16:00:03 #startmeeting openstack_ansible_meeting 16:00:04 Meeting started Tue Jun 19 16:00:03 2018 UTC and is due to finish in 60 minutes. The chair is mnaser. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:05 o/ 16:00:06 #topic rollcall 16:00:07 The meeting name has been set to 'openstack_ansible_meeting' 16:00:12 o/ 16:00:12 o/ 16:00:15 \o/ 16:00:17 o/ 16:00:47 while people check in, feel free to look over https://wiki.openstack.org/wiki/Meetings/openstack-ansible#Agenda and review 'last week highlights' if you have something to bring to the table :) 16:01:16 mnaser: it would be better to write it here, so that's recorded in the minutes. 16:01:30 yes, i'll do that once i get to it, just wanted to give folks a chance to add something if they wanted to add :) 16:01:35 but i'll get us started 16:01:43 #topic what happened since last meeting 16:01:54 so i had a pretty productive weekend. 16:01:55 o/ 16:02:00 mnaser: it looks so :0 16:02:02 the mess i made with missing epel packages is mostly cleaned up 16:02:02 :) 16:02:17 https://review.openstack.org/#/c/576305/ this is one that might be one of the last 16:02:19 o/ 16:02:35 so hopefully with that merged, we have very minimal epel dependency 16:02:49 also, i've worked with rdo and we just got a +1 from jenkins on the job to bring uwsgi into rocky 16:03:05 so keep an eye out on https://review.rdoproject.org/r/#/c/14240/ .. once that's in, we'll swap out some repos and we'll get uwsgi. 16:03:13 so: yay, that'll help hwoarang's work on distro packages 16:03:32 so, just got this message 16:03:34 amoralej> mnaser, uwsgi is now tagged in rocky-testing, it will be in the repos in next 15 minutes or so 16:03:35 yay! 16:03:44 anyways, next up, cleaning up ci jobs 16:04:17 you'll see me push a lot of changes, i'm hoping that we can adopt using project-templates rather than job inside every repo. it'll keep things a lot more consistent. long term, i'd like to work on stabilizing our integrated repo jobs and run those against roles 16:04:30 so we don't have wild breakages. we're still a bit far away from that, but that's something to work on 16:04:59 today a change merged to add percona repos to infra caches which means no more unreliable percona mirrors, yay. https://review.openstack.org/#/c/576241/ adds that and it has passed, confirmed with infra that it's using it 16:05:31 and finally, nspawn jobs in integrated are failing, i added the fix but the jobs are broken and it's beyond my understanding.. so.. https://review.openstack.org/#/c/576307/ -- if anyone knows nspawn, please halp :) 16:05:46 that's it for me. just wanted to get everyone up to date at what i've been up to. 16:05:51 anyone else? :) 16:06:04 nothing to add 16:06:09 I have my doc patch up from last week's triage 16:06:10 nice work on the percona bits! 16:06:15 sorry pretty busy week :( 16:06:24 spotz: yay! did it get a chance to get reviewed? 16:06:38 andymccr: thank you :) 16:06:49 evrardjp: no worries, you're keeping up with some good reviews so ++ :) 16:06:52 mnaser: Not sure let me go find it:) 16:07:06 i got the Elastic perf monitor tool working with horizon/django, anyone who fancies appkying that to the back end api please go for it.... 16:07:30 Question folks, i am configuring F5 for OSAD so do i need to configure them to use SSL or non-SSL? 16:07:34 jrosser: i saw those patches, pretty sweet, wish i had time to hack on that :p 16:08:05 spotz: while you find the patch, we'll dive into bug triage and we can get back to it in open discussion :) 16:08:17 #topic bug triage 16:08:20 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777635 16:08:21 Launchpad bug 1777635 in openstack-ansible "lxc-veth-wiring fails to find interface" [Undecided,New] 16:08:41 "Unknown interface eth1" 16:09:24 looks like the submitter added more details 16:10:09 it looks like we're using "source-directory" and somehow that's not long supported, we need to use "source" 16:10:10 cloudnull: or jamesdenton ^^ 16:10:25 * mnaser looks at lxc_container_create quickly 16:10:28 Any idea, should i set SSL or non SSL? 16:10:35 Amy Marrich (spotz) proposed openstack/openstack-ansible master: Remove IP range 1.2.3.4 from doc examples https://review.openstack.org/574930 16:10:53 :) 16:10:59 i wonder if the images recently changed? 16:11:25 tux_: we are in a meeting, could we discuss this later? 16:11:52 mnaser: do we see the version? 16:11:55 is that master? 16:12:00 "This happens with Openstack-Ansible 16.0.13 and 17.0.4." 16:12:24 evrardjp: sure, i can wait.. 16:12:36 master uses systemd_network i think with the work cloudnull did 16:13:19 yeah 16:13:28 i'm trying to see if we manage the /etc/network/interfaces file 16:13:51 I know we used to but haven't looked at that bit in ages 16:14:06 i'm trying to dig quickly via github 16:14:10 https://github.com/openstack/openstack-ansible-lxc_container_create/tree/stable/queens 16:14:50 https://github.com/openstack/openstack-ansible-lxc_container_create/blob/stable/queens/tasks/lxc_container_config.yml#L119-L127 16:15:18 looks like we add things to interfaces.d 16:15:22 but we don't touch interfaces 16:15:53 we need more info, what's the best status to set it to? 16:16:08 I am not convinced we have a bug yet, but I am not sure I have a full understanding of the bug. 16:16:13 incomplete 16:16:18 I agree 16:16:26 with you, on the classification 16:16:41 cool 16:16:42 next 16:16:47 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777465 16:16:48 Launchpad bug 1777465 in openstack-ansible "Prepare the deployment host in OpenStack-Ansible" [Undecided,New] 16:17:14 hmm 16:17:19 pointing to a pip_install that doesn't exist? 16:17:27 network issue? 16:17:42 I tried 17.0.5 a few days ago I didn't get that. 16:17:49 or was it .4? 16:17:53 I can recheck if you like 16:17:58 .5 16:18:02 it seems like it's trying to checkout 407931dcfe5bd15153ddafc0d96467a7ae5b6524 16:18:05 which isnt a real commit 16:18:14 yeah but 16:18:16 why 16:18:18 :p 16:18:45 maybe we tagged the wrong thing 16:18:51 let me verify 16:19:05 https://github.com/openstack/openstack-ansible-pip_install/commit/407931dcfe5bd15153ddafc0d96467a7ae5b6524 is a commit 16:19:08 it's fine 16:19:10 and I tried it 16:19:19 and it's gate tested :p 16:19:21 ok cool 16:19:25 invalid? 16:19:30 yup 16:20:01 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777462 16:20:02 Launchpad bug 1777462 in openstack-ansible "tempest plugins should not be targetting master in stable branches" [Undecided,New] 16:20:07 * mnaser sighs 16:20:08 this one 16:20:22 i'm thinking we convince tempest to release a new version 16:20:29 and we drop the whole master stuff 16:20:30 thoughts? 16:21:00 i think we've had that chat 16:21:28 andymccr: what was the final verdict? 16:21:44 tl;dr - tempest wouldnt version because the tests should work against all versions 16:21:47 or something along those lines 16:21:54 perhaps that stance has changed 16:22:02 well 16:22:10 it doesnt work on all versions >:( 16:22:11 but yeah 16:22:14 i hear you about that 16:22:20 yeah i agree with that 16:22:26 i will assign to me and try to get a release out from tempest 16:22:28 and we can unpin that stuff 16:22:54 i'll try to convince them to push out 18.0.1 or something 16:23:08 sounds good 16:23:08 assigned to me 16:23:15 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777456 16:23:16 Launchpad bug 1777456 in openstack-ansible "Not possible to get all logs into syslog" [Undecided,New] 16:23:16 i imagine if there was a reason i dont see why they would block it 16:23:25 andymccr: yep esp when they broke their own rules, teehee. 16:23:41 hmm 16:24:19 that bug seems invalid to me because you can use rsyslog_client_user_defined_targets in your user_variables 16:24:35 assuming that's what the user means? 16:26:10 i'm not sure, i guess i'll put it down as incomplete and ask if `rsyslog_client_user_defined_targets` solves their issues [y/n]? 16:27:09 incomplete and added comment 16:27:22 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777403 16:27:23 Launchpad bug 1777403 in openstack-ansible "galera_server: configuration file name and path" [Undecided,New] 16:28:01 i think that is not an issue, because we use our own filesystem/config files rather than depending on upstream 16:28:22 hmm 16:28:43 cause we drop our own https://github.com/openstack/openstack-ansible-galera_server/blob/master/templates/my.cnf.j2 16:28:56 yeah we could ask if its actually causing an issue? 16:29:06 like to say "hey the file isnt name the same" isnt really an issue if everything runs correctly 16:29:10 German Eichberger proposed openstack/openstack-ansible-os_octavia master: [WIP] Adds the issuer to the CAs https://review.openstack.org/576234 16:29:23 andymccr: yep, let me ask 16:29:55 andymccr: added comment, will set to 'incomplete' for now 16:30:03 sounds good 16:30:06 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777303 16:30:07 Launchpad bug 1777303 in openstack-ansible "rndc command not present on Centos 7" [Undecided,New] 16:30:53 looks like jrosser pushed up a fix 16:31:02 oh sweet 16:31:03 for that one I think the deployer has to deal with the installation of the backend itself 16:31:11 but that's good news 16:31:22 but centos doesn't seem to ship rndc in bind-utils 16:31:24 mnaser: link to that fix? 16:31:32 but in bind instead 16:31:35 https://review.openstack.org/#/c/570353/7 16:31:46 but looks like in redhat-7.yaml we should s/bind-utils/bind/ 16:32:11 pretty low hanging fruit, anyone wanna volunteer? :) 16:32:21 come on ppl :) 16:32:24 that's an easy fix! 16:32:40 i thought i handled all the distros? maybe forgot one? 16:32:41 go team go! 16:32:43 you don't want mnaser to beat the commit count openstack world record, right? 16:32:55 jrosser: you did actually but apparently centos wants bind instead of bind-utils 16:32:57 jrosser: different package needed :) 16:33:20 grrr :) 16:33:32 jrosser: if you're busy, i can push it up quickly 16:33:52 sure im on mobile right now 16:34:36 ok ill look into it 16:34:46 confirmed 16:34:50 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777254 16:34:51 Launchpad bug 1777254 in openstack-ansible "using ceph backend with lvm strategy leads to error" [Undecided,New] 16:35:18 haha 16:35:20 ok 16:35:22 so... 16:35:38 i saw a patch about this today 16:35:40 andymccr: has a patch, and I know the fixes for the failures of his patch. 16:35:50 did I push it? 16:35:59 I can't remember it, I was in a series of meetings. 16:36:13 anyway 16:36:14 so yeah 16:36:15 I will fix that. 16:36:19 that bug will only be fixed on master of ceph-ansible 16:36:24 assign that to me. 16:36:36 for backports, that's gonna be another thing. 16:36:37 Mohammed Naser proposed openstack/openstack-ansible-os_designate master: Use correct package for CentOS rndc https://review.openstack.org/576579 16:36:42 evrardjp: cool, done 16:36:48 let me know if you need help, ceph is important for me :-) 16:36:56 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777072 16:36:57 Launchpad bug 1777072 in openstack-ansible "In inventory/group_vars/repo_all.yml, internal_lb_vip_address needs to use ipwrap to handle IPv6" [Undecided,New] 16:37:04 nah I figured it out, just need work. 16:37:09 coolios 16:37:19 its not really a ceph issue anyway ;P 16:37:28 ok so 16:37:41 for that one, I think I suggested that patch. 16:37:48 to blinkniz 16:37:57 blinkiz 16:38:11 however, I think there is another patch of nicolasbock that should fix it in another way 16:38:19 so there are open fixes for that 16:38:21 so that might become invalid. 16:38:24 yeah 16:38:49 i see https://review.openstack.org/#/c/574893/ 16:39:06 yeah the patch of nicolasbock :) 16:39:08 looks like it's in a merge conflict 16:39:17 The gate is having issues though 16:39:36 Yes, I'll fix that, but before then the gate checks kept running into seemingly unrelated issues 16:39:46 those should mostly be addressed 16:39:53 I'll rebase 16:40:04 i threw a recheck your way, any ci issues you have => reach out please :) 16:40:15 Ok, thanks! 16:40:23 well rebase is needed anyway :p 16:40:33 so for the bug 16:40:36 duplicate? 16:40:43 duplicate would work. 16:40:46 of 1775621 16:40:51 ofc 16:41:05 done 16:41:15 #link https://bugs.launchpad.net/openstack-ansible/+bug/1777066 16:41:16 Launchpad bug 1777066 in openstack-ansible "os_cinder: Adding 'lvm' volume type fails: ERROR: No volumetype with a name or ID of 'lvm' exists" [Undecided,New] 16:41:25 yeah, i have no idea. 16:41:29 i tried to talk to the cinder team 16:41:31 they have no idea. 16:41:59 hwoarang: are you there? 16:42:06 it's blocking os_cinder right now 16:42:07 and os_magnum 16:42:45 probably worth checking in their documentation, to see if there are things that have changed, like deploy process change 16:42:55 i spoke with the ptl 16:42:57 nothing :( 16:43:02 is that exclusively broken for suse or for everything? 16:43:04 so i have no idea. i'll do some more digging 16:43:06 everything andymccr 16:43:23 https://review.openstack.org/#/c/575297/ 16:43:26 all functional broken there 16:43:30 do the cinder role tests break too? 16:43:45 i think it's something that changed in cinder though. beacuse distro deploys still work 16:43:56 i'll have to work a bit more with their team 16:44:01 yeah that's what I meant 16:44:12 what do you mean nothing mnaser ? 16:44:22 according to them, no commits that should have affected that behaviour 16:44:37 interestingly the upgrade works 16:44:41 at the expense of time, ill assign to me + confirmed and dig deeper, we got a few more to go through 16:44:47 type-set is where it breaks 16:44:57 yeah, maybe that moved outside cinder cli? 16:45:00 to openstack cli? 16:45:09 maybe we should use the openstack cli too 16:45:13 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776944 16:45:14 Launchpad bug 1776944 in openstack-ansible "horizon translations update playbook misses openstack_auth" [Undecided,New] 16:45:15 or this can be used in an ansiblemodule? 16:45:29 evrardjp: that's an option too, probably the cleanest 16:45:49 that seems pretty easy fix 16:45:57 the last bug seems important to me 16:46:07 we have a job that's critical for us to make sure translations work 16:46:19 if there is anything wrong, we should fix it. 16:46:35 evrardjp: ok cool. it seems to be pretty straightforward fix of adding an extra item there 16:46:38 andymccr: maybe you have an insight on this though? 16:47:06 mnaser: if that's adding a folder into that list, yes :) 16:47:18 but I am not sure about why it happened or it was missed 16:47:24 openstack_auth is a new thing afaik in queens 16:47:31 we run the script provided by horizon repo against the directories to update 16:47:39 im not sure if a new one is needed, but its an easy fix if that is the case 16:48:13 anyone volunteers to push up? 16:48:29 andymccr: ? :p 16:48:37 I am fixing your thing, you can fix mine :D 16:49:00 (assuming this is "mine") 16:49:09 (and the other was "yours") 16:49:14 ok assign that to me mnaser 16:49:16 if eumel8 was around we could confirm 16:49:18 I will fix that right now 16:49:29 oh i mean i can push a pr up in a sec if you want me to 16:50:05 i found andymccr easier than evrardjp so assigned to andymccr i'll follow the patch in ci :p 16:50:09 hahaha 16:50:11 ok cool 16:50:14 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776902 16:50:15 Launchpad bug 1776902 in openstack-ansible "galera_server: hardcoded database service name" [Undecided,New] 16:50:15 hehe 16:50:35 hm 16:50:48 i mean.. we gate all this stuff so it works across suse/redhat/ubuntu 16:50:51 i'm inclined this is invalid 16:51:42 we always pull from upstream by default 16:51:51 this could be behaviour when not pulling from upstream 16:52:28 Andy McCrae proposed openstack/openstack-ansible-os_horizon master: Add openstack_auth to translations update https://review.openstack.org/576584 16:52:36 lol andymccr 16:52:38 Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: [WIP] Add the openstack_auth folder to translations https://review.openstack.org/576585 16:52:54 ok, marked that one as incomplete. added comment asking if they're using distro pkgs 16:52:58 tbh, it makes sense i checked the directory tree from horizon 16:53:04 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776818 16:53:05 Launchpad bug 1776818 in openstack-ansible "Unknown lvalue 'PrivateUsers' in section 'Service'" [Undecided,New] 16:53:10 I had an ssh key issue so that was uploaded after a few reconnections, what's your excuse for a patch happening so late? 16:53:27 evrardjp: i checked whether its needed and didnt do a WIP :P 16:53:43 mine is PRODUCTION READY 16:53:43 cloudnull isn't around but maybe when he is, he can comment on that one. 16:53:52 yeah that is a weird one 16:54:03 only thing i'll ask is 16:54:09 if it's causing warnings or failing 16:54:11 yeah 16:54:16 what is the downside 16:54:22 mnaser: worth asking indeed 16:54:27 incomplete, asked 16:54:31 without htat even if we did push a PR im not sure how we'd know if its better or fixed 16:54:33 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776799 16:54:34 Launchpad bug 1776799 in openstack-ansible "Instance can ping gw and google, but Fails to Reach Metadata server with ODL+OVS" [Undecided,New] 16:54:59 * mnaser doesn't know anything about ODL 16:55:31 hahaha 16:55:37 judging by the silence neither do we :( 16:55:39 welcome tot he club mnaser 16:55:59 you could assign this to manuel buil? 16:56:14 mbuil? 16:56:17 yeah 16:56:25 hopefully he can help him 16:56:29 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776606 16:56:29 Launchpad bug 1776606 in openstack-ansible "qemu-kvm missing on new compute hosts" [Undecided,New] 16:56:53 ocata bug gr 16:57:03 if qemu-kvm is not installed 16:57:07 no vm would ever get booted.. 16:57:50 i'll ask for compute node deploy logs if possible 16:57:54 it'll help us troubleshoot 16:58:12 Tahvok: are you there? 16:58:20 that's your bug :) 16:58:56 incomplete, asking for logs 16:59:02 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775873 16:59:03 Launchpad bug 1775873 in openstack-ansible "gnocchi missing configuration to use ceph backend" [Undecided,New] 16:59:24 eh, it probably needs ceph_client to run on it 16:59:46 i dont have bandwidth to look into that right now, but it's very much confirmed, i've seen it before 16:59:54 yeah id believe that 16:59:59 didn't we see that last week? 17:00:06 i feel like we did 17:00:06 lol 17:00:11 haha 17:00:29 sorry I haven't checked last week logs 17:00:40 anyways, set to confirmed, hopefully someone can pick it up.. asked them if they would like to help us fix it 17:00:59 we're over time but 17:01:02 #topic open discussion 17:01:06 spotz: pushed up a fix for docs? 17:01:20 yeah 17:01:33 https://review.openstack.org/#/c/574930/2 17:01:44 it LGTM and passes ci 17:02:20 i'll also ask to keep eyes out on https://review.openstack.org/#/q/starredby:mnaser@vexxhost.com .. i'm always pushing stuff up and they're small and easy and ready to merge most of the time 17:02:23 anyone has anything else? :) 17:03:02 done 17:03:19 I can't chair next meeting :( 17:03:21 again 17:03:22 ETIMEOUT 17:03:24 #topic next chair 17:03:35 i can do it unless someone really feels like they want to ;) 17:04:05 i'll chair the next one 17:04:35 thanks for coming evrardjp andymccr spotz nicolasbock jrosser ansmith mgariepy 17:04:43 #endmeeting