| *** sawblade6 has joined #openstack-ansible | 01:05 | |
| *** sawblade6 has quit IRC | 01:10 | |
| *** sawblade6 has joined #openstack-ansible | 01:32 | |
| *** sawblade6 has quit IRC | 01:45 | |
| *** sawblade6 has joined #openstack-ansible | 01:51 | |
| *** sawblade6 has quit IRC | 01:55 | |
| *** dxiri has joined #openstack-ansible | 02:46 | |
| *** sawblade6 has joined #openstack-ansible | 02:47 | |
| *** cshen has joined #openstack-ansible | 02:49 | |
| *** sawblade6 has quit IRC | 02:52 | |
| *** cshen has quit IRC | 02:53 | |
| *** sawblade6 has joined #openstack-ansible | 03:02 | |
| *** sawblade6 has quit IRC | 03:08 | |
| *** sawblade6 has joined #openstack-ansible | 03:14 | |
| *** Bhujay has joined #openstack-ansible | 03:18 | |
| *** sawblade6 has quit IRC | 03:22 | |
| *** Bhujay has quit IRC | 03:26 | |
| *** dxiri has quit IRC | 03:43 | |
| *** Bhujay has joined #openstack-ansible | 04:16 | |
| *** sawblade6 has joined #openstack-ansible | 04:19 | |
| *** sawblade6 has quit IRC | 04:24 | |
| *** dxiri has joined #openstack-ansible | 04:25 | |
| *** sawblade6 has joined #openstack-ansible | 04:54 | |
| *** sawblade6 has quit IRC | 04:58 | |
| *** dxiri has quit IRC | 05:05 | |
| *** dxiri has joined #openstack-ansible | 05:09 | |
| *** sawblade6 has joined #openstack-ansible | 06:02 | |
| *** sawblade6 has quit IRC | 06:10 | |
| *** sawblade6 has joined #openstack-ansible | 06:26 | |
| *** sawblade6 has quit IRC | 06:31 | |
| *** luksky has joined #openstack-ansible | 06:42 | |
| *** pcaruana has joined #openstack-ansible | 06:46 | |
| *** sawblade6 has joined #openstack-ansible | 06:46 | |
| *** cshen has joined #openstack-ansible | 06:50 | |
| *** sawblade6 has quit IRC | 06:50 | |
| jrosser | Morning | 07:08 |
|---|---|---|
| *** cshen has quit IRC | 07:30 | |
| *** flaviosr has quit IRC | 07:30 | |
| *** tosky has joined #openstack-ansible | 07:31 | |
| *** flaviosr has joined #openstack-ansible | 07:32 | |
| *** shardy has joined #openstack-ansible | 07:33 | |
| *** rpittau has joined #openstack-ansible | 07:41 | |
| *** sawblade6 has joined #openstack-ansible | 07:53 | |
| *** sawblade6 has quit IRC | 07:58 | |
| *** dxiri has quit IRC | 08:05 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_panko master: Add bionic to galaxy_info metadata https://review.openstack.org/593536 | 08:25 |
| evrardjp | morning | 08:26 |
| *** DanyC has joined #openstack-ansible | 08:27 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/rocky: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593538 | 08:27 |
| *** electrofelix has joined #openstack-ansible | 08:28 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/queens: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593539 | 08:32 |
| *** sawblade6 has joined #openstack-ansible | 08:33 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/queens: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593539 | 08:34 |
| noonedeadpunk | evrardjp: Hi, Please, take a look at https://review.openstack.org/#/c/593092/ when you'll have some time | 08:35 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove the last remnants of get-pip.py https://review.openstack.org/591392 | 08:41 |
| *** sawblade6 has quit IRC | 08:42 | |
| *** luksky has quit IRC | 08:44 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Ensure that tests fail early if network interfaces are down https://review.openstack.org/593546 | 08:48 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: AIO - Improve robustness of network configuration. https://review.openstack.org/590424 | 09:09 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Ensure that tests fail early if network interfaces are down https://review.openstack.org/593546 | 09:17 |
| *** luksky has joined #openstack-ansible | 09:22 | |
| *** Pradiprwt has joined #openstack-ansible | 09:27 | |
| Pradiprwt | Hi Everyone, I have installed openstack setup using openstack-ansible and created a flat provider network, the problem is if I am lunching a instance in provider network it is not getting ip... Is there any way to make it work ? | 09:30 |
| evrardjp | Pradiprwt: the multi-node AIO has an example of network creation -- it's in the ops repo, multi-node aio section, service setup. Have a look . | 09:35 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Introduce OpenStack Healthchecks https://review.openstack.org/587544 | 09:37 |
| odyssey4me | I have loads of reviews ready folks, please review if you have some time: https://goo.gl/k9cCfF | 09:45 |
| *** dxiri has joined #openstack-ansible | 09:50 | |
| odyssey4me | also a few other key reviews: https://review.openstack.org/591672 & https://review.openstack.org/571368 & https://review.openstack.org/593127 | 09:50 |
| Pradiprwt | evrardjp: I am not able to find service setup in mlti-node aio section in git,...... Floating IP is working for me, only problem comes if I lunch the vm direct in provider network | 09:54 |
| *** dxiri has quit IRC | 09:57 | |
| *** sawblade6 has joined #openstack-ansible | 10:03 | |
| *** sawblade6 has quit IRC | 10:08 | |
| *** shyamb has joined #openstack-ansible | 10:12 | |
| *** shyamb has quit IRC | 10:23 | |
| *** shardy has quit IRC | 10:31 | |
| *** shardy has joined #openstack-ansible | 10:31 | |
| jrosser | oh whoops what did we do to ceph http://logs.openstack.org/04/591304/1/check/openstack-ansible-deploy-ceph-ubuntu-xenial/9c92a86/job-output.txt.gz#_2018-08-20_09_46_59_866918 | 10:31 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Fix to setup notify mq service https://review.openstack.org/593127 | 10:40 |
| noonedeadpunk | jrosser: it seems because of https://review.openstack.org/#/c/583943/ but ansible_port seems to be a default variable.... | 10:40 |
| *** sawblade6 has joined #openstack-ansible | 10:42 | |
| noonedeadpunk | I'll place patch now then | 10:42 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_aodh master: Updated from OpenStack Ansible Tests https://review.openstack.org/593573 | 10:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_barbican master: Updated from OpenStack Ansible Tests https://review.openstack.org/593574 | 10:47 |
| *** sawblade6 has quit IRC | 10:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_ceilometer master: Updated from OpenStack Ansible Tests https://review.openstack.org/593575 | 10:47 |
| jrosser | noonedeadpunk: i guess this should really have been caught in the role test - can you see if you can extend the test coverage a bit? | 10:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_cinder master: Updated from OpenStack Ansible Tests https://review.openstack.org/593576 | 10:47 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Setting default port if ansible_port is not defined https://review.openstack.org/593577 | 10:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_cloudkitty master: Updated from OpenStack Ansible Tests https://review.openstack.org/593578 | 10:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_congress master: Updated from OpenStack Ansible Tests https://review.openstack.org/593579 | 10:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_designate master: Updated from OpenStack Ansible Tests https://review.openstack.org/593580 | 10:48 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_glance master: Updated from OpenStack Ansible Tests https://review.openstack.org/593581 | 10:48 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_heat master: Updated from OpenStack Ansible Tests https://review.openstack.org/593582 | 10:48 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_ironic master: Updated from OpenStack Ansible Tests https://review.openstack.org/593583 | 10:48 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_keystone master: Updated from OpenStack Ansible Tests https://review.openstack.org/593584 | 10:48 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_magnum master: Updated from OpenStack Ansible Tests https://review.openstack.org/593585 | 10:49 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_masakari master: Updated from OpenStack Ansible Tests https://review.openstack.org/593586 | 10:49 |
| noonedeadpunk | jrosser: I really didn't catch these such problem, as ansible_port was always defined for my environment... I've placed https://review.openstack.org/593577 as for quick fix of this | 10:49 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_neutron master: Updated from OpenStack Ansible Tests https://review.openstack.org/593587 | 10:49 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_nova master: Updated from OpenStack Ansible Tests https://review.openstack.org/592045 | 10:49 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_octavia master: Updated from OpenStack Ansible Tests https://review.openstack.org/592046 | 10:49 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_sahara master: Updated from OpenStack Ansible Tests https://review.openstack.org/593588 | 10:50 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/queens: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593539 | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_swift master: Updated from OpenStack Ansible Tests https://review.openstack.org/593589 | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_tacker master: Updated from OpenStack Ansible Tests https://review.openstack.org/593590 | 10:50 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/rocky: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593538 | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_trove master: Updated from OpenStack Ansible Tests https://review.openstack.org/593591 | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_watcher master: Updated from OpenStack Ansible Tests https://review.openstack.org/593592 | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_zun master: Updated from OpenStack Ansible Tests https://review.openstack.org/593593 | 10:51 |
| noonedeadpunk | jrosser: I'll take a look on tests now | 10:51 |
| jrosser | noonedeadpunk: perhaps related to this? https://github.com/ansible/ansible/issues/43026 <- would be worth seeing if that also affects 2.5 | 10:52 |
| evrardjp | odyssey4me: I see you voted https://review.openstack.org/#/c/593577/1 | 10:54 |
| odyssey4me | evrardjp: yep, just saw a fail related to that and was about to patch it when I saw it already done :) | 10:54 |
| evrardjp | maybe we should think about other cases, like I explained in the comment | 10:54 |
| evrardjp | what's your opinion? | 10:55 |
| evrardjp | I checked in our code, there is no place where we use ansible_ssh_port | 10:55 |
| evrardjp | so ansible_port is more likely | 10:55 |
| odyssey4me | yeah, we cleared that up some time ago | 10:55 |
| jrosser | is there a good reason for it being undefined? | 10:55 |
| odyssey4me | I think in the move to ansible 2.0 | 10:55 |
| evrardjp | I thought it was the other way around odyssey4me | 10:55 |
| evrardjp | and that was a remnant | 10:55 |
| odyssey4me | ansible_ssh_port is pre 2.0 | 10:56 |
| *** sawblade6 has joined #openstack-ansible | 10:56 | |
| noonedeadpunk | evrardjp: isn't ansible_ssh_port deprecated since 2.0? | 10:57 |
| odyssey4me | it is, yes | 10:57 |
| evrardjp | ahah ok I confused the two | 10:57 |
| evrardjp | it's still in code so... | 10:58 |
| odyssey4me | well, for now providing a default is fine - at least to unblock things, the question of why it's undefined is something worth considering once the bleeding stops | 10:58 |
| noonedeadpunk | jrosser: according to error message this bug differs a bit... | 10:58 |
| *** schwicht has joined #openstack-ansible | 10:59 | |
| evrardjp | odyssey4me: because we don't need it defined? | 10:59 |
| evrardjp | if it's not defined, it's defaulting to 22. | 11:00 |
| odyssey4me | apparentyl nothttp://logs.openstack.org/92/591392/5/check/openstack-ansible-deploy-ceph-ubuntu-xenial/15bb758/logs/ara-report/result/edfa25a0-f551-467d-b53b-572f50cb9d06/ | 11:00 |
| odyssey4me | bah | 11:00 |
| evrardjp | we are always using connection plugin to not ssh, or ssh to port 22, so no issue there | 11:00 |
| odyssey4me | http://logs.openstack.org/92/591392/5/check/openstack-ansible-deploy-ceph-ubuntu-xenial/15bb758/logs/ara-report/result/edfa25a0-f551-467d-b53b-572f50cb9d06/ | 11:00 |
| *** sawblade6 has quit IRC | 11:00 | |
| evrardjp | odyssey4me: nah that's not what I meant | 11:00 |
| evrardjp | odyssey4me: I mean for ansible to run we never needed ansible_port to be defined, until we had as task to explicitly require it | 11:01 |
| odyssey4me | so you're thinking that perhaps we should just remove its mention? | 11:01 |
| evrardjp | well I think this is error prone: | 11:03 |
| evrardjp | https://github.com/openstack/openstack-ansible-ceph_client/blob/master/tasks/ceph_get_mon_host.yml#L19-L24 | 11:03 |
| odyssey4me | ok, what's the better alternative? | 11:03 |
| odyssey4me | wait_for ? | 11:03 |
| evrardjp | I think we should instead build dynamic host list (add_hosts) based on ceph_mons var, and wait for connection | 11:03 |
| noonedeadpunk | evrardjp: I was thinking about providing ceph conf and keys in config, but had no time for this | 11:04 |
| evrardjp | or wait_for | 11:04 |
| evrardjp | or we stop messing with the ssh connection ourselves | 11:04 |
| evrardjp | and require users to fetch the configuration beforehand | 11:04 |
| evrardjp | I know some ppl that will be unhappy | 11:04 |
| odyssey4me | wait_for or wait_for_connection would be better - depending on how that connection is used | 11:05 |
| evrardjp | odyssey4me: indeed | 11:05 |
| evrardjp | and you can delegate to the appropriate node if necessary | 11:05 |
| odyssey4me | I think that the way it's used is through the ceph client to fetch the keys, so wait_for_connection is not really appropriate given that ansible doesn't access the mons. | 11:05 |
| *** shyamb has joined #openstack-ansible | 11:05 | |
| evrardjp | I am also surprised this is not run only once | 11:05 |
| odyssey4me | IT's the ceph client that accesses the mons. | 11:06 |
| odyssey4me | So I think that wait_for is perhaps the better option here, but meh - I don't really know the depths of this role. | 11:06 |
| evrardjp | odyssey4me: I guess wait_for is less reliable than ssh command | 11:06 |
| evrardjp | it's just a tcp thing | 11:06 |
| odyssey4me | yep, could also do ssh-keyscan | 11:07 |
| evrardjp | that's probably better. | 11:07 |
| evrardjp | but why is this a local action? | 11:07 |
| odyssey4me | lol, legacy | 11:08 |
| noonedeadpunk | Does wait_for_connection establishes SSH connection? | 11:08 |
| odyssey4me | noonedeadpunk: yes it does, through ansible's transport | 11:08 |
| evrardjp | odyssey4me: haha | 11:08 |
| evrardjp | https://github.com/openstack/openstack-ansible-ceph_client/commit/f30ee47ed0e423a0555888b45b14999e7da517bd | 11:08 |
| noonedeadpunk | So, I may test it and apply a patch | 11:08 |
| odyssey4me | lol, same purpose - so funny | 11:09 |
| evrardjp | odyssey4me: yeah | 11:09 |
| evrardjp | I'd rather be implicit after all for once | 11:09 |
| odyssey4me | and yes, it does seem that this should be run from the target host - rather than the deploy host | 11:09 |
| odyssey4me | given that it's the target host that has to fetch the keys, I think? | 11:10 |
| evrardjp | odyssey4me: that depends on where you put your ssh config file :) | 11:10 |
| evrardjp | odyssey4me: makes sense to me | 11:10 |
| evrardjp | so for me we should replace this with ssh-keyscan and get over with | 11:10 |
| odyssey4me | ok, I'll leave the improvements to jrosser, logan- and noonedeadpunk given that they use this stuff :) | 11:10 |
| evrardjp | let me check if ssh-keyscan takes the ssh config file by default | 11:10 |
| evrardjp | yeah, but it's good to dig deep to understand what we should be achieving | 11:11 |
| * odyssey4me runs away :p | 11:11 | |
| noonedeadpunk | Some further actions are delegated to ceph mon, like retrieving auth keys, so yes, deployment host should be able to login to at least one mon | 11:11 |
| evrardjp | odyssey4me: hahaha | 11:11 |
| odyssey4me | aha, if that is the case then wait_for_connection is better - for that to work you'd have to do it in the playbook I think? | 11:12 |
| evrardjp | noonedeadpunk: aren't these two different things? | 11:12 |
| evrardjp | odyssey4me: yeah | 11:13 |
| *** ansmith has quit IRC | 11:13 | |
| evrardjp | After reading the code, it should be done differently: ceph_mons are added into the play, you fetch things from there, and test connection | 11:13 |
| evrardjp | basically I don't consider this to be a feature of the ceph client role | 11:14 |
| evrardjp | this doesn't configure ceph_client in itself, it merely find the things needed to ceph_client to function | 11:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/rocky: Fix to setup notify mq service https://review.openstack.org/593597 | 11:14 |
| evrardjp | the same way you can pass those things as arguments :) | 11:14 |
| evrardjp | so it deservers to be a separate thing IMO. | 11:15 |
| evrardjp | but that's a larger refactor | 11:15 |
| *** vnogin has joined #openstack-ansible | 11:15 | |
| noonedeadpunk | evrardjp: yeah, you can provide all auth keys and ceph config in user_variables, but's it's not as convenient, as it's retrievement from ceph mons. And without auth keys ceph client won't be really operational | 11:17 |
| *** Bhujay has quit IRC | 11:17 | |
| noonedeadpunk | actually, there's not option to provide ceph keys in variables right now | 11:18 |
| noonedeadpunk | evrardjp: but probably I've missed smth, or get it wrong. | 11:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia master: Updated from OpenStack Ansible Tests https://review.openstack.org/592046 | 11:20 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/queens: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593539 | 11:22 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/rocky: Added possibility to specify ssh username and port for ceph mons. https://review.openstack.org/593538 | 11:23 |
| evrardjp | noonedeadpunk: please add where you cherry pick it from in the commit message:) | 11:23 |
| evrardjp | done | 11:23 |
| evrardjp | sorry | 11:23 |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client master: Setting default port if ansible_port is not defined https://review.openstack.org/593577 | 11:23 |
| evrardjp | it's just not written the way I am used to : p | 11:23 |
| noonedeadpunk | evrardjp: should I write it the master patch in the commit message as well? | 11:25 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/ocata: Ensure get-pip cannot be modified https://review.openstack.org/593411 | 11:27 |
| evrardjp | noonedeadpunk: no it's fine I could track it. But it's a combined cherry pick of two patches, so two references should be provided, not only one. | 11:27 |
| noonedeadpunk | evrardjp: ok, got it. should I abbandon cherry-picks to stable branches? | 11:28 |
| noonedeadpunk | as actually it's not enough to specify ceph_mon in your users ssh config, as ansible seems to ignore it. Or I just did smth wrong... | 11:29 |
| evrardjp | ansible reads the ssh configuration | 11:29 |
| evrardjp | maybe the name didn't match a section of your config? | 11:30 |
| evrardjp | this was the reason the ssh command was used in the first place | 11:30 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/rocky: Ensure get-pip cannot be modified https://review.openstack.org/593408 | 11:30 |
| noonedeadpunk | ok, I'll try it one more time then | 11:30 |
| evrardjp | noonedeadpunk: if needed, we can revert both patches in master branch | 11:31 |
| evrardjp | it is fine, that's just code | 11:31 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/pike: Ensure get-pip cannot be modified https://review.openstack.org/593410 | 11:31 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/queens: Ensure get-pip cannot be modified https://review.openstack.org/593409 | 11:31 |
| noonedeadpunk | evrardjp: but I think changing ceph user is still pretty usefull | 11:32 |
| evrardjp | noonedeadpunk: yeah but I think it needs a spec if we want to rewrite things properly | 11:33 |
| evrardjp | or at least you should discuss it with jrosser logan- and mnaser :) | 11:33 |
| noonedeadpunk | for the next release? | 11:33 |
| evrardjp | for master indeed | 11:33 |
| evrardjp | here I'd say I don't see any bug to fix | 11:33 |
| evrardjp | I can be convinced otherwise | 11:34 |
| evrardjp | I only see a new feature | 11:34 |
| *** sawblade6 has joined #openstack-ansible | 11:34 | |
| evrardjp | new feature -> master except if we decide together to release a major version bump | 11:34 |
| *** dave-mccowan has joined #openstack-ansible | 11:35 | |
| noonedeadpunk | evrardjp: ok, and how/when master appears in stable relase? Only with new release? | 11:36 |
| noonedeadpunk | evrardjp: btw, here's patch for horizon we've discussed https://review.openstack.org/#/c/593092/ | 11:39 |
| *** Bhujay has joined #openstack-ansible | 11:40 | |
| *** schwicht has quit IRC | 11:40 | |
| *** sawblade6 has quit IRC | 11:42 | |
| kaiokmo | morning | 11:43 |
| evrardjp | morning | 11:45 |
| openstackgerrit | Merged openstack/openstack-ansible-os_congress master: Updated from OpenStack Ansible Tests https://review.openstack.org/593579 | 11:56 |
| *** elmaciej has joined #openstack-ansible | 11:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_watcher master: Updated from OpenStack Ansible Tests https://review.openstack.org/593592 | 11:59 |
| *** shyamb has quit IRC | 11:59 | |
| *** jroll has quit IRC | 12:00 | |
| *** jroll has joined #openstack-ansible | 12:01 | |
| *** weshay|ruck is now known as weshay | 12:03 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cloudkitty master: Updated from OpenStack Ansible Tests https://review.openstack.org/593578 | 12:03 |
| openstackgerrit | Merged openstack/openstack-ansible-os_masakari master: Updated from OpenStack Ansible Tests https://review.openstack.org/593586 | 12:03 |
| *** elmaciej has quit IRC | 12:04 | |
| *** elmaciej has joined #openstack-ansible | 12:05 | |
| *** jamesdenton has joined #openstack-ansible | 12:06 | |
| *** elmaciej has quit IRC | 12:06 | |
| *** jamesden_ has joined #openstack-ansible | 12:07 | |
| *** kaiokmo has quit IRC | 12:07 | |
| *** schwicht has joined #openstack-ansible | 12:07 | |
| *** jamesdenton has quit IRC | 12:08 | |
| *** jamesden_ is now known as jamesdenton | 12:08 | |
| odyssey4me | evrardjp: with regards to the ceph_client thing - I guess if it was possible to already do this before, the issue is that it's not documented how to | 12:08 |
| openstackgerrit | Merged openstack/openstack-ansible-os_trove master: Updated from OpenStack Ansible Tests https://review.openstack.org/593591 | 12:09 |
| odyssey4me | perhaps the next step here is to actually revert the change to that one task, and to add docs for how to do it the way that works in the stable branches today | 12:09 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker master: Updated from OpenStack Ansible Tests https://review.openstack.org/593590 | 12:09 |
| odyssey4me | then the low impact 'feature' change to allow a different user can be ported back with the other adjustments | 12:10 |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat master: Updated from OpenStack Ansible Tests https://review.openstack.org/593582 | 12:12 |
| *** shyamb has joined #openstack-ansible | 12:12 | |
| *** Bhujay has quit IRC | 12:16 | |
| *** Bhujay has joined #openstack-ansible | 12:16 | |
| *** schwicht has quit IRC | 12:18 | |
| *** dxiri has joined #openstack-ansible | 12:19 | |
| *** schwicht has joined #openstack-ansible | 12:19 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: Updated from OpenStack Ansible Tests https://review.openstack.org/593584 | 12:19 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Updated from OpenStack Ansible Tests https://review.openstack.org/593575 | 12:23 |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Updated from OpenStack Ansible Tests https://review.openstack.org/593574 | 12:24 |
| *** schwicht has quit IRC | 12:26 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate master: Updated from OpenStack Ansible Tests https://review.openstack.org/593580 | 12:28 |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: Updated from OpenStack Ansible Tests https://review.openstack.org/593581 | 12:28 |
| openstackgerrit | Merged openstack/openstack-ansible-os_zun master: Updated from OpenStack Ansible Tests https://review.openstack.org/593593 | 12:31 |
| *** kaiokmo has joined #openstack-ansible | 12:32 | |
| *** mgariepy has joined #openstack-ansible | 12:33 | |
| *** shyamb has quit IRC | 12:35 | |
| *** shyamb has joined #openstack-ansible | 12:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Updated from OpenStack Ansible Tests https://review.openstack.org/593576 | 12:37 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara master: Updated from OpenStack Ansible Tests https://review.openstack.org/593588 | 12:39 |
| *** jbadiapa has joined #openstack-ansible | 12:39 | |
| mgariepy | good morning everyone | 12:40 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Bump SHAs for rocky release https://review.openstack.org/593058 | 12:41 |
| odyssey4me | howdy mgariepy | 12:41 |
| *** Pradiprwt has quit IRC | 12:43 | |
| odyssey4me | cores - I have loads of reviews ready, please review if you have some time: https://goo.gl/k9cCfF | 12:43 |
| *** elmaciej has joined #openstack-ansible | 12:43 | |
| *** shyamb has quit IRC | 12:47 | |
| *** ansmith has joined #openstack-ansible | 12:53 | |
| *** shyamb has joined #openstack-ansible | 12:53 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Updated from OpenStack Ansible Tests https://review.openstack.org/593583 | 12:58 |
| mnaser | Well speaking of that | 13:06 |
| mnaser | Who wants to see a fun surprise | 13:06 |
| *** elmaciej has quit IRC | 13:06 | |
| mnaser | Go look at ceph_client job logs | 13:06 |
| evrardjp | odyssey4me: I agree with you on the ceph client | 13:06 |
| evrardjp | mnaser: something new? | 13:07 |
| *** schwicht has joined #openstack-ansible | 13:07 | |
| mnaser | I don’t know how long it’s been this way | 13:08 |
| mnaser | But I feel it’s more fun if you discover it yourself evrardjp :p | 13:08 |
| odyssey4me | mnaser: the test does nothing - that's well known | 13:08 |
| odyssey4me | unless you're talking about something else | 13:08 |
| evrardjp | yeah it's vague | 13:08 |
| evrardjp | how can I know what you mean without a log url? :p | 13:08 |
| mnaser | Yeah it was the fact it isn’t running any tests | 13:10 |
| mnaser | With our EPEL removal we lost support for ceph deploys on CentOS | 13:10 |
| *** spatel has joined #openstack-ansible | 13:11 | |
| odyssey4me | two reasons why there are no tests - it's hard; and we've been hoping for a long time to get rid of the role | 13:11 |
| mnaser | Also: is anyone looking at our periodics? The reason I removed them was because they’ve been just failing non stop for a really long time | 13:12 |
| mnaser | If someone is actually looking at them and fixing them I’m all keeping them but it looks like they’re neglected | 13:12 |
| odyssey4me | the ones you removed have not been, because we gat eon the same jobs | 13:12 |
| odyssey4me | the ones remaining have been, and yes I look at them sometimes - but not all the time | 13:12 |
| odyssey4me | we do know that p->q upgrades are broken due to timeouts - because pike takes forever to deploy | 13:13 |
| odyssey4me | I'm hoping that using the compute kit for all tests as a base may help us overcome that | 13:13 |
| noonedeadpunk | odyssey4me: ok, than let's revert commits for ceph_client. Feature for different user must have specs? | 13:13 |
| odyssey4me | noonedeadpunk: no, the different use is nice - but that change to that one command was unnecessary because apparently it can be done another way (I have no idea how) | 13:14 |
| spatel | jamesdenton: You there? | 13:14 |
| jamesdenton | spatel sure am | 13:15 |
| mnaser | odyssey4me: A’s someone who looks at these jobs, which ones would you want to stay in at least? | 13:15 |
| spatel | Found something interesting in F5 | 13:15 |
| jamesdenton | oh? | 13:15 |
| spatel | Hey morning | 13:15 |
| spatel | Did you read my email? | 13:15 |
| spatel | if not then i can explain here | 13:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Updated from OpenStack Ansible Tests https://review.openstack.org/593589 | 13:15 |
| odyssey4me | mnaser: the bitrot jobs are good for stable branches, and upgrade jobs can't be done in gate due to the long runtime - so periodics is the only option | 13:15 |
| * jamesdenton looking... | 13:15 | |
| spatel | Cool! | 13:16 |
| odyssey4me | ideally I'd like upgrades in gate instead, but until we can cut runtime down, that's not going to happen | 13:16 |
| evrardjp | the purpose of those periodics is to be able to test because it takes too much time to run them on gates. | 13:16 |
| jamesdenton | ok spatel. anything new? | 13:16 |
| spatel | so since i switch to F5 i am seeing that error everywhere in logs "(2006, "MySQL server has gone away (error(104, 'Connection reset by peer'))") [SQL: u'SELECT 1'] (Background on this error at: http://sqlalche.me/e/e3q8)" | 13:17 |
| evrardjp | for ppl who followed it, I said on last PTG, summit, and ML we need ppl to help there. | 13:17 |
| spatel | I have tried everything, tried oneconnect profile... tried standard vip everything but issue was there.. | 13:18 |
| evrardjp | s/for ppl who followed it/for the question of who is working on the periodics/g | 13:18 |
| spatel | Finally when i switch from "SNAT" to "automap" that resolved that issue... | 13:18 |
| spatel | jamesdenton: ^^ | 13:18 |
| spatel | I have noticed in tcpdump, F5 was sending RST tcp flag to client every minute | 13:19 |
| evrardjp | I did it for almost 3/4 of a cycle in Queens, and I got almost nothing done in that cycle, except fixing other ppl things. THat's why I wanted to split responsibility on those. | 13:19 |
| mnaser | Well the thing is most periodics are just failing so if once they fail there is no ownership for it, no point in them using up CI cycles | 13:19 |
| evrardjp | now, it's up to you mnaser -- you keep them, you leave them, it's up to you. | 13:19 |
| spatel | jamesdenton: Do you have any explanation, that why SNAT causing issue here? | 13:19 |
| mnaser | It’s up to us :) I don’t take decisions I just try to help us all to figure out what’s ideal | 13:20 |
| evrardjp | mnaser: ownership is not mine to decide anymore. | 13:20 |
| evrardjp | :0 | 13:20 |
| evrardjp | :) * | 13:20 |
| odyssey4me | mnaser: ideal is to have attention on those periodics, because they expose more coverage issues | 13:20 |
| evrardjp | well what's ideal is to have ppl on it | 13:20 |
| mnaser | I do think however that right now my focus is on getting OSA integrated jobs stable and slowly adding back the things that were dropped out | 13:20 |
| mnaser | Yeah, do we however have a volunteer to monitor those jobs? If someone steps up I’m all for making it | 13:20 |
| mnaser | Or maybe 2 people to share that responsibility | 13:21 |
| evrardjp | that's been the problem for all queens and rocky -- nobody wanted to take that. | 13:21 |
| evrardjp | At some point I just stopped doing it. | 13:21 |
| odyssey4me | well, it might be easier if we discuss it in the meeting - and also discuss what the resposibility entails | 13:21 |
| jamesdenton | spatel I'm not sure of your SNAT configuration, unless you were using the SNAT pool from the script. It's very possible there is/was an IP conflict. Automap uses the self-ip or floating self-ip of the F5 interface | 13:21 |
| odyssey4me | I would suggest that it not be the responsibility to fix it, but instead just register bugs for it | 13:21 |
| evrardjp | odyssey4me: then we need someone that actively fixes things too. | 13:22 |
| mnaser | odyssey4me: that’s reasonable too (both on the fact of not having to fix it and discussing in meeting) | 13:22 |
| evrardjp | it's reasonable to me | 13:22 |
| odyssey4me | triage of the bugs can be done in the weekly meeting, and fixes done by whoever picks it up | 13:22 |
| spatel | I have cross check SNAT IP its not used anywhere.. if its used anywhere F5 won't let me put there also other application working fine.. only MySQL doesn't like this.. | 13:22 |
| spatel | jamesdenton: ^^ | 13:22 |
| mnaser | I can work on CI stuff. It’s been mostly what I’ve been doing | 13:22 |
| jamesdenton | QQ: What is the preferred styling for a long if statement/conditional > 79 chars? Specifically https://review.openstack.org/#/c/592661/1/library/provider_networks, line 211 and others like it? | 13:23 |
| odyssey4me | in order to scale the resources we have, we have to make each responsibility work within the time and skills people have | 13:23 |
| mnaser | Given the split in users however, I feel like we either gotta agree to never put jobs in non voting and be forced to fix other operating systems if they need to land their stuff | 13:23 |
| mnaser | Or allow anyone to put jobs to non voting for any other OS if there isn’t reasonable progress in fixing it | 13:23 |
| evrardjp | jamesdenton: use black :p | 13:23 |
| spatel | jamesdenton: I am going to ask this question to F5 community to see if they have something to say on this.. | 13:24 |
| jamesdenton | spatel I can't really say. I would expect the behavior to be the same. automap is fine. A snat pool just gives you flexibility to not run out of ephemeral ports | 13:24 |
| evrardjp | jamesdenton: https://github.com/ambv/black | 13:24 |
| mnaser | Because I’ve noticed the more a job is non voting the more likely it will further break and be harder to fix | 13:24 |
| odyssey4me | mnaser: that only works if we have active people available on a consistent basis with the expertise to fix on each platform | 13:24 |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: Updated from OpenStack Ansible Tests https://review.openstack.org/593585 | 13:24 |
| jamesdenton | evrardjp thanks!!! | 13:24 |
| evrardjp | jamesdenton: tool of choice when not caring about formatting. | 13:25 |
| evrardjp | it's a beta thing but I like it | 13:25 |
| spatel | jamesdenton: I totally agreed about SNAT solution, even i have 4 million connection running on same F5 with 20 IPs of SNAT on same hardware :) I know the pain of running out of ports :) | 13:25 |
| odyssey4me | mnaser: and if that's an approach you'd like to take, then quite honestly we need to reduce our test matrix to things the active contribution base want... because right now we support too many options to reasonably cover with the active contribution base | 13:26 |
| spatel | jamesdenton: ^^ I will try to switch SNAT IP and check if incase something funky going on | 13:27 |
| jamesdenton | spatel sure. or hop into a container now and see if you can ping or arp for that other snat addr | 13:27 |
| jamesdenton | the f5 may very well still respond, tho | 13:28 |
| spatel | I did that same test, before used SNAP IP but i will do it again.. | 13:28 |
| mnaser | odyssey4me: i agree with you 100% | 13:28 |
| mnaser | i think given the amount of deployment choices we offer and the resources we have, it's too much | 13:29 |
| spatel | jamesdenton: I asked to F5 folks and they said this "You can try to configure SNAT and attache new FastL4 profile with PVA acceleration to None or Offload State to EST." | 13:29 |
| odyssey4me | it's fine to have them - but then we need to be clear about the things we support and test, and the things we have code for, but don't test or actively maintain | 13:29 |
| mnaser | odyssey4me: that's very fair, and things that are actively tested and maintained are ones where there are relatively quick-to-respond maintainers | 13:30 |
| mnaser | so it doesn't block our gate forever | 13:30 |
| odyssey4me | as a community we can choose a time-boxed timeframe to invest in a new option, like we did for centos, but then it needs to bring in new contributors by the close of that time-box otherwise we should abandon it or put it into sunset mode or something | 13:30 |
| jamesdenton | spatel i've never really had an issue with any snat configuration that couldn't be explained by something at a lower level (i.e. arp conflict) | 13:30 |
| jamesdenton | but who knows! | 13:31 |
| odyssey4me | if we don't do that, the burden just grows on the active contribution base and the pain escalates | 13:31 |
| spatel | jamesdenton: give me little time to find out :) | 13:31 |
| mnaser | odyssey4me: i agree, though for centos given that we deploy it, i kinda understand we might be a bit of an outlier and i'm always here to fix stuff up to make it work | 13:31 |
| odyssey4me | mnaser: sure, but you're one person | 13:31 |
| mnaser | mgariepy deploys centos, spatel i think deploys centos too :) | 13:32 |
| odyssey4me | quite honestly, the model only works if there're at least two | 13:32 |
| mnaser | (and in other news, guilhermesp is officially with us and he'll be working on OSA and focusing on centos too) | 13:32 |
| mnaser | so i'm making sure that there is right resources on it | 13:32 |
| mnaser | so it doesnt lag behind | 13:32 |
| odyssey4me | for those of us who don't use centos for production, we need those who do to be accessible to help figure things out and do active improvements | 13:33 |
| odyssey4me | so yeah, thanks for persisting to getting it back into the green | 13:33 |
| odyssey4me | hopefully that can be maintained | 13:33 |
| guilhermesp | o/ it is a pleasure to be with this awesome team! Thanks mnaser | 13:34 |
| odyssey4me | It might be worth considering limiting the support to distribution builds only if the contributor base is too small. | 13:34 |
| odyssey4me | ah, fantastic :) | 13:34 |
| mnaser | odyssey4me: i think that decision could be up to the maintainers of the distro | 13:36 |
| mnaser | if they want to support both and can carry/support the extra load, sure | 13:36 |
| mnaser | though the only reason i still have/need deploy from source is because distro deploys are not fully there yet | 13:36 |
| jamesdenton | evrardjp Thanks: https://black.now.sh/ | 13:37 |
| odyssey4me | mnaser: yep, makes sense - perhaps it'd be good to somehow formalise who the core maintainers are for different subsections of work - distributions, or perhaps other things like nspawn, lxc, etc | 13:38 |
| mnaser | odyssey4me: ++ yes, i agree | 13:38 |
| mnaser | we need some sort of processs to handle all of that in a reasonable way that allows for folks to integrate but ensures accountability to not slow down the existing other contributors | 13:39 |
| odyssey4me | yeah, that'd be nice | 13:39 |
| evrardjp | mnaser: I started the idea of having SIG in the past, including one for upgrades | 13:40 |
| evrardjp | not openstack SIGs, but you understand the idea | 13:40 |
| evrardjp | odyssey4me: was all in IIRC :p | 13:40 |
| odyssey4me | yeah, everyone trying to do everything just doesn't work | 13:42 |
| evrardjp | The goal was double: Introduce more ppl to responsibilities based on what they care, even not core (and a path to core), and dilute the responsibilities into groups instead of individuals | 13:42 |
| odyssey4me | if you have a smaller focus area, it's much easier to come up with ways to cut tech debt and improve things | 13:42 |
| mnaser | odyssey4me: https://etherpad.openstack.org/p/osa-stein-ptg i added something there | 13:43 |
| odyssey4me | oh bother, apparently ansible_user isn't defined either: http://logs.openstack.org/92/591392/5/check/openstack-ansible-deploy-ceph-ubuntu-xenial/8ad2ae2/logs/ara-report/result/f08f1cad-4e3d-41f7-ac18-df31c24b2784/ | 13:43 |
| mnaser | it might not be a template-able field | 13:43 |
| odyssey4me | it may be time to revert that whole change now | 13:43 |
| mnaser | kinda how like the python interpreter was not templated for a whiel | 13:44 |
| odyssey4me | noonedeadpunk: ^ I think it's best we revert those two patches to get ceph back to working | 13:48 |
| odyssey4me | whatever patch comes next will need to be tested using a dummy patch to the integrated repo and a depends-on | 13:48 |
| noonedeadpunk | odyssey4me: very strange thing, but yes, lets revert it | 13:50 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Revert "Setting default port if ansible_port is not defined" https://review.openstack.org/593622 | 13:50 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-plugins master: Modify provider_network plugin to compare group_binds to group_names https://review.openstack.org/592661 | 13:50 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Revert "Added possibility to specify ssh username and port for ceph mons." https://review.openstack.org/593623 | 13:51 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible master: Provides group_names for use with provider_network plugin https://review.openstack.org/592659 | 13:52 |
| odyssey4me | noonedeadpunk: I'd suggest doing a combined revert, and please state the reason for the revert in the commit message. | 13:52 |
| noonedeadpunk | odyssey4me: so commit revert just makes another commit? And how to make combined revert? | 13:54 |
| odyssey4me | noonedeadpunk: you'd need to revert from the git cli | 13:55 |
| *** aicilamd has joined #openstack-ansible | 13:55 | |
| noonedeadpunk | and then use git review? | 13:55 |
| noonedeadpunk | ok, should I abandon reverts made from gerrit in favor of git cli? | 13:56 |
| odyssey4me | noonedeadpunk: you can download the current revert, then 'git revert <sha>' the other one, then 'git rebase -i HEAD~3' and squash them, then git review | 13:56 |
| noonedeadpunk | Sorry, I've never used gerrit before and some things are really new for me | 13:57 |
| odyssey4me | no worries, this is how we learn :) | 13:57 |
| mnaser | noonedeadpunk: i have been using gerrit for over 5 years now | 13:58 |
| mnaser | and i'm still learning and breaking things :D | 13:58 |
| jrosser | i wonder how many people actually know where/how the periodic jobs come to be and how to see whats working / not working | 14:01 |
| jrosser | ^ btw i am *not* one of those poeple | 14:01 |
| *** Bhujay has quit IRC | 14:02 | |
| guilhermesp | gerrit is love in form of a review system <3 | 14:02 |
| guilhermesp | but I'm still breaking things to learn hahah | 14:02 |
| mnaser | jrosser: neither am I.. i dont really know how they work | 14:03 |
| noonedeadpunk | odyssey4me: probably I'd add default user for https://github.com/openstack/openstack-ansible-ceph_client/blob/master/defaults/main.yml#L60 ? | 14:04 |
| jrosser | i would love to see i screen with a grid of red/green vs job and time <- will be totally unsurprised if this already exists | 14:04 |
| mgariepy | jrosser, result are here : http://zuul.openstack.org/builds.html?pipeline=periodic&project=openstack%2Fopenstack-ansible | 14:06 |
| *** chandankumar has quit IRC | 14:12 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Revert "Added possibility to specify ssh username and port for ceph mons." https://review.openstack.org/593623 | 14:22 |
| *** shyamb has quit IRC | 14:25 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Revert "Added possibility to specify ssh username and port for ceph mons." Revert "Setting default port if ansible_port is not defined" https://review.openstack.org/593623 | 14:27 |
| noonedeadpunk | odyssey4me mnaser ^ | 14:28 |
| noonedeadpunk | I suppose I did revert right now, but it's not 100% | 14:28 |
| *** aicilamd has quit IRC | 14:31 | |
| *** aicilamd has joined #openstack-ansible | 14:32 | |
| *** aicilamd has quit IRC | 14:37 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: When saving VM images RC=2 is not a fail https://review.openstack.org/593652 | 14:47 |
| *** aicilamd has joined #openstack-ansible | 14:48 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ceph_client master: Revert "Added possibility to specify ssh username and port for ceph mons." https://review.openstack.org/593623 | 14:49 |
| odyssey4me | noonedeadpunk: ^ I just added a reason to the commit msg and removed the second title | 14:49 |
| cloudnull | mornings | 14:50 |
| jamesdenton | hello cloudnull | 14:52 |
| *** dxiri has quit IRC | 14:52 | |
| cloudnull | o/ | 14:53 |
| odyssey4me | o/ | 14:53 |
| openstackgerrit | Jacob Wagner proposed openstack/openstack-ansible-ops master: Add ability to deploy Ceph into a Multinode AIO https://review.openstack.org/585003 | 14:53 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Ensure package cache is updated for nspawn containers https://review.openstack.org/588430 | 14:54 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: When gating nspawn containers use a simplified config https://review.openstack.org/593230 | 14:54 |
| spatel | mnaser: are you talking about running openstack-ansible on CentOS ? | 14:57 |
| spatel | If that was the question then yes we are running Openstack-ansible QUEENS on CentOS 7.5 | 14:58 |
| spatel | and its on production | 14:58 |
| *** macza has joined #openstack-ansible | 15:02 | |
| odyssey4me | d34dh0r53 cloudnull got a min to review https://review.openstack.org/593652 ? | 15:02 |
| jamesdenton | evrardjp are you responsible for these opensuse failures??? | 15:03 |
| *** luksky has quit IRC | 15:07 | |
| *** DanyC has quit IRC | 15:09 | |
| *** pcaruana has quit IRC | 15:11 | |
| evrardjp | jamesdenton: if the police officer comes, it's not me. | 15:12 |
| evrardjp | hahah | 15:12 |
| evrardjp | how can I help? | 15:12 |
| jamesdenton | :) | 15:13 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add FS detection in metricbeat for containers https://review.openstack.org/593196 | 15:14 |
| jamesdenton | For me, there have been some recurring failures with the opensuse tests. http://logs.openstack.org/10/590810/2/check/openstack-ansible-functional-distro_install-opensuse-423/4aacba2/ara-report/ and specifically http://logs.openstack.org/10/590810/2/check/openstack-ansible-functional-distro_install-opensuse-423/4aacba2/logs/openstack/server2/neutron/neutron-server.log.txt.gz | 15:14 |
| *** tailDASHeF has quit IRC | 15:15 | |
| *** cshen has joined #openstack-ansible | 15:15 | |
| odyssey4me | cores - need another review on this revert asap: https://review.openstack.org/593623 | 15:21 |
| mnaser | odyssey4me: +W | 15:26 |
| mnaser | thank you for following up with that | 15:26 |
| *** dxiri has joined #openstack-ansible | 15:27 | |
| mnaser | jamesdenton: pushing up all sorts of cool networking stuff to os_neutron :D | 15:27 |
| jamesdenton | i'm trying not to break too much ;) | 15:27 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: MNAIO: When saving VM images RC=2 is not a fail https://review.openstack.org/593652 | 15:30 |
| *** aicilamd has quit IRC | 15:36 | |
| *** aicilamd has joined #openstack-ansible | 15:37 | |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client master: Revert "Added possibility to specify ssh username and port for ceph mons." https://review.openstack.org/593623 | 15:38 |
| *** aicilamd has quit IRC | 15:38 | |
| *** gyee has joined #openstack-ansible | 15:38 | |
| *** aicilamd has joined #openstack-ansible | 15:38 | |
| cloudnull | odyssey4me sorry missed the message | 15:39 |
| cloudnull | here now | 15:39 |
| kaiokmo | guys, which one of those expressions should work properly for a "suse-42.yml" named vars file? http://paste.openstack.org/show/728431/ | 15:44 |
| *** dxiri has quit IRC | 15:45 | |
| jrosser | kaiokmo: you might want to have a poke around some of the other roles, i noticed before that list is not consistent everywhere | 15:45 |
| kaiokmo | jrosser: that's what I was thinking... | 15:46 |
| kaiokmo | jrosser: will do. thanks :) | 15:46 |
| mrhillsman | i just finished a new install where the setup-openstack playbook failed so i re-ran and now it keeps saying ansible_env is not set | 15:55 |
| mrhillsman | the only way to resolve was to modify os-keystone-install.yaml and set gather_facts to yes as it says no currently | 15:55 |
| mrhillsman | is this a known thing? | 15:55 |
| *** vnogin has quit IRC | 15:56 | |
| *** aicilamd has quit IRC | 15:56 | |
| odyssey4me | https://github.com/openstack/openstack-ansible/blob/master/playbooks/os-keystone-install.yml#L36 default is true, not false? | 15:56 |
| *** vnogin has joined #openstack-ansible | 15:56 | |
| odyssey4me | the only time it's set to false is when gate-check-commit is set as far as I know | 15:56 |
| mrhillsman | task above that one | 15:56 |
| mrhillsman | that's where it fails | 15:56 |
| *** aicilamd has joined #openstack-ansible | 15:56 | |
| odyssey4me | ah, interesting | 15:57 |
| odyssey4me | looks like we either need to switch these to use a lookup, or we should gather facts: https://github.com/openstack/openstack-ansible-openstack_openrc/blob/master/defaults/main.yml#L29-L47 | 15:57 |
| odyssey4me | great find, thanks mrhillsman | 15:58 |
| mrhillsman | very welcome | 15:58 |
| odyssey4me | yeah, seems gathering facts would be best | 15:58 |
| mrhillsman | was not sure if a bug should be filed so figured i would ask first | 15:58 |
| odyssey4me | want to patch that up, or shall I? | 15:58 |
| mrhillsman | you would probably be faster lol | 15:59 |
| mrhillsman | if you have something now you're focused on i can though | 15:59 |
| *** aicilamd has quit IRC | 16:01 | |
| *** cshen has quit IRC | 16:02 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Gather facts for openstack_openrc role https://review.openstack.org/593684 | 16:03 |
| odyssey4me | mrhillsman: if you can test and verify that's good, that'd be most useful - you can then comment/vote as such | 16:03 |
| mrhillsman | ++ doing that now | 16:03 |
| *** aicilamd has joined #openstack-ansible | 16:07 | |
| evrardjp | jamesdenton: I have seen that in the past | 16:08 |
| evrardjp | but I thought this was fixed. Let me try your patch and find it out. | 16:08 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Adds DPDK support for Open vSwitch https://review.openstack.org/590810 | 16:09 |
| evrardjp | odyssey4me: that's interesting, but lookup isn't doing exactly the same thing. It would lookup the user of the user running ansible on the deploy, where this was, I guess, intended for the destination user after connection to the host. | 16:13 |
| evrardjp | which is probably root | 16:13 |
| odyssey4me | evrardjp: yeah, that's why I opted to gather facts instead | 16:13 |
| evrardjp | odyssey4me: well | 16:14 |
| evrardjp | it's nice but it is the default | 16:14 |
| evrardjp | so utility shouldn't be impacted | 16:15 |
| evrardjp | I prefer explicit though : ) | 16:15 |
| *** cshen has joined #openstack-ansible | 16:17 | |
| *** DanyC has joined #openstack-ansible | 16:21 | |
| *** shardy has quit IRC | 16:21 | |
| dw | i've got an ubuntu 16.04 AIO vm i'd like to preserve (dev tools and junk installed), does anyone know why ssh is hanging up on me after some idle time? it does not appear to be TMOUT variable inserted into /etc/profile by ansible-hardening | 16:22 |
| dw | ..aaand i lied, somehow the variable reappeared :) | 16:23 |
| dw | seems i can just tweak the default to something huge, and no problem if it gets reapplied | 16:24 |
| mgariepy | dw ? https://github.com/openstack/ansible-hardening/blob/master/defaults/main.yml#L329 | 16:27 |
| *** KeithMnemonic has joined #openstack-ansible | 16:28 | |
| jrosser | dw: if you stick that var ^ into /etc/openstack_deploy/user_variables.yml you can set it to whatever you need | 16:29 |
| dw | ClientAliveInterval is just checking for hung connections, the smoking gun is security_rhel7_session_timeout | 16:30 |
| dw | thx jrosser, still no clue how these files are laid out :) | 16:30 |
| jrosser | dw: theres a match for those user_<>.yml file and they all get turned into -e ansible vars files, so win over everything else | 16:31 |
| *** Darcidride_ has joined #openstack-ansible | 16:32 | |
| jamesdenton | thanks evrardjp | 16:41 |
| d34dh0r53 | is stein going to be bionic only or both bionic and xenial? | 16:43 |
| mnaser | d34dh0r53: i think we don't have much of a say in that | 16:43 |
| d34dh0r53 | ok | 16:43 |
| mnaser | d34dh0r53: because canonical said no xenial packages for stein | 16:43 |
| mnaser | i remember odyssey4me telling me the details of that | 16:44 |
| odyssey4me | mnaser d34dh0r53 yeah, I think openstack is going for bionic only, so we will have to - otherwise UCA will be too far behind | 16:44 |
| odyssey4me | rocky will have to be our transition release | 16:44 |
| d34dh0r53 | yep, that's what I figured, just wanted to double check | 16:46 |
| odyssey4me | we'll have to start flipping the bits to remove xenial from master quite soon | 16:46 |
| *** aicilamd has quit IRC | 16:49 | |
| *** cshen has quit IRC | 16:54 | |
| * jrosser still bashing away at bionic | 16:58 | |
| jrosser | i've spent the entire day so far trying to get networkd to bring the boostrap-hosts interfaces up in a way that is functional | 16:58 |
| odyssey4me | jrosser: :( sounds ! fun | 16:59 |
| spatel | cloudnull: hey i want to set MTU 9000 on lxcbr0 interface who does that? | 17:00 |
| jrosser | its certainly not. sys-subsystem-net-devices-br\x2dvlan.device <- i mean what kind of nonsense is that :(((((( | 17:00 |
| odyssey4me | spatel: there's no need to do that - that bridge is only used for internet access from the containers | 17:01 |
| spatel | I am seeing strange MTU issue... | 17:02 |
| spatel | From HOST machine i can ping -s 9000 | 17:02 |
| spatel | but from container i am not able to do that only 1450 MTU is working | 17:03 |
| spatel | even in container i set 9000 MTU and i can see its there | 17:03 |
| odyssey4me | what're you trying to ping? something on one of the bridged networks, or on the other side of L3? | 17:03 |
| spatel | Trying to set MTU 9000 for br-storage to access ceph | 17:04 |
| spatel | I did set MTU=9000 on br-storage on infra node so that is good and i can ping 9000 bytes packet from infra node | 17:04 |
| spatel | I have ceph-mon container which is map with br-storage and i want to set MTU=9000 on ceph-mon container | 17:05 |
| odyssey4me | ok, and are you doing a ping from a container which is bridged to that? | 17:05 |
| odyssey4me | the higher mtu shouldn't be needed for a mon though, FYI - data is accessed directly from compute to storage... the mon just tells it where each OSD is | 17:05 |
| spatel | odyssey4me: from container (ceph-mon) to bridge interface i am not able to ping 9000 bytes packet | 17:06 |
| *** cshen has joined #openstack-ansible | 17:07 | |
| spatel | odyssey4me: my ceph-mon and all OSD running on same network so they need to communicate.. I had very big discussion with someone here about that.. | 17:07 |
| jrosser | i think you need 9k mtu from infra nodes to ceph for cinder/glance | 17:08 |
| -spatel- # Ceph network | 17:08 | |
| -spatel- monitor_address_block: "10.50.0.0/21" | 17:08 | |
| -spatel- public_network: "10.50.0.0/21" | 17:08 | |
| -spatel- cluster_network: "10.51.0.0/21" | 17:08 | |
| spatel | This is my Ceph configuration, monitor_address & public_network on same LAN so they need to talk right? | 17:09 |
| *** joabdearaujo has joined #openstack-ansible | 17:11 | |
| spatel | technically if i set MTU on br-storage and container then it should work... | 17:11 |
| spatel | look like its not working | 17:11 |
| spatel | br-storage: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 9000 | 17:11 |
| odyssey4me | ok, sorry I need to run - I think jrosser and logan- may be best positioned to advise you | 17:11 |
| spatel | odyssey4me: no worry! i will keep digging | 17:12 |
| odyssey4me | I'm out folks - have a great day/evening! | 17:14 |
| d34dh0r53 | see ya odyssey4me | 17:18 |
| spotz | seeya odyssey4me | 17:24 |
| *** aicilamd has joined #openstack-ansible | 17:26 | |
| jamesdenton | evrardjp guess that rebase did the trick. thanks! | 17:28 |
| *** aicilamd has quit IRC | 17:30 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/rocky: Revert "Revert "Work around nested virt issues on OVH test nodes"" https://review.openstack.org/593414 | 17:31 |
| spatel | odyssey4me: after reboot full infra node that fixed my MTU issue... sometime i hate computer :( | 17:32 |
| openstackgerrit | Merged openstack/openstack-ansible stable/rocky: Drop non-voting jobs and increase timeout by 20 minutes https://review.openstack.org/593416 | 17:34 |
| *** Darcidride_ has quit IRC | 17:37 | |
| *** Darcidride_ has joined #openstack-ansible | 17:38 | |
| evrardjp | jamesdenton: it's just the conditions that changed :) | 17:43 |
| evrardjp | I will check in my env | 17:43 |
| *** tailDASHeF has joined #openstack-ansible | 17:43 | |
| jamesdenton | k | 17:45 |
| evrardjp | but glad it worked better :) | 17:46 |
| jamesdenton | Like a good lobbyist, now i need to solicit reviews | 17:48 |
| jamesdenton | free beer in Berlin! | 17:48 |
| Tahvok | Hey guys! Having some production issues after latest osa run.. http://paste.openstack.org/show/9pC2FA0g42EUo9BfPrpY/ | 17:53 |
| Tahvok | After checking /etc/neutron/plugins/ml2/openvswitch_agent.ini I saw that local_ip is wrong | 17:53 |
| Tahvok | It's different on each compute, but the ip is completely wrong | 17:54 |
| Tahvok | For example, in the above log the ip should be 10.63.2.145 and not 10.63.2.34 | 17:55 |
| Tahvok | I'm trying to understand how it takes the ip, and so having some hard time.. Hope you could help me with it | 17:55 |
| noonedeadpunk | does anybody experienced problems with live migration on libvirt? As I'm having not more than 8mbit for live migration on 17.0.5 on one installation. Probably I've missed something... These all configs related to live migration http://paste.openstack.org/show/728424/ | 17:56 |
| *** electrofelix has quit IRC | 17:59 | |
| *** poopcat has joined #openstack-ansible | 18:01 | |
| Tahvok | Ok, found the culprit. The machine we ran from, had some old ansible facts cache, with some old computes.. | 18:02 |
| jamesdenton | right on | 18:03 |
| *** vnogin has quit IRC | 18:15 | |
| *** cshen has quit IRC | 18:23 | |
| *** DanyC has quit IRC | 18:29 | |
| *** DanyC has joined #openstack-ansible | 18:29 | |
| *** KeithMnemonic has quit IRC | 18:33 | |
| *** DanyC has quit IRC | 18:34 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Convert test network setup to use the networkd role https://review.openstack.org/592596 | 18:35 |
| *** aicilamd has joined #openstack-ansible | 18:43 | |
| *** aicilamd has quit IRC | 18:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Clean up tests a bit https://review.openstack.org/589259 | 18:47 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 18:48 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Updated from OpenStack Ansible Tests https://review.openstack.org/593587 | 18:49 |
| noonedeadpunk | when do we plan to add new version for queens? I'm asking, as 1 critical bug (as for me) has been fixed in nova recently https://bugs.launchpad.net/nova/+bug/1786346 | 18:51 |
| openstack | Launchpad bug 1786346 in OpenStack Compute (nova) rocky "live migrations slow" [High,Fix committed] - Assigned to Matt Riedemann (mriedem) | 18:51 |
| mnaser | noonedeadpunk: evrardjp usually takes care of bumps, if this is urgent, we can do a push for one of them, but you can also override it locally | 18:53 |
| mnaser | nova_git_repo: .. | 18:53 |
| mnaser | nova_git_branch: COMMIT_ID | 18:53 |
| noonedeadpunk | mnaser: yep, you're right, forgotten about this) | 18:53 |
| *** aicilamd has joined #openstack-ansible | 19:00 | |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/ansible-hardening master: Add makestep to chronyd config. https://review.openstack.org/593864 | 19:02 |
| mnaser | everyone: if you'll be in attendance at the ptg, can you please make sure you're added here: https://etherpad.openstack.org/p/osa-stein-ptg | 19:03 |
| mnaser | mgariepy: will you be attending this time as well? | 19:03 |
| mgariepy | mnaser, nop | 19:03 |
| mnaser | aw, bummer, okay :( | 19:03 |
| mgariepy | i won't be there. | 19:03 |
| *** luksky has joined #openstack-ansible | 19:04 | |
| *** aicilamd has quit IRC | 19:04 | |
| mgariepy | mnaser, i would like to attend at some point. but not this time. maybe the next one. | 19:05 |
| mnaser | mgariepy: might be easier next time with the most-probable-merge of both events | 19:05 |
| mgariepy | yeah maybe | 19:06 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Remove octavia_requires_pip_packages https://review.openstack.org/588177 | 19:06 |
| mgariepy | if anyone has a comment on https://review.openstack.org/593864 i would appreciate it. | 19:07 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/ansible-hardening master: Add makestep to chronyd config. https://review.openstack.org/593864 | 19:08 |
| prometheanfire | when is the next sha bump scheduled for stable branches? friday? | 19:13 |
| *** DanyC has joined #openstack-ansible | 19:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Updated from OpenStack Ansible Tests https://review.openstack.org/592046 | 19:20 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Set quota for octavia service project https://review.openstack.org/571368 | 19:20 |
| spotz | d34dh0r53: Hey if you know Denver, add more recommendations for places. Even if we don't use them other groups might like them:) | 19:26 |
| d34dh0r53 | spotz: I only lived there for a year and mostly hit up small dives, and it was 20+ years ago but I still have some friends in the area. I'll ask around | 19:27 |
| spotz | d34dh0r53: Thanks! I went by google and no national chains. Doesn't seem like there's much within walking under 30 | 19:28 |
| spatel | folks what is "/openstack/venvs/nova-17.0.8/bin/privsep-helper" ? | 19:29 |
| d34dh0r53 | spotz: denver does has pretty good public transportation, the busses are nasty but the light rail is nice (at least it was 20 years ago :) | 19:29 |
| spotz | d34dh0r53: Small enough group it'll work coordinating 20.. we're walking! | 19:30 |
| d34dh0r53 | hah, yep | 19:30 |
| spotz | d34dh0r53: I still haven't figured out Berlin yet:( hotels are pretty far from the citycube | 19:32 |
| *** aicilamd has joined #openstack-ansible | 19:34 | |
| *** DanyC has quit IRC | 19:39 | |
| mnaser | spatel: thats a nova detail | 19:45 |
| mnaser | evrardjp, odyssey4me, cloudnull, prometheanfire, cjloader, spotz, d34dh0r53: asking as you've put yourself down at the ptg, is there any specific time that you can't be free for taking team photos? | 19:46 |
| spotz | mnaser: M/T I'll be at OPs meetup, right now it's looking like UC will be meeting Wed AM, and fly out about noon on Fri | 19:46 |
| prometheanfire | mnaser: atm no | 19:47 |
| d34dh0r53 | Not that I know of at this point | 19:47 |
| prometheanfire | first come first serve | 19:47 |
| mnaser | any preferences in time? :) | 19:47 |
| prometheanfire | tuesday right after relmgmt? | 19:47 |
| mnaser | we've got anywhere ffrom 9:30 to 3:30 on tuesday and wendesday | 19:47 |
| mnaser | bahaha | 19:47 |
| mnaser | i'll put us down as that and we can always look into it later | 19:48 |
| prometheanfire | I like batching them, I may schedle reqs around then too | 19:48 |
| spatel | mnaser: thanks, do i need to put that in monitor "privsep-helper" ? | 19:48 |
| spatel | I am setting up monitoring | 19:48 |
| mnaser | probably not | 19:48 |
| mnaser | spatel: ^ | 19:48 |
| cjloader | mnaser: not atm | 19:49 |
| cloudnull | mnaser im in sunday out early friday | 19:49 |
| cjloader | i'm flying back Friday evening | 19:49 |
| spotz | I vote for Wed afternoon | 19:51 |
| prometheanfire | mnaser: actually, can we switch? | 19:51 |
| spatel | mnaser: I am setting up monitoring for openstack and checking all running PID on compute node.. i found "/openstack/venvs/nova-17.0.8/bin/privsep-helper" so wonder is this something i should worry about or bot | 19:51 |
| mnaser | prometheanfire: so requirement then osa? | 19:51 |
| prometheanfire | mnaser: since there's a lot of crossover between relmgmt and reqs | 19:51 |
| prometheanfire | ya | 19:51 |
| mnaser | yeah sure | 19:51 |
| mnaser | voila | 19:51 |
| mnaser | oops i messed up the editing | 19:52 |
| prometheanfire | ya, saw that :P | 19:52 |
| prometheanfire | thanks | 19:52 |
| mnaser | np | 19:52 |
| *** mwe1 has quit IRC | 19:55 | |
| *** Leo_m has joined #openstack-ansible | 20:04 | |
| Leo_m | changed: [localhost] => (item={u'device': u'/dev/sdb1', u'mount_point': u'/openstack', u'fstype': u'ext4'}) | 20:05 |
| Leo_m | failed: [localhost] (item={u'device': u'/dev/sdb2', u'mount_point': u'/var/lib/lxc', u'fstype': u'ext4'}) => {"changed": fa | 20:05 |
| Leo_m | lse, "failed": true, "item": {"device": "/dev/sdb2", "fstype": "ext4", "mount_point": "/var/lib/lxc"}, "msg": "Error mounti | 20:05 |
| Leo_m | ng /var/lib/lxc: mount: wrong fs type, bad option, bad superblock on /dev/sdb2,\n missing codepage or helper program, | 20:05 |
| Leo_m | or other error\n\n In some cases useful info is found in syslog - try\n dmesg | tail or so.\n"} | 20:05 |
| Leo_m | hi, has somebody seen this issue while deploying an AIO ^^ | 20:05 |
| mnaser | did you see anything useful in syslog | 20:06 |
| Leo_m | ERROR: can't access '/var/lib/machines': No such file or directory | 20:07 |
| *** v1k0d3n has quit IRC | 20:08 | |
| *** v1k0d3n has joined #openstack-ansible | 20:10 | |
| Leo_m | [1118696.056220] EXT4-fs (sdb2): VFS: Can't find ext4 filesystem | 20:11 |
| *** vnogin has joined #openstack-ansible | 20:20 | |
| *** spatel has quit IRC | 20:23 | |
| *** KeithMnemonic has joined #openstack-ansible | 20:23 | |
| *** spatel has joined #openstack-ansible | 20:24 | |
| *** spatel has quit IRC | 20:28 | |
| *** hamzaachi has joined #openstack-ansible | 20:32 | |
| *** ansmith has quit IRC | 20:34 | |
| *** aicilamd has quit IRC | 20:37 | |
| *** aicilamd has joined #openstack-ansible | 20:37 | |
| cloudnull | Leo_m was this an AIO ? | 20:42 |
| cloudnull | or something in prod? | 20:42 |
| Leo_m | yea | 20:42 |
| Leo_m | nah an AIO | 20:42 |
| *** aicilamd has quit IRC | 20:42 | |
| cloudnull | was it running a the gate script to setup ? | 20:42 |
| cloudnull | gate-check-commit.sh | 20:42 |
| FrankZhang | Leo_m: did you meet this problem on phobos lab? | 20:42 |
| Leo_m | I was just running the ./script.sh | 20:42 |
| Leo_m | yes in phobos | 20:43 |
| FrankZhang | if so, I got a magic script for you | 20:43 |
| cloudnull | ^ boom ! | 20:43 |
| Leo_m | didn't saw it last time I did the AIO | 20:43 |
| cloudnull | community :) | 20:43 |
| Leo_m | cool! lol | 20:43 |
| FrankZhang | circle it back to RAX | 20:43 |
| *** aicilamd has joined #openstack-ansible | 20:49 | |
| *** vnogin has quit IRC | 20:52 | |
| dw | been working with ansible for years, and OSA if the first playbook where i've been genuinely mystified :P~ | 20:56 |
| dw | i'm trying to grok how group_by functions, but none of the ansible docs, ansible src, or OSA example are particularly useful to understanding what's going on | 20:56 |
| d34dh0r53 | dw: :) you're welcome? | 20:56 |
| cloudnull | dw group_by in the inventory ? | 20:57 |
| dw | at the top of /opt/openstack-ansible/playbooks/containers-lxc-create.yml in this case | 20:57 |
| cloudnull | IE https://github.com/os-cloud/os-ansible-deployment/blob/master/etc/openstack_deploy/openstack_user_config.yml.aio#L53-L54 | 20:57 |
| cloudnull | oh, looking | 20:57 |
| dw | struggling to replicate a bug, and this dynamic inventory stuff might be related, it's the only thing (i think) my local test doesn't replicate | 20:58 |
| cloudnull | oh, the ansible task "group_by" ? | 20:58 |
| dw | yup. the docs are written like "lxc_containers" is a per-host fact, but i can't even find it defined anywhere, and besides it doesn't seem to be named like a fact | 20:59 |
| cloudnull | this https://docs.ansible.com/ansible/2.5/modules/group_by_module.html ? | 20:59 |
| dw | that's the one | 20:59 |
| cloudnull | I think its set dynamically in facts but idk if its accessible other than in a playbook | 20:59 |
| dw | oh! i'm reading things wrong. it adds the /current/ host to the group, and in OSA's case, when the when: is true | 21:00 |
| cloudnull | maybe dump the hostvars to see if tis something set there? | 21:00 |
| cloudnull | ++ yea we're sorting the container types by OS and CPU architecture. | 21:01 |
| dw | yeppers, i understand what it's intending to do, but not quite how :) i think it comes down to the 'key' parameter would be better named 'group_name', because that's exactly what it is. i thought it was some sort of indirect lookup key | 21:01 |
| cloudnull | ++ | 21:02 |
| dw | the action API part, you don't actually specify the host as being added, that's implicit, the actual host-added-to-group bit happens way down elsewhere in the guts of ansible | 21:02 |
| cloudnull | that would be better! | 21:02 |
| dw | is there a way to run e.g. 'ansible-inventory' with all the OSA environment setup, like 'openstack-ansible' does for 'ansible-playbook'? | 21:05 |
| cloudnull | you can execute the dynamic-inventory.py script directly | 21:06 |
| dw | of course :) thanks | 21:06 |
| cloudnull | you can also run `ansible-playbook -e @$USERVARIABLE_PATH ...` | 21:07 |
| cloudnull | the `openstack-ansible` command is just a convenience+ wrapper. | 21:07 |
| cloudnull | but `opt/openstack-ansible/inventory/dynamic_inventory.py` will return the inventory in JSON format | 21:08 |
| *** aicilamd has quit IRC | 21:10 | |
| *** aicilamd has joined #openstack-ansible | 21:11 | |
| *** aicilamd has quit IRC | 21:16 | |
| *** tailDASHeF has quit IRC | 21:18 | |
| *** spatel has joined #openstack-ansible | 21:19 | |
| *** DanyC has joined #openstack-ansible | 21:20 | |
| *** ansmith has joined #openstack-ansible | 21:27 | |
| dw | i can't replicate the bug.. because i'd fixed it already | 21:32 |
| dw | stopping for dinner = endgame :( hehe | 21:32 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Increase Amp Active timeout for tests https://review.openstack.org/590474 | 21:35 |
| *** hamzaachi has quit IRC | 21:52 | |
| cloudnull | dw there's anything I can do to help let me know, I'd be happy to help wherever possible | 21:55 |
| *** aicilamd has joined #openstack-ansible | 21:56 | |
| dw | cloudnull: thx! i just have to grin and bear it for the time being. but that's one bug out of 4 out of the way, so off to a decent start | 21:57 |
| cloudnull | nice! | 21:57 |
| dw | cloudnull: i want to get used to OSA anyway, the setup is huge, it's a perfect test for the next piece of work i have (and it'll also probably benefit from it) | 21:57 |
| cloudnull | ++ absolutly looking forward to it | 21:57 |
| dw | the new work is a total rewrite of the strategy, rather than a monkeypatch. it needs a lot of exercise | 21:57 |
| dw | (but it's also a hell of a lot faster! esp for multi-target) | 21:58 |
| cloudnull | sounds like itll be amazing ! | 21:58 |
| dw | it gets ansible's process model into sufficient shape that the ssh client can be moved in-process, itching to do that, but haven't finished the previous step yet | 21:59 |
| cloudnull | baby steps :) | 21:59 |
| *** aicilamd has quit IRC | 21:59 | |
| dw | indeed! | 21:59 |
| *** aicilamd has joined #openstack-ansible | 21:59 | |
| *** luksky has quit IRC | 22:05 | |
| *** Leo_m has quit IRC | 22:05 | |
| *** Leo_m has joined #openstack-ansible | 22:09 | |
| *** Leo_m has quit IRC | 22:10 | |
| *** Leo_m_ has joined #openstack-ansible | 22:10 | |
| *** Leo_m_ has quit IRC | 22:12 | |
| *** rabies_ has joined #openstack-ansible | 22:15 | |
| *** rabies_ has left #openstack-ansible | 22:15 | |
| *** dxiri has joined #openstack-ansible | 22:16 | |
| *** Laszlo-74 has joined #openstack-ansible | 22:16 | |
| dw | is it normal for a string of "Terminated" to be printed on the console when the run finishes? is that some wrapper, or some bug mitogen is tickling? | 22:17 |
| Laszlo-74 | hello everybody! | 22:17 |
| Laszlo-74 | I need some help with openstack-ansible | 22:17 |
| Laszlo-74 | I need to add an entri to the nova.conf and neutron conf files | 22:18 |
| Laszlo-74 | I'm using the user_variables.yml file to define the neutron_neutron_conf_overrides and nova_nova_conf_overrides as follows: | 22:19 |
| Laszlo-74 | neutron_neutron_conf_overrides: DEFAULT: host: "{{ inventory_hostname }}.example.intra" | 22:20 |
| Laszlo-74 | neutron_neutron_conf_overrides: | 22:20 |
| Laszlo-74 | DEFAULT: | 22:20 |
| Laszlo-74 | host: "{{ inventory_hostname }}.example.intra" | 22:20 |
| Laszlo-74 | this is adding the corresponding line in each host and container but I have a problem with the underscores form the names of the containers. Neutron is complaining that the name that I got with the mentioned expression is not a valid hostname | 22:22 |
| Laszlo-74 | I would like to replace the underscores to hyphens (network1_neutron_server_container -> network1-neutron-server-container) | 22:23 |
| Laszlo-74 | can I use the following line: | 22:23 |
| Laszlo-74 | host: "{{ inventory_hostname | regex_replace('_','-') }}.example.intra" | 22:24 |
| Laszlo-74 | ? | 22:24 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Document adding "extra" networks https://review.openstack.org/593967 | 22:24 |
| *** aicilamd has quit IRC | 22:25 | |
| Laszlo-74 | any ideas are welcome ... | 22:25 |
| *** aicilamd has joined #openstack-ansible | 22:26 | |
| *** aicilamd has quit IRC | 22:30 | |
| cloudnull | dw I've never seen "terminated" | 22:31 |
| cloudnull | could be something conflicting with our default plugins ? | 22:31 |
| cloudnull | terminated I would assume is coming from a thread being killed? | 22:31 |
| dw | it sounds like something SSH might print when sent a SIGTERM, like the extension does during shutdown :) but i've never seen it before | 22:31 |
| dw | will run the whole thing under strace and go digging | 22:32 |
| cloudnull | we have a couple ssh options we run by default | 22:32 |
| dw | do you know where those options are set? i see nothign related in openstack-ansible.rc | 22:33 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/all/all.yml#L114-L124 | 22:33 |
| dw | aaha :) | 22:33 |
| cloudnull | maybe something there is causing the issue? | 22:33 |
| dw | nothing immediately obvious. sometimes ssh prints extra stuff with multiplexing enabled, but nothing special going on ehre | 22:34 |
| cloudnull | Laszlo-74 can you use: host: "{{ ansible_hostname }}.example.intra" | 22:34 |
| cloudnull | Laszlo-74 you could also set it to the fqdn option - host: "{{ ansible_fqdn }}" | 22:35 |
| cloudnull | dw we also run with pipelinning on, would that cause any known issue? | 22:35 |
| *** Talion has joined #openstack-ansible | 22:35 | |
| Laszlo-74 | cloudnull what are the values of ansible_hostname and ansible fqdn? | 22:36 |
| cloudnull | Laszlo-74 those are facts generated by ansible. | 22:37 |
| cloudnull | should be the hostname / fqdn of your hosts | 22:37 |
| cloudnull | which should be RFC-1034/35 | 22:38 |
| dw | cloudnull: nopes :) have a rather slowww run right now with strace, it should catch it being printed somewhere in the logs | 22:40 |
| Laszlo-74 | cloudnull I see ... that's what I want to avoid. my hostname is in the form datacentername-chassis-blade (dc-c1-bl01) and the name that I would like to use for the host in openstack is compute1 | 22:40 |
| cloudnull | so http://paste.openstack.org/show/728467/ should probably get you want you need | 22:40 |
| cloudnull | ah | 22:40 |
| cloudnull | Laszlo-74 there's ansible_nodename ? | 22:41 |
| cloudnull | run `ansible -m setup $inventory_hostname` and look for those options | 22:42 |
| Laszlo-74 | how could I check ? | 22:42 |
| cloudnull | my hope is that one of them will get you there. | 22:42 |
| cloudnull | otherwise you can do the regex replace. | 22:42 |
| cloudnull | it won't always be an RFC compliant hostname, but it should work 90% of the time | 22:43 |
| Laszlo-74 | is this syntax correct for the regex? host: "{{ inventory_hostname | regex_replace('_','-') }}.example.intra" | 22:44 |
| Laszlo-74 | cloudnull An other option for me would be if I could restrict the override to be applied only to the compute nodes. Is that possible somehow? | 22:49 |
| cloudnull | Laszlo-74 the syntax looks good to me | 22:49 |
| cloudnull | you can restrict it to just the hosts. if you put in the openstack_user_config for all compute nodes under "host_vars" | 22:50 |
| cloudnull | something like so | 22:51 |
| cloudnull | http://paste.openstack.org/show/728468/ | 22:51 |
| cloudnull | but that could be a lot more config | 22:51 |
| cloudnull | depending on the number of compute nodes you have | 22:52 |
| dw | found it.. it's coming from an intermediary /bin/sh process that's injected as part of the container namespace transition. grmbl | 22:52 |
| cloudnull | is that from our ssh plugin ? | 22:53 |
| cloudnull | here https://github.com/openstack/openstack-ansible-plugins/blob/master/connection/ssh.py#L363-L390 | 22:54 |
| cloudnull | or local to mitogen> | 22:54 |
| cloudnull | 's/>/?/' | 22:54 |
| Laszlo-74 | cloudnull that means that I have to set the override for each compute host individually. I was thinking whether there is a way to do it for the entire group of compute nodes at once? | 22:54 |
| cloudnull | or both :) | 22:55 |
| cloudnull | Laszlo-74 what release? | 22:55 |
| dw | cloudnull: it's local to mitogen. actually this connection method shouldn't be used for here, but it's the only one jrosser could get working :) the real lxc plugin is broken for you somehow, i haven't looked at that yet | 22:55 |
| Laszlo-74 | queens | 22:55 |
| cloudnull | you could create /etc/openstack_deploy/group_vars/nova_compute.yml | 22:56 |
| cloudnull | and put that config there | 22:56 |
| cloudnull | I believe that still works. | 22:56 |
| * cloudnull hasnt tried in a while | 22:56 | |
| cloudnull | dw ah ok | 22:57 |
| *** schwicht has quit IRC | 23:04 | |
| openstackgerrit | Kaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca master: Add support for openSUSE https://review.openstack.org/588584 | 23:06 |
| *** Darcidride_ has quit IRC | 23:08 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Change AIO ZFS pool to be at /pool https://review.openstack.org/591304 | 23:08 |
| *** tosky has quit IRC | 23:08 | |
| *** DanyC has quit IRC | 23:26 | |
| *** pmannidi has joined #openstack-ansible | 23:29 | |
| *** DanyC has joined #openstack-ansible | 23:29 | |
| *** markvoelker has joined #openstack-ansible | 23:29 | |
| *** markvoelker has quit IRC | 23:32 | |
| *** DanyC has quit IRC | 23:33 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Document adding "extra" networks https://review.openstack.org/593967 | 23:50 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Add documentation for routed interfaces https://review.openstack.org/593974 | 23:50 |
| *** schwicht has joined #openstack-ansible | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!