| *** britthouser has joined #openstack-kolla | 00:04 | |
| *** Pavo has joined #openstack-kolla | 00:04 | |
| *** sdake has quit IRC | 00:08 | |
| Jeffrey4l_ | i am in, too. | 00:38 |
|---|---|---|
| *** jtriley has joined #openstack-kolla | 00:45 | |
| *** jtriley has quit IRC | 00:50 | |
| Pavo | anyone here? having an issue with kolla-build | 01:23 |
| Pavo | http://pastebin.com/9Gb4EM4Y | 01:24 |
| Pavo | my kolla-build.conf looks like this | 01:24 |
| Pavo | http://pastebin.com/qHTT3SYK | 01:24 |
| Pavo | any suggestions? | 01:25 |
| mark-casey | Pavo: I usually just pass command line options to kolla-build, but a bit of searching suggests you need "[DEFAULT]" as the first line | 01:30 |
| Pavo | oh ok ty | 01:30 |
| mark-casey | Pavo: welcome... hope it works | 01:32 |
| Pavo | yes that worked | 01:32 |
| Pavo | ty | 01:32 |
| mark-casey | cool, welcome | 01:32 |
| Pavo | now I am having an issue with adding [Service] | 01:45 |
| Pavo | MountFlags=shared | 01:45 |
| Pavo | to /etc/systemd/system/docker.service.d/kolla.conf | 01:45 |
| Pavo | not seeing docker.service.d/ | 01:45 |
| *** sdake has joined #openstack-kolla | 01:47 | |
| Pavo | lol now ERROR:kolla.cmd.build:base:Tag mitaka not found in repository docker.io/library/centos | 01:48 |
| mark-casey | Pavo: base OS is centos and you're building centos containers, right? | 01:50 |
| Pavo | yes | 01:50 |
| Pavo | following this guide | 01:50 |
| Pavo | http://egonzalez.org/openstack-kolla-deployment-from-rdo-packages/ | 01:51 |
| mark-casey | you may have to create docker.service.d | 01:51 |
| Pavo | yeah figured that one out lol | 01:53 |
| mark-casey | IMO that guide is less than ideal because pushing to dockerhub will probably take a long time. are you deploying kolla multinode or all in one? | 01:54 |
| Pavo | 2 node deployment on VMs | 01:56 |
| Pavo | just trying it out | 01:56 |
| mark-casey | sure. are you running kolla-build and kolla-ansible and such inside one of these VMs on on another machine? | 01:57 |
| mark-casey | *or on | 01:57 |
| Pavo | seems like its trying to look for tag Mitaka in docker.io/library/centos when its trying to pull the images | 01:59 |
| *** Jeffrey4l_ has quit IRC | 02:00 | |
| *** daneyon_ has joined #openstack-kolla | 02:08 | |
| Pavo | I see all the images here https://hub.docker.com/u/kolla/ | 02:09 |
| *** daneyon_ has quit IRC | 02:12 | |
| mark-casey | Pavo: when you ran 'kolla-build -n pavo --push' you would actually find those images at https://hub.docker.com/u/pavo. The guide is having you make your own copy. That is why it said run 'docker login' and why it said the build/push would take 3-9 hours. | 02:13 |
| mark-casey | probably easier to run a local registry | 02:14 |
| *** rhallisey has quit IRC | 02:17 | |
| *** sdake has quit IRC | 02:17 | |
| mark-casey | Pavo: did you run kolla-build inside one of the VMs or doing that on your physical machine/another machine? | 02:19 |
| mark-casey | also, you'll find plenty of help here if you do want to deploy to a few VMs, but there is also an all-in-one Vagrant environment for trying Kolla out. I just don't use it much so I didn't recommend it first thing. | 02:22 |
| Pavo | yeah I made a local registry and used kolla-build --registry 192.168.1.40:4000 --push | 02:27 |
| Pavo | still saying the same thing though | 02:28 |
| mark-casey | oh, awesome. did you add the --insecure-registry flag to the docker systemd dropin? | 02:29 |
| mark-casey | in addition to MountFlags? | 02:29 |
| Pavo | yeah | 02:29 |
| Pavo | ok its working now | 02:30 |
| Pavo | just took the tag = mitaka out | 02:30 |
| mark-casey | yeah. sorry was going through the list coming around to that :D | 02:30 |
| Pavo | guess it will use mitaka since its the latest build | 02:30 |
| Pavo | I tell ya though, I am loving this openstack container idea though | 02:31 |
| mark-casey | not an expert on that part but it follows the version of kolla you check out | 02:31 |
| mark-casey | yea, me too | 02:31 |
| Pavo | or I gues I should say openstack containized | 02:31 |
| mark-casey | so if you check out stable/mitaka of kolla you deploy that version, if you check out (or pip install) master you build/deploy images that track master of the other projects | 02:32 |
| Pavo | I have been deploying openstack using packstack myself for about a year now | 02:33 |
| Pavo | but figured I would see how this goes | 02:33 |
| Pavo | and just starting in docker | 02:33 |
| Pavo | so its a BIG step for me | 02:33 |
| mark-casey | it's not perfect but I like this better than fuel which I was using before. The image model is better IMO and cleaning up and re-deploying is really easy because almost all changes on nodes were containerized | 02:34 |
| *** klint has joined #openstack-kolla | 02:35 | |
| Pavo | oh I started out with fuel | 02:35 |
| Pavo | f that noise | 02:35 |
| mark-casey | indeed :D | 02:35 |
| Pavo | had so........... many issues with fuel | 02:35 |
| Pavo | and with the new version the nics flip flop around sometimes which is a pain to TS | 02:35 |
| mark-casey | yeah. having not used docker a ton myself in the past kolla had a decent learning curve as well. but community support is pretty good | 02:37 |
| Pavo | ok hit another issue | 02:37 |
| Pavo | ERROR:kolla.image.build.kolla-toolbox:Get https://192.168.1.40:4000/v1/_ping: http: server gave HTTP response to HTTPS client | 02:37 |
| Pavo | why is it trying v1 of registry when I am running v2 | 02:38 |
| Pavo | anyway to fix that? | 02:38 |
| mark-casey | can you ps -ef | grep docker on the VMs? | 02:39 |
| Pavo | http://pastebin.com/J1qtHFay | 02:39 |
| Pavo | f0686b06c57b registry:2 "/entrypoint.sh /etc/" 50 minutes ago Up 48 minutes 0.0.0.0:4000->5000/tcp registry | 02:40 |
| mark-casey | I believe the dockerd line is missing --insecure-registry ... | 02:42 |
| Pavo | oh so you mean I need kolla-build --insecure-registry 192.168.1.40:4000 --push instead? | 02:43 |
| mark-casey | no. that part was right. what should happen is on the VMs when you add the drop in inside /etc/systemd/system/docker.service.d you include that flag on the ExecStart. Then systemctl daemon-reload and systemctl restart dockerd | 02:44 |
| Pavo | oh ok | 02:44 |
| Pavo | ok how do I do that | 02:46 |
| Pavo | I only have kolla.conf in that dir | 02:46 |
| mark-casey | yeah was trying to find it in docs or in some of my stuff | 02:47 |
| *** jtriley has joined #openstack-kolla | 02:47 | |
| Pavo | found it | 02:48 |
| Pavo | under /etc/sysconfig/docker | 02:48 |
| Pavo | but I have that | 02:48 |
| Pavo | whoops was missing an I | 02:49 |
| Pavo | lol | 02:49 |
| *** williamc_ has quit IRC | 02:52 | |
| mark-casey | oh ok. the best practice may have changed. you can do it where you did mountflags by setting execstart blank and then putting the original command with the flag as the following setting. Shouldn't matter as long as it shows in ps -ef | docker on your VMs | 02:53 |
| Pavo | well I guess I could have used the registry I am running on ym mac | 02:54 |
| Pavo | lol | 02:54 |
| mark-casey | perhaps. IIRC any non-localhost registry must be HTTPS and authenticated unless it is listed as that flag | 02:55 |
| mark-casey | of which you can list multiple | 02:55 |
| Pavo | well I changed it to localhost:4000 and its working it seems | 02:56 |
| mark-casey | is .40 the mac? | 02:57 |
| Pavo | nah a VM | 02:57 |
| Pavo | I will re-run it again and push it to my mac registry so if I blow these VMs away I will already have them in my other regsitry | 02:58 |
| mark-casey | sounds good. you'll know quickly if it's happy. that error about v1 registry (it was more mad about HTTP!=HTTPS) is the symptom of it not liking the registry security | 03:01 |
| Pavo | yeah but if I have it in that file I should be good though right | 03:02 |
| *** haplo37- has quit IRC | 03:03 | |
| *** jtriley has quit IRC | 03:04 | |
| mark-casey | yep | 03:05 |
| Pavo | wonder if I could build a centos repo container, so I can host my own repos | 03:07 |
| *** haplo37_ has joined #openstack-kolla | 03:09 | |
| *** haplo37_ has quit IRC | 03:09 | |
| mark-casey | probably. I've never done it on centos. I usually try to find fast public mirrors in a nearby city | 03:09 |
| *** haplo37_ has joined #openstack-kolla | 03:09 | |
| mark-casey | though I'm on ubntu | 03:09 |
| Pavo | yeah but if I wanted a completely offline system it would ideal | 03:10 |
| mark-casey | right | 03:10 |
| Pavo | should be able to just make a apache docker container and map the volume for /var/www/html to a filesystem share with all the repos | 03:10 |
| Pavo | I would think | 03:11 |
| *** dwalsh has joined #openstack-kolla | 03:33 | |
| *** dwalsh has quit IRC | 03:38 | |
| *** jtriley has joined #openstack-kolla | 03:56 | |
| *** daneyon_ has joined #openstack-kolla | 03:56 | |
| *** daneyon_ has quit IRC | 04:00 | |
| *** jtriley has quit IRC | 04:01 | |
| *** sdake has joined #openstack-kolla | 04:14 | |
| mark-casey | sdake: ping | 04:22 |
| sdake | sup mark-casey | 04:22 |
| mark-casey | on container portability... if you're on centos and ubuntu build fails, would you expect that bindmounting docker socket into an ubuntu container and trying again would help? | 04:23 |
| mark-casey | i.e. is it the kernel or the os that is the problem | 04:23 |
| sdake | kernel syscall interface, rtnetlink | 04:24 |
| sdake | bindimoutning the docker socket wont fix those things | 04:24 |
| mark-casey | yea, was afraid of that :) | 04:25 |
| sdake | mark-casey got a project underway: https://drive.google.com/open?id=0B8q6xDPETSkHZk01M1Z4YlN4d1E | 04:26 |
| sdake | i've moved far beyond that image - thank god | 04:26 |
| mark-casey | heh, sweet! | 04:26 |
| mark-casey | I remember you mentioning the... bah... head/pickup thing :D is it using a needle? | 04:27 |
| sdake | nah its the tonearm | 04:29 |
| sdake | the thing that holds the cartridge | 04:29 |
| *** rstarmer has joined #openstack-kolla | 04:29 | |
| sdake | the cartridge holds the stylus (needle) | 04:29 |
| mark-casey | right. makes sense | 04:29 |
| sdake | this is a a closeup of the bearing assembly: https://drive.google.com/open?id=0B8q6xDPETSkHX0gyN18zMUhBOW8 | 04:31 |
| mark-casey | looks very well engineered | 04:35 |
| Pavo | is this going to be an issue? http://pastebin.com/CRUYk2PD | 04:43 |
| Pavo | when I try to build those images it says binary not yet available for centos' | 04:44 |
| rstarmer | anyone seen an error in an all-in-one deployment TASK [common : Creating log volume], I'm getting a TypeError... http://pastebin.com/PuiU4vZy | 04:48 |
| *** sdake has quit IRC | 04:49 | |
| mark-casey | Pavo: I don't think so, not unless you need those services now; I don't think they'll be used for most "compute kit" or "compute kit+ceph" type setups | 04:51 |
| Pavo | oh ok | 04:52 |
| Pavo | getting 2 fails on deploy | 04:52 |
| Pavo | http://pastebin.com/0E0FN93Y | 04:52 |
| Pavo | and | 04:52 |
| Pavo | http://pastebin.com/uHMn4tif | 04:52 |
| mark-casey | rstarmer: retry with kolla-ansible -vvvv ... ? maybe more details there. | 04:55 |
| mark-casey | Pavo: it can't reach the registry. For example if you'd moved the images to your mac registry and hadn't updated globals.yml | 04:56 |
| mark-casey | well, no. it said it got a 500 | 04:56 |
| mark-casey | oh, yea. ok. no route to host | 04:57 |
| Pavo | yeah I seen that now | 05:00 |
| Pavo | tryikng to remove those images and getting (cannot be forced) - image has dependent child images | 05:00 |
| mark-casey | I had that earlier tonight and there was one named <none> | 05:00 |
| mark-casey | which was the child I needed to remove | 05:00 |
| rstarmer | mark-casey: I've not tried to debug the kolla_docker module, but that seems to be where it's complaining... http://pastebin.com/91DXugyW | 05:01 |
| mark-casey | rstarmer: https://bugs.launchpad.net/kolla/+bug/1587279 | 05:06 |
| openstack | Launchpad bug 1587279 in kolla "kolla-ansible deploy fail" [Critical,Confirmed] | 05:06 |
| rstarmer | mark-casey: # TODO(SamYaple): Replace with required_if when Ansible 2.0 lands | 05:08 |
| rstarmer | if (module.params.get('action') != 'pull_image' | 05:08 |
| rstarmer | I guess that means I just need to make sure I have 1.9.4 installed... which gets me to my next error, but I'll rebuild and force 1.9.4. | 05:09 |
| rstarmer | are there "clean this up to nothing" instructions somewhere? | 05:09 |
| mark-casey | there are clean-containers and clean-images scripts that need to run on the target nodes even though they're usually on the deployer node | 05:10 |
| mark-casey | I do it like this: ansible -i multinode -m script -a '/kolla/tools/cleanup-containers' all | 05:10 |
| mark-casey | ansible -i multinode -m script -a '/kolla/tools/cleanup-images' all | 05:10 |
| mark-casey | the thing in -a being the path on deploy host | 05:11 |
| rstarmer | thx | 05:11 |
| *** fragatina has joined #openstack-kolla | 05:13 | |
| *** fragatina has quit IRC | 05:17 | |
| *** Pavo has quit IRC | 05:19 | |
| *** unicell has quit IRC | 05:19 | |
| *** unicell has joined #openstack-kolla | 05:19 | |
| *** sdake has joined #openstack-kolla | 05:31 | |
| *** zhubingbing has joined #openstack-kolla | 05:42 | |
| *** daneyon_ has joined #openstack-kolla | 05:44 | |
| *** daneyon_ has quit IRC | 05:48 | |
| *** sdake has quit IRC | 05:50 | |
| zhubingbing | hello | 05:51 |
| *** sdake has joined #openstack-kolla | 06:03 | |
| *** sdake has quit IRC | 06:07 | |
| *** senk_ has joined #openstack-kolla | 06:13 | |
| rstarmer | any idea why, if I didn't set up a local registry, 'kolla-ansible deploy' is trying to pull a kolla-toolbox container from docker.io, and even a different version than in my globals.yml? | 06:14 |
| rstarmer | nevermind, my mistake, I apparently _did_ set the version in globals... | 06:21 |
| rstarmer | all that trouble because of ansible 2.0... wow :( | 06:43 |
| mark-casey | yeah. it was a pretty substantial refactor from 1.9.x | 06:44 |
| mark-casey | however the porting has already been done in kolla master | 06:45 |
| mark-casey | to use ansible 2 | 06:45 |
| mark-casey | later all | 06:49 |
| *** mark-casey has quit IRC | 06:50 | |
| *** unicell has quit IRC | 06:51 | |
| *** zhurong has joined #openstack-kolla | 07:11 | |
| *** zhurong has quit IRC | 07:24 | |
| *** bootsha has joined #openstack-kolla | 07:28 | |
| *** daneyon_ has joined #openstack-kolla | 07:32 | |
| *** daneyon_ has quit IRC | 07:37 | |
| *** bootsha has quit IRC | 08:18 | |
| zhubingbing | Shaun Smekel | 08:27 |
| zhubingbing | hi | 08:27 |
| *** Jeffrey4l has joined #openstack-kolla | 08:30 | |
| *** Jeffrey4l has quit IRC | 08:38 | |
| *** bootsha has joined #openstack-kolla | 08:41 | |
| *** bootsha has quit IRC | 08:43 | |
| *** bootsha has joined #openstack-kolla | 08:45 | |
| *** zhubingbing_ has joined #openstack-kolla | 08:52 | |
| *** zhubingbing has quit IRC | 08:54 | |
| *** bootsha has quit IRC | 09:00 | |
| zhubingbing_ | Anybody here? who can help me？ | 09:11 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the memcache location value in horizon local settings https://review.openstack.org/355232 | 09:17 |
| *** Jeffrey4l has joined #openstack-kolla | 09:23 | |
| *** hieulq_ has joined #openstack-kolla | 09:28 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: repair gnocchi dockerfile https://review.openstack.org/348388 | 09:32 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add gnocchi ansible role https://review.openstack.org/349351 | 09:41 |
| *** hieulq_ has quit IRC | 09:43 | |
| *** hieulq_ has joined #openstack-kolla | 09:44 | |
| *** bootsha has joined #openstack-kolla | 09:50 | |
| *** opennode has joined #openstack-kolla | 10:09 | |
| *** duonghq has joined #openstack-kolla | 10:13 | |
| *** bootsha has quit IRC | 10:16 | |
| *** zhubingbing_ has quit IRC | 10:21 | |
| *** bootsha has joined #openstack-kolla | 10:27 | |
| *** senk__ has joined #openstack-kolla | 10:37 | |
| *** senk_ has quit IRC | 10:37 | |
| *** hieulq_ has quit IRC | 10:38 | |
| *** hieulq_ has joined #openstack-kolla | 10:48 | |
| *** vbel has quit IRC | 11:00 | |
| *** vbel has joined #openstack-kolla | 11:00 | |
| *** bootsha has quit IRC | 11:06 | |
| *** kbaegis has joined #openstack-kolla | 11:16 | |
| kbaegis | Hi all | 11:16 |
| *** zhubingbing has joined #openstack-kolla | 11:25 | |
| *** zhurong has joined #openstack-kolla | 11:33 | |
| *** hieulq_ has quit IRC | 11:43 | |
| duonghq | hello there | 11:46 |
| *** bootsha has joined #openstack-kolla | 11:56 | |
| kbaegis | Anyone know what the docker_namespace globals variable is used for? | 11:58 |
| kbaegis | I have a registry setup, but I can't find them with a curl | 11:58 |
| kbaegis | "curl https://crucible.lab:4000/v2/_catalog {"repositories":["ubuntu"]}" | 11:58 |
| kbaegis | Oh, my fault | 11:59 |
| kbaegis | nvm | 11:59 |
| *** zhubingbing has quit IRC | 11:59 | |
| *** klint has quit IRC | 12:40 | |
| kbaegis | Hey, I can set the "openstack_release" to a blank string, right? | 12:42 |
| kbaegis | kolla-build didn't tag my images by the looks of it | 12:43 |
| kbaegis | "The requested image does not exist: crucible.lab:4000/kolla/centos-source-heka:" | 12:43 |
| kbaegis | whereas: "curl https://crucible.lab:4000/v2/_catalog yields "kolla/centos-binary-heka" | 12:44 |
| kbaegis | Oh | 12:44 |
| *** kbaegis has quit IRC | 12:54 | |
| *** kbaegis has joined #openstack-kolla | 12:54 | |
| kbaegis | Evidently forgot to reconfigure kolla-build.conf :) | 12:55 |
| kbaegis | At least I got to configure a full 32 threads this time around :) | 12:55 |
| duonghq | kbaegis: So, I made that? | 12:56 |
| *** zhubingbing has joined #openstack-kolla | 12:56 | |
| duonghq | *you | 12:56 |
| zhubingbing | hi | 12:56 |
| zhubingbing | duonghq | 12:57 |
| duonghq | nice to see you, zhubingbing | 12:57 |
| zhubingbing | How do you spend your weekends | 12:57 |
| *** kbaegis has quit IRC | 12:58 | |
| duonghq | this weekend is really busy for me, hope that it is easier in Sep | 12:59 |
| duonghq | and you? | 12:59 |
| *** kbaegis has joined #openstack-kolla | 13:02 | |
| duonghq | kbaegis: is your Kolla ok right now? | 13:02 |
| zhubingbing | me too，so busy | 13:05 |
| kbaegis | duonghq Yeah, it's chugging along :) Forgot to run it inside screen though | 13:06 |
| duonghq | nice | 13:07 |
| kbaegis | I'm the crazy guy trying to deploy kolla on gentoo. HAProxy was broken earlier, so I'm trying the compile-from-source approach to see if the centos compiler chain doesn't miss something the packaged builds do | 13:07 |
| *** tfukushima has joined #openstack-kolla | 13:07 | |
| kbaegis | Been a great learning opportunity for some of dockers edge-cases, configs, etc | 13:08 |
| duonghq | kbaegis: which branch do you use? master? | 13:09 |
| kbaegis | Yeah. Tried rolling back to mitaka, but that didn't work either | 13:10 |
| kbaegis | container was stuck in an ugly loop | 13:10 |
| duonghq | hmm, you pushed this to new land, very interesting | 13:12 |
| kbaegis | :) I'll keep you guys updated | 13:12 |
| kbaegis | Provided it remains interesting | 13:12 |
| duonghq | Hope that you write something like blog entry after everything is up and run. | 13:13 |
| kbaegis | I'll try to do that. :) I only have 3 machines in my lab: 2 d525 atoms, and my new 16 core xeon box | 13:15 |
| kbaegis | The next step is to see if I can run the controllers in 2-4 gb of ram w/ swap | 13:15 |
| kbaegis | Purely as standby nodes | 13:15 |
| duonghq | all 3 machines run gentoo, huh? | 13:16 |
| kbaegis | Yes :D | 13:18 |
| duonghq | look forward to your result, I used to be Gentoo user :) | 13:20 |
| *** bootsha has quit IRC | 13:23 | |
| *** bootsha has joined #openstack-kolla | 13:23 | |
| *** chrichip has quit IRC | 13:25 | |
| *** chrichip has joined #openstack-kolla | 13:25 | |
| *** dwalsh has joined #openstack-kolla | 13:25 | |
| kbaegis | Not anymore? | 13:27 |
| duonghq | Yeah, now I use Arch and Ubuntu | 13:27 |
| kbaegis | It's a PITA if you don't have your own build server | 13:27 |
| kbaegis | If you do, it's just like running canonical :) | 13:27 |
| duonghq | but I really like emerge's slot of Gentoo | 13:28 |
| kbaegis | Yes | 13:28 |
| kbaegis | Those are handy. My personal favorite though are the unlicensed binaries- ffmpeg and all | 13:28 |
| kbaegis | You can compile them, but you can't redistribute | 13:28 |
| kbaegis | Lots of codecs, etc | 13:29 |
| duonghq | yes | 13:30 |
| duonghq | someday I'll comeback to Gentoo | 13:33 |
| *** bootsha has quit IRC | 13:33 | |
| kbaegis | Yeah, it's a cool community. It really makes you understand how stuff works. Gentoo: because linux isn't hard enough ;) | 13:34 |
| duonghq | agree | 13:36 |
| duonghq | tuning compiling process is really fun | 13:37 |
| duonghq | (compile arguments indeed) | 13:37 |
| kbaegis | I'm going to be setting up some heat templates and mistral flows for an autoscaling package compiler. I think gentoo really needs something like that. Step 1, though, is getting kolla up | 13:39 |
| duonghq | Do you think it's quite overkill? | 13:40 |
| duonghq | (IMO) | 13:40 |
| kbaegis | Yeah, absolutely. :) | 13:40 |
| kbaegis | But it's good fun and practice | 13:40 |
| duonghq | absolutely | 13:41 |
| *** duonghq has left #openstack-kolla | 13:46 | |
| *** duonghq has quit IRC | 13:46 | |
| openstackgerrit | Merged openstack/kolla: Add --user-domain when adding heat role admin https://review.openstack.org/353471 | 13:57 |
| openstackgerrit | Merged openstack/kolla: Add upgrade tasks for Ceilometer https://review.openstack.org/351007 | 14:08 |
| openstackgerrit | Merged openstack/kolla: Minor Readability Adjustments to Operating Kolla https://review.openstack.org/347099 | 14:08 |
| kbaegis | Soo. Here's my kolla-build.conf: http://pastebin.com/016hg8ti | 14:13 |
| kbaegis | kolla build completes/pushes successfully. The images show up appropriately under $(docker images) but when I do curl https://crucible.lab:4000/v2/_catalog, they're nowhere to be fou d | 14:14 |
| kbaegis | *found | 14:14 |
| openstackgerrit | Merged openstack/kolla: Customizations for Tempest https://review.openstack.org/354397 | 14:14 |
| *** dwalsh has quit IRC | 14:17 | |
| kbaegis | IOError: [Errno 2] No such file or directory: '/var/lib/kolla/config_files/config.json' in haproxy | 14:19 |
| kbaegis | weird. It looks like it's not copying the directories in | 14:20 |
| kbaegis | Does that sound right? | 14:20 |
| *** Pavo has joined #openstack-kolla | 14:21 | |
| kbaegis | Huh. looking at the dockerfile, this should work. Somehow it's not executing the && operator | 14:25 |
| kbaegis | Or it's reading the \\n ineffectively | 14:25 |
| kbaegis | The group gets created, but the mkdir -p doesn't | 14:30 |
| kbaegis | weird | 14:30 |
| kbaegis | when you run kolla-build, is that reading from the docker files under /usr/share/kolla/docker/? | 14:47 |
| kbaegis | I'm trying to figure out why the keepalived image is hanging "Restarting (1)" | 14:54 |
| *** dave-mccowan has joined #openstack-kolla | 14:55 | |
| *** zhurong has quit IRC | 14:56 | |
| kbaegis | Seems like it's crashing on INFO:__main__:Validating config file | 14:56 |
| *** dave-mcc_ has joined #openstack-kolla | 14:58 | |
| *** dave-mccowan has quit IRC | 15:01 | |
| kbaegis | is there an option in globals to disable the VIP? | 15:05 |
| *** dave-mccowan has joined #openstack-kolla | 15:07 | |
| *** dave-mcc_ has quit IRC | 15:09 | |
| kbaegis | well, centos-source-keepalived is the ONLY thing I can't get working | 15:12 |
| kbaegis | haproxy is actually working | 15:13 |
| Pavo | ok I have on the deployer node a running registry but when I do a kolla-build --registry 192.168.1.50:4000 --push, getting an error during build that that host is trying to use v1 and is not responding, I have INSECURE_REGISTRY="--insecure-registry 192.168.1.50:4000" in /etc/sysconfig/docker | 15:21 |
| Pavo | any suggestions | 15:21 |
| *** kbaegis has quit IRC | 15:32 | |
| *** kbaegis has joined #openstack-kolla | 15:32 | |
| kbaegis | Hey Pavo. Only way I got it working was by setting up a secure registry: https://docs.docker.com/registry/deploying/#get-a-certificate | 15:33 |
| Pavo | ok ty I'll take a look at that | 15:34 |
| kbaegis | That seems to work fine. Otherwise you'll get errors like "https sent to http endpoint" or something | 15:34 |
| kbaegis | My actual command: docker run -d -p 192.168.254.3:4000:5000 --restart=always --name registry -v $(pwd)/certs:/certs -e REGISTRY_HTTP_TLS_CERTIFICATE=/certs/domain.crt -e REGISTRY_HTTP_TLS_KEY=/certs/domain.key registry:2 | 15:35 |
| Pavo | how would that work for Kolla when it tries to use 4000 instead of 5000 | 15:35 |
| kbaegis | See above. It's the same | 15:35 |
| Pavo | oh ok I see now | 15:35 |
| Pavo | that should be in Kolla documentation | 15:35 |
| kbaegis | You may need DNS for a FQDN | 15:35 |
| kbaegis | Well, there's some complexity :) | 15:35 |
| kbaegis | I run BIND locally, but not everyone has that | 15:36 |
| Pavo | ok so how do I get the certs? | 15:39 |
| kbaegis | mkdir certs&&openssl req -newkey rsa:8192 -nodes -sha512 -keyout certs/domain.key -x509 -days 3650 -out certs/domain.crt | 15:39 |
| kbaegis | from ~/ | 15:39 |
| Pavo | ah ok yeah forgot about oppenssl | 15:40 |
| kbaegis | Your OS has a specific way to then update the certs | 15:40 |
| kbaegis | I use gentoo, so mine is different | 15:40 |
| Pavo | yeah I use centos | 15:41 |
| kbaegis | ca-update or something | 15:41 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the ubuntu xenial gate https://review.openstack.org/355149 | 16:23 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Get packages from cdn rather than github https://review.openstack.org/329060 | 16:26 |
| *** zhubingbing has quit IRC | 16:30 | |
| *** daneyon_ has joined #openstack-kolla | 16:34 | |
| *** daneyon_ has quit IRC | 16:38 | |
| kbaegis | So when I execute a kolla container with 'bash', I can't find a /var/lib/kolla directory | 17:10 |
| kbaegis | is this supposed to be in an attached volume, copied, in or created? | 17:11 |
| kbaegis | All I see is that keepalived container is perpetually restarting. I'm trying to fix it, but the logs are useless | 17:11 |
| kbaegis | there's a bunch of junk under /etc/keepalived/keepalived.conf | 17:13 |
| kbaegis | I can't even tell if that's used though | 17:14 |
| kbaegis | If anyone could explain this to me, it would be greatly appreciated | 17:17 |
| kbaegis | done with a -v looks like | 17:21 |
| *** tfukushima has quit IRC | 17:23 | |
| kbaegis | Hmm | 17:27 |
| kbaegis | Well, finally got it fully deployed. HAProxy isn't assigning the VIP appropriately. Workaround: sudo ip addr add <VIP CIDR> dev <IF> | 17:48 |
| kbaegis | Going to report it on the repo as a bug | 17:48 |
| *** opennode has quit IRC | 18:01 | |
| *** senk__ has quit IRC | 18:04 | |
| *** dave-mccowan has quit IRC | 18:18 | |
| *** daneyon_ has joined #openstack-kolla | 18:22 | |
| *** tfukushima has joined #openstack-kolla | 18:24 | |
| *** daneyon_ has quit IRC | 18:27 | |
| *** tfukushima has quit IRC | 18:28 | |
| kbaegis | neutron-server returning the following: http://pastebin.com/7KzJSMpD | 18:46 |
| kbaegis | build log: INFO:kolla.image.build.openstack-base:Cannot retrieve metalink for repository: epel/x86_64. Please verify its path and try again | 18:47 |
| *** dave-mccowan has joined #openstack-kolla | 18:55 | |
| *** dave-mcc_ has joined #openstack-kolla | 18:57 | |
| *** opennode has joined #openstack-kolla | 18:59 | |
| *** dave-mccowan has quit IRC | 19:00 | |
| *** kbaegis has quit IRC | 19:24 | |
| *** tfukushima has joined #openstack-kolla | 19:25 | |
| *** tfukushima has quit IRC | 19:30 | |
| *** kbaegis has joined #openstack-kolla | 19:49 | |
| Pavo | kbaegis ok I tried the solution you told me about the registry being TLS and still not working on deploy | 19:49 |
| Pavo | getting this error | 19:49 |
| Pavo | http://pastebin.com/R6rX7pwA | 19:50 |
| Pavo | and I used this docker command | 19:50 |
| Pavo | docker run -d -p 192.168.1.50:4000:5000 --restart=always --name registry -v `pwd`/certs:/certs -e REGISTRY_HTTP_TLS_CERTIFICATE=/certs/domain.crt -e REGISTRY_HTTP_TLS_KEY=/certs/domain.key registry:2 | 19:50 |
| kbaegis | "192.168.1.50:4000:"? | 19:55 |
| kbaegis | Did you have a spare : at the end? | 19:55 |
| Pavo | yeah I copied your command you posted | 19:55 |
| kbaegis | You should be able to curl 192.168.1.50:4000/v2/_ping and get a 404 | 19:55 |
| kbaegis | Try that | 19:55 |
| Pavo | curl: (7) Failed connect to 192.168.1.50:4000; Connection refused | 19:56 |
| Pavo | bae090cb7b99 registry:2 "/entrypoint.sh /etc/" 11 minutes ago Restarting (1) 4 minutes ago 192.168.1.50:4000->5000/tcp registry | 19:56 |
| kbaegis | docker log bae090cb7b99 | 19:56 |
| *** opennode has quit IRC | 19:57 | |
| kbaegis | I'd also check your routing | 19:57 |
| kbaegis | If your bridge has a default route out from the container | 19:57 |
| Pavo | docker: 'log' is not a docker command. | 19:57 |
| kbaegis | logs? | 19:57 |
| kbaegis | plural I think | 19:57 |
| kbaegis | try docker run -it ubuntu:latest bash and do a route print | 19:58 |
| Pavo | hmmm seems like its not seeing my certs | 19:58 |
| kbaegis | Oh | 19:59 |
| kbaegis | Well did you put them in ./certs/? | 19:59 |
| Pavo | time="2016-08-14T19:59:16Z" level=fatal msg="open /certs/domain.crt: no such file or directory" | 19:59 |
| Pavo | they are in root dir | 19:59 |
| kbaegis | From whereever you ran | 19:59 |
| kbaegis | docker run -d -p 192.168.1.50:4000:5000 --restart=always --name registry -v `pwd`/certs:/certs -e REGISTRY_HTTP_TLS_CERTIFICATE=/certs/domain.crt -e REGISTRY_HTTP_TLS_KEY=/certs/domain.key registry:2? | 19:59 |
| kbaegis | You put them in absolute? | 19:59 |
| kbaegis | pwd grabs your current path | 20:00 |
| Pavo | ah | 20:00 |
| kbaegis | docker run -d -p 192.168.1.50:4000:5000 --restart=always --name registry -v /certs:/certs -e REGISTRY_HTTP_TLS_CERTIFICATE=/certs/domain.crt -e REGISTRY_HTTP_TLS_KEY=/certs/domain.key registry:2? | 20:00 |
| Pavo | still not seeing them | 20:01 |
| Pavo | grrrr | 20:01 |
| kbaegis | Hmm | 20:02 |
| kbaegis | permissions? | 20:02 |
| Pavo | http://pastebin.com/EAtbK1Ai | 20:02 |
| Pavo | maybe | 20:02 |
| kbaegis | regenerate the self-signed cert | 20:02 |
| Pavo | ok | 20:02 |
| kbaegis | cd /&&mkdir certs&&openssl req -newkey rsa:8192 -nodes -sha512 -keyout certs/domain.key -x509 -days 3650 -out certs/domain.crt | 20:03 |
| Pavo | ok done | 20:05 |
| kbaegis | rerun the container :) | 20:06 |
| Pavo | ok think its working now | 20:06 |
| Pavo | but when I do curl 192.168.1.50:4000/v2/_ping nothing is returning | 20:07 |
| Pavo | ERROR:kolla.image.build.mariadb:Get https://192.168.1.50:4000/v1/_ping: x509: cannot validate certificate for 192.168.1.50 because it doesn't contain any IP SANs | 20:08 |
| *** daneyon_ has joined #openstack-kolla | 20:10 | |
| *** daneyon_ has quit IRC | 20:15 | |
| *** kbaegis has quit IRC | 20:21 | |
| *** kbaegis has joined #openstack-kolla | 20:22 | |
| *** Pavo has quit IRC | 20:24 | |
| *** tfukushima has joined #openstack-kolla | 20:27 | |
| *** tfukushima has quit IRC | 20:31 | |
| *** Jeffrey4l_ has joined #openstack-kolla | 20:34 | |
| *** Jeffrey4l has quit IRC | 20:35 | |
| *** dave-mcc_ has quit IRC | 20:50 | |
| kbaegis | Hey Pavo, still around? | 21:01 |
| kbaegis | Okay, next problem | 21:12 |
| kbaegis | Path /run/netns is mounted on /run but it is not a shared mount | 21:12 |
| kbaegis | neutron-dhcp-agent | 21:12 |
| kbaegis | well, tried chmod 777 /run/netns | 21:17 |
| kbaegis | Let's see if that absurdity fixes it | 21:18 |
| kbaegis | idempotent indeed | 21:18 |
| *** tfukushima has joined #openstack-kolla | 21:28 | |
| *** tfukushima has quit IRC | 21:32 | |
| kbaegis | http://pastebin.com/vN1VRgCX | 21:33 |
| kbaegis | Is that really correct? Keep getting "linux mounts: Path /run/netns is mounted on /run but it is not a shared mount" | 21:33 |
| rstarmer | has anoyone run into a collision between haproxy and mariadb on an all-in-one install running from master? | 21:40 |
| kbaegis | Yeah, I had issues | 21:42 |
| kbaegis | I'm running gentoo, though, which is the unsupportable red-headed stepchild | 21:42 |
| rstarmer | kbaegis: how did you resolve it? I think it's because I'm trying to deploy All-In-One but haproxy/keepalived are trying to run as well. | 21:59 |
| sbezverk | rstarmer: you can diasable keepalive in all-in-one | 21:59 |
| kbaegis | How? I was trying to figure that one out? | 22:00 |
| kbaegis | So for the keepalived/haproxy | 22:00 |
| kbaegis | It only looks for the VIP, which you can assign manually | 22:00 |
| kbaegis | My workaround: sudo ip addr add <CIDR addr> dev <IF>) | 22:01 |
| sbezverk | in globals.yml: enable_keepalive: "no" | 22:01 |
| kbaegis | tyvm | 22:03 |
| kbaegis | Couldn't find that anywhere | 22:03 |
| rstarmer | ah, I see what happened, my setup ansible didn't fill in the VIP addr... fixing (I think...) | 22:03 |
| kbaegis | ovs is still broken for me, and I can't figure out the shared /run/netns mount for dhcp agent | 22:04 |
| sbezverk | have you changed mount=shared for docker? | 22:04 |
| sbezverk | make sure this | 22:05 |
| sbezverk | mount --make-shared /run | 22:05 |
| sbezverk | command is persistent | 22:05 |
| kbaegis | Oh, nice | 22:07 |
| kbaegis | '/run/netns is not mountpoint or bad option' | 22:08 |
| sbezverk | kbaegis: correction, the actual command is enable_haproxy: "no" | 22:08 |
| kbaegis | sbezverk tyvm | 22:08 |
| kbaegis | So if /run is part of my root fs, how do I use mount to remount it? | 22:09 |
| sbezverk | in globals, it will disable both haproxy and keepalived since yyou do not really need then with aio installation | 22:09 |
| rstarmer | thanks, that I can add to by build ansible... | 22:09 |
| sbezverk | you need to run mount command exactl as I pasted | 22:10 |
| sbezverk | on the host (target) where you run kolla | 22:10 |
| kbaegis | Didn't return erros, tyvm :) | 22:10 |
| kbaegis | Let's see if this works | 22:10 |
| kbaegis | Disabled my manual vip assignment | 22:11 |
| kbaegis | disabled ovs | 22:11 |
| kbaegis | may actually deploy throughout now :) | 22:11 |
| kbaegis | *crosses fingers* | 22:16 |
| rstarmer | kbaegis: likewise... I think I've just re-set things back to a single interfaces (with bridge/veth pair for my neutron external env.) | 22:26 |
| rstarmer | hope to have a running system here shortly | 22:26 |
| kbaegis | day 3 for me | 22:26 |
| kbaegis | horizon is up | 22:26 |
| kbaegis | :D | 22:26 |
| kbaegis | 105 containers | 22:27 |
| kbaegis | Nice | 22:27 |
| rstarmer | neutron keeps failing for me | 22:27 |
| kbaegis | What's the error you're getting? | 22:28 |
| kbaegis | I had to recompile my kernel with netfilter options | 22:28 |
| rstarmer | fatal: [localhost -> localhost]: FAILED! => {"changed": true, "failed": true, "msg": "Container exited with non-zero return code"} | 22:28 |
| kbaegis | and docker logs? | 22:28 |
| rstarmer | there's no container to get logs from | 22:28 |
| kbaegis | docker ps -a | 22:28 |
| kbaegis | There should be one that's just "created" or "restarting" | 22:29 |
| kbaegis | YAY!!! First Kolla build | 22:29 |
| kbaegis | :) | 22:29 |
| kbaegis | Albeit with 4 critical workarounds over 3 days | 22:29 |
| rstarmer | ImportError: No module named fixtures | 22:29 |
| kbaegis | Yeah, I had that issue too | 22:29 |
| kbaegis | I think master is broken | 22:29 |
| kbaegis | Here's my fix | 22:29 |
| kbaegis | neutron_plugin_agent: "linuxbridge" | 22:30 |
| kbaegis | I also went from centos-source to centos-binary, and rolled back to release 2.0.1 | 22:30 |
| rstarmer | I've got the agent for linux bridge already | 22:30 |
| kbaegis | I'm going to go check on my beautifully provisioned horizon setup and I'll be back | 22:30 |
| rstarmer | will go try 2.0.1 | 22:30 |
| kbaegis | Did you set release tag to 2.0.1? | 22:31 |
| rstarmer | thanks, | 22:31 |
| kbaegis | sure | 22:31 |
| rstarmer | no, I'll do that now | 22:31 |
| kbaegis | hope that helps :) | 22:31 |
| rstarmer | me too! | 22:31 |
| kbaegis | I need to dig through the git logs next and send a email/bill for whoever wrote the regression | 22:31 |
| rstarmer | ha! :D | 22:32 |
| kbaegis | Actually, totally my fault for syncing from master | 22:32 |
| kbaegis | But still :) | 22:32 |
| kbaegis | Great learning opportunity | 22:32 |
| kbaegis | Horizon is listening on port 80, right? | 22:35 |
| rstarmer | well, I tried building stable/mitaka with ansible 2.0... also a bad idea :P | 22:35 |
| rstarmer | horizon should be on 80 | 22:35 |
| rstarmer | should redirect to /dashboard/ | 22:35 |
| kbaegis | beautiful | 22:35 |
| kbaegis | Have to set up a new subdomain, but whatever :) | 22:40 |
| rstarmer | for my next trick (once I get this process stabilized), I'm going to try to do a multi-machine config, will see what fun that creates. Anyone know how to tell neutron to use a smaller MTU with this model? Is there somewhere that I can pass the MTU parameter to the container? | 22:40 |
| kbaegis | Forgot, horizon doesn't implement an easy relative path that I'm aware of | 22:40 |
| kbaegis | I think openvswitch is the appropriate place for that. Why change MTU/MSS? | 22:40 |
| rstarmer | looking at my env. it looks like the've dropped the /dashboard path anyway | 22:40 |
| *** Pavo has joined #openstack-kolla | 22:41 | |
| *** dwalsh has joined #openstack-kolla | 22:41 | |
| rstarmer | Good question. I was thinking I'd have to tell neutron to use a smaller MTU, but if I do tenant networks over VXLAN, I think that's already baked in. My "external" network is unfortunately only a bridge on the network node(s) anyway. | 22:42 |
| *** dwalsh has quit IRC | 22:49 | |
| kbaegis | Hmm | 22:50 |
| kbaegis | So I'm having issues with horizon. I don't have access to that subnet (by design). Anyone experienced with hiding horizon behind apache? | 22:50 |
| kbaegis | I need apache to work as a proxy | 22:50 |
| kbaegis | kolla is completely deployed :) | 22:51 |
| rstarmer | have you tried an ssh tunnel just to verify? | 22:51 |
| rstarmer | ssh user@host -L 8080:localhost:80 | 22:51 |
| rstarmer | ^ sort of thing? | 22:51 |
| kbaegis | Yeah, I can curl it | 22:51 |
| kbaegis | It's not routable, nor is it supposed to be | 22:51 |
| kbaegis | mod_rewrite will be a last resort | 22:52 |
| kbaegis | But should work | 22:52 |
| rstarmer | I thought it was just mod_proxy that you could configure | 22:52 |
| rstarmer | in a container... | 22:52 |
| rstarmer | :D | 22:53 |
| rstarmer | same error with 2.0.1 images | 22:53 |
| rstarmer | ImportError: No module named fixtures | 22:53 |
| kbaegis | Weird. I'm not getting that anymore after disabling | 22:54 |
| kbaegis | Hmm | 22:54 |
| kbaegis | I can't remember the specific workaround, but I definitely remember running into that one | 22:54 |
| kbaegis | I thought I just reverted | 22:55 |
| kbaegis | Here's my functional config | 22:57 |
| kbaegis | http://pastebin.com/raqnSdmB | 22:57 |
| kbaegis | Isn't mod_proxy supposed to alter the return addresses as well? | 23:01 |
| rstarmer | hmm, it's been a long time since I actually configured mod_proxy... | 23:01 |
| rstarmer | thanks for the config though, I'm going to reset to stable/mitaka ansible 1.9.4, etc. and will re-try with 2.0.1 | 23:02 |
| Pavo | man getting this private registry to work is a pain in the ass | 23:08 |
| kbaegis | Yes it is | 23:08 |
| kbaegis | Sorry, you dc'ed before I could help | 23:08 |
| Pavo | no problem | 23:08 |
| kbaegis | So there's a way to update your os'es certs | 23:09 |
| Pavo | the tls way is giving me nothing but 509 errors | 23:09 |
| kbaegis | update-ca-certificates or somesuch | 23:09 |
| kbaegis | Yeah, if you can't curl https://<IP>:<PORT> without the -k flag, it won't work | 23:09 |
| kbaegis | https://docs.docker.com/v1.7/registry/authentication/ | 23:10 |
| Pavo | this is what I don't understand | 23:11 |
| Pavo | for the CN myregistrydomain.com | 23:11 |
| Pavo | where to set that? | 23:11 |
| kbaegis | "cp auth/domain.crt /etc/pki/ca-trust/source/anchors/myregistrydomain.com.crt | 23:11 |
| kbaegis | update-ca-trust" | 23:11 |
| kbaegis | Oh | 23:11 |
| Pavo | I just used my IP | 23:11 |
| kbaegis | That's the resolvable | 23:11 |
| kbaegis | IP won't work | 23:11 |
| kbaegis | You can use hostname and /etc/hosts file | 23:11 |
| kbaegis | Or you can set up DNS | 23:11 |
| Pavo | well I have everything in hosts file | 23:11 |
| kbaegis | because the certificate encodes the fqdn or hostname in the cn field | 23:11 |
| kbaegis | Basically, it checks that "bob.com" belongs to bob | 23:12 |
| Pavo | so I should use something like deployer.localhost? | 23:12 |
| kbaegis | sure | 23:12 |
| kbaegis | That'll work | 23:12 |
| Pavo | ok I'll try that after this push gets done | 23:12 |
| kbaegis | registry.local | 23:12 |
| kbaegis | Sure | 23:12 |
| kbaegis | You have to write that into cn when you gen your cert | 23:12 |
| Pavo | ok hang on | 23:13 |
| kbaegis | It'll ask for common name, fqdn or somethign | 23:13 |
| Pavo | so like this | 23:14 |
| Pavo | http://pastebin.com/BSg8iMks | 23:14 |
| Pavo | and my /etc/hosts file looks like this | 23:15 |
| Pavo | http://pastebin.com/4hHF64Pa | 23:15 |
| Pavo | lol ok now getting new error | 23:20 |
| Pavo | 0cc151c648af | 23:20 |
| Pavo | oops | 23:20 |
| Pavo | ERROR:kolla.image.build.elasticsearch:Get https://deploy.local:4000/v1/_ping: x509: certificate signed by unknown authority | 23:20 |
| Pavo | why does it keep trying to use v1 of registry | 23:21 |
| Pavo | I think thats the main issue | 23:21 |
| rstarmer | I got that working.https://github.com/kumulustech/kolla-multinode/blob/master/deploy.yml#L73-L124 | 23:25 |
| kbaegis | Pavo, I think it's just that it's an "unknown authority", i.e. self-signed cert | 23:26 |
| kbaegis | Please read this: https://docs.docker.com/v1.7/registry/authentication/ | 23:26 |
| rstarmer | https://github.com/kumulustech/kolla-multinode/blob/master/deploy.yml#L149-L152 | 23:26 |
| kbaegis | the command you need is: cp auth/domain.crt /etc/pki/ca-trust/source/anchors/myregistrydomain.com.crt ; update-ca-trust | 23:26 |
| rstarmer | basically, if you set the cert up right, and tell docker to use it, you don't have to deal with the insecure registry. | 23:26 |
| rstarmer | that seems to work for me as I can build my images and push them | 23:27 |
| *** jtriley has joined #openstack-kolla | 23:27 | |
| Pavo | kbaegis so I would use cp certs/domain.crt /etc/pki/ca-trust/source/anchors/domain.crt ; update-ca-trust | 23:28 |
| kbaegis | Yeah | 23:29 |
| kbaegis | Or | 23:29 |
| kbaegis | Just the crt | 23:29 |
| kbaegis | Yes | 23:29 |
| Pavo | ok I did cp /certs/domain.crt /etc/pki/ca-trust/source/anchors/deploy.local.crt ; update-ca-trust | 23:29 |
| kbaegis | Sweet | 23:29 |
| Pavo | same error | 23:29 |
| kbaegis | now do a curl https://<ip>:<port>/v2/_ping | 23:30 |
| kbaegis | Really? | 23:30 |
| kbaegis | Hmm | 23:30 |
| kbaegis | And it's centos? | 23:30 |
| kbaegis | the directions I gave you were actually for RHEL | 23:30 |
| kbaegis | I'll be annoyed if they differ | 23:30 |
| rstarmer | try curl https://ip:port/v2/_containers | 23:30 |
| rstarmer | I don't think there's a _ping for v2 | 23:30 |
| kbaegis | You're right | 23:30 |
| kbaegis | what he shouldn't get either way is an ssl rejection | 23:30 |
| kbaegis | w/o -k | 23:30 |
| rstarmer | true | 23:30 |
| rstarmer | did you restart docker daemon after fiddling with cets? | 23:31 |
| rstarmer | certs | 23:31 |
| Pavo | [root@deploy ~]# curl https://deploy.local:4000/v2/containers | 23:31 |
| Pavo | 404 page not found | 23:31 |
| rstarmer | _containers | 23:32 |
| Pavo | no but I will right nos | 23:32 |
| Pavo | now | 23:32 |
| rstarmer | _ is important | 23:32 |
| rstarmer | But even the 404 is better, you are not getting a cert error | 23:32 |
| rstarmer | but dockerd needs to pick that up as well | 23:32 |
| Pavo | [root@deploy ~]# curl https://deploy.local:4000/v2/_containers | 23:32 |
| Pavo | 404 page not found | 23:32 |
| Pavo | e5dfca0c25c9 registry:2 "/entrypoint.sh /etc/" 9 minutes ago Up 27 seconds 0.0.0.0:4000->5000/tcp registry | 23:33 |
| rstarmer | I used to get '{[]}' as a response | 23:34 |
| rstarmer | anything in the log? | 23:34 |
| rstarmer | does it complain about certs/TLS issues? | 23:35 |
| Pavo | time="2016-08-14T23:34:48Z" level=warning msg="No HTTP secret provided - generated random secret. This may cause problems with uploads if multiple registries are behind a load-balancer. To provide a shared secret, fill in http.secret in the configuration file or set the REGISTRY_HTTP_SECRET environment variable." go.version=go1.6.3 instance.id=04674853-2f7e-4852-bd13-e93ac029d8cf version=v2.5.0 | 23:36 |
| Pavo | 192.168.1.50 - - [14/Aug/2016:23:35:04 +0000] "GET /v2/_containers HTTP/1.1" 404 19 "" "curl/7.29.0" | 23:36 |
| Pavo | curl still giving me 404 | 23:36 |
| Pavo | guess I'm gonna start all over | 23:37 |
| Pavo | good thing I took a snapshot before I started lol | 23:37 |
| rstarmer | did you try pushing something to the repo: | 23:38 |
| rstarmer | Pull (or build) some image from the hub | 23:38 |
| rstarmer | docker pull ubuntu | 23:38 |
| rstarmer | Tag the image so that it points to your registry | 23:38 |
| rstarmer | docker tag ubuntu localhost:4000/myfirstimage | 23:38 |
| rstarmer | Push it | 23:38 |
| rstarmer | docker push localhost:4000/myfirstimage | 23:38 |
| rstarmer | that's paraphrased from the docker instructions | 23:39 |
| rstarmer | change the localhost:4000 to https://repo:4000/ | 23:39 |
| rstarmer | see if it lets you push/pull | 23:39 |
| rstarmer | though it really seems like it should have just been happy | 23:39 |
| Pavo | no I guess I should have tried that before reverting lol | 23:39 |
| Pavo | I'll try that if I have an issue again | 23:39 |
| rstarmer | also, you can also launch the repo with a compose file, I felt that was more repeatable than the comand line (where I tended to forget something)... | 23:40 |
| kbaegis | its _catalog | 23:40 |
| kbaegis | Not _containers | 23:40 |
| rstarmer | and then there's that... | 23:41 |
| rstarmer | my compose file (which I called a template, even though there's nothing templated): https://github.com/kumulustech/kolla-multinode/blob/master/templates/registry_yml.j2 | 23:41 |
| rstarmer | note the cert directory mapping, which may be different for you | 23:41 |
| rstarmer | and I set a "secret" just so I wouldn't get the complaint about not having set one.... | 23:42 |
| Pavo | nice templates | 23:45 |
| Pavo | have you tried using pci passthrough or sr-iov with kolla yet? | 23:46 |
| rstarmer | haven't, though there is no reason that it _shouldn't_ work. Maybe my @Cisco friends will give me access to a system again where I could try something like that :D | 23:47 |
| *** zhurong has joined #openstack-kolla | 23:51 | |
| *** jtriley has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!