
| pbourke | yeah | 10:02 |
| pbourke | im marking each triaged | 10:02 |
| pbourke | once I triage them :) | 10:03 |
| *** coolsvap has joined #openstack-kolla | 10:03 | |
| sdake | i am super surprised someone ninja added me to a talk without consulting me first | 10:06 |
| pbourke | sdake: do you know off hand what package provides delorean.repo | 10:10 |
| sdake | we manually install that via the build.py script | 10:11 |
| sdake | https://github.com/openstack/kolla/blob/master/kolla/common/config.py#L30 | 10:12 |
| pbourke | perfect thanks | 10:14 |
| Jeffrey4l | pbourke, https://review.openstack.org/#/c/331420/ please review this PS. | 10:18 |
| patchbot | Jeffrey4l: patch 331420 - kolla (stable/mitaka) - Fix the ubuntu binary on mitaka branch | 10:18 |
| pbourke | curl http://trunk.rdoproject.org/centos7/current-passed-ci/delorean.repo | 10:19 |
| pbourke | i dont know how this isn't breaking the gate | 10:19 |
| pbourke | as its returning 302 redirect for me | 10:19 |
| pbourke | Jeffrey4l: why not upgrade pip? | 10:19 |
| Jeffrey4l | pbourke, it is binary install. we never upgrade the pip in the binary. | 10:20 |
| pbourke | we did at one stage | 10:20 |
| pbourke | the cache bloats the image size a lot | 10:21 |
| Jeffrey4l | OK. i will upgrade the pip. | 10:21 |
| *** shardy has joined #openstack-kolla | 10:24 | |
| *** sdake_ has joined #openstack-kolla | 10:28 | |
| *** dmk0202 has quit IRC | 10:29 | |
| *** sdake has quit IRC | 10:32 | |
| *** diga has left #openstack-kolla | 10:32 | |
| *** phuongnh has quit IRC | 10:35 | |
| *** bootsha has quit IRC | 10:37 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Fix docker daemon proxy support in vagrant https://review.openstack.org/333238 | 10:41 |
| *** bootsha has joined #openstack-kolla | 10:41 | |
| *** daneyon has joined #openstack-kolla | 10:42 | |
| *** coolsvap has quit IRC | 10:42 | |
| *** coolsvap has joined #openstack-kolla | 10:44 | |
| *** daneyon has quit IRC | 10:46 | |
| mewald | What's the rule Kolla follows when deciding whether to do something in Dockerfile or via Ansible task? I am looking into code a bit today and was wondering that because all of what happens in Dockerfiles could just as well be done in Ansible or not? | 10:52 |
| *** Mr_Broken has quit IRC | 10:53 | |
| *** Mr_Broken has joined #openstack-kolla | 10:54 | |
| sdake_ | mewald we try to do very little in the dockerfilees themselves | 10:55 |
| *** Mr_Broken has quit IRC | 10:55 | |
| sdake_ | mewald we have an ABI | 10:55 |
| *** Mr_Broken has joined #openstack-kolla | 10:56 | |
| mewald | sdake_: what's ABI? | 10:56 |
| sdake_ | abi = application binary interface | 10:56 |
| sdake_ | https://github.com/openstack/kolla/blob/master/ansible/roles/heat/templates/heat-api.json.j2 | 10:56 |
| sdake_ | we use that heat-api.json file to tell the contianer how to operate | 10:56 |
| sdake_ | heat-api.json gets copied into the ocntainer on startup | 10:57 |
| sdake_ | if you could tell me what you want to add, I can help guide you :) | 10:57 |
| sdake_ | the model of the abi is other applications could provide their own json files and control the container however they chose | 10:58 |
| sdake_ | choose | 10:58 |
| *** salv-orl_ has joined #openstack-kolla | 10:58 | |
| *** coolsvap has quit IRC | 10:58 | |
| mewald | not trying to add anything :) just trying to figure out how it all works. So that file you just referred to is used to build the command that is ran in the container, right? | 10:59 |
| sdake_ | and the list of files to be copied | 10:59 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Follow redirects for repo files in build https://review.openstack.org/333252 | 11:00 |
| mewald | yeah ok, but how does that have to do with the choice of whether to do CMD apt-get install in Dockerfile or run an ansible apt task? | 11:00 |
| sdake_ | mewald the apt commands are for installing the container | 11:00 |
| sdake_ | our general policy is data that should always be there should be installed in the dockerfile | 11:00 |
| sdake_ | data that changes (non-immutable) should be copied in | 11:01 |
| sdake_ | that way we can make it immutable if we want | 11:01 |
| *** coolsvap has joined #openstack-kolla | 11:01 | |
| sdake_ | that json file is parsed by: | 11:01 |
| sdake_ | https://github.com/openstack/kolla/blob/master/docker/base/set_configs.py | 11:01 |
| *** salv-orlando has quit IRC | 11:01 | |
| sdake_ | every dockerfile in the system starts with this script: | 11:01 |
| sdake_ | https://github.com/openstack/kolla/blob/master/docker/base/start.sh | 11:01 |
| sdake_ | we sepraate build from run | 11:01 |
| mewald | puhh for me this is a lot of stuff to glue together in my mind | 11:02 |
| mewald | have to keep digging and reading ^^ | 11:02 |
| sdake_ | so the flow is ansible puts the json file and any config files on the deploy target, runs start.sh, which runs set_configs.py which parses the json | 11:02 |
| sdake_ | finally start.sh runs the proper command as specified in the json | 11:03 |
| mewald | does all that happen inside the container or inside the docker host? | 11:03 |
| sdake_ | half and half | 11:04 |
| sdake_ | the copying of the json to the target happens via ansible | 11:04 |
| sdake_ | the copying of the json nto the container happens via a bindmount inside the container | 11:04 |
| *** achanda has joined #openstack-kolla | 11:05 | |
| mewald | I just notice I have to start somewhere before all that :D Even the docker files are not static but generated :D damn :D | 11:05 |
| sdake_ | ya that is how we handle multi distro | 11:06 |
| *** stvnoyes has quit IRC | 11:09 | |
| *** achanda has quit IRC | 11:09 | |
| coolsvap | pbourke, we need to update the delorean.repo url | 11:10 |
| coolsvap | to http://buildlogs.centos.org/centos/7/cloud/x86_64/rdo-trunk-master/ | 11:10 |
| pbourke | coolsvap: I have a feeling it switches back and forth though | 11:10 |
| pbourke | coolsvap: buildlogs is the staging area before it goes to trunk afaik | 11:11 |
| coolsvap | pbourke, did you read this https://www.redhat.com/archives/rdo-list/2016-June/msg00206.html | 11:13 |
| coolsvap | i will double check again | 11:13 |
| pbourke | coolsvap: ah | 11:14 |
| pbourke | coolsvap: that seems conclusive then | 11:14 |
| pbourke | I can update the url | 11:14 |
| *** reddydodda has joined #openstack-kolla | 11:17 | |
| *** sdake_ has quit IRC | 11:21 | |
| *** williamcaban has joined #openstack-kolla | 11:26 | |
| *** dmk0202 has joined #openstack-kolla | 11:33 | |
| pbourke | coolsvap: the second deloran-deps.repo is not at the buildlogs url... | 11:34 |
| coolsvap | let me check | 11:36 |
| pbourke | i hate this shit | 11:37 |
| coolsvap | pbourke, i dont see it in earlier url as well :/ | 11:37 |
| pbourke | will ask in #rdo I suppose | 11:37 |
| *** v1k0d3n has joined #openstack-kolla | 11:41 | |
| *** fragatina has joined #openstack-kolla | 11:43 | |
| pbourke | coolsvap: the problem is both these repo files used to be under trunk | 11:44 |
| pbourke | now one has moved to buildlogs but delorean-deps has not | 11:44 |
| pbourke | so we would have to maintain two urls | 11:44 |
| *** Mr_Broken has quit IRC | 11:45 | |
| *** Mr_Broken has joined #openstack-kolla | 11:46 | |
| coolsvap | pbourke, i agree but both urls are currently active | 11:46 |
| *** williamcaban has quit IRC | 11:46 | |
| coolsvap | they were discussing a plan of about 3 months or N cycle before completely moving overe | 11:47 |
| *** rhallisey has joined #openstack-kolla | 11:49 | |
| *** Mr_Broken has quit IRC | 11:50 | |
| *** salv-orl_ has quit IRC | 11:53 | |
| *** coolsvap has quit IRC | 11:53 | |
| *** coolsvap_ has joined #openstack-kolla | 11:53 | |
| *** coolsvap has joined #openstack-kolla | 11:57 | |
| *** rhallisey has quit IRC | 11:57 | |
| *** rhallisey has joined #openstack-kolla | 11:58 | |
| *** coolsvap_ has quit IRC | 11:58 | |
| *** reddydodda has quit IRC | 12:02 | |
| *** ppowell has joined #openstack-kolla | 12:04 | |
| *** achanda has joined #openstack-kolla | 12:06 | |
| *** fragatina has quit IRC | 12:07 | |
| *** coolsvap has quit IRC | 12:08 | |
| *** achanda has quit IRC | 12:10 | |
| *** fragatina has joined #openstack-kolla | 12:22 | |
| *** zhiwei has quit IRC | 12:26 | |
| *** daneyon has joined #openstack-kolla | 12:30 | |
| *** Mr_Broken has joined #openstack-kolla | 12:30 | |
| *** daneyon has quit IRC | 12:34 | |
| *** zhiwei has joined #openstack-kolla | 12:39 | |
| *** neilus1 has joined #openstack-kolla | 12:39 | |
| *** neilus has quit IRC | 12:40 | |
| *** zhiwei has quit IRC | 12:43 | |
| *** salv-orlando has joined #openstack-kolla | 12:47 | |
| *** sdake has joined #openstack-kolla | 12:48 | |
| sdake | pbourke sweet delorean via the cdn | 12:49 |
| sdake | that will help | 12:49 |
| *** zhiwei has joined #openstack-kolla | 12:50 | |
| *** Mr_Broken has quit IRC | 12:50 | |
| *** Mr_Broken has joined #openstack-kolla | 12:51 | |
| *** fragatina has quit IRC | 12:51 | |
| *** signed8bit has joined #openstack-kolla | 12:54 | |
| *** zhiwei has quit IRC | 12:54 | |
| dmsimard | hi o/ | 12:55 |
| dmsimard | RDO | 12:55 |
| *** Mr_Broken has quit IRC | 12:55 | |
| dmsimard | We've moved to a static instance that can still be reached through trunk.rdoproject.org like you're used to for current and consistent repositories | 12:55 |
| dmsimard | current-passed-ci is now synchronized with a CDN and we encourage using it | 12:55 |
| dmsimard | the current-passed-ci on the CDN is slightly out of date but that will be updated soon | 12:56 |
| sdake | dmsimard we will be using the cdn - that gives us ha of our storage ;) | 12:58 |
| sdake | rather of our rpms | 12:58 |
| *** godleon has joined #openstack-kolla | 12:59 | |
| dmsimard | sdake: yes, keep in mind that current-passed-ci was validated by tripleo, packstack and puppet-openstack ci | 13:00 |
| dmsimard | so it's not necessarily updated every day if there are ongoing trunk issues | 13:00 |
| dmsimard | If you need to test something that isn't in current-passed-ci yet, you can try the consistent repository | 13:00 |
| *** zhiwei has joined #openstack-kolla | 13:01 | |
| *** bootsha has quit IRC | 13:01 | |
| *** shardy has quit IRC | 13:03 | |
| *** fragatina has joined #openstack-kolla | 13:05 | |
| *** zhiwei has quit IRC | 13:09 | |
| mewald | In task "Fetching Ceph keyrings" I get "ValueError: No JSON object could be decoded" I have no idea where to look. Any hints? | 13:15 |
| *** zhiwei has joined #openstack-kolla | 13:16 | |
| dmsimard | sdake: fyi CDN is up to date now. | 13:17 |
| sdake | mewald groan | 13:18 |
| dmsimard | sdake, pbourke: let me know if you happen to have any questions re: RDO repos | 13:18 |
| *** Mr_Broken has joined #openstack-kolla | 13:18 | |
| sdake | mewald are you running as sudo? | 13:19 |
| mewald | sdake: no everything as root | 13:19 |
| *** signed8bit is now known as signed8bit_Zzz | 13:21 | |
| *** zhiwei has quit IRC | 13:21 | |
| pbourke | dmsimard: is there a CDN location for http://trunk.rdoproject.org/centos7/delorean-deps.repo ? | 13:21 |
| *** signed8bit_Zzz is now known as signed8bit | 13:21 | |
| *** signed8bit is now known as signed8bit_Zzz | 13:21 | |
| dmsimard | pbourke: no but that repo file is mostly static | 13:22 |
| pbourke | ok thanks dmsimard | 13:22 |
| dmsimard | the one that changes is delorean.repo because it points to a specific hashed repository | 13:22 |
| pbourke | mewald: on the node that the ansible task failed on, try 'docker exec ceph_mon fetch_ceph_keys.py' | 13:23 |
| *** signed8bit_Zzz is now known as signed8bit | 13:26 | |
| *** coolsvap has joined #openstack-kolla | 13:29 | |
| *** david-lyle_ has joined #openstack-kolla | 13:32 | |
| mewald | pbourke: there is no ceph_mon container running on the controller nodes | 13:32 |
| *** inc0 has joined #openstack-kolla | 13:32 | |
| pbourke | mewald: that means you must have had a previous failure bootstrapping the mons | 13:33 |
| pbourke | mewald: if you're unsure Id recommend a cleanup and redeploy | 13:34 |
| *** dwalsh has joined #openstack-kolla | 13:34 | |
| inc0 | good morning | 13:34 |
| mewald | pbourke: cleanup the entire controller? :D | 13:34 |
| pbourke | mewald: just the ceph parts | 13:34 |
| *** jtriley has joined #openstack-kolla | 13:34 | |
| inc0 | mewald, what is issue? | 13:35 |
| *** achanda has joined #openstack-kolla | 13:35 | |
| *** david-lyle has quit IRC | 13:36 | |
| pbourke | inc0: the fetching keys task is failing | 13:36 |
| pbourke | inc0: and he has no monitors running | 13:36 |
| pbourke | which says to me they failed to bootstrap | 13:36 |
| inc0 | mewald, can you ssh from controllers to storage nodes | 13:36 |
| inc0 | ? | 13:36 |
| inc0 | and from controllers to each other? | 13:37 |
| inc0 | btw guys I have different issue which baffles me. I deploy neutron, run init-runonce, it creates vxlan net | 13:38 |
| sbezverk | rhallisey ping | 13:38 |
| inc0 | then I spawn a vm | 13:38 |
| rhallisey | hye | 13:38 |
| rhallisey | hey | 13:38 |
| inc0 | and I see tunnel at controller node/dhcp | 13:38 |
| inc0 | but no tunnel is created on compute | 13:38 |
| inc0 | no errors in logs, no nothing...just missing tunnel | 13:39 |
| sdake | mewald i'd recommend a complete cleanup | 13:40 |
| sdake | mewald that includes deleting everything in /etc/kolla that isnt a yml file | 13:40 |
| sbezverk | rhallisey qq for genconfig, vip address in globals.yml should be from the service range allocated to k8s cluster right? | 13:40 |
| sdake | cleanup-host should do it | 13:40 |
| *** salv-orlando has quit IRC | 13:40 | |
| *** salv-orlando has joined #openstack-kolla | 13:41 | |
| *** stvnoyes has joined #openstack-kolla | 13:42 | |
| *** zhiwei has joined #openstack-kolla | 13:43 | |
| *** gfidente is now known as gfidente|afk | 13:44 | |
| rhallisey | sbezverk, I'm not sure we use vip address for much any more. You need to set api_interface_address | 13:44 |
| rhallisey | net=host is being dropped everywhere | 13:45 |
| *** Mr_Broken has quit IRC | 13:45 | |
| wirehead_ | Yeah. | 13:46 |
| *** neilus1 has quit IRC | 13:46 | |
| *** achanda has quit IRC | 13:46 | |
| *** Mr_Broken has joined #openstack-kolla | 13:46 | |
| kfox1111 | sdake: no, I was not upgrading to trunk. and wasn't using kolla to do it. just foudn an old cloud and upgraded it. just saying, I ran into yet another case where being an op requires the flexability to fix things like that when unexpected breakage happens. | 13:46 |
| *** pmisiak has joined #openstack-kolla | 13:46 | |
| pmisiak | Hi, guys | 13:46 |
| wirehead_ | Nova / Neutron / Swift all are still net=host, so they will act.. interestingly... in the multi-node case. | 13:46 |
| sdake | kfox1111 remind me what we were talking about | 13:47 |
| *** zhiwei has quit IRC | 13:47 | |
| kfox1111 | wirehead_: you see my statements above? | 13:47 |
| sbezverk | wirehead_ I figured out the issue with my cluster, now it is wroking :-) | 13:47 |
| mewald | pbourke: why does kolla-deploy now show any other errors than this? | 13:48 |
| kfox1111 | sdake: the case where I was upgrading a keystone and the db_sync failed. | 13:48 |
| wirehead_ | sbezverk: Awe sum. Anything worth writing up for the docs? | 13:48 |
| pbourke | mewald: are sure no other errors scrolled by | 13:48 |
| wirehead_ | kfox1111: Yeah. I'm diving into Neutron and Nova next.. | 13:48 |
| pbourke | *are you | 13:48 |
| kfox1111 | sdake: the need of ops to sometimes be able to stop an upgrade workflow and manually play with it to fix things. :/ | 13:48 |
| kfox1111 | wirehead_: cool. | 13:48 |
| sbezverk | rhallisey could you give me example of globals.yml with this option? | 13:49 |
| *** coolsvap has quit IRC | 13:49 | |
| wirehead_ | kfox1111: One thing is that Kube isn't set to randomly bounce pods around without reason, so the only reason why you need to be 'termination safe' is to work around infra failure. | 13:49 |
| pmisiak | Have you fixed a glance issue in multinode deployment? When I uploaded image to glance it is stored on one controller node and you will have 404 when HAproxy direct your request to another glance. | 13:49 |
| *** huikang has joined #openstack-kolla | 13:50 | |
| sdake | pmisiak are you using cinder or nova on the frontend? | 13:50 |
| kfox1111 | wirehead_: still, hardware failure, data center power failure, or rolling upgrade might move them aroudn to differnet nodes, | 13:50 |
| sbezverk | wirehead_ it was my fault, it seems when I first brought up cluster it came up with wrong ip service range and created wrong cluster ip, so after I deleted wrong entry, a new one was created and strated working :-) | 13:50 |
| kfox1111 | and the rpc service list might get a lot of dead entries that need cleaning. | 13:50 |
| *** Mr_Broken has quit IRC | 13:50 | |
| *** coolsvap has joined #openstack-kolla | 13:50 | |
| rhallisey | sbezverk, https://review.openstack.org/#/c/327925/5 | 13:51 |
| patchbot | rhallisey: patch 327925 - kolla - Enable kolla k8s to override bind api bind address... | 13:51 |
| sbezverk | wirehead_ now trying to bring up mariadb in multinode, btw skydns is now working without restarts | 13:51 |
| *** huikang has quit IRC | 13:51 | |
| pmisiak | I can't spawn a VM sometimes (when image get request is directed to another glance instance). | 13:51 |
| rhallisey | inc0, ^ need that patch merged | 13:51 |
| sbezverk | rhallisey thank you | 13:51 |
| mewald | pbourke: yes I am sure, I just scrolled up | 13:51 |
| rhallisey | sbezverk, no problem | 13:51 |
| kfox1111 | heats going to have the same issue I think. heat-engine does the same kind of rpc server too. | 13:51 |
| *** huikang has joined #openstack-kolla | 13:51 | |
| mewald | pbourke: this is the stack trace, maybe helps: https://gist.github.com/mewald1/44f1a774cc2f360cc0f18ac25e82d6e2 | 13:51 |
| wirehead_ | kfox1111: Yeah. So how does the addition and deletion of RPC Workers happen in OpenStack outside of our weird environment? It's just plain old not termination safe? | 13:52 |
| pmisiak | sdake: as a workaround I run rsync to sync images files between contoller nodes) | 13:52 |
| inc0 | rhallisey, can't +2 since I'm author (rather co-author). But gates fails on deploy, could you take a look plz? | 13:53 |
| kfox1111 | I've been assured by the neutron folks restarting a neutron-server is safe, even though it contains rpc servers too. | 13:53 |
| kfox1111 | I think the request just times out and rabbit makes it available again. | 13:53 |
| pbourke | mewald: do you have the full output | 13:53 |
| inc0 | kfox1111, thats not how oslo.messaging work:( | 13:53 |
| inc0 | messages are acked when received afair | 13:54 |
| inc0 | unless they changed srg | 13:54 |
| kfox1111 | but if you ever delete a neutron-server I think you have to remove it with neutron agent-delete | 13:54 |
| rhallisey | inc0, ya | 13:54 |
| *** huikang has quit IRC | 13:54 | |
| wirehead_ | 'suse me, flashing back to the same conversation at the Portland summit. | 13:54 |
| *** huikang has joined #openstack-kolla | 13:55 | |
| kfox1111 | inc0: I've wondered how true it was... heat has clamed the same, that if you just shoot a heat engine, another would pick up the task. | 13:55 |
| inc0 | with heat case it's even harder | 13:55 |
| wirehead_ | Well, I can guarantee that This Will Be Tested, at the very least. :D | 13:55 |
| *** dave-mccowan has joined #openstack-kolla | 13:55 | |
| inc0 | but they have sth called convergence now | 13:55 |
| wirehead_ | Convergence should be restart-stable. | 13:55 |
| inc0 | where they track status of task in DB | 13:55 |
| inc0 | but I don't think it will work if you shoot engine in middle of parsing rabbitmq message. | 13:56 |
| inc0 | this is very narrow window, but it's there | 13:56 |
| kfox1111 | so, inc0: I've been wondering if anyone has put their openstack api services in a wsgi server rather then standaline. | 13:56 |
| inc0 | keystone | 13:56 |
| kfox1111 | for k8s termination to work, it needs a way to block the termination process until all connections are drained out. | 13:57 |
| kfox1111 | k8s itself has no functionality for that. :/ | 13:57 |
| kfox1111 | nginx or apache do though. | 13:57 |
| wirehead_ | kfox1111: there is support for an orderly termination of a pod in 1.3 | 13:57 |
| *** bootsha has joined #openstack-kolla | 13:57 | |
| kfox1111 | I'm using it, thats why I'm asking. I've got it working for one of my services, but it required hooking in nginx quit I think. | 13:58 |
| kfox1111 | otherwise, the terminating pod quit too quickly and sometimes killed connections. | 13:58 |
| mewald | pbourke: here is the entire output of deploy https://gist.github.com/15d12b9d59e4759880dd609ed6243b29 | 13:58 |
| sbezverk | rhallisey when I run genconfig all services gets assigned ip address of the interface where I generate this config, this beahvior does not make any sense.. | 13:58 |
| kfox1111 | I looked at using something line netstat to just look for open connections to block until they were drined, to do it more generically, | 13:58 |
| kfox1111 | but if the api process uses a connection pool, like to mysql, then it will never drop all connections. :/ | 13:59 |
| *** zhiwei has joined #openstack-kolla | 14:00 | |
| kfox1111 | so I think keystone will work since they are recommending using apache anyway. | 14:00 |
| sbezverk | rhallisey when you gen config for your hyperkube, which ip get assgined to let's say mariadb? | 14:00 |
| pbourke | mewald: what does your inventory look like | 14:01 |
| kfox1111 | I don't know if anyone embeds say, glance, cinder, neutron, etc apis though. | 14:01 |
| *** bootsha has quit IRC | 14:02 | |
| mewald | pbourke: inventory is this https://gist.github.com/anonymous/5003e566730a54ea50a6f600e83d3937 | 14:04 |
| *** zhiwei has quit IRC | 14:04 | |
| *** achanda has joined #openstack-kolla | 14:05 | |
| *** dwalsh has quit IRC | 14:05 | |
| pbourke | mewald: looks correct. trying to figure out why its skipping deploying the mons | 14:07 |
| *** achanda has quit IRC | 14:07 | |
| pbourke | mewald: can you try your deploy with -vvv | 14:07 |
| mewald | yep mom | 14:07 |
| *** neilus has joined #openstack-kolla | 14:07 | |
| *** huikang has quit IRC | 14:09 | |
| *** shardy has joined #openstack-kolla | 14:09 | |
| mewald | pbourke: with -vvv https://gist.github.com/anonymous/3e24cf654cc750ea475b72038f666f92 | 14:11 |
| pbourke | mewald: docker ps -a | 14:12 |
| pbourke | mewald: on the control | 14:12 |
| pbourke | mewald: any instances of ceph_mon ? | 14:12 |
| mewald | pbourke: no absolutely nothing | 14:12 |
| *** huikang has joined #openstack-kolla | 14:12 | |
| *** sdake has quit IRC | 14:13 | |
| mewald | pbourke: docker images shows mon and rgw image | 14:16 |
| *** sdake has joined #openstack-kolla | 14:16 | |
| *** zhiwei has joined #openstack-kolla | 14:16 | |
| *** kfox1111 is now known as kfox1111_away | 14:17 | |
| sdake | today is the day when i work on customization | 14:17 |
| sdake | ;) | 14:17 |
| inc0 | guys, can you do me a favor and deploy multinode kolla? | 14:18 |
| inc0 | and see if vxlans works well? | 14:18 |
| *** daneyon has joined #openstack-kolla | 14:18 | |
| inc0 | or rather, deploy kolla, run init-runonce | 14:18 |
| inc0 | spawn a vm | 14:18 |
| inc0 | and see if tunnel between dhcp agent and compute node was created successfully | 14:19 |
| pbourke | mewald: on control node, docker volume ls | 14:20 |
| *** zhiwei has quit IRC | 14:21 | |
| mewald | pbourke: there is a ceph_mon_config volume if that's what you want to hear :) | 14:22 |
| mewald | well, actually not on all of them | 14:22 |
| pbourke | mewald: try removing it on all control nodes, 'docker volume rm ceph_mon_config' | 14:22 |
| *** daneyon has quit IRC | 14:23 | |
| *** v1k0d3n has quit IRC | 14:23 | |
| mewald | pbourke: looks good now! thanks! Any ideas why this happened? | 14:27 |
| pbourke | mewald: tasks in kolla aim to be idempotent except bootstraps which generally dont take well to failures midway | 14:28 |
| pbourke | im assuming you either had a failure earlier on or halted it during a deploy | 14:29 |
| *** JeroenBo has quit IRC | 14:31 | |
| *** daneyon has joined #openstack-kolla | 14:32 | |
| rhallisey | sbezverk, 0.0.0.0 | 14:33 |
| *** zhiwei has joined #openstack-kolla | 14:33 | |
| sdake | sbezverk have you fixed https://bugs.launchpad.net/kolla/+bug/1572267 already? | 14:35 |
| openstack | Launchpad bug 1572267 in kolla mitaka "swift-rsyncd build breaks on Ubuntu if Docker uses AUFS" [Critical,Confirmed] - Assigned to Serguei Bezverkhi (sbezverk) | 14:35 |
| *** salv-orlando has quit IRC | 14:35 | |
| *** salv-orlando has joined #openstack-kolla | 14:36 | |
| *** daneyon_ has joined #openstack-kolla | 14:36 | |
| sbezverk | sdake it was fixed by not using capabilities | 14:36 |
| sdake | so that bug is invalid? | 14:37 |
| sbezverk | sdake and using high port for rsyncd | 14:37 |
| *** ayoung has quit IRC | 14:37 | |
| sbezverk | sdake I guess yes | 14:37 |
| *** zhiwei has quit IRC | 14:38 | |
| sbezverk | rhallisey what creates this file: /usr/local/bin/kolla_extend_start: line 22: /var/run/secrets/kubernetes.io/serviceaccount/token: No such file or directory | 14:38 |
| sbezverk | rhallisey I see mariadb is failing because it does not exist | 14:38 |
| rhallisey | weird. Kube usually creates it | 14:39 |
| *** daneyon has quit IRC | 14:39 | |
| *** inc0 has quit IRC | 14:41 | |
| mewald | I would be nice if post-deploy also gathered ceph keyring and config file onto the deploy node - what do you think? | 14:41 |
| *** Lyncos has left #openstack-kolla | 14:41 | |
| *** jrist has quit IRC | 14:43 | |
| pbourke | mewald: for backup purposes? | 14:43 |
| mewald | yeah that and to customize the crushmap, rulesets etc. That sort of stuff | 14:46 |
| *** inc0 has joined #openstack-kolla | 14:51 | |
| *** coolsvap has quit IRC | 14:52 | |
| sdake | so many bugs.. | 14:52 |
| sdake | all minor things | 14:53 |
| sdake | but holistically its a shit ton of work to fix | 14:53 |
| *** coolsvap has joined #openstack-kolla | 14:55 | |
| inc0 | right now I'm trying to pinpoint bug that, if in kolla, will be critical as fuck | 14:55 |
| *** zhiwei has joined #openstack-kolla | 14:55 | |
| inc0 | pardon my language | 14:55 |
| *** jrist has joined #openstack-kolla | 14:55 | |
| inc0 | sdake, you have multinode env right? | 14:56 |
| sdake | i do | 14:56 |
| inc0 | can you plz deploy latest master and see if vm will get ip? | 14:56 |
| inc0 | do 1 controller and 2 computes | 14:56 |
| inc0 | and vxlan network | 14:56 |
| sdake | vxlan is default, no? | 14:57 |
| inc0 | if you run init-runonce it will create vxlan | 14:57 |
| inc0 | demo-net will be vxlan | 14:57 |
| *** sacharya has joined #openstack-kolla | 14:57 | |
| sdake | what do you see, the vms don't get an ip addressvia dhcp? | 14:57 |
| mewald | Deploying a network node I see this: https://gist.github.com/mewald1/f6b916cca99ea957c82aaea6c3c9050f I set neutron_external_interface for the nodes to "ens5" which is the proper interface and it exists. Any ideas? | 14:57 |
| inc0 | I see that tunnel doesn't build between dhcp-agent and compute-node | 14:57 |
| inc0 | I see tunnel enter on dhcp node, but no tunnel exit at compute | 14:58 |
| sdake | inc0 so vm does get an ip address? | 14:58 |
| inc0 | if I spawn 2 vms on different computes, vxlan builds between them with no issues | 14:58 |
| inc0 | no...vm doesn't have L2 connectivity to dhcp | 14:58 |
| inc0 | so no ip | 14:58 |
| inc0 | or traffic | 14:58 |
| sdake | so nova list doesn't show an ip address assigned? | 14:59 |
| *** david-lyle_ is now known as david-lyle | 14:59 | |
| *** shardy has quit IRC | 14:59 | |
| sdake | i cannot ssh into my vms atm | 14:59 |
| sdake | my network is not properly setup for that | 14:59 |
| inc0 | sdake, nova list will show ip | 15:00 |
| *** b_bezak has quit IRC | 15:00 | |
| sdake | so if I can't 'inspect via nova list, then I wont be able to tell | 15:00 |
| *** zhiwei has quit IRC | 15:00 | |
| inc0 | but vm won't get it | 15:00 |
| sdake | inc0 i can't verify that one way or another | 15:00 |
| sdake | my network is not setup properly | 15:00 |
| *** dave-mccowan has quit IRC | 15:02 | |
| *** fragatina has quit IRC | 15:06 | |
| *** shardy has joined #openstack-kolla | 15:08 | |
| *** ppowell has quit IRC | 15:12 | |
| *** dwalsh has joined #openstack-kolla | 15:13 | |
| *** sdake_ has joined #openstack-kolla | 15:14 | |
| *** sdake has quit IRC | 15:16 | |
| *** dave-mccowan has joined #openstack-kolla | 15:19 | |
| *** vhosakot has joined #openstack-kolla | 15:19 | |
| *** sdake_ has quit IRC | 15:20 | |
| *** gfidente|afk is now known as gfidente | 15:20 | |
| *** belmoreira has quit IRC | 15:23 | |
| *** daneyon_ has quit IRC | 15:25 | |
| *** shardy has quit IRC | 15:25 | |
| *** zhiwei has joined #openstack-kolla | 15:28 | |
| *** JoseMello has joined #openstack-kolla | 15:28 | |
| *** ppowell has joined #openstack-kolla | 15:31 | |
| *** sdake has joined #openstack-kolla | 15:31 | |
| Mech422__ | inc0: as of a week ago, it worked multinode, dhcp and vxlan | 15:32 |
| *** v1k0d3n has joined #openstack-kolla | 15:32 | |
| *** salv-orlando has quit IRC | 15:33 | |
| inc0 | hmm | 15:33 |
| Mech422__ | inc0: however, the 'demo-net' is built with an arbritrary IP range having nothing to do the actual network | 15:33 |
| *** zhiwei has quit IRC | 15:33 | |
| inc0 | well that's not it | 15:33 |
| Mech422__ | inc0: so I had to manually hack the script to use correct ip/netmask | 15:33 |
| inc0 | vxlan isn't built at all | 15:33 |
| inc0 | that might be neutron issue | 15:34 |
| *** sdake_ has joined #openstack-kolla | 15:34 | |
| Mech422__ | inc0: let me see what I got running adm | 15:34 |
| inc0 | but regardless, it's sttrange | 15:34 |
| Mech422__ | err..atm | 15:34 |
| pbourke | sdake: 24 new bugs now | 15:34 |
| pbourke | down almost half | 15:34 |
| *** coolsvap has quit IRC | 15:35 | |
| Mech422__ | inc0: I got 1 vm running atm - want me to launch a few more ? | 15:36 |
| inc0 | Mech422__, can you redeploy with latest master? | 15:36 |
| *** sdake has quit IRC | 15:36 | |
| openstackgerrit | Merged openstack/kolla: Update kollaglue repo on docker to kolla https://review.openstack.org/329289 | 15:36 |
| Mech422__ | inc0: take me about an hour to rebuild and redeploy, but yes I can | 15:37 |
| inc0 | run a vm -> docker exec -it openvswitch_vswitchd ovs-vsctl show on every node | 15:37 |
| inc0 | see if br-tun has all the tunnels right | 15:37 |
| Mech422__ | inc0: should prolly do that before I tear down then, just to compare the before and after | 15:37 |
| *** ayoung has joined #openstack-kolla | 15:37 | |
| inc0 | nah, if it's working, thats cool | 15:37 |
| inc0 | I already see what's wrong with my env | 15:38 |
| inc0 | now I'm trying to find out why that happends | 15:38 |
| Mech422__ | inc0: remember - my networking is 'different' - I build all the ovs bridges myself | 15:38 |
| inc0 | vxlan tunnels should be setup anyhow right? | 15:38 |
| *** ssurana has joined #openstack-kolla | 15:38 | |
| *** darkmike has quit IRC | 15:38 | |
| *** godleon has quit IRC | 15:39 | |
| Mech422__ | inc0: it depends.... | 15:39 |
| Mech422__ | inc0: you get 1 vxlan for each 'user network' | 15:39 |
| Mech422__ | inc0: so if your not using user networks, you might not get it ? | 15:40 |
| Mech422__ | inc0: (eg, my 'external' network is vlan, not vxlan) | 15:40 |
| Mech422__ | inc0: you should get the bridge though I think | 15:41 |
| inc0 | br-tun was created | 15:41 |
| *** v1k0d3n has quit IRC | 15:41 | |
| inc0 | so vxlan tunnel is built when you add network with provider_type:vxlan | 15:41 |
| inc0 | or rather --provider network_type:vxlan | 15:41 |
| inc0 | then when you spawn a vm OVS creates 2 endpoints to a tunnel | 15:41 |
| inc0 | one on dhcp-agent node and another one on vm node | 15:42 |
| inc0 | pretty much if port in this net is on node, there will be vxlan tunnel to any other node with port | 15:42 |
| inc0 | issue is | 15:42 |
| Mech422__ | yeah - though dhcp is optional, as is a router | 15:43 |
| inc0 | I create vm, vxlan is created on dhcp node, not created on vm node | 15:43 |
| *** pmisiak has quit IRC | 15:43 | |
| Mech422__ | inc0: you assigned the vxlan network to the VM, right ? | 15:43 |
| *** g3ek has quit IRC | 15:44 | |
| sbezverk | rhallisey any chance to get k8s 1.2.4 rpm from RH? | 15:44 |
| *** haplo37 has quit IRC | 15:45 | |
| rhallisey | sbezverk, 1.2.0 is the latest I can see | 15:45 |
| *** dmk0202 has quit IRC | 15:46 | |
| inc0 | Mech422__, yeah | 15:46 |
| *** haplo37 has joined #openstack-kolla | 15:47 | |
| *** daneyon has joined #openstack-kolla | 15:47 | |
| *** coolsvap has joined #openstack-kolla | 15:47 | |
| *** g3ek has joined #openstack-kolla | 15:48 | |
| Mech422__ | inc0: build it running now - I cheated and didn't re-os the boxes | 15:48 |
| *** bootsha has joined #openstack-kolla | 15:48 | |
| *** bootsha has quit IRC | 15:48 | |
| inc0 | thanks a lot | 15:48 |
| *** huikang has quit IRC | 15:48 | |
| *** bootsha has joined #openstack-kolla | 15:48 | |
| *** zhiwei has joined #openstack-kolla | 15:50 | |
| *** athomas has quit IRC | 15:53 | |
| *** zhiwei has quit IRC | 15:54 | |
| sdake_ | vhosakot ping | 16:00 |
| vhosakot | sdake_: pong | 16:00 |
| sdake_ | i was going to abandon https://review.openstack.org/#/c/300747/ sincee you took over the work | 16:00 |
| patchbot | sdake_: patch 300747 - kolla - [WIP] Fix the Magnum failed to create trustee | 16:00 |
| sdake_ | could you point me at your patchset for magnum pls? | 16:01 |
| *** ppowell has quit IRC | 16:03 | |
| *** diogogmt has joined #openstack-kolla | 16:03 | |
| sdake_ | vhosakot ^^ | 16:04 |
| sbezverk | rhallisey right, so what I do is install 1.2.0 and then replace binaries from 1.2.4 | 16:05 |
| vhosakot | sdake_: yep, I told Jeffrey4l that his PS must be abandoned.. I think he agreed... my PS is at https://review.openstack.org/#/c/326163/ | 16:05 |
| patchbot | vhosakot: patch 326163 - kolla - Fix Magnum trustee issues | 16:05 |
| *** tfukushima has quit IRC | 16:05 | |
| sbezverk | rhallisey concerned that this hackish way could break things | 16:05 |
| sdake_ | vhosakot looks like need some action on that patch | 16:06 |
| vhosakot | sdake_: I will address the comments in my PS today... sure | 16:06 |
| rhallisey | sbezverk, idk when 1.2.4 will be packaged | 16:07 |
| vhosakot | sdake_: will make sure it gets merged today/tomorrow.. I have another PS as well I need to address comments in for prechecks on Ubuntu.. will do both today.. cool | 16:07 |
| sdake_ | wfm | 16:07 |
| sdake_ | thanks :) | 16:07 |
| vhosakot | sdake_: cool, np | 16:07 |
| Mech422__ | pbourke: Hey paul - whats the next step with that ceph patch? | 16:10 |
| Mech422__ | pbourke: we never really got a consensus on next steps? | 16:10 |
| pbourke | Mech422__: i have a local implementation that allows you to specify disks to use via globals.yml | 16:10 |
| pbourke | Mech422__: seems to work pretty well, just need to put it up for review. its less work than rewriting find_disks.py | 16:11 |
| pbourke | Mech422__: also I think its good to have that option rather than requiring operators to label all their disks | 16:11 |
| Mech422__ | pbourke: yeah - there was some talk about supporting both the labeled partitions AND passing a disk map via inventory vars ... | 16:12 |
| *** sacharya has quit IRC | 16:12 | |
| *** sacharya_ has joined #openstack-kolla | 16:12 | |
| Mech422__ | pbourke: that would give operators a release to 'port' to the inventory var system | 16:12 |
| *** zhiwei has joined #openstack-kolla | 16:12 | |
| Mech422__ | pbourke: but not sure if it was officially decided thats the way to go | 16:12 |
| pbourke | Mech422__: yeah, well we wouldnt even necessarily have to deprecate the disk labelling mechanism | 16:13 |
| pbourke | Mech422__: its good when it works | 16:13 |
| pbourke | but nice to have an alternative | 16:13 |
| Mech422__ | pbourke: we could try merging our patches, so we get support for both ? | 16:13 |
| pbourke | i'll see if I can submit it tomorrow, currently rebuilding my upstream dev environment and it takes ages | 16:14 |
| Mech422__ | pbourke: heh - I know your pain, my brother :-) | 16:14 |
| Mech422__ | pbourke: I usually drop a rebuild of the full env as the last thing I do before bed... | 16:15 |
| Mech422__ | pbourke: gives me time to catch up on slashdot and hacker news, and a wake up to fresh env. | 16:15 |
| pbourke | Mech422__: can you address Jeffrey4l's comments on the patch to see if he can remove his -2 | 16:15 |
| Mech422__ | pbourke: I've talked till I'm blue in the face | 16:15 |
| pbourke | lol | 16:15 |
| Mech422__ | pbourke: inc0 seemed to be making progress with him...but the -2 is still there | 16:16 |
| sdake_ | vhosakot https://review.openstack.org/#/c/314060/ | 16:16 |
| *** zhiwei has quit IRC | 16:16 | |
| patchbot | sdake_: patch 314060 - magnum - LBaaS API v2 | 16:16 |
| sdake_ | vhosakot can you work with that review with your current real deployment | 16:16 |
| vhosakot | sdake_: I pulled that PS and did not work in the past... yes, once my magnum fix is merged, I will pull https://review.openstack.org/#/c/314060/ and test with lbaas v2.. cool | 16:17 |
| patchbot | vhosakot: patch 314060 - magnum - LBaaS API v2 | 16:17 |
| Mech422__ | pbourke: the only thing find_disks uses udev for with my patch is finding the device, given a partition | 16:18 |
| Mech422__ | pbourke: but it has no support for doing file backed storage - which would be nice for AIO/dev envs. | 16:18 |
| pbourke | that's another topic i think | 16:19 |
| pbourke | hmm launchpad is down | 16:19 |
| Mech422__ | pbourke: if we're gonna do ceph config via vars - I actually think it should default to 'undefine' or file backed... | 16:20 |
| sdake_ | vhosakot can you comment on that lbaas v2 patch that it is nonfunctional for you | 16:20 |
| sdake_ | and report what you saw | 16:20 |
| Mech422__ | pbourke: otherwise, someone is gonna forget to edit the vars, and blow away a disk they are using | 16:20 |
| pbourke | Mech422__: i want to review Jeffrey4l's comments on the launchpad bug. he had some concerns you were renaming mounted drives | 16:20 |
| vhosakot | sdake_: cool, will pull the latest PS, test with my fix, and comment on it.. cool | 16:20 |
| sdake_ | thanks | 16:20 |
| vhosakot | sdake_: cool | 16:20 |
| Mech422__ | pbourke: I keep telling people - I didn't do that - thats sam yapple's original, existing, all ready in use code... | 16:21 |
| *** sdake_ is now known as sdake | 16:21 | |
| *** strigazi has joined #openstack-kolla | 16:21 | |
| Mech422__ | pbourke: all I did was change it to use sgdisk to read the partition table instead of udev | 16:21 |
| *** b_bezak has joined #openstack-kolla | 16:21 | |
| pbourke | Jeffrey4l: can you please reconsider your -2 on https://review.openstack.org/#/c/326609/ | 16:21 |
| patchbot | pbourke: patch 326609 - kolla - Read partition tables directly, bypassing stale ke... | 16:21 |
| Daviey | Has anyone got an example of using the neutron provider network as part of a vlan? | 16:23 |
| *** zhiwei has joined #openstack-kolla | 16:29 | |
| *** huikang has joined #openstack-kolla | 16:29 | |
| Mech422__ | Daviey: umm - sorta | 16:29 |
| Mech422__ | Daviey: My entire native host network stack is based on OVS | 16:30 |
| Mech422__ | Daviey: and I use vlans for pretty much everything | 16:30 |
| Mech422__ | Daviey: but I have to patch/hack Kolla to make it work | 16:30 |
| Mech422__ | Daviey: The trick is to expose ethX.YY or a veth connected to the vlan, to kolla | 16:33 |
| *** zhiwei has quit IRC | 16:33 | |
| Mech422__ | Daviey: and if your using OVS, don't let kolla start up the ovs-db container | 16:33 |
| *** huikang has quit IRC | 16:34 | |
| Mech422__ | inc0: Ok - rebuilt, post-deploy run, public and demo-nets created | 16:34 |
| Mech422__ | inc0: what do you want to check first ? | 16:34 |
| inc0 | spawn a vm | 16:35 |
| inc0 | and see if dhcp will go through | 16:35 |
| inc0 | in demo-net | 16:35 |
| *** matrohon has quit IRC | 16:35 | |
| Mech422__ | inc0: yep - got an addy in the correct range | 16:38 |
| Mech422__ | inc0: can ping the gateway | 16:38 |
| Mech422__ | inc0: suppose I could check the log to see if I got it via dhcp, or config-disk ... | 16:39 |
| inc0 | hmm | 16:39 |
| inc0 | ovs-vsctl show in compute node plz | 16:39 |
| inc0 | with vm | 16:39 |
| Mech422__ | kk - sec | 16:40 |
| *** buddy4321 has joined #openstack-kolla | 16:40 | |
| buddy4321 | having issues deploying kolla all-in-one node | 16:41 |
| buddy4321 | "Unknown error message: Tag 3.0.0 not found in repository docker.io/kolla/centos-binary-kolla-toolbox" | 16:41 |
| Mech422__ | inc0: http://paste.openstack.org/show/521706/ | 16:41 |
| inc0 | hmm looks good | 16:42 |
| inc0 | so it might be my issue | 16:42 |
| Mech422__ | buddy4321: 3.0.0 is for newton branch - did you checkout stable/mitaka by chance ? | 16:42 |
| inc0 | maybe something due to a fact of running it on vms | 16:42 |
| *** b_bezak has quit IRC | 16:43 | |
| Mech422__ | inc0: now triage is done, I can nuke the boxes back to metal, and rebuild/redploy - just to make sure it wasn't something left over from prior deploy ? | 16:43 |
| inc0 | Mech422__, I don't think it matters | 16:43 |
| inc0 | just keep an eye on this plz | 16:43 |
| inc0 | I'll try to nuke my env and see what happends | 16:44 |
| Mech422__ | inc0: I have been remiss...careless...unmindful....and playing with K8s | 16:44 |
| Mech422__ | inc0: I'll try to get back to rebuilding on a regular basis | 16:44 |
| inc0 | k8s does that to people | 16:44 |
| *** bootsha has quit IRC | 16:44 | |
| Mech422__ | inc0: :-) | 16:45 |
| Mech422__ | inc0: remember though - we lobotimized most of the network setup in my build... | 16:45 |
| *** hanchao has quit IRC | 16:45 | |
| Mech422__ | inc0: I pre-build the br-ex and stuff manually | 16:45 |
| buddy4321 | @Mech422 am just using kolla_install_type: "source" | 16:45 |
| Mech422__ | buddy4321: when you git 'cloned' kolla - which branch did you clone ? | 16:46 |
| buddy4321 | https://github.com/openstack/kolla.git | 16:46 |
| Mech422__ | hmm - thats master... | 16:46 |
| Mech422__ | Oh! coolsvap uploaded new images... | 16:47 |
| buddy4321 | Can I switch to stable kilo version? | 16:47 |
| Mech422__ | coolsvap: the images you uploaded - does that mean we need to bump the tag to 3.0.1 ? | 16:47 |
| inc0 | buddy4321, first...don't kilo is eol | 16:47 |
| inc0 | second, earliest stable kolla is liberty | 16:47 |
| Mech422__ | buddy4321: not sure that kilo/libery work - stable/mitaka worked for me | 16:47 |
| inc0 | liberty is ok now | 16:47 |
| Mech422__ | my bad :-P | 16:48 |
| Mech422__ | Give me Liberty....or Give me Mitaka...IDC :-P | 16:48 |
| Mech422__ | meeting...bbiab | 16:49 |
| buddy4321 | @Mech422 Master branch deploy Newton version? | 16:50 |
| Mech422__ | buddy4321: yes, they're working on newton support now :-) | 16:51 |
| *** zhiwei has joined #openstack-kolla | 16:51 | |
| Mech422__ | buddy4321: you can use stable/liberty or stable/mitaka (change the 'branch' drop down in github to get the url to clone) | 16:52 |
| Mech422__ | if you want 'stable' versions | 16:52 |
| buddy4321 | got it | 16:53 |
| buddy4321 | Thanks Mech422 | 16:53 |
| *** ppowell has joined #openstack-kolla | 16:53 | |
| *** jogam has joined #openstack-kolla | 16:55 | |
| *** zhiwei has quit IRC | 16:56 | |
| *** jogam82 has quit IRC | 16:56 | |
| *** fragatina has joined #openstack-kolla | 16:58 | |
| Daviey | Mech422__: Sorry for missing your response... So I have bond1.160, which is a bonded interface on vlan 160. I tried to use that as the external interface, and it didn't seem to have connectivity. | 16:59 |
| *** daneyon_ has joined #openstack-kolla | 17:01 | |
| *** daneyon_ has quit IRC | 17:05 | |
| *** jtriley has quit IRC | 17:06 | |
| *** jtriley has joined #openstack-kolla | 17:09 | |
| *** Jeffrey4l has quit IRC | 17:11 | |
| *** coolsvap has quit IRC | 17:11 | |
| *** Serlex has quit IRC | 17:12 | |
| jogam | pbourke :ping | 17:12 |
| *** zhiwei has joined #openstack-kolla | 17:14 | |
| *** signed8bit is now known as signed8bit_Zzz | 17:14 | |
| *** g3ek has quit IRC | 17:14 | |
| *** haplo37 has quit IRC | 17:14 | |
| *** vhosakot has quit IRC | 17:16 | |
| *** signed8bit_Zzz is now known as signed8bit | 17:16 | |
| *** haplo37 has joined #openstack-kolla | 17:17 | |
| *** g3ek has joined #openstack-kolla | 17:17 | |
| *** zhiwei has quit IRC | 17:19 | |
| *** huikang has joined #openstack-kolla | 17:19 | |
| *** coolsvap has joined #openstack-kolla | 17:20 | |
| *** sdake has quit IRC | 17:22 | |
| *** vhosakot has joined #openstack-kolla | 17:25 | |
| *** ayoung has quit IRC | 17:28 | |
| *** zhiwei has joined #openstack-kolla | 17:30 | |
| *** vhosakot has quit IRC | 17:33 | |
| *** mliima_ has joined #openstack-kolla | 17:34 | |
| *** zhiwei has quit IRC | 17:35 | |
| *** huikang has quit IRC | 17:35 | |
| Mech422__ | Daviey: yeah - kolla is gonna do some 'fun' stuff with your existing networking | 17:36 |
| inc0 | Daviey, it's not about kolla really, it's ovs | 17:36 |
| Mech422__ | Daviey: it'll try to add bond1.160 to the bridge (not sure - can linux native bonds be added to a bridge?) | 17:37 |
| *** mliima has quit IRC | 17:37 | |
| Daviey | Hmm.. is it OVS for provider network | 17:37 |
| Mech422__ | Daviey: I had to add a veth pair to my existing ovs bridge...ahh! | 17:37 |
| Mech422__ | Daviey: ok - so ovs bridges can't be directly added to another bridge | 17:37 |
| Mech422__ | Daviey: you either manually configure the bridges and use patch pairs | 17:38 |
| Daviey | Mech422__: yeah.. i created a new bridge and added the bond to that.. and declared the bridge.. but didn't get any dice | 17:38 |
| inc0 | yeah, you need a patch iface | 17:38 |
| Mech422__ | Daviey: or you add a veth pair and hand one 1 to ovs and the other end to kolla | 17:38 |
| Daviey | ah, i tried using a dummy interface | 17:38 |
| Mech422__ | Daviey: also, when kolla fires up the ovs-db container, existing ovs setup goes to shit | 17:38 |
| Daviey | :) | 17:38 |
| Mech422__ | Daviey: so you have to stop that container | 17:39 |
| Daviey | Mech422__: but i'm not sure OVS is relevant for the provider network? | 17:39 |
| Mech422__ | err....stop it from starting | 17:39 |
| Mech422__ | Daviey: anything that uses ovs will get hosed as soon as the new empty db is started | 17:39 |
| Daviey | as in.. demo1 network, is working as expected (non-vlan) | 17:39 |
| Daviey | but provider network is the phys network.. so shouldn't be using ovs, should it? | 17:40 |
| Mech422__ | ok - and how is provider network setup ? linux native or ovs | 17:40 |
| Daviey | (when added two interfaces to a VM) | 17:40 |
| inc0 | Daviey, demo-net is vxlan going over mgmt network | 17:40 |
| inc0 | so it wont go over your bond | 17:40 |
| Daviey | Mech422__: from the init-runonce script.. that | 17:40 |
| Daviey | right | 17:40 |
| Daviey | inc0: demo-net is working | 17:41 |
| Mech422__ | I'm lost - you have a bond on the host native network stack right ? | 17:41 |
| Daviey | yes | 17:41 |
| Mech422__ | is that bond created via linux native bonding ? or OVS ? | 17:41 |
| Daviey | native | 17:41 |
| Daviey | with vlan | 17:41 |
| Mech422__ | Daviey: ok - so the ovs stuff shouldn't effect it | 17:42 |
| Mech422__ | Daviey: and the vlan is untagged/native right ? | 17:42 |
| *** haplo37 has quit IRC | 17:42 | |
| *** salv-orlando has joined #openstack-kolla | 17:42 | |
| Daviey | network_interface: "eno2.101" <-- working as expected with demo-net | 17:42 |
| Daviey | neutron_external_interface: "bond1.160" <-- not working with phys net | 17:42 |
| Mech422__ | Daviey: and the bond is setup to automagically add the vlan tagg to traffic right ? | 17:43 |
| Daviey | butchered lines from init-runonce, https://gist.github.com/anonymous/f58925db8db4d4ee9a89df8acf78c26d | 17:43 |
| Mech422__ | not run once - the actual bond config - it set up the vlan tagging right ? | 17:44 |
| Daviey | Mech422__: yeah, i have bond1 which is the raw device, and bond.160 which is tagged as vlan 160 | 17:44 |
| inc0 | Daviey, can you remove tag from this iface and use neutron tagged network? | 17:44 |
| Daviey | sure | 17:44 |
| Mech422__ | Daviey: ok - so kolla is gonna take bond.160 and try to add it to an OVS bridge | 17:44 |
| inc0 | yeah I'm not too sure how this behaves in ovs | 17:45 |
| *** vhosakot has joined #openstack-kolla | 17:45 | |
| inc0 | how ovs handles native tagged ifaces | 17:45 |
| Daviey | I didn't think phys net would use ovs outside of the host | 17:45 |
| inc0 | however what I'm sure works is to add raw device and make neutron tag is as 160 | 17:45 |
| Mech422__ | Daviey: here's my working setup on a compute node: http://paste.openstack.org/show/521706/ | 17:46 |
| Daviey | inc0: Yeah.. I guess i'll do that. | 17:46 |
| Daviey | inc0: Do you have just enough config for neutron tagged interface? | 17:46 |
| *** g3ek has quit IRC | 17:46 | |
| Daviey | Mech422__: thanks | 17:47 |
| Mech422__ | Daviey: extrnl/extrnlh is my veth pair | 17:47 |
| *** gfidente is now known as gfidente|afk | 17:47 | |
| Mech422__ | Daviey: and extrnl gets added to br-ex | 17:47 |
| inc0 | Daviey, yeah, I'm pretty sure kolla undersands vlan provider network | 17:47 |
| *** zhiwei has joined #openstack-kolla | 17:47 | |
| Daviey | inc0: hmm.. I'm still a bit puzzled.. but i'll look more at this tommorrow | 17:47 |
| *** g3ek has joined #openstack-kolla | 17:48 | |
| Daviey | Considering the bond vlan was already functional, i was going to try and re-use that.. but i might need to look at neutron tagging | 17:48 |
| *** salv-orl_ has joined #openstack-kolla | 17:49 | |
| *** salv-orlando has quit IRC | 17:49 | |
| *** haplo37 has joined #openstack-kolla | 17:51 | |
| *** zhiwei has quit IRC | 17:52 | |
| *** ppowell_ has joined #openstack-kolla | 17:52 | |
| Daviey | thanks chaps.. i'll look more at this tomorrow | 17:54 |
| *** ppowell has quit IRC | 17:56 | |
| openstackgerrit | David Wang proposed openstack/kolla: WIP: Enable kolla k8s to override bind api bind address in genconfig https://review.openstack.org/327925 | 17:57 |
| *** dave-mccowan has quit IRC | 17:58 | |
| *** dwalsh_ has joined #openstack-kolla | 17:59 | |
| Mech422__ | Anyone got experience exposing a fairly large number (10-20) of services/ports to the internet thru a home consumer cable modem/AP (netgear in my case) | 18:00 |
| *** dwalsh has quit IRC | 18:00 | |
| Mech422__ | I gotta expose my horizon and stuff from my home network to the internet so I can demo it at a BoF | 18:00 |
| wirehead_ | Expand your horizons, so to speak. | 18:03 |
| wirehead_ | Lately, I’ve been using autossh: https://www.everythingcli.org/ssh-tunnelling-for-fun-and-profit-autossh/ | 18:04 |
| Mech422__ | wirehead_: my boss mentioned that - but ssh tunneling isn't something I'm familiar with | 18:06 |
| Mech422__ | wirehead_: you think that would be easier then port forwarding/snat ? | 18:06 |
| wirehead_ | Depends on how secure you want things to be, I suppose. | 18:06 |
| wirehead_ | A autossh tunnel from home to cloud server, then from your laptop to cloud server, gets complex. | 18:07 |
| Mech422__ | wirehead_: I just need it for like 1 day | 18:07 |
| Mech422__ | wirehead_: though, if it works - they'll prolly expect me to do it all the time :-P | 18:07 |
| wirehead_ | Heh. | 18:07 |
| wirehead_ | So, previously I used OpenVPN which works quite well but is actually a real pain in the butt to orchestrate. | 18:08 |
| wirehead_ | And it’s kinda stale. | 18:08 |
| wirehead_ | By comparison, AutoSSH devolves down to just SSH port forwarding. | 18:08 |
| *** fragatina has quit IRC | 18:08 | |
| Mech422__ | wirehead_: thanks - I'll RTFM on AutoSSH and see how hard it looks | 18:09 |
| wirehead_ | I’ve never tried SNAT. | 18:09 |
| wirehead_ | Pretty much, you can look how to do it with SSH. | 18:09 |
| Mech422__ | wirehead_: I didn't like the fact that I had to configure it on every device thats gonna access though | 18:09 |
| *** zhiwei has joined #openstack-kolla | 18:09 | |
| wirehead_ | And then autoSSH makes the connection reconnect in the face of inevitable network failures. | 18:09 |
| wirehead_ | I’ve never had to set up snat. | 18:10 |
| Mech422__ | wirehead_: ugghh - yeah, I get disconnected from the work vpn at least 1/day | 18:10 |
| *** sdake has joined #openstack-kolla | 18:11 | |
| *** zhiwei has quit IRC | 18:13 | |
| wirehead_ | The SSH protocol is quite an amazingly sophisticated thing, where just using it for encrypted terminal access just scratches the surface. | 18:15 |
| *** haplo37 has quit IRC | 18:15 | |
| wirehead_ | More-or-less, it’s an encrypted network tunnel mascarading as terminals, because X Windows connection forwarding. | 18:16 |
| *** jtriley has quit IRC | 18:17 | |
| *** dave-mccowan has joined #openstack-kolla | 18:19 | |
| *** g3ek has quit IRC | 18:20 | |
| *** g3ek has joined #openstack-kolla | 18:20 | |
| *** haplo37 has joined #openstack-kolla | 18:24 | |
| *** salv-orl_ has quit IRC | 18:28 | |
| *** jtriley has joined #openstack-kolla | 18:34 | |
| *** salv-orlando has joined #openstack-kolla | 18:36 | |
| *** huikang has joined #openstack-kolla | 18:36 | |
| *** vhosakot has quit IRC | 18:37 | |
| *** huikang has quit IRC | 18:41 | |
| *** coolsvap has quit IRC | 18:41 | |
| *** coolsvap has joined #openstack-kolla | 18:42 | |
| *** zhiwei has joined #openstack-kolla | 18:43 | |
| *** dwalsh_ has quit IRC | 18:44 | |
| *** dwalsh_ has joined #openstack-kolla | 18:47 | |
| *** vhosakot has joined #openstack-kolla | 18:47 | |
| *** g3ek has quit IRC | 18:47 | |
| *** vhosakot has quit IRC | 18:47 | |
| *** zhiwei has quit IRC | 18:47 | |
| *** vhosakot has joined #openstack-kolla | 18:48 | |
| *** huikang has joined #openstack-kolla | 18:48 | |
| *** huikang has quit IRC | 18:49 | |
| *** vhosakot has quit IRC | 18:55 | |
| *** vhosakot has joined #openstack-kolla | 18:56 | |
| *** g3ek has joined #openstack-kolla | 18:56 | |
| *** ssurana has quit IRC | 18:57 | |
| *** ssurana has joined #openstack-kolla | 18:57 | |
| *** zhiwei has joined #openstack-kolla | 19:00 | |
| *** zhiwei has quit IRC | 19:04 | |
| *** huikang has joined #openstack-kolla | 19:05 | |
| *** zhiwei has joined #openstack-kolla | 19:16 | |
| *** dwalsh_ has quit IRC | 19:18 | |
| *** daneyon has quit IRC | 19:18 | |
| *** g3ek has quit IRC | 19:20 | |
| *** zhiwei has quit IRC | 19:21 | |
| *** dwalsh_ has joined #openstack-kolla | 19:21 | |
| *** daneyon has joined #openstack-kolla | 19:25 | |
| *** g3ek has joined #openstack-kolla | 19:29 | |
| *** salv-orlando has quit IRC | 19:30 | |
| *** ayoung has joined #openstack-kolla | 19:31 | |
| *** dwalsh_ has quit IRC | 19:31 | |
| *** zhiwei has joined #openstack-kolla | 19:33 | |
| *** zhiwei has quit IRC | 19:38 | |
| *** daneyon_ has joined #openstack-kolla | 19:43 | |
| *** dwalsh_ has joined #openstack-kolla | 19:45 | |
| *** daneyon_ has quit IRC | 19:48 | |
| *** zhiwei has joined #openstack-kolla | 19:50 | |
| *** salv-orlando has joined #openstack-kolla | 19:52 | |
| *** zhiwei has quit IRC | 19:54 | |
| *** haplo37 has quit IRC | 19:56 | |
| *** g3ek has quit IRC | 19:56 | |
| *** fragatina has joined #openstack-kolla | 19:59 | |
| *** vhosakot has quit IRC | 20:02 | |
| *** vhosakot has joined #openstack-kolla | 20:02 | |
| *** haplo37 has joined #openstack-kolla | 20:03 | |
| *** g3ek has joined #openstack-kolla | 20:03 | |
| *** ayoung has quit IRC | 20:04 | |
| *** dwalsh_ has quit IRC | 20:07 | |
| *** ssurana has quit IRC | 20:08 | |
| *** ssurana has joined #openstack-kolla | 20:09 | |
| *** fragatina has quit IRC | 20:11 | |
| *** zhiwei has joined #openstack-kolla | 20:12 | |
| *** jogam has quit IRC | 20:16 | |
| *** zhiwei has quit IRC | 20:16 | |
| *** huikang has quit IRC | 20:16 | |
| *** ayoung has joined #openstack-kolla | 20:18 | |
| dcwangmit01_ | So, when recheck'ing the CI tests in Kolla https://review.openstack.org/#/c/327925/ , I believe there is a system failure that isn't to do with the code, but with building the images. Anyone know how to get beyond this: 2016-06-23 19:25:09.383281 | File contains no section headers. | 20:19 |
| patchbot | dcwangmit01_: patch 327925 - kolla - WIP: Enable kolla k8s to override bind api bind ad... | 20:19 |
| dcwangmit01_ | 2016-06-23 19:25:09.383300 | file: file:///etc/yum.repos.d/delorean.repo, line: 1 | 20:20 |
| *** dwalsh_ has joined #openstack-kolla | 20:21 | |
| *** huikang has joined #openstack-kolla | 20:24 | |
| *** vhosakot has quit IRC | 20:26 | |
| rhallisey | dcwangmit01_, ya looks like it | 20:27 |
| rhallisey | https://review.openstack.org/#/c/333252/ | 20:27 |
| patchbot | rhallisey: patch 333252 - kolla - Update urls for Delorean repo files | 20:27 |
| rhallisey | might have something to do with that? | 20:27 |
| rhallisey | idk | 20:27 |
| *** g3ek has quit IRC | 20:27 | |
| *** mewald has quit IRC | 20:27 | |
| *** mdorman has quit IRC | 20:27 | |
| *** NobodyCam has quit IRC | 20:27 | |
| *** sgordon has quit IRC | 20:27 | |
| *** Daviey has quit IRC | 20:27 | |
| *** pumaranikar has quit IRC | 20:27 | |
| *** zigo has quit IRC | 20:27 | |
| *** DuncanT has quit IRC | 20:27 | |
| *** smekel_ has quit IRC | 20:27 | |
| *** mdorman has joined #openstack-kolla | 20:27 | |
| *** mewald has joined #openstack-kolla | 20:27 | |
| *** sgordon has joined #openstack-kolla | 20:27 | |
| *** huikang has quit IRC | 20:28 | |
| *** zigo has joined #openstack-kolla | 20:29 | |
| *** zhiwei has joined #openstack-kolla | 20:29 | |
| *** haplo37 has quit IRC | 20:30 | |
| *** NobodyCam has joined #openstack-kolla | 20:30 | |
| *** ediardo has quit IRC | 20:30 | |
| *** DuncanT has joined #openstack-kolla | 20:31 | |
| *** smekel_ has joined #openstack-kolla | 20:31 | |
| *** jtriley has quit IRC | 20:32 | |
| *** haplo37 has joined #openstack-kolla | 20:33 | |
| *** zhiwei has quit IRC | 20:34 | |
| *** Daviey has joined #openstack-kolla | 20:34 | |
| dcwangmit01_ | Thanks rhallisey | 20:34 |
| *** pumaranikar has joined #openstack-kolla | 20:36 | |
| *** g3ek has joined #openstack-kolla | 20:37 | |
| *** esmiurium has joined #openstack-kolla | 20:38 | |
| *** buddy4321 has quit IRC | 20:40 | |
| *** vhosakot has joined #openstack-kolla | 20:44 | |
| *** ediardo has joined #openstack-kolla | 20:47 | |
| *** daneyon has quit IRC | 20:50 | |
| *** ayoung has quit IRC | 20:51 | |
| *** zhiwei has joined #openstack-kolla | 20:51 | |
| *** daneyon has joined #openstack-kolla | 20:51 | |
| *** jogam has joined #openstack-kolla | 20:54 | |
| *** zhiwei has quit IRC | 20:55 | |
| *** sacharya has joined #openstack-kolla | 21:01 | |
| sdake | groan | 21:02 |
| sdake | dcwangmit01_ ya we need that patch to merge - unfortunately it has a jenkins -1 vote on it tm | 21:02 |
| sdake | dmsimard ping re deps repo | 21:03 |
| *** haplo37 has quit IRC | 21:03 | |
| dmsimard | ? | 21:03 |
| sdake | https://review.openstack.org/#/c/333252/2/kolla/common/config.py | 21:03 |
| patchbot | sdake: patch 333252 - kolla - Update urls for Delorean repo files | 21:03 |
| sdake | dmsimard can you look at that question | 21:03 |
| sdake | and providee an answwer in the review or give me the anser and i'll provide it | 21:03 |
| dmsimard | replied | 21:04 |
| *** g3ek has quit IRC | 21:04 | |
| *** inc0 has quit IRC | 21:04 | |
| *** sacharya_ has quit IRC | 21:05 | |
| sdake | thanks reading | 21:05 |
| sdake | dmsimard moving deps woudl rock - then we would get cdn on that too | 21:06 |
| sdake | which means HA of our gate sources | 21:06 |
| dmsimard | sdake: we'll consider it thanks | 21:06 |
| sdake | understood | 21:06 |
| sdake | i know red hat gets a whole lot of requests for a whole lot of things | 21:07 |
| sdake | try to fit it in if possible tia :) | 21:07 |
| *** ozialien10 has quit IRC | 21:08 | |
| *** rhallisey has quit IRC | 21:08 | |
| *** haplo37 has joined #openstack-kolla | 21:09 | |
| *** g3ek has joined #openstack-kolla | 21:09 | |
| *** ppowell_ has quit IRC | 21:13 | |
| *** zhiwei has joined #openstack-kolla | 21:13 | |
| sdake | dmsimard hows ara coming along | 21:14 |
| sdake | dmsimard anything new and exciting | 21:14 |
| sdake | dcwangmit01_ that repo change will be merged within 24 hours | 21:15 |
| sdake | in the meantime i'd just be patient about the rechecks | 21:15 |
| sdake | Mech422__ liberty and mitaka are nearly the same codebase - so they both work almost identically | 21:16 |
| openstackgerrit | David Wang proposed openstack/kolla-kubernetes: Updated quickstart doc dependencies section https://review.openstack.org/333591 | 21:16 |
| dmsimard | sdake: I'm probably going to cut a release after I'm back from vacation in early july | 21:17 |
| dmsimard | nearing a first stable release | 21:17 |
| Mech422__ | sdake: ? | 21:17 |
| sdake | Mech422__ conversation from prior | 21:17 |
| sdake | not sure when | 21:17 |
| sdake | was in scrollback, you said kilo/liberty on't work | 21:17 |
| dmsimard | sdake: we sort of addressed your feedback around figuring out roles and stuff. | 21:17 |
| sdake | dmsimard did it end up in openstack git? | 21:17 |
| *** zhiwei has quit IRC | 21:17 | |
| Mech422__ | oh yeah - I can't remember if I actually got kolla/liberty working | 21:17 |
| Mech422__ | we run liberty in our existing cluster, but its not kolla based | 21:18 |
| sdake | dmsimard i don't recall what feedback i gave - hope it was helpful :) | 21:18 |
| dmsimard | sdake: if you look at http://logs.openstack.org/09/331509/7/gate/gate-ara-integration-centos-7/de2e330/logs/build/playbook/84a98bab-d55d-4ce9-bed1-65178c33dcbb/index.html | 21:18 |
| *** huikang has joined #openstack-kolla | 21:18 | |
| Mech422__ | kolla/mitaka seemed to work nice, and kolla/master/newton seems to work well too | 21:18 |
| dmsimard | sdake: you can now filter by file, see the file and click on the line number to see that particular action | 21:18 |
| sdake | dmsimard cool looking | 21:18 |
| dcwangmit01_ | sdake: Thanks. I've posted an update in the "override bind address" review. I've been unable to reproduce the issue on my local machine, but have a good theory about why it fails on the CI system and thus uploaded patch set 6. Even if the delorean patch is fixed, there still may be work to do. | 21:19 |
| sdake | dmsimard looks good | 21:19 |
| sdake | dmsimard go for brroke - ask infra for ara.openstack.org | 21:20 |
| dmsimard | sdake: well actually I submitted https://review.openstack.org/#/c/330874/ | 21:20 |
| patchbot | dmsimard: patch 330874 - openstack-infra/zuul - WIP: Integrate ARA reporting inside Zuul Launcher | 21:20 |
| sdake | the worst theey can do is ssay no :) | 21:20 |
| *** vhosakot has quit IRC | 21:21 | |
| dmsimard | so ara reports would be provided by default in http://logs.openstack.org/09/331509/7/gate/gate-ara-integration-centos-7/de2e330/_zuul_ansible/ | 21:21 |
| sdake | dmsimard you will have to figure out what that would look like | 21:21 |
| dmsimard | it's kinda hard though, cause they don't have a lot of testing around zuul | 21:22 |
| dmsimard | they test things in prod and revert if need be :p | 21:22 |
| dmsimard | that scares me a bit | 21:22 |
| sdake | what project does that gate job run against? | 21:22 |
| dmsimard | that particular one is ARA's own gate | 21:22 |
| dmsimard | we run integration tests that runs some test tasks and runs cli commands/generate the static site you saw | 21:22 |
| sdake | would be cool to have a kolla ara gate | 21:22 |
| sdake | or publish job | 21:23 |
| sdake | or something like that | 21:23 |
| dmsimard | sure, I can work on it when I'm back from pto :) | 21:23 |
| sdake | enjoy pto :) | 21:23 |
| *** huikang has quit IRC | 21:24 | |
| sdake | dmsimard that zuul launcher integratoin looks pretty cool - but what owuld be better is to see the kolla gate run (from my pov of course;) | 21:28 |
| dmsimard | porque no los dos | 21:28 |
| *** daneyon_ has joined #openstack-kolla | 21:31 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Adding debugging documentation for Keystone https://review.openstack.org/330788 | 21:33 |
| sdake | i onl yspeak english | 21:33 |
| sdake | unfortunatley | 21:33 |
| *** daneyon_ has quit IRC | 21:36 | |
| *** sdake has quit IRC | 21:38 | |
| *** haplo37 has quit IRC | 21:40 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Adding debugging documentation for Keystone https://review.openstack.org/330788 | 21:40 |
| *** g3ek has quit IRC | 21:40 | |
| *** huikang has joined #openstack-kolla | 21:40 | |
| *** haplo37 has joined #openstack-kolla | 21:41 | |
| *** vhosakot has joined #openstack-kolla | 21:41 | |
| *** g3ek has joined #openstack-kolla | 21:41 | |
| *** zhiwei has joined #openstack-kolla | 21:42 | |
| *** sdake has joined #openstack-kolla | 21:42 | |
| *** zhiwei has quit IRC | 21:46 | |
| *** huikang has quit IRC | 22:05 | |
| *** godleon has joined #openstack-kolla | 22:06 | |
| *** JoseMello has quit IRC | 22:09 | |
| *** g3ek has quit IRC | 22:10 | |
| *** signed8bit is now known as signed8bit_Zzz | 22:11 | |
| *** wznoinsk is now known as wznoinsk_ | 22:13 | |
| *** zhiwei has joined #openstack-kolla | 22:15 | |
| *** g3ek has joined #openstack-kolla | 22:19 | |
| *** zhiwei has quit IRC | 22:19 | |
| *** wznoinsk has joined #openstack-kolla | 22:21 | |
| *** wznoinsk_ has quit IRC | 22:22 | |
| *** ravig has joined #openstack-kolla | 22:23 | |
| *** dwalsh_ has quit IRC | 22:32 | |
| *** aNupoisc has joined #openstack-kolla | 22:33 | |
| *** diogogmt has quit IRC | 22:33 | |
| *** vhosakot has quit IRC | 22:39 | |
| *** zhiwei has joined #openstack-kolla | 22:43 | |
| *** diogogmt has joined #openstack-kolla | 22:44 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Glance works without host network https://review.openstack.org/333022 | 22:44 |
| *** zhiwei has quit IRC | 22:47 | |
| sdake | man our review queue is stuffed up | 22:53 |
| *** fragatina has joined #openstack-kolla | 22:57 | |
| *** salv-orl_ has joined #openstack-kolla | 22:59 | |
| *** zhiwei has joined #openstack-kolla | 23:00 | |
| *** salv-orlando has quit IRC | 23:02 | |
| *** fragatina has quit IRC | 23:03 | |
| *** fragatina has joined #openstack-kolla | 23:03 | |
| *** zhiwei has quit IRC | 23:04 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Adding documentation for services with multiple server types. https://review.openstack.org/333621 | 23:13 |
| *** zhiwei has joined #openstack-kolla | 23:22 | |
| *** vhosakot has joined #openstack-kolla | 23:22 | |
| *** bootsha has joined #openstack-kolla | 23:26 | |
| *** zhiwei has quit IRC | 23:26 | |
| *** fragatina has quit IRC | 23:30 | |
| *** bootsha has quit IRC | 23:30 | |
| *** sdake_ has joined #openstack-kolla | 23:31 | |
| *** sdake has quit IRC | 23:35 | |
| *** dmsimard is now known as dmsimard|pto | 23:48 | |
| *** daneyon has quit IRC | 23:50 | |
| *** zhiwei has joined #openstack-kolla | 23:50 | |
| *** sdake_ has quit IRC | 23:53 | |
| *** haplo37 has quit IRC | 23:53 | |
| *** g3ek has quit IRC | 23:54 | |
| *** zhiwei has quit IRC | 23:54 | |
| *** sdake has joined #openstack-kolla | 23:58 | |
| *** haplo37 has joined #openstack-kolla | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!