16:01:10 #startmeeting openstack_ansible_meeting 16:01:11 Meeting started Tue Jun 12 16:01:10 2018 UTC and is due to finish in 60 minutes. The chair is hwoarang. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:12 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:15 The meeting name has been set to 'openstack_ansible_meeting' 16:01:24 #topic Rollcall and last week highlights 16:01:30 o/ 16:01:30 Tahvok: :) 16:01:33 o/ 16:01:33 o/ 16:01:33 andymccr that's where you're wrong, I can not not show up if I want to :) 16:01:35 o/ 16:01:44 o/ 16:01:46 cloudnull: thats cos you're special 16:01:48 o/ 16:01:56 andymccr you know it :P 16:01:58 andymccr: :) 16:02:25 * tosky pretends to be around 16:03:11 ok so since last week no major things have happened except mnaser has started working on the epel cleanup! 16:03:34 Just a question, what's the reasoning of the epel cleanup? 16:03:36 Andy Smith proposed openstack/openstack-ansible-os_trove master: Update to use oslo.messaging service for RPC and Notify https://review.openstack.org/574789 16:03:49 Tahvok: package conflicts between epel and rdo repositories 16:04:08 Tahvok, epel is evil as dmsimard would say. 16:04:12 anyone, if you feel that something should have been listed on the last week highlights, please add it to the wiki. 16:04:19 So potentially users should be warned to not install epel by themselves on the hosts? 16:04:48 yeah wiki is at https://wiki.openstack.org/wiki/Meetings/openstack-ansible 16:04:50 #link https://wiki.openstack.org/wiki/Meetings/openstack-ansible 16:05:09 anything else worth mentioning since last week? 16:05:10 mnaser stated breaking all the things with an epel cleanup 16:05:22 but i'll be sure to do all the clean up needed 16:05:26 ok, i've been working on leap 15 enablement and usual distro stuff 16:05:28 and make sure that things get fixed 16:05:35 mgariepy: EVIL 16:05:41 thank you mnaser 16:05:49 thanks hwoarang on the leap 15 16:05:54 I've found a potential bug... qemu-kvm package is not installed on compute hosts.. But someone else to confirm it.. 16:05:57 and the use of distro packages 16:06:08 Tahvok: if you have a bug, you can push it up and we can check it out in the triage section of our meeting 16:06:09 Tahvok: did you file a bug? We're gonna go for bug triage now 16:06:09 o/ 16:06:19 Tahvok, epel packages change all the time and shouldn't be installed along with RDO 16:06:59 I'm still checking it.. But what I found is that this package is not installed on latest compute nodes we've added. All older (4+ months) nodes have this installed 16:07:28 And I have not found any mention of it in osa, so I have no idea how it got into the older nodes in the first place.. 16:07:39 Not having this package breaks live-migration 16:07:49 Tahvok: please file a bug so we can move on with the meeting. or wait and we can discuss after it 16:07:59 will fill a bug 16:08:12 thank you. moving on 16:08:27 #topic bugtriage 16:08:30 the fun part 16:08:57 lol, netsplit 16:08:58 #link https://etherpad.openstack.org/p/osa-bugtriage 16:09:08 oh noes 16:09:36 #link https://bugs.launchpad.net/openstack-ansible/+bug/1776212 16:09:37 Launchpad bug 1776212 in openstack-ansible "ceph_client: cephkeys_access_group group creation" [Undecided,New] 16:09:50 who speaks ceph here? 16:09:55 * mnaser does 16:09:56 * hwoarang is looking at andymccr 16:10:02 that makes sense 16:10:05 logan- and andymccr too 16:10:31 https://github.com/openstack/openstack-ansible-ceph_client/blob/master/tasks/ceph_auth.yml#L34-L38 16:10:34 https://github.com/openstack/openstack-ansible-ceph_client/blob/master/tasks/main.yml#L75 16:10:35 that does seem like a ceph-config thing 16:10:41 sorry doorbell went! 16:10:44 L75 is ceph config 16:10:47 andymccr: hahahahaha 16:10:49 ok confirmed i suppose. anyone willing to work on it? 16:10:50 \o/ - but ping if specifically needed! 16:11:02 hwoarang: I don't think it's confirmed 16:11:06 And no not volunteering hwoarang:) 16:11:07 i'll take a look at that one 16:11:33 i have no powerzz to assign people on it so andymccr feel free to grab it 16:11:38 haha ok done 16:11:44 awesome :) 16:11:45 hwoarang: are you not on the bug triage group or w/e 16:11:47 I think https://github.com/openstack/openstack-ansible-ceph_client/blob/master/tasks/ceph_auth.yml#L38 should be removed 16:11:54 but except that see main L75 16:12:03 andymccr: yeah maybe i am missing some privs on launchpad 16:12:24 thanks for looking to fix it andymccr 16:12:29 we can deal with that later hwoarang :) 16:12:34 sure 16:12:37 so yes 16:12:41 confirmed 16:12:45 but... 16:12:49 low 16:12:51 I'd say 16:13:01 next? 16:13:03 yeah, low - it only affects you if you're using the tags 16:13:04 sure 16:13:06 easy workaround 16:13:10 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775938 16:13:11 Launchpad bug 1775938 in openstack-ansible "haproxy template file causes haproxy to fail to start" [Undecided,New] 16:13:12 odyssey4me: and we never said you should 16:13:38 hmm 16:13:53 hmm 16:13:55 that's a rough one 16:13:57 need a trace 16:14:06 * odyssey4me has no idea what that is - is that a var that's settable? 16:14:06 https://github.com/openstack/openstack-ansible-haproxy_server/blob/master/defaults/main.yml#L98-L107 16:14:09 i assume it has to do with that 16:14:09 i think this was fixed recently, but the commit message must not have made it back to launchpad 16:14:11 I suspect the number is marked as string 16:14:28 https://github.com/openstack/openstack-ansible-haproxy_server/commit/c85793ed2162e357eaf99058b2a2162ba30cb26b 16:14:34 I can have a look 16:14:36 that 16:14:39 is the fix most likely 16:14:42 assign it to me :) 16:14:50 https://review.openstack.org/#/c/573843/ 16:14:50 oh nice thinking evrardjp :) 16:14:51 do not mark it as confirmed for now I'd say 16:15:21 let's see next week 16:15:30 evrardjp: what jmccrory said about https://review.openstack.org/#/c/573843/ seems legit 16:15:45 oh yeah, bad commit msg - but looks appropriate 16:15:56 this looks like it's fixed but the commit message is bad so launchpad did not notice 16:15:57 :/ 16:16:04 probably worth porting that back too if applicable to queens- 16:16:17 evrardjp: either way feel free to grab it 16:16:46 moving on 16:16:48 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775873 16:16:49 Launchpad bug 1775873 in openstack-ansible "gnocchi missing configuration to use ceph backend" [Undecided,New] 16:17:37 Merged openstack/openstack-ansible-os_sahara master: Add release notes link to README https://review.openstack.org/574594 16:17:51 Merged openstack/openstack-ansible-os_swift master: Add release notes link to README https://review.openstack.org/574590 16:18:00 so by reading the bug i think it's valid but not sure who wants to work on gnocchi right now 16:18:41 I am not sure, the rados lib has to be symlinked 16:18:43 seems likely to be legit, and easy enough to patch - but harder to test 16:19:01 not installed then symlinked 16:19:08 or am I wrong? 16:19:13 it's just partial? 16:19:27 it could be, just partially linking the c bindings for example 16:19:31 yeah, I think gnocchi already has 'include_role: ceph-client' so it should just be a patch to ceph-client 16:19:33 i guess we can follow the pattern from os_nova aand stuff 16:19:40 yeah 16:19:53 maybe the pattern is incomplete 16:19:58 because the grep is too restrictive 16:20:01 that can happen 16:20:12 python-apt for example, wouldn't work with the current code 16:20:43 i think that's confirmed 16:20:49 ok 16:21:11 we need to make sure the binaries are installed 16:21:11 in any case, the expectations are wrong 16:21:14 ceph_client should be doing that i think 16:21:19 yeah 16:21:30 there is probably something funky going on that needs fixing 16:21:50 i can look into it but not until a week or so, this week is super hectic for me 16:21:55 seems like a straight forward fix 16:22:03 "Once python-rados installed and ceph keyring imported ((manually in my case), the deployment is achieved." is for me the selling point to a "we need to fix this" 16:22:17 yep it looks like that stuff should probably happen automatically 16:22:32 Merged openstack/openstack-ansible-openstack_hosts master: Install EPEL repositories manually and exclude Python packages https://review.openstack.org/573057 16:22:34 mnaser: if we triage this as low, and we assign that to you, you won't forget, and you fix it when you can 16:22:41 ++ 16:23:01 great :) 16:23:17 next 16:23:19 :) 16:23:20 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775621 16:23:21 Launchpad bug 1775621 in openstack-ansible "IPv6 addresses need to be wrapped in square brackets ([ ])" [Undecided,New] 16:23:26 hwoarang: you actually can't change anything on the bugs right now, can you? 16:23:29 ouch 16:23:30 this one 16:23:40 odyssey4me: i think i can confirm bugs but not assign people 16:23:40 Unless someone is dying to touch this one I could try 16:23:47 nicolasbock: woot 16:23:50 that bug seems valid 16:23:50 :) 16:23:52 hwoarang: ok, continue - lemme sort that out quickly 16:23:53 nicolasbock: :) 16:24:02 congratulations on your bug 16:24:04 next ;p 16:24:06 It said low hanging ;) 16:24:19 that's not low hanging 16:24:22 :) 16:24:30 but we can help you :) 16:24:31 AAAHHHHHH 16:25:05 I'll definitely ask for help then ;) 16:25:06 well it seems it has fixed the issue for the author 16:25:27 so it's a good start -- and that would be the low hanging fruit I guess :D 16:25:43 yeah, I reckon it'd be pretty simple change to the plugins 16:25:49 I'll see how tall I am :) 16:25:57 well it all depends if we add testing or not I'd say, but that's for the review 16:26:06 I'd say the bug is confirmed and medium 16:26:09 let me mark as confirmed 16:26:30 next 16:26:37 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775584 16:26:38 Launchpad bug 1775584 in openstack-ansible "OSA-AIO deploy failed cause haproxy didn't start" [Undecided,New] - Assigned to Ilya Popov (ilya-p) 16:27:01 hwoarang: you should be able to do things in LP now - welcome to the drivers team :p 16:27:13 yeah more powerzzz 16:27:20 mnaser: you too :p 16:27:24 over 9000 16:27:43 https://www.youtube.com/watch?v=ygBP7MtT3Ac 16:27:45 ok back to work 16:28:17 * spotz points at hwoarang and laughs, "More work for you!" 16:28:20 sounds like 16:28:23 that is a deployer issue 16:28:32 something was using port 80? 16:28:43 why did the play not stop though 16:29:31 so 1) something was using port 80 already and b) our playbooks didn't hard stop 16:29:46 David Wilde proposed openstack/openstack-ansible-os_sahara master: Add trustee section by default https://review.openstack.org/569886 16:29:53 Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_cinder master: Start cinder-backup last https://review.openstack.org/574795 16:30:32 I have seen a failure to hard restart didn't rc = 0 in the past but I will analyse this with my patch on the nbproc I guess 16:30:51 ok then 16:30:56 ++ 16:31:02 thanks 16:31:03 evrardjp mnaser hwoarang spotz - also made you administrators of the drivers group now too, so you can add/remove/change people's access as need be 16:31:18 I would really like to thank evrardjp for that bug :) 16:31:27 Wow evrardjp wasn't a member? 16:31:36 evrardjp: feel free to take it 16:31:45 nope I wasn't 16:31:51 lets move on 16:31:53 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775544 16:31:54 Launchpad bug 1775544 in openstack-ansible "1.2.3.4 is not supposed to be used as example" [Undecided,New] 16:31:54 but I never cared, it didn't prevent me to do the work :) 16:32:08 he was a member, but not an admin - apologies, I've tried to keep the cores as admins over time, but obviously haven't keep it up to date 16:32:31 well the bug seems legit and low fruit 16:32:47 agree 16:33:10 tha's valid 16:33:32 spotz: could you deal with that bug? 16:33:40 it's a docs issue 16:33:42 valid + low i guess 16:33:45 or bring someone in 16:33:53 :D 16:33:54 wow, people read the docs ;) 16:34:00 hwoarang: agreed. 16:34:03 ok next 16:34:03 evrardjp: Yeah 16:34:05 odyssey4me: yeah, surprising. 16:34:11 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775494 16:34:12 Launchpad bug 1775494 in openstack-ansible "Queens Insallation fails: github spice-html5 link broken" [Undecided,New] 16:34:17 meh that's prob already fixed 16:34:33 it has been and odyssey4me pushed up a release note too to help with work arounds 16:34:35 yeah and released 16:34:52 maybe refer the user to the release note 16:35:12 https://docs.openstack.org/releasenotes/openstack-ansible/queens.html#known-issues 16:35:20 thank you mnaser 16:35:39 are you doing it mnaser ? 16:35:44 i am 16:35:48 ok jesse did it 16:35:53 :P 16:36:05 haha, my ninja fingers beat yours :p 16:36:05 we all had the link, odyssey4me beat us to it 16:36:16 Rock always wins! 16:36:19 heheh 16:36:31 ok great moving on 16:36:43 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775315 16:36:44 Launchpad bug 1775315 in openstack-ansible "ODL : ERROR networking_odl.common.odl_features ConnectionError: HTTPConnectionPool" [Undecided,New] 16:37:00 that's for mbuil :D 16:38:01 true 16:38:03 no questions asked 16:38:17 hahah 16:38:20 next 16:38:29 mbuil: could you have a look? 16:38:44 he may not be around 16:39:14 I still assigned it to him : ) 16:39:18 yep 16:39:20 next 16:39:22 #link https://bugs.launchpad.net/openstack-ansible/+bug/1775014 16:39:23 Launchpad bug 1775014 in openstack-ansible "OpenStack Dashboard Overview page gives 504 Gateway Time-Out(The server didn't respond in time)" [Undecided,New] 16:39:23 I am not marking confirmed, so we can track 16:40:11 we get this too :( 16:40:20 depending on # of projects 16:40:35 I asked a few questions 16:40:41 please add more questions 16:40:53 and in the meantime we should mark it as incomplete, IMO 16:41:06 i agree 16:41:07 for now 16:41:55 thanks! 16:41:56 moving on 16:42:03 #link https://bugs.launchpad.net/openstack-ansible/+bug/1774999 16:42:04 Launchpad bug 1774999 in openstack-ansible "os_gnocchi: Broken DB connection" [Undecided,New] 16:42:10 well that's my bug but this seems to be fixed 16:42:21 at least i dont think we see this anymore 16:42:27 cool 16:42:29 i dont know which magical commit fix it though 16:42:30 thanks hwoarang I guess :) 16:42:33 hahaha 16:42:45 maybe we should add tests there. 16:43:48 evrardjp: yes 16:44:20 Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: common-tasks: test-set-nodepool-vars: Populate fact for EPEL mirror https://review.openstack.org/574801 16:44:35 ok last one 16:44:42 #link https://bugs.launchpad.net/openstack-ansible/+bug/1774984 16:44:43 Launchpad bug 1774984 in openstack-ansible "Run playbooks in OpenStack-Ansible" [Undecided,New] 16:45:14 Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: common-tasks: test-set-nodepool-vars: Populate fact for EPEL mirror https://review.openstack.org/574801 16:45:21 looks incomplete to me 16:45:37 and maybe a client-side issue 16:45:39 perhaps we can ask for more info 16:45:44 i dont understand the bug 16:45:49 maybe a network issue ? 16:45:55 i think haproxy failed to install to what seems like a network issue to me 16:46:22 hard to tell if it's haproxy or a genuine network thing 16:47:16 yeah, incomplete - or perhaps a bad mirror 16:47:47 yeah nothing very osa related from what i see 16:47:50 if an ansible task fails to download the package for ansible 2.4 or less, that error comes up... it's terrible 16:47:58 2.5 onwards it gives more detailed error reports 16:48:18 we should get rid of apt package on repo server, but that's completely unlinked to our conversation. 16:48:32 Markos Chandras (hwoarang) proposed openstack/openstack-ansible-lxc_hosts master: redhat: Add the EPEL repository for aria2 and python2-lxc packages https://review.openstack.org/574382 16:48:32 I tried backporting the patch to bring that back, but it was rejected because 2.4 is now in security-only fix mode 16:48:35 package reverse proxy* 16:48:49 odyssey4me: yes I am not surprised 16:49:23 Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Convert role to use a common systemd mount role https://review.openstack.org/552185 16:50:00 I'm kinda inclined to remove the cache on the repo server and rather build a user story for how you can make a mirror if you want to. 16:50:08 But that's a story for another day. 16:50:25 ok guess we are done 16:50:27 with bugzz 16:50:41 coolos hwoarang 16:50:47 next topic 16:50:48 #topic next meeting chair 16:50:57 I can't run next week 16:51:02 any volunteers or we will let the computers decide 16:51:05 oh wow 16:51:09 where is that magic script 16:51:15 no need 16:51:15 let me give that 16:51:16 i got it 16:51:21 ah great! 16:51:23 ok cool 16:51:25 woot! 16:51:29 just wait a sec 16:51:46 sec is over 16:51:48 #link https://gist.github.com/evrardjp/cc8cb51a38e72204041b5cea3411ec5f 16:51:55 for the record 16:51:57 :) 16:52:00 put it on gerrit! 16:52:01 ;p 16:52:22 haha yeah 16:52:25 Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance stable/queens: Adding systemd dependencies for NFS mounts https://review.openstack.org/574806 16:52:26 ok thank you all for attending. enjoy your evening/day 16:52:26 I wonder if we can make the openstack bot run the script with a key word. 16:52:34 omg the hardest part is the core list :p 16:52:34 hehe, if you pick me when I'm out it'll be a quiet meeting:) 16:52:46 #endmeeting