| *** salv-orlando has quit IRC | 00:07 | |
| *** salv-orlando has joined #openstack-kolla | 00:08 | |
| *** yangyape_ has quit IRC | 00:17 | |
| *** yangyapeng has joined #openstack-kolla | 00:18 | |
| *** lukepatrick has joined #openstack-kolla | 00:18 | |
| *** yangyapeng has quit IRC | 00:22 | |
| *** Pavo has quit IRC | 00:31 | |
| *** hieulq has joined #openstack-kolla | 00:43 | |
| *** huanxie has joined #openstack-kolla | 00:54 | |
| *** nguyentrihai has joined #openstack-kolla | 01:04 | |
| *** lukepatrick has quit IRC | 01:06 | |
| *** salv-orl_ has joined #openstack-kolla | 01:07 | |
| *** salv-orlando has quit IRC | 01:10 | |
| *** kevzha01 has joined #openstack-kolla | 01:14 | |
| *** kevzha01 is now known as kevinz | 01:14 | |
| *** jinke_ has joined #openstack-kolla | 01:14 | |
| *** genek has quit IRC | 01:16 | |
| *** yangyapeng has joined #openstack-kolla | 01:20 | |
| *** tovin07_ has joined #openstack-kolla | 01:27 | |
| *** zhubingbing_ has joined #openstack-kolla | 01:44 | |
| *** caowei has joined #openstack-kolla | 02:02 | |
| *** zhangfei has joined #openstack-kolla | 02:07 | |
| *** kiennt26 has joined #openstack-kolla | 02:08 | |
| *** xinliang has quit IRC | 02:23 | |
| *** salv-orl_ has quit IRC | 02:25 | |
| *** salv-orlando has joined #openstack-kolla | 02:25 | |
| *** xinliang has joined #openstack-kolla | 02:35 | |
| *** xinliang has joined #openstack-kolla | 02:35 | |
| *** salv-orlando has quit IRC | 02:43 | |
| *** ArminderSingh- has quit IRC | 02:44 | |
| *** salv-orlando has joined #openstack-kolla | 02:44 | |
| *** masber has quit IRC | 02:47 | |
| *** ArminderSingh has joined #openstack-kolla | 02:48 | |
| *** duonghq has joined #openstack-kolla | 02:52 | |
| *** dave-mccowan has quit IRC | 03:05 | |
| *** hieulq has quit IRC | 03:23 | |
| *** hieulq has joined #openstack-kolla | 03:23 | |
| *** masber has joined #openstack-kolla | 03:33 | |
| masber | good afternoon, I am having this error while deploying kolla-ansible. fatal: [test-openstack-controller]: FAILED! => {"changed": false, "elapsed": 301, "failed": true, "msg": "Timeout when waiting for 192.168.20.5:3306"} | 03:34 |
|---|---|---|
| masber | any thoughts? | 03:34 |
| *** gkadam has joined #openstack-kolla | 03:35 | |
| *** gkadam_ has joined #openstack-kolla | 03:40 | |
| *** gkadam has quit IRC | 03:41 | |
| *** lukepatrick has joined #openstack-kolla | 03:56 | |
| *** mdnadeem has joined #openstack-kolla | 03:59 | |
| *** janki has joined #openstack-kolla | 04:22 | |
| kolla-slack | <lxdong.007> How to find macros.install_packages() , we need to customize Horizon. | 04:28 |
| *** skramaja has joined #openstack-kolla | 04:39 | |
| openstackgerrit | Heetae Ahn proposed openstack/kolla master: Add RHEL subscription registration https://review.openstack.org/495148 | 04:41 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: implement build jobs in kolla https://review.openstack.org/513971 | 04:58 |
| *** jaosorior has joined #openstack-kolla | 05:02 | |
| *** lukepatrick has quit IRC | 05:03 | |
| *** pratapagoutham has quit IRC | 05:07 | |
| *** pratapagoutham has joined #openstack-kolla | 05:24 | |
| pratapagoutham | Hi all this is with reference to this link to deploy openstack-kolla multinode "https://docs.openstack.org/project-deploy-guide/kolla-ansible/ocata/multinode.html#deploy-a-registry" | 05:26 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: enable bifrost in oraclelinux+source https://review.openstack.org/506945 | 05:28 |
| pratapagoutham | should we also do pip install kolla after doing pip install kolla-ansible ?? | 05:29 |
| *** pratapagoutham has left #openstack-kolla | 05:33 | |
| *** pratapagoutham has joined #openstack-kolla | 05:33 | |
| kolla-slack | <lxdong.007> How to customize source of horizon image building?Like we have a customized directory override the original. | 05:36 |
| *** yangyapeng has quit IRC | 05:36 | |
| *** yangyapeng has joined #openstack-kolla | 05:36 | |
| pratapagoutham | hi all when i execute tools/start-registry why am i getting this error "Could not reach any registry endpoint" "?? any help ? | 05:40 |
| *** yangyapeng has quit IRC | 05:41 | |
| *** danardelean has joined #openstack-kolla | 05:57 | |
| *** danardelean has quit IRC | 05:57 | |
| *** dardelean has joined #openstack-kolla | 06:16 | |
| *** danardelean has joined #openstack-kolla | 06:16 | |
| *** danardelean has quit IRC | 06:21 | |
| pratapagoutham | Hi all i have a doubt is there way where we can specify the services required to be installed? | 06:21 |
| *** magicboiz has joined #openstack-kolla | 06:21 | |
| pratapagoutham | for eg: say i dont want all the services i just want only nova, neutron, glance and ceilometer to be installed can i do it if yes how to do that??? | 06:22 |
| pratapagoutham | any help ? | 06:22 |
| pratapagoutham | thanks in advance :) | 06:23 |
| *** cah_link has joined #openstack-kolla | 06:25 | |
| *** magicboiz has quit IRC | 06:26 | |
| *** cah_link has quit IRC | 06:29 | |
| *** magicboiz has joined #openstack-kolla | 06:35 | |
| *** athomas has joined #openstack-kolla | 06:41 | |
| *** hrw has quit IRC | 06:44 | |
| *** tvignaud has joined #openstack-kolla | 06:45 | |
| *** cah_link has joined #openstack-kolla | 06:47 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: implement build jobs in kolla https://review.openstack.org/513971 | 06:47 |
| *** danardelean has joined #openstack-kolla | 06:51 | |
| *** hrw has joined #openstack-kolla | 06:59 | |
| *** pcaruana has joined #openstack-kolla | 07:00 | |
| *** serlex has joined #openstack-kolla | 07:02 | |
| *** jmccarthy has joined #openstack-kolla | 07:18 | |
| *** zhangfei has quit IRC | 07:18 | |
| masber | pratapagoutham, yes you have to specify the services you want to install in the globals.yml file | 07:26 |
| *** jistr is now known as jistr|mtgs | 07:26 | |
| bjolo | morning | 07:28 |
| masber | pratapagoutham, use this document for kolla-ansible deployment on multiple nodes https://docs.openstack.org/kolla-ansible/latest/user/multinode.html | 07:28 |
| masber | pratapagoutham, you can also use this document as a reference https://docs.openstack.org/kolla-ansible/latest/reference/ceph-guide.html#simple-3-node-example | 07:29 |
| *** zhangfei has joined #openstack-kolla | 07:32 | |
| openstackgerrit | yanpeifei proposed openstack/kolla master: update "mitaka" to "ocata" https://review.openstack.org/511670 | 07:35 |
| openstackgerrit | yanpeifei proposed openstack/kolla master: update "opendaylight" tar "distribution-karaf-0.6.2-Carbon.tar.gz" has the latest features. https://review.openstack.org/514175 | 07:35 |
| *** salv-orlando has quit IRC | 07:50 | |
| *** salv-orlando has joined #openstack-kolla | 07:50 | |
| openstackgerrit | yanpeifei proposed openstack/kolla master: fix "openvswitch-base-plugin-dpdk" url. https://review.openstack.org/514178 | 07:51 |
| hrw | morning | 07:53 |
| *** dougsz has joined #openstack-kolla | 07:54 | |
| *** salv-orl_ has joined #openstack-kolla | 07:55 | |
| *** salv-orlando has quit IRC | 07:58 | |
| kolla-slack | <aslancimbom100> Morning | 08:02 |
| spsurya | sup all | 08:04 |
| spsurya | duonghq: sup | 08:04 |
| spsurya | morning hrw | 08:04 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Optimize zuulv2 jobs https://review.openstack.org/513612 | 08:11 |
| *** Zophar has joined #openstack-kolla | 08:14 | |
| Zophar | Good morning | 08:15 |
| Zophar | Could we go ahead with this one: https://review.openstack.org/#/c/509792/2 ? | 08:15 |
| Zophar | I want to push a new feature on Horizon for image sources... but i need this one to be push in master and stable first | 08:16 |
| hrw | Zophar: make a patch and use "Depends-on: CHANGEID" in commit | 08:18 |
| Zophar | Hum.... the point is that my actual patch https://review.openstack.org/#/c/509784/ will need a rebase after this one.... | 08:19 |
| Zophar | I'll will update it once 509792 is merged | 08:22 |
| duonghq | spsurya, hi | 08:22 |
| pratapagoutham | thanq masber :) | 08:27 |
| magicboiz | morning | 08:42 |
| magicboiz | anyone can check this? https://bugs.launchpad.net/kolla-ansible/+bug/1723155 | 08:42 |
| openstack | Launchpad bug 1723155 in kolla-ansible "No valid host was found error" [Undecided,Incomplete] | 08:42 |
| openstackgerrit | Merged openstack/kolla master: Update Image Builder Links in Readme https://review.openstack.org/512702 | 08:47 |
| openstackgerrit | Merged openstack/kolla master: Add EC2-API to Kolla providing images list https://review.openstack.org/510466 | 08:47 |
| magicboiz | is there any difference on Ceph version installed with kolla-ansible/Pike from Centos to Ubuntu/debian?? | 09:01 |
| *** sambetts|afk is now known as sambetts | 09:15 | |
| *** jiriprox has joined #openstack-kolla | 09:17 | |
| flaper87 | hey guys, have you started migrating kolla CI to zuul? I checked last week and it didn't seem like it but I wanted to double check if you had. | 09:18 |
| flaper87 | I'd like to take a look and see if there are some things we can share :) | 09:19 |
| flaper87 | kfox1111: inc0 ^ | 09:19 |
| openstackgerrit | zhongshengping proposed openstack/kolla-ansible stable/pike: Add ipc_mode=host for nova-compute conainer https://review.openstack.org/507028 | 09:21 |
| *** janki has quit IRC | 09:21 | |
| openstackgerrit | zhongshengping proposed openstack/kolla-ansible stable/ocata: Add ipc_mode=host for nova-compute conainer https://review.openstack.org/507029 | 09:21 |
| *** zhangfei has quit IRC | 09:25 | |
| pratapagoutham | Hi while running the prechecks im getting this error https://hastebin.com/owaxorucir.tex can anyone guide me ?? thanks in advance | 09:35 |
| *** salv-orl_ has quit IRC | 09:36 | |
| *** salv-orlando has joined #openstack-kolla | 09:37 | |
| pbourke | flaper87: pretty sure its done, change merged late last week | 09:37 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: implement build jobs in kolla https://review.openstack.org/513971 | 09:37 |
| *** zhangfei has joined #openstack-kolla | 09:38 | |
| pbourke | pratapagoutham: please cross check your ansible version with https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html#install-dependencies | 09:38 |
| pratapagoutham | my version is 2.0.0 but still it is throwing up this error | 09:41 |
| *** salv-orlando has quit IRC | 09:41 | |
| flaper87 | kolla-ansible or kolla-kubernetes? | 09:46 |
| flaper87 | pbourke: ^ | 09:46 |
| pbourke | flaper87: kolla-ansible, unsure about kolla-k8s | 09:46 |
| flaper87 | gotcha, thanks | 09:49 |
| *** zhangfei has quit IRC | 09:51 | |
| openstackgerrit | zenghui.shi proposed openstack/kolla master: Add ptp service docker image https://review.openstack.org/513703 | 09:52 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla master: Add lbaas dashboard in horizon ubuntu binary https://review.openstack.org/508191 | 09:58 |
| *** ioni has joined #openstack-kolla | 09:59 | |
| openstackgerrit | yanpeifei proposed openstack/kolla master: update "copy_links:true" ARG. synchronize options of the "copy_link" required "yes". https://review.openstack.org/514202 | 10:00 |
| ioni | hello guys | 10:00 |
| ioni | i want to install pike and i have an error | 10:00 |
| ioni | https://paste.xinu.at/s2p/ | 10:00 |
| ioni | manifest for kolla/ubuntu-binary-fluentd:5.0.0 not found | 10:00 |
| ioni | i do see that the tag for fluentd doesn't exists upstream | 10:00 |
| ioni | i'm not sure how to resolve this problem | 10:01 |
| ioni | https://hub.docker.com/r/kolla/ubuntu-binary-fluentd/tags/ | 10:01 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla master: nova-libvirt: fix kvm permission issue https://review.openstack.org/501653 | 10:01 |
| *** paramite has joined #openstack-kolla | 10:04 | |
| *** kiennt26 has quit IRC | 10:05 | |
| *** jinke_ has quit IRC | 10:08 | |
| *** zhangfei has joined #openstack-kolla | 10:08 | |
| *** jinke_ has joined #openstack-kolla | 10:08 | |
| *** magicboiz has quit IRC | 10:10 | |
| *** calbers has joined #openstack-kolla | 10:11 | |
| hrw | ioni: I do not know did they got uploaded at all (:5.0.0 images) | 10:12 |
| *** tovin07_ has quit IRC | 10:12 | |
| ioni | hrw, right. what's the best way to deploy 5.0 using kolla? | 10:13 |
| hrw | ioni: spend 2-3h and build them locally? | 10:13 |
| hrw | ioni: with local registry deploy goes quick | 10:14 |
| *** magicboiz has joined #openstack-kolla | 10:15 | |
| openstackgerrit | James McCarthy proposed openstack/kolla master: Create murano applications package for import https://review.openstack.org/513632 | 10:17 |
| *** magicboiz has quit IRC | 10:19 | |
| *** caowei has quit IRC | 10:20 | |
| *** magicboiz has joined #openstack-kolla | 10:26 | |
| ioni | hrw, ok, thanks | 10:29 |
| *** pbourke has quit IRC | 10:37 | |
| *** pbourke has joined #openstack-kolla | 10:38 | |
| *** yangyapeng has joined #openstack-kolla | 10:38 | |
| *** janki has joined #openstack-kolla | 10:40 | |
| *** yangyapeng has quit IRC | 10:43 | |
| *** yangyapeng has joined #openstack-kolla | 10:44 | |
| *** zhubingbing has joined #openstack-kolla | 10:51 | |
| *** zhubingbing_ has quit IRC | 10:51 | |
| *** zhubingbing has quit IRC | 10:58 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-kubernetes master: Updated from global requirements https://review.openstack.org/509443 | 11:01 |
| *** sean-k-mooney has joined #openstack-kolla | 11:03 | |
| *** zhangfei has quit IRC | 11:03 | |
| *** dave-mccowan has joined #openstack-kolla | 11:04 | |
| pratapagoutham | Hi all can anyone help me with this please?? | 11:34 |
| pratapagoutham | https://askubuntu.com/questions/967990/upgrade-docker-installed-via-pip | 11:34 |
| pratapagoutham | please | 11:34 |
| pratapagoutham | onlything left out in my installation | 11:34 |
| *** jiriprox has quit IRC | 11:38 | |
| *** jiriprox has joined #openstack-kolla | 11:38 | |
| *** lpetrut has joined #openstack-kolla | 11:39 | |
| *** erlon has joined #openstack-kolla | 11:42 | |
| hrw | pratapagoutham: check (with strace) where from it loaded docker | 11:43 |
| hrw | you may have python-docker installed systemwide while pip installed into your home directory. | 11:44 |
| *** janki has quit IRC | 11:44 | |
| hrw | python checks system dirs and then user one | 11:44 |
| hrw | import sys | 11:45 |
| hrw | print sys.path | 11:45 |
| hrw | and then check | 11:45 |
| pratapagoutham | yes strangly why is it taking this ??? " /home/kolla/.local/lib/python2.7/site-packages/docker/__init__.py" | 11:46 |
| pratapagoutham | should i delete that docker directory in that location hrw ?? | 11:47 |
| hrw | pratapagoutham: if you do not understand how it happened then I suggest you to use python-virtualenv instead of randomly playing with pip | 11:47 |
| pratapagoutham | ok but can i delete that directory in that location?? | 11:47 |
| hrw | pratapagoutham: you used pip as a user. so it stored in user homedir | 11:47 |
| pratapagoutham | oh ok | 11:47 |
| hrw | remove /home/kolla/.local/lib/python2.7/ and use virtualenv | 11:48 |
| hrw | that way you will have separate python installation in which you have permissions | 11:48 |
| pratapagoutham | hrw: thanks it worked :) thankyou | 11:48 |
| hrw | yw | 11:49 |
| pratapagoutham | yes yes i will do it from now on :) | 11:49 |
| hrw | I was in your situation few months ago ;D | 11:49 |
| pratapagoutham | haha :D | 11:53 |
| pratapagoutham | i now get it | 11:53 |
| openstackgerrit | Chen proposed openstack/kolla master: Update links in README.rst https://review.openstack.org/510854 | 12:01 |
| *** jmccarthy has left #openstack-kolla | 12:03 | |
| *** salv-orlando has joined #openstack-kolla | 12:04 | |
| *** huanxie has quit IRC | 12:16 | |
| rwellum | hrw: ioni we have the tarballs correct? That's what I've been using for Pike/Master | 12:17 |
| ioni | rwellum, i didn't use source, just binary | 12:18 |
| ioni | i kinda failed to use local registry, i have them all built but cannot be used by ansible | 12:18 |
| rwellum | I think these are binary - just tarballed up. | 12:19 |
| ioni | rwellum, i don't know then, all that failed(binary) don't have tags | 12:19 |
| rwellum | http://tarballs.openstack.org/kolla/images/ | 12:19 |
| ioni | on hub docker | 12:19 |
| ioni | maybe i failed to read the documentation somehow | 12:20 |
| rwellum | Right that's correct, images post ocata are not on docker hub until infra steps are done | 12:20 |
| rwellum | tarballs an alternative - not sure if kolla-ansible folks are using | 12:21 |
| ioni | rwellum, ok, until know i used openstack ansible | 12:21 |
| ioni | but i wanted something fast and tried kolla | 12:21 |
| *** gnupyx has joined #openstack-kolla | 12:32 | |
| hrw | rwellum: do not know. I usually use aarch64 images so have to build them anyway | 12:36 |
| *** pcaruana has quit IRC | 12:41 | |
| pratapagoutham | Hi all i have been trying to deploy multi-node openstack-kolla in the precheck it shows this error https://hastebin.com/yuhimudapo.php any help?? | 12:43 |
| *** gkadam_ has quit IRC | 12:44 | |
| pratapagoutham | hrw: any idea on this?? | 12:44 |
| openstackgerrit | Merged openstack/kolla master: Add lbaas dashboard in horizon ubuntu binary https://review.openstack.org/508191 | 12:46 |
| openstackgerrit | Merged openstack/kolla stable/pike: Bump OpenStack service version https://review.openstack.org/513605 | 12:47 |
| *** huanxie has joined #openstack-kolla | 12:56 | |
| *** pcaruana has joined #openstack-kolla | 12:57 | |
| *** ktibi has joined #openstack-kolla | 12:58 | |
| *** pcaruana has quit IRC | 13:01 | |
| *** jtriley has quit IRC | 13:07 | |
| kolla-slack | <aslancimbom100> pratapagoutham what is running on the IP/port which is currently inuse? | 13:12 |
| kolla-slack | <aslancimbom100> if you are running local registry, then re-create it with another port, ie 5443 | 13:13 |
| pratapagoutham | kolla-slack: https://hastebin.com/ocanucezav.vbs | 13:13 |
| *** salv-orl_ has joined #openstack-kolla | 13:15 | |
| pratapagoutham | kolla-slack: i changed that to 4000 | 13:15 |
| pratapagoutham | how to check on what port my regsitry is running?? | 13:16 |
| *** salv-orlando has quit IRC | 13:16 | |
| *** salv-orl_ has quit IRC | 13:17 | |
| *** salv-orlando has joined #openstack-kolla | 13:18 | |
| pratapagoutham | kolla-slack: https://hastebin.com/jadebizipu.nginx | 13:19 |
| pratapagoutham | i think this is a bit clear . | 13:19 |
| *** jtriley has joined #openstack-kolla | 13:21 | |
| *** xinliang has quit IRC | 13:23 | |
| *** jtriley has quit IRC | 13:26 | |
| *** jinke_ has quit IRC | 13:27 | |
| *** huanxie has quit IRC | 13:27 | |
| *** goldyfruit has joined #openstack-kolla | 13:32 | |
| kolla-slack | <aslancimbom100> You need to change the registry container to run on different port, do you not have the command in your history? | 13:33 |
| kolla-slack | <aslancimbom100> You have changed the client to accept registry running on port 4000 | 13:34 |
| *** cah_link has quit IRC | 13:37 | |
| pratapagoutham | kolla-slack: thanks it worked | 13:37 |
| pratapagoutham | while deploying im getting this error " https://hastebin.com/vefijivahi.tex " | 13:37 |
| *** mdnadeem has quit IRC | 13:39 | |
| kolla-slack | <aslancimbom100> ok so that fails because you haven't changed the registry container port | 13:41 |
| kolla-slack | <aslancimbom100> do "docker ps -a" is there a registry container? | 13:41 |
| pratapagoutham | https://hastebin.com/ehevuyocos.rb | 13:42 |
| pratapagoutham | yes i redirected it to 5443 | 13:42 |
| pratapagoutham | donno why 5000,tcp is still there .. :( | 13:43 |
| *** salv-orl_ has joined #openstack-kolla | 13:44 | |
| *** salv-orlando has quit IRC | 13:45 | |
| *** jtriley has joined #openstack-kolla | 13:46 | |
| *** jiriprox has quit IRC | 13:47 | |
| *** jiriprox has joined #openstack-kolla | 13:47 | |
| *** yangyapeng has quit IRC | 13:50 | |
| pratapagoutham | https://hastebin.com/uvoqoseloq.pas any help why im getting this error ??? | 13:50 |
| *** yangyapeng has joined #openstack-kolla | 13:51 | |
| kolla-slack | <aslancimbom100> what is the output of docker ps -a | 13:51 |
| pratapagoutham | https://hastebin.com/exuyazowem.nginx | 13:52 |
| pratapagoutham | kolla-slack: the above one is the output | 13:52 |
| pratapagoutham | kolla-build --base ubuntu --type source --registry 172.16.69.71:5443 --push I tried this and i got error like this "https://hastebin.com/uvoqoseloq.pas" | 13:55 |
| pratapagoutham | any help on how to do ??? :'( please | 13:55 |
| *** yangyapeng has quit IRC | 13:55 | |
| *** Zophar has quit IRC | 13:55 | |
| *** skramaja has quit IRC | 13:56 | |
| *** hongbin has joined #openstack-kolla | 13:56 | |
| *** hrw has quit IRC | 13:59 | |
| *** hrw has joined #openstack-kolla | 14:01 | |
| *** Zophar has joined #openstack-kolla | 14:04 | |
| *** jaosorior has quit IRC | 14:08 | |
| *** hachi_ has joined #openstack-kolla | 14:18 | |
| *** salv-orl_ has quit IRC | 14:20 | |
| *** salv-orlando has joined #openstack-kolla | 14:20 | |
| *** david-lyle has joined #openstack-kolla | 14:21 | |
| kolla-slack | <aslancimbom100> hey, can you show me the command you use to create the registry | 14:31 |
| *** zhangfei has joined #openstack-kolla | 14:34 | |
| pratapagoutham | https://hastebin.com/nifezekizu.bash | 14:36 |
| pratapagoutham | kolla-slack: https://hastebin.com/nifezekizu.bash this is the command | 14:36 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Add the api works num for watcher https://review.openstack.org/514340 | 14:45 |
| *** jiriprox has quit IRC | 14:50 | |
| pratapagoutham | kolla-slack: ping | 14:53 |
| *** zhangfei has quit IRC | 14:53 | |
| *** jaosorior has joined #openstack-kolla | 14:57 | |
| *** hachi_ has quit IRC | 14:58 | |
| *** hachi_ has joined #openstack-kolla | 14:58 | |
| *** yangyapeng has joined #openstack-kolla | 15:03 | |
| *** jtriley_ has joined #openstack-kolla | 15:03 | |
| *** yangyapeng has quit IRC | 15:04 | |
| *** yangyapeng has joined #openstack-kolla | 15:04 | |
| *** jtriley has quit IRC | 15:06 | |
| *** lukepatrick has joined #openstack-kolla | 15:10 | |
| *** hachi_ has quit IRC | 15:19 | |
| *** salv-orlando has quit IRC | 15:19 | |
| *** salv-orlando has joined #openstack-kolla | 15:20 | |
| kfox1111 | morning. | 15:33 |
| ktibi | Hi all, I try to deploy infra with kolla. For now, all works except the vxlan network. I can create VM, console VNC works, all services works | 15:45 |
| ktibi | someone have same bug I guess https://ask.openstack.org/en/question/110570/vxlan-peers-not-being-created-on-compute-node-vm-not-getting-dhcp/ | 15:45 |
| ktibi | I use stable/pike | 15:45 |
| ktibi | I don't know why but the creation of vxlan port on br-tun is not called :/ | 15:46 |
| *** rmart04 has joined #openstack-kolla | 15:49 | |
| kolla-slack | <aslancimbom100> @pratapagoutham try something like this "docker run -d -p 5443:5000..." | 16:06 |
| *** erlon has left #openstack-kolla | 16:09 | |
| *** lebauce has quit IRC | 16:11 | |
| inc0 | good morning | 16:16 |
| inc0 | ktibi: ansible? | 16:16 |
| ktibi | inc0 yep | 16:16 |
| inc0 | let me read through this bug, is it accurate description of what you're seeing? | 16:17 |
| ktibi | hum, in my link, user can create vxlan port with restart container. for me no works :/ | 16:17 |
| inc0 | you mean network itself isn't created? | 16:18 |
| ktibi | you can see my conf here : http://paste.openstack.org/show/624372/ | 16:18 |
| *** itlinux has joined #openstack-kolla | 16:18 | |
| inc0 | this is all in one? | 16:18 |
| ktibi | I can create network, I can create VM, I can console in my VM, dhcp and VM port is up in br-int | 16:18 |
| ktibi | but vxlan port in br-tun are not created | 16:19 |
| ktibi | nop 1 ctrl 1 compute | 16:19 |
| inc0 | what's your globals.yml? and check /var/lib/docker/volumes/kolla_logs on hosts for logs | 16:19 |
| ktibi | all service with debug=true but no error | 16:19 |
| ktibi | It looks like some openvswitch flows are not being applied | 16:20 |
| inc0 | to me worrisome is that you have 127.0.0.1 as tunnel endpoints.. | 16:21 |
| ktibi | inc0 globals.yml http://paste.openstack.org/show/624390/ | 16:23 |
| inc0 | right, so tunnels should default to brbond0.1287 | 16:25 |
| ktibi | inc0 neutron try to insert endpoint vxlan | 16:25 |
| ktibi | Endpoint with ip 172.16.7.3 already exists: DBDuplicateEntry: (pymysql.err.IntegrityError) (1062, u"Duplicate entry '172.16.7.3' for key 'PRIMARY'") | 16:26 |
| ktibi | yes brbond0.1287 | 16:26 |
| inc0 | hmm interesting | 16:26 |
| inc0 | seems like bloated db? | 16:26 |
| ktibi | they can ping themselves | 16:26 |
| ktibi | the end : [SQL: u'INSERT INTO ml2_vxlan_endpoints (ip_address, udp_port, host) VALUES (%(ip_address)s, %(udp_port)s, %(host)s)'] [parameters: {'host': u'compute01', 'ip_address': u'172.16.7.3', 'udp_port': 4789}] | 16:27 |
| ktibi | it's just a warning | 16:27 |
| inc0 | right, that's what is issue | 16:27 |
| inc0 | still seems to be problematic | 16:28 |
| ktibi | http://paste.openstack.org/show/624391/ | 16:28 |
| inc0 | can you clean up your env and try to redeploy? | 16:29 |
| ktibi | already try | 16:29 |
| inc0 | did you run kolla-ansible destroy? | 16:30 |
| ktibi | destroy all, container, volumes | 16:30 |
| ktibi | yep | 16:30 |
| ktibi | destroy and deploy | 16:30 |
| *** lpetrut has quit IRC | 16:30 | |
| inc0 | ehh dunno, but this endpoirt in db seems to me what's broken | 16:32 |
| ktibi | error come when I restart neutron agent on compute | 16:32 |
| ktibi | I'll try to delete line in DB and restart agent | 16:32 |
| ktibi | error is in neutron-server.log on ctrl | 16:33 |
| ktibi | inc0 no error but vxlan port not crreated | 16:35 |
| *** jaosorior has quit IRC | 16:35 | |
| *** jaosorior has joined #openstack-kolla | 16:36 | |
| *** tvignaud has quit IRC | 16:39 | |
| *** jmccarthy has joined #openstack-kolla | 16:39 | |
| inc0 | ktibi: uhh I don't know:/ | 16:44 |
| ktibi | yeah me too ^^ | 16:44 |
| ktibi | But with the bug on ask, i'am not alone :D | 16:44 |
| inc0 | right, it might be good to ask neutron people too | 16:45 |
| ktibi | yeah I tried | 16:45 |
| inc0 | hmm, try to do kolla-ansible destroy, reboot all machines | 16:47 |
| inc0 | and deploy fresh | 16:47 |
| *** rmart04 has quit IRC | 16:49 | |
| ktibi | ok | 16:50 |
| *** serlex has quit IRC | 16:58 | |
| *** salv-orlando has quit IRC | 17:09 | |
| *** salv-orlando has joined #openstack-kolla | 17:10 | |
| *** salv-orlando has quit IRC | 17:14 | |
| *** lpetrut has joined #openstack-kolla | 17:18 | |
| *** sambetts is now known as sambetts|afk | 17:23 | |
| *** sambetts has joined #openstack-kolla | 17:24 | |
| *** lpetrut has quit IRC | 17:29 | |
| *** danardelean has quit IRC | 17:33 | |
| *** danardelean has joined #openstack-kolla | 17:34 | |
| *** danardelean has quit IRC | 17:38 | |
| *** jmccarthy has left #openstack-kolla | 17:42 | |
| *** vhosakot has joined #openstack-kolla | 17:47 | |
| *** danardelean has joined #openstack-kolla | 18:08 | |
| *** danardelean has quit IRC | 18:12 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Enable multinode scenarios in zuulv3 https://review.openstack.org/512779 | 18:16 |
| *** aagate has joined #openstack-kolla | 18:18 | |
| iniazi | maybe a stupid question, but what network (kolla-ansible config wise) would live migration use? api_network? | 18:23 |
| inc0 | not a stupid question and yes, api_iface | 18:25 |
| inc0 | you'd need to hack your way around nova configuration out of this | 18:26 |
| *** rwsu has quit IRC | 18:29 | |
| iniazi | no thats fine. i'm trying to figure out (and add additional) 1gbe and 10gbe interfaces. i'm kind of limited to 4 interfaces per node (trying to be also cheap, using off-lease equipemnt etc.) | 18:31 |
| iniazi | 2x 1gbe, and 2x10gbe is easy. but i'm trying to figure out how I'd configure kolla-ansible so it effectively uses those ports. | 18:32 |
| inc0 | one way would be to use bonding across 10gbe | 18:33 |
| inc0 | and vlan + quos | 18:33 |
| inc0 | qos | 18:33 |
| *** lukepatrick has quit IRC | 18:33 | |
| iniazi | on the bond, send ceph + api_iface? | 18:33 |
| inc0 | well if you configure qos for each, you'd be safe | 18:34 |
| inc0 | and you want bonding anyway for HA | 18:34 |
| iniazi | i'll have 2 network nodes, 2 controller, and 4 computes. network nodes will actually have 6 interfaces, 4x1gbe, and 2x10gbe. but the other nodes I'm kind of limited to 4 total ports | 18:35 |
| inc0 | but also depends on your planned usecase - if it's all pet vms, you want good lm network | 18:35 |
| inc0 | if it's cattle, you might not need lm at all | 18:35 |
| iniazi | its mostly web servers, and then some internal db/elasticsearch and light hadoop type of clusters | 18:36 |
| inc0 | well it's always a judgement call | 18:37 |
| iniazi | yeah. i think i have plenty of bandwidth. | 18:37 |
| iniazi | so besides ceph & api subnet, there would be external/public ip network per node, and then the last 1gbe... for tunnel? or stick that on 10gbe too? | 18:40 |
| inc0 | tunnel usually is heaviest one tho | 18:40 |
| iniazi | oh need one for internal non-routable subnet | 18:40 |
| inc0 | it's vm-to-vm communication | 18:40 |
| inc0 | non routable is for external network in neutron | 18:41 |
| inc0 | technically it's only required in network node unless you use DVR | 18:41 |
| iniazi | i was thinking of using dvr. going to configure it and see how it works. | 18:42 |
| inc0 | then you need neutron_external_interface for every node | 18:42 |
| iniazi | so fi i have dvr, i'd need 1 interface for public and 1 for internal subnet (non-ha) | 18:43 |
| inc0 | aka compute nodes gonna external network access | 18:43 |
| iniazi | yeah | 18:44 |
| *** dardelean_ has joined #openstack-kolla | 18:45 | |
| iniazi | on some nodes I can have 4x1gbe (i.e. most of the compute nodes) so I can put bonded 1gbe on those for neutron net and then 1 bond for internal public | 18:45 |
| iniazi | well its really these two nodes i have that i'm running into 4 interface limitation (2x1gbe and 2x10gbe nic on pcie) | 18:45 |
| iniazi | I can just use those for controller only. | 18:46 |
| iniazi | inc0: thanks a bunch. I think I know what to buy/upgrade | 18:47 |
| *** dardelean has quit IRC | 18:48 | |
| inc0 | any time, designing cables is always fun ;) | 18:48 |
| *** lukepatrick has joined #openstack-kolla | 18:48 | |
| *** jgriffith has quit IRC | 18:51 | |
| *** jgriffith_ has joined #openstack-kolla | 18:52 | |
| *** itlinux has quit IRC | 18:52 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla master: WIP Dockerhub publisher job https://review.openstack.org/513855 | 18:52 |
| *** harlowja has joined #openstack-kolla | 18:53 | |
| *** jamesbenson has joined #openstack-kolla | 18:54 | |
| *** itlinux has joined #openstack-kolla | 18:55 | |
| *** jascott1 has joined #openstack-kolla | 18:58 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla master: WIP Dockerhub publisher job https://review.openstack.org/513855 | 19:02 |
| jascott1 | inc0 did you stash any Ubuntu specific notes from yesterday? | 19:02 |
| inc0 | jascott1: nothing much to it tbh | 19:03 |
| SamYaple | o/ jascott1 | 19:03 |
| jascott1 | whatup SamYaple! | 19:03 |
| inc0 | only difference was that I didn't add any ceph repos | 19:03 |
| inc0 | just ran apt-get install ceph-deploy | 19:03 |
| jascott1 | ok im working by that etherpad | 19:03 |
| jascott1 | want to make it centos and ubuntu | 19:04 |
| inc0 | I'm juggling few things, but let me know if you need any help | 19:04 |
| jascott1 | k thanks | 19:04 |
| *** jgriffith_ has quit IRC | 19:06 | |
| *** jgriffith has joined #openstack-kolla | 19:10 | |
| *** jgriffith has quit IRC | 19:11 | |
| *** jgriffith has joined #openstack-kolla | 19:11 | |
| *** jgriffith has quit IRC | 19:11 | |
| *** jgriffith has joined #openstack-kolla | 19:15 | |
| kfox1111 | jascott1: cool. let me know if you run into any issues. | 19:18 |
| kfox1111 | I threw it together pretty quickly, so there is probably a bump or two in the road somewhere. | 19:18 |
| *** jinke_ has joined #openstack-kolla | 19:23 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla master: WIP Dockerhub publisher job https://review.openstack.org/513855 | 19:26 |
| jamesbenson | jascott1 Can I ask what's going on with ceph/working on? | 19:27 |
| *** jinke_ has quit IRC | 19:27 | |
| *** lukepatrick has quit IRC | 19:28 | |
| *** hachi_ has joined #openstack-kolla | 19:33 | |
| inc0 | jamesbenson: yesterday we made little hackathon for ceph+k8s | 19:36 |
| *** jgriffith has quit IRC | 19:37 | |
| jamesbenson | oh nice! | 19:37 |
| *** newmember has quit IRC | 19:38 | |
| *** jgriffith has joined #openstack-kolla | 19:38 | |
| *** itlinux has quit IRC | 19:39 | |
| jamesbenson | I have 2 questions... if anyone is free to help. Networking/cinder :-) | 19:39 |
| inc0 | ask away | 19:41 |
| jamesbenson | I have DVR setup and I can ping my router, but can't access/ping my instances. I usually don't do DVR, so I'm not sure if this is a config I messed up or something else... | 19:43 |
| inc0 | tcpdump your way through network stack to see where ping gets lost? | 19:44 |
| jamesbenson | (with DVR you should have L3 agents, metadata-agents, and openvswitch-agents on all of your compute, correct?) | 19:45 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Enable multinode scenarios in zuulv3 https://review.openstack.org/512779 | 19:47 |
| jamesbenson | cinder: booting with volume always fails, is there a way to either get cinder working properly so this works, or disable it by default so it wont always have that option enabled? | 19:48 |
| kfox1111 | jascott1: ping | 19:53 |
| jascott1 | kfox1111 pong | 19:53 |
| kfox1111 | jascott1: quick go question. | 19:54 |
| kfox1111 | and my go's rusty. | 19:54 |
| jascott1 | "rusty" ha | 19:54 |
| jascott1 | go on :) | 19:54 |
| kfox1111 | say I wanted to have a list which contains hashes. whats the prototype for that? | 19:54 |
| kfox1111 | ie: | 19:54 |
| kfox1111 | foo: | 19:54 |
| kfox1111 | - key: value | 19:54 |
| jascott1 | the values are hashes or? | 19:55 |
| kfox1111 | key and value are both strings. | 19:55 |
| kfox1111 | but each list entry could have abritrary numbers of unique key/values. | 19:55 |
| kfox1111 | so is that []map[string]string ? | 19:56 |
| jascott1 | map[string][]string | 19:57 |
| kfox1111 | so its inner type -> outer type order? | 19:57 |
| jascott1 | map["foo"] = []string{key:value} | 19:57 |
| jascott1 | ^ just for illustration | 19:57 |
| jascott1 | map["foo"] = []string{"key:value"} | 19:58 |
| jascott1 | so would really be | 19:58 |
| kfox1111 | oh.I think thats: | 19:58 |
| kfox1111 | foo: | 19:58 |
| jascott1 | map[string][]string{"mykey1:myval1","mykey2:myval2"} | 19:58 |
| kfox1111 | - "key: value" | 19:59 |
| jascott1 | if you want the key and value not to be a string you have to split | 19:59 |
| kfox1111 | which isn't the same as - key: value (at least in yaml) | 19:59 |
| jascott1 | then you would have to use something there | 19:59 |
| kfox1111 | lets switch to json as it may be more clear... | 19:59 |
| jascott1 | yeah if it wasnt a list it then you would use nested ma | 20:00 |
| jascott1 | map | 20:00 |
| kfox1111 | foo: [{"key":"value"}] | 20:00 |
| kfox1111 | and I wanted to parse foo's data type. | 20:00 |
| jascott1 | thats a map not a string | 20:00 |
| kfox1111 | yes. | 20:00 |
| jascott1 | then its a map[string]interface | 20:00 |
| kfox1111 | so, foo = []map[string]string | 20:01 |
| jascott1 | map[string][]map[string]interface | 20:01 |
| kfox1111 | k. let me try that. thanks. | 20:01 |
| jascott1 | a map with key strings, that containts a slice of maps taking string key and having interface value | 20:01 |
| kfox1111 | I think we're up a level. I just need the decoder for foo itself, not the thing containing foo. | 20:02 |
| jascott1 | ah | 20:02 |
| kfox1111 | like in the strict definition: Foo map[string]interface `json:"foo" yaml:"foo"` | 20:03 |
| jascott1 | value of foo is still a slice of map[string]interface | 20:03 |
| jascott1 | then you have to check the type for interface :( but thats the way its done | 20:04 |
| *** bmace has joined #openstack-kolla | 20:06 | |
| kfox1111 | hmm. ok. | 20:06 |
| jascott1 | its all over helm, im trying to find one | 20:06 |
| sbezverk | kfox1111: as soon as api changed gets merged, I will push changes for rbd external provisioner to support namespace.. | 20:08 |
| kfox1111 | sbezverk: which api change? | 20:08 |
| sbezverk | kfox1111: rbdvolumesource which was used for persistent volume, did not support user's secret name space | 20:09 |
| sbezverk | it is using LocalObject reference | 20:10 |
| kfox1111 | hmm.... | 20:10 |
| kfox1111 | can a pod be launched by a user with a reference to a secret in another namespace anyway? | 20:10 |
| sbezverk | new API rbdpersistentvolumesource will use secret reference where you can specify secret name and namespace | 20:10 |
| kfox1111 | I didn't think rbac would allow that. | 20:10 |
| kfox1111 | hmm... | 20:10 |
| sbezverk | kfox1111: that part I am not sure | 20:10 |
| kfox1111 | have a link for that? | 20:10 |
| sbezverk | need to check the code | 20:11 |
| sbezverk | sure | 20:11 |
| kfox1111 | thx. | 20:11 |
| sbezverk | kfox1111: https://github.com/kubernetes/kubernetes/pull/54302/ | 20:11 |
| kfox1111 | cool. thanks. | 20:11 |
| sbezverk | so we do in tree first because of api change | 20:12 |
| sbezverk | next step out of tree provisioner | 20:12 |
| kfox1111 | I'd definatly double check this against rbac enabled k8s. as that is increasingly common. | 20:13 |
| kfox1111 | there may need to be an exception that allows a dynamically created pv's rbd secret to come from a different namespace whether the user has access to it or not. | 20:13 |
| kfox1111 | I'm pretty sure it will be broken out of the box right now. :( | 20:13 |
| kfox1111 | otherwise, I think a user could just put in any volum request for an rbd volume specifying any users creds. | 20:20 |
| sbezverk | rootfs just tested it in his test bed and it works | 20:21 |
| *** hachi_ has quit IRC | 20:22 | |
| kfox1111 | with a non admin user? | 20:22 |
| sbezverk | yep | 20:22 |
| kfox1111 | have him try just a bare pod with rbd volume directly specified. | 20:22 |
| kfox1111 | that one shouldn't work. | 20:23 |
| sbezverk | that will not work | 20:23 |
| kfox1111 | are you sure? | 20:23 |
| sbezverk | direct reference from pod uses different spec | 20:23 |
| kfox1111 | ah. | 20:23 |
| kfox1111 | so the ps allows it as its a different route and its under the control of the creater of the pv. | 20:24 |
| kfox1111 | should be ok then. | 20:24 |
| sbezverk | right | 20:24 |
| jamesbenson | inc0: my instances can't ping the external gateway, but can ping google.com | 20:26 |
| jamesbenson | and outside of the cloud, I can ping the external gateway IP | 20:27 |
| *** salv-orlando has joined #openstack-kolla | 20:33 | |
| *** lukepatrick has joined #openstack-kolla | 20:34 | |
| *** salv-orlando has quit IRC | 20:37 | |
| *** salv-orlando has joined #openstack-kolla | 20:37 | |
| *** lpetrut has joined #openstack-kolla | 20:45 | |
| *** jascott1_ has joined #openstack-kolla | 20:49 | |
| *** jascott1 has quit IRC | 20:49 | |
| *** ktibi has quit IRC | 20:52 | |
| *** aagate has quit IRC | 20:59 | |
| *** jgriffith has quit IRC | 20:59 | |
| *** jgriffith has joined #openstack-kolla | 21:00 | |
| *** danardelean has joined #openstack-kolla | 21:08 | |
| *** dardelean_ has quit IRC | 21:08 | |
| *** jtriley_ has quit IRC | 21:10 | |
| *** spiette has joined #openstack-kolla | 21:14 | |
| *** Pavo has joined #openstack-kolla | 21:15 | |
| *** lukepatrick has quit IRC | 21:17 | |
| *** lukepatrick has joined #openstack-kolla | 21:19 | |
| *** lukepatrick has quit IRC | 21:20 | |
| jamesbenson | inc0: is it possible that this could be soemthing with iptables: | 21:22 |
| jamesbenson | ? | 21:22 |
| *** lukepatrick has joined #openstack-kolla | 21:22 | |
| *** dave-mccowan has quit IRC | 21:28 | |
| *** dougsz has quit IRC | 21:28 | |
| *** dougsz has joined #openstack-kolla | 21:29 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Enable multinode scenarios in zuulv3 https://review.openstack.org/512779 | 21:31 |
| *** dougsz has quit IRC | 21:33 | |
| *** MasterOfBugs has joined #openstack-kolla | 21:34 | |
| *** pramodrj07 has joined #openstack-kolla | 21:34 | |
| *** Pavo has quit IRC | 21:35 | |
| *** Pavo has joined #openstack-kolla | 21:37 | |
| *** jgriffith has quit IRC | 21:50 | |
| SamYaple | inc0: did you figure out the mariadb issue? | 21:51 |
| *** sambetts has quit IRC | 21:53 | |
| kolla-slack | <inc0> Negative, trying few things to confirm that net works | 21:54 |
| *** lpetrut has quit IRC | 21:55 | |
| *** jgriffith has joined #openstack-kolla | 21:58 | |
| *** jamesbenson has quit IRC | 22:03 | |
| danardelean | hi guys | 22:07 |
| danardelean | there is a slack channel for kolla? | 22:07 |
| masber | danardelean, yes it is a channel within the kubernetes.slack.com | 22:09 |
| danardelean | ty | 22:10 |
| masber | good morning, I am having an issue deploying kolla, where the kolla_internal_vip_address is not created on the control/network host, as a consequence I am getting this error during deployment --> fatal: [test-openstack-controller]: FAILED! => {"changed": false, "elapsed": 301, "failed": true, "msg": "Timeout when waiting for 192.168.20.5:3306"} | 22:11 |
| masber | my question is, on which interface should kolla_internal_vip_address live? | 22:12 |
| *** Pavo has quit IRC | 22:13 | |
| masber | these are the details http://paste.openstack.org/raw/624412/ | 22:19 |
| *** salv-orlando has quit IRC | 22:34 | |
| *** salv-orlando has joined #openstack-kolla | 22:34 | |
| *** hrw has quit IRC | 22:38 | |
| *** salv-orlando has quit IRC | 22:39 | |
| *** vhosakot has quit IRC | 22:44 | |
| *** dave-mccowan has joined #openstack-kolla | 22:49 | |
| inc0 | hmm masber interesting, that might be same issue I'm experiencing in gates | 22:53 |
| SamYaple | inc0: im pretty sure it is. its the issue i was saying keeps popping up in the channel | 22:54 |
| inc0 | masber: but you're saying IP doesn't appear on controller? | 22:55 |
| inc0 | it should live in api interface | 22:55 |
| inc0 | that would mean there is issue with keepalived | 22:55 |
| masber | inc0, correct, if you see the paste log, kolla_internal_vip_address: "192.168.20.5" is missing on the controller/network node | 22:56 |
| masber | inc0, yes I assumed same thing, the 192.168.20.5 belongs to the management segment assigned to the api_interface | 22:57 |
| inc0 | masber: just to confirm it's keepalived issue, try putting kolla_internal_vip_address : << actual address of one of controllers >> | 22:57 |
| inc0 | and enable_haproxy: false | 22:57 |
| *** janki has joined #openstack-kolla | 22:57 | |
| inc0 | that would bypass all keepalived mechanisms and deployment should be successful | 22:57 |
| masber | inc0, ok, let me check | 22:57 |
| inc0 | that would also explain why our previous gates works | 22:58 |
| inc0 | or why my local environment worked | 22:59 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Enable multinode scenarios in zuulv3 https://review.openstack.org/512779 | 23:01 |
| *** jgriffith has quit IRC | 23:01 | |
| *** janki has quit IRC | 23:02 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Enable multinode scenarios in zuulv3 https://review.openstack.org/512779 | 23:03 |
| *** jgriffith has joined #openstack-kolla | 23:08 | |
| *** danardelean has quit IRC | 23:14 | |
| *** danardelean has joined #openstack-kolla | 23:15 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla master: WIP Dockerhub publisher job https://review.openstack.org/513855 | 23:15 |
| inc0 | masber: any progress? | 23:16 |
| masber | inc0, yes, it works | 23:16 |
| inc0 | so keepalived is broken | 23:16 |
| masber | inc0, deployment is successful if haproxy is disabled | 23:16 |
| inc0 | tomorrow I'll prep env for myself to enable keepalived and try to reproduce | 23:17 |
| inc0 | this is critical | 23:17 |
| inc0 | SamYaple: we both were right, it's not networking but it's not mariadb as well | 23:17 |
| SamYaple | heh | 23:17 |
| SamYaple | mariadb is just the first time the vip is used eh | 23:18 |
| *** duonghq has quit IRC | 23:18 | |
| inc0 | right | 23:18 |
| inc0 | masber: you're using ubuntu? | 23:18 |
| masber | in my case issue was that kolla_internal_vip_address was missing, the creating of that subinterface was part of the hapoxy? | 23:18 |
| *** duonghq has joined #openstack-kolla | 23:18 | |
| masber | inc0, I am using centos binary and host is centos 7 | 23:19 |
| SamYaple | inc0: you said you deployed multinode successfully, no? | 23:19 |
| SamYaple | by hand | 23:19 |
| inc0 | masber: it's not about interface, keepalived adds new IP address to existing iface | 23:19 |
| inc0 | SamYaple: yes, but without haproxy | 23:19 |
| inc0 | I have it disabled locally because neutron address pairs | 23:19 |
| SamYaple | oh... well thats no multinode then | 23:20 |
| *** danardelean has quit IRC | 23:20 | |
| inc0 | not full ha, but good for most of clustering tests | 23:20 |
| inc0 | but yeah, I'll fix that up tomorrow | 23:20 |
| SamYaple | clearly not lol | 23:20 |
| SamYaple | but yes this clears up the confusion | 23:20 |
| inc0 | I'd wish you could just tell neutron to not care about address pairs -.- | 23:21 |
| SamYaple | inc0: you can | 23:21 |
| inc0 | in configs? | 23:21 |
| SamYaple | ya | 23:21 |
| SamYaple | or as admin on the cluster per port | 23:21 |
| inc0 | I don't have acccess to this cloud:/ | 23:21 |
| SamYaple | well yea if you are just a user then that would be really bad to expose | 23:22 |
| inc0 | I know | 23:22 |
| SamYaple | as a user, no. you get what youre given | 23:22 |
| inc0 | even on your tenant network you created? | 23:22 |
| SamYaple | yes because you could still do evil things on the hypervisor otherwise | 23:23 |
| inc0 | yeah I guess | 23:23 |
| inc0 | well, I screwed:/ | 23:23 |
| SamYaple | this is why i vxlan in my cluster of hosts on clouds | 23:23 |
| SamYaple | full l2 capabilities | 23:23 |
| inc0 | yeah I might add this to my cluster too | 23:24 |
| inc0 | Jeffrey4l: when you're awake, can you try to deploy multinode with master? Keepalived seems to be broken | 23:24 |
| inc0 | I'll deal with this tomorrow morning when ops for my cloud comes to work | 23:25 |
| SamYaple | according to my logs, im pretty sure its been broken for the life of 5.0.0 too, just fyi | 23:25 |
| SamYaple | give you a place to git blame | 23:25 |
| inc0 | unlikely since people did deploy 5.0 already and with success | 23:25 |
| SamYaple | with haproxy :) | 23:26 |
| SamYaple | could just be a subset of distros that are borked | 23:26 |
| *** dave-mccowan has quit IRC | 23:27 | |
| *** goldyfruit has quit IRC | 23:31 | |
| masber | I would like to ask if I can use this guide https://docs.openstack.org/kolla-ansible/latest/reference/external-ceph-guide.html with ceph luminus? | 23:32 |
| SamYaple | masber: kolla code doesnt have support for lumonous yet | 23:33 |
| SamYaple | so you would have to use external, but youll have to modify the sources to pull in loumous in the build | 23:34 |
| masber | SamYaple, yes, I mean deploy external ceph using ceph-ansible and connect it to openstack | 23:35 |
| kolla-slack | <britthouser> @masber you'll still need to pull luminus client into the kolla containers. | 23:37 |
| SamYaple | masber: right im saying youll have to modify the sources.list or the centos stuff when you build the containers | 23:37 |
| SamYaple | for the lumionous clients | 23:37 |
| masber | SamYaple, ok, yes the clients! | 23:38 |
| SamYaple | sorry, ill attempt to be more clear :) | 23:38 |
| masber | np, thank you for your help!, I still remember you were the one who helped me deploying my first kolla environment | 23:39 |
| *** hongbin has quit IRC | 23:39 | |
| SamYaple | i remember that too | 23:40 |
| *** jascott1_ has quit IRC | 23:44 | |
| *** jascott1 has joined #openstack-kolla | 23:44 | |
| *** lukepatrick has quit IRC | 23:44 | |
| SamYaple | inc0: btw, i have a significantly better way (in my opinoin) t odeploy ceph now | 23:45 |
| inc0 | which is? | 23:45 |
| *** jascott1 has quit IRC | 23:45 | |
| SamYaple | it involves creating the ceph keys and saving them as passwords. they are generic not-linked-to-the-cluster base64 encoded 16 byte strings | 23:46 |
| SamYaple | that way we dont have to do alot of the fetch distribute dances | 23:46 |
| SamYaple | and they really do act like password | 23:46 |
| SamYaple | to change kolla to it though would require some hands on migration | 23:46 |
| SamYaple | because you would need ot update teh secrets | 23:47 |
| SamYaple | with teh existing secrets from teh existing clusters | 23:47 |
| SamYaple | but basically it lets you bring up new nodes or new clusters at the same time | 23:47 |
| SamYaple | ceph can work itself out | 23:47 |
| SamYaple | no more monmap copying or anything like that | 23:47 |
| SamYaple | im doing it in salt now, but i basically start all the tasks at once. osds. mons. mgrs. | 23:48 |
| SamYaple | doesnt matter if its a new cluster or not | 23:48 |
| SamYaple | i suppose you could build that into the kolla-ansible command | 23:49 |
| SamYaple | like a fetch-ceph-secrets thing | 23:49 |
| inc0 | probably, yes | 23:49 |
| SamYaple | then use that output to update secrets.yml | 23:49 |
| SamYaple | but anyway. its way better and easier to work with | 23:49 |
| SamYaple | alot less ordering (read: none) | 23:50 |
| inc0 | although we're thinking of dropping ceph in general | 23:50 |
| SamYaple | do you recall me saying "im not sure if ceph should like in kolla since its basically like another project" | 23:50 |
| SamYaple | i stand by that :P | 23:50 |
| SamYaple | s/like/live/ | 23:51 |
| inc0 | we'll upgrade to L now and probably deprecate it alltogether or at least externalise ceph deployment | 23:51 |
| SamYaple | if you provide a migration path to ceph-docker that would be best | 23:51 |
| SamYaple | not the best images... but at least you could galaxy in the ceph-ansible playbooks | 23:51 |
| inc0 | it's really cool to have simple auto-orchestration of cinder and such | 23:52 |
| inc0 | so it's going to be fun, but we'll have whole next release for that | 23:52 |
| inc0 | I'm worried about upgrade tho | 23:53 |
| SamYaple | i was happy with my ceph work in ansible there, but its not standard, ceph-ansible is standard | 23:53 |
| SamYaple | moving to that would be best | 23:54 |
| inc0 | yeah, I agree, ideally we'd have migration path from one to another | 23:54 |
| inc0 | only reason I'm hesitant is that people already have ceph done by us and I don't want them to be screwed | 23:54 |
| SamYaple | a bit of parted magic to recombobulate the drives and youll be fine | 23:55 |
| SamYaple | i brought my home cluster from bare-metal -> kolla-ceph -> baremetal -> ceph-salt | 23:55 |
| SamYaple | so its doable for sure | 23:55 |
| inc0 | maybe it's documentation issue | 23:57 |
| inc0 | we will upgrade to L this release, but maybe deprecate it too | 23:58 |
| SamYaple | documentation and a playbook to assist and itll be easy | 23:58 |
| SamYaple | your bigger problem will be doing the ceph-ansible integration | 23:58 |
| SamYaple | because you arent using any outside galaxy roles now | 23:59 |
| SamYaple | so thats a whole other piece oyu gotta work out | 23:59 |
| inc0 | yup | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!