16:00:13 <tiffanie> #startmeeting openstack_ansible_meeting 16:00:14 <openstack> Meeting started Tue Sep 17 16:00:13 2019 UTC and is due to finish in 60 minutes. The chair is tiffanie. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:17 <openstack> The meeting name has been set to 'openstack_ansible_meeting' 16:00:32 <tiffanie> #topic office hours 16:00:33 <guilhermesp> o/ 16:00:51 <cjloader> o/ 16:01:12 <jrosser> o/ 16:04:09 <chandankumar> \o/ 16:04:38 <noonedeadpunk> o/ 16:05:23 <noonedeadpunk> so, upgrade jobs passing https://review.opendev.org/#/c/671105/ !!! 16:06:07 <cjloader> nice! 16:06:29 <noonedeadpunk> and, milestone https://review.opendev.org/#/c/681002/ was updated today 16:06:56 <noonedeadpunk> But due to keystone bug we can't merge it now 16:07:10 <guilhermesp> nice noonedeadpunk 16:07:39 <jrosser> we are going to need to work on tempest to land the keystone bug 16:08:03 <jrosser> becasue it needs this https://review.opendev.org/678379 16:08:04 <noonedeadpunk> jrosser: so, they're waiting our CI to pass before merging the patch? 16:08:19 <jrosser> no, i think it just needs review on the keystone side tbh 16:08:23 <jrosser> we have other issues :) 16:08:36 <cjloader> like swift 16:08:37 <mnaser> awesome work on ugprades, noonedeadpunk 16:08:39 <cjloader> lol 16:08:42 <mnaser> this is _very_ good 16:09:38 <noonedeadpunk> probably it's worth to include upgrade to milestone as well? 16:10:05 <spotz> o/ 16:10:05 <mnaser> i personally think its very valuable to have that locked in 16:10:23 <mnaser> what are thoughts from others? 16:11:59 <jrosser> the centos upgrade is a long job - i hope we don't hit timeouts 16:12:58 <noonedeadpunk> yeah... 16:13:04 <noonedeadpunk> but I know how to reduce time 16:13:40 <noonedeadpunk> we need https://review.opendev.org/#/c/682103/ 16:14:02 <noonedeadpunk> to be able to skip tempest tests after stein deployment 16:15:44 <noonedeadpunk> https://review.opendev.org/#/c/671105/33/scripts/gate-check-commit.sh 16:17:25 <gillesMo> Hello ! I've just added a compute node on my cluster (Rocky/Ubuntu16.04). This node is Ubuntu 18.04, and I don't see how to make the repo build venvs and co for Ubuntu 18.04, it does not running all the playbooks 16:19:24 <jrosser> noonedeadpunk: so i guess the upgrade is surviving a py2 -> py3 transition of some sort? 16:19:45 <jrosser> i figured that the ansible-runtime venv won't be recreated as py3 though 16:20:27 <noonedeadpunk> I guess I haven't rebased on top of py3 yet 16:20:42 <noonedeadpunk> :( 16:24:13 <jrosser> actually you may be ok https://zuul.opendev.org/t/openstack/build/8922b05644d6462ca472b071ba824187/log/job-output.txt#12667 16:25:19 <spatel> gillesMo: You should hold on for sometime, look like weekly meeting going on. 16:25:31 <noonedeadpunk> oh, indeed 16:25:51 <noonedeadpunk> but it's for ansible. Dunno about projects 16:26:07 <noonedeadpunk> But seems that uwsgi has python3 suffix as well 16:26:47 <jrosser> cool 16:27:35 <noonedeadpunk> since aio don't contain swift or ironic - I'm ok :p 16:27:48 <cjloader> ... 16:27:50 <cjloader> lol 16:29:06 <cjloader> speaking of swift 16:29:22 <noonedeadpunk> jrosser: btw, regarding swift - should we edit our template as for now and keep track on the reported bug? 16:29:34 <cjloader> +1 noonedeadpunk 16:29:35 <noonedeadpunk> To rollback if they'll decide to use raw 16:29:56 <jrosser> you could propose a patch a revert at the same time so we remember 16:30:01 <jrosser> *and a 16:30:51 <jrosser> there are generally a few broken things we need to address 16:30:55 <cjloader> i'm all for that jrosser 16:31:05 <noonedeadpunk> sure:) 16:31:09 <jrosser> swift like we just mentioned which we can deal with 16:31:22 <cjloader> it'd help unblock the ironic efforts as well 16:31:32 <jrosser> neutron/calico is a python3 problem too 16:31:45 <jrosser> keystone bug + related tempest update 16:32:00 <jrosser> and i would like to land bind-to-mgmt too which needs some more work 16:32:14 <jrosser> plus i think we are about to learn what breaks with centos7.7 16:32:40 <noonedeadpunk> :) 16:32:50 <jrosser> and haproxy on metal looks very close too 16:33:02 <guilhermesp> Yeah I put a hold on the galera issue but will come back to it 16:33:03 <noonedeadpunk> I guess it's ok 16:33:13 <noonedeadpunk> regarding haproxy) 16:34:33 <jrosser> and one all that is sorted out there is a bunch of stuff folk have contributed which has got a bit backed up 16:38:14 <noonedeadpunk> jrosser: btw, I guess we need to walk through distro variables and ensure that ubuntu at least uses py3 everywhere 16:42:21 <noonedeadpunk> or leave it afterwards 16:46:09 <jrosser> so for bind-to-mgmt i want to tidy this up https://review.opendev.org/#/c/670051/ 16:46:42 <jrosser> far far too many variables, that needs cutting back and a global one introducing a bit like the py2/3 switch variable 16:50:40 <guilhermesp> jrosser: you mean by like adding a global var in each role correspondent to the mgmt-adddr? 16:50:59 <jrosser> yes, it's a bunch of one line patches to os_* i think 16:51:10 <jrosser> i was going to leave all the infrastructure things as they are 16:52:17 <openstackgerrit> Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_swift master: Update fallocate_reserve to work with py3 https://review.opendev.org/682697 16:53:01 <jrosser> guilhermesp: nova_service_bind_address: "{{ openstack_service_bind_address | default('0.0.0.0') }}" 16:53:05 <jrosser> something like that 16:53:08 <noonedeadpunk> jrosser: for swift I've updated your's dnm if you don't mind:) 16:53:18 <guilhermesp> sounds clearer jrosser 16:53:46 <guilhermesp> for galera, we have abandoned the patch to replace the bind addrs but i need to figure out why it is still breaking on the integrated repo 16:54:08 <guilhermesp> practically on a production env it works as it is proposed in the patch 16:54:43 <jrosser> noonedeadpunk: thats fine 16:55:19 <jrosser> mnaser: anything from you on all this stuff? :) 16:55:29 <noonedeadpunk> guilhermesp: https://review.opendev.org/#/c/677830/ seems to be passing CI. Going to check what';s wrong with CLI 16:56:03 <guilhermesp> yeah noonedeadpunk it was passing all the time 16:56:07 <mnaser> jrosser: about galera.. everything we've used to deploy is all that i've handed to guilhermesp .. so im not sure ~how~ 16:56:14 <guilhermesp> now I guess murano-api is up, but cli returns bad request 16:56:15 <noonedeadpunk> not the last one:) 16:56:43 <guilhermesp> yeah I kinda of reproduced the vars we needed to set in 2 metal deployments 16:57:01 <guilhermesp> but was a mutlinode deployment 16:57:09 <guilhermesp> things could bahave differenctly in aio 16:57:19 <guilhermesp> in terms of galera 16:58:25 <mnaser> yeah, im not sure how/why it would behave different, but it does. 17:00:11 <tiffanie> #endmeeting