16:01:47 <evrardjp> #startmeeting openstack_ansible_meeting 16:01:47 <openstack> Meeting started Tue Apr 11 16:01:47 2017 UTC and is due to finish in 60 minutes. The chair is evrardjp. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:50 <openstack> The meeting name has been set to 'openstack_ansible_meeting' 16:01:59 <andymccr> woo bug triage :D 16:02:10 * lbragstad lingers 16:02:10 <asettle> \o/ 16:02:15 <asettle> lbragstad: it's too fun to leave 16:02:31 <lbragstad> much fun #amaze 16:02:41 <asettle> lbragstad: just wait until evrardjp and andymccr get started 16:02:43 <asettle> It's a rollercoaster 16:02:44 <andymccr> lbragstad: we don't have any opinionated things i think - we would carry an empty template that can be "template config'd) 16:02:55 <lbragstad> andymccr ++ 16:02:59 * lbragstad grabs some popcorn 16:03:11 <evrardjp> #topic Review of last week action points 16:03:16 <evrardjp> well. 16:03:24 <evrardjp> I'm sad to say I didn't do my work properly 16:03:33 <evrardjp> #action evrardjp review https://bugs.launchpad.net/openstack-ansible/+bug/1678165 16:03:34 <openstack> Launchpad bug 1678165 in openstack-ansible "dynamic_inventory.py container_mtu bug" [Undecided,New] 16:03:36 <andymccr> you should've been here in the beginning weeks of the cycle because that was a rollercoaster - but now its manageable :P 16:03:55 <evrardjp> No other action points 16:04:08 <evrardjp> #topic this week triage \o/ 16:04:24 <evrardjp> so our first for today is: 16:04:26 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1681742 16:04:27 <openstack> Launchpad bug 1681742 in openstack-ansible "Securing services with SSL certificates link broken" [Undecided,In progress] - Assigned to Chason (chen-xing) 16:04:57 <evrardjp> that was easy. 16:05:13 <evrardjp> we just have to review what was done 3 minutes ago! 16:05:24 <evrardjp> next 16:05:27 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1681714 16:05:28 <openstack> Launchpad bug 1681714 in openstack-ansible " Ceph cluster’s performance are not monitored by Telegraf" [Undecided,New] - Assigned to Bertrand Lallau (bertrand-lallau) 16:05:42 <evrardjp> confirmed wishlist ? 16:05:55 <andymccr> yeah 16:05:56 <andymccr> agreed 16:05:57 <spotz> yeah 16:06:04 <evrardjp> next 16:06:04 <andymccr> also looks like Bertrand is taking care of it which is awesome :) 16:06:05 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1681695 16:06:06 <openstack> Launchpad bug 1681695 in openstack-ansible "Incorrect keystone with multiple memcache configuration" [Undecided,New] - Assigned to Jean-Philippe Evrard (jean-philippe-evrard) 16:06:22 <evrardjp> what??? I was assigned this? 16:06:38 <evrardjp> k 16:06:40 <spotz> hehehe, enjoy?:) 16:06:46 <evrardjp> let's see how this one goes. 16:07:01 <andymccr> evrardjp: its cos you wrote a patch :) 16:07:02 <evrardjp> I'm not sure if my fix gonna fix everything, so let's see with expert keystone advice 16:07:06 <andymccr> so far so good on bug triage! 16:07:10 <evrardjp> yeah. 16:07:14 <evrardjp> but it's related 16:07:17 <evrardjp> not fixed 16:07:23 <andymccr> damnit! fix it properly :P 16:07:27 <evrardjp> haha 16:07:59 <evrardjp> let's move on we'll see later if that needs a change of status 16:08:13 <evrardjp> next 16:08:13 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1681372 16:08:14 <openstack> Launchpad bug 1681372 in openstack-ansible "Having a pre installed Ansible in a different location causes AIO issues." [Undecided,New] 16:08:46 <andymccr> ahh yeah this is kinda interesting 16:09:15 <evrardjp> agreed. 16:09:20 <stevelle> medium or high? 16:09:23 <evrardjp> I think we should let ppl use their own ansible 16:09:30 <andymccr> agreed 16:09:35 <andymccr> although im wondering if we shouldn't deploy our ansible in a venv? 16:09:35 <evrardjp> I'd be enclined to make it medium at minimum 16:09:37 <andymccr> we create a binary anyway? 16:09:47 <openstackgerrit> chen.xing proposed openstack/openstack-ansible-os_keystone master: Fix the dead link https://review.openstack.org/455766 16:10:15 <evrardjp> well the question is do we need to andymccr? 16:10:36 <evrardjp> we could leave the responsibility to the deployer if he doesn't want to use bootstrap_ansible 16:10:39 <andymccr> evrardjp: well that way we can allow you to use whatever versions of whatever you want. 16:10:49 <andymccr> agreed - but perhaps if you want to use the bootstrap (and know it'll work) we can support it in a venv 16:11:25 <andymccr> i think hughsaunders came across a slightly similar but unrelated issue around the versioning of pip packages we install which also wouldn't have happened if we used a venv 16:11:30 <evrardjp> oh you mean adding a new feature of "bring your own ansible" and we just link to it, as we do now? 16:11:47 <stevelle> haven't we tried ansible in venv and found it didn't behave nicely? 16:11:51 <evrardjp> and then continue as we do (the roles etc) 16:12:00 <evrardjp> stevelle: I think it's in a venv now :p 16:12:07 <andymccr> stevelle: not sure - im mostly just theorizing. 16:12:13 <andymccr> evrardjp: i dont think it is in integrated build 16:12:16 <evrardjp> ok 16:12:24 <andymccr> in role builds it is though since tox runs environments :P 16:12:54 <evrardjp> I thought it was in /opt/ansible-runtine 16:12:57 <evrardjp> runtime* 16:13:12 <evrardjp> anyway 16:13:15 <evrardjp> The question is 16:13:20 <andymccr> the key issue, imo, is that you (as a deployer) could have any sort of tools that have package version issues with our deploy - and since its not in a venv that would cause problems 16:13:25 <evrardjp> how do we prioritize this 16:13:34 <andymccr> hmm 16:13:49 <evrardjp> is that a new feature we want, or do we think we should always have this 16:13:49 <andymccr> cloudnull: ^ thoughts if you're about? 16:14:24 <evrardjp> IMO it's a bug, and everyone should be able to use its own ansible. But not using bootstrap_ansible then. So the bug only becomes a docs bug. 16:14:34 <evrardjp> Which changes the impact. 16:14:41 <spotz> Which means me! 16:14:54 <evrardjp> :) 16:15:11 <spotz> And clear comments on the bug as to what we need to fix:) 16:15:18 <andymccr> evrardjp: hmm - i guess you could say it that way. an alternative fix is to only link the binaries we know exist, e.g. which ansible-playbook and then link from inside that dir rather than the ansible-runtime/bin 16:15:44 <evrardjp> yes, in the script we have a || ansible-playbook we could maybe remove 16:16:00 <evrardjp> so confirmed medium + docs? 16:16:07 <andymccr> evrardjp: yeah lets say that 16:16:31 <stevelle> why can't they use bootstrap_ansible though 16:16:39 <openstackgerrit> chen.xing proposed openstack/openstack-ansible master: Fix the dead link https://review.openstack.org/455771 16:16:45 <stevelle> our wrapper turns out to be pretty helpful 16:17:15 <evrardjp> stevelle: IMO if the deployer wants to have its own ansible, we should consider (s)he skips the ansible building 16:17:29 <evrardjp> the deployer should be smart enough to install roles, and load variables. 16:17:53 <stevelle> that breaks every bit of doc we have, which assumes the wrapper 16:17:54 <evrardjp> well that's my opinion :p 16:18:06 <stevelle> it's a high hurdle 16:18:32 <evrardjp> it's just a question of documenting that if you're not using bootstrap ansible (and using your own), you can't use the wrapper and the nice tooling we've made 16:18:51 <andymccr> so can we not just rewrite that code to check where ansible-playbook binary exists and not just link with an assumption 16:19:12 <evrardjp> but I already see issues coming with older ansible versions that are breaking things because we just "re-use" the old ansible binary 16:19:17 <andymccr> i still quite like the idea of having it in a venv since it means we can have our own requirements and the deployer can use whatever requirements/packages they want outisde of that binary - but im not sure how possible that is :) 16:19:18 <evrardjp> binaries* 16:19:37 <evrardjp> I don't think that will be that hard 16:19:39 <andymccr> evrardjp: i think thats the benefit of the bootstrap process - we can ensure you are using a tested/up to date version that works with our plays 16:19:44 <evrardjp> maybe we should build a test there 16:19:46 <evrardjp> haha. 16:19:46 <andymccr> yeah 16:19:50 <evrardjp> good thought ;) 16:20:04 <openstackgerrit> chen.xing proposed openstack/openstack-ansible master: Fix the dead link https://review.openstack.org/455771 16:20:06 <andymccr> stevelle may be right though around the ansible version stuff 16:20:15 <evrardjp> there is probably something in the meta that could be used if you see what I mean :D 16:20:59 <stevelle> we can add an ansible version check to the bootstrap, for existing packages pretty easily 16:21:21 <openstackgerrit> Amy Marrich (spotz) proposed openstack/openstack-ansible master: [DOCS] Fix the dead link https://review.openstack.org/455771 16:21:53 <evrardjp> I think we all agree there 16:22:02 <evrardjp> let's continue then! 16:22:04 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1680948 16:22:05 <openstack> Launchpad bug 1680948 in openstack-ansible "cannot see status of mysql cluster with -h localhost on centos" [Undecided,New] 16:22:31 <andymccr> im sure this is a centos issue - we had problems when we first added centos support because localhost didnt work on mysql where 127.0.0.1 did 16:22:47 <evrardjp> probably a my.cnf change? 16:22:50 <evrardjp> I don't know 16:23:15 <evrardjp> mgariepy: are you there? 16:24:02 <andymccr> i think -h 127.0.0.1 works 16:24:23 <evrardjp> Is this someone you did directly (inspired by our docs), or just because you needed it? In the former could you point to docs, in the latter did you try 127.0.0.1 ? 16:24:42 <evrardjp> let's continue before confirming the source of the issue 16:24:55 <andymccr> ok cool 16:24:57 <evrardjp> let's continue* 16:25:06 <evrardjp> we'll see when he'll be back :) 16:25:11 <evrardjp> next 16:25:14 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1680233 16:25:15 <openstack> Launchpad bug 1680233 in openstack-ansible "re-running os-glance with NFS mount breaks" [Undecided,New] 16:25:39 <andymccr> i put a comment in there 16:25:47 <andymccr> (the centos mysql one) 16:26:02 <openstackgerrit> Amy Marrich (spotz) proposed openstack/openstack-ansible-os_horizon master: [DOCS] Fix the dead link https://review.openstack.org/455763 16:26:14 <evrardjp> andymccr: thanks. 16:27:15 <andymccr> that sounds real and i'd say high 16:28:00 <evrardjp> Agreed. 16:28:19 <evrardjp> do you have cycles for it? 16:28:32 <andymccr> evrardjp: yeah i'll take it 16:28:39 <evrardjp> cool thanks 16:28:44 <evrardjp> next 16:28:45 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1680075 16:28:47 <openstack> Launchpad bug 1680075 in openstack-ansible "keystone fails with haproxy in maint mode" [Undecided,New] 16:29:02 <evrardjp> logan-: ? 16:29:17 <logan-> looking 16:29:26 <evrardjp> it could be useful for you I think :) 16:30:10 <logan-> oh you know what 16:30:12 <logan-> I need to wrap up https://review.openstack.org/#/c/438224/ 16:30:16 <logan-> ill do that asa 16:30:18 <logan-> asap* 16:30:41 <evrardjp> haha great! 16:30:48 <evrardjp> don't forget to tag the bug there :) 16:30:52 <logan-> yep will do 16:30:56 <logan-> can you assign the bug to me please 16:30:58 <evrardjp> thanks! 16:31:00 <evrardjp> sure 16:31:47 <evrardjp> last one was the one of last week 16:31:50 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1678165 16:31:51 <openstack> Launchpad bug 1678165 in openstack-ansible "dynamic_inventory.py container_mtu bug" [Undecided,New] 16:31:57 <evrardjp> I didn't got the chance to triage this 16:32:14 <andymccr> hmm 16:32:15 <evrardjp> did someone got the chance to do it? 16:32:30 <evrardjp> I added it again to my action points for next week. 16:32:56 <evrardjp> I think Nolan is on holiday 16:33:07 <evrardjp> cloudnull: opinion ? 16:33:29 <cloudnull> confirmed 16:33:42 <evrardjp> k 16:33:53 <cloudnull> if that option is set when the inventory is first created it works 16:34:00 <cloudnull> otherwise it does not 16:34:10 <cloudnull> and requires the deployer to go in and manually set it 16:34:17 <cloudnull> I worked with BjoernT on that 16:34:20 <evrardjp> ok 16:34:31 <evrardjp> importance? 16:34:38 <evrardjp> Is there a fix somewhere? 16:34:57 <openstackgerrit> Andy McCrae proposed openstack/openstack-ansible-os_glance master: Default to "omit" mode from directory creation https://review.openstack.org/455776 16:35:00 <BjoernT> probably prio2 16:35:05 <evrardjp> (or in the work) 16:35:18 <evrardjp> BjoernT: haha :) 16:35:23 <BjoernT> lol 16:35:39 <BjoernT> yeah I was wondering what else is broken... 16:35:41 <evrardjp> it sounds serious because of expactations 16:35:43 <evrardjp> -a 16:35:50 <evrardjp> yeah 16:36:14 <mgariepy> andymccr, i'll take care of https://bugs.launchpad.net/openstack-ansible/+bug/1680948 when I have 2 minutes. 16:36:16 <openstack> Launchpad bug 1680948 in openstack-ansible "cannot see status of mysql cluster with -h localhost on centos" [Undecided,New] - Assigned to Marc Gariépy (mgariepy) 16:36:22 <evrardjp> I'd rather say high 16:36:23 <andymccr> ok cool thanks mgariepy :) 16:36:31 <evrardjp> thanks mgariepy! 16:36:36 <mgariepy> small doc update :) 16:36:45 <evrardjp> ahah I knew it! :p 16:36:56 <evrardjp> anyway, for the last bug? 16:36:57 <spotz> ooh ohh docs to me!!!! 16:37:09 <evrardjp> Has someone cycles to work on this? 16:37:16 <evrardjp> That's gonna be a fun one and important ones 16:37:36 <evrardjp> cloudnull: you are already familiar with it? 16:37:59 <evrardjp> s/ones/one/ 16:37:59 <cloudnull> on what ? 16:38:04 <evrardjp> on https://bugs.launchpad.net/openstack-ansible/+bug/1678165 16:38:05 <openstack> Launchpad bug 1678165 in openstack-ansible "dynamic_inventory.py container_mtu bug" [High,Confirmed] 16:38:20 <evrardjp> after that we can close the bug triage 16:38:24 <evrardjp> for today 16:38:56 <evrardjp> Ok I'll stop pushing :) 16:39:13 <evrardjp> That's all for today 16:39:15 <evrardjp> Thanks everyone! 16:39:19 <evrardjp> #endmeeting