| nomatics | Should Install cinder services play be ran against a host? It first runs the play in the container and then on the host. I have is_metal set to false in env.d/cinder-volumes.yml. | 00:00 |
|---|---|---|
| *** yifei has joined #openstack-ansible | 00:03 | |
| *** exodusftw has quit IRC | 00:13 | |
| *** marc_ab has quit IRC | 00:17 | |
| *** exodusftw has joined #openstack-ansible | 00:17 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible stable/newton: Added a common tag to the common tasks https://review.openstack.org/498992 | 00:19 |
| *** thorst_afk has joined #openstack-ansible | 00:22 | |
| *** nomatics_ has joined #openstack-ansible | 00:27 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: migrate_openstack_vars.py: Correct folder name https://review.openstack.org/498871 | 00:28 |
| *** nomatics has quit IRC | 00:28 | |
| *** thorst_afk has quit IRC | 00:29 | |
| *** thorst_afk has joined #openstack-ansible | 00:30 | |
| *** thorst_afk has quit IRC | 00:34 | |
| *** lbragstad has joined #openstack-ansible | 00:44 | |
| *** exodusftw has quit IRC | 00:49 | |
| *** basilAB has left #openstack-ansible | 00:51 | |
| *** woodard has quit IRC | 00:51 | |
| *** woodard has joined #openstack-ansible | 00:51 | |
| *** exodusftw has joined #openstack-ansible | 00:55 | |
| *** thorst_afk has joined #openstack-ansible | 01:02 | |
| *** thorst_afk has quit IRC | 01:03 | |
| *** galstrom_zzz is now known as galstrom | 01:41 | |
| *** nomatics_ has quit IRC | 01:53 | |
| *** kukacz has quit IRC | 02:00 | |
| *** kukacz has joined #openstack-ansible | 02:01 | |
| *** thorst_afk has joined #openstack-ansible | 02:04 | |
| *** lbragstad has quit IRC | 02:07 | |
| *** thorst_afk has quit IRC | 02:09 | |
| *** dxiri has joined #openstack-ansible | 02:31 | |
| *** dxiri has quit IRC | 02:32 | |
| *** dxiri has joined #openstack-ansible | 02:32 | |
| *** dxiri has quit IRC | 02:33 | |
| *** dxiri_ has quit IRC | 02:34 | |
| *** dxiri has joined #openstack-ansible | 02:35 | |
| *** gouthamr has quit IRC | 02:46 | |
| *** galstrom is now known as galstrom_zzz | 02:56 | |
| *** dxiri has quit IRC | 02:59 | |
| *** dxiri has joined #openstack-ansible | 03:02 | |
| *** thorst_afk has joined #openstack-ansible | 03:05 | |
| *** thorst_afk has quit IRC | 03:10 | |
| cloudnull | nomatics if you have ismetal false it should only run in the container | 03:11 |
| cloudnull | if it is you may need to clean up the inventory | 03:12 |
| cloudnull | checkout the inventory-manage script | 03:13 |
| cloudnull | and remove the host from the cinder group | 03:14 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Run gate playbooks in parallel https://review.openstack.org/497742 | 03:17 |
| *** udesale has joined #openstack-ansible | 03:18 | |
| prometheanfire | I feel like I should submit reverts to osa-ops for https://github.com/openstack/openstack-ansible-ops/commit/9049479dda62e845e87c6cb22050f82272cc542b and https://github.com/openstack/openstack-ansible-ops/commit/53d548dfba6a8a45c6c7afa6c8550fe76cfd4663 | 03:22 |
| prometheanfire | even with the 'fix' leaps are broken | 03:22 |
| cloudnull | from those patches ? | 03:25 |
| prometheanfire | cloudnull: yes | 03:25 |
| prometheanfire | also | 03:25 |
| prometheanfire | looks like the db migration fixes didn't work | 03:25 |
| cloudnull | it doesn't look like those prs do anything by default ? | 03:26 |
| cloudnull | being that the vars are unset ? | 03:26 |
| prometheanfire | re-deploy.sh: line 65: syntax error in conditional expression | 03:26 |
| prometheanfire | maybe? | 03:27 |
| prometheanfire | I reverted them locally | 03:27 |
| cloudnull | if you remove the -u https://github.com/openstack/openstack-ansible-ops/blob/master/leap-upgrades/re-deploy.sh#L21 | 03:28 |
| cloudnull | ? | 03:28 |
| prometheanfire | maybe fine | 03:28 |
| prometheanfire | I'll test it in a bit | 03:28 |
| prometheanfire | cloudnull: /win 1 | 03:28 |
| prometheanfire | bah | 03:28 |
| prometheanfire | cloudnull: would this work for leap upgrades? It's my understanding that the leap upgrades don't use anything from the rpc-openstack or it's submodule https://github.com/rcbops/rpc-openstack/pull/2491/files | 03:29 |
| prometheanfire | I'm thinking that the leap upgrades don't care about the rpc-o ansible-role-requirements.yml file | 03:31 |
| cloudnull | I'm not sure, the leap upgrades were not designed with rpc-o specifically in mind. | 03:32 |
| prometheanfire | ya | 03:32 |
| cloudnull | the inline modifications rpco is making to the process change a lot of the functionality | 03:33 |
| prometheanfire | it looks like the rpc-o wraping of the leap upgrades doesn't refrence ansible-role-requirements at all | 03:33 |
| prometheanfire | yep... | 03:33 |
| cloudnull | quite likely | 03:33 |
| prometheanfire | wish I didn't waste the last 12 hours then | 03:33 |
| *** gkadam_ has joined #openstack-ansible | 03:33 | |
| cloudnull | on that file ? | 03:33 |
| prometheanfire | ya | 03:34 |
| prometheanfire | well, I was testing an upgrade | 03:34 |
| prometheanfire | between those two commits and that modification... | 03:34 |
| *** lbragstad has joined #openstack-ansible | 03:37 | |
| *** dave-mccowan has quit IRC | 03:40 | |
| prometheanfire | cloudnull: you should like the cruftyness, but here's my osa-ops patch :D http://sprunge.us/iGRf | 03:53 |
| prometheanfire | I'm fairly sure later definitions in ansible-role-requirements.yml override the earlier ones | 03:53 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Use separate interfaces for lb traffic https://review.openstack.org/498862 | 03:57 |
| cloudnull | it'll pull down the osa roles when the bootstrap is run | 04:00 |
| prometheanfire | I wonder if the leap runs bootstrap for each step | 04:02 |
| prometheanfire | I'd expect it to | 04:02 |
| * prometheanfire shrugs | 04:02 | |
| cloudnull | https://github.com/openstack/openstack-ansible-ops/blob/ee85d37ce9fed010d54b2c95d3078cd6f0f4bbe9/leap-upgrades/lib/functions.sh#L122 | 04:03 |
| cloudnull | it does | 04:03 |
| prometheanfire | cool | 04:04 |
| prometheanfire | so I should be good with that patch then | 04:04 |
| prometheanfire | as dirty as it is | 04:04 |
| cloudnull | rpc-o should leap and then apply its differences and stop injecting nonsense at runtime. | 04:05 |
| prometheanfire | http://i0.kym-cdn.com/photos/images/facebook/000/210/119/9b3.png | 04:05 |
| *** thorst_afk has joined #openstack-ansible | 04:06 | |
| cloudnull | things like https://github.com/rcbops/rpc-openstack/blob/master/scripts/bootstrap-ansible.sh#L74-L80 should be stopped | 04:06 |
| cloudnull | hahaha | 04:06 |
| cloudnull | the image applies to the second comment too | 04:06 |
| prometheanfire | btw, mind reviewing https://review.openstack.org/498987 ? | 04:08 |
| prometheanfire | it's a one character fix :P | 04:08 |
| prometheanfire | well, two maybe with spacing | 04:09 |
| *** lbragstad has quit IRC | 04:09 | |
| *** thorst_afk has quit IRC | 04:10 | |
| prometheanfire | cloudnull: email sent to rax email explaining it all | 04:11 |
| cloudnull | https://github.com/rcbops/rpc-openstack/pull/2492 | 04:13 |
| prometheanfire | :D | 04:13 |
| cloudnull | ^ should help start remove the injections | 04:13 |
| *** chhavi has joined #openstack-ansible | 04:15 | |
| prometheanfire | looks likt I'm staying up for the night to babysit this | 04:17 |
| cloudnull | https://github.com/rcbops/rpc-openstack/pull/2493 | 04:34 |
| cloudnull | prometheanfire: ^ | 04:34 |
| bhujay | trying to build multios repo. repo1 - 3 are Ubuntu containers and repo4 is centos . build on repo 4 went smooth but repo1 is failing at wheel building stage . IOError:[Errorno 2 ] No such file or directory : /var/www/repo/pools/centos-7.3-x86_64/ldappool/ldappool-2.1.0.py2.py3-none-any.whl | 04:38 |
| bhujay | any clue ? | 04:38 |
| cloudnull | bhujay: o/ | 04:40 |
| bhujay | I am using OSA 15.1.6 cloud that be a problem ? shd I upgrade to 15.1.8? | 04:40 |
| cloudnull | I'd not expect so , however using the latest stable is a good way to rule out issues. | 04:41 |
| bhujay | sure , let me try | 04:41 |
| *** rstarmer has quit IRC | 04:43 | |
| *** rstarmer has joined #openstack-ansible | 04:43 | |
| *** chyka has quit IRC | 04:58 | |
| *** thorst_afk has joined #openstack-ansible | 05:07 | |
| *** thorst_afk has quit IRC | 05:11 | |
| *** chhavi has quit IRC | 05:12 | |
| *** chhavi has joined #openstack-ansible | 05:12 | |
| *** udesale__ has joined #openstack-ansible | 05:13 | |
| *** a0x3e8 has quit IRC | 05:13 | |
| *** udesale has quit IRC | 05:15 | |
| *** zcourts has quit IRC | 05:20 | |
| *** zcourts has joined #openstack-ansible | 05:21 | |
| *** yolanda has quit IRC | 05:21 | |
| *** zcourts has quit IRC | 05:25 | |
| prometheanfire | turns out cats don't like the taste of espresso | 05:43 |
| cloudnull | who knew :D | 05:49 |
| prometheanfire | me | 05:52 |
| prometheanfire | now :P | 05:52 |
| prometheanfire | just about time to test another leap | 05:52 |
| *** vishwana_ has quit IRC | 06:02 | |
| *** vishwanathj has joined #openstack-ansible | 06:03 | |
| *** jwitko has quit IRC | 06:03 | |
| *** jwitko has joined #openstack-ansible | 06:04 | |
| *** poopcat has quit IRC | 06:04 | |
| *** drifterza has joined #openstack-ansible | 06:05 | |
| *** poopcat has joined #openstack-ansible | 06:06 | |
| *** thorst_afk has joined #openstack-ansible | 06:07 | |
| *** thorst_afk has quit IRC | 06:12 | |
| *** Oku_OS-away is now known as Oku_OS | 06:15 | |
| hw_wutianwei | cloudnull: hi, I meet some issue http://paste.openstack.org/show/619849/ | 06:17 |
| hw_wutianwei | I use master | 06:17 |
| hw_wutianwei | did I miss something? | 06:17 |
| *** huxinhui_ has joined #openstack-ansible | 06:21 | |
| cloudnull | hw_wutianwei: maybe missed something in the user config | 06:26 |
| cloudnull | have a look at https://docs.openstack.org/openstack-ansible-haproxy_server/latest/ | 06:26 |
| cloudnull | for more on the haproxy configs | 06:26 |
| hw_wutianwei | cloudnull: I check the keepalived code {% for name, sync_group in keepalived_sync_groups.iteritems() %}, 'iteritems()' seem to be a function. | 06:33 |
| *** cshen has quit IRC | 06:33 | |
| *** cshen has joined #openstack-ansible | 06:35 | |
| hw_wutianwei | cloudnull: I have no idea which I missed. In my opinion, there are default when I missed | 06:36 |
| cloudnull | hw_wutianwei: can you check the python version | 06:40 |
| cloudnull | in /opt/ansible-runtime/bin/python | 06:40 |
| cloudnull | hw_wutianwei: If you have py3 | 06:41 |
| cloudnull | i wonder if you have an old version of the keepalived role ? | 06:41 |
| cloudnull | https://github.com/evrardjp/ansible-keepalived/commit/2553ddffd948f3e8e8196cd017d27716b396554a | 06:41 |
| cloudnull | I got a change in that role a while back to address the py3 issues | 06:41 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/ansible-role-requirements.yml#L29-L32 | 06:42 |
| cloudnull | you should have that role out of master | 06:42 |
| hw_wutianwei | the python version is 3.5.2 | 06:42 |
| cloudnull | so you shouldn't see that specific error | 06:42 |
| cloudnull | when was this master repo cloned? | 06:42 |
| *** pcaruana has joined #openstack-ansible | 06:43 | |
| hw_wutianwei | commit 4d39f2cc29417153780210fc0bb86223387e9968 | 06:43 |
| hw_wutianwei | Merge: b4f0401 c5b317d | 06:43 |
| hw_wutianwei | Author: Jenkins <jenkins@review.openstack.org> | 06:43 |
| hw_wutianwei | Date: Wed Aug 23 00:06:31 2017 +0000 | 06:43 |
| hw_wutianwei | cloudnull: this is the latest git log. | 06:44 |
| cloudnull | was this an old checkout before ? | 06:45 |
| cloudnull | if so have you rerun ./scripts/bootstrap-ansible.sh | 06:45 |
| cloudnull | that will pull down the latest roles. | 06:45 |
| neith | hey guys when rebuilding venvs after updating from 14.04 to 16.04 I get | 06:46 |
| neith | https://www.irccloud.com/pastebin/BaONKje9/ | 06:46 |
| neith | any idea? | 06:46 |
| hw_wutianwei | cloudnull: ok, I will try | 06:47 |
| hw_wutianwei | thanks | 06:47 |
| cloudnull | hw_wutianwei: np | 06:48 |
| cloudnull | neith: running repo-build? | 06:48 |
| cloudnull | do you still have a 14.04 host ? | 06:48 |
| neith | cloudnull: TASK [repo_build : Create OpenStack-Ansible requirement wheels] **************** | 06:48 |
| neith | cloudnull: yes | 06:48 |
| cloudnull | http://paste.openstack.org/show/619850/ | 06:49 |
| cloudnull | looks like that's the error | 06:50 |
| cloudnull | seems the gnocci checkout is missing? | 06:50 |
| cloudnull | s'/checkout/wheel/ | 06:50 |
| cloudnull | # /var/www/repo/pools/ubuntu-14.04-x86_64/gnocchiclient/gnocchiclient-2.8.2-py2.py3-none-any.whl | 06:50 |
| cloudnull | maybe that's a broken link? | 06:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Correct cinder online migrations command https://review.openstack.org/499044 | 06:51 |
| neith | cloudnull: maybe | 06:51 |
| neith | cloudnull: the file is present | 06:53 |
| cloudnull | and its not a broken symlink ? | 06:54 |
| cloudnull | if no, try rerunning the repo-build | 06:54 |
| *** gtrxcb has quit IRC | 06:54 | |
| cloudnull | also do you have both 14.04 and 16.04 infra hosts? | 06:55 |
| neith | cloudnull: my bad the pools dir is empty | 06:55 |
| cloudnull | ah. ok. | 06:55 |
| neith | though /var/www/repo/ubuntu-14.04-x86_64/gnocchiclient/gnocchiclient-2.8.2-py2.py3-none-any.whl exists | 06:55 |
| cloudnull | you'll need to comb through the dirs and clean up any busted symlinks | 06:55 |
| cloudnull | or nuke /var/www/repo/ubuntu-14.04-x86_64 and rerun the repo-build | 06:56 |
| neith | cloudnull: I can nuke it | 06:56 |
| cloudnull | so long as you have 1 infra 14.04 and 1 infra 16.04 it will rebuild | 06:56 |
| neith | :) | 06:56 |
| neith | its the first infra i'm upgrading | 06:57 |
| *** yolanda has joined #openstack-ansible | 06:57 | |
| neith | yet, I dont have any working unfra on 16.04 | 06:57 |
| *** drifterza has quit IRC | 06:57 | |
| cloudnull | ah. you will need at least 1 16.04 infra | 06:59 |
| *** a0x3e8 has joined #openstack-ansible | 06:59 | |
| cloudnull | to build the different system wheels | 06:59 |
| a0x3e8 | hello. problem: I finally managed to deploy openstack using ansible but I had to create cinder volume manually. anyway, I want to access to my instances from outside of the stack so I need network. in my setup I used 3 physical interfaces instead of 3 vlans in default configurations. my management ip is 10.1.10.0/24, network if: 10.1.20.0/24 and st | 07:00 |
| a0x3e8 | orage network: 10.1.30.0/24. I tried to create a network with subnet 10.1.20.0/24 and made it external! but no access! how should I create a network with external access which will connect to my 10.1.20.0/24 vlan in router? | 07:00 |
| neith | cloudnull: infra1 is on 16.04 | 07:00 |
| neith | cloudnull: but the repo container get the task I previously mentionned | 07:00 |
| cloudnull | oh. | 07:00 |
| cloudnull | ok | 07:00 |
| cloudnull | it should go serially and build everything for each os type it encounters and then sync | 07:01 |
| *** arbrandes has joined #openstack-ansible | 07:01 | |
| neith | cloudnull: destroying the container and replay all the playbboks should help? | 07:03 |
| *** arbrandes1 has quit IRC | 07:04 | |
| cloudnull | neith: you can destroy just that one. then rebuild it and run the repo buts | 07:05 |
| cloudnull | **bits | 07:05 |
| cloudnull | something like `openstack-ansible lxc-container-destroy.yml lxc-container-create.yml repo-install.yml --limit repo_all` | 07:05 |
| cloudnull | I'm off. | 07:05 |
| * cloudnull time for sleep | 07:05 | |
| cloudnull | take care all | 07:06 |
| neith | cloudnull: sleep tight ;) | 07:06 |
| *** a0x3e8 has left #openstack-ansible | 07:08 | |
| *** thorst_afk has joined #openstack-ansible | 07:08 | |
| *** jvidal has joined #openstack-ansible | 07:09 | |
| *** jamielennox has quit IRC | 07:09 | |
| *** gus has quit IRC | 07:09 | |
| *** gus has joined #openstack-ansible | 07:10 | |
| *** thorst_afk has quit IRC | 07:13 | |
| *** jamielennox has joined #openstack-ansible | 07:14 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:15 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:17 |
| *** sxc731 has joined #openstack-ansible | 07:20 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:23 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:24 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:27 |
| *** drifterza has joined #openstack-ansible | 07:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Pin erlang packages to version 19.3 https://review.openstack.org/498952 | 07:32 |
| *** coolkil has joined #openstack-ansible | 07:32 | |
| coolkil | morning | 07:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 07:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/newton: Pin erlang packages to version 19.3 https://review.openstack.org/498952 | 07:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Update documentation redirects https://review.openstack.org/499059 | 07:35 |
| coolkil | im looking at https://github.com/openstack/openstack-ansible-repo_server/blob/master/tasks/repo_post_install.yml#L71-L89 I'm wondering why the base venvs directory is not created during this stage? | 07:36 |
| *** mbuil has joined #openstack-ansible | 07:38 | |
| odyssey4me | coolkil it gets done here instead: https://github.com/openstack/openstack-ansible-repo_build/blob/master/tasks/repo_build_venvs.yml#L16-L24 | 07:38 |
| coolkil | the problem is that when somone builds the repo for the very first time while using multiple architectures the sync_repo.yml wil try to sync to /var/www/repo/vens/<openstack_versionnr> but because the venvs dir is nonexistant on the target server this fails | 07:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Update documentation redirects https://review.openstack.org/499064 | 07:40 |
| odyssey4me | coolkil aha, well that's a bug then | 07:40 |
| odyssey4me | it does seem sensible to create it in the repo_server role along with the others | 07:40 |
| coolkil | the only differance is the permissions | 07:41 |
| coolkil | venvs is 755 and the rest is 775 | 07:42 |
| coolkil | scratch that | 07:42 |
| coolkil | read the code wrong | 07:42 |
| coolkil | I'l make the change! | 07:43 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: Correct leap hook closing brackets https://review.openstack.org/499065 | 07:47 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-specs master: (WIP) Blueprint for Congress integration with OSA. https://review.openstack.org/499066 | 07:48 |
| odyssey4me | thanks coolkil | 07:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Correct cinder online migrations command https://review.openstack.org/499068 | 07:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Correct cinder online migrations command https://review.openstack.org/499069 | 07:50 |
| bhujay | a0x3e8:While creating the external net using openstack network create , have you mentioned a gateway for 1.20 netwrok , that is physically existing and connected physical net 1.20 | 07:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: Correct cinder online migrations command https://review.openstack.org/499071 | 07:57 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/pike: Correct cinder online migrations command https://review.openstack.org/499072 | 07:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/ocata: Correct cinder online migrations command https://review.openstack.org/499073 | 07:59 |
| *** drifterza has quit IRC | 08:05 | |
| bhujay | My repo build job completed ok with Ubuntu and Centos . I updated OSA to 15.1.8 and also re - created clean repo containers . | 08:05 |
| *** thorst_afk has joined #openstack-ansible | 08:09 | |
| *** jafeha has joined #openstack-ansible | 08:11 | |
| bhujay | a0x3e8: Are you saying on the server side there is no VLAN or subinterface but at the switch side you have VLAN ? | 08:13 |
| *** thorst_afk has quit IRC | 08:14 | |
| bhujay | this could be a problem forwarding traffic without appropriate VLAN tagging . I suggest you create a router in the external net , note down the routers interface ip in 1.20 net , connect external net in a switch without VLAN and from external net try to ping the routers IP . If this goes through the problem is with VLAN . You may also change the switch port type to trunk port / 802.1q see if that helps | 08:19 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-specs master: (WIP) Blueprint for Congress integration with OSA. Congress is a policy framework for OpenStack. https://review.openstack.org/499066 | 08:19 |
| *** maybebuggy has joined #openstack-ansible | 08:20 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-specs master: (WIP) Blueprint for Congress integration with OSA. Congress is a policy framework for OpenStack. https://review.openstack.org/499066 | 08:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499079 | 08:37 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Enable verbose logging when fetching roles via git-clone https://review.openstack.org/498788 | 08:38 |
| *** vnogin has joined #openstack-ansible | 08:41 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499079 | 08:43 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Enable verbose logging when fetching roles via git-clone https://review.openstack.org/498788 | 08:43 |
| *** andreas_s has joined #openstack-ansible | 08:44 | |
| *** andreas_s has quit IRC | 08:44 | |
| *** andreas_s has joined #openstack-ansible | 08:45 | |
| *** andreas_s has quit IRC | 08:46 | |
| *** andreas_s has joined #openstack-ansible | 08:47 | |
| *** andreas_s has quit IRC | 08:47 | |
| openstackgerrit | Chris Beukers proposed openstack/openstack-ansible-repo_server master: Create venvs directory when repo server container is created https://review.openstack.org/499083 | 08:49 |
| openstackgerrit | Chris Beukers proposed openstack/openstack-ansible-repo_build master: Removes creation of venvs directory during repo build https://review.openstack.org/499084 | 08:59 |
| *** drifterza has joined #openstack-ansible | 09:03 | |
| coolkil | what is neccesary to get this accepted? https://review.openstack.org/#/c/479844/ besides removing the wip part? it still needs a change to https://github.com/openstack/openstack-ansible/blob/master/group_vars/all/nova.yml#L31-L33 but that is a different repository | 09:05 |
| *** sxc731 has quit IRC | 09:07 | |
| *** andreas_s has joined #openstack-ansible | 09:10 | |
| *** thorst_afk has joined #openstack-ansible | 09:10 | |
| *** thorst_afk has quit IRC | 09:14 | |
| odyssey4me | coolkil no-one will touch it until the WIP is removed | 09:15 |
| odyssey4me | and you can submit the other patch with 'Depends-On: <change ID>' in the commit message to have them tested together | 09:15 |
| *** sxc731 has joined #openstack-ansible | 09:17 | |
| *** electrofelix has joined #openstack-ansible | 09:23 | |
| coolkil | ah thnx odyssey4me wil run the other change localy first if it works here i will submit it | 09:26 |
| *** a0x3e8 has joined #openstack-ansible | 09:26 | |
| *** hw_wutianwei has quit IRC | 09:37 | |
| *** yifei has quit IRC | 09:39 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Allow overrides of bootstrap ansible script https://review.openstack.org/499102 | 10:11 |
| *** thorst_afk has joined #openstack-ansible | 10:11 | |
| *** thorst_afk has quit IRC | 10:15 | |
| *** askb has quit IRC | 10:15 | |
| *** m3rl1n has joined #openstack-ansible | 10:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499079 | 10:26 |
| *** dave-mccowan has joined #openstack-ansible | 10:29 | |
| *** stuartgr has joined #openstack-ansible | 10:29 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Enable verbose logging when fetching roles via git-clone https://review.openstack.org/498788 | 10:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499079 | 10:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Enable verbose logging when fetching roles via git-clone https://review.openstack.org/498788 | 10:33 |
| *** sxc731 has quit IRC | 10:43 | |
| odyssey4me | evrardjp andymccr alright, it looks like https://review.openstack.org/499079 is actually resolving the zuul-cloner on pike issue | 10:48 |
| odyssey4me | once that merges I'll forward port to master if that's ok? | 10:48 |
| *** oneswig has joined #openstack-ansible | 11:00 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Removes creation of venvs directory during repo build https://review.openstack.org/499084 | 11:08 |
| *** thorst_afk has joined #openstack-ansible | 11:12 | |
| *** ivveh has quit IRC | 11:12 | |
| *** thorst_afk has quit IRC | 11:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Fix venv build in when mariadb is installed https://review.openstack.org/498987 | 11:20 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Allow overrides of bootstrap ansible script https://review.openstack.org/499102 | 11:27 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Allow overrides of bootstrap ansible script https://review.openstack.org/499102 | 11:28 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Sync apt preferences during lxc host configuration https://review.openstack.org/498802 | 11:29 |
| coolkil | odyssey4me you made a change to the commit message how do i pull that change to my enviroment or is that not neccesary? | 11:34 |
| odyssey4me | coolkil git review -d <review number> | 11:35 |
| odyssey4me | that downloads the review in its current state | 11:35 |
| coolkil | thnx!! | 11:35 |
| *** fxpester has joined #openstack-ansible | 11:36 | |
| openstackgerrit | Chris Beukers proposed openstack/openstack-ansible-repo_build master: Removes creation of venvs directory during repo build https://review.openstack.org/499084 | 11:39 |
| *** bauruine has quit IRC | 11:40 | |
| *** bauruine has joined #openstack-ansible | 11:41 | |
| *** rstarmer has quit IRC | 11:41 | |
| *** rstarmer has joined #openstack-ansible | 11:42 | |
| *** sxc731 has joined #openstack-ansible | 11:45 | |
| *** sxc731 has quit IRC | 11:51 | |
| *** gkadam_ has quit IRC | 11:54 | |
| *** thorst_afk has joined #openstack-ansible | 11:56 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Update role SHA's for online migration fixes https://review.openstack.org/499121 | 11:57 |
| *** yifei has joined #openstack-ansible | 11:57 | |
| *** yifei has quit IRC | 12:01 | |
| *** yifei has joined #openstack-ansible | 12:04 | |
| *** maybebuggy has quit IRC | 12:07 | |
| *** a0x3e8 has quit IRC | 12:07 | |
| *** ianychoi has quit IRC | 12:07 | |
| *** brad[] has quit IRC | 12:07 | |
| *** mrda has quit IRC | 12:07 | |
| *** Jeffrey4l has quit IRC | 12:07 | |
| *** logan- has quit IRC | 12:07 | |
| *** hamzy has quit IRC | 12:07 | |
| *** jrosser has quit IRC | 12:07 | |
| *** fdegir has quit IRC | 12:07 | |
| *** nyloc has quit IRC | 12:07 | |
| *** mrhillsman has quit IRC | 12:07 | |
| *** neillc has quit IRC | 12:07 | |
| *** ianychoi_ has joined #openstack-ansible | 12:07 | |
| *** jrosser has joined #openstack-ansible | 12:07 | |
| *** brad[]` has joined #openstack-ansible | 12:07 | |
| *** Jeffrey4l has joined #openstack-ansible | 12:07 | |
| *** hamzy has joined #openstack-ansible | 12:07 | |
| *** a0x3e8 has joined #openstack-ansible | 12:12 | |
| *** sxc731 has joined #openstack-ansible | 12:19 | |
| *** maybebuggy has joined #openstack-ansible | 12:22 | |
| *** mrda has joined #openstack-ansible | 12:22 | |
| *** logan- has joined #openstack-ansible | 12:22 | |
| *** fdegir has joined #openstack-ansible | 12:22 | |
| *** nyloc has joined #openstack-ansible | 12:22 | |
| *** mrhillsman has joined #openstack-ansible | 12:22 | |
| *** neillc has joined #openstack-ansible | 12:22 | |
| *** yifei has quit IRC | 12:27 | |
| *** huxinhui_ has quit IRC | 12:30 | |
| *** pester has joined #openstack-ansible | 12:35 | |
| *** hw_wutianwei has joined #openstack-ansible | 12:36 | |
| *** fxpester has quit IRC | 12:38 | |
| *** hachi has joined #openstack-ansible | 12:42 | |
| *** oneswig has quit IRC | 12:43 | |
| *** woodard has quit IRC | 12:51 | |
| *** woodard has joined #openstack-ansible | 12:52 | |
| *** japestinho has joined #openstack-ansible | 12:53 | |
| *** a0x3e8 has quit IRC | 12:54 | |
| odyssey4me | @andymccr mhayden FYI https://review.openstack.org/496671 is failing consistently with a functional test failure for swift | 12:54 |
| odyssey4me | mhayden https://review.openstack.org/497293 is failing consistently with an idempotence failure for debian | 12:54 |
| mhayden | i was seeing that too :/ | 12:54 |
| mhayden | odyssey4me: https://review.openstack.org/#/c/498468/ | 12:55 |
| mhayden | that will fix it | 12:55 |
| odyssey4me | we'll need https://review.openstack.org/499079 in to resolve pike integrated build failures | 12:55 |
| odyssey4me | ah yes, that one - need another vote :/ | 12:55 |
| *** a0x3e8 has joined #openstack-ansible | 12:55 | |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: Updated from OpenStack Ansible Tests https://review.openstack.org/497293 | 12:55 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/496671 | 12:57 |
| *** drifterza has quit IRC | 13:02 | |
| mhayden | andymccr: could i request a bit of gandering at https://review.openstack.org/498468 ? :) | 13:07 |
| mhayden | si vous plait ;) | 13:07 |
| andymccr | will take a look! | 13:07 |
| mhayden | andymccr: gracias | 13:08 |
| mgariepy | morning everyone. | 13:10 |
| *** esberglu has joined #openstack-ansible | 13:10 | |
| *** dxiri has quit IRC | 13:10 | |
| andymccr | morning mgariepy! | 13:10 |
| *** dxiri has joined #openstack-ansible | 13:11 | |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/newton: Install specific version of Erlang from ESL https://review.openstack.org/498951 | 13:11 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/newton: Pin erlang packages to version 19.3 https://review.openstack.org/498952 | 13:11 |
| mgariepy | how are you guys doing ? | 13:11 |
| *** andreas_s has quit IRC | 13:11 | |
| mgariepy | hmm mhayden seen this ?http://logs.openstack.org/79/499079/4/check/gate-openstack-ansible-openstack-ansible-ceph-centos-7-nv/80dbf0a/console.html#_2017-08-30_11_44_13_330366 | 13:12 |
| mgariepy | \n\nFailed:\n python2-urllib3.noarch 0:1.16-1.el7 | 13:12 |
| * mhayden toots | 13:12 | |
| mhayden | whoa | 13:12 |
| mhayden | error unpacking? | 13:12 |
| mhayden | has it happened more than once? i wonder if it's something broken in the mirror perhaps | 13:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Update role SHA's for online migration fixes https://review.openstack.org/499121 | 13:13 |
| mgariepy | yep happenned for other review in the centos-Ceph test | 13:14 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening stable/pike: Update vars and test tooling for Pike https://review.openstack.org/496626 | 13:16 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/496915 | 13:18 |
| mhayden | odyssey4me: okay, i think i have the ansible-hardening patches stacked properly | 13:19 |
| *** gouthamr has joined #openstack-ansible | 13:19 | |
| mhayden | pike may need the apparmor idempotency test backported to make those work consistently -- they will fail intermittently until that gets in | 13:19 |
| odyssey4me | then we wait :) | 13:20 |
| *** Apsu has joined #openstack-ansible | 13:24 | |
| *** chyka has joined #openstack-ansible | 13:25 | |
| *** chyka has quit IRC | 13:30 | |
| cloudnull | mornings | 13:40 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Allow overrides of bootstrap ansible script https://review.openstack.org/499102 | 13:40 |
| mgariepy | is the leap upgrade from K > N ready yet ? | 13:41 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Correct leap hook closing brackets https://review.openstack.org/499065 | 13:41 |
| cloudnull | +/- ready | 13:41 |
| cloudnull | I know folks have been testing with it recently | 13:41 |
| mgariepy | cloudnull, what's the overall process like ? | 13:42 |
| evrardjp | mgariepy: no | 13:43 |
| *** sxc731 has quit IRC | 13:44 | |
| cloudnull | assuming you have a happy health, stable kilo cloud, you would clone the ops repo, cd into the leap dir and run the scripts. | 13:44 |
| *** lbragstad has joined #openstack-ansible | 13:44 | |
| cloudnull | it would migrate your dbs, then redeploy on newton | 13:44 |
| *** mrch has joined #openstack-ansible | 13:44 | |
| cloudnull | it aggregates most of our upgrade utilities and adds a few more to the mix | 13:45 |
| cloudnull | evrardjp: It more or less works when deploying with OSA. | 13:46 |
| cloudnull | mgariepy: if you have a test environment or the hardware to do a multi-node-aio you can try it out | 13:46 |
| cloudnull | which I'd recommend using your configs | 13:46 |
| *** chhavi has quit IRC | 13:47 | |
| *** Donckers has joined #openstack-ansible | 13:47 | |
| mgariepy | well, I could install a testbed to test it. | 13:47 |
| cloudnull | I think jmccrory has done some work with it too ? | 13:47 |
| mgariepy | I probably will.. :) | 13:47 |
| cloudnull | but it'd be great to get additional feedback on it | 13:47 |
| *** lucasxu has joined #openstack-ansible | 13:48 | |
| odyssey4me | the first challenge is to deploy kilo :/ | 13:49 |
| odyssey4me | unfortunately that will require a fork to patch up a few things to get it working | 13:49 |
| evrardjp | more or less. | 13:51 |
| cloudnull | https://github.com/rcbops/rpc-maas/blob/master/tests/aio-create.sh -- odyssey4me mgariepy -- pass in IRR_CONTEXT=kilo and it should go | 13:52 |
| cloudnull | that works fairly reliably for an AIO but if you need kilo those couple of additions should make ti work | 13:53 |
| mgariepy | cool. | 13:54 |
| mgariepy | it will probably be a month or 2 before I start digging deeply into this. | 13:55 |
| cloudnull | cool | 13:55 |
| mgariepy | I hop it won't be too much trouble :D | 13:56 |
| mgariepy | at least once it's done I won't have to maintain that kitten anymore :D | 13:57 |
| cloudnull | odyssey4me: http://logs.openstack.org/42/497742/18/check/gate-openstack-ansible-openstack-ansible-aio-ubuntu-xenial/0af9d0f/console.html#_2017-08-30_04_04_14_453839 - is that something we're aware of ? | 13:58 |
| *** hw_wutianwei has quit IRC | 13:59 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Run gate playbooks in parallel https://review.openstack.org/497742 | 13:59 |
| cloudnull | ah https://review.openstack.org/#/c/499071/ | 14:00 |
| cloudnull | odyssey4me: ^ is that all we need? | 14:00 |
| *** mrch has quit IRC | 14:00 | |
| odyssey4me | cloudnull that's the role fix, there's also a playbook fix | 14:01 |
| *** a0x3e8 has quit IRC | 14:01 | |
| cloudnull | ok | 14:01 |
| odyssey4me | https://review.openstack.org/#/q/topic:rolling-upgrades+status:open | 14:01 |
| odyssey4me | also need this to fix pike https://review.openstack.org/499079 | 14:02 |
| *** marst has joined #openstack-ansible | 14:02 | |
| dmsimard | andymccr, cloudnull, mgariepy, odyssey4me, evrardjp: heads up, the -testing repos for RDO will be fully populated in preparation for imminent release today | 14:02 |
| cloudnull | coool | 14:03 |
| *** chhavi has joined #openstack-ansible | 14:04 | |
| evrardjp | cool indeed | 14:04 |
| *** vakuznet has joined #openstack-ansible | 14:04 | |
| *** pcaruana has quit IRC | 14:06 | |
| *** rstarmer has quit IRC | 14:07 | |
| neith | how are the directory /var/www/repo/pools/ubuntu-14.04-x86_64 is populated? | 14:12 |
| neith | *is | 14:12 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: redeploy should be done in the proper folder https://review.openstack.org/499164 | 14:13 |
| *** aludwar has quit IRC | 14:15 | |
| *** woodard has quit IRC | 14:20 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Added a common tag to the common tasks https://review.openstack.org/498992 | 14:22 |
| *** pester has quit IRC | 14:22 | |
| MasterofJOKers | neith, the role repo_build should be responsible for that. look into /etc/ansible/roles/repo_build on your deploy host | 14:23 |
| *** fxpester has joined #openstack-ansible | 14:24 | |
| *** weezS has joined #openstack-ansible | 14:27 | |
| neith | MasterofJOKers: thks | 14:28 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Correct cinder online migrations command https://review.openstack.org/499071 | 14:32 |
| *** hachi has quit IRC | 14:33 | |
| *** hachi has joined #openstack-ansible | 14:33 | |
| jafeha | hey everbody. we have deployed pike rc1 with haproxy as external load balancer and set the address to the fqdn (externally resolved). we're now having an issue that the external lb address is not excluded from the lxc containers list. i'm not sure if this should be considered a bug or just a non-intended setup. | 14:33 |
| jafeha | for the moment we're asking ourselfs: how do we get out of this mess? :) (destroy / rebuild the container and blacklist the ip wrongly used ip) | 14:33 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: redeploy should be done in the proper folder https://review.openstack.org/499164 | 14:34 |
| neith | the weird thing is that when I rebuild the repo containers it fails only on my ubuntu 16.04 host not on ubuntu 14.04 with a No such file or directory: '/var/www/repo/pools/ubuntu-14.04-x86_64/gnocchiclient/gnocchiclient-2.8.2-py2.py3-none-any.whl error. any idea? | 14:36 |
| *** galstrom_zzz is now known as galstrom | 14:37 | |
| odyssey4me | looking for reviews for https://review.openstack.org/499121 please | 14:38 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/pike: Correct cinder online migrations command https://review.openstack.org/499072 | 14:39 |
| *** kjw3 has joined #openstack-ansible | 14:39 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/ocata: Correct cinder online migrations command https://review.openstack.org/499073 | 14:39 |
| *** drifterza has joined #openstack-ansible | 14:40 | |
| *** hachi has quit IRC | 14:43 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: Create venvs directory when repo server container is created https://review.openstack.org/499083 | 14:50 |
| firebat | Hey guys quick dumbo question: in the ovs documentation physnet1 represents the physical interface the bridge and OVS port sit on top of? | 14:52 |
| firebat | Additionally leveraging vlan inside openstack to separate tenant private traffic can those ranges be for openstack only and not exist outside if the ports are trunked? | 14:54 |
| *** hachi has joined #openstack-ansible | 14:54 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_tacker master: Update roles & vars for stable/pike https://review.openstack.org/499183 | 14:58 |
| *** dxiri has quit IRC | 14:59 | |
| *** woodard has joined #openstack-ansible | 15:03 | |
| *** vnogin has quit IRC | 15:06 | |
| *** vnogin has joined #openstack-ansible | 15:07 | |
| *** thegreenhundred has joined #openstack-ansible | 15:10 | |
| *** chyka has joined #openstack-ansible | 15:16 | |
| *** chyka has quit IRC | 15:17 | |
| lbragstad | mhayden: you run fedora 26 still, right? | 15:17 |
| *** Oku_OS is now known as Oku_OS-away | 15:18 | |
| *** hachi has quit IRC | 15:19 | |
| *** hachi has joined #openstack-ansible | 15:19 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Bump compute RAM to 8192MB https://review.openstack.org/498917 | 15:19 |
| *** chyka has joined #openstack-ansible | 15:19 | |
| mgariepy | anyone having some issue with galera when performing minor upgrade | 15:20 |
| *** woodard has quit IRC | 15:21 | |
| cloudnull | mgariepy: which release ? | 15:21 |
| mgariepy | newton | 15:21 |
| cloudnull | I've not had issues recently | 15:21 |
| cloudnull | what are you seeing ? | 15:21 |
| *** woodard has joined #openstack-ansible | 15:21 | |
| mgariepy | from 14.2.4 to .8 | 15:21 |
| mgariepy | if I do a dist-upgrade on 1 galera node | 15:22 |
| cloudnull | is it failing to install the new packages or just sart post upgrade? | 15:22 |
| mgariepy | it fail to start mysql | 15:22 |
| cloudnull | is it a wsrep issue? | 15:22 |
| mgariepy | need to install percona-xtrabackup | 15:22 |
| mgariepy | first. | 15:22 |
| mgariepy | if I install this pkg then do the dist-upgrade it passes without issue. | 15:23 |
| mgariepy | when installing the percona-extrabackup pkg it removes : percona-xtrabackup-22 | 15:25 |
| *** aludwar has joined #openstack-ansible | 15:26 | |
| *** marc_ab has joined #openstack-ansible | 15:26 | |
| *** udesale__ has quit IRC | 15:26 | |
| *** coolkil has quit IRC | 15:31 | |
| *** marc_ab has quit IRC | 15:31 | |
| *** thegreenhundred has quit IRC | 15:48 | |
| *** manjaroi3 has joined #openstack-ansible | 15:48 | |
| mhayden | lbragstad: yessir | 15:50 |
| *** hachi has quit IRC | 15:52 | |
| *** gouthamr has quit IRC | 15:53 | |
| *** pcaruana has joined #openstack-ansible | 15:54 | |
| lbragstad | mhayden: do you run containers locally at all? | 15:55 |
| lbragstad | using lxc? | 15:55 |
| mhayden | i usually go with systemd-nspawn | 15:55 |
| lbragstad | aha - got it | 15:55 |
| *** manjaroi3 has quit IRC | 15:56 | |
| *** thegreenhundred has joined #openstack-ansible | 15:56 | |
| *** thegreenhundred has quit IRC | 15:58 | |
| *** thegreenhundred has joined #openstack-ansible | 15:58 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: L3 Support for ODL deployments https://review.openstack.org/497458 | 15:58 |
| *** dxiri has joined #openstack-ansible | 15:59 | |
| *** esberglu has quit IRC | 16:04 | |
| *** esberglu has joined #openstack-ansible | 16:05 | |
| *** esberglu has quit IRC | 16:06 | |
| *** marc_ab has joined #openstack-ansible | 16:11 | |
| *** japestinho has quit IRC | 16:11 | |
| *** dxiri has quit IRC | 16:15 | |
| *** m3rl1n has quit IRC | 16:16 | |
| *** esberglu has joined #openstack-ansible | 16:18 | |
| *** gouthamr has joined #openstack-ansible | 16:19 | |
| tasker | ok -- once I did the OSA newton upgrade I am having a heck of a time fixing these SSL problems. changing the "public" endpoints to SSL just seems to make it worse. horizon can't connect to neutron endpoint. neutron CLI works if I give it "--insecure". the openstack CLI seems to work just fine. | 16:24 |
| tasker | searching out SSL on the bug launchpad doesn't show much help. it seemed that setting the public endpoints to https as recommended by the Newton release notes didn't help. | 16:24 |
| tasker | has anyone else in here wrestled with this? | 16:24 |
| tasker | if I use the "admin" or "internal" endpoints in curl calls, it's OK. | 16:27 |
| tasker | of course, this is all probably going to be a fault of my configuration because I have two different IPs: one for public and one for andmin and internal. | 16:27 |
| tasker | if I use "--insecure / -k" on a curl poke to the public endpoint it works. otherwise it doesn't. | 16:29 |
| *** germs has joined #openstack-ansible | 16:29 | |
| tasker | ahh . well, this is helping (but not optimal): OS_INTERFACE="admin" | 16:31 |
| tasker | so it looks like things are pointing to the "public" url and SSL. | 16:31 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build master: Removes creation of venvs directory during repo build https://review.openstack.org/499084 | 16:31 |
| *** markvoelker has joined #openstack-ansible | 16:32 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible master: [DOC] Check Galera/Rabbitmq versions during cycle https://review.openstack.org/499226 | 16:34 |
| *** weezS has quit IRC | 16:37 | |
| tasker | an https poke direct to the internal IP of the service (not the haproxy endpoint) does not work: "SSL23_GET_SERVER_HELLO:unknown protocol". using http works just fine. | 16:40 |
| *** zcourts_ has joined #openstack-ansible | 16:40 | |
| odyssey4me | tasker are you using a real cert, or a self-signed cert, or a cert from some sort of internal CA? | 16:42 |
| tasker | self-signed. i was under the impression that was an OSA thing. | 16:42 |
| odyssey4me | oh no - the self-signed is for demo purposes | 16:42 |
| tasker | . ( | 16:43 |
| odyssey4me | you should ideally use a real cert of some sort | 16:43 |
| *** vikrant has joined #openstack-ansible | 16:43 | |
| odyssey4me | you can use self-signed if you really want to, but then your client needs to be able to verify it | 16:43 |
| odyssey4me | if it can't, then the client has to use --insecure | 16:43 |
| tasker | which echos what I'm seeing. | 16:44 |
| mhayden | cloudnull: would you have a moment to gander at https://review.openstack.org/498468 today? :) | 16:44 |
| tasker | thanks for the clarification, odyssey4me. | 16:44 |
| odyssey4me | tasker to allow horizon to work with a self-signed cert, this var needs to be set to disable ssl verification: https://github.com/openstack/openstack-ansible-os_horizon/blob/master/templates/horizon_local_settings.py.j2#L210 | 16:44 |
| odyssey4me | *but* horizon should be set to use the *internal* keystone interface by default, which by default is non-SSL | 16:45 |
| odyssey4me | take a peek at the value of this in the horizon container: https://github.com/openstack/openstack-ansible-os_horizon/blob/master/templates/horizon_local_settings.py.j2#L169 | 16:45 |
| odyssey4me | is that pointing at the internal or public endpoint? | 16:45 |
| tasker | that is the internal / admin endpoint in my cluster. | 16:46 |
| odyssey4me | ok, but it's getting redirected to the public one? | 16:47 |
| tasker | and OPENSTACK_SSL_NO_VERIFY = False. | 16:47 |
| *** vnogin has quit IRC | 16:47 | |
| tasker | keystone works. it's the various other endpoints: neutron, nova, etc. | 16:47 |
| tasker | I can log into horizon just fine and see quite a bit of info. | 16:48 |
| odyssey4me | what's the value of https://github.com/openstack/openstack-ansible-os_horizon/blob/master/templates/horizon_local_settings.py.j2#L376 ? | 16:48 |
| tasker | "internalURL" | 16:48 |
| *** jwitko has quit IRC | 16:48 | |
| odyssey4me | ok, so horizon is then set to use the internal endpoint when it can | 16:48 |
| odyssey4me | the way the clients work is that they will speak to keystone, get the service catalog, then use the public endpoint by default - unless you've told it to use something else | 16:49 |
| odyssey4me | and only an admin can use endpoints that aren't public | 16:49 |
| odyssey4me | so when you set 'OS_INTERFACE' in your env vars, you're telling it to use another interface | 16:50 |
| *** vikrant has quit IRC | 16:50 | |
| odyssey4me | so basically you've got to decide whether you're going to replace the self-signed certs with real ones, whether you're going to keep the self-signed cert and alias your commands (as we do in the utility container), or whether you disable ssl for public endpoints | 16:51 |
| *** zcourts has joined #openstack-ansible | 16:54 | |
| *** vikrant has joined #openstack-ansible | 16:54 | |
| *** gkadam_ has joined #openstack-ansible | 16:57 | |
| *** zcourts_ has quit IRC | 16:57 | |
| *** markvoelker has quit IRC | 17:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build stable/pike: Reduce package list for CentOS https://review.openstack.org/496741 | 17:09 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: Update upgrade role for Pike https://review.openstack.org/496679 | 17:09 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron master: Update upgrade role for Pike https://review.openstack.org/496680 | 17:09 |
| *** mbuil has quit IRC | 17:10 | |
| *** strigazi has joined #openstack-ansible | 17:11 | |
| admin0 | hi all | 17:12 |
| admin0 | we have pike stable ? | 17:14 |
| *** strigazi has quit IRC | 17:14 | |
| admin0 | this soon :) | 17:14 |
| admin0 | awesome | 17:14 |
| odyssey4me | admin0 not released yet, RC for now | 17:15 |
| odyssey4me | note https://docs.openstack.org/openstack-ansible/latest/#pike-release-candidate | 17:15 |
| admin0 | how does it handle multi cells ? can i give 3 controllers for 1 cell and only 1 controller for another ? | 17:15 |
| odyssey4me | no idea how cells work | 17:15 |
| odyssey4me | not even sure whether it's ready from a nova standpoint just yet | 17:16 |
| admin0 | its marketed as its ready with nova cells v2 | 17:16 |
| odyssey4me | cells v2 has been there since ocata | 17:17 |
| odyssey4me | but don't know how feature complete it is | 17:17 |
| odyssey4me | as with most things, I expect it depends on your use case | 17:17 |
| odyssey4me | we, for now, enroll all computes into the first cell | 17:17 |
| odyssey4me | I don't think we've done anything to handle more cells at this point. Someone would have to figure out how it's supposed to work and build in the capability to extend to more cells. | 17:18 |
| *** vnogin has joined #openstack-ansible | 17:18 | |
| admin0 | my use case is this.. i have 2 tenants using 1 cloud env inside office .. one is very stable .. has production stuff runining .. other tenant has demo/training running .. so they create like 50-100 machines every 5 mins .. do some demo, destroy it and repeat the cycle continously .. so with multi cells, i was hoping to isolate the db/rabbit chatter to a different cell dedicted to this tenant | 17:19 |
| *** zcourts_ has joined #openstack-ansible | 17:19 | |
| *** zcourts has quit IRC | 17:22 | |
| *** vnogin has quit IRC | 17:22 | |
| *** dxiri has joined #openstack-ansible | 17:33 | |
| bhujay | odyssey4me : As per openstack security guide both internal and external endpoints should be ssl enabled . I am using self signed certs . Will it not be possible through OSA configs ? | 17:36 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Correct cinder online migrations command https://review.openstack.org/499068 | 17:36 |
| SamYaple | bhujay: let me just inject here... internal endpoints being ssl'd may be a good idea, but you still have services like memcached that has *no* auth/security/etc and can be accessed and modified by anyone | 17:37 |
| odyssey4me | bhujay self-signed certs give you no security whatsoever, easy to man-in-the-middle, and therefore pointless | 17:37 |
| openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Correct cinder online migrations command https://review.openstack.org/499069 | 17:37 |
| SamYaple | that requires network security to ensure that data isn't red in the clear | 17:37 |
| odyssey4me | bhujay however, external - absolutely, internal - probably with some var tweaks and requiring the shipping of your CA cert all over the place | 17:38 |
| SamYaple | odyssey4me: thats not entirely true, a cert of some kind would be needed to encrypt the data. if the in flight traffic was teh concern, self-signed helps there | 17:39 |
| odyssey4me | SamYaple fair enough | 17:40 |
| bhujay | Self signed is not a constraint for me , we can get one CA cert once I establish all SSL communication. But What Sam says is worrysome , We will se that. First if you can help me understand 1) whether all the internal comms will work properly once SSL is enabled and 2) its possible to enable through OSA | 17:41 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499079 | 17:41 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Enable verbose logging when fetching roles via git-clone https://review.openstack.org/498788 | 17:41 |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Update role SHA's for online migration fixes https://review.openstack.org/499121 | 17:41 |
| bhujay | Yeah Sam that's the point our security team first look into that the data is not passing cleartext | 17:42 |
| odyssey4me | bhujay you'll have to try it to find out - in our history of working with SSL we have found the support from a client standpoint to be a bit hit and miss | 17:42 |
| odyssey4me | sometimes the service middleware doesn't work with it either | 17:42 |
| SamYaple | bhujay: there is no getting around the memcache in the clear thing. that said, *most* of the data can be encrypted, but not all of it (nova-consoleauth, im looking at you) | 17:42 |
| odyssey4me | these have nothing to do with deployment tooling, and everything to do with the upstream services/libraries | 17:43 |
| SamYaple | you can technically run without memcached but with huge performance losses (and not HA for some services) | 17:43 |
| bhujay | I will try this in few weeks , building a separate env just for this test so that my current installation works . I had lot of issues in my initial installation and then I followed osa docs where it suggests disabling SSL for self signed certs . I had hope that I will be able to overcome once I learn a little more about OSA... lets see | 17:47 |
| bhujay | As an alternate approach ... | 17:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Update role SHA's for online migration fixes https://review.openstack.org/499244 | 17:49 |
| *** DanyC has joined #openstack-ansible | 17:49 | |
| odyssey4me | bhujay as with the discussion with tasker above - self-signed certs are not great unless you know what you're in for and happy with that | 17:50 |
| bhujay | In case internal SSL really fails , I have to segregate the network into multiple segments with firewalls in between such as haproxy , api , db/mq and compute . I was wondering if using the openstack-user_config , containers_network section I should be able to do that , any suggestions ? | 17:50 |
| bhujay | point noted about self signed certs | 17:51 |
| bhujay | as of now we define one ip range for br-mgmt but shd it not be possible to provide multiple range of IP's ? | 17:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Update role SHA's for online migration fixes https://review.openstack.org/499246 | 17:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Correct zuul-cloner cloning of pinned SHA's https://review.openstack.org/499247 | 17:53 |
| odyssey4me | bhujay you may be interested to take a look at the routed environment example configs: https://docs.openstack.org/project-deploy-guide/openstack-ansible/pike/app-config-pod.html | 17:57 |
| jrosser | we do it more simply, the haproxy public vip is on a completely different net / interface to the internal one | 17:58 |
| jrosser | so external things can never ever get at the internal vip | 17:59 |
| jrosser | and there is no connection between the external net and the mgmt net, nice and simple | 17:59 |
| *** gkadam_ has quit IRC | 18:01 | |
| *** electrofelix has quit IRC | 18:01 | |
| bhujay | Many thanks odyssey, looks like a great match with what I was thinking . Need to study this tomorrow . But it is only for pike or shd be applicable for stable/ocata as well ? | 18:02 |
| *** poopcat has quit IRC | 18:03 | |
| odyssey4me | bhujay not sure personally - haven't looked in great detail... OSA is infinitely flexible for the most part if you have the skills and determination to do it... | 18:03 |
| odyssey4me | jmccrory wrote up that documentation, so he might be able to shed some light on what it'll work with | 18:04 |
| odyssey4me | otherwise, yeah jrosser and others have implemented alternative ways of doing things :) | 18:04 |
| *** weezS has joined #openstack-ansible | 18:04 | |
| jrosser | i was also scratching my head on the SSL stuff :) self signed is of course an excellent way to ensure that only trusted internal things can talk to internal endpoints | 18:05 |
| jrosser | so some kind of self-signed internal / public-signed external setup would be interesting | 18:05 |
| *** vikrant has quit IRC | 18:05 | |
| *** poopcat has joined #openstack-ansible | 18:05 | |
| bhujay | I thought it was lot to do with the dynamic inventory management.py which actually creates the network address for the containers | 18:06 |
| jrosser | but understood on tool trouble with self-signed | 18:06 |
| odyssey4me | jrosser yeah, I would think an internal CA would be better | 18:06 |
| jrosser | something very much like that is done with radius servers in wpa-enterprise wifi | 18:07 |
| bhujay | jrosser : I have done that too , haproxy is on a separate host and network , however that host itself is a bridge between two network | 18:07 |
| jrosser | you just cant use a public cert there | 18:07 |
| odyssey4me | if someone implemented a playbook to use a role not maintained by this community (ideally) to setup a CA which can do all the right things then that would be far better | 18:07 |
| bhujay | therefore it is best to ensure all internal communicatiosn are also encrypted . This has been very well described in openstack security guide how the domains overlaps in an cloud environment | 18:08 |
| odyssey4me | I'm out for the night - cheers all! | 18:09 |
| jrosser | i'm not sure i've seen any special treatment of the haproxy host to prevent it ip forwarding between networks | 18:09 |
| *** markvoelker has joined #openstack-ansible | 18:10 | |
| bhujay | thanks for the doc link and suggestions on ssl odyssey , have a great night out :) | 18:10 |
| *** albertcard1 has joined #openstack-ansible | 18:11 | |
| *** poopcat has quit IRC | 18:14 | |
| *** chhavi has quit IRC | 18:17 | |
| *** dxiri has quit IRC | 18:26 | |
| *** gouthamr has quit IRC | 18:28 | |
| *** jwitko has joined #openstack-ansible | 18:33 | |
| *** stuartgr has quit IRC | 18:34 | |
| firebat | My god guys I'm almost done deploying I just need to get networking to work :) | 18:36 |
| firebat | Does anyone have a good link on how to debug the networking services? | 18:39 |
| *** markvoelker has quit IRC | 18:43 | |
| *** gouthamr has joined #openstack-ansible | 18:44 | |
| *** dxiri has joined #openstack-ansible | 18:48 | |
| *** weezS_ has joined #openstack-ansible | 19:00 | |
| *** weezS has quit IRC | 19:01 | |
| *** weezS_ is now known as weezS | 19:01 | |
| admin0 | firebat: whats wrong with networking ? | 19:02 |
| *** admin0 has left #openstack-ansible | 19:03 | |
| *** admin0 has joined #openstack-ansible | 19:03 | |
| firebat | admin0: Trying to figure that out now | 19:03 |
| *** markvoelker has joined #openstack-ansible | 19:06 | |
| mhayden | andymccr: for what it's worth, i have a patch in there for upgrading to RabbitMQ 3.6.11 -> https://review.openstack.org/#/c/496801/ | 19:21 |
| openstackgerrit | Merged openstack/openstack-ansible master: [DOC] Check Galera/Rabbitmq versions during cycle https://review.openstack.org/499226 | 19:29 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: Add release note for F26 support https://review.openstack.org/498847 | 19:30 |
| *** markvoelker has quit IRC | 19:37 | |
| *** markvoelker has joined #openstack-ansible | 19:38 | |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: Add release note for F26 support https://review.openstack.org/498847 | 19:40 |
| tasker | here's a fun one, if you're following my SSL tribulations: if I set "OPENSTACK_ENDPOINT_TYPE='adminURL'" (originally 'internalURL') in the horizon settings (and restart) it works just fine. the fun piece is that 'internal' and 'admin' are the same IP/port at the same haproxy. | 19:49 |
| tasker | it does nothing for my CLI utilities or the actual problem, but at least my cluster is not "broken". | 19:52 |
| admin0 | tasker: whats the issue with SSL ( was not following ) | 19:52 |
| admin0 | you need to use 2 different IPs and URLs which seperates it properly | 19:53 |
| admin0 | i have NFS and CEPH cinder ,, so the is_metal is set to false .. now I want to add a physical linux LVM to the mix .. is it posible to say this is metal for just this storage host ? | 19:54 |
| *** markvoelker has quit IRC | 19:56 | |
| *** woodard_ has joined #openstack-ansible | 20:01 | |
| tasker | omg. nevermind. turns out that my horizon problems were because my internal neturon endpoint had 'https' instead of 'http'. | 20:01 |
| *** woodard has quit IRC | 20:05 | |
| *** woodard_ has quit IRC | 20:06 | |
| *** weezS has quit IRC | 20:13 | |
| *** weezS has joined #openstack-ansible | 20:13 | |
| *** vakuznet has quit IRC | 20:25 | |
| *** pcaruana has quit IRC | 20:40 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-galera_server master: Update to MariaDB 10.2 https://review.openstack.org/499300 | 20:41 |
| SamYaple | mhayden: i had issues when testing with mariadb 10.2, specfically during migrations. are you seeing anything liek that? | 20:43 |
| mhayden | not so far | 20:43 |
| SamYaple | i was testing against newton with 10.2, maybe that was it then. ill have to retest with master | 20:44 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-tests master: [TEST] Test with MariaDB 10.2 https://review.openstack.org/499317 | 20:45 |
| *** germs has quit IRC | 20:47 | |
| *** lucasxu has quit IRC | 20:49 | |
| *** gouthamr has quit IRC | 20:54 | |
| *** germs has joined #openstack-ansible | 20:54 | |
| *** tots has joined #openstack-ansible | 20:57 | |
| tots | hello . how would you remote scp some files from a container to localhost ? | 20:58 |
| cloudnull | admin0: yes. you can set things that way | 21:03 |
| cloudnull | tasker: good to go now? | 21:03 |
| cloudnull | tots: are the contianers using a file store or lvm ? | 21:03 |
| cloudnull | SamYaple: mgariepy: was having maria migration issues in newton . maybe related? | 21:04 |
| tots | its a fresh installation cloudnull and i want to copy something from the utility container , so i guess its file store | 21:04 |
| cloudnull | the filesystem for the container is located under /var/lib/lxc/$CONTAINER_NAME/rootfs/ | 21:05 |
| cloudnull | so from the host you can just cp the data elsewhere | 21:05 |
| cloudnull | or you can scp directly from the deploynode. | 21:05 |
| cloudnull | scp root@@CONTAINER_NAME:/things/and/stuff /place/locally/ | 21:06 |
| *** gouthamr has joined #openstack-ansible | 21:10 | |
| *** thorst_afk has quit IRC | 21:12 | |
| *** weezS has quit IRC | 21:16 | |
| *** DanyC_ has joined #openstack-ansible | 21:18 | |
| SamYaple | cloudnull: might be. but newton is about to go eol so ill just test against ocata/pike/master and see what happens | 21:20 |
| tasker | cloudnull: for the most part, yes. still need to do something about the "public" endpoints, but that can wait until later. | 21:20 |
| *** DanyC has quit IRC | 21:22 | |
| admin0 | cloudnull: is there a sample i can see ? | 21:22 |
| admin0 | where we do it on a per-container/host basis | 21:22 |
| cloudnull | admin0: I think you can set that in the openstack_user_config using host_vars | 21:24 |
| cloudnull | though you may need to mangle the env.d/cinder file to setup a new is_metal: false group | 21:25 |
| cloudnull | SamYaple: ++ | 21:25 |
| tasker | SamYaple: when is newton slated for EOL? | 21:28 |
| *** woodard has joined #openstack-ansible | 21:29 | |
| cloudnull | i think next month | 21:30 |
| tasker | fantastic! | 21:31 |
| tasker | does osa newton support enabling nova placement service? | 21:31 |
| *** tots has quit IRC | 21:31 | |
| cloudnull | yes i believe thats done by default ? | 21:32 |
| tasker | my nova-compute log states that there is "No authentication information found for placement API". this is not something I thought prior to doing the upgrade. | 21:33 |
| cloudnull | I am mistaken | 21:37 |
| cloudnull | in newton its not there. | 21:37 |
| cloudnull | in ocata is it | 21:37 |
| cloudnull | **it is | 21:37 |
| tasker | ok, thanks! | 21:37 |
| cloudnull | tasker: https://releases.openstack.org/ | 21:37 |
| cloudnull | newton is dead 17.10.11 | 21:37 |
| cloudnull | the branch will likely be around for a while after it goes EOL | 21:38 |
| cloudnull | like it was with mitaka | 21:38 |
| cloudnull | but not much will go back | 21:39 |
| DimGR | openstack-ansible has to be the best software humans ever produced period | 21:43 |
| *** esberglu has quit IRC | 21:45 | |
| *** askb has joined #openstack-ansible | 21:47 | |
| *** nomatics has joined #openstack-ansible | 21:47 | |
| *** kjw3 has quit IRC | 21:48 | |
| *** vnogin has joined #openstack-ansible | 21:48 | |
| cloudnull | lol... | 21:49 |
| *** vnogin has quit IRC | 21:53 | |
| nomatics | After a failed build, lxc_hosts : Prepare cached image setup commands is now failing for me. Packages are trying to be installed from a node via yum proxy and the connection is refused. | 21:54 |
| nomatics | Can someone provide insight on whats going or a place to start debuging? | 21:54 |
| cloudnull | nomatics: was the repo container up | 21:55 |
| cloudnull | and you've since deletred it ? | 21:55 |
| cloudnull | we use a local proxy through the repo containers maybe something that needs to be cleaned up ? | 21:56 |
| *** thegreenhundred has quit IRC | 21:56 | |
| nomatics | The play is in setup-hosts, which is before buildign the containers | 21:56 |
| nomatics | Before this I was running setup-openstack and made a typo in in openstack_user_config. I knew i made a mistake and killed the build, fixed the typo and re-rand all playbooks. Now the error is coming up. | 21:57 |
| cloudnull | nomatics: https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-tasks/package-cache-proxy.yml#L58-L83 | 21:57 |
| cloudnull | I think your bumping up against that | 21:58 |
| cloudnull | check the /etc/yum.conf file | 21:58 |
| cloudnull | and remove the local proxy line | 21:58 |
| cloudnull | or dnf.conf if you're using that | 21:58 |
| nomatics | Yea, I have it commneted out right now | 21:59 |
| nomatics | But I assume that it puts it back? | 21:59 |
| cloudnull | it will once the repo servers are back online | 21:59 |
| cloudnull | the cache will inherit a couple files from the host then build | 22:00 |
| cloudnull | maybe its inherited this one but without the commented proxy config ? | 22:01 |
| cloudnull | I would destory the cache and just rerun `openstack-ansible lxc-hosts.yml lxc-container-create.yml` | 22:01 |
| cloudnull | you should be able to delete the base cache container and remove the base image from machinectl | 22:02 |
| cloudnull | which would start 100% fresh | 22:02 |
| nomatics | I've destroyed all the containers and cleaned up as much as I know | 22:02 |
| cloudnull | the base lxc cache is under /var/cache/lxc | 22:02 |
| nomatics | lxc-ls shows no running containers and proxy is removed from the host yum.conf | 22:02 |
| cloudnull | there's likely a pre-constucted tar ball there | 22:03 |
| nomatics | Ah okah. Can I just nuke /var/cache/lxc/*? | 22:04 |
| cloudnull | yup | 22:04 |
| cloudnull | leave the lxc dir in place | 22:04 |
| cloudnull | but all the files within can go away | 22:04 |
| openstackgerrit | Merged openstack/openstack-ansible master: Correct cinder online migrations command https://review.openstack.org/499044 | 22:04 |
| cloudnull | also, just to nuke it from orbit run `machinectl list-images` and remove any images found there too | 22:05 |
| cloudnull | its the only way to be sure :) | 22:05 |
| nomatics | Thanks | 22:05 |
| cloudnull | https://media.giphy.com/media/ISAHN6dnrJHry/giphy.gif | 22:06 |
| nomatics | Lol thanks | 22:06 |
| nomatics | Is there any docs on cleaning up / full restart? | 22:11 |
| nomatics | So far I've been destroying all containers, cleaning up /etc/hosts and yum config, and now I know to purge the lxc cache. | 22:12 |
| *** dave-mccowan has quit IRC | 22:12 | |
| nomatics | Okay so nuking /var/cache/lxc didn't work ? | 22:16 |
| nomatics | I also deleted /etc/openstack_deploy and reboot strapped. :/ | 22:17 |
| *** thegreenhundred has joined #openstack-ansible | 22:18 | |
| tasker | i think I've just bumped into this: https://bugs.launchpad.net/openstack-ansible/+bug/1630953. however, I'm not fully following. there are no console endpoints registerd within keystone's catalog and all of the various nova config files do indeed have 'https' for "html5proxy_base_url", but i am still not getting a console. | 22:20 |
| openstack | Launchpad bug 1630953 in openstack-ansible "haproxy SSL, nova console is blank . " [Medium,Fix released] - Assigned to Jesse Pretorius (jesse-pretorius) | 22:20 |
| *** marst has quit IRC | 22:20 | |
| *** zcourts has joined #openstack-ansible | 22:20 | |
| tasker | is an endpoint suppsoed to be registered to keystone? where does the resultant url generated by "nova_spice_html5proxy_base_proto" end up? | 22:21 |
| *** DanyC_ has quit IRC | 22:21 | |
| tasker | I'm under the impression (from the release notes) that the OSA upgrade does not regenerate the catalog. | 22:21 |
| *** zcourts_ has quit IRC | 22:23 | |
| tasker | so, the setting given by nova_spice_html5proxy_base_url is used in /etc/ansible/roles/os_nova/templates/nova.conf.j2 and applied to the nova conf. it's not registerd with keystone. | 22:32 |
| tasker | anyone know what other roles or repos pulled in by OSA would make reference to nova_spice_html5proxy_base_url? | 22:33 |
| *** rstarmer has joined #openstack-ansible | 22:35 | |
| *** thegreenhundred has quit IRC | 22:39 | |
| cloudnull | I think there is, one sec | 22:40 |
| cloudnull | nomatics: ^ | 22:40 |
| *** galstrom is now known as galstrom_zzz | 22:44 | |
| cloudnull | tasker: nova_spice_html5proxy_base_url is only used in os_nova to the best of my knowledge | 22:44 |
| tasker | thanks. | 22:45 |
| cloudnull | nomatics: https://docs.openstack.org/openstack-ansible/pike/contributor/quickstart-aio.html#rebuilding-an-aio - thats what i remember | 22:47 |
| cloudnull | more geared toward the aio | 22:47 |
| nomatics | cloudnull: thanks, I think i found the problem. | 22:47 |
| cloudnull | but similar non-the-less | 22:47 |
| nomatics | I had a container named 'LXC_NAME' which I deleted. | 22:47 |
| nomatics | I thought it was just the output of the lxc-ls command, so I over looked it :/ Strange. | 22:48 |
| nomatics | After deleting that the play continues tho | 22:49 |
| cloudnull | cool. that is initialized container cache | 22:49 |
| cloudnull | so if it gets angry when creating the cache its likely going to be a point of frustration. | 22:50 |
| DimGR | cloudnull tasker speaking of spike , one time i had it working , another time not . Both deployments were using the exact same configs | 22:51 |
| tasker | did you ever figure out why it didn't work? | 22:53 |
| DimGR | never bothered :P | 22:53 |
| tasker | lol. | 22:53 |
| tasker | if I told my CIO that, I'd get the literal ax. | 22:54 |
| tasker | so, i kinda need to bother. <G> | 22:54 |
| DimGR | i checked just now and its a no go when in fact was working a month ago when i last checked it but i suspect the reason is that i did an upgrade to stable/ocata and somehow it decided not to work now | 22:55 |
| DimGR | tasker tell him to ssh instead :) | 22:55 |
| tasker | he'd reply with some variation of "it's for our customers". | 22:55 |
| DimGR | i think its network related | 22:56 |
| DimGR | i think | 22:56 |
| tasker | I'm outta here for the day. I'll be back to throw curses at this tomorrow. | 22:57 |
| tasker | cloudnull: thanks for your help! | 22:57 |
| cloudnull | get what to work ? | 22:58 |
| DimGR | spike | 22:58 |
| cloudnull | spice? | 22:58 |
| cloudnull | the console | 22:58 |
| cloudnull | ? | 22:58 |
| cloudnull | spice | 22:59 |
| DimGR | spice yes | 22:59 |
| cloudnull | oh ok . | 22:59 |
| cloudnull | tasker: the endpoint should be restered within the service catalog | 22:59 |
| * cloudnull going to look it up | 22:59 | |
| cloudnull | but its not the full url | 23:00 |
| DimGR | checking my instances and console is loading , they were loading before the minor upgrade | 23:00 |
| DimGR | is not loading * | 23:00 |
| cloudnull | it hits the nova proxy, then passing the traffic back to the console service | 23:00 |
| cloudnull | it is NOT loading after the upgrade? | 23:00 |
| DimGR | it is not | 23:01 |
| cloudnull | is it running the latest release? | 23:05 |
| cloudnull | can you login to the nova console contianer and see if the console service is running ? | 23:05 |
| DimGR | ok hold on | 23:06 |
| cloudnull | anything interesting in th elogs | 23:06 |
| DimGR | 01:03:31.915 13494 INFO nova.consoleauth.manager [req-83fd2027-523e-4e0e-8b94-d2300aa27936 f1c34db4c8f5488da1b8acc8e57e958c 04a241a0bfc24deebeb5ea53e2581aa6 - - -] Received Token: 2caaf684-6f9a-47d1-9862-0b29b22dfa87, {'instance_uuid': u'5b18c940-50f1-483e-b8fb-e6a22abe6c42', 'access_url': u'http://172.29.248.106:6082/spice_auto.html?token=2caaf684-6f9a-47d1-9862-0b29b22dfa87', 'token': u'2caaf684-6f9a-47d1-9862-0b | 23:08 |
| DimGR | 29b22dfa87', 'last_activity_at': 1504134211.914083, 'internal_access_path': None, 'console_type': u'spice-html5', 'host': u'172.29.248.106', 'port': u'5902'} | 23:08 |
| DimGR | it is running openstack/venvs/nova-15.1.8/bin/python /openstack/venvs/nova-15.1.8/bin/nova-consoleauth --log-file=/var/log/nova/nova-consoleauth.log | 23:09 |
| *** chyka has quit IRC | 23:10 | |
| DimGR | cloudnull ^^ | 23:12 |
| DimGR | i dont like the internal access path : none | 23:14 |
| cloudnull | can you curl that URL ? | 23:16 |
| cloudnull | are you using haproxy ? | 23:17 |
| DimGR | it is haproxy | 23:17 |
| cloudnull | `curl http://172.29.248.106:6082/spice_auto.html` | 23:17 |
| DimGR | curl: (52) Empty reply from server | 23:18 |
| *** weezS has joined #openstack-ansible | 23:18 | |
| cloudnull | `curl -D - http://172.29.248.106:6082` | 23:18 |
| DimGR | same exact error | 23:18 |
| cloudnull | whats the response code? | 23:19 |
| DimGR | curl: (52) Empty reply from server | 23:19 |
| cloudnull | and is haproxy reporting the backend is up ? | 23:19 |
| *** nomatics has quit IRC | 23:19 | |
| cloudnull | also is memcached up ? | 23:19 |
| DimGR | checking | 23:20 |
| DimGR | memcache is up and running | 23:21 |
| cloudnull | the proxy server and console auth work together and require the use of memcached to store / generate a token, so if any one of those things is down or otherwise unavailable the console will not work. | 23:21 |
| DimGR | haproxy logs are not showing any errors | 23:21 |
| cloudnull | hatop -s /var/run/haproxy.sock | 23:22 |
| cloudnull | ^ its an interactive console for haproxy and your backends | 23:22 |
| DimGR | http://paste.openstack.org/show/619955/ | 23:23 |
| DimGR | there you are , some down | 23:24 |
| DimGR | its haproxy.stats :) | 23:24 |
| cloudnull | so is the console auth down ? | 23:25 |
| DimGR | it says so | 23:26 |
| cloudnull | can you checkout or restart the console containers to verify ? | 23:26 |
| cloudnull | yyou might need to just restart the console service? | 23:27 |
| DimGR | give me the full ansible syntax please :) | 23:27 |
| DimGR | will it cause any running instances to go crazy ? | 23:27 |
| cloudnull | no. | 23:28 |
| cloudnull | it wont go crazy | 23:28 |
| cloudnull | ansible -m shell -a 'systemctl restart nova-consoleauth' nova_console | 23:28 |
| DimGR | ok they are restarted | 23:29 |
| DimGR | but still down when i view them with hatop | 23:30 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Fix LXC container start order https://review.openstack.org/497721 | 23:30 |
| cloudnull | still down ? | 23:36 |
| cloudnull | is the service running ? | 23:36 |
| cloudnull | maybe something interesting in the log? | 23:36 |
| DimGR | ansible -m shell -a 'systemctl status nova-consoleauth' nova_console and everything is running | 23:38 |
| *** rstarmer has quit IRC | 23:38 | |
| DimGR | checking logs | 23:38 |
| DimGR | http://paste.openstack.org/show/619956/ i see these cloudnull not sure if they are errors though | 23:40 |
| DimGR | warnings | 23:40 |
| cloudnull | yea | 23:40 |
| *** woodard has quit IRC | 23:41 | |
| *** woodard has joined #openstack-ansible | 23:42 | |
| *** markvoelker_ has joined #openstack-ansible | 23:45 | |
| DimGR | memcache service is running too | 23:45 |
| cloudnull | and hap is still saying its down | 23:46 |
| cloudnull | ? | 23:46 |
| DimGR | yes | 23:47 |
| cloudnull | if you curl the ip address of the container | 23:47 |
| cloudnull | on 6082 is it up ? | 23:47 |
| DimGR | nova console container ? | 23:47 |
| cloudnull | yes | 23:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker master: Update roles & vars for stable/pike https://review.openstack.org/499183 | 23:48 |
| *** woodard has quit IRC | 23:48 | |
| *** woodard has joined #openstack-ansible | 23:48 | |
| *** woodard has quit IRC | 23:48 | |
| *** woodard has joined #openstack-ansible | 23:48 | |
| DimGR | connection refused | 23:49 |
| *** woodard has quit IRC | 23:49 | |
| *** woodard has joined #openstack-ansible | 23:50 | |
| cloudnull | within the console container run `ss -ntlp` | 23:52 |
| cloudnull | is there anything listening on port 6082 ? | 23:53 |
| *** woodard has quit IRC | 23:54 | |
| DimGR | http://paste.openstack.org/show/619957/ | 23:54 |
| DimGR | nope | 23:54 |
| DimGR | could this be a bug ? | 23:57 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: [WIP] Further hyper-converge our container usage https://review.openstack.org/499396 | 23:58 |
| cloudnull | no, I wouldnt think so | 23:58 |
| cloudnull | can you restart the containers. | 23:58 |
| cloudnull | and see if it comes back ? | 23:58 |
| cloudnull | I have to run | 23:58 |
| cloudnull | but will be back online later. | 23:58 |
| DimGR | ok i will , i will update you tomorrow your time | 23:58 |
| cloudnull | ok | 23:58 |
| DimGR | its 3 am | 23:58 |
| DimGR | safe to restart them eh | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
