16:01:25 #startmeeting openstack_ansible_meeting 16:01:26 Meeting started Tue Oct 30 16:01:25 2018 UTC and is due to finish in 60 minutes. The chair is mnaser. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:27 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:29 The meeting name has been set to 'openstack_ansible_meeting' 16:01:35 #topic rollcall 16:01:36 o/ 16:01:42 o/ 16:01:48 o/ 16:02:28 o/ 16:02:39 jp mentioned he's out sick this week :( 16:02:41 \o 16:02:52 so everyone hopefully sending gooooooood things his way 16:03:21 #topic last week highlights 16:03:34 i dont see any new highlights added, i believe jrosser has mostly resolved these issues by dropping apt-cacher-ng ? 16:03:43 o/ 16:03:49 hmm yes thats a couple of weeks oldnow 16:03:52 (fwiw i have done 2 rocky deployments on bionic and they're running flawlessly :)) 16:04:05 doing a few more before the summit and it's rock-y solid 16:04:09 so thanks for your work jrosser 16:04:26 :) nice to see it all working 16:04:40 i dont think there's much other highlights afaik 16:04:50 just things humming, i've been pushing more and more fixes of things i see in prod 16:04:59 we'll probably have to sha bump cinder and nova soon 16:05:02 we caught a few critical rocky bugs 16:05:18 (in the projects themselves, not OSA, we're obviously writing flawless code, ha) 16:05:25 i started some work on massive reduction in packages installed arbitrarily everywhere 16:05:37 that'll be tremendous in speeding things up 16:05:52 we should schedule a sprint maybe together 16:05:58 to hack on making roles test using integrated repo 16:06:07 o/ attending two meetings at once, so only half here 16:06:12 odyssey4me: might be interested too :) so we can kill openstack-ansible-tests 16:06:16 but that's probably for later discussions 16:07:00 onto bugzz 16:07:03 #topic bug triage 16:07:08 #link https://bugs.launchpad.net/openstack-ansible/+bug/1800655 16:07:08 Launchpad bug 1800655 in openstack-ansible "Utility container cannot reach repo in play: name: Get list of repo packages" [Undecided,New] 16:07:13 yeah, I think a sprint for distro installs would be good to consider too - given only opensuse is really working right now 16:08:07 odyssey4me: agreed. we should plan something over the ML 16:08:15 also would help improve our upgrade jobs too 16:08:33 ok that bug it looks like the user said it was all fixed when they did iptables -F 16:08:42 and getting icmp "unreachable - admin prohibited, length 68" 16:08:54 we don't really put in any iptables rules eh? 16:09:03 it must have been created by the user 16:11:12 ..anyone 16:11:14 :) 16:11:32 i guess ill go for invalid 16:12:01 theres some discussion of this earlier wasnt there - firewalld related? 16:12:07 up there ^^^^ somewhere 16:12:22 buy yes it looks invalid 16:12:44 yep thats what i did and added a comment 16:12:50 #link https://bugs.launchpad.net/openstack-ansible/+bug/1800425 16:12:50 Launchpad bug 1800425 in openstack-ansible "[os_keystone] Error when installing keystone with LDAP" [Undecided,New] 16:13:39 mnaser there was discussion earlier - it seems like there may be a package installing/depending on firewalld - so ideally we need to identify that and figure out how to surgcally fix it 16:13:49 i believe that is fixed by I0d7e4804bdc92ffe7a679060a686e684c01fcd1b 16:13:51 odyssey4me: oh gross 16:14:05 i've been deploying on bionic here with no problems and our gate works? so im not sure 16:15:44 yeah, this is centos specific - the firewalld thing I mean 16:16:03 the ldap thing, I think there was another follow on patch after I0d7e4804bdc92ffe7a679060a686e684c01fcd1b which was needed 16:16:03 firewalld comes on the host by default, i havent seen it get pulled into the containers 16:16:20 odyssey4me: the user commented that https://github.com/openstack/openstack-ansible/blob/stable/rocky/playbooks/os-keystone-install.yml#L117 fixed for them 16:16:23 yeah, it's not the container - it's a block on the host 16:16:31 which was backported on here https://github.com/openstack/openstack-ansible/commit/d350df49f30f38d180422af7b7b4559ac6b31b30 16:16:35 so just not relesed yet? 16:16:49 yeah, part of what we do when we prep centos hosts is get rid of firewalld and networkmanager 16:16:59 should probably document that 16:16:59 oh yes, https://review.openstack.org/#/q/I0d7e4804bdc92ffe7a679060a686e684c01fcd1b was the follow on patch :p 16:17:20 i set that to fix released and added the reference 16:17:21 mnaser yeah, we should either document it or automate it 16:17:27 errr fix committed 16:17:41 i think document is solution, automate feels dirty in case someone actually depends on firewalld 16:17:53 then surrpise your firewall rules are gone 16:18:13 ill update that first bug 16:18:51 Agreed. 16:19:03 And for the latest bug mentioned, 17.1.2 includes the fix - I commented in the bug. 16:19:35 ah cool, my bad 16:19:44 great, i updated the first one, tagged as low-hanging-fruit too, easy simple doc change 16:19:52 #link https://bugs.launchpad.net/openstack-ansible/+bug/1800185 16:19:52 Launchpad bug 1800185 in openstack-ansible "Upgrades from hmm 16:20:20 this is a good one 16:20:28 so many envs i run into with env.d/.yml 16:21:07 it seems like a bit of a hard problem to solve because if we just verify checksums, the yaml content might be the same 16:21:11 Yep, I don't know the best way to check for exact copies and to remove those - but hopefully someone can pick that up and figure it out. 16:21:12 so it might involve writing some sort of utility 16:21:39 it's plausible that rsync can do it, or perhaps diff 16:22:17 yeah but any rearranging that we'd do of the file would make it look different on disk 16:22:21 even tho it evaluates to the same thing 16:22:31 that's just nitpicking at the problem though, i agree it does exist 16:22:46 i dunno how critical we can consider it though, i'm inclined to put under wishlist.. 16:23:31 yep, it's more of a wishlist thing 16:23:46 ok cool 16:23:51 #link https://bugs.launchpad.net/openstack-ansible/+bug/1800169 16:23:51 Launchpad bug 1800169 in openstack-ansible "ansible-hardening role runs out of memory when running for ~60 nodes" [Undecided,New] 16:25:47 odyssey4me: i know you might be busy but i feel like you're experienced at the include/include_tasks/import_tasks thing 16:26:56 oo 16:26:58 "I realized huge memory increases when the tasks was doing include_tasks. I checked the imports with static import_tasks and the memory usage remained 1.5GB." 16:27:24 yep, interesting huh - and given rocky onwards has switched to using include_tasks I think we have a pretty big problem looming 16:27:54 i'm going to propose to leave it open for now 16:27:58 I'll try and find some time to work on this - this is one of those things that's going to bite us. 16:28:01 to discuss next week with more folks hopefully 16:28:13 but yeah this might be a nasty surprise 16:28:30 #link https://bugs.launchpad.net/openstack-ansible/+bug/1800100 16:28:30 Launchpad bug 1800100 in openstack-ansible "Aodh has wrong admin and internal endpoint in group vars" [Undecided,New] - Assigned to KimMinsik (for-beatitudo) 16:28:45 i think that got fixed? 16:28:48 include_tasks also doesn't let the tags work as operators might be used to, so we really need to get the right thing done in the roles 16:28:49 yay merged 16:29:39 #link https://bugs.launchpad.net/openstack-ansible/+bug/1799910 16:29:39 Launchpad bug 1799910 in openstack-ansible "glance_cors_allowed_origin should include protocol " [Undecided,New] 16:30:13 aw man 16:30:15 with a patch too 16:30:34 i'm going to ask the patch submitted if they're interested in working to push it to gerrit 16:31:17 done, will leave it open till next week in case 16:31:26 #link https://bugs.launchpad.net/openstack-ansible/+bug/1799514 16:31:26 Launchpad bug 1799514 in openstack-ansible "Test environment example in openstack-ansible" [Undecided,New] 16:31:26 * hwoarang was on an intenal meeting. sorry for being late 16:32:27 yup 16:32:30 that is indeed a mistake 16:32:56 ill keep it open and ask if the bug reporter would be interested in getting some help pushing up the patch 16:33:43 done 16:33:49 * mnaser is trying to foster helping people push patches 16:34:03 #topic open discussion 16:34:20 well 16:34:25 https://etherpad.openstack.org/p/OSA-berlin-planning 16:34:28 that doesnt seem to load for me 16:34:32 okay it finally is now 16:35:29 some cool sessions with OSA 16:35:39 let's try to mobilize and split/concur :) 16:36:20 chandankumar, arxcruz: any progress from IBM for merging our tempest testing? :p 16:36:25 any reviews you want eyes on, etc? 16:36:34 mnaser: LOL 16:36:40 it's a work in progress 16:36:50 mnaser: so, tempestconf initial patch is almost done 16:36:54 cool, look forward for power9 support patch too :p 16:37:02 some tests already pass using tempestconf instead of tempest 16:37:11 w00t 16:37:13 that sounds awesome 16:37:17 https://review.openstack.org/#/c/612968/ 16:37:34 thanks for your work 16:37:44 now i'm checking why tempest.conf isn't being collected 16:38:02 so you get 1-2 hour attempts 16:38:03 fun 16:38:11 and checking with other jobs if tempest.conf created by tempestconf is the same as the old tempest.conf 16:38:12 might be an openstack-ansible-tests change most likely 16:38:41 mnaser: after that, we will add the options to remove and add specific config parameters to tempestconf 16:38:48 and of course, rpm installation 16:39:37 cool 16:39:50 and of course, i'll be in OSA berlin meeting :) 16:40:20 awesome! 16:40:28 yes quite a few of us will be there 16:40:42 well, I have to, I live here :P 16:41:19 ahaha 16:41:23 didnt know but now i do :p 16:42:01 cool, anything else for open discussion?! 16:42:06 o/ 16:42:28 so, that's all I have for now, I hope to have tempestconf up and running till next meeting 16:42:28 I wanna ask reviews for https://review.openstack.org/#/c/613752/ and https://review.openstack.org/#/c/604567/ for translated Ansible documentation publication! 16:43:28 the second patch seems to be merged 16:43:55 oh sorry wrong link 16:43:56 arxcruz: You're in charge of finding a place for us to have our outing!:) 16:43:59 oh, there is another point to discuss - placement API 16:44:01 For https://review.openstack.org/#/c/604567/ , seems that eumel8 is so busy with preparing Berlin Summit.. Would some Ansible team members help make templatization for For https://review.openstack.org/#/c/604567/ , seems that eumel8 is so busy with preparing Berlin Summit.. Would some Ansible team members help make templatization for https://review.openstack.org/#/c/612510/ patch? 16:44:27 sorry https://review.openstack.org/#/c/612510/ is correct :) 16:44:45 spotz: ack! 16:44:47 ianychoi: for the second one i'd be inclined to agree with jp's comment 16:44:56 o/ 16:45:18 ianychoi yeah I'll revise the patch after the meeting 16:45:38 odyssey4me, thanks a lot! :) I agree with mnaser :) 16:45:43 the first one is an "honestly, i'll +W because it looks ok but i'm not too sure" 16:45:47 I just want you let you know that there is still an issue with open-iscsi: 16:45:51 odyssey4me: yes, i wanted to bring that up, gah good call. 16:45:53 #link https://bugs.launchpad.net/openstack-ansible/+bug/1704161 16:45:54 Launchpad bug 1704161 in openstack-ansible "open-iscsi post-installation script fails in cinder-api-container" [High,New] 16:46:06 dcdamien: let's get to that in a second :) 16:46:08 mnaser yeah, I chose to just trust that ianychoi knows what he wants and tested it - 'cos argh 16:46:16 i wanted to draft up a response to matt regarding OSA 16:46:25 so #1 i think we need to get a placement role written first 16:46:55 mnaser yes, definitely - if someone can pick that up I'd appreciate it... I'm short on time 16:46:56 Thanks a lot for kind help on two reviews! For the first one, I tested it locally and hopefully it solves all I18n publication issues 16:46:57 which i believe shouldn't be all that difficult 16:47:47 we'll then need to get it used across our role tests 16:47:48 honestly, i'm so short on time these days too. i could hack on it this weekend but i cant guarantee much. 16:47:56 and then figure out upgrades. 16:48:27 which quite honestly is the most complicated bit 16:48:56 because i dont think we even have much of functional upgrade jobs to start with :x 16:49:38 in the integrated build we have a good working upgrade test, although not enough eyes are on them and actively doing fixes for where they break. 16:49:50 oh but "it works"? 16:49:57 i thought it was pretty much broken 16:50:09 thats pretty good to hear 16:50:29 Given our job changes have settled down a bit, I think I should be able to rebase my patch to enable roles to use the integrated build which will make using placement easier. 16:51:09 oh you have a wip patch for that? i think we should prioritize that, the # of things biting us from openstack-ansible-tests is just too much 16:51:20 I do, yes. 16:51:47 https://review.openstack.org/551791 16:52:16 I'll probably simplify it a bit now that I've had time to percolate and try a few things. 16:52:27 okay if you can take time getting that to pass, i will try to take up sometime to build os_placement 16:52:36 i'll probably go through the whole 16:52:39 add a new project in infra thing 16:53:40 if anyone else has time, please feel free to pick it up 16:53:48 as this will be sitting on a very big list :( 16:54:34 mnaser: I might be able to work on something just finishing a peer review at work 16:55:26 spotz: ping me when you're done with that and i can try to defer 16:55:45 Merged openstack/openstack-ansible-os_cinder master: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/561218 16:56:15 Damian Cikowski (dcdamien) proposed openstack/openstack-ansible master: Add extra volume types to AIO https://review.openstack.org/561266 16:58:05 mnaser: Ok, plotting with arxcruz for outing:) https://www.hofbraeu-wirtshaus.de/en/berlin/ takes reservations:) 16:58:27 reservations can be done on the website 16:58:37 they have people speaking english as well 16:58:45 i think that wraps it up 16:58:56 #endmeeting