| cloudnull | jmccrory automagically: so for the tests repo we're wanting to keep the assert tests in the roles but move all of the exercise tasks into the central repo right ? | 00:00 |
|---|---|---|
| jmccrory | my understanding was central repo for any plays used by multiple roles | 00:05 |
| cloudnull | ok | 00:05 |
| cloudnull | makes sense to m e | 00:05 |
| cloudnull | *me | 00:05 |
| *** eil397__ has joined #openstack-ansible | 00:06 | |
| cloudnull | are you just cloning the test repo and then running the common bits as includes leaving the group_vars, inventory, etc in the role specific repo? | 00:07 |
| * cloudnull wants to make sure I'm implementing a pattern that makes sense | 00:07 | |
| *** david-lyle has quit IRC | 00:08 | |
| *** Guest20708 has quit IRC | 00:08 | |
| *** eil397_ has quit IRC | 00:09 | |
| *** eil397 has quit IRC | 00:09 | |
| cloudnull | because {host,group}_vars all exist within the IRR, curious if we're moving those bits too | 00:09 |
| *** hblah has joined #openstack-ansible | 00:12 | |
| *** gouthamr_ has joined #openstack-ansible | 00:13 | |
| *** eil397__ has quit IRC | 00:15 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests: Update for openstack-hosts https://review.openstack.org/360129 | 00:16 |
| *** gouthamr has quit IRC | 00:16 | |
| *** eil397_ has joined #openstack-ansible | 00:17 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated to use the openstack-ansible-tests repo https://review.openstack.org/360130 | 00:17 |
| jmccrory | cloudnull: yep, cloning the repo as part of the ansible commands in tox, created a new inventory, groups_vars, and a few test var files | 00:17 |
| cloudnull | idk if im doing that right but there's my first pass | 00:17 |
| jmccrory | https://review.openstack.org/#/c/355155/ has been my only attempt so far | 00:18 |
| jmccrory | oh ok, i put the clone inside tox.ini https://review.openstack.org/#/c/355155/11/tox.ini, i think odyssey4me mentioned wanting to bring in the tests repo some other way at the midcycle but can't remember what that was now | 00:20 |
| * odyssey4me lurks | 00:32 | |
| *** kylek3h has joined #openstack-ansible | 00:37 | |
| cloudnull | odyssey4me: clone from tox or test playbook ? | 00:41 |
| cloudnull | -cc automagically ^ | 00:41 |
| automagically | cloudnull: I’m doing it in tox too | 00:41 |
| automagically | Following jmccrory lead for the moment. I don’t recall the discussion during mid-cycle | 00:42 |
| cloudnull | ^ me neither. | 00:42 |
| cloudnull | if that's the pattern i'll adjust . | 00:42 |
| cloudnull | can we add something to git ignore or clone outside of the test repo? | 00:43 |
| cloudnull | maybe in "{homedir}/.ansible/test-playbooks" ? | 00:45 |
| cloudnull | then those playbooks wont dirty the git tree when we run a test? | 00:45 |
| pabelanger | cloudnull: So, were you having APT packaging issues in openstack-ansible jobs? | 00:46 |
| pabelanger | recently having issues | 00:46 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-tests: Don't use Rabbit SSL for role tests https://review.openstack.org/360142 | 00:46 |
| cloudnull | or we could add to gitignore "tests/playbooks" | 00:46 |
| cloudnull | pabelanger: we were | 00:46 |
| pabelanger | cloudnull: what did you do to fix it? | 00:47 |
| automagically | cloudnull: ++ .gitignore | 00:47 |
| cloudnull | we had issues with libssl1.0.0 | 00:47 |
| pabelanger | stale package? | 00:47 |
| automagically | pabelanger: Thx for checking in on that issue | 00:47 |
| cloudnull | which was because the mirror was behind what the lxc image cache was providing us | 00:47 |
| cloudnull | it was kinda a stale package | 00:48 |
| pabelanger | Ya, so turns out our mirror.ubuntu volume wasn't released in 4 days. I had to manually kick it off | 00:48 |
| pabelanger | so, in fact, your package issues were likely exposing that but didn't make the connection until today | 00:49 |
| pabelanger | So, if you have the problem again, feel free to ping openstack-infra to confirm our mirrors are in sync! | 00:49 |
| cloudnull | ah. likley . | 00:49 |
| cloudnull | will do . | 00:49 |
| cloudnull | i pinged last night about it however i just thought ubuntu was doing something stupid with their images. | 00:50 |
| pabelanger | Ah | 00:50 |
| pabelanger | I must have missed it | 00:50 |
| cloudnull | so we can revert this https://review.openstack.org/#/c/359586/ | 00:50 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts: Revert "[GATE UNBLOCK] Purge/reinstall broken package from the xenial base image" https://review.openstack.org/360143 | 00:50 |
| cloudnull | ^ that | 00:50 |
| pabelanger | Ya | 00:50 |
| cloudnull | pabelanger: thanks ! | 00:50 |
| pabelanger | try a revert | 00:50 |
| cloudnull | yea it'll go through the build process and if it works then "great success" :) | 00:51 |
| pabelanger | I'm going to add datetime file to http://mirror.dfw.rax.openstack.org/ubuntu/ with info about our last sync times | 00:51 |
| pabelanger | so projects can quickly check how stale it is | 00:51 |
| cloudnull | cool! | 00:51 |
| odyssey4me | oh, nice idea pabelanger - thanks for following up on it | 00:52 |
| odyssey4me | in fact a datetime file in all mirrors is probably a good idea | 00:53 |
| odyssey4me | ie part of the update process | 00:53 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated to use the openstack-ansible-tests repo https://review.openstack.org/360130 | 00:53 |
| pabelanger | Ya, for what ever reason I see that patch but didn't think there was an issue in openstack-infra | 00:53 |
| pabelanger | odyssey4me: yes, should be easy to add. Will hack on it tomorrow | 00:53 |
| odyssey4me | pabelanger hopefully with https://review.openstack.org/354203 too :) | 00:54 |
| automagically | pabelanger: You folks in infra rock, really appreciate the support and follow up | 00:54 |
| *** markvoelker has joined #openstack-ansible | 00:55 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_keystone: Use the centralized role tests repo https://review.openstack.org/360147 | 00:56 |
| cloudnull | jmccrory automagically if you have a moment -- https://review.openstack.org/#/c/360129/ | 00:56 |
| cloudnull | ah, nvmd | 00:56 |
| cloudnull | i have to change a thing | 00:56 |
| odyssey4me | cloudnull automagically jmccrory so I haven't thought too much about the tests repo and how best to implement its usage. I don't think anyone's established a pattern yet either. | 00:57 |
| cloudnull | jmccrory: has a pattern :) | 00:57 |
| odyssey4me | The only thing that I have thought about is that the role being tested should have its own inventory, its own install and test plays, and its own variables which are unique for the test needs). | 00:58 |
| odyssey4me | so yeah, I'm interested to see how that shakes out | 00:58 |
| automagically | odyssey4me: jmccrory pattern follows that model | 00:59 |
| odyssey4me | Once N3 is out I'll be revisiting getting cross-repo testing working - that may have to change things up a bit. | 00:59 |
| automagically | So be it, we’ll shake out the pattern a bit by getting a few more roles implemented | 00:59 |
| automagically | Alright, I’m officially done for the evening. Time for scotch….really time for scotch | 00:59 |
| automagically | Cheers y'all | 00:59 |
| odyssey4me | but I only expect that it will switch us from doing git clones to using zuul cloner... and we may have to do it outside of tox but I'm not certain yet | 00:59 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests: Update for openstack-hosts https://review.openstack.org/360129 | 01:00 |
| odyssey4me | automagically FYI thanks for the monasca repo patch - I've fixed it up and given my +1 | 01:00 |
| automagically | Appreciate the update on that odyssey4me | 01:01 |
| cloudnull | automagically jmccrory odyssey4me https://review.openstack.org/360129 -- if you have a moment | 01:01 |
| odyssey4me | automagically oh yeah, one more thing - that rabbit patch seems to fail at odd times, not always timing out | 01:01 |
| odyssey4me | you may want to dig into the logs to figure out if something isn't perhaps going wrong there | 01:01 |
| *** kypenguin has joined #openstack-ansible | 01:02 | |
| *** sdake has quit IRC | 01:06 | |
| *** eil397_ has quit IRC | 01:08 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests: Update for openstack-hosts https://review.openstack.org/360129 | 01:09 |
| *** thorst has quit IRC | 01:09 | |
| *** thorst has joined #openstack-ansible | 01:10 | |
| odyssey4me | cloudnull did you figure out the lxc stop/start issue on centos? I saw you mention a cgroups issue of some sort | 01:10 |
| cloudnull | no, i've not figured it out | 01:11 |
| *** jamielennox is now known as jamielennox|away | 01:11 | |
| cloudnull | we were looking into it | 01:11 |
| *** jamielennox|away is now known as jamielennox | 01:11 | |
| odyssey4me | it's likely a race condition of some kind, probably related to the kernel in use | 01:11 |
| cloudnull | it may be a regression or a race | 01:11 |
| odyssey4me | the issue was there before we updated to 2.0.4 FYI | 01:11 |
| odyssey4me | from 2.0.1 | 01:11 |
| odyssey4me | I wonder if a simply IO flush between the stop and the start would be good and help make that go away | 01:12 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated to use the openstack-ansible-tests repo https://review.openstack.org/360130 | 01:12 |
| mrda | odyssey4me: I was about to -1 https://review.openstack.org/#/c/360096/2 and you just +W'd it :( | 01:12 |
| odyssey4me | mrda what's wrong with it? | 01:12 |
| odyssey4me | ohh | 01:13 |
| cloudnull | odyssey4me: its possible? | 01:13 |
| odyssey4me | I see :) | 01:13 |
| mrda | openstack/openstack-ansible-os_ironic / examples/playbook.yml should have in is role: openstack-ansible-os_ironic not role: openstack-ansible-ironic | 01:13 |
| mrda | I was just ckicking -1 when the +W came thru :) | 01:13 |
| mrda | oh well | 01:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic: [DOCS] Syntax updates to the docs https://review.openstack.org/360096 | 01:13 |
| odyssey4me | mrda ok fixed now :) | 01:14 |
| odyssey4me | now you can +w if you like | 01:14 |
| mrda | I'll hit it | 01:14 |
| odyssey4me | good catch - I totally missed that | 01:14 |
| mrda | not a big deal | 01:15 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add play to deploy Rally to the utility containers https://review.openstack.org/357495 | 01:17 |
| *** thorst has quit IRC | 01:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-specs: Updated from global requirements https://review.openstack.org/360162 | 01:21 |
| openstackgerrit | Merged openstack/openstack-ansible-tests: Remove duplicate galera_server_id https://review.openstack.org/360110 | 01:21 |
| odyssey4me | cloudnull have you managed to put together a binary lookup yet? | 01:22 |
| cloudnull | no. we still need to decide on a bulk update pattern. I figured I'd wait for tomorrow. | 01:22 |
| cloudnull | w/ the meeting | 01:22 |
| odyssey4me | Why not just propose one to a single repo - then we can use that as a discussion point. Once it merges, we replicate. | 01:23 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated to use the openstack-ansible-tests repo https://review.openstack.org/360130 | 01:25 |
| *** markvoelker has quit IRC | 01:36 | |
| *** sdake has joined #openstack-ansible | 01:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-specs: Updated from global requirements https://review.openstack.org/360162 | 01:36 |
| *** galstrom_zzz is now known as galstrom | 01:37 | |
| *** sdake_ has joined #openstack-ansible | 01:39 | |
| cloudnull | any chance cores are around that might want to shove this through https://review.openstack.org/#/c/360129/ | 01:40 |
| cloudnull | which will let me continue w/ https://review.openstack.org/#/c/360130/ | 01:41 |
| cloudnull | and move on to the next setup-hosts role | 01:41 |
| *** sdake has quit IRC | 01:42 | |
| *** eil397_ has joined #openstack-ansible | 01:43 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible: Unbound DNS resolution containers https://review.openstack.org/336936 | 01:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Limit LXC hosts playbook to container hosts only https://review.openstack.org/360081 | 01:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic: [DOCS] Syntax updates to the docs https://review.openstack.org/360096 | 01:47 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 01:48 |
| cloudnull | also jmccrory odyssey4me if you have a moment https://review.openstack.org/#/c/345572/ | 01:49 |
| cloudnull | which will remove the ansible.cfg file | 01:49 |
| *** wadeholler has joined #openstack-ansible | 01:50 | |
| odyssey4me | cloudnull one query in that review | 01:52 |
| *** sdake_ has quit IRC | 01:54 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 01:54 |
| cloudnull | maybe missed in a rebase | 01:55 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Remove the ansible.cfg file https://review.openstack.org/345572 | 01:56 |
| cloudnull | odyssey4me: ^ added back | 01:56 |
| *** eil397_ has quit IRC | 01:59 | |
| *** thorst has joined #openstack-ansible | 01:59 | |
| odyssey4me | thanks LGTM | 01:59 |
| *** thorst has quit IRC | 01:59 | |
| *** thorst has joined #openstack-ansible | 02:00 | |
| *** sdake has joined #openstack-ansible | 02:03 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests: Add tests for the openstack_hosts role https://review.openstack.org/360129 | 02:04 |
| *** gouthamr_ is now known as gouthamr | 02:05 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Remove the ansible.cfg file https://review.openstack.org/345572 | 02:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 02:06 |
| odyssey4me | cloudnull ^ rebased those on top of the calico patch which is the blocker at this point | 02:07 |
| cloudnull | cool | 02:07 |
| *** thorst has quit IRC | 02:08 | |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts: Remove LVM config from openstack_hosts https://review.openstack.org/360077 | 02:12 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible: Unbound DNS resolution containers https://review.openstack.org/336936 | 02:13 |
| *** eil397_ has joined #openstack-ansible | 02:19 | |
| *** thorst has joined #openstack-ansible | 02:25 | |
| *** rackertom has quit IRC | 02:25 | |
| *** rackertom has joined #openstack-ansible | 02:27 | |
| *** thorst has quit IRC | 02:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Move container bind-mounts and command settings to the inventory https://review.openstack.org/347400 | 02:28 |
| *** wadeholler has quit IRC | 02:28 | |
| *** serverascode has quit IRC | 02:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Move container bind-mounts and command settings to the inventory https://review.openstack.org/347400 | 02:29 |
| *** serverascode has joined #openstack-ansible | 02:31 | |
| *** sdake has quit IRC | 02:34 | |
| *** sdake has joined #openstack-ansible | 02:35 | |
| *** jamesdenton has joined #openstack-ansible | 02:44 | |
| *** sdake has quit IRC | 02:45 | |
| *** sdake has joined #openstack-ansible | 02:47 | |
| odyssey4me | cloudnull does https://github.com/ansible/ansible/issues/17236 express the problem properly? | 02:56 |
| odyssey4me | I'll have to work up an example to go with it to illustrate the problem, but I wanted to get the basic description right first. | 02:56 |
| cloudnull | someone tried to implement this https://github.com/ansible/ansible/pull/6666 | 02:57 |
| cloudnull | it was rejected | 02:57 |
| odyssey4me | if it relates, maybe add a comment to the issue? | 02:57 |
| odyssey4me | cloudnull a quick review of https://review.openstack.org/360081 ? | 02:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Move container bind-mounts and command settings to the inventory https://review.openstack.org/347400 | 03:01 |
| odyssey4me | thanks - that should speed up multi-node deployments quite a bit! | 03:02 |
| *** jamielennox is now known as jamielennox|away | 03:09 | |
| *** sdake has quit IRC | 03:12 | |
| cloudnull | i still hate the bindmount things | 03:17 |
| cloudnull | i get we want to move things to group_vars but the pain that can and will cause isn't worth it in my opinion. | 03:18 |
| cloudnull | a few seconds savings (even if its a few min) on the initial deployment vs a disaster caused by variable collisions that goes totally unnoticed for half the life of a deployment isn't worth it imo | 03:19 |
| odyssey4me | it's not about saving time - it's about having all config change happen at once, reducing container restarts - thus creating a choke point on upgrades, and improving the success of initial deployments | 03:23 |
| *** jamielennox|away is now known as jamielennox | 03:23 | |
| odyssey4me | your concern is valid, but at this stage there is no problem - just the potential for a problem if we aren't vigilent about it | 03:23 |
| cloudnull | the config still isnt in one place | 03:25 |
| odyssey4me | not yet - that's a stepping stone | 03:25 |
| *** eil397 has joined #openstack-ansible | 03:25 | |
| cloudnull | nor can it be | 03:25 |
| cloudnull | the conditional includes wont allow it | 03:25 |
| odyssey4me | I'll be looking at the conditional bits each in turn and seeing if there's a way to get that right. | 03:26 |
| cloudnull | despite additional container restarts at the initial deployment time what does this solve. | 03:26 |
| cloudnull | 's/./?/ | 03:26 |
| odyssey4me | But for now this at least covers the large majority of the containers. It's only a few special cases that remain. | 03:26 |
| odyssey4me | not just initial deployment - every major upgrade too | 03:27 |
| cloudnull | if the binds dont change there's no restart | 03:27 |
| cloudnull | so major upgrades shouldn't be effected very much, or at all. | 03:27 |
| cloudnull | and on upgrade even if a mount did change it'd only be changing on a specific containers. | 03:27 |
| odyssey4me | but the binds are changing, at least if we ever manage to get https://review.openstack.org/351188 in | 03:28 |
| cloudnull | so the restart caused by a changed mount would be the same regardless. | 03:28 |
| cloudnull | thats true | 03:28 |
| cloudnull | if that goes they'd change | 03:28 |
| odyssey4me | the point is to get those changes effected when executing lxc-containers-create... not some other play, if possible | 03:28 |
| *** thorst has joined #openstack-ansible | 03:29 | |
| odyssey4me | it makes the upgrade process much easier and more intuitive | 03:29 |
| cloudnull | how? | 03:29 |
| odyssey4me | but yes, there are still conditionals that need moving for that to become a full reality | 03:29 |
| cloudnull | i'd say not at all. it does however create a potential for veriable collision. | 03:29 |
| odyssey4me | any variable collission has to be entirely intentional | 03:30 |
| cloudnull | we still have to look in more than once place to see where the mounts come from | 03:30 |
| odyssey4me | yes, we now do - but I'm trying to move them al to one place | 03:30 |
| odyssey4me | *all | 03:30 |
| odyssey4me | I want to move the defaults out of the lxc-container-create role and into the group vars too | 03:31 |
| odyssey4me | then some of the conditionals look simple enough to get right | 03:31 |
| odyssey4me | we have a mess of logic in some tasks that need revisiting | 03:31 |
| *** weezS has joined #openstack-ansible | 03:31 | |
| odyssey4me | once those are cleaned up, I think it'll be easier | 03:32 |
| cloudnull | i think we need to realize that until something like https://github.com/ansible/ansible/pull/6666 gets into ansible proper group_vars are a bad idea. | 03:32 |
| cloudnull | unless the var is namespaced. | 03:32 |
| odyssey4me | sure, which is why we namespace - and why when we don't - they are treated like globals | 03:33 |
| cloudnull | if we dont then it should go in all.yml | 03:34 |
| odyssey4me | most of them are in all | 03:34 |
| odyssey4me | or maybe all_containers.yml for the container things | 03:34 |
| cloudnull | this is our current set of collisions: http://cdn.pasteraw.com/hrkst8mo2ymealahfyqousp8kcxwm0o | 03:35 |
| odyssey4me | I looked through that list - those are all completely intentional | 03:35 |
| odyssey4me | it's also misleading - some of those are *consuming* vars, not setting them | 03:37 |
| cloudnull | lxc_container_config_list is intentional | 03:37 |
| *** thorst has quit IRC | 03:37 | |
| cloudnull | the point is that so long as there are conflicts in group_vars unpredictable things can and will happen | 03:38 |
| odyssey4me | an example - service_region is set in all.yml, and consumed in ceilometer_all.yml: https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/ceilometer_all.yml#L18 | 03:38 |
| odyssey4me | so there's no collission there | 03:39 |
| *** finchd has quit IRC | 03:39 | |
| cloudnull | revised list http://cdn.pasteraw.com/91tn2a9bwqrk13utt4wcwv0dgwamkr | 03:43 |
| odyssey4me | it still has duplication too | 03:45 |
| odyssey4me | galera_client_drop_config_file is listed twice | 03:45 |
| odyssey4me | lxc_container_config_list is listed 3 times | 03:45 |
| cloudnull | thats because its in three files | 03:45 |
| openstackgerrit | Merged openstack/openstack-ansible: Limit LXC hosts playbook to container hosts only https://review.openstack.org/360081 | 03:46 |
| cloudnull | inventory/group_vars/cinder_volume.yml:16:lxc_container_config_list: | 03:46 |
| cloudnull | inventory/group_vars/all_containers.yml:18:lxc_container_config_list: | 03:46 |
| cloudnull | inventory/group_vars/neutron_agent.yml:16:lxc_container_config_list: | 03:46 |
| odyssey4me | yep - all intentional | 03:46 |
| cloudnull | all bad form | 03:46 |
| odyssey4me | all_containers holds the default, the others have the override | 03:46 |
| odyssey4me | what's the problem? | 03:46 |
| cloudnull | the collide | 03:47 |
| odyssey4me | it's an override | 03:47 |
| odyssey4me | it's supposed to have the same name, otherwise it would not be an override | 03:47 |
| cloudnull | its an override if its in the play or as a role based variable. | 03:48 |
| cloudnull | as a group var its a collision because the vars are flattened and the last load wins | 03:49 |
| odyssey4me | that's silly - why change the location when the group_var achieves the job perfectly | 03:49 |
| odyssey4me | yes exactly - the last one is the leaf group, we want it to win | 03:49 |
| cloudnull | we dont if the user changes the leaf which we do have | 03:49 |
| odyssey4me | it's only a problem if we end up setting something in, say neutron_all and neutron_agents. | 03:50 |
| cloudnull | or bind mounts. | 03:50 |
| odyssey4me | if a user is changing those settings, then they should be doing it in group_vars too | 03:50 |
| odyssey4me | if you have a moment, https://review.openstack.org/358852 needs another review | 03:53 |
| cloudnull | have a read here https://github.com/ansible/ansible/pull/6666 the ansible core devs talk about what we're doing specifically as being bad form and that the var should be set in the play or as a role var. | 03:54 |
| cloudnull | specifically https://github.com/ansible/ansible/pull/6666#issuecomment-136738132 | 03:54 |
| cloudnull | using role vars. | 03:54 |
| cloudnull | now, if we can get that change in, or implement it ourself in some other mechanism +1 | 03:54 |
| cloudnull | but I'm against building in known pain points | 03:54 |
| *** abelur has quit IRC | 03:57 | |
| cloudnull | also i know of 1 VERY large deployment (not the OSIC) where they are changing the env.d bits, and what we have is going to cause them a headache. | 03:57 |
| *** abelur has joined #openstack-ansible | 03:57 | |
| *** abelur is now known as 5EXAA7WD3 | 03:57 | |
| *** eil397 has quit IRC | 03:59 | |
| *** eil397_ has quit IRC | 04:00 | |
| *** hybridpollo has left #openstack-ansible | 04:02 | |
| cloudnull | in that PR all of the examples are what we want, but Ansible rejects that as bad form. | 04:05 |
| *** galstrom is now known as galstrom_zzz | 04:05 | |
| cloudnull | and until that's fixed in one way or another I do too (mainly because of the potential for pain). | 04:06 |
| odyssey4me | maybe I'm reading it wrong, but it seems that most of those examples are people trying to do host overrides through group vars | 04:06 |
| odyssey4me | but anyway, we have a pain now which that review resolves - it creates the *potential* for other pain... but it's only potential | 04:07 |
| odyssey4me | so it's going to come down to which is seen to be a bigger problem | 04:07 |
| cloudnull | idk that it resolves anything. | 04:07 |
| cloudnull | it just moves the binds into the contianer create process | 04:08 |
| odyssey4me | we can discuss it and decide in the meeting later | 04:08 |
| cloudnull | which is a first time deploy issue. | 04:08 |
| odyssey4me | and upgrade | 04:08 |
| cloudnull | i dont see how | 04:08 |
| odyssey4me | every single major upgrade has involved some sort of container config/bind mount change | 04:08 |
| cloudnull | true | 04:09 |
| cloudnull | which requires a restart | 04:09 |
| odyssey4me | exactly | 04:09 |
| cloudnull | which has to happen in the container create play, or in the service playbook | 04:09 |
| cloudnull | its the same time. | 04:09 |
| odyssey4me | right now it requires multiple restarts | 04:09 |
| odyssey4me | because some change made in lxc-c-c gets done, then restart... then another in the service play... and a restart again | 04:10 |
| *** 5EXAA7WD3 has quit IRC | 04:10 | |
| odyssey4me | every time the container restarts is an opportunity for failure | 04:10 |
| *** abelur has joined #openstack-ansible | 04:10 | |
| odyssey4me | either in service uptime, or in the container failing to restart properly for whatever reason | 04:10 |
| cloudnull | thats not true. | 04:11 |
| cloudnull | within the lxc-container-create play it restarts multiple times too | 04:11 |
| cloudnull | https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/container_create.yml#L160 | 04:11 |
| cloudnull | every iterable it restarts | 04:11 |
| cloudnull | its the same process | 04:11 |
| odyssey4me | no it doesn't | 04:11 |
| cloudnull | in that role or in the common-task | 04:11 |
| odyssey4me | that executes a container command - no restart | 04:11 |
| odyssey4me | the notifier handles the restart | 04:12 |
| cloudnull | ah. so change the common-task to do exactly that. | 04:12 |
| cloudnull | then no collision | 04:12 |
| cloudnull | and you achieve the same goal | 04:12 |
| cloudnull | and everything happens in once place | 04:13 |
| odyssey4me | you still then end up with multiple restarts, which is silly when that can be cut down to multiple restarts only for special cases | 04:13 |
| odyssey4me | or yes, the other option is that we move all the container config a bind mounting out into the service plays | 04:13 |
| cloudnull | or just restart the container twice | 04:14 |
| odyssey4me | *every* container | 04:14 |
| odyssey4me | this is the point - by implementing that change we then only have to restart a small few containers twice | 04:14 |
| odyssey4me | and only in special circumstances | 04:14 |
| cloudnull | but have the side effect of creating issues with variable collisions. | 04:16 |
| cloudnull | and in the case of a bind mount thats super bad | 04:17 |
| cloudnull | like data loss bad | 04:17 |
| cloudnull | where now we restart the containers potentially twice but nothing is lost | 04:18 |
| cloudnull | and the handling is explicite | 04:18 |
| odyssey4me | but we have no variable collissions - only overrides for special cases | 04:19 |
| odyssey4me | so there is no data loss | 04:19 |
| cloudnull | you do have collissions because the namespaced variable is defined in multiple groups | 04:20 |
| cloudnull | which is squashed | 04:20 |
| cloudnull | and we cant control how folks will deploy. but we can control how we lay thing out. | 04:20 |
| odyssey4me | the last one wins, which is exactly what we want | 04:20 |
| odyssey4me | anyway, time for me to try and get some sleep | 04:21 |
| cloudnull | take care. get some sleep | 04:22 |
| *** hblah has quit IRC | 04:22 | |
| *** albertcard has quit IRC | 04:24 | |
| *** eil397_ has joined #openstack-ansible | 04:28 | |
| *** thorst has joined #openstack-ansible | 04:35 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 04:37 |
| *** hblah has joined #openstack-ansible | 04:38 | |
| *** gouthamr has quit IRC | 04:39 | |
| *** thorst has quit IRC | 04:42 | |
| *** shausy has joined #openstack-ansible | 04:56 | |
| *** hblah has quit IRC | 05:01 | |
| *** admin0 has joined #openstack-ansible | 05:04 | |
| *** johnmilton has quit IRC | 05:07 | |
| *** admin0 has quit IRC | 05:23 | |
| *** janki has joined #openstack-ansible | 05:23 | |
| *** johnmilton has joined #openstack-ansible | 05:27 | |
| *** Prabhu has joined #openstack-ansible | 05:32 | |
| *** shausy has quit IRC | 05:33 | |
| *** shausy has joined #openstack-ansible | 05:33 | |
| Prabhu | Hi csmart | 05:34 |
| *** admin0 has joined #openstack-ansible | 05:36 | |
| csmart | Prabhu: hey | 05:36 |
| Prabhu | got your email. but i couldnt come online. sorry for that. | 05:37 |
| csmart | nps, I looked into it a bit more, firstly, I was in a rush yesterday and gave you the wrong instruction - httplib2 needed to be updated in the container, not the host | 05:37 |
| csmart | so no wonder it still failed | 05:37 |
| csmart | so I tried myself with the latest httplib2 in the container and it still failed for me | 05:38 |
| csmart | after some more testing, it looks like a bug in httplib2 when the comma separated list of IPs in no_proxy have spaces | 05:38 |
| *** admin0 has quit IRC | 05:38 | |
| csmart | in otherwords, for me, this fails: | 05:39 |
| csmart | no_proxy="localhost, 127.0.0.1, 172.29.236.100" | 05:39 |
| csmart | this works: | 05:39 |
| csmart | no_proxy="localhost,127.0.0.1,172.29.236.100" | 05:39 |
| *** rgogunskiy has joined #openstack-ansible | 05:40 | |
| Prabhu | okay. good to see that you found the issue and fix. | 05:40 |
| *** thorst has joined #openstack-ansible | 05:40 | |
| csmart | so the work around is to make sure that our no_proxy setting has no spaces | 05:41 |
| csmart | that's the bit where you write to user_variables.yml | 05:41 |
| *** abelur has quit IRC | 05:42 | |
| *** david-lyle has joined #openstack-ansible | 05:43 | |
| Prabhu | csmart: can I remove the space in user_variables.yml file and try re-run ? | 05:43 |
| csmart | Prabhu: yes I expect so | 05:43 |
| csmart | give that a shot at least | 05:43 |
| Prabhu | on it. | 05:43 |
| csmart | Prabhu: actually, you'll need to write the actual values there | 05:44 |
| csmart | Prabhu: as in, you'll need to set it to something like: | 05:45 |
| csmart | no_proxy_env: "\"localhost,127.0.0.1,172.29.236.100,192.168.122.120,172.29.237.80,172.29.239.37,172.29.238.230,172.29.236.182,172.29.239.106,172.29.237.213,172.29.237.151,172.29.238.235,172.29.236.134,172.29.238.124,172.29.236.121,172.29.237.85,172.29.238.31,172.29.238.64,172.29.238.87,172.29.239.36,172.29.236.240,172.29.236.87,172.29.236.215,172.29.237.131,172.29.236.226,172.29.238.34,172.29.236.78,172.29.238.13,172.29.236.107\"" | 05:45 |
| csmart | based on whatever is in /etc/environment | 05:45 |
| *** thorst has quit IRC | 05:48 | |
| csmart | here's my proof of the problem - I run this, sshed into the container http://paste.openstack.org/show/563253/ | 05:48 |
| Prabhu | great. I have updated my /etc/environment file. and restarted playbook. | 05:52 |
| csmart | Prabhu: I really hope this works ;-) | 05:53 |
| Prabhu | I believe too ;-) | 05:53 |
| csmart | the only other difference is that I also updated httplib2 | 05:55 |
| csmart | but we'll see what happens first | 05:55 |
| *** abelur has joined #openstack-ansible | 05:56 | |
| *** weezS has quit IRC | 05:58 | |
| *** abelur has quit IRC | 06:04 | |
| *** david-lyle has quit IRC | 06:09 | |
| *** McMurlock1 has joined #openstack-ansible | 06:14 | |
| *** abelur has joined #openstack-ansible | 06:16 | |
| *** abelur has quit IRC | 06:18 | |
| *** johnmilton has quit IRC | 06:19 | |
| csmart | Prabhu: I think I have a new version of the line to generate that list without spaces, just need to test | 06:22 |
| *** shausy has quit IRC | 06:24 | |
| *** shausy has joined #openstack-ansible | 06:24 | |
| *** johnmilton has joined #openstack-ansible | 06:26 | |
| *** eil397_ has quit IRC | 06:34 | |
| csmart | Prabhu: looks like it's my screw up actually, I broke out the variable onto multiple lines for my blog, that seems to be causing it | 06:39 |
| Prabhu | csmart: it failed again | 06:44 |
| csmart | Prabhu: ok, let me test this tonight and email you back | 06:44 |
| *** thorst has joined #openstack-ansible | 06:44 | |
| csmart | you can try to ssh into your container if you want and check it | 06:45 |
| csmart | I'll test tonight and get back to you in a few hours | 06:46 |
| Prabhu | sure csmart | 06:48 |
| csmart | sorry for the trouble, I'm not sure why this started happening or why I missed it before | 06:49 |
| Prabhu | no problem sir. I am wondering how it was successful for you earlier. | 06:49 |
| Prabhu | ;-) | 06:49 |
| csmart | hehe well me too.. | 06:50 |
| *** thorst has quit IRC | 06:52 | |
| Prabhu | :) | 06:54 |
| *** johnmilton has quit IRC | 06:56 | |
| *** admin0 has joined #openstack-ansible | 07:02 | |
| *** abelur_ has joined #openstack-ansible | 07:08 | |
| *** DrifterZA has joined #openstack-ansible | 07:11 | |
| *** johnmilton has joined #openstack-ansible | 07:20 | |
| *** abelur_ has quit IRC | 07:25 | |
| *** abelur_ has joined #openstack-ansible | 07:31 | |
| csmart | Prabhu: the other piece of the puzzle might be the needing to upgrade httplib2 because when I tested 0.8 it seemed to ignore NO_PROXY | 07:31 |
| *** bootsha has joined #openstack-ansible | 07:32 | |
| *** yarkot1 has quit IRC | 07:40 | |
| *** jduhamel has quit IRC | 07:42 | |
| *** jduhamel has joined #openstack-ansible | 07:43 | |
| *** thorst has joined #openstack-ansible | 07:50 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/openstack-ansible: DOC - note that stable/mitaka on Ubuntu works at most with 14.04 https://review.openstack.org/359714 | 07:55 |
| *** thorst has quit IRC | 07:58 | |
| *** chas has joined #openstack-ansible | 07:59 | |
| *** berendt has joined #openstack-ansible | 08:00 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add the BGP dynamic routing neutron plugin https://review.openstack.org/359859 | 08:01 |
| *** Oku_OS-away is now known as Oku_OS | 08:02 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: DOC - note that stable/mitaka on Ubuntu works at most with 14.04 https://review.openstack.org/359714 | 08:05 |
| *** eil397_ has joined #openstack-ansible | 08:10 | |
| Prabhu | okay csmart, let me try that also. ;-) | 08:17 |
| *** eil397_ has quit IRC | 08:18 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 08:18 |
| Prabhu | cmsart: it was already upgraded. do you want me to try it on NEW VM ? | 08:19 |
| *** asettle has joined #openstack-ansible | 08:20 | |
| *** abelur_ has quit IRC | 08:37 | |
| *** vnogin__ has joined #openstack-ansible | 08:43 | |
| *** vnogin_ has quit IRC | 08:46 | |
| *** openstackgerrit has quit IRC | 08:48 | |
| *** openstackgerrit has joined #openstack-ansible | 08:49 | |
| *** abelur_ has joined #openstack-ansible | 08:51 | |
| *** abelur_ has quit IRC | 08:54 | |
| *** thorst has joined #openstack-ansible | 08:56 | |
| *** abelur_ has joined #openstack-ansible | 08:57 | |
| *** admin0 has quit IRC | 08:58 | |
| openstackgerrit | Merged openstack/openstack-ansible: Project Calico integration https://review.openstack.org/340175 | 09:01 |
| openstackgerrit | Merged openstack/openstack-ansible: Remove the ansible.cfg file https://review.openstack.org/345572 | 09:01 |
| openstackgerrit | Merged openstack/openstack-ansible: Add play to deploy Rally to the utility containers https://review.openstack.org/357495 | 09:02 |
| *** Andrew_jedi has joined #openstack-ansible | 09:02 | |
| evrardjp | good morning | 09:03 |
| *** abelur_ has quit IRC | 09:03 | |
| *** thorst has quit IRC | 09:03 | |
| DrifterZA | Hi evrardjp | 09:03 |
| *** bootsha has quit IRC | 09:07 | |
| *** kylek3h has quit IRC | 09:10 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts: Revert "[GATE UNBLOCK] Purge/reinstall broken package from the xenial base image" https://review.openstack.org/360143 | 09:12 |
| *** bsv has joined #openstack-ansible | 09:13 | |
| DrifterZA | Nice odyssey4me | 09:13 |
| DrifterZA | -re monasca | 09:14 |
| *** abelur_ has joined #openstack-ansible | 09:15 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_keystone: Add a test Scenario for uwsgi & nginx https://review.openstack.org/359801 | 09:18 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Trust the roles to specify the correct inventory group https://review.openstack.org/359339 | 09:19 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi: Fix bare variables in with_items https://review.openstack.org/360056 | 09:21 |
| *** kylek3h has joined #openstack-ansible | 09:26 | |
| *** abelur_ has quit IRC | 09:29 | |
| *** thetrav has quit IRC | 09:36 | |
| *** admin0 has joined #openstack-ansible | 09:38 | |
| *** abelur_ has joined #openstack-ansible | 09:42 | |
| openstackgerrit | Merged openstack/openstack-ansible: DOC - note that stable/mitaka on Ubuntu works at most with 14.04 https://review.openstack.org/359714 | 09:46 |
| *** bootsha has joined #openstack-ansible | 09:47 | |
| openstackgerrit | git-harry proposed openstack/openstack-ansible: Print remaining tasks on failed upgrade correctly https://review.openstack.org/360385 | 09:49 |
| *** admin0 has left #openstack-ansible | 09:51 | |
| *** admin0 has joined #openstack-ansible | 09:51 | |
| admin0 | \o | 09:51 |
| *** admin0 has quit IRC | 09:52 | |
| *** abelur_ has quit IRC | 09:53 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Use a templated network conf https://review.openstack.org/360389 | 09:56 |
| *** thorst has joined #openstack-ansible | 10:00 | |
| *** hblah has joined #openstack-ansible | 10:03 | |
| *** abelur_ has joined #openstack-ansible | 10:07 | |
| *** thorst has quit IRC | 10:07 | |
| *** eil397_ has joined #openstack-ansible | 10:10 | |
| *** eil397_ has quit IRC | 10:14 | |
| *** electrofelix has joined #openstack-ansible | 10:17 | |
| *** abelur_ has quit IRC | 10:25 | |
| csmart | Prabhu: to do that you have to edit the ansible roles file on your host | 10:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Consume new data structure for venv builds https://review.openstack.org/360406 | 10:30 |
| *** abelur_ has joined #openstack-ansible | 10:39 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Support multiple rabbitmq clusters https://review.openstack.org/344884 | 10:40 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Consume new data structure for venv builds https://review.openstack.org/360406 | 10:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add Magnum deployment to setup-openstack playbook https://review.openstack.org/360428 | 10:53 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Use a templated network conf https://review.openstack.org/360389 | 10:55 |
| *** strobelight has quit IRC | 10:55 | |
| odyssey4me | DrifterZA you around? | 10:58 |
| odyssey4me | DrifterZA FYI https://github.com/openstack/openstack-ansible-os_monasca is now the official repository for monasca and you've been added as a core developer for that repo... please ensure that the openstack-ansible-core group remains, but the allocation of additional cores is up to you | 11:00 |
| odyssey4me | as a starting point you'll need to add scaffolding, much like https://github.com/openstack/openstack-ansible-os_watcher/commit/c62beb2372390291b66d535369612c51c807edd3 so that tests work | 11:01 |
| *** asettle has quit IRC | 11:01 | |
| *** asettle has joined #openstack-ansible | 11:02 | |
| *** thorst has joined #openstack-ansible | 11:05 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updated from global requirements https://review.openstack.org/360370 | 11:09 |
| *** hblah has quit IRC | 11:11 | |
| *** thorst has quit IRC | 11:12 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Consume new data structure for venv builds https://review.openstack.org/360406 | 11:17 |
| *** kencjohnston_ has quit IRC | 11:23 | |
| *** homerp_ has quit IRC | 11:23 | |
| *** sonus has quit IRC | 11:23 | |
| *** homerp has joined #openstack-ansible | 11:23 | |
| *** kencjohnston has joined #openstack-ansible | 11:24 | |
| *** sonus has joined #openstack-ansible | 11:24 | |
| DrifterZA | okay jesse | 11:25 |
| DrifterZA | I realise now why no one has written a trove playbook | 11:25 |
| DrifterZA | lol | 11:25 |
| DrifterZA | its not easy. | 11:25 |
| odyssey4me | heh, as far as I recall Trove and Octavia suffer from the same horrible disease of requiring VM's to be spun up for the service to run | 11:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add Magnum deployment to setup-openstack playbook https://review.openstack.org/360428 | 11:27 |
| DrifterZA | Well, I'm getting it to a point where the api,conductor and taskmanager will run in lxc | 11:28 |
| DrifterZA | But basing the play on the source code is hard | 11:28 |
| odyssey4me | andymccr can you please take a look through https://review.openstack.org/351188 | 11:29 |
| odyssey4me | nvm, I see that automagically got it | 11:31 |
| *** eil397_ has joined #openstack-ansible | 11:34 | |
| DrifterZA | +1 | 11:35 |
| DrifterZA | This is the one I did no ? | 11:35 |
| *** strobelight has joined #openstack-ansible | 11:36 | |
| automagically | Yep | 11:36 |
| automagically | Figured it made sense to get it pulled in so we can go through the review process with any changes | 11:37 |
| DrifterZA | good idea | 11:37 |
| *** abelur_ has quit IRC | 11:38 | |
| *** deadnull has joined #openstack-ansible | 11:39 | |
| *** eil397_ has quit IRC | 11:40 | |
| *** thorst has joined #openstack-ansible | 11:44 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 11:45 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 11:49 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 11:51 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible: Unbound DNS resolution containers https://review.openstack.org/336936 | 11:54 |
| mhayden | morning | 11:54 |
| *** asettle has quit IRC | 11:55 | |
| *** asettle has joined #openstack-ansible | 11:56 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_watcher: Replace references to lxc_hosts role in Makefile https://review.openstack.org/360473 | 11:59 |
| automagically | o/ mhayden | 11:59 |
| mhayden | ohai | 11:59 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 11:59 |
| mhayden | automagically: i saw something yesterday about comcast doing 2 gigabit residential service in some markets -- that's wild | 12:01 |
| automagically | Yep, cable labs has figured out how to push some serious bandwidth using DOCSIS | 12:01 |
| mhayden | indeedy | 12:01 |
| DrifterZA | automagically : LTGM dude | 12:04 |
| DrifterZA | LGTM I mean | 12:04 |
| odyssey4me | DrifterZA heh, you should leave +w for the second reviewer | 12:05 |
| mhayden | evrardjp / odyssey4me: i can't get the containers to restart with https://review.openstack.org/358852 | 12:05 |
| mhayden | there's no "notify:" on that task | 12:05 |
| odyssey4me | mhayden yep, there are missing parts to make it happen | 12:05 |
| mhayden | oh, did it restart containers for you? | 12:06 |
| odyssey4me | right now that one task that changes the config for the container will force a restart in the task itself - we have to break it out | 12:06 |
| *** spotz_zzz is now known as spotz | 12:09 | |
| mhayden | odyssey4me: but that one task has no notify, and i cannot get a container to restart if I remove the bind mount line and run that task | 12:10 |
| odyssey4me | mhayden it's not restarted by a notify - it's restarted by the module itself | 12:10 |
| mhayden | wouldn't containers be restarting in that playbook 10 times if that was true? :) | 12:11 |
| *** gouthamr has joined #openstack-ansible | 12:11 | |
| odyssey4me | mhayden yes - it looks like the back/forward ports were incomplete | 12:12 |
| mhayden | okay, i am going to run through this again to verify | 12:12 |
| odyssey4me | but that's exactly what we've been trying to resolve recently - restarts happening far too often and causing failures | 12:12 |
| matias | it's offtopic, but nobody answered when I asked in #openstack-monasca.. how does monasca play with ceilometer/gnocchi? it seems to have a lot of overlap in its purpose, is it a dropin replacement? | 12:14 |
| odyssey4me | ceilometer/aodh/gnocchi provide telemetry - not monitoring | 12:15 |
| odyssey4me | I *think* that Monasca consumes stuff from the Telemetry subsystems. | 12:15 |
| matias | hmm I mean, from what I understood, monasca has its own time-event database (influxdb), so this e.g. overlaps with gnocchi; also, it wasn't clear for me if it uses data from ceilometer collectors or if it has its own collectors | 12:17 |
| *** mpjetta has quit IRC | 12:18 | |
| odyssey4me | matias personally, I have no idea - DrifterZA ? | 12:18 |
| *** mpjetta has joined #openstack-ansible | 12:18 | |
| *** prometheanfire has quit IRC | 12:19 | |
| DrifterZA | Realtime event processing really, so monasca has its own agents which is mostly written in python | 12:19 |
| *** sc68cal has quit IRC | 12:20 | |
| *** arif-ali has quit IRC | 12:20 | |
| DrifterZA | So think of monasca as Nagios with real time event processing | 12:20 |
| DrifterZA | As I understand it. | 12:20 |
| matias | hm, so it like preprocesses what is inserted into influxdb? | 12:20 |
| matias | thanks DrifterZA | 12:20 |
| mhayden | odyssey4me: well, nvm, i got it to restart by removing the line | 12:20 |
| mhayden | that's extremely weird | 12:20 |
| *** mfisch has quit IRC | 12:21 | |
| odyssey4me | not weird at all - just unintuitive | 12:21 |
| *** sc68cal has joined #openstack-ansible | 12:21 | |
| *** prometheanfire has joined #openstack-ansible | 12:21 | |
| matias | odyssey4me: is there a queue of roles expected to be integrated into the main OSA playbook before newton is released? I'm curious if designate is going to come in | 12:21 |
| DrifterZA | I suppose a picture says a thousand words : https://wiki.openstack.org/wiki/Monasca#Architecture | 12:21 |
| *** kjw3 has joined #openstack-ansible | 12:23 | |
| *** hpe-hj has joined #openstack-ansible | 12:23 | |
| *** hj-hpe has quit IRC | 12:23 | |
| *** arif-ali has joined #openstack-ansible | 12:23 | |
| *** h1nch has quit IRC | 12:23 | |
| DrifterZA | but I would agree with odyssey4me, monasca is monitoring where gnocchi is metrics into timeseries db | 12:23 |
| *** bootsha has quit IRC | 12:24 | |
| *** sigmavirus has quit IRC | 12:24 | |
| DrifterZA | But sure there may be some overlap | 12:25 |
| *** mfisch has joined #openstack-ansible | 12:25 | |
| *** h1nch has joined #openstack-ansible | 12:25 | |
| matias | DrifterZA: what was confusing me is that you can also monitor the cloud health etc. from the metric stored in gnocchi (using e.g. grafana); but the conceptual difference from what you explained appears to be that monasca preprocesses what is stores or alarms on | 12:25 |
| *** mfisch has quit IRC | 12:25 | |
| *** mfisch has joined #openstack-ansible | 12:25 | |
| DrifterZA | Essentially the way I understand it is preprocessing vs postprocessing | 12:27 |
| DrifterZA | and a little more granularity with the agent checks | 12:27 |
| DrifterZA | than a global view | 12:27 |
| *** _sigmavirus24 has joined #openstack-ansible | 12:27 | |
| matias | nice, thanks for the explanation :D | 12:27 |
| DrifterZA | Sure anytime, I hope it made sense :D | 12:28 |
| evrardjp | logan-: interesting commit (the unbound one) I'll check how you deal with rabbit and all these fun stuff | 12:28 |
| evrardjp | added to my todo list | 12:28 |
| *** bootsha has joined #openstack-ansible | 12:28 | |
| *** _sigmavirus24 is now known as sigmavirus | 12:30 | |
| *** sigmavirus has joined #openstack-ansible | 12:30 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Update repo source download to use processed requirements https://review.openstack.org/360500 | 12:35 |
| odyssey4me | andymccr stevelle automagically is the plan to keep Apache/mod_wsgi and also have the option to move to Apache/uWSGI? OR are we shifting over wholesale? | 12:37 |
| *** sbezverk_ has joined #openstack-ansible | 12:37 | |
| odyssey4me | oh sorry - uWSGI/nginx | 12:38 |
| odyssey4me | ie are we keeping both options | 12:38 |
| automagically | I _hope_ its to keep Apache in place and add uwsgi/mginx | 12:38 |
| automagically | That’s the pattern I’m seeing | 12:38 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-lxc_container_create: Fix incorrect galera logs bind mount bug https://review.openstack.org/358852 | 12:38 |
| automagically | That gives us some risk insulation as we come up to the N3 deadline | 12:38 |
| matias | back into the designate etc. subject, how do you people are used to deploy roles which aren't integrated into the main playbook yet? do you fork the openstack-ansible repo and merge into it the stuff from the role's extra directory, or do you copy the repos to /etc/openstack_deploy/defaults/repo_packages/openstack_services.yml and so on? | 12:38 |
| andymccr | odyssey4me: that's what stevelle has done for the keystone role as a start | 12:38 |
| andymccr | worked ok in my testing | 12:39 |
| *** sbezverk has quit IRC | 12:39 | |
| mhayden | odyssey4me: this should be safer -> https://review.openstack.org/358852 | 12:39 |
| odyssey4me | matias the extras stuff is meant to be a hold over for testing purposes - ideally that gets used until everything is working, then that stuff moves over into the integrated repo | 12:39 |
| odyssey4me | matias unfortunately no-one has really spent time finalising the designate role yet | 12:39 |
| matias | perhaps it is too late now for me to catch it? we are almost in freeze | 12:40 |
| automagically | matias: I may go back to chipping away at it once RabbitMQ multi-cluster stuff is squared away | 12:40 |
| odyssey4me | mhayden hmm ok - lemme point you at something | 12:40 |
| matias | automagically: nice :D | 12:41 |
| odyssey4me | matias if you can validate that it works, even if it is just a basic working system, then please feel free to push the patches up to get it integrated | 12:41 |
| automagically | But, I don’t think its a Newton priority | 12:41 |
| odyssey4me | I would love to see that happen | 12:41 |
| automagically | odyssey4me matias ++ on testing and patches | 12:41 |
| odyssey4me | no, it's not a Newton priority - but if you have the time to do it, then go for it! | 12:42 |
| asettle | ugh odyssey4me there's inconsistencies between cinder-volume and cinder_volumes ... are they different things or the same? | 12:42 |
| asettle | Asking for a friend | 12:42 |
| asettle | :p | 12:42 |
| odyssey4me | asettle 'cinder-volume' is the service name... not sure what 'conder_volume' refers to... it might be a group name or property of some sort | 12:42 |
| asettle | Sigh. | 12:43 |
| asettle | Hahaha | 12:43 |
| asettle | Context help? | 12:43 |
| asettle | "The ``cinder_volumes`` service cannot run in a highly available configuration." | 12:43 |
| matias | is there any role which is currently considered a newton priority? I ask because we are also interested in barbican (mainly to use together with lbaasv2 in order to get tls support) | 12:43 |
| stevelle | odyssey4me: automagically andymccr you got it | 12:43 |
| odyssey4me | no, that should be 'cinder-volume' | 12:44 |
| asettle | odyssey4me: cheers :) | 12:44 |
| odyssey4me | matias all the newton priority roles are in - barbican would be nice, and yes probably nicer than designate | 12:44 |
| *** Andrew_jedi has quit IRC | 12:44 | |
| odyssey4me | barbican could be used for swift encryption at rest, magnum and LBaaS | 12:44 |
| *** Andrew_jedi has joined #openstack-ansible | 12:45 | |
| *** vnogin__ has quit IRC | 12:46 | |
| automagically | Agreed on barbican > designate | 12:46 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Use a templated network conf https://review.openstack.org/360389 | 12:47 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Add convergence test for test repo. https://review.openstack.org/360509 | 12:47 |
| *** fxpester has joined #openstack-ansible | 12:47 | |
| matias | the last question, I promise: is there any way to add new items to a list or dict (e.g. haproxy_service_configs) via /etc/openstack_deploy instead of overriding it completely? I mean, something like what is done to env.d | 12:47 |
| odyssey4me | mhayden see how https://github.com/openstack/openstack-ansible-lxc_container_create/commit/f7fb95fbe13db979a073b29d484997e591f54f07 split the container config and container commands into two tasks, and made the container config use the notifier... it looks like that didn't get backported properly (or at all maybe) | 12:48 |
| *** markvoelker has joined #openstack-ansible | 12:48 | |
| odyssey4me | matias no, not at this time | 12:48 |
| mhayden | odyssey4me: i see | 12:49 |
| matias | odyssey4me: thanks | 12:49 |
| *** bsv has quit IRC | 12:50 | |
| odyssey4me | mhayden so ideally we want to ensure that mechanism is properly backported - otherwise the rolling restart for galera won't work at all anyway | 12:51 |
| mhayden | hmm, looks like that one is a forward port from liberty -> newton | 12:52 |
| cloudnull | mornings | 12:52 |
| matias | morning | 12:52 |
| matias | nice to know we are in similar TZs :D | 12:53 |
| mhayden | odyssey4me: so that change is in mitaka for sure | 12:53 |
| mhayden | i'm looking at it right now | 12:53 |
| *** bootsha has quit IRC | 12:54 | |
| *** mgagne_ is now known as mgagne | 12:54 | |
| automagically | o/ cloudnull | 12:54 |
| *** vnogin has joined #openstack-ansible | 12:55 | |
| odyssey4me | mhayden ah, I see what's going on | 12:55 |
| *** bootsha has joined #openstack-ansible | 12:55 | |
| odyssey4me | liberty was only changed in terms of the config entries that have nothing to do with bind mounts | 12:55 |
| odyssey4me | it was a minimal change | 12:55 |
| odyssey4me | master has been changed to ensure that *any* config change uses the notifier | 12:56 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Ensure AIDE initializes on subsequent runs https://review.openstack.org/359554 | 12:56 |
| odyssey4me | mhayden considering the type of problem being handled here, it would be better to learn from master than from liberty | 12:56 |
| mhayden | cloudnull: added |lower to https://review.openstack.org/359554 | 12:57 |
| cloudnull | tyvm ! | 12:58 |
| cloudnull | :) | 12:58 |
| odyssey4me | mhayden or, of course, we could go with your frankenstein patch :p | 12:58 |
| mhayden | haha | 12:58 |
| mhayden | i tried to just make it as safe as possible | 12:59 |
| mhayden | :) | 12:59 |
| mhayden | i used manual stop/start on each container and they're working fine now with the right mounts | 12:59 |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone: Add a test Scenario for uwsgi & nginx https://review.openstack.org/359801 | 12:59 |
| mhayden | in 13.3.1 | 12:59 |
| *** eil397_ has joined #openstack-ansible | 13:00 | |
| *** bootsha has quit IRC | 13:00 | |
| * mhayden tips his hat to cloudnull | 13:00 | |
| DrifterZA | Hey guys I have an initial commit for openstack-ansible-os_trove , could someone help get it in an official repo so we can have more eyes to work on it ? | 13:00 |
| DrifterZA | https://github.com/drifterza/openstack-ansible-os_trove | 13:01 |
| *** psilvad has joined #openstack-ansible | 13:01 | |
| odyssey4me | DrifterZA yeah sure, I'll put the request in | 13:02 |
| DrifterZA | Thanks odyssey4me, its a start at least | 13:02 |
| *** janki has quit IRC | 13:03 | |
| *** Prabhu has quit IRC | 13:03 | |
| *** yarkot1 has joined #openstack-ansible | 13:03 | |
| *** eil397_ has quit IRC | 13:04 | |
| *** janki has joined #openstack-ansible | 13:05 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara: Role cleanup/changes for newton OSA https://review.openstack.org/359551 | 13:06 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_magnum: Add basic functional tests to role https://review.openstack.org/355155 | 13:07 |
| *** psilvad has quit IRC | 13:11 | |
| *** johnmilton has quit IRC | 13:14 | |
| *** johnmilton has joined #openstack-ansible | 13:15 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 13:15 |
| odyssey4me | DrifterZA the request is in: https://review.openstack.org/360530 | 13:16 |
| *** klamath has joined #openstack-ansible | 13:19 | |
| *** klamath_ has joined #openstack-ansible | 13:20 | |
| DrifterZA | Thanks odyssey4me | 13:22 |
| odyssey4me | cloudnull so after some thinking and discussion with evrardjp we came up with another way to achieve https://review.openstack.org/347400 using fully namespaced vars | 13:22 |
| odyssey4me | we would have to move the container create role execution into the service install plays | 13:23 |
| evrardjp | many other ways | 13:23 |
| evrardjp | :p | 13:23 |
| *** klamath has quit IRC | 13:24 | |
| odyssey4me | if they're in the service install plays then we can easily set the appropriate namespaced var to use for executing the role | 13:24 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [WIP] Add hostname adjustment to liberty>mitaka upgrades https://review.openstack.org/360539 | 13:24 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [WIP] Add hostname adjustment to liberty>mitaka upgrades https://review.openstack.org/360539 | 13:29 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_designate: DOCS - point homepage to role developer docs https://review.openstack.org/360542 | 13:30 |
| *** woodard has joined #openstack-ansible | 13:32 | |
| automagically | JIT container creation that way as well | 13:32 |
| automagically | Interesting idea | 13:32 |
| *** KLevenstein has joined #openstack-ansible | 13:32 | |
| odyssey4me | automagically yeah, I'm going to try and put a comparative patch together now | 13:33 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_designate: Remove apt update from Vagrantfile https://review.openstack.org/360546 | 13:33 |
| odyssey4me | what I like is that it will achieve the goal of having all container changes happen at the same time, and therefore only one restart... the vars will be in an obvious place... and things will change when you expect them to | 13:34 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_magnum: Add basic functional tests to role https://review.openstack.org/355155 | 13:34 |
| *** asettle has quit IRC | 13:35 | |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 13:35 |
| *** asettle has joined #openstack-ansible | 13:35 | |
| *** michaelgugino has joined #openstack-ansible | 13:36 | |
| *** GMAzrael has joined #openstack-ansible | 13:39 | |
| logan- | Evrardjp: cool it just places both rfc and non rfc names in for now | 13:40 |
| evrardjp | ? | 13:40 |
| evrardjp | yes atm you have both after an upgrade | 13:40 |
| matias | I keep shooting myself in the foot by doing ifdown br-mgmt, which destroys all veth pairs with the containers | 13:40 |
| evrardjp | It's still wip | 13:40 |
| logan- | Matias: check the ha proxy playbook I added an extra services var to extend without having to override the whole dict | 13:40 |
| matias | logan-: nice :D thank you | 13:41 |
| logan- | Evrardjp: re your earlier msg on how the unbound config handles rabbit. I figured you were talking about the rfc naming change | 13:41 |
| matias | /usr/local/bin/lxc-veth-check comes to the rescue :D | 13:43 |
| *** DrifterZA has quit IRC | 13:43 | |
| evrardjp | logan-: erlang vm doesn't use a proper resolver nor glibc resolution | 13:44 |
| evrardjp | I think it's recommended to have /etc/hosts | 13:44 |
| evrardjp | for that particular case | 13:44 |
| evrardjp | and I think it's only for the rabbitmq cluster nodes | 13:44 |
| evrardjp | (and no IPs -- they want their short hostnames in /etc/hosts) | 13:45 |
| evrardjp | at least if I remember correctly | 13:45 |
| *** hblah has joined #openstack-ansible | 13:45 | |
| *** janki has quit IRC | 13:45 | |
| evrardjp | so if unbound is additional that's fine | 13:46 |
| evrardjp | but if unbound is to be the only one way to do name resolution we have to care about that use case | 13:46 |
| odyssey4me | mhayden want me to take a crack at revising that galera patch, or are you on it? | 13:47 |
| *** jperry has joined #openstack-ansible | 13:47 | |
| jperry | le | 13:47 |
| mhayden | feel free to, odyssey4me -- i'm trapped in meetings for the forseaable future | 13:48 |
| logan- | that's interesting evrardjp. I am going to check my rabbit containers right now but I have not experienced any clustering issues and I doubt the other containers are in hosts | 13:48 |
| logan- | actually they are. interesting.. maybe the rabbit role manages hosts separately from openstack_hosts | 13:49 |
| evrardjp | logan-: good to know :D | 13:50 |
| evrardjp | I didn't check on these role for a while | 13:50 |
| evrardjp | it makes sense because that's a rabbit requirement | 13:50 |
| logan- | yep it sure does. rabbit_pre_install does it | 13:53 |
| *** eil397_ has joined #openstack-ansible | 13:54 | |
| cloudnull | odyssey4me: looking | 13:59 |
| odyssey4me | cloudnull I haven't done it yet - right now it's just a concept. | 13:59 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 13:59 |
| *** eil397_ has quit IRC | 14:00 | |
| cloudnull | automagically: ^ updated based on your feedback | 14:00 |
| cloudnull | odyssey4me: ^ also updated so that the container only restarts once | 14:00 |
| *** NachoDuck has quit IRC | 14:01 | |
| *** rgogunskiy has quit IRC | 14:02 | |
| odyssey4me | cloudnull small implementation change to make sure we log the stop/start | 14:02 |
| *** NachoDuck has joined #openstack-ansible | 14:04 | |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 14:05 |
| odyssey4me | looking for reviews on https://review.openstack.org/360406 and https://review.openstack.org/360500 please | 14:06 |
| * automagically looking | 14:07 | |
| qwang | good morning | 14:07 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 14:07 |
| cloudnull | odyssey4me: good looking out. updated | 14:08 |
| cloudnull | o/ qwang | 14:08 |
| *** janki has joined #openstack-ansible | 14:08 | |
| qwang | i have a patch that was failed on jenkins due to broken gate in aodh. now aodh has been fixed and i want to submit the patch again, how should i do it without changing anything? | 14:09 |
| openstackgerrit | Merged openstack/openstack-ansible: Support multiple rabbitmq clusters https://review.openstack.org/344884 | 14:09 |
| *** galstrom_zzz is now known as galstrom | 14:09 | |
| qwang | https://review.openstack.org/#/c/356772/ | 14:09 |
| odyssey4me | qwang you can just rebase from gerrit | 14:09 |
| odyssey4me | or if there's a merge conflict you'll have to resolve it | 14:10 |
| odyssey4me | (manually) | 14:10 |
| odyssey4me | automagically w00t! the rabbit patch is in! | 14:11 |
| qwang | odyssey4me: thank you very much. will do it | 14:11 |
| automagically | odyssey4me: Saw that. I’m pretty excited. Not to revive the os_ceilometer patch that had been reverted | 14:11 |
| openstackgerrit | Merged openstack/openstack-ansible: Add the BGP dynamic routing neutron plugin https://review.openstack.org/359859 | 14:12 |
| openstackgerrit | Merged openstack/openstack-ansible: Print remaining tasks on failed upgrade correctly https://review.openstack.org/360385 | 14:12 |
| openstackgerrit | Merged openstack/openstack-ansible: Add Magnum deployment to setup-openstack playbook https://review.openstack.org/360428 | 14:12 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 14:13 |
| michaelgugino | is there a blueprint for the testing repo? Or etherpad? I'm trying to figure out what the consensus direction is. | 14:13 |
| *** deadnull is now known as _deadnull | 14:15 | |
| *** hblah has quit IRC | 14:15 | |
| odyssey4me | michaelgugino I think right now it's kinda being felt out. I didn't expect any real work to start until N3. | 14:15 |
| odyssey4me | oh, a bp was registered: central-test-repository | 14:16 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_ceilometer: Revert "Revert "RabbitMQ config rework"" https://review.openstack.org/360578 | 14:16 |
| *** Mudpuppy has joined #openstack-ansible | 14:17 | |
| *** michauds has joined #openstack-ansible | 14:21 | |
| openstackgerrit | Qin Wang (qwang) proposed openstack/openstack-ansible-os_aodh: Add SNI support via OS packages for os_aodh role https://review.openstack.org/356772 | 14:21 |
| *** shausy has quit IRC | 14:21 | |
| openstackgerrit | Qin Wang (qwang) proposed openstack/openstack-ansible-os_aodh: Add SNI support via OS packages for os_aodh role https://review.openstack.org/356772 | 14:23 |
| *** berendt has quit IRC | 14:28 | |
| *** janki has quit IRC | 14:29 | |
| *** chas has quit IRC | 14:31 | |
| *** karimb has joined #openstack-ansible | 14:33 | |
| *** spotz is now known as spotz_zzz | 14:35 | |
| *** spedione|AWAY is now known as spedione | 14:35 | |
| kencjohnston | hi team, I'm at the Ops Summit, do OSA roles install upper constraints into service venvs? | 14:38 |
| odyssey4me | kencjohnston yes, we've been using upper constraints since liberty (when they were created) | 14:38 |
| kencjohnston | odyssey4me: thanks | 14:38 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_designate: DOCS - Cleanup/comment defaults/main.yml https://review.openstack.org/360601 | 14:39 |
| kencjohnston | For those interested in following along, Ops Mid-Cycle in NYC is talking about containerising the control plane right now - https://etherpad.openstack.org/p/NYC-ops-Containerising-the-control-plane | 14:40 |
| *** sdake has joined #openstack-ansible | 14:42 | |
| *** spedione is now known as spedione|AWAY | 14:43 | |
| *** DrifterZA has joined #openstack-ansible | 14:43 | |
| *** galstrom is now known as galstrom_zzz | 14:46 | |
| DrifterZA | odyssey4me | 14:46 |
| DrifterZA | may I +1 workflow on https://review.openstack.org/#/c/360459/ ? | 14:47 |
| odyssey4me | DrifterZA yep, if you've reviewed and it looks good you can +2+w | 14:47 |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 14:48 |
| odyssey4me | generally the principle is that two people other than the author review and vote - the second can then +w | 14:48 |
| DrifterZA | I think it looks good now | 14:48 |
| *** openstackgerrit has quit IRC | 14:48 | |
| *** jperry has quit IRC | 14:49 | |
| *** openstackgerrit has joined #openstack-ansible | 14:49 | |
| michaelgugino | that containers etherpad is interesting. OSA solves a lot of the problems mentioned already. Of course, OSA doesn't use docker, so I guess we don't get the fanboys | 14:50 |
| kencjohnston | michaelgugino: in the room there is generally a sentiment of, don't use docker, not mature enough, a new security attack service. | 14:51 |
| cloudnull | kencjohnston: ++ | 14:51 |
| cloudnull | :) | 14:51 |
| * cloudnull loves docker | 14:51 | |
| michaelgugino | good to know | 14:52 |
| kencjohnston | cloudnull: fanboy | 14:52 |
| cloudnull | you know it | 14:52 |
| kencjohnston | something about docker driving you to use their ecosystem of rooted containers on dockerhub... | 14:53 |
| automagically | Damn, now I’m regretting not attending the ops midcycle | 14:53 |
| *** david-lyle has joined #openstack-ansible | 14:53 | |
| *** finchd has joined #openstack-ansible | 14:54 | |
| DrifterZA | can anyone +1w on https://review.openstack.org/#/c/360530/2 ? | 14:54 |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca: Role skeleton for os_monasca https://review.openstack.org/360459 | 14:54 |
| odyssey4me | DrifterZA wait patiently grasshopper - that's not curated by this team - the infra team has to vote there | 14:55 |
| odyssey4me | DrifterZA if it's not done by tomorrow, I'll ask nicely | 14:55 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [WIP] Add hostname adjustment to liberty>mitaka upgrades https://review.openstack.org/360539 | 14:55 |
| DrifterZA | haha | 14:55 |
| DrifterZA | OKay :D | 14:55 |
| *** weezS has joined #openstack-ansible | 14:57 | |
| *** jperry has joined #openstack-ansible | 14:57 | |
| *** finchd has left #openstack-ansible | 14:58 | |
| *** david-lyle has quit IRC | 14:59 | |
| *** matias has quit IRC | 15:00 | |
| automagically | cloudnull: Can config_template overrides reference vars? | 15:04 |
| cloudnull | what do you mean ? | 15:05 |
| *** matias has joined #openstack-ansible | 15:05 | |
| *** matias is now known as Guest78429 | 15:06 | |
| automagically | Something like https://gist.github.com/trumant/04363f77e2685d3a6627b6679edb1f75 where nova_compute_pin_set is a host_var | 15:06 |
| cloudnull | yes. that should work | 15:06 |
| automagically | Figured, just wanted to double-check | 15:06 |
| cloudnull | the jinja template is rendered, then the overrides applied and jinja rendered again | 15:07 |
| *** spedione|AWAY is now known as spedione | 15:07 | |
| cloudnull | so the items in config_template will have the most precedence | 15:07 |
| cloudnull | and the values can be set as static items or vars. | 15:07 |
| *** jamesden_ has joined #openstack-ansible | 15:10 | |
| *** jamesdenton is now known as Guest1506 | 15:11 | |
| *** janki has joined #openstack-ansible | 15:14 | |
| *** furlongm_ has joined #openstack-ansible | 15:19 | |
| DrifterZA | So anyone thought about the fact that mariadb has gone commercial and what the plan is ? | 15:19 |
| odyssey4me | heh, cue cloudnull rant :p | 15:19 |
| *** furlongm has quit IRC | 15:20 | |
| cloudnull | DrifterZA: https://twitter.com/cloudnull/status/767396775854432256 | 15:20 |
| cloudnull | ^ rant | 15:20 |
| * cloudnull thinking that Percona is a good option, fully open source and is not in the buisness of selling SQL server they sell proserv. | 15:22 | |
| cloudnull | postgres is another option. but may be too big a leap for most. | 15:22 |
| DrifterZA | Percona + Galera | 15:24 |
| *** eil397_ has joined #openstack-ansible | 15:24 | |
| cloudnull | XtraDB | 15:27 |
| cloudnull | which is their galera thingame | 15:27 |
| cloudnull | https://www.percona.com/software/mysql-database/percona-xtradb-cluster | 15:28 |
| *** _deadnull is now known as deadnull | 15:29 | |
| *** deadnull has quit IRC | 15:30 | |
| michaelgugino | seeing as we don't use a database proxy (which typically intercepts and relays commands / replies), I don't see a problem with them having a commercial product. | 15:30 |
| *** Andrew_jedi has quit IRC | 15:31 | |
| DrifterZA | I suppose its just upstream patches / bug fixes, but I doubt they would leave the core project hanging like that. | 15:32 |
| DrifterZA | Unless they are utter douche bags | 15:32 |
| DrifterZA | But I would hate to migrate on a 1000+ node cluster | 15:33 |
| odyssey4me | I would not be averse to having another option on the table next cycle to give deployers choice. | 15:33 |
| DrifterZA | makes me think | 15:33 |
| odyssey4me | That said, it costs. | 15:33 |
| DrifterZA | Ignoring the elephant also costs | 15:34 |
| DrifterZA | its a major concern actually | 15:34 |
| odyssey4me | we'd have to do a new role with equivalent functionality, then figure out what it takes to migrate from the one DB to the other and build that into the upgrade path... and it's likely we'll be forced to have hard down time | 15:36 |
| odyssey4me | so it's not trivial | 15:36 |
| DrifterZA | Precisley | 15:36 |
| DrifterZA | postgres has been around for like ever | 15:37 |
| DrifterZA | but no guarantee's that it could suffer the same fate later on. | 15:38 |
| *** gouthamr is now known as gouthamr|afk | 15:38 | |
| odyssey4me | it's also quite different to administer | 15:38 |
| *** v1k0d3n has joined #openstack-ansible | 15:39 | |
| DrifterZA | and clustering.. jeeze im not so sure | 15:40 |
| *** Andrew_jedi has joined #openstack-ansible | 15:42 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Implement container bind mount for all logs https://review.openstack.org/351188 | 15:42 |
| cloudnull | michaelgugino DrifterZA: I see the move toward BSL as a stepping stone to something like Oracle has w/ mysql. | 15:44 |
| cloudnull | IE they have multiple versions most of which are "free" while operating at a low threshold. | 15:45 |
| *** BjoernT has joined #openstack-ansible | 15:46 | |
| cloudnull | w/ mysql CE its all free all the time, but it lacks features as well as bug fixes. | 15:46 |
| cloudnull | and now that MariaDB has a framework for doing the same I fear that they actually will. | 15:47 |
| cloudnull | i'm also a tin foil hat person so feel free to ignore me | 15:47 |
| cloudnull | :) | 15:47 |
| michaelgugino | cloudnull: you might be right. Cash talks though. If people want it to continue to be free software, someone's got to pay the bills. | 15:48 |
| cloudnull | ++ | 15:48 |
| thotypous | postgresql doesn't have a multi-master mode, you need to do failover to implement HA instead of load balancing | 15:49 |
| *** Guest78429 has quit IRC | 15:49 | |
| *** Guest78429 has joined #openstack-ansible | 15:49 | |
| cloudnull | michaelgugino: I get why they need/want to do it. I just dont like it | 15:49 |
| cloudnull | thotypous: postgresql has XC, pg_pool1/2, and BDR but to odyssey4me point its a quite different to admin | 15:50 |
| DrifterZA | If I may interject, Percona has been around a long while. I'm thinking it will have the least resistance in implentation and admin. | 15:52 |
| cloudnull | DrifterZA: +1 | 15:53 |
| cloudnull | we could do a drop-in replacement of the galera_server package w/ percona. | 15:53 |
| cloudnull | and it **should** upgrade cleanly. | 15:53 |
| DrifterZA | yeah I mean xtradb is just a performance mod ontop of innodb no ? | 15:53 |
| cloudnull | ^ that would require a lot of testing | 15:53 |
| cloudnull | yup | 15:53 |
| *** admin0 has joined #openstack-ansible | 15:54 | |
| cloudnull | its version compatible too | 15:54 |
| cloudnull | and we already use xtrabackup | 15:54 |
| DrifterZA | True. | 15:54 |
| DrifterZA | we should raise it in the mid-week | 15:54 |
| DrifterZA | if its not on the agenda | 15:54 |
| cloudnull | today :) | 15:54 |
| palendae | 9 minutes :) | 15:55 |
| admin0 | hey all | 15:55 |
| michaelgugino | I think we're being awfully hasty here. There's no reason to switch yet. If they go down the path of broken CE and great EE, then I think the community at large will switch. | 15:55 |
| admin0 | for some reasons, my default install fails On cinder .. tried debugging for a few days already .. anyone with some free time can guide | 15:55 |
| DrifterZA | michaelgugino : Hasty perhaps, but I think it would be a prudent exercise to get it on the drawing board | 15:56 |
| DrifterZA | and like odyssey4me said, give devs the option at least | 15:56 |
| DrifterZA | I mean if needs be I can start the work after I've done Trove. | 15:57 |
| *** jchhatbar has joined #openstack-ansible | 15:58 | |
| *** janki has quit IRC | 15:59 | |
| *** eil397_ has quit IRC | 15:59 | |
| spedione | Does anyone have any idea why networking wouldn't be working correctly after running the gate_check_commit script to set up an AIO? | 15:59 |
| palendae | Is the keystone_auth_admin_token allowed to do almost anything in Keystone? | 16:00 |
| *** eil397_ has joined #openstack-ansible | 16:00 | |
| *** jchhatbar is now known as janki | 16:00 | |
| DrifterZA | admin0 : what seems to be the issue ? | 16:00 |
| mhayden | Meeting in #openstack-meeting-4! cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden, scarlisle, luckyinva, ntt, javeriak, automagically, spotz, vdo, jmccrory, alextricity25, jasondotstar, admin0, | 16:00 |
| mhayden | michaelgugino, ametts, v1k0d3n, severion, bgmccollum, darrenc, JRobinson__, asettle, colinmcnamara, thorst, adreznec, eil397 | 16:00 |
| *** eil397 has joined #openstack-ansible | 16:01 | |
| *** eil397 has left #openstack-ansible | 16:01 | |
| *** Guest78429 is now known as matias | 16:01 | |
| *** matias has quit IRC | 16:02 | |
| *** matias has joined #openstack-ansible | 16:02 | |
| cloudnull | michaelgugino: to be clear, I'm VERY much against switching anything this cycle, and maybe even the next but i do think its something we should be aware of. | 16:03 |
| automagically | ^ that ^ | 16:04 |
| michaelgugino | yeah, I think we're just a long way off. I think we're all on the same page. If they start pulling shenanigans, we jump ship. But, I'm pretty sure we won't be the only ones doing so. | 16:04 |
| odyssey4me | yeah, no sweat if some work gets done this cycle - but we won't integrate any of it or focus on it until next | 16:04 |
| automagically | FYI - there has been talk in other projects of shutting down postgres gates | 16:04 |
| automagically | If you care about being able to go that route, speak up when those topics hit the ML | 16:05 |
| cloudnull | automagically: it's likely those gates will go away. just like pypy did. they're not being maintained. | 16:08 |
| cloudnull | and at this point i dont think it's important to most. | 16:09 |
| cloudnull | but it may be in the near-ish future :) | 16:09 |
| palendae | O.o | 16:10 |
| * michaelgugino has to google every command for postgres | 16:10 | |
| cloudnull | pg user :) | 16:10 |
| palendae | That makes me a little sad that postgres isn't being maintained, but also understand it | 16:11 |
| palendae | Not even because I prefer postgres, but because having options is good | 16:11 |
| michaelgugino | everything is driven through sqlalchemy afaik, so nbd. Most queries should just /work/ | 16:12 |
| palendae | 'should' :) | 16:12 |
| cloudnull | -tm | 16:12 |
| michaelgugino | upgrade scripts and bootstrapping is what will probably get broken. | 16:12 |
| palendae | Not tested, doesn't work | 16:12 |
| michaelgugino | ++ | 16:12 |
| *** spotz_zzz is now known as spotz | 16:16 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_ceilometer: Revert "Revert "RabbitMQ config rework"" https://review.openstack.org/360578 | 16:19 |
| *** michauds has quit IRC | 16:21 | |
| *** michauds has joined #openstack-ansible | 16:21 | |
| admin0 | DrifterZA: i am facing the trouble to get cinder installed ( this worked like 50+ times ) because i destroy and test a lot .. sudeenly it stopped working .. tried all debug mode and verbose mode, but cannot understand where it actually fails | 16:22 |
| admin0 | @ some SNI name thing | 16:22 |
| cloudnull | admin0: mitaka? | 16:23 |
| DrifterZA | could you post the debug output to pastebin or something | 16:23 |
| admin0 | mitaka | 16:23 |
| cloudnull | recently there have been a rash of SNI problems upstream. | 16:24 |
| *** weezS has quit IRC | 16:25 | |
| admin0 | and how are people coping with it or fixing it ? | 16:25 |
| cloudnull | we may need to do backporting of some, or all, of https://review.openstack.org/#/q/status:merged+branch:master+topic:add-sni-pkgs to mitaka | 16:25 |
| cloudnull | admin0: the big one to fix for deployment is https://review.openstack.org/#/c/348887/ | 16:27 |
| nishpatwa | Hi I am getting jenkins failure everytime https://review.openstack.org/#/c/356713/ | 16:27 |
| automagically | nishpatwa: Yep, we believe their is an LXC bug in CentOS that is biting us | 16:27 |
| nishpatwa | Oho | 16:28 |
| cloudnull | automagically: i think i've narrowed that down finally. sadly i think the fix is to add a retry | 16:28 |
| automagically | Bah….(ノಠ益ಠ)ノ彡┻━┻ | 16:29 |
| cloudnull | it looks like the cgroup isnt ready and a restart then fixes it | 16:30 |
| admin0 | cloudnull: would adding that patch to the current one fix it ? | 16:31 |
| automagically | Hmm, can we do a http://docs.ansible.com/ansible/wait_for_module.html on the cgroup path? | 16:31 |
| *** chas has joined #openstack-ansible | 16:31 | |
| cloudnull | admin0: it should | 16:32 |
| admin0 | thanks | 16:32 |
| admin0 | will give it a try | 16:32 |
| *** chas has quit IRC | 16:36 | |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 16:37 |
| admin0 | its moving forward | 16:40 |
| DrifterZA | nice | 16:44 |
| admin0 | hmm.. didn’t worked :( | 16:45 |
| admin0 | fails in : TASK: [os_cinder | Add in cinder devices types] ******************************* | 16:46 |
| *** gouthamr|afk is now known as gouthamr | 16:46 | |
| *** jamesdenton has joined #openstack-ansible | 16:48 | |
| *** jamesden_ has quit IRC | 16:48 | |
| *** david-lyle has joined #openstack-ansible | 16:50 | |
| cloudnull | same SNI issues there? | 16:52 |
| admin0 | yeah | 16:52 |
| cloudnull | SNI issues connecting to the LB ? | 16:53 |
| *** jamesdenton has quit IRC | 16:53 | |
| *** jamesdenton has joined #openstack-ansible | 16:54 | |
| *** DrifterZA has quit IRC | 16:55 | |
| *** weezS has joined #openstack-ansible | 16:56 | |
| palendae | automagically, Test cases on that bug would be helpful, though I can reliably recreate the bug by simply deploying master or mitaka with Keystone's policy.json replaced | 16:56 |
| automagically | K, will give the bug a re-read palendae and see if I can be of any assistance | 16:58 |
| palendae | automagically, Thanks. Keystone's the one failing now, but I assume that's only cause it's the first service to try to register a tenant/project | 16:58 |
| odyssey4me | palendae the keystone stuff registers directly into the DB, so it can't be | 16:59 |
| odyssey4me | it uses the bootstrap | 16:59 |
| palendae | odyssey4me, Newp | 16:59 |
| palendae | odyssey4me, post bootstrap we register keystone into the service catalog. That's what gets the 403. http://git.openstack.org/cgit/openstack/openstack-ansible-os_keystone/tree/tasks/keystone_service_setup.yml#n50 | 17:00 |
| odyssey4me | cloudnull IIRC we already backported the main SNI patches back | 17:00 |
| odyssey4me | ie the pip_install and bootstrap-ansible patches | 17:00 |
| odyssey4me | it is possible that admin0 is running something from before then though | 17:01 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Add retry to container start/stop handlers https://review.openstack.org/360681 | 17:05 |
| cloudnull | ^ that "should" fix the cent container start issues. | 17:05 |
| odyssey4me | haha, beat it with a hammer | 17:06 |
| cloudnull | if you check the runs where it failed, the container was running before being restarted | 17:06 |
| cloudnull | and it seems like the system is just not ready. | 17:07 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create: Fix tag typo https://review.openstack.org/359288 | 17:07 |
| *** asettle has quit IRC | 17:08 | |
| cloudnull | or its a race caused by two containers starting at the exact same time. | 17:08 |
| *** asettle has joined #openstack-ansible | 17:09 | |
| *** janki has quit IRC | 17:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron: Update paste, policy and rootwrap configurations 2016-08-25 https://review.openstack.org/360685 | 17:10 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Update all SHAs for Newton 2016-08-25 https://review.openstack.org/360686 | 17:12 |
| *** asettle has quit IRC | 17:13 | |
| *** weezS has quit IRC | 17:18 | |
| odyssey4me | palendae odd though - we always create a heat domain - so every deployment has more than one domain by default... so I'm struggling to understand how this hasn't been a problem before | 17:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Update all SHAs for 13.3.3 https://review.openstack.org/360691 | 17:20 |
| palendae | odyssey4me, yeah, same | 17:20 |
| palendae | Though the heat domain's created later than keystone | 17:21 |
| odyssey4me | sure, but so is this other domain - right? | 17:21 |
| odyssey4me | unless this is an upgrade issue? | 17:21 |
| palendae | It'll happen on a greenfield with https://github.com/openstack/keystone/blob/stable/mitaka/etc/policy.v3cloudsample.json in place | 17:21 |
| odyssey4me | ah, so the key is the policy change | 17:22 |
| palendae | I'll poke the reporters some more to see if this is an upgraded env or not, but I do run in to the same problem | 17:22 |
| odyssey4me | why not just deploy that policy change later? | 17:22 |
| palendae | Good question - no idea | 17:22 |
| palendae | I wonder if that breaks anything | 17:22 |
| palendae | If it's an upgrade it may already be in place | 17:22 |
| odyssey4me | if it's a new env then it's an acceptable workaround. If it's an upgrade then that's an issue | 17:23 |
| Andrew_jedi | In mitaka, the admin_token is missing from keystone.conf. Is it intentional ? | 17:23 |
| odyssey4me | Andrew_jedi yes - the admin_token middleware has been deprecated | 17:23 |
| palendae | odyssey4me, I'll ask and in the meantime spin another AIO where I try adding it later | 17:23 |
| odyssey4me | it's a security hole of note | 17:23 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Update all SHAs for 12.2.3 https://review.openstack.org/360694 | 17:24 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-plugins: Implement an opportunistic strategy and connection plugin https://review.openstack.org/358904 | 17:24 |
| openstackgerrit | Merged openstack/openstack-ansible: Updated from global requirements https://review.openstack.org/360370 | 17:25 |
| Andrew_jedi | odyssey4me: AFAIK, admin_token middleware was essential for Keystone and Ceph integration. I wonder how to integrate Keystone and Ceph now ? | 17:25 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron: Add the BGP dynamic routing neutron plugin https://review.openstack.org/359860 | 17:25 |
| odyssey4me | Andrew_jedi Uh, you can add it in - it's not not enabled by default. | 17:25 |
| odyssey4me | Andrew_jedi just use the config_override for keystone.conf to add it in | 17:26 |
| odyssey4me | I'm guessing that you mean for radosgw? | 17:26 |
| *** AnarchyAo has joined #openstack-ansible | 17:26 | |
| Andrew_jedi | odyssey4me: Yes radosgw, but if i execute the keystone playbook again. Is it possible that i will loose any data? | 17:27 |
| odyssey4me | Andrew_jedi not unless you made manual changes to config in the containers | 17:27 |
| odyssey4me | Andrew_jedi if you're paranoid you can execute the playbook limited to one container at a time | 17:27 |
| odyssey4me | using --limit | 17:28 |
| Andrew_jedi | odyssey4me: Gotcha! let me try that :) | 17:28 |
| *** McMurlock1 has quit IRC | 17:28 | |
| *** chas has joined #openstack-ansible | 17:33 | |
| odyssey4me | ok, I'm out - it looks like the gate queues are hosed for a while anyway | 17:36 |
| odyssey4me | catch y'all tomorrow | 17:36 |
| openstackgerrit | Flávio Ramalho proposed openstack/openstack-ansible: Adding a playbook for deploying Sahara https://review.openstack.org/360700 | 17:36 |
| cloudnull | odyssey4me: cheers | 17:37 |
| cloudnull | have a good one | 17:37 |
| matias | cya | 17:37 |
| *** chas has quit IRC | 17:37 | |
| cloudnull | anyone around that might want to give this a shove https://review.openstack.org/#/c/360129/ | 17:37 |
| matias | I just realized that cinderbackup fails if you backup from a ceph volume into a swift store (it tries to import rbd and rados python libraries, but python-cephlibs is missing from the venv), can I add it to cinder_pip_packages? | 17:38 |
| matias | probably all goes fine if you are backuping from ceph INTO ceph | 17:39 |
| cloudnull | yea, if thats missing we should add it | 17:39 |
| odyssey4me | matias we already have a link from the venv to the host libraries | 17:39 |
| odyssey4me | we can't add the python libs from pypi because there aren't any... and each version of ceph has different libs | 17:39 |
| odyssey4me | so we detect the libs on the host and symlink them into the venv | 17:40 |
| matias | odyssey4me: hm, so we should install the python libs into the host manually? | 17:41 |
| odyssey4me | matias I don't think so - I think the ceph_client role does that.. although I don't know | 17:41 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Enable the opportunistic strategy https://review.openstack.org/349485 | 17:41 |
| *** karimb has quit IRC | 17:41 | |
| *** Jeffrey4l__ has quit IRC | 17:41 | |
| matias | well, something is missing, the backup only ran when I rebuilt the venv with python-cephlibs | 17:41 |
| odyssey4me | matias I wonder if the cinder venv is missing the links | 17:42 |
| odyssey4me | check in the venv to see if there are symlinks after deploying | 17:42 |
| matias | let me investigate it, thanks for the pointers | 17:42 |
| cloudnull | mhayden odyssey4me: updated https://review.openstack.org/358904 w/ a release node. if you have a moment to review it and the other patch that makes that go -- https://review.openstack.org/349485 -- it'd be appreciated. | 17:42 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible: Unbound DNS resolution containers https://review.openstack.org/336936 | 17:42 |
| odyssey4me | if gerrit wouldn't be crashing I would | 17:43 |
| odyssey4me | cloudnull missing a reno in https://review.openstack.org/#/c/349485/ ? | 17:44 |
| odyssey4me | trailing whitespace in https://review.openstack.org/#/c/358904/3/releasenotes/notes/opportunistic-strategy-and-connection-plugin-bc476fa3607dcc4a.yaml | 17:44 |
| cloudnull | do we need two ? | 17:44 |
| palendae | odyssey4me, Question - you removed the mitaka target from https://bugs.launchpad.net/openstack-ansible/+bug/1614211; I get not targetting backports til master merges, but why remove it from bugs when it is affected? | 17:44 |
| openstack | Launchpad bug 1614211 in openstack-ansible trunk "Playbook Runs Fail in Multi-Domain Environments" [Medium,Confirmed] - Assigned to Nolan Brubaker (nolan-brubaker) | 17:44 |
| odyssey4me | oh fair enough - the one would get pulled in | 17:45 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-plugins: Implement an opportunistic strategy and connection plugin https://review.openstack.org/358904 | 17:45 |
| odyssey4me | palendae two reasons - one is that series and milestones aren't really generally targeted in launchpad any more in openstack in general | 17:48 |
| odyssey4me | you can target all you like, but it doesn't really get maintained properly | 17:48 |
| odyssey4me | secondly, it sets a false expectation | 17:48 |
| odyssey4me | until something to fix it merges in master it's not worth even talking about or setting the expectation of a backport | 17:48 |
| odyssey4me | if a backport happens, comments will show up in the bug when they are released anyway - so notifications happen | 17:49 |
| palendae | Hm, ok | 17:51 |
| palendae | I expect the bug to say "these branches are affected," actually getting a backport isn't necessarily tied to that | 17:51 |
| palendae | A branch could be affected, but won't necessarily get fixed | 17:51 |
| odyssey4me | the bug usually mentins branches affected | 17:51 |
| odyssey4me | the point is that when you add the series, people expect things to happen in them | 17:52 |
| palendae | Guess I have different expectations | 17:52 |
| palendae | Just based on the body, it gets really hard to search for bugs affecting, say, "mitaka" | 17:52 |
| odyssey4me | it's easy enough to note things in the comments/description | 17:52 |
| odyssey4me | but yeah, I'm not a fan of this silly limbo | 17:53 |
| palendae | Sounds like OpenStack's machinery does some weird stuff with that | 17:53 |
| odyssey4me | it's annoying that I can't use storyboard because it's not ready yet, and I can't use launchpad properly because we're somewhere in-between. | 17:53 |
| palendae | The branch affected is noted in the comments, just seems like a hassle to get an overview | 17:53 |
| *** palendae has left #openstack-ansible | 17:54 | |
| *** palendae has joined #openstack-ansible | 17:54 | |
| odyssey4me | if you'd like to volunteer to be a bug triager/maintainer then please go ahead. | 17:54 |
| palendae | Nope :p | 17:54 |
| palendae | Mostly curious | 17:54 |
| palendae | Was storyboard revived? | 17:54 |
| odyssey4me | I personally am not prepared to spend the time pruning all that. | 17:54 |
| odyssey4me | I have tried in the past and it's just too much overhead for me. | 17:55 |
| odyssey4me | And yes, storyboard never stopped being developed... but it's back in revival mode since the Tokyo summit. | 17:55 |
| palendae | I had heard rumblings that it wasn't being actively developed for a little bit | 17:55 |
| palendae | Not that launchpad really is | 17:56 |
| odyssey4me | the developers are quite responsive and like to talk about it :) | 17:56 |
| odyssey4me | I asked a question once, and a few days later I got this: https://storyboard.openstack.org/#!/story/2000667 | 17:56 |
| palendae | Cool | 17:56 |
| odyssey4me | it still needs a bit of work to be ready for more generl things - like there's no exposure of the kanban boards other than someone sending you a link. | 17:56 |
| odyssey4me | But it does look like a nifty tool. | 17:57 |
| odyssey4me | muck around a bit, then pop into #storyboard and give them feedback :) | 17:57 |
| odyssey4me | ok, now I really am out - night all | 17:59 |
| *** david-lyle has quit IRC | 18:00 | |
| *** KLevenstein has quit IRC | 18:02 | |
| *** sdake has quit IRC | 18:03 | |
| openstackgerrit | Flávio Ramalho proposed openstack/openstack-ansible: Adding a playbook for deploying Sahara https://review.openstack.org/360700 | 18:04 |
| *** david-lyle has joined #openstack-ansible | 18:04 | |
| *** sdake has joined #openstack-ansible | 18:04 | |
| *** david-lyle has quit IRC | 18:11 | |
| *** david-lyle has joined #openstack-ansible | 18:12 | |
| *** weezS has joined #openstack-ansible | 18:14 | |
| *** david-lyle_ has joined #openstack-ansible | 18:15 | |
| *** david-lyle_ has quit IRC | 18:15 | |
| *** schwicht has joined #openstack-ansible | 18:19 | |
| *** Andrew_jedi has quit IRC | 18:29 | |
| *** Mudpuppy_ has joined #openstack-ansible | 18:29 | |
| *** Mudpuppy has quit IRC | 18:30 | |
| spedione | Does anyone have any idea why networking wouldn't be working correctly after running the gate_check_commit script to set up an AIO? | 18:30 |
| *** electrofelix has quit IRC | 18:30 | |
| *** spedione is now known as chris_hultin | 18:30 | |
| openstackgerrit | Qin Wang (qwang) proposed openstack/openstack-ansible-os_aodh: Add SNI support via OS packages for os_aodh role https://review.openstack.org/356772 | 18:34 |
| *** jperry has quit IRC | 18:35 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Change nova_virt_type from lxc to lxd https://review.openstack.org/360730 | 18:38 |
| *** asettle has joined #openstack-ansible | 18:38 | |
| *** asettle has quit IRC | 18:38 | |
| *** asettle has joined #openstack-ansible | 18:39 | |
| *** Mudpuppy_ has quit IRC | 18:46 | |
| *** Mudpuppy has joined #openstack-ansible | 18:49 | |
| *** weezS has quit IRC | 18:50 | |
| *** krotscheck has left #openstack-ansible | 18:51 | |
| *** chas has joined #openstack-ansible | 18:54 | |
| *** hybridpollo has joined #openstack-ansible | 18:56 | |
| *** chas has quit IRC | 18:59 | |
| *** michaelgugino has quit IRC | 19:01 | |
| cloudnull | spedione. what are you seeing? | 19:03 |
| cloudnull | chris_hultin: just ping'd me about something similar. maybe related? | 19:03 |
| *** jperry has joined #openstack-ansible | 19:08 | |
| *** david-lyle has quit IRC | 19:10 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Change nova_virt_type from lxc to lxd https://review.openstack.org/360730 | 19:11 |
| *** guilherme has joined #openstack-ansible | 19:12 | |
| logan- | Andrew_jedi: what version of ceph are you on? Recent versions support keystone token auth in lieu of admin tokens | 19:13 |
| logan- | See http://docs.ceph.com/docs/master/radosgw/keystone/ | 19:14 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create: Add retry to container start/stop handlers https://review.openstack.org/360681 | 19:20 |
| *** jheroux has joined #openstack-ansible | 19:21 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create: Add SNI support to lxc container create role via OS packages https://review.openstack.org/356713 | 19:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create: Ansible 2.1.1 role testing https://review.openstack.org/354373 | 19:23 |
| *** johnmilton has quit IRC | 19:32 | |
| *** messy has joined #openstack-ansible | 19:33 | |
| chris_hultin | cloudnull: Sorry, my name reset for some reason. Same person, same question :D | 19:34 |
| matias | if, for some reason, you destroy the repo container or make it inacessible, you can't run setup-hosts anymore because the apt commands fail | 19:36 |
| matias | is there some auxiliary playbook to temporarily disable the apt cache in all nodes? | 19:36 |
| *** McMurlock1 has joined #openstack-ansible | 19:36 | |
| matias | hm I will just remove /etc/apt/apt.conf.d/00apt-cacher-proxy | 19:43 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 19:44 |
| -openstackstatus- NOTICE: The Gerrit service on review.openstack.org is restarting to implement some performance tuning adjustments, and should return to working order momentarily. | 19:45 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 19:46 |
| *** jheroux has quit IRC | 19:48 | |
| *** johnmilton has joined #openstack-ansible | 19:48 | |
| *** weezS has joined #openstack-ansible | 19:49 | |
| *** Mudpuppy_ has joined #openstack-ansible | 19:49 | |
| *** Mudpupp__ has joined #openstack-ansible | 19:50 | |
| *** Mudpuppy_ has quit IRC | 19:51 | |
| *** jheroux has joined #openstack-ansible | 19:51 | |
| *** Mudpupp__ is now known as Mudpuppy_ | 19:52 | |
| *** Mudpuppy has quit IRC | 19:53 | |
| *** johnmilton has quit IRC | 19:53 | |
| cloudnull | chris_hultin: ah! | 19:55 |
| cloudnull | makes sense | 19:55 |
| chris_hultin | Yep. Sorry about that | 19:55 |
| *** chas has joined #openstack-ansible | 19:55 | |
| cloudnull | matias: thats how i'd do it | 19:56 |
| cloudnull | simply remove the proxy file until the apt / repo container bits are made happy | 19:56 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create: Add SNI support to lxc container create role via OS packages https://review.openstack.org/356713 | 19:59 |
| *** jheroux has quit IRC | 19:59 | |
| *** Mudpuppy_ is now known as Mudpuppy | 19:59 | |
| *** chas has quit IRC | 20:00 | |
| *** sdake has quit IRC | 20:03 | |
| *** jheroux has joined #openstack-ansible | 20:03 | |
| *** KLevenstein has joined #openstack-ansible | 20:03 | |
| *** sdake has joined #openstack-ansible | 20:04 | |
| *** johnmilton has joined #openstack-ansible | 20:09 | |
| *** johnmilton has quit IRC | 20:15 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 20:17 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-lxc_container_create: Ansible 2.1.1 role testing https://review.openstack.org/354373 | 20:24 |
| cloudnull | well... thats bad git://labs.feurix.org/feurix/admin/hatop.git == errno=Connection refused and http://hatop.googlecode.com/files/hatop-0.7.7.tar.gz == 404 | 20:25 |
| cloudnull | seems the storage url changed | 20:27 |
| cloudnull | it is now https://storage.googleapis.com/google-code-archive-downloads/v2/code.google.com/hatop/hatop-0.7.7.tar.gz | 20:27 |
| *** johnmilton has joined #openstack-ansible | 20:28 | |
| *** sdake has quit IRC | 20:29 | |
| *** asettle has quit IRC | 20:32 | |
| openstackgerrit | Merged openstack/openstack-ansible-security: Ensure AIDE initializes on subsequent runs https://review.openstack.org/359554 | 20:33 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 20:33 |
| *** sdake has joined #openstack-ansible | 20:35 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-haproxy_server: fix broken hatop URL https://review.openstack.org/360778 | 20:36 |
| *** asettle has joined #openstack-ansible | 20:36 | |
| cloudnull | odyssey4me automagically jmccrory stevelle andymccr hughsaunders d34dh0r53 ^ that needs to go in | 20:36 |
| stevelle | wow, that makes me sad cloudnull. dead project? | 20:37 |
| cloudnull | seems that way | 20:37 |
| cloudnull | the script is still super useful | 20:38 |
| cloudnull | but we might want to cache an archive in various places just in case. | 20:38 |
| *** asettle has quit IRC | 20:45 | |
| *** chris_hultin is now known as spedione|AWAY | 20:46 | |
| *** Mudpuppy has quit IRC | 20:48 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 20:50 |
| *** Mudpuppy has joined #openstack-ansible | 20:52 | |
| stevelle | automagically: did you get a chance to test ssl termination on https://review.openstack.org/#/c/358798/? | 20:54 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Disable automatic ToC generation https://review.openstack.org/359547 | 20:54 |
| automagically | Nope, haven’t done so | 20:54 |
| stevelle | is that a feature we even care about in keystone or do we only care about LB termination now? | 20:56 |
| stevelle | I feel like I haven't kept up to speed on that detail | 20:56 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-lxc_container_create: Fix incorrect galera logs bind mount bug https://review.openstack.org/358852 | 20:57 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 20:58 |
| *** Andrew_jedi has joined #openstack-ansible | 21:00 | |
| *** cfarquhar has joined #openstack-ansible | 21:01 | |
| *** matias has quit IRC | 21:02 | |
| *** Mudpuppy has quit IRC | 21:05 | |
| *** sdake_ has joined #openstack-ansible | 21:05 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 21:07 |
| *** schwicht has quit IRC | 21:08 | |
| automagically | stevelle: We added LB termination without removing Apache termination, so probably best to leave it | 21:08 |
| *** sdake has quit IRC | 21:08 | |
| *** woodard_ has joined #openstack-ansible | 21:09 | |
| *** thorst has quit IRC | 21:09 | |
| *** thorst has joined #openstack-ansible | 21:12 | |
| *** abelur_ has joined #openstack-ansible | 21:12 | |
| *** admin0 has quit IRC | 21:12 | |
| *** woodard has quit IRC | 21:12 | |
| *** woodard_ has quit IRC | 21:13 | |
| thotypous | br-vxlan 8000.1a898668c134 no 6c550730_eth1 | 21:14 |
| thotypous | there is some regex bug in the veth pairing script | 21:14 |
| thotypous | anyway, need to go now.. cya.. i'll hunt it in a few hours | 21:15 |
| *** thorst has quit IRC | 21:16 | |
| *** abelur_ has quit IRC | 21:17 | |
| *** abelur_ has joined #openstack-ansible | 21:18 | |
| *** sdake_ has quit IRC | 21:20 | |
| mrda | Morning OSA | 21:21 |
| rromans | b3rnard0: http://screenful.com/trello/power-up | 21:22 |
| *** sdake has joined #openstack-ansible | 21:22 | |
| *** guilherme has quit IRC | 21:22 | |
| b3rnard0 | rromans: trello is dead; long live Jira | 21:24 |
| cloudnull | bah... | 21:24 |
| rromans | b3rnard0: *mumble* | 21:25 |
| *** david-lyle has joined #openstack-ansible | 21:26 | |
| *** McMurlock1 has quit IRC | 21:26 | |
| *** ChanServ sets mode: +o cloudnull | 21:27 | |
| *** b3rnard0 was kicked by cloudnull (User terminated!) | 21:27 | |
| cloudnull | be-gone jira ! | 21:27 |
| cloudnull | :P | 21:27 |
| cjloader | go jira | 21:27 |
| cjloader | lol | 21:27 |
| cloudnull | cjloader: silense | 21:28 |
| cjloader | lol | 21:28 |
| cloudnull | jira is not all that bad | 21:28 |
| cjloader | no, long live juno.... | 21:29 |
| cjloader | lol | 21:29 |
| cloudnull | hahah | 21:29 |
| cjloader | cloudnull: you know I'm no juno fan | 21:30 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-haproxy_server: fix broken hatop URL https://review.openstack.org/360778 | 21:30 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-haproxy_server: Fix broken hatop URL and role gate https://review.openstack.org/360778 | 21:31 |
| *** david-lyle has quit IRC | 21:31 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate: DOCS - point homepage to role developer docs https://review.openstack.org/360542 | 21:34 |
| *** jperry has quit IRC | 21:35 | |
| openstackgerrit | Merged openstack/openstack-ansible: [DOCS} Further edits, corrects to draft install https://review.openstack.org/360548 | 21:35 |
| cloudnull | we need another core to make https://review.openstack.org/#/c/360778 go | 21:37 |
| cloudnull | thanks automagically for fixing it up | 21:37 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 21:39 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server: Fix broken hatop URL and role gate https://review.openstack.org/360778 | 21:39 |
| *** jamesdenton has quit IRC | 21:40 | |
| *** schwicht has joined #openstack-ansible | 21:43 | |
| *** klamath_ has quit IRC | 21:52 | |
| *** strobelight has quit IRC | 21:55 | |
| *** chas has joined #openstack-ansible | 21:56 | |
| *** schwicht has quit IRC | 21:57 | |
| *** gouthamr has quit IRC | 22:00 | |
| *** gouthamr has joined #openstack-ansible | 22:00 | |
| *** gouthamr has quit IRC | 22:00 | |
| *** chas has quit IRC | 22:01 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server: Remove apt update from Vagrantfile https://review.openstack.org/358054 | 22:02 |
| *** BjoernT has quit IRC | 22:03 | |
| *** Andrew_jedi has quit IRC | 22:05 | |
| *** michauds has quit IRC | 22:05 | |
| *** spotz is now known as spotz_zzz | 22:11 | |
| *** woodard has joined #openstack-ansible | 22:11 | |
| stevelle | automagically: my question was more about the nginx solution and not having tested termination in the container yet. Doesn't seem that important with that context. | 22:11 |
| cloudnull | stevelle: i've been looking over your uwsgi bits, that's well done. | 22:12 |
| cloudnull | if we can apply that pattern to all of the services, then interal SSL for all services is really a non-issue | 22:12 |
| stevelle | good to hear, I wont fuss over the ssl termination then | 22:15 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_keystone: Install and configure Nginx https://review.openstack.org/358798 | 22:17 |
| *** schwicht has joined #openstack-ansible | 22:17 | |
| *** sdake has quit IRC | 22:24 | |
| *** gouthamr has joined #openstack-ansible | 22:25 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_keystone: Implement CentOS 7 support in os_keystone https://review.openstack.org/320216 | 22:25 |
| cloudnull | stevelle: on that note, when might you thing that the keystone nginx patch will be ready? | 22:25 |
| cloudnull | is it ready now? | 22:25 |
| *** jheroux has quit IRC | 22:26 | |
| cloudnull | asking because i've been working on keystone+cent and dont want to stop on work inflight | 22:26 |
| cloudnull | nad if i can rebase off your existing patch then all good. | 22:26 |
| cloudnull | etc | 22:26 |
| *** schwicht has quit IRC | 22:27 | |
| stevelle | cloudnull: the nginx patch is ready. I have one or two patches that need to follow for splitting apache from mod_wsgi but I can follow your patches with that | 22:28 |
| stevelle | I'd say rebase off the existing patch and go cloudnull | 22:28 |
| *** strobelight has joined #openstack-ansible | 22:29 | |
| cloudnull | ok | 22:32 |
| cloudnull | will do | 22:32 |
| *** KLevenstein has quit IRC | 22:35 | |
| *** strobelight_ has joined #openstack-ansible | 22:35 | |
| *** strobelight is now known as Guest21496 | 22:35 | |
| *** strobelight_ is now known as strobelight | 22:36 | |
| *** sdake has joined #openstack-ansible | 22:37 | |
| *** strobelight has quit IRC | 22:39 | |
| *** Guest21496 has quit IRC | 22:40 | |
| *** thorst has joined #openstack-ansible | 22:42 | |
| *** Guest1506 is now known as jamesdenton | 22:45 | |
| *** thorst has quit IRC | 22:47 | |
| *** thorst has joined #openstack-ansible | 22:47 | |
| *** cfarquhar has quit IRC | 22:52 | |
| *** thorst has quit IRC | 22:56 | |
| *** schwicht has joined #openstack-ansible | 22:57 | |
| *** klamath has joined #openstack-ansible | 22:58 | |
| *** gouthamr_ has joined #openstack-ansible | 22:59 | |
| *** gouthamr has quit IRC | 23:03 | |
| *** hblah has joined #openstack-ansible | 23:22 | |
| *** AnarchyAo has quit IRC | 23:41 | |
| *** klamath has quit IRC | 23:52 | |
| *** david-lyle has joined #openstack-ansible | 23:52 | |
| *** thorst has joined #openstack-ansible | 23:54 | |
| *** chas has joined #openstack-ansible | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!