16:01:00 #startmeeting openstack_ansible_meeting 16:01:01 Meeting started Tue May 2 16:01:00 2017 UTC and is due to finish in 60 minutes. The chair is odyssey4me. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:04 The meeting name has been set to 'openstack_ansible_meeting' 16:01:08 Here is our bug list for today https://etherpad.openstack.org/p/osa-bugtriage 16:01:22 #topic Last week AP 16:01:34 andymccr triage https://bugs.launchpad.net/openstack-ansible/+bug/1682169 16:01:34 Launchpad bug 1682169 in openstack-ansible "Upgrade N->O nova-manage cell_v2 issue" [Undecided,New] - Assigned to Andy McCrae (andrew-mccrae) 16:02:28 andymccr are you around? 16:02:43 is anyone around and paying attention? :) 16:02:50 i am 16:02:56 o/ 16:02:57 :) 16:04:00 ok, it looks like andymccr hasn't yet found a working connection at his new desk 16:04:06 so I guess the action carries 16:04:19 #action andymccr triage https://bugs.launchpad.net/openstack-ansible/+bug/1682169 16:04:19 Launchpad bug 1682169 in openstack-ansible "Upgrade N->O nova-manage cell_v2 issue" [Undecided,New] - Assigned to Andy McCrae (andrew-mccrae) 16:04:25 still carrying! 16:04:36 haha :) 16:04:40 next up https://bugs.launchpad.net/openstack-ansible/+bug/1687447 16:04:41 Launchpad bug 1687447 in openstack-ansible "neutron-ha-tool running of outdated venv when upgrading intra release" [Undecided,New] 16:04:43 4 mins late :P not bad. 16:05:22 hmm 16:05:24 humm. 16:05:28 hahah 16:05:37 is this a matter of us needing to respawn the cron? 16:06:04 surely the cron should point to the correct place? 16:06:09 its executed via cron, maybe that's not getting recreated 16:07:06 https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tasks/neutron_l3_ha.yml#L19 16:07:10 it should be: https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tasks/neutron_l3_ha.yml#L22 16:07:24 so the cron calls that script so it should be pointed to the right place 16:08:21 so maybe Bjoern was seeing an old process still being executed ? 16:08:32 when cron refires post upgrade it will have the right path 16:08:45 but this does call in a good point 16:08:52 maybe its time to let that tool go 16:08:53 ? 16:09:01 it would seem sensible to me that we hould disable the cron job and kill any running process while an upgrade is being executed 16:09:13 its technically been deprecated since kilo 16:09:15 https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tasks/neutron_l3_ha.yml#L16-L17 16:09:32 but we've kept it around due to bugs in l3ha ? 16:10:14 maybe we should standardize on l3ha and leave this alone ? 16:10:46 well, it's still part of mitaka, newton and ocata 16:11:01 we could perhaps remove it from pike, but the issue still stands, assuming that it gets confirmed 16:11:05 yea i mean its there. but in pike should we drop it? 16:11:12 ok 16:11:33 yea. as for the issue at hand, i dont see how the script would not have been updated 16:11:36 could we adjust the cron somehow? 16:11:49 hmm 16:11:49 the cron calls the script and the script is a template file. 16:11:58 so it should be pointed all to the right place 16:12:02 cloudnull: but the cron runs a bash script htat may not have changed 16:12:03 yep 16:12:24 the bash script doesn't need changing as it doesn't refer to the venv 16:12:29 yeah 16:13:29 ok, so it sounds like we need more details or someone needs to try and replicate the issue 16:13:55 yeah seems so 16:14:07 so here's the tool https://github.com/openstack/openstack-ansible-os_neutron/blob/master/templates/neutron-ha-tool.py.j2#L1 16:14:16 which would have the right pathing assuming the playbook ran 16:14:35 the job is set here https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tasks/neutron_l3_ha.yml#L37 16:14:43 and then executed https://github.com/openstack/openstack-ansible-os_neutron/blob/master/templates/neutron_ha_tool.sh.j2 16:15:08 so on a minor upgrade there'd be a new venv and that venv path is based on the release version 16:15:10 yeah. hmm 16:15:28 but it would still point to the correct location once that file is retemplated which it would be because that var has changed 16:15:40 well assuming it ran successfully 16:15:42 I have this issue on upgrade from mitaka to newton ! 16:15:50 hmm 16:15:51 which is defined here https://github.com/openstack/openstack-ansible-os_neutron/blob/master/defaults/main.yml#L61 and set in the playbooks. 16:16:08 mgariepy: what did you need to do to solve it ? 16:16:21 did dig into it yet. 16:16:25 didn't dig into it yet. 16:16:40 mgariepy: do you think you'll get a chance to look into it? 16:16:51 yeah i can take a look 16:16:59 I wonder if the script is changed but the job is just running in memory ? 16:17:19 for mitaka to newton makes sense, because the default changed from this being enabled to this being disabled 16:17:30 the script isn't modified 16:17:41 hmm 16:18:01 ha.. 16:18:02 mgariepy: the shell wouldnt be but it should just call the py script which should change 16:18:05 ah, that's a good point. maybe the option is disabled? 16:18:26 if disabled, maybe it should be removed then ? 16:18:43 https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tasks/main.yml#L89-L94 16:19:31 ok so maybe the issue is its disabled now 16:19:34 and it should just be removed 16:19:53 ok, I've asked for confirmation of the value of the legacy tool enablement 16:19:53 we could run a cleanup that removes the cron job if not legacy_ha_tool_enabled | bool 16:20:05 yes, that is a good idea 16:20:35 remove the cron job and the scripts related to it 16:21:03 ok i'll make a patch for this one. 16:21:39 awesome, thanks mgariepy 16:22:12 #action mgariepy to create neutron-ha-tool cleanup if neutron_legacy_ha_tool_enabled is false 16:22:20 Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Clean up cron if l3ha legacy tool is not enabled https://review.openstack.org/461843 16:22:26 hhahahahaha 16:22:28 hahaha 16:22:33 next up https://bugs.launchpad.net/openstack-ansible/+bug/1687369 16:22:33 Launchpad bug 1687369 in openstack-ansible "ocata installation fails for aodh,ceilometer and designate when sqlalchemy version is 1.1.9 version " [Undecided,New] 16:22:35 oh... my bad. was not watching the screen 16:22:43 i think there is a patch in for this 16:22:59 alextricity25: ^ 16:23:19 sorry mgariepy ... 16:23:41 its ok cloudnull you can buy him a beer next week to say sorry :P 16:23:49 done. 16:23:52 beers all around . 16:25:01 lol 16:25:33 it looks like there's an outstanding question for this already 16:27:00 ok, let me look into this 16:27:03 yeah i know alextricity25 did some fixes on master for sqlalchemy 16:27:06 but i dont htink it applies to stable/ocata 16:27:13 yep, when I tried them they didn't work ;) 16:27:24 I'll take another look and see what I can find. 16:27:44 hmm 16:27:48 #action odyssey4me to confirm/triage https://bugs.launchpad.net/openstack-ansible/+bug/1687369 16:27:48 Launchpad bug 1687369 in openstack-ansible "ocata installation fails for aodh,ceilometer and designate when sqlalchemy version is 1.1.9 version " [Undecided,New] - Assigned to Jesse Pretorius (jesse-pretorius) 16:27:50 weird it should've worked based on constraints still 16:28:11 yep, it should have - no idea why he's trying to increase the constraint 16:28:18 anyway, next up https://bugs.launchpad.net/openstack-ansible/+bug/1687594 16:28:18 Launchpad bug 1687594 in openstack-ansible "no configuration for osa neutron to configure external dns yet" [Undecided,New] 16:28:42 pjm6 you around? 16:28:53 Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Clean up cron if l3ha legacy tool is not enabled https://review.openstack.org/461843 16:29:43 this should be pretty easy to fix up 16:29:54 interesting. i didn't know neutron configs were needed for designate. 16:30:01 wishlist? 16:30:06 it's just a conditional portion of neutron's conf 16:30:19 you could use config overrides atm? 16:30:38 i mean it would be cool to enable it if you are deploying designate by default - but right now if you wanted there is an option so yeah wishlist 16:30:39 yeah, of course - but as we ship designate we should auto-wire it if you have it 16:31:31 Hey no ping?:) 16:31:44 odyssey4me: +1 16:31:48 ok, marked low hanging fruit and wishlist 16:32:25 that looks like it for the new bugs 16:32:33 easy 16:32:36 thanks all 16:32:41 anyone got a bug to raise that needs more attention? 16:33:27 spotz_ the ping was for spotz - somehow you've gained a trailing underscore 16:33:30 reviews on https://review.openstack.org/#/c/461802/ would be appreciated :D 16:33:52 odyssey4me: Yeah when I changed bouncers. Let me see if I can fix 16:35:37 okie dokey, SHA's confirmed GTG 16:35:40 grrr 16:35:48 thanks! 16:35:50 right, thanks all! 16:35:53 #endmeeting