cloudnull | evenings | 00:04 |
---|---|---|
*** NWOutlier_ has joined #openstack-ansible | 00:08 | |
NWOutlier_ | anyone else having issues creating images 15.1.4 OSA w/Ocata. We get TypeError: Cannot read property 'data' of undefined - the research we did brought us to edit local_settings.py to include REST_API_REQUIRED_SETTINGS = ['OPENSTACK_HYPERVISOR_FEATURES', 'LAUNCH_INSTANCE_DEFAULTS', 'OPENSTACK_IMAGE_FORMATS'] unfortunately we still have the error. | 00:11 |
cloudnull | NWOutlier_: trying to upload a remote image? | 00:15 |
NWOutlier_ | correct - tahnks for bringing that up; we also made a change to another attribute to allow the remote images to be uploaded. I can find the file and AVP in a sec if you need me too | 00:18 |
logan- | is your horizon url different than your glance api endpoint | 00:18 |
NWOutlier_ | IMAGES_ALLOW_LOCATION .. let me check the horizon url and glance end point | 00:19 |
NWOutlier_ | glance and horizon are the same url (different ports) | 00:20 |
NWOutlier_ | here is one resource we've read through https://bugs.launchpad.net/horizon/+bug/1630833 | 00:21 |
openstack | Launchpad bug 1630833 in OpenStack Dashboard (Horizon) "Horizon image upload via URL fails with "TypeError: Cannot read property 'data' of undefined"" [Medium,Fix released] - Assigned to Gloria Gu (gloria-gu) | 00:21 |
NWOutlier_ | it describes our issue very well (e.g. exactly). | 00:23 |
cloudnull | NWOutlier_: you have the horizon local_settings.py setup like so https://bugs.launchpad.net/horizon/+bug/1630833/comments/5 ? | 00:23 |
openstack | Launchpad bug 1630833 in OpenStack Dashboard (Horizon) "Horizon image upload via URL fails with "TypeError: Cannot read property 'data' of undefined"" [Medium,Fix released] - Assigned to Gloria Gu (gloria-gu) | 00:23 |
NWOutlier_ | ok let me take a look at that Cloudnull | 00:24 |
cloudnull | and is "show_multiple_locations" enabled within glance? | 00:24 |
cloudnull | it looks like the bug was fixed in master. | 00:25 |
NWOutlier_ | I'll check my file for comment 5... what file would i look at for show_multiple_locations? | 00:25 |
cloudnull | i dont see a backport to stable at this time. | 00:25 |
cloudnull | glance-api i believe. | 00:26 |
cloudnull | idk if it will go but i initiated the cherry-pick https://review.openstack.org/#/c/474396/ | 00:26 |
NWOutlier_ | ok this will take me a few minutes... my first time working with lxc attach and navigate | 00:26 |
cloudnull | `ansible -m shell -a 'grep -rni show_multiple_locations /etc/glance/' glance_all` | 00:27 |
cloudnull | ^ distributed search :) | 00:27 |
NWOutlier_ | will that work within the glance container? | 00:28 |
cloudnull | it will . | 00:28 |
cloudnull | you can run any ansible command and target various containers/hosts. | 00:28 |
cloudnull | if you ever need to run commands within containers there's a whole mess of groups you can target | 00:29 |
NWOutlier_ | glance-api.conf:show_multiple_locations = True | 00:30 |
cloudnull | generally all of the openstack things have ${SERVICENAME}_all | 00:30 |
cloudnull | so glance_all, nova_all, neutron_all ect | 00:30 |
cloudnull | that's good. | 00:30 |
NWOutlier_ | ok let me check comment 5.. I'm sure we did it.. but need to check | 00:31 |
*** SerenaFeng has joined #openstack-ansible | 00:31 | |
*** SerenaFeng has quit IRC | 00:31 | |
*** SerenaFeng has joined #openstack-ansible | 00:32 | |
*** SerenaFeng has quit IRC | 00:33 | |
NWOutlier_ | ok - comment 5 was completed as part of our debugging and at some point we've restarted the container and httpd | 00:33 |
NWOutlier_ | seems we've been doing everything correctly - but still get the same error when creating image from ISO, QCOW2 - anything. | 00:34 |
logan- | look in your browser debug console for clues | 00:35 |
logan- | the error happens when it is contacting the wrong api endpoint or if the browser is blocking the request | 00:35 |
cloudnull | if you upload an image using the glance client do you get an error ? | 00:35 |
*** bauruine has quit IRC | 00:37 | |
NWOutlier_ | I haven't tried that I can. I might have to pick this up tomorrow, I'm runing about an hour late. Logan - you're right, we were getting a 403 error early this morning (forbidden) | 00:38 |
NWOutlier_ | we found that via the browser console and in the logs on the server side | 00:39 |
*** bauruine has joined #openstack-ansible | 00:42 | |
*** fyu has joined #openstack-ansible | 00:43 | |
*** manheim has joined #openstack-ansible | 00:56 | |
*** armaan has quit IRC | 00:56 | |
*** cuongnv has joined #openstack-ansible | 00:57 | |
*** dxiri has joined #openstack-ansible | 01:00 | |
*** dxiri has quit IRC | 01:01 | |
*** dxiri has joined #openstack-ansible | 01:02 | |
*** armaan has joined #openstack-ansible | 01:05 | |
*** dxiri has quit IRC | 01:06 | |
*** SerenaFeng has joined #openstack-ansible | 01:07 | |
*** SerenaFeng has quit IRC | 01:07 | |
*** cNilesh has joined #openstack-ansible | 01:11 | |
*** acormier has joined #openstack-ansible | 01:12 | |
*** lucasxu has joined #openstack-ansible | 01:19 | |
*** lucasxu has quit IRC | 01:21 | |
*** jamesdenton has joined #openstack-ansible | 01:23 | |
*** woodard_ has joined #openstack-ansible | 01:26 | |
*** woodard has quit IRC | 01:26 | |
*** manheim has quit IRC | 01:30 | |
*** piliman974 has joined #openstack-ansible | 01:32 | |
*** jamesdenton has quit IRC | 01:34 | |
*** bauruine has quit IRC | 01:39 | |
*** acormier has quit IRC | 01:44 | |
openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Fix default_floating_pool option https://review.openstack.org/473634 | 01:48 |
*** SerenaFeng has joined #openstack-ansible | 01:51 | |
*** bauruine has joined #openstack-ansible | 01:53 | |
*** SerenaFeng has quit IRC | 01:53 | |
*** SerenaFeng has joined #openstack-ansible | 02:00 | |
*** SerenaFeng has quit IRC | 02:03 | |
*** manheim has joined #openstack-ansible | 02:27 | |
*** markvoelker has quit IRC | 02:32 | |
*** markvoelker has joined #openstack-ansible | 02:33 | |
*** markvoelker has quit IRC | 02:37 | |
*** piliman974 has quit IRC | 02:43 | |
*** piliman974 has joined #openstack-ansible | 02:45 | |
*** woodard_ has quit IRC | 02:53 | |
*** woodard has joined #openstack-ansible | 02:54 | |
*** woodard has quit IRC | 02:58 | |
*** manheim has quit IRC | 03:00 | |
*** dxiri has joined #openstack-ansible | 03:03 | |
*** dxiri has quit IRC | 03:09 | |
*** bauruine has quit IRC | 03:28 | |
*** bauruine has joined #openstack-ansible | 03:31 | |
*** hybridpollo has quit IRC | 03:55 | |
*** SerenaFeng has joined #openstack-ansible | 04:05 | |
*** tobberydberg has joined #openstack-ansible | 04:09 | |
*** SerenaFeng has quit IRC | 04:11 | |
*** tobberydberg has quit IRC | 04:13 | |
*** bauruine has quit IRC | 04:13 | |
*** bauruine has joined #openstack-ansible | 04:14 | |
*** dxiri has joined #openstack-ansible | 04:15 | |
*** dxiri has quit IRC | 04:20 | |
*** gkadam has joined #openstack-ansible | 04:21 | |
*** gouthamr has quit IRC | 04:24 | |
*** markvoelker has joined #openstack-ansible | 04:33 | |
*** piliman974 has quit IRC | 04:40 | |
*** armaan has quit IRC | 04:50 | |
*** markvoelker has quit IRC | 05:07 | |
*** udesale has joined #openstack-ansible | 05:11 | |
*** Amit82 has joined #openstack-ansible | 05:14 | |
*** phalmos_ has quit IRC | 05:15 | |
*** phalmos has joined #openstack-ansible | 05:18 | |
Amit82 | Hi All | 05:19 |
Amit82 | Virtual Machine console is not accessible through horizon | 05:24 |
Amit82 | I had used "openstack_service_publicuri_proto: http" option in user_variables.yml | 05:25 |
Amit82 | Has anybody observed such behavior | 05:25 |
*** udesale has quit IRC | 05:30 | |
*** vnogin has joined #openstack-ansible | 05:34 | |
oanson | Amit82, I think you need to enable novnc with: nova_console_type: novnc (Or whatever type you're trying to use). But I don't have much experience in that. | 05:34 |
*** vnogin has quit IRC | 05:38 | |
*** udesale has joined #openstack-ansible | 05:49 | |
*** tobberydberg has joined #openstack-ansible | 05:50 | |
*** markvoelker has joined #openstack-ansible | 06:04 | |
*** udesale has quit IRC | 06:04 | |
*** dxiri has joined #openstack-ansible | 06:08 | |
*** dxiri has quit IRC | 06:09 | |
*** udesale has joined #openstack-ansible | 06:09 | |
*** galstrom_zzz is now known as galstrom | 06:10 | |
*** Oku_OS-away is now known as Oku_OS | 06:11 | |
*** armaan has joined #openstack-ansible | 06:31 | |
*** armaan_ has joined #openstack-ansible | 06:36 | |
*** armaan has quit IRC | 06:36 | |
odyssey4me | Amit82 verify that the endpoint in the service catalog is http | 06:36 |
odyssey4me | also, if the VM was created before you changed your endpoints then I'm not sure if you can change the console reference for it | 06:36 |
Amit82 | odyssey4me: All the service endpoints as observed in Horizon are http | 06:37 |
odyssey4me | finally, check your browser debug console to verify that it's using the right endpoint | 06:37 |
*** markvoelker has quit IRC | 06:37 | |
Amit82 | VM was created after | 06:37 |
odyssey4me | Amit82 you should observe in the catalog itself using the utility container via the CLI | 06:37 |
Amit82 | I recreated the setup | 06:37 |
odyssey4me | the browser only gives you an end-user point of view | 06:37 |
*** SerenaFeng has joined #openstack-ansible | 06:39 | |
Amit82 | Here is the list of endpoints, all are on http: http://paste.openstack.org/show/612635/ | 06:39 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server stable/newton: Dynamically include upgrade_check task files https://review.openstack.org/474457 | 06:39 |
odyssey4me | Amit82 you'll need to do some digging if you can't figure it out via the browser | 06:40 |
*** ArchiFleKs has quit IRC | 06:44 | |
openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Python fix to the connection issue https://review.openstack.org/473337 | 06:52 |
*** pcaruana has joined #openstack-ansible | 06:58 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Make a "test_branch" variable set to master https://review.openstack.org/471440 | 07:02 |
*** armaan_ has quit IRC | 07:02 | |
*** armaan has joined #openstack-ansible | 07:03 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: scripts: Introduce sources-branch-updater-lib library https://review.openstack.org/465465 | 07:03 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: scripts: sources-branch-updater: Allow role requirements updates on master https://review.openstack.org/465466 | 07:03 |
*** NWOutlier_ has quit IRC | 07:07 | |
*** dxiri has joined #openstack-ansible | 07:10 | |
evrardjp | fun, https://review.openstack.org/#/c/473368/ failed again | 07:12 |
*** dxiri has quit IRC | 07:15 | |
*** armaan has quit IRC | 07:16 | |
*** manheim has joined #openstack-ansible | 07:20 | |
*** manheim has quit IRC | 07:24 | |
*** udesale has quit IRC | 07:26 | |
*** udesale has joined #openstack-ansible | 07:27 | |
*** SerenaFeng has quit IRC | 07:30 | |
*** SerenaFeng has joined #openstack-ansible | 07:30 | |
*** armaan has joined #openstack-ansible | 07:31 | |
*** armaan has quit IRC | 07:32 | |
*** armaan has joined #openstack-ansible | 07:33 | |
*** markvoelker has joined #openstack-ansible | 07:34 | |
*** shardy has joined #openstack-ansible | 07:36 | |
*** ArchiFleKs has joined #openstack-ansible | 07:39 | |
*** manheim has joined #openstack-ansible | 07:40 | |
*** armaan has quit IRC | 07:43 | |
*** manheim has quit IRC | 07:45 | |
*** SerenaFeng has quit IRC | 07:46 | |
*** gkadam is now known as gkadam-brb | 07:48 | |
*** vnogin has joined #openstack-ansible | 08:01 | |
*** armaan has joined #openstack-ansible | 08:03 | |
*** SerenaFeng has joined #openstack-ansible | 08:04 | |
*** SerenaFeng has quit IRC | 08:05 | |
evrardjp | vars vars vars :) | 08:06 |
*** markvoelker has quit IRC | 08:08 | |
*** SerenaFeng has joined #openstack-ansible | 08:11 | |
*** dxiri has joined #openstack-ansible | 08:11 | |
*** dxiri has quit IRC | 08:15 | |
*** armaan has quit IRC | 08:17 | |
*** vnogin has quit IRC | 08:24 | |
*** pbandark has joined #openstack-ansible | 08:26 | |
*** cNilesh has quit IRC | 08:29 | |
*** shardy has quit IRC | 08:31 | |
odyssey4me | evrardjp yeah, I just realised that for tests we aren't activating the thing to place the override | 08:32 |
*** cNilesh has joined #openstack-ansible | 08:32 | |
evrardjp | I will have a look at the tool in tests/ to update this globally | 08:32 |
evrardjp | openstack-ansible-tests | 08:33 |
odyssey4me | you're going to do it now? | 08:33 |
evrardjp | never touched it, it's time | 08:33 |
evrardjp | finishing something quick first | 08:33 |
odyssey4me | ok, thanks - I'll wait for your patch then | 08:33 |
*** rstarmer_ has joined #openstack-ansible | 08:34 | |
evrardjp | thanks | 08:37 |
evrardjp | and sorry for the wait :p | 08:37 |
*** armaan has joined #openstack-ansible | 08:40 | |
*** cNilesh has quit IRC | 08:40 | |
*** rstarmer_ has quit IRC | 08:41 | |
*** rstarmer_ has joined #openstack-ansible | 08:42 | |
*** cNilesh has joined #openstack-ansible | 08:43 | |
odyssey4me | evrardjp does this patch make any sense to you? https://review.openstack.org/470944 | 08:48 |
evrardjp | oh my god the test updating script does tons of changes | 08:48 |
evrardjp | I'll do it manually I think | 08:49 |
odyssey4me | test updating script? | 08:49 |
evrardjp | sync test repos.sh | 08:49 |
odyssey4me | oh no, that's not needed to fix the galera thing at all | 08:49 |
evrardjp | Well I wanted to test it anyway, if we have many repos to fix | 08:50 |
odyssey4me | all we need is to activate galera_disable_privatedevices in test-vars | 08:50 |
evrardjp | yes | 08:50 |
odyssey4me | there's other work going on to make that update automatically, so it's better if you don't | 08:50 |
evrardjp | oh yes, not for all the repo, for the global one | 08:50 |
*** electrofelix has joined #openstack-ansible | 08:51 | |
evrardjp | memcached too? | 08:51 |
evrardjp | let's start with galera :p | 08:51 |
odyssey4me | I'm working on a patch for memcache now - only galera required at this time | 08:51 |
evrardjp | ok. Gates run everything under containers right? I can't remember one that doesn't. | 08:53 |
odyssey4me | mostly, for role testing it doesn't matter if we activate it globally | 08:53 |
odyssey4me | for galera it's always in a container anyway | 08:54 |
*** piliman974 has joined #openstack-ansible | 08:54 | |
*** manheim has joined #openstack-ansible | 08:55 | |
*** manheim has quit IRC | 08:56 | |
*** manheim has joined #openstack-ansible | 08:56 | |
*** yolanda_ has joined #openstack-ansible | 08:59 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Set PrivateDevices=false for Galera https://review.openstack.org/474316 | 09:00 |
odyssey4me | evrardjp ^ like so? | 09:00 |
openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: Disable private devices https://review.openstack.org/474500 | 09:01 |
evrardjp | I will have a look | 09:01 |
evrardjp | yup | 09:01 |
evrardjp | sounds safer | 09:01 |
evrardjp | if anyone sets is_metal: no for example | 09:01 |
evrardjp | or is_metal: yes | 09:01 |
odyssey4me | yep | 09:01 |
evrardjp | we love our | bool | 09:02 |
evrardjp | >< | 09:02 |
odyssey4me | mgagne according to a memcache install i have, PrivateDevices=no | 09:05 |
*** markvoelker has joined #openstack-ansible | 09:05 | |
odyssey4me | oops, not mgagne I meant mgariepy | 09:05 |
evrardjp | mgariepy: ? | 09:05 |
evrardjp | :p | 09:05 |
odyssey4me | https://pastebin.com/ps1LMrcJ | 09:06 |
odyssey4me | let me put it into a container to confirm | 09:06 |
evrardjp | should be good | 09:06 |
*** vnogin has joined #openstack-ansible | 09:07 | |
evrardjp | if there is anyone else to vote on https://review.openstack.org/#/c/474500/ that would be great. | 09:07 |
odyssey4me | let's see if it passes gate - if it does, then perhaps we push it through considering it's a gate blocker | 09:08 |
odyssey4me | and we've been sitting with this problem for two weeks now | 09:08 |
*** jwitko__ has quit IRC | 09:10 | |
*** SerenaFeng has quit IRC | 09:11 | |
*** vnogin has quit IRC | 09:12 | |
evrardjp | odyssey4me: about the https://review.openstack.org/#/c/470944 ... | 09:17 |
evrardjp | well it makes sense to me to want to separate the packages, but I think the whole thing is overblown | 09:17 |
odyssey4me | agreed | 09:17 |
evrardjp | packages install should be simple as | 09:17 |
odyssey4me | part of me wants to let this happen to allow hwoarang to get on with the suse work | 09:18 |
evrardjp | name: "{{ list1 + list2 + list3 }}" | 09:18 |
odyssey4me | but another part of me wants to rewrite the whole way the packages are installed | 09:18 |
evrardjp | well yes it's not readable | 09:18 |
evrardjp | you did a patch recently that used this kind of pattern right? | 09:19 |
odyssey4me | I did some, yes - but the pattern is quite different | 09:19 |
*** rstarmer_ has quit IRC | 09:19 | |
odyssey4me | https://review.openstack.org/#/q/topic:distro-pkg-optimise | 09:20 |
*** rstarmer_ has joined #openstack-ansible | 09:21 | |
evrardjp | did we merge the lookup change with the group vars at the end? | 09:21 |
odyssey4me | lookup change? | 09:21 |
evrardjp | yes for py_pkg | 09:22 |
evrardjp | to load from group_vars | 09:22 |
odyssey4me | no, that's still in review | 09:22 |
evrardjp | I know I didn't like it | 09:22 |
evrardjp | well | 09:22 |
evrardjp | let me change the vote, it's not because I don't like it we shouldn't go forward | 09:22 |
evrardjp | could you ping me the review? | 09:23 |
odyssey4me | https://review.openstack.org/466295 | 09:23 |
evrardjp | thanks | 09:23 |
*** vnogin has joined #openstack-ansible | 09:23 | |
odyssey4me | do you want to add the /all bit there to make it clearer that it's a global only lookup? | 09:23 |
odyssey4me | I think it'd be better actually | 09:24 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins master: Add group_vars to py_pkgs lookup https://review.openstack.org/466295 | 09:24 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins master: Add group_vars to py_pkgs lookup https://review.openstack.org/466295 | 09:24 |
evrardjp | maybe add /repo too | 09:25 |
odyssey4me | like that - which means it will only read group vars targeted for all | 09:25 |
odyssey4me | we're doing this all over the place to make use of the requirements repo sha: https://github.com/openstack/openstack-ansible/blob/master/playbooks/lxc-hosts-setup.yml#L55 | 09:25 |
odyssey4me | so I'd want to put the git sources into all, not the repo group | 09:26 |
evrardjp | I meant /group_vars/repo | 09:26 |
odyssey4me | I know, and I still do not think that has any use at all. | 09:26 |
evrardjp | that's fine, but I think the future could be brighter if we target groups properly | 09:26 |
odyssey4me | it is not global | 09:27 |
*** SerenaFeng has joined #openstack-ansible | 09:27 | |
odyssey4me | this is a global lookup | 09:27 |
odyssey4me | yes, ultimately we could do a lot of things better | 09:27 |
odyssey4me | I was thinking this morning that we should actually pull all repo setup, apt cache updates, etc into its own role | 09:27 |
odyssey4me | perhaps even use a galaxy role | 09:27 |
evrardjp | so someone that sets something in group_vars/repo would not even see his changed applied there? | 09:28 |
odyssey4me | no | 09:28 |
odyssey4me | because the repo group is not universal | 09:28 |
evrardjp | that's nonsense! | 09:28 |
odyssey4me | it needs to be clear that repo vars are global | 09:28 |
evrardjp | Ok now I remember why we arrived to a deadlock. | 09:28 |
*** SerenaFeng has quit IRC | 09:29 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins master: Add group_vars to py_pkgs lookup https://review.openstack.org/466295 | 09:30 |
evrardjp | I'd rather see the var for the git_repo for keystone in group_vars/keystone_all .... but we can iterate | 09:30 |
evrardjp | I think the part I don't get is what you call "global" | 09:31 |
odyssey4me | the values returned by the lookup are used by the repo in a global fashion | 09:32 |
odyssey4me | there is absolutely nothing group focused aboutit | 09:32 |
evrardjp | move its usage into a default variable, and you'll see it respects the groups :p | 09:33 |
evrardjp | joking aside | 09:33 |
evrardjp | I agree we should go forward | 09:34 |
evrardjp | Added mhayden to the review. | 09:34 |
evrardjp | Him and cloudnull should be good enough to get this merged :p | 09:35 |
odyssey4me | :( http://logs.openstack.org/00/474500/1/check/gate-openstack-ansible-tests-ansible-func-centos-7/a82abb2/console.html#_2017-06-15_09_18_21_440569 | 09:35 |
odyssey4me | hmm, actually that's an access denied - which is different to before | 09:35 |
evrardjp | nooo | 09:35 |
evrardjp | I think I tried on my machine the galera_server change, I think it worked | 09:36 |
odyssey4me | lemme finish verifying that memcache doesn't need the private thingy override, then I'll fire this test up too | 09:37 |
*** markvoelker has quit IRC | 09:38 | |
evrardjp | well I'll fire up the galera + cinder just to see | 09:39 |
openstackgerrit | Merged openstack/openstack-ansible master: scripts: Introduce sources-branch-updater-lib library https://review.openstack.org/465465 | 09:40 |
*** galstrom is now known as galstrom_zzz | 09:52 | |
*** rstarmer_ has quit IRC | 10:12 | |
evrardjp | odyssey4me: Jun 15 09:59:00 infra1 systemd[5546]: Failed at step NAMESPACE spawning /bin/sh: Invalid argument | 10:13 |
odyssey4me | evrardjp where is that? | 10:14 |
evrardjp | on my test machine | 10:14 |
odyssey4me | hmm, is the override file there then? | 10:14 |
evrardjp | I've updated the test vars, and deleted all my previous roles | 10:14 |
evrardjp | let me double check, but I edited the file beforehand. | 10:14 |
odyssey4me | is the task to put the file there activating? | 10:15 |
evrardjp | yup it is | 10:15 |
*** rstarmer_ has joined #openstack-ansible | 10:16 | |
odyssey4me | hmm, it appears that the vars plugin is not doing host_vars properly for ansible 2.3 | 10:17 |
odyssey4me | well, not on centos at least | 10:18 |
odyssey4me | let me check for ubuntu | 10:18 |
evrardjp | yup it ran | 10:19 |
evrardjp | mmm ? | 10:19 |
odyssey4me | nvm, let me not distract you | 10:19 |
odyssey4me | figuring out the centos mariadb issue is something that needs solving | 10:19 |
*** cuongnv has quit IRC | 10:20 | |
evrardjp | yes. Here are patching here and there, but nothing seems very conclusive | 10:23 |
evrardjp | I am not sure we fixed the problem with the disable thingy | 10:24 |
odyssey4me | it's working on the role tests | 10:24 |
odyssey4me | that was the first time we had a gate pass for months | 10:24 |
odyssey4me | well, weeks | 10:24 |
*** abhishekk has joined #openstack-ansible | 10:25 | |
*** piliman974 has quit IRC | 10:26 | |
*** abhishekk has quit IRC | 10:26 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: sources-branch-updater: Allow role requirements updates on master https://review.openstack.org/465466 | 10:27 |
*** abhishekk has joined #openstack-ansible | 10:27 | |
*** cNilesh has quit IRC | 10:27 | |
*** stuartgr has joined #openstack-ansible | 10:29 | |
evrardjp | I wasn't aware of the issue to be honest | 10:33 |
evrardjp | :/ | 10:33 |
*** markvoelker has joined #openstack-ansible | 10:34 | |
odyssey4me | ok, I've confirmed that memcache has the same issue in a container | 10:40 |
*** armaan has quit IRC | 10:46 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: sources-branch-updater-lib.sh: Fix typo https://review.openstack.org/474537 | 10:48 |
*** shardy has joined #openstack-ansible | 10:50 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 10:52 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 10:53 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 10:54 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 10:55 |
*** rstarmer_ has quit IRC | 10:55 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 10:56 |
odyssey4me | sorry about the spam, clearly I'm being a little thick this morning | 10:56 |
*** fyu has quit IRC | 10:58 | |
*** gkadam-brb is now known as gkadam | 11:00 | |
*** armaan has joined #openstack-ansible | 11:05 | |
*** tobberyd_ has joined #openstack-ansible | 11:06 | |
*** tobberyd_ has quit IRC | 11:07 | |
*** tobberyd_ has joined #openstack-ansible | 11:07 | |
*** markvoelker has quit IRC | 11:08 | |
*** smatzek has joined #openstack-ansible | 11:09 | |
*** tobberydberg has quit IRC | 11:09 | |
*** tobberyd_ has quit IRC | 11:10 | |
*** tobberydberg has joined #openstack-ansible | 11:11 | |
*** ivveh has quit IRC | 11:15 | |
spotz | Quick reminder for the early crowd no meeting today | 11:16 |
*** ivveh has joined #openstack-ansible | 11:16 | |
Tahvok | Is it ok to ignore the log_hosts? Or is it required? | 11:16 |
Tahvok | I don't want to have seperate log_hosts in my env | 11:17 |
spotz | Tahvok: Where do you want your logs to go? | 11:17 |
evrardjp | why no meeting today? | 11:19 |
Tahvok | The default? On the controller/compute? | 11:19 |
Tahvok | Which logs are we talking about anyway? The service logs? | 11:20 |
evrardjp | if no meeting, we have to think about who's gonna handle the bug triage meeting for the next weeks. | 11:20 |
spotz | evrardjp: We decided to cancel last meeting cause no andymccr | 11:22 |
spotz | Tahvok: then that's your log host | 11:23 |
spotz | be back need to go take care of pony | 11:23 |
evrardjp | well we can't postpone every week I guess | 11:25 |
evrardjp | if ppl want to postpone fine for me | 11:25 |
*** udesale has quit IRC | 11:26 | |
*** armaan has quit IRC | 11:27 | |
evrardjp | odyssey4me: mysql -u root -p secrete ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) | 11:28 |
evrardjp | fun. | 11:28 |
*** udesale has joined #openstack-ansible | 11:28 | |
*** yolanda__ has joined #openstack-ansible | 11:30 | |
*** yolanda_ has quit IRC | 11:30 | |
*** tobberyd_ has joined #openstack-ansible | 11:31 | |
*** tobberydberg has quit IRC | 11:34 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: sources-branch-updater-lib.sh: Fix typos https://review.openstack.org/474537 | 11:36 |
*** armaan has joined #openstack-ansible | 11:37 | |
*** armaan has quit IRC | 11:37 | |
odyssey4me | evrardjp for centos, perhaps instead of executing the securing bits via our own script we should be using the mariadb provided script? | 11:39 |
evrardjp | I didn't know any of this. | 11:39 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: sources-branch-updater-lib.sh: Avoid using 'dirname $0' https://review.openstack.org/474551 | 11:39 |
odyssey4me | also, is the root user's password being set at all | 11:39 |
odyssey4me | I just discovered something interesting - using the vars plugin you cannot override localhost's vars | 11:40 |
odyssey4me | you can override any other host, but not localhost | 11:40 |
hwoarang | hmm what is the vars plugin | 11:41 |
evrardjp | hwoarang: something I wrote | 11:41 |
hwoarang | aha | 11:41 |
evrardjp | it will be deprecated now that ansible is becoming better :p | 11:42 |
evrardjp | don't worry too much :p | 11:42 |
evrardjp | odyssey4me: interesting | 11:42 |
evrardjp | I'd like to see the env, I don't think it's directly related to the plugins itself | 11:42 |
evrardjp | plugin* | 11:43 |
evrardjp | I think I tried it on localhost | 11:43 |
evrardjp | could you run it with -vvvvv ? | 11:43 |
odyssey4me | yep, want to pop into my test env? | 11:43 |
evrardjp | You should see why and how it overrides | 11:43 |
evrardjp | sure thing | 11:43 |
evrardjp | I just want to understand the mysql too | 11:43 |
evrardjp | you were talking about that right? https://github.com/openstack/openstack-ansible-galera_server/blob/master/tasks/galera_setup.yml | 11:43 |
Tahvok | network_hosts | 11:45 |
Tahvok | Is it the neutron api, or the network agent hosts? | 11:45 |
odyssey4me | evrardjp nope https://github.com/openstack/openstack-ansible-galera_server/blob/master/tasks/galera_install_yum.yml#L109-L135 | 11:45 |
odyssey4me | Tahvok both | 11:46 |
Tahvok | If I have a dedicated network nodes, is it it? | 11:46 |
evrardjp | odyssey4me: haha ok. | 11:46 |
Tahvok | odyssey4me: but I want it seperated.. The neutron server on the controllers, and the l3 agents on the network nodes. | 11:46 |
Tahvok | This is actually per the networking guide: https://docs.openstack.org/ocata/networking-guide/deploy-ovs-ha-vrrp.html | 11:47 |
*** tobberyd_ has quit IRC | 11:49 | |
odyssey4me | Tahvok yeah, I have no idea why our grouping is like that right now and haven't had a chance to figure out how to seperate them, which I expect is possible with our defaults | 11:49 |
Tahvok | odyssey4me: so how should I do this? | 11:49 |
*** markvoelker has joined #openstack-ansible | 11:51 | |
*** tobberydberg has joined #openstack-ansible | 11:52 | |
*** yolanda__ is now known as yolanda | 12:05 | |
mhayden | morning | 12:07 |
*** abhishekk has quit IRC | 12:12 | |
*** scott_wd has joined #openstack-ansible | 12:13 | |
mgariepy | https://pastebin.com/zWNQubJv | 12:14 |
mgariepy | odyssey4me, ^^ | 12:14 |
scott_wd | hi, I have following interfaces available on my vms, what should I specify in openstack_user_config.yml for network_providers, http://paste.openstack.org/show/612680/, please help | 12:14 |
odyssey4me | mgariepy yes, I have a patch up for it already | 12:15 |
odyssey4me | :) | 12:15 |
odyssey4me | in a container that setting is there, outside a container it is not | 12:15 |
odyssey4me | weird | 12:15 |
odyssey4me | mgariepy https://review.openstack.org/474538 | 12:16 |
odyssey4me | mhayden ^ | 12:16 |
*** dxiri has joined #openstack-ansible | 12:16 | |
mgariepy | huh ? | 12:16 |
mgariepy | when installing baremetal the setting it not there ? | 12:16 |
mhayden | odyssey4me: one of the best written patches ever, it looks like it was inspired by greatness :P | 12:17 |
odyssey4me | mgariepy nope - it's off | 12:17 |
mhayden | i gave it +2.5 | 12:17 |
odyssey4me | mhayden I learn from the best. | 12:18 |
odyssey4me | I hope to do great things. :) | 12:18 |
mgariepy | odyssey4me, on 16.04 in container is it on or off ? | 12:19 |
odyssey4me | mgariepy I didn't check | 12:20 |
mgariepy | if I install memcached on my host it's the same private output. | 12:20 |
*** dxiri has quit IRC | 12:20 | |
openstackgerrit | Merged openstack/openstack-ansible-memcached_server master: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474538 | 12:23 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-memcached_server stable/ocata: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474561 | 12:25 |
odyssey4me | mgariepy so, I've been looking into the issue with the selinux bindings for localhost... and as it turns out it appears that you cannot use the inventory to override the python interpreter for localhost | 12:25 |
odyssey4me | for any other host, sure, but never localhost | 12:25 |
mgariepy | is there a way to do something similar for localhost ? | 12:27 |
*** bauruine has quit IRC | 12:28 | |
neith | someone can explain me whe the "cinder-volume" service is deployed on infra hosts? | 12:29 |
scott_wd | hi all , I have these (http://paste.openstack.org/show/612680/) interfaces available on my vms, what should I specify in openstack_user_config.yml for network_providers, please help | 12:29 |
odyssey4me | mgariepy busy trying to figure it out | 12:29 |
odyssey4me | neith it is not by default - it is deployed on the hosts you designate as storage hosts | 12:29 |
mgariepy | from the end of : https://github.com/ansible/ansible/issues/16724 | 12:29 |
mgariepy | -- delegate_to DOES change the host execution context | 12:29 |
mgariepy | so delegate to localhost ? | 12:30 |
openstackgerrit | Merged openstack/openstack-ansible-plugins master: Remove neutron library https://review.openstack.org/467797 | 12:30 |
odyssey4me | mgariepy we are delegating to localhost | 12:31 |
*** schwicht has joined #openstack-ansible | 12:31 | |
odyssey4me | it's just that for localhost the ansible_python_interpreter var is ptotected | 12:31 |
mgariepy | hmm | 12:32 |
neith | according to https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/app-config-prod.html the storage hosts are also the infra hosts. And they are handling the nfs mount for cinder? I mean, how the infra hosts can communicate with the nfs server if they don"t even have an IP on the same subnet than the nfs server? | 12:33 |
evrardjp | I think this var was never intended to be one in the inventory | 12:33 |
openstackgerrit | Merged openstack/monitorstack master: Remove inline comments from example config https://review.openstack.org/474392 | 12:33 |
openstackgerrit | Merged openstack/monitorstack master: Verify https connections by default https://review.openstack.org/474366 | 12:34 |
evrardjp | neith: in all cases you have to ensure the proper host for storage host can speak with its backend | 12:34 |
*** jamesdenton has joined #openstack-ansible | 12:34 | |
*** Amit82 has quit IRC | 12:35 | |
neith | evrardjp: I do understand that, I just raise the fact that the doc might be wrong | 12:35 |
evrardjp | oh yes. | 12:35 |
odyssey4me | neith that is just an example configuration | 12:35 |
evrardjp | where exactly? | 12:35 |
*** schwicht has quit IRC | 12:35 | |
evrardjp | and FYI you don't have to be on the same network to do NFS :p | 12:36 |
neith | evrardjp: its said that infra hosts dont need IP on br storage | 12:36 |
evrardjp | nfs is working perfectly fine routed. | 12:36 |
evrardjp | I will not say it's smart | 12:36 |
evrardjp | :p | 12:36 |
*** bauruine has joined #openstack-ansible | 12:36 | |
neith | evrardjp: yep but in the same page its shown as isolated vlans | 12:36 |
odyssey4me | neith ah, that's an oversight in the config example | 12:36 |
evrardjp | indeed. neith and odyssey4me | 12:36 |
odyssey4me | could you either submit a bug or patch for that | 12:36 |
neith | I will | 12:37 |
evrardjp | neith: are you willing to patch this? that could be good :) | 12:37 |
neith | count on me | 12:37 |
evrardjp | hehe good | 12:37 |
neith | so last question, If I add an IP static conf to br-storage will it interfere with lxc containers already assigned with an ip on this bridge? | 12:38 |
evrardjp | maybe it's good to multiply the example interfaces config file too? One for infra, one for compute, what do you think odyssey4me? | 12:38 |
odyssey4me | mgariepy evrardjp so it appears that the way it's supposed to be done is to override it in the play, or in the task itself | 12:38 |
evrardjp | odyssey4me: yes | 12:38 |
odyssey4me | evrardjp too many files then | 12:38 |
odyssey4me | evrardjp I'd rather see the example file commented with any extra info | 12:38 |
mgariepy | odyssey4me, why is it overidable by -e "ansible_python_interpreter=/usr/bin/python" | 12:38 |
evrardjp | mgariepy: that's because ansible is lenient :) | 12:38 |
odyssey4me | mgariepy that's a global override, which works | 12:39 |
odyssey4me | so the simplest solution here is to implement the override on the task, which I think is fair - any comments before I push the patch? | 12:39 |
evrardjp | No I am not, I am even strongly in favor, because I wanted to use that everywhere :p | 12:39 |
evrardjp | for this I'd rather be explicit | 12:40 |
mgariepy | go ahead | 12:40 |
evrardjp | and like we discussed live, it opens the door for venvs | 12:40 |
evrardjp | like utility venvs :) | 12:40 |
mgariepy | build an venv for the utility container ? | 12:41 |
evrardjp | I meant not building a utility container and having a venv somewhere where we do all our stuff :D | 12:41 |
mgariepy | ha. | 12:42 |
evrardjp | instead of installing keystone client everywhere | 12:42 |
evrardjp | it's not my fault, odyssey4me shown me that we are installing stuff everywhere. | 12:42 |
*** schwicht has joined #openstack-ansible | 12:43 | |
evrardjp | :D | 12:43 |
*** oldcrowew has joined #openstack-ansible | 12:43 | |
oldcrowew | hello, i was wondering if anyone had experience using jenkins-job-builder? I am having issues find good examples for builders: ansible-playbook. Any help would be appreciated! | 12:44 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts master: Use the host python interpreter when delegating to localhost https://review.openstack.org/474565 | 12:44 |
odyssey4me | mgariepy mhayden ^ that should get the integrated build back on track for centos | 12:45 |
odyssey4me | well, for that one issue | 12:45 |
mhayden | oh i bumped into this the other day | 12:46 |
mhayden | was doing a fair amount of head scratching | 12:46 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove localhost host_vars file https://review.openstack.org/474567 | 12:47 |
mgariepy | great success :D http://paste.openstack.org/show/612696/ | 12:48 |
odyssey4me | huzzah | 12:48 |
odyssey4me | mhayden mgariepy the integreated build patch is to ensure that no-one is led down this path of madness again | 12:49 |
odyssey4me | having that file there is misleading | 12:49 |
*** esberglu has joined #openstack-ansible | 12:50 | |
mhayden | learned something new | 12:50 |
odyssey4me | as did I | 12:50 |
odyssey4me | now, back to the centos mariadb debacle | 12:50 |
odyssey4me | mhayden mgariepy FYI for some reason https://review.openstack.org/474500 does not work | 12:52 |
odyssey4me | it's very odd | 12:52 |
* mhayden looks | 12:52 | |
openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: Use the host python interpreter when delegating to localhost https://review.openstack.org/474565 | 12:53 |
mhayden | odyssey4me: well, MariaDB is *up* but something else is broen | 12:53 |
mhayden | broken, rather | 12:53 |
*** kylek3h has joined #openstack-ansible | 12:53 | |
odyssey4me | oh yes, the root password either is not being set or is being set incorrectly | 12:53 |
mhayden | ERROR 1045 (28000): Access denied for user 'root'@'localhost <-- server had to be online to say that | 12:53 |
mhayden | that's unusual | 12:54 |
odyssey4me | I was wondering whether https://github.com/openstack/openstack-ansible-galera_server/blob/master/tasks/galera_install_yum.yml#L109-L135 should be changed | 12:54 |
odyssey4me | odd though, because it's obviously working in the galera role | 12:54 |
*** schwicht has quit IRC | 12:55 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts stable/ocata: Use the host python interpreter when delegating to localhost https://review.openstack.org/474569 | 12:56 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts stable/newton: Use the host python interpreter when delegating to localhost https://review.openstack.org/474570 | 12:56 |
mgariepy | odyssey4me, mhayden missing /run/mysqld ? | 12:57 |
mgariepy | so mysql cannot create the socket file. | 12:57 |
*** armaan has joined #openstack-ansible | 12:58 | |
*** qcrivera has joined #openstack-ansible | 12:59 | |
* mhayden fired up a vm to test it in the background | 13:04 | |
*** udesale has quit IRC | 13:05 | |
evrardjp | mhayden: mgariepy let's stay in sync there | 13:05 |
evrardjp | I have my own instance too | 13:05 |
evrardjp | odyssey4me: suggested to cleanup the extra tasks we are doing for yum | 13:06 |
evrardjp | https://github.com/openstack/openstack-ansible-galera_server/blob/1da7340a99328d4d7d89bbcdb9986b4e6c975a67/tasks/galera_install_yum.yml#L112-L129 | 13:06 |
evrardjp | I doubt that's the issue | 13:07 |
evrardjp | mgariepy: I see the socket file on my machine | 13:08 |
mgariepy | started a new one with the private stuff patch in ? | 13:09 |
evrardjp | yes | 13:10 |
evrardjp | I'm testing cinder role | 13:10 |
evrardjp | FYI | 13:10 |
evrardjp | not that matters much | 13:10 |
evrardjp | that it* | 13:10 |
openstackgerrit | Major Hayden proposed openstack/ansible-hardening stable/ocata: [Docs] Replace security role references https://review.openstack.org/473837 | 13:10 |
*** gouthamr has joined #openstack-ansible | 13:11 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: sources-branch-updater-lib.sh: Avoid using 'dirname $0' https://review.openstack.org/474551 | 13:12 |
mhayden | if anyone has a moment to gander at my docs patch, i'd be much obliged: https://review.openstack.org/473823 | 13:12 |
odyssey4me | I have some ocata backports waiting for review if anyone has a gap: https://review.openstack.org/#/q/owner:jesse-pretorius+status:open+branch:stable/ocata | 13:13 |
odyssey4me | same for newton: https://review.openstack.org/#/q/owner:jesse-pretorius+status:open+branch:stable/newton | 13:13 |
*** oldcrowew has quit IRC | 13:16 | |
*** schwicht has joined #openstack-ansible | 13:16 | |
evrardjp | mhayden: and... reviewed. | 13:17 |
mhayden | evrardjp: thanks sir! :) | 13:17 |
evrardjp | you didn't see the review yet! | 13:17 |
evrardjp | haha no don't worry. | 13:17 |
mhayden | nice | 13:17 |
neith | so cinder-volume service is creating the /var/lib/cinder/mnt/7288a12b3319e2d9ff0da4e3b2b7b5aa directory but does not set permission for user cinder to use it. does this path is randomly generated? | 13:22 |
neith | when using nfs | 13:23 |
odyssey4me | neith you have to ensure that your nfs system allows access to all users | 13:23 |
odyssey4me | and yes, for volumes the path is random | 13:23 |
neith | odyssey4me: All users? not just cinder? | 13:23 |
evrardjp | odyssey4me: I'm having a look at the ansible_python_environment | 13:24 |
odyssey4me | well, is the cinder user (or whichever use mounts the disk) consistent across hosts? | 13:24 |
odyssey4me | if so, then just that | 13:24 |
evrardjp | It should be an inventory thing apparently :/ | 13:24 |
evrardjp | except for pip tasks | 13:24 |
*** cpuga has joined #openstack-ansible | 13:25 | |
openstackgerrit | Merged openstack/ansible-hardening master: Doc updates https://review.openstack.org/473823 | 13:25 |
*** askb has quit IRC | 13:26 | |
*** cpuga has quit IRC | 13:26 | |
*** lucasxu has joined #openstack-ansible | 13:26 | |
*** cpuga has joined #openstack-ansible | 13:27 | |
openstackgerrit | Marc Gariépy proposed openstack/openstack-ansible-ceph_client master: Use the host python interpreter when delegating to localhost https://review.openstack.org/474577 | 13:27 |
openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts stable/newton: Use the host python interpreter when delegating to localhost https://review.openstack.org/474570 | 13:27 |
openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts stable/ocata: Use the host python interpreter when delegating to localhost https://review.openstack.org/474569 | 13:27 |
odyssey4me | nice catch there mgariepy | 13:28 |
mgariepy | nothing grep cannot find :) | 13:28 |
*** klamath has joined #openstack-ansible | 13:29 | |
*** klamath has quit IRC | 13:29 | |
mgariepy | + I need ceph to work on centos :D | 13:29 |
*** klamath has joined #openstack-ansible | 13:29 | |
*** mariusv has joined #openstack-ansible | 13:33 | |
*** George has joined #openstack-ansible | 13:34 | |
*** jamesden_ has joined #openstack-ansible | 13:35 | |
*** jamesdenton has quit IRC | 13:36 | |
*** thorst has joined #openstack-ansible | 13:36 | |
*** whiteveil has joined #openstack-ansible | 13:42 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Implement a data plane down time test for upgrades https://review.openstack.org/467661 | 13:44 |
*** scott_wd has quit IRC | 13:44 | |
odyssey4me | I'd appreciate it if I could get some reviews for ^ to round out the tests during upgrades. They'll be very useful to improve our rolling upgrade capabilities. | 13:45 |
neith | odyssey4me: alright I squashed user to cinder users | 13:47 |
*** woodard has joined #openstack-ansible | 13:47 | |
*** woodard has quit IRC | 13:48 | |
*** woodard has joined #openstack-ansible | 13:48 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Set PrivateDevices=false for MemcacheD https://review.openstack.org/474583 | 13:51 |
odyssey4me | mhayden ^ I am once again inspired by your work. | 13:52 |
odyssey4me | mgariepy ^ | 13:52 |
mhayden | woot | 13:52 |
mgariepy | magical, only missing a unicorn ! Thanks odyssey4me | 13:54 |
*** mariusv has quit IRC | 13:55 | |
spotz | forgot to say back but I don't think you guys missed me:) | 13:58 |
*** armaan has quit IRC | 13:58 | |
mgariepy | spotz, we always miss you | 13:58 |
spotz | aww mgariepy | 13:58 |
*** George has quit IRC | 14:00 | |
*** schwicht has quit IRC | 14:04 | |
cloudnull | mornings | 14:06 |
*** qcrivera has quit IRC | 14:06 | |
palendae | Morning | 14:09 |
*** phalmos has quit IRC | 14:10 | |
*** cjloader_ has joined #openstack-ansible | 14:12 | |
*** Tahvok has quit IRC | 14:12 | |
openstackgerrit | Merged openstack/openstack-ansible-memcached_server stable/ocata: Disable PrivateDevices for MemcacheD on CentOS 7 https://review.openstack.org/474561 | 14:13 |
openstackgerrit | Merged openstack/openstack-ansible-plugins master: Python fix to the connection issue https://review.openstack.org/473337 | 14:13 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: defaults: main: Set default host disk size to 16G https://review.openstack.org/474150 | 14:15 |
neith | the ressource usage menu in horizon is the gnocchi GUI? | 14:15 |
*** SerenaFeng has joined #openstack-ansible | 14:16 | |
*** SerenaFeng has quit IRC | 14:17 | |
robcresswell | neith: No, its Ceilometer, and was removed a long time ago | 14:19 |
*** acormier has joined #openstack-ansible | 14:20 | |
*** SerenaFeng has joined #openstack-ansible | 14:20 | |
neith | robcresswell: ceilometer is deprecated ? | 14:21 |
openstackgerrit | Merged openstack/openstack-ansible-ceph_client master: Use the host python interpreter when delegating to localhost https://review.openstack.org/474577 | 14:22 |
robcresswell | neith: No, the GUI part in Horizon was removed a long time ago though. | 14:22 |
robcresswell | neith: It was barely usable for many cycles. | 14:23 |
*** cjloader_ has quit IRC | 14:23 | |
neith | robcresswell: so how to remove it from horizon ? | 14:23 |
*** cjloader_ has joined #openstack-ansible | 14:23 | |
odyssey4me | neith gnocchi is simply like RRD - it's a time series database... it doesn't make sense to have a GUI | 14:23 |
openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/newton: Dynamically include upgrade_check task files https://review.openstack.org/474457 | 14:24 |
odyssey4me | I have no idea whether it actually does. | 14:24 |
odyssey4me | And I have no idea whether there's any sort of telemetry system GUI for horizon. | 14:24 |
mhayden | evrardjp: so the openstack-ansible-tests test that i ran pulled in the latest galera server changes | 14:27 |
*** jamesden_ has quit IRC | 14:27 | |
mhayden | but the galera_disable_privatedevices variable isn't being set | 14:27 |
*** cjloader_ has quit IRC | 14:28 | |
evrardjp | mhayden: did you use my patch? | 14:28 |
evrardjp | https://review.openstack.org/#/c/474500/ | 14:28 |
evrardjp | mhayden: I used this patch, it sets the proper value, but still fails | 14:28 |
*** qcrivera has joined #openstack-ansible | 14:29 | |
mhayden | yeah, i have that patched in | 14:29 |
*** jamesdenton has joined #openstack-ansible | 14:29 | |
*** udesale has joined #openstack-ansible | 14:30 | |
mhayden | the variable in test-vars.yml isn't being applied, for some reason | 14:30 |
evrardjp | what? | 14:31 |
evrardjp | I see the task running | 14:31 |
mhayden | well it did get applied, but i checked the system journal and it looks like one of the mariadb startups failed with the namespace error | 14:32 |
mhayden | subsequent startups are okay, but the first one failed | 14:33 |
evrardjp | oh | 14:34 |
mhayden | https://gist.github.com/major/5d7bf972ad9c562b33fd3693e41addcb | 14:34 |
mhayden | so there was a task that enabled the service | 14:34 |
mhayden | and then the next one that runs all of the user setup commands failed | 14:34 |
mhayden | because mariadb couldn't start | 14:34 |
*** Tahvok has joined #openstack-ansible | 14:34 | |
evrardjp | interesting output service mysql stop removes=None creates=/etc/mysql/rhel_configured _uses_shell=True | 14:35 |
openstackgerrit | Charles Farquhar proposed openstack/openstack-ansible-os_swift master: Add support for swift3 middleware https://review.openstack.org/374416 | 14:36 |
evrardjp | ok I've messed up too much these things for other purposes, let me destroy my infra1 and debug it properly | 14:36 |
neith | odyssey4me: well, I understadn the usage of time series DB but how gnocchi is now integrated to OSA? whats its goal? | 14:38 |
*** armaan has joined #openstack-ansible | 14:38 | |
evrardjp | integration for ceilometer/aodh | 14:39 |
openstackgerrit | Charles Farquhar proposed openstack/openstack-ansible-os_swift master: Add support for swift3 middleware https://review.openstack.org/374416 | 14:40 |
openstackgerrit | Merged openstack/openstack-ansible master: Add API availability tests to upgrade https://review.openstack.org/471468 | 14:44 |
openstackgerrit | Merged openstack/openstack-ansible stable/newton: Add additional role path for lint testing https://review.openstack.org/473895 | 14:44 |
openstackgerrit | Merged openstack/openstack-ansible stable/newton: If repo build fails, force all hosts to fail https://review.openstack.org/473905 | 14:45 |
*** woodard has quit IRC | 14:45 | |
odyssey4me | mhayden is there a daemond reload being done? | 14:46 |
odyssey4me | *daemon | 14:46 |
*** stewie925 has joined #openstack-ansible | 14:46 | |
odyssey4me | neith I don't know the telemetry systems all that well, but as I understand it it goes something like this: | 14:47 |
odyssey4me | ceilometer collects metrics from hosts | 14:47 |
odyssey4me | gnocchi stores those metrics | 14:47 |
odyssey4me | aodh interprets them and stores alarms | 14:47 |
odyssey4me | and I can't recall what panko does | 14:48 |
odyssey4me | alextricity25 ^ | 14:48 |
neith | ok | 14:48 |
neith | i'll dig into that later | 14:48 |
*** Tahvok has quit IRC | 14:49 | |
*** armaan has quit IRC | 14:51 | |
openstackgerrit | Merged openstack/openstack-ansible master: Set PrivateDevices=false for Galera https://review.openstack.org/474316 | 14:52 |
*** bauruine has quit IRC | 14:53 | |
*** Tahvok has joined #openstack-ansible | 14:55 | |
*** manheim has quit IRC | 14:55 | |
*** bauruine has joined #openstack-ansible | 14:57 | |
*** dxiri has joined #openstack-ansible | 14:57 | |
*** schwicht has joined #openstack-ansible | 14:59 | |
*** rstarmer has joined #openstack-ansible | 14:59 | |
*** gfa has quit IRC | 15:01 | |
neith | to modify spice keymap I can only re run os-nova-install.yml ? | 15:04 |
*** gkadam has quit IRC | 15:04 | |
odyssey4me | neith yep | 15:04 |
*** dxiri has quit IRC | 15:07 | |
*** rstarmer has quit IRC | 15:09 | |
*** dxiri has joined #openstack-ansible | 15:11 | |
*** schwicht has quit IRC | 15:11 | |
*** Oku_OS is now known as Oku_OS-away | 15:14 | |
openstackgerrit | Charles Farquhar proposed openstack/openstack-ansible-os_swift master: Add support for swift3 middleware https://review.openstack.org/374416 | 15:15 |
*** phalmos has joined #openstack-ansible | 15:15 | |
*** schwicht has joined #openstack-ansible | 15:15 | |
*** tobberyd_ has joined #openstack-ansible | 15:22 | |
*** tobberydberg has quit IRC | 15:25 | |
*** tobberyd_ has quit IRC | 15:28 | |
*** pbandark has quit IRC | 15:28 | |
*** pbandark1 has joined #openstack-ansible | 15:28 | |
*** pbandark1 is now known as pbandark | 15:31 | |
*** vnogin has quit IRC | 15:32 | |
*** vnogin has joined #openstack-ansible | 15:32 | |
*** manheim has joined #openstack-ansible | 15:35 | |
*** dxiri has quit IRC | 15:37 | |
openstackgerrit | Merged openstack/openstack-ansible master: scripts: sources-branch-updater-lib.sh: Fix typos https://review.openstack.org/474537 | 15:38 |
openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Add additional role path for lint testing https://review.openstack.org/473894 | 15:38 |
openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Include strategy/connection plugin fix https://review.openstack.org/473938 | 15:38 |
openstackgerrit | Merged openstack/openstack-ansible stable/ocata: If repo build fails, force all hosts to fail https://review.openstack.org/473904 | 15:38 |
*** SerenaFeng has quit IRC | 15:39 | |
*** dxiri has joined #openstack-ansible | 15:39 | |
*** manheim has quit IRC | 15:44 | |
*** lenserpgo has joined #openstack-ansible | 15:45 | |
evrardjp | mhayden: mgariepy could you have a look at the centos thing? I have to go for today, and it doesn't look that the var change is enough | 15:47 |
mgariepy | http://paste.openstack.org/show/612725/ | 15:47 |
mhayden | i'll try to gander in a bit | 15:47 |
mgariepy | the mysql create user fails | 15:47 |
*** craigs has quit IRC | 15:53 | |
*** tobberydberg has joined #openstack-ansible | 15:53 | |
*** dxiri has quit IRC | 15:53 | |
evrardjp | Ok guys, enjoy the next weeks, and see you in July! :p | 15:54 |
*** dxiri has joined #openstack-ansible | 15:56 | |
odyssey4me | cheerio evrardjp - have a great holiday! | 15:57 |
mgariepy | cya evrardjp | 15:57 |
*** tobberydberg has quit IRC | 15:57 | |
alextricity25 | How can I run just the os_ceilometer playbook when running the role tests? Is there an environment variable I can set? | 15:58 |
odyssey4me | alextricity25 just use ansible in the tox venv | 16:00 |
alextricity25 | That'll work. Thanks odyssey4me | 16:00 |
odyssey4me | .tox/functional/bin/ansible-playbook -i tests/inventory <playbook name> | 16:00 |
*** udesale__ has joined #openstack-ansible | 16:05 | |
*** udesale has quit IRC | 16:07 | |
*** dxiri has quit IRC | 16:08 | |
EmilienM | folks, please look [deployment][kolla][openstack-ansible][kolla][openstack-helm] ansible role to produce oslo.config files for openstack services | 16:09 |
EmilienM | odyssey4me: ^ | 16:09 |
*** dxiri has joined #openstack-ansible | 16:10 | |
*** jamesdenton has quit IRC | 16:16 | |
cloudnull | oh cool | 16:23 |
alextricity25 | Turns out the uri module writes out json with the unicode indicators: {'telemetry:create_samples': u'', 'telemetry:query_sample': u'', 'telemetry:get_resources': u'', 'telemetry:get_meters': u'', 'segregation': u'rule:context_is_admin', 'telemetry:get_resource': u'', 'context_is_admin': u'role:admin', 'telemetry:get_sample': u'', 'telemetry:compute_statistics': u'', 'telemetry:get_samples': u''} | 16:25 |
alextricity25 | So ceilometer flops when trying to parse that | 16:25 |
alextricity25 | Does anybody know of a way to convert that to a format without the unicode strings? | 16:26 |
*** gfa has joined #openstack-ansible | 16:27 | |
*** dxiri has quit IRC | 16:32 | |
*** manheim has joined #openstack-ansible | 16:33 | |
odyssey4me | alextricity25 I think there's a filter that does that | 16:33 |
odyssey4me | to_nice_json or something | 16:33 |
*** dxiri has joined #openstack-ansible | 16:34 | |
*** Mannheim has joined #openstack-ansible | 16:37 | |
*** ThiagoCMC has joined #openstack-ansible | 16:37 | |
*** manheim has quit IRC | 16:37 | |
*** qcrivera has quit IRC | 16:37 | |
alextricity25 | odyssey4me: Yup. that's it! Thanks! I was using to_json, but that was escaping all of the double quotes and printing the escaped quotes | 16:38 |
alextricity25 | i.e. "{\"telemetry:create_samples\": \"\", \"telemetry:query_sample\": \"\", \"telemetry:get_meters\": \"\", \"segregation\": \"rule:context_is_admin\", \"context_is_admin\": \"role:admin\" | 16:39 |
odyssey4me | yeah, it thinks it's a string | 16:39 |
odyssey4me | which it is | 16:39 |
alextricity25 | Ah, that makes sense | 16:41 |
*** dxiri has quit IRC | 16:44 | |
ThiagoCMC | Hey guys, is it possible to deploy Nova LXD on compute nodes deployed by OSA? | 16:44 |
ThiagoCMC | I want to create a host aggregate with Nova LXD... | 16:45 |
ThiagoCMC | And not the whole cloud | 16:45 |
odyssey4me | ThiagoCMC yes | 16:46 |
ThiagoCMC | Just found this: https://medium.com/walmartlabs/deploying-nova-lxd-hypervisors-with-openstack-ansible-39525157879d but, looks like that it is for the whole thing... | 16:46 |
*** dxiri has joined #openstack-ansible | 16:46 | |
ThiagoCMC | odyssey4me, niiice! | 16:46 |
ThiagoCMC | :-D | 16:46 |
ThiagoCMC | Is there a way to have 2 AZs (2 Host Aggregates), 1 with KVM, another with LXD? | 16:46 |
odyssey4me | yes | 16:47 |
ThiagoCMC | Awesome! | 16:47 |
*** udesale__ has quit IRC | 16:47 | |
logan- | o/ | 16:48 |
logan- | ThiagoCMC: you could use group vars to split out the compute_hosts group to a kvm_compute_hosts and a lxd_compute_hosts | 16:48 |
ThiagoCMC | Hmm... Thanks! | 16:49 |
odyssey4me | ThiagoCMC this may be useful to you: https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiVkpiep8DUAhUsIcAKHfdEC_YQFggoMAA&url=https%3A%2F%2Fmedium.com%2Fwalmartlabs%2Fdeploying-nova-lxd-hypervisors-with-openstack-ansible-39525157879d&usg=AFQjCNHeeVQdp3a8-befaUI7FsPHc7uIVg&sig2=yVSYFQHg3QOjHVykao7YwQ | 16:49 |
odyssey4me | dammit | 16:49 |
odyssey4me | https://medium.com/walmartlabs/deploying-nova-lxd-hypervisors-with-openstack-ansible-39525157879d | 16:49 |
ThiagoCMC | lol | 16:50 |
ThiagoCMC | Tks! | 16:50 |
odyssey4me | I would suggest just starting with a dedicated region for a test environment - get familiar with it, then try a mixed hypervisor setup | 16:50 |
logan- | i do a similar thing where i have some hypervisors that configure nova to use rbd backed storage, and other hypervisors that use local disk storage. let me see if i can pull an example together for this | 16:50 |
ThiagoCMC | You guys are very helpful! | 16:51 |
ThiagoCMC | ^_^ | 16:51 |
*** woodard has joined #openstack-ansible | 16:51 | |
odyssey4me | cloudnull you are terrible: https://cloudnull.io/2017/05/rolling-restart-using-openstack-ansible/ | 16:51 |
*** woodard has quit IRC | 16:52 | |
*** dxiri has quit IRC | 16:52 | |
odyssey4me | ThiagoCMC this may be useful - not the same hypervisor mix, but conceptually useful: https://cloudnull.io/2016/11/multi-hypervisor-openstack-cloud/ | 16:52 |
ThiagoCMC | That's cool! | 16:53 |
*** woodard has joined #openstack-ansible | 16:53 | |
ThiagoCMC | Thank you! | 16:53 |
*** dxiri has joined #openstack-ansible | 16:55 | |
*** Mannheim has quit IRC | 17:05 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Set PrivateDevices=false for MemcacheD https://review.openstack.org/474583 | 17:06 |
*** n3v3rm0r3r has quit IRC | 17:07 | |
*** Manheim has joined #openstack-ansible | 17:07 | |
*** whiteveil has quit IRC | 17:10 | |
*** dxiri_ has joined #openstack-ansible | 17:12 | |
*** dxiri has quit IRC | 17:13 | |
*** pcaruana has quit IRC | 17:16 | |
*** greghaynes is now known as greghayn1 | 17:18 | |
*** n3v3rm0r3r has joined #openstack-ansible | 17:19 | |
mgariepy | ho the diff between the galera-server test and the roles tests it that galera_server install an old version first. which is not started by systemd. | 17:21 |
mgariepy | then upgrade it. | 17:22 |
*** bauruine has quit IRC | 17:23 | |
*** greghayn1 has quit IRC | 17:25 | |
cloudnull | odyssey4me: you like that :) | 17:29 |
*** Mannheim has joined #openstack-ansible | 17:29 | |
cloudnull | I had to do that to the osic when we changed the domain name | 17:29 |
*** dxiri_ has quit IRC | 17:30 | |
*** bauruine has joined #openstack-ansible | 17:30 | |
*** craigs_ has joined #openstack-ansible | 17:31 | |
*** Manheim has quit IRC | 17:32 | |
*** dxiri has joined #openstack-ansible | 17:36 | |
*** whiteveil has joined #openstack-ansible | 17:41 | |
mgariepy | mhayden, odyssey4me , i'm debugging galera stuff on centos | 17:41 |
*** whiteveil_ has joined #openstack-ansible | 17:45 | |
*** whiteveil has quit IRC | 17:45 | |
*** whiteveil_ is now known as whiteveil | 17:45 | |
*** shardy has quit IRC | 17:47 | |
*** dxiri has quit IRC | 17:51 | |
mgariepy | mhayden, odyssey4me are you still around ? | 17:51 |
*** dxiri has joined #openstack-ansible | 17:51 | |
mhayden | yes, but a bit sidetracked right now | 17:51 |
logan- | ThiagoCMC: https://gist.github.com/Logan2211/88a591fcf5ea54d0f17cc98123810851 something like that to split the compute_hosts group up should work. | 17:52 |
odyssey4me | yep | 17:52 |
mgariepy | mhayden, the issue with musql is that it's starting correctly via systemd without having galera_new_cluster executed. | 17:52 |
mgariepy | http://logs.openstack.org/00/474500/1/check/gate-openstack-ansible-tests-ansible-func-centos-7/a82abb2/logs/ara/ | 17:52 |
mgariepy | task :: galera_server : Start galera clusterinfra1command0:12:250:00:00SKIPPED | 17:53 |
ThiagoCMC | logan-, wooooow!!!! :-D | 17:53 |
ThiagoCMC | Looks so simple! | 17:53 |
ThiagoCMC | O_O | 17:53 |
logan- | ThiagoCMC: then in openstack_user_config, add hosts to lxd-compute_hosts and kvm-compute_hosts instead of compute_hosts | 17:54 |
ThiagoCMC | :-O | 17:54 |
ThiagoCMC | Thank you so much! | 17:54 |
logan- | no probs. my setup is a bit different from that so it might not work out of the box, but if you run into problems ping away and ill try to help | 17:55 |
cloudnull | we should just do something like this by default ^ | 17:55 |
cloudnull | we kinda have that for ironic now | 17:55 |
ThiagoCMC | You guys are awsome! | 17:55 |
cloudnull | it'd be nice to just expand the pattern | 17:55 |
ThiagoCMC | cloudnull, like for Hyper-V Compute Nodes, ESXi, Xen, etc? | 17:56 |
cloudnull | ++ | 17:56 |
cloudnull | for ironic we do this https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/env.d/ironic.yml#L60-L65 | 17:56 |
*** admin0 has joined #openstack-ansible | 17:56 | |
cloudnull | which is similar to what logan- has | 17:57 |
logan- | i think i used the ironic skel as a reference actually | 17:57 |
openstackgerrit | Merged openstack/openstack-ansible master: Remove localhost host_vars file https://review.openstack.org/474567 | 18:10 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/ocata: Correct heat developer mode constraint https://review.openstack.org/474720 | 18:14 |
odyssey4me | I'd appreciate a quick review for ^ - it's a very simple patch. | 18:14 |
odyssey4me | cloudnull logan- ^ | 18:16 |
odyssey4me | jmccrory ^ | 18:16 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add breakout groups for hypervisor types https://review.openstack.org/474721 | 18:16 |
cloudnull | logan-: ThiagoCMC ^ | 18:16 |
cloudnull | odyssey4me: looking | 18:16 |
odyssey4me | lol, thanks all of you | 18:17 |
odyssey4me | now we hope the heat gate is working | 18:17 |
cloudnull | much workflow | 18:17 |
cloudnull | such code-review | 18:17 |
cloudnull | :) | 18:17 |
*** dxiri has quit IRC | 18:22 | |
*** dxiri has joined #openstack-ansible | 18:23 | |
*** stuartgr has quit IRC | 18:24 | |
*** electrofelix has quit IRC | 18:24 | |
*** rstarmer has joined #openstack-ansible | 18:25 | |
*** lbragstad has quit IRC | 18:28 | |
*** qcrivera has joined #openstack-ansible | 18:28 | |
*** vnogin has quit IRC | 18:29 | |
*** rstarmer has quit IRC | 18:36 | |
*** lbragstad has joined #openstack-ansible | 18:36 | |
*** manheim has joined #openstack-ansible | 18:37 | |
*** NWOutlier_ has joined #openstack-ansible | 18:39 | |
*** rstarmer has joined #openstack-ansible | 18:40 | |
odyssey4me | mgariepy hmm, but it's not starting that way because there's only one galera host | 18:46 |
odyssey4me | oh, maybe not | 18:47 |
odyssey4me | it's possibly because 'Start mysql node normally' succeeded? | 18:48 |
odyssey4me | see https://github.com/openstack/openstack-ansible-galera_server/blob/master/tasks/galera_bootstrap.yml#L42 | 18:48 |
mgariepy | yes | 18:51 |
*** dxiri has quit IRC | 18:52 | |
odyssey4me | oh I see | 18:53 |
odyssey4me | so it starts normally, which means the cluster is never initialised | 18:53 |
odyssey4me | in a multi-node situation there would be failures and we're covered there | 18:53 |
odyssey4me | so perhaps one of those tasks need an OR condition for when there's only one node | 18:53 |
mgariepy | yes I guess so. | 18:54 |
odyssey4me | or perhaps we add https://github.com/openstack/openstack-ansible-galera_server/commit/ff4e9c6ece0849d08f36dda93315181fe5a75339 to that initial task (or something like it) | 18:55 |
odyssey4me | this whole role needs simplification | 18:57 |
odyssey4me | it's way too complicated - difficult to follow | 18:57 |
openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible-os_glance master: Implement db sync and restart facts for upgrades https://review.openstack.org/474299 | 18:58 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add breakout groups for hypervisor types https://review.openstack.org/474721 | 19:00 |
mgariepy | +3 on thos odyssey4me | 19:01 |
mgariepy | :) | 19:01 |
*** dxiri has joined #openstack-ansible | 19:05 | |
mgariepy | odyssey4me, it's really confusing to me haha | 19:05 |
mgariepy | galera_server works because of the galera using init.d instead of systemd and cannot start with the initd can we have a better test to know if there is a cluster up ? | 19:06 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server master: Split user create and ssh key generation https://review.openstack.org/474730 | 19:06 |
mgariepy | and it works on ubuntu because we pressed debconf stuff here https://github.com/openstack/openstack-ansible-galera_server/blob/master/tasks/galera_install_apt.yml#L83-L91 | 19:08 |
*** lucasxu has quit IRC | 19:10 | |
*** lucasxu has joined #openstack-ansible | 19:11 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone master: Split user create and ssh key generation https://review.openstack.org/474734 | 19:12 |
odyssey4me | ok, I need to get outta here | 19:14 |
odyssey4me | too many late nights trying ot work this out | 19:14 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add breakout groups for hypervisor types https://review.openstack.org/474721 | 19:14 |
odyssey4me | keep at it mgariepy - hopefully mhayden can help you out some time today too | 19:14 |
mgariepy | i'll try to submit a working patch talk to you tomorrow. | 19:15 |
odyssey4me | it's driving me crazy | 19:15 |
odyssey4me | thank you so much for keeping at it | 19:15 |
*** dxiri has quit IRC | 19:22 | |
*** vnogin has joined #openstack-ansible | 19:25 | |
*** dxiri has joined #openstack-ansible | 19:26 | |
*** vnogin has quit IRC | 19:29 | |
alextricity25 | cloudnull: You around? | 19:30 |
cloudnull | depends :) | 19:30 |
alextricity25 | hehe | 19:30 |
alextricity25 | Question about v2_config_template | 19:30 |
alextricity25 | more specificly the json bits | 19:30 |
alextricity25 | specifically* | 19:30 |
cloudnull | and then ? | 19:31 |
alextricity25 | cloudnull: Do you think we need to specifiy an encoding here: https://github.com/openstack/openstack-ansible-plugins/blob/master/action/_v2_config_template.py#L363? The module splits out unicode strings into policy files, which causes parsing issues :/ | 19:32 |
palendae | alextricity25: Yes | 19:32 |
palendae | I have an example | 19:32 |
palendae | https://review.openstack.org/#/c/471531/ | 19:32 |
alextricity25 | I love examples | 19:32 |
palendae | Unicode writing - encode. Unicode reading - decode | 19:33 |
palendae | Note, though, that I was super lazy and just did ascii for all that stuff | 19:33 |
alextricity25 | So we need an encode('ascii') there, it seems | 19:33 |
*** manheim has quit IRC | 19:33 | |
cloudnull | alextricity25: what does the override look like? | 19:34 |
cloudnull | and are you running w/ py3? | 19:34 |
palendae | the .encode will work in p2.7 | 19:34 |
palendae | But yeah, curiuous what you're giving it | 19:34 |
alextricity25 | cloudnull: In this case there is no override. The module is being fed content via the 'content' attribute. So the module still processes the string | 19:34 |
alextricity25 | Let me show you the task | 19:34 |
palendae | If it's got UTF-8 somehow the ascii encoding won't work | 19:34 |
alextricity25 | sec.. | 19:34 |
* cloudnull f5 | 19:35 | |
* cloudnull f5 | 19:35 | |
* cloudnull f5 | 19:35 | |
cloudnull | :P | 19:35 |
*** vnogin has joined #openstack-ansible | 19:35 | |
alextricity25 | cloudnull: So demanding! | 19:35 |
alextricity25 | cloudnull: https://github.com/openstack/openstack-ansible-os_ceilometer/blob/master/tasks/ceilometer_post_install.yml#L75-L78 | 19:35 |
palendae | Why isn't it here yet? | 19:35 |
* cloudnull f5 | 19:35 | |
* cloudnull f5 | 19:35 | |
cloudnull | hahaha | 19:35 |
cloudnull | when will then be now!? | 19:35 |
palendae | Hm, that looks suspiciously like everything else | 19:35 |
palendae | I mean, the encode call shouldn't hurt | 19:36 |
alextricity25 | All of the other services don't pull down policy content from git though. I think ceilometer is the only one that does it. The other services get the content from templates | 19:36 |
palendae | Ahhhhh | 19:36 |
palendae | That may be it | 19:36 |
cloudnull | are we seeing that across all services? | 19:37 |
*** tobberydberg has joined #openstack-ansible | 19:37 | |
alextricity25 | cloudnull: I'm only seeing this with ceilometer/policy.json. It looks like the other services get the content from templates, which is probably why they aren't having this problem. | 19:37 |
alextricity25 | Let me show you what it looks like currently.. | 19:38 |
cloudnull | this makes sense | 19:38 |
cloudnull | I'd give the encoding a try | 19:38 |
cloudnull | as palendae mentioned | 19:38 |
alextricity25 | {'telemetry:create_samples': u'', 'telemetry:query_sample': u'', 'telemetry:get_resources': u'', 'telemetry:get_meters': u'', 'segregation': u'rule:context_is_admin', 'telemetry:get_resource': u'', 'context_is_admin': u'role:admin', 'telemetry:get_sample': u'', 'telemetry:compute_statistics': u'', 'telemetry:get_samples': u''} | 19:38 |
alextricity25 | ^ that's written out into the policy file | 19:38 |
palendae | Interesting | 19:38 |
palendae | Yeah, I'd guess it has something to do with coming from git | 19:38 |
palendae | vs our own templates | 19:39 |
alextricity25 | I'll write up a patch to add the encode there | 19:39 |
alextricity25 | hopefully that does the trick. | 19:39 |
cloudnull | ++ | 19:39 |
palendae | I didn't turn on failing for python3 compatability because our tox envs are kinda weird | 19:39 |
palendae | Can't just do tox -e py27,py36 without redefining stuff | 19:40 |
openstackgerrit | Merged openstack/openstack-ansible master: Set PrivateDevices=false for MemcacheD https://review.openstack.org/474583 | 19:42 |
*** dxiri has quit IRC | 19:45 | |
alextricity25 | Are these modules cached anywhere? I'm changing vi /root/.ansible/plugins/action/_v2_config_template.py locally but it doesn't make a difference. | 19:45 |
*** dxiri has joined #openstack-ansible | 19:45 | |
palendae | alextricity25: Maybe drop an import pdb; pdb.set_trace() there to see if it works? | 19:48 |
palendae | May also be in the ansible runtime venv | 19:48 |
*** dxiri has quit IRC | 19:49 | |
*** pbandark has quit IRC | 19:53 | |
alextricity25 | palendae: Will the set_trace stop the playbook execution? | 19:54 |
palendae | alextricity25: Should | 19:54 |
palendae | set_trace starts an interactive prompt | 19:54 |
palendae | Could toss that into a blank py file and run it to see what happens | 19:55 |
alextricity25 | palendae: Hmm...stange. It's not doing anything for me when I add it in /root/.ansible/plugins/action/_v2_config_template.py | 19:55 |
palendae | Are you on master? | 19:56 |
alextricity25 | palendae: Ocata | 19:56 |
openstackgerrit | Marc Gariépy proposed openstack/openstack-ansible-galera_server master: Move PrivateDevices before mysql password set https://review.openstack.org/474760 | 19:57 |
mgariepy | mhayden, ^^ | 19:57 |
alextricity25 | palendae: this json.dumps code must not be running at all | 19:58 |
palendae | alextricity25: Maybe try in /etc/ansible/roles/plugins/action? | 19:58 |
alextricity25 | maybe my understanding of the 'content' attribute is wrong | 19:58 |
palendae | https://github.com/openstack/openstack-ansible/blob/stable/ocata/scripts/openstack-ansible.rc#L27 | 19:59 |
palendae | Could also move that pdb line up to the top of the module and see if it's happening | 19:59 |
palendae | Do you have a stack trace? | 19:59 |
alextricity25 | No :/ I'm still working at it. it looks like the module reads the content, writes it out to a tempfile(using encode()), then set's it to the 'source' attribute | 20:02 |
palendae | Hm | 20:03 |
palendae | I haven't looked closely at that code | 20:03 |
openstackgerrit | Marc Gariépy proposed openstack/openstack-ansible-tests master: [test] do not merge (it's pointless) https://review.openstack.org/474765 | 20:04 |
mgariepy | on that, have a good evening everyone | 20:05 |
palendae | alextricity25: https://gist.github.com/nrb/4f9b76c7074782073a4754d96d36ccb8 is what I got dropping a pdb at the top of the file in /etc/ansible/roles/plugins/actions then running os-install-keystone | 20:09 |
palendae | So it should interrupt the playbook and show you the pdb shell | 20:09 |
alextricity25 | palendae: hmm...my playbook execution failed after showing the pdb shell, then kicked me out :( | 20:10 |
palendae | That's a start | 20:10 |
palendae | But also I'm on master | 20:10 |
alextricity25 | lol | 20:10 |
palendae | At least now you're seeing which file's getting run | 20:10 |
alextricity25 | looks like you're running an intergraded build too | 20:10 |
*** qcrivera has quit IRC | 20:10 | |
palendae | I am | 20:10 |
palendae | Well | 20:10 |
palendae | Kind of | 20:10 |
palendae | I did setup-hosts, setup-infra, setup-keystone | 20:11 |
palendae | And now working on geting the glance playbooks to do rolling upgrades | 20:11 |
palendae | But I had keystone there so that's what I tried | 20:11 |
*** manheim has joined #openstack-ansible | 20:24 | |
*** thorst has quit IRC | 20:27 | |
*** manheim has quit IRC | 20:28 | |
odyssey4me | cloudnull you were looking into https://bugs.launchpad.net/openstack-ansible/+bug/1696802 at some point? | 20:28 |
openstack | Launchpad bug 1696802 in openstack-ansible "lxc-containers-create fails when run on a host with existing containers" [Critical,Confirmed] | 20:28 |
odyssey4me | I'm seeing it on newton minor upgrades :/ | 20:28 |
*** dxiri has joined #openstack-ansible | 20:29 | |
cloudnull | i've not | 20:30 |
cloudnull | I've seen chatter on it | 20:30 |
cloudnull | but spent no time looking into why | 20:30 |
*** rstarmer has quit IRC | 20:31 | |
*** dxiri has quit IRC | 20:37 | |
*** dxiri has joined #openstack-ansible | 20:39 | |
* odyssey4me disappears again | 20:40 | |
palendae | Hm, there's no filter/test like |changed but for brand new information, is there? I need to distinguish from a brand new glance install and an updated one | 20:41 |
alextricity25 | I think odyssey4me used local facts for that kind of stuff | 20:42 |
alextricity25 | on the cinder ps | 20:42 |
palendae | Yyyyeah | 20:42 |
palendae | Well, glance needs some more logic :( | 20:42 |
alextricity25 | bleh | 20:42 |
palendae | It's db sync command has no --check flag | 20:42 |
palendae | afaict it has no way to communicate to me that a migration is necessary unless I figure it out myself | 20:43 |
palendae | Right now my change tries to sync, expand, migrate, and contract all at once | 20:43 |
palendae | But if sync was performed, I don't need the others | 20:43 |
logan- | i think cinder is the same way. iirc we were discussing storing a local fact for the last deployed tag and running migrations if the tag we're deploying is newer than the stored fact | 20:43 |
palendae | Ok, maybe I missed that | 20:44 |
palendae | Cinder doesn't have expand/migrate/contract yet | 20:44 |
palendae | Keystone does, but it also uses db sync --check to signal if they're needed | 20:44 |
logan- | gotcha | 20:44 |
palendae | https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tasks/keystone_db_setup.yml#L31 - doesn't exist in the glance stuff | 20:45 |
*** smatzek has quit IRC | 20:46 | |
palendae | So yeah, maybe I need to write local facts with the version of glance installed =\ | 20:46 |
palendae | Ah, I can use the venv checksum | 20:47 |
palendae | Cool | 20:47 |
palendae | Although changes could have happened that don't necessitate migrations... | 20:49 |
palendae | I think I'll go ask the glance devs | 20:49 |
logan- | yeah its kind of impossible to know really if they don't expose a way to check. pretty much just have to assume we need to run migrations any time we roll a new venv imo | 20:50 |
*** penguintamer has joined #openstack-ansible | 20:50 | |
alextricity25 | palendae: So I found where it writes to the file, and it does look like it's already using encode: https://github.com/openstack/openstack-ansible-plugins/blob/stable/ocata/action/_v2_config_template.py#L449 | 20:50 |
alextricity25 | I'm not sure why it would still write out unicode indicators | 20:51 |
*** dxiri has quit IRC | 20:52 | |
alextricity25 | ugh. I'm going to have to come back to this | 20:53 |
*** dxiri has joined #openstack-ansible | 20:53 | |
*** lucasxu has quit IRC | 20:55 | |
*** hamzy has quit IRC | 20:58 | |
*** thorst has joined #openstack-ansible | 20:59 | |
*** dxiri has quit IRC | 21:01 | |
palendae | alextricity25: Ahh, so the issue is it's writing strings like u'thing' instead of 'thing'? | 21:01 |
*** klamath has quit IRC | 21:01 | |
palendae | logan-: Yeah, which would be fine, except when I run migrations that aren't necessary, it raises an exception | 21:02 |
palendae | And I don't wanna blanket ignore errors | 21:02 |
logan- | yuck | 21:02 |
palendae | I'll jump in their channel and bug them about it tomorrow though | 21:02 |
logan- | yeah | 21:02 |
palendae | Maybe that can still be added for Pike | 21:02 |
logan- | what do we do right now? ignore them? | 21:02 |
palendae | For glance? looks like >.> | 21:02 |
palendae | But also, in Ocata they changed their migration model | 21:02 |
NWOutlier_ | I have a newbie question; We've completed an OSA 15.1.5 - with Ocata: When attempting to launch an instance we get The requested availability zone is not available (HTTP 400) (Request-ID: req-13fd98cf-2311-4ddc-8c50-75dfb30cd84b) ... In the past, I don't recall needed the availability zones to be configured, or I assumed there would be demo availability zones created during hte OSA execution. Anyone else run | 21:03 |
NWOutlier_ | into this? | 21:03 |
palendae | logan-: https://github.com/openstack/openstack-ansible-os_glance/blob/master/tasks/glance_db_setup.yml | 21:03 |
palendae | vs their docs: https://docs.openstack.org/developer/glance/db.html | 21:03 |
palendae | Which don't mention checks...you'd have to look at their migrations directories from what I can tell | 21:04 |
logan- | i guess db_sync doesn't throw an exception if there are no migrations to do | 21:04 |
palendae | It does for me | 21:04 |
palendae | It's complaining cause there's duplicate info | 21:04 |
palendae | e.g. columns already exist | 21:04 |
palendae | Same with expand right after sync (because I'm on the latest already) | 21:05 |
logan- | we aren't ignoring errors there though | 21:05 |
palendae | Oh, with their old one you mean | 21:05 |
palendae | Yeah, good point | 21:05 |
logan- | yeah | 21:05 |
palendae | I'm using master | 21:05 |
palendae | Specifically, https://review.openstack.org/#/c/474299/ and whatever it checks out | 21:06 |
logan- | i see | 21:06 |
*** smatzek has joined #openstack-ansible | 21:06 | |
logan- | seems like a bug in glance for sure. they ought to either provide a check or at the very least not fail if no migrations are needed | 21:07 |
palendae | https://gist.github.com/nrb/6f3788c333f717e601b0ded35137acee | 21:07 |
palendae | Yep | 21:07 |
palendae | I'll poke around their channel tomorrow and ask about it | 21:08 |
logan- | cool | 21:08 |
palendae | I also didn't look at their bug tracker/open patches | 21:08 |
*** bauruine has quit IRC | 21:08 | |
*** tobberydberg has quit IRC | 21:10 | |
*** bauruine has joined #openstack-ansible | 21:10 | |
*** tobberydberg has joined #openstack-ansible | 21:10 | |
*** michaelb has joined #openstack-ansible | 21:12 | |
*** tobberydberg has quit IRC | 21:15 | |
*** woodard_ has joined #openstack-ansible | 21:16 | |
*** qcrivera has joined #openstack-ansible | 21:19 | |
*** woodard has quit IRC | 21:19 | |
*** woodard_ has quit IRC | 21:20 | |
*** smatzek has quit IRC | 21:22 | |
*** vnogin has quit IRC | 21:22 | |
*** oanson has quit IRC | 21:23 | |
*** gouthamr has quit IRC | 21:23 | |
*** oanson has joined #openstack-ansible | 21:26 | |
*** whiteveil has quit IRC | 21:29 | |
*** Mannheim has quit IRC | 21:31 | |
*** manheim has joined #openstack-ansible | 21:38 | |
openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: Move unsetting of group and host env vars https://review.openstack.org/474787 | 21:39 |
*** vnogin has joined #openstack-ansible | 21:41 | |
*** gouthamr has joined #openstack-ansible | 21:42 | |
*** esberglu has quit IRC | 21:43 | |
*** esberglu has joined #openstack-ansible | 21:44 | |
*** yolanda has quit IRC | 21:45 | |
*** esberglu has quit IRC | 21:48 | |
*** cpuga has quit IRC | 21:53 | |
*** eaxiom has joined #openstack-ansible | 21:57 | |
eaxiom | hi all | 21:57 |
eaxiom | i do not want to use a separate network for storage | 21:58 |
eaxiom | is it still required to mention a storage network in the openstack_user_config.yml file ? | 21:58 |
*** eaxiom has quit IRC | 22:04 | |
*** manheim has quit IRC | 22:09 | |
*** askb has joined #openstack-ansible | 22:12 | |
*** deadnull has joined #openstack-ansible | 22:14 | |
*** thorst has quit IRC | 22:14 | |
*** woodard has joined #openstack-ansible | 22:18 | |
*** woodard has quit IRC | 22:18 | |
*** woodard has joined #openstack-ansible | 22:19 | |
*** kylek3h has quit IRC | 22:27 | |
*** admin0 has quit IRC | 22:32 | |
*** thorst has joined #openstack-ansible | 22:33 | |
*** thorst has quit IRC | 22:34 | |
*** oanson has quit IRC | 22:38 | |
*** esberglu has joined #openstack-ansible | 22:38 | |
*** schwicht has quit IRC | 22:39 | |
*** oanson has joined #openstack-ansible | 22:39 | |
*** bauruine has quit IRC | 22:42 | |
*** esberglu has quit IRC | 22:43 | |
*** bauruine has joined #openstack-ansible | 22:43 | |
*** deadnull has quit IRC | 22:54 | |
*** schwicht has joined #openstack-ansible | 22:54 | |
*** esberglu has joined #openstack-ansible | 22:54 | |
*** acormier has quit IRC | 23:03 | |
*** acormier has joined #openstack-ansible | 23:04 | |
*** acormier has quit IRC | 23:08 | |
*** stewie925 has quit IRC | 23:13 | |
*** schwicht has quit IRC | 23:20 | |
*** bauruine has quit IRC | 23:22 | |
*** schwicht has joined #openstack-ansible | 23:22 | |
*** phalmos has quit IRC | 23:24 | |
*** schwicht has quit IRC | 23:26 | |
*** bauruine has joined #openstack-ansible | 23:30 | |
*** vnogin has quit IRC | 23:44 | |
*** thorst has joined #openstack-ansible | 23:53 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!