| *** klamath has quit IRC | 00:05 | |
| *** klamath has joined #openstack-ansible | 00:08 | |
| *** nurdie_ has joined #openstack-ansible | 00:14 | |
| *** nurdie has quit IRC | 00:15 | |
| *** cjloader has quit IRC | 00:15 | |
| *** cjloader has joined #openstack-ansible | 00:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/pike: Adds constraints to all pip installs https://review.openstack.org/571282 | 00:18 |
|---|---|---|
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/queens: Adds constraints to all pip installs https://review.openstack.org/571283 | 00:19 |
| *** dave-mccowan has joined #openstack-ansible | 00:19 | |
| *** nurdie_ has quit IRC | 00:23 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/pike: Update repo_build test-vars data https://review.openstack.org/568810 | 00:23 |
| *** cjloader has quit IRC | 00:23 | |
| *** nurdie has joined #openstack-ansible | 00:23 | |
| *** dave-mccowan has quit IRC | 00:30 | |
| *** jeferobles has joined #openstack-ansible | 00:31 | |
| *** kstev has joined #openstack-ansible | 00:36 | |
| *** openstacking_123 has joined #openstack-ansible | 00:41 | |
| *** openstacking_123 has quit IRC | 00:45 | |
| mnaser | odyssey4me: thanks for your fix | 00:48 |
| mnaser | osa-cores: https://review.openstack.org/#/c/571274/2 good to fix things :) | 00:48 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/queens: An OSA password needs to end with _password https://review.openstack.org/571323 | 00:49 |
| *** chyka has joined #openstack-ansible | 00:52 | |
| *** chyka has quit IRC | 00:57 | |
| *** chigang__ has joined #openstack-ansible | 01:02 | |
| *** cshen has joined #openstack-ansible | 01:17 | |
| *** kstev has quit IRC | 01:21 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Adds the certificate client key password to user_secrets https://review.openstack.org/571319 | 01:27 |
| *** jra has quit IRC | 01:31 | |
| *** nurdie_ has joined #openstack-ansible | 01:31 | |
| *** jra has joined #openstack-ansible | 01:31 | |
| *** nurdie has quit IRC | 01:34 | |
| *** cshen has quit IRC | 01:49 | |
| *** mma has joined #openstack-ansible | 01:51 | |
| *** markvoelker has joined #openstack-ansible | 01:54 | |
| *** mma has quit IRC | 01:56 | |
| *** tux_ has joined #openstack-ansible | 02:01 | |
| *** markvoelker has quit IRC | 02:02 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Use the distro package for libvirt-python https://review.openstack.org/571274 | 02:03 |
| *** jwitko has joined #openstack-ansible | 02:06 | |
| *** markvoelker has joined #openstack-ansible | 02:08 | |
| *** markvoelker has quit IRC | 02:14 | |
| *** markvoelker has joined #openstack-ansible | 02:15 | |
| *** markvoelker has quit IRC | 02:19 | |
| *** cjloader has joined #openstack-ansible | 02:19 | |
| *** cjloader has quit IRC | 02:23 | |
| *** klamath has quit IRC | 02:39 | |
| *** jwitko_ has joined #openstack-ansible | 02:42 | |
| *** jwitko has quit IRC | 02:44 | |
| *** klamath has joined #openstack-ansible | 03:13 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 03:19 |
| *** klamath has quit IRC | 03:29 | |
| *** jeferobles has quit IRC | 03:32 | |
| *** nurdie_ has quit IRC | 03:32 | |
| *** nurdie has joined #openstack-ansible | 03:33 | |
| *** nurdie has quit IRC | 03:37 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 03:39 |
| *** bhujay has quit IRC | 03:47 | |
| *** ianychoi has quit IRC | 03:51 | |
| *** zul has quit IRC | 03:58 | |
| *** udesale has joined #openstack-ansible | 04:06 | |
| *** nurdie has joined #openstack-ansible | 04:07 | |
| *** radeks__ has joined #openstack-ansible | 04:11 | |
| *** gyee has quit IRC | 04:21 | |
| *** jeferobles has joined #openstack-ansible | 04:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/pike: Create Octavia policy roles and legacy admin-owner rules https://review.openstack.org/568874 | 04:39 |
| *** poopcat has quit IRC | 04:47 | |
| *** jeferobles has quit IRC | 04:49 | |
| *** nurdie has quit IRC | 04:54 | |
| *** nurdie has joined #openstack-ansible | 04:54 | |
| *** nurdie has quit IRC | 04:59 | |
| *** jeferobles has joined #openstack-ansible | 05:03 | |
| *** nurdie has joined #openstack-ansible | 05:19 | |
| *** hassenius has joined #openstack-ansible | 05:21 | |
| *** jmnk has quit IRC | 05:25 | |
| *** hassenius has quit IRC | 05:25 | |
| *** jeferobles has quit IRC | 05:26 | |
| *** jeferobles has joined #openstack-ansible | 05:26 | |
| *** jmnk has joined #openstack-ansible | 05:29 | |
| *** jeferobles has quit IRC | 05:41 | |
| *** jeferobles has joined #openstack-ansible | 05:44 | |
| *** jmnk has quit IRC | 05:47 | |
| *** tux_ has quit IRC | 05:51 | |
| *** jwitko_ has quit IRC | 05:57 | |
| *** sar has quit IRC | 05:57 | |
| *** jwitko has joined #openstack-ansible | 06:01 | |
| *** markvoelker has joined #openstack-ansible | 06:05 | |
| *** bhujay has joined #openstack-ansible | 06:06 | |
| *** jmnk has joined #openstack-ansible | 06:16 | |
| *** sar has joined #openstack-ansible | 06:26 | |
| *** jeferobles has quit IRC | 06:38 | |
| *** pcaruana has joined #openstack-ansible | 06:43 | |
| *** radeks__ has quit IRC | 06:43 | |
| *** klamath has joined #openstack-ansible | 06:44 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: This patch ensures that the functional tests for Blazar pass. https://review.openstack.org/571388 | 06:49 |
| *** klamath has quit IRC | 06:49 | |
| *** pvradu has joined #openstack-ansible | 06:56 | |
| Taseer | evrardjp: Good morning. The Ubuntu job is still not being triggered. https://review.openstack.org/#/c/571388/ | 06:58 |
| *** jwitko_ has joined #openstack-ansible | 06:59 | |
| *** mma has joined #openstack-ansible | 07:02 | |
| *** jwitko has quit IRC | 07:03 | |
| *** radeks__ has joined #openstack-ansible | 07:24 | |
| *** mbuil has joined #openstack-ansible | 07:25 | |
| *** Taseer has quit IRC | 07:32 | |
| *** shardy has joined #openstack-ansible | 07:32 | |
| *** ianychoi has joined #openstack-ansible | 07:32 | |
| *** Taseer has joined #openstack-ansible | 07:34 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: Do not pip install keystoneclient on the host https://review.openstack.org/571393 | 07:35 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_ceilometer master: Do not pip install keystoneclient on the host https://review.openstack.org/571394 | 07:38 |
| *** nurdie has quit IRC | 07:39 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_glance master: Do not pip install keystoneclient on the host https://review.openstack.org/571395 | 07:41 |
| *** tux_ has joined #openstack-ansible | 07:41 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: [WIP] Add OpenStack Zun support to OSA https://review.openstack.org/571400 | 07:48 |
| evrardjp | Taseer: edit something else than a documentation text file and you'll see | 07:48 |
| evrardjp | cloudnull: I will try to bring os_zun this week, as next week is the deadline | 07:48 |
| cloudnull | excellent. | 07:48 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_heat master: Do not pip install keystoneclient on the host https://review.openstack.org/571401 | 07:48 |
| cloudnull | hopefully folks find it useful | 07:49 |
| * cloudnull is just working on this on his freetime because it was asked for a couple times and I found it interesting. | 07:49 | |
| cloudnull | IDK how many users are out there in the wild | 07:50 |
| cloudnull | but it does look fairly complete and in local testing seems to work pretty well | 07:50 |
| cloudnull | OK, im off to sleep. | 07:50 |
| cloudnull | take care all | 07:50 |
| *** tux_ has quit IRC | 07:52 | |
| olivierbourdon38 | good morning everyone. I have an issue understanding networking settings required for OSA. My setup is as follows | 07:54 |
| olivierbourdon38 | I have a baremetal node with 2 physical network interfaces, 1 primary connected to internet (configured as a bridge) with proper GW/DNS servers/... and a secondary which is unconfigured but connected to the same switch | 07:57 |
| evrardjp | cloudnull: amazing job | 08:00 |
| openstackgerrit | Chason Chan proposed openstack/openstack-ansible-os_aodh master: Remove keystone_authtoken section for aodh https://review.openstack.org/571402 | 08:01 |
| evrardjp | olivierbourdon38: ok ? | 08:01 |
| evrardjp | and your first network is meant for external usage too? | 08:01 |
| olivierbourdon38 | evrardjp yes so this is where br-mgmt bridge should be right ? | 08:03 |
| olivierbourdon38 | on this second interface I create a 1st bridge (manual) on which I create VLANS for OSA and 1 extra for provisioning and other bridges on top of these VLANS | 08:04 |
| evrardjp | olivierbourdon38: up to you if you make it a mgmt bridge | 08:04 |
| evrardjp | the mgmt bridge can be something else. | 08:05 |
| olivierbourdon38 | evrardjp ok so I can create br-mgmt on the second interface like all the others, I won't get internet connectivity ? | 08:06 |
| evrardjp | the container connectivity is done with dnsmasq and iptables. | 08:07 |
| evrardjp | on lxcbr0 | 08:07 |
| evrardjp | by default | 08:07 |
| evrardjp | towards the outside | 08:07 |
| evrardjp | br-mgmt is an internal network | 08:07 |
| evrardjp | to do mgmt trafic of your openstack cloud | 08:07 |
| olivierbourdon38 | ok got it now (I think) thanks | 08:08 |
| evrardjp | you can probably do br-mgmt, br-storage, br-vxlan on this second nic, each of them having one vlan | 08:08 |
| evrardjp | the br-vlan will be a different case | 08:08 |
| evrardjp | because, IIRC, in ubuntu 14 (maybe still in ubuntu 16) you couldn't pass the device nic that's already used at other place to neutron | 08:09 |
| evrardjp | so you'd have to do some veth pluggin | 08:10 |
| evrardjp | or doing q-in-q | 08:10 |
| evrardjp | both are fine, but pay attention to your mtus | 08:10 |
| *** admin0 has joined #openstack-ansible | 08:10 | |
| admin0 | \o | 08:11 |
| hwoarang | evrardjp: i don't understand your -1 here https://review.openstack.org/#/c/571401/ because this has already been approved on the keystone role :/ | 08:12 |
| *** hassenius has joined #openstack-ansible | 08:12 | |
| *** hasseniu_ has joined #openstack-ansible | 08:13 | |
| *** markvoelker has quit IRC | 08:15 | |
| *** hassenius has quit IRC | 08:17 | |
| *** radeks__ has quit IRC | 08:19 | |
| odyssey4me | andymccr unfortunately it looks like you broke the integrated build - we must be missing a nuance in tests: http://logs.openstack.org/14/570914/3/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/1523ae8/job-output.txt.gz#_2018-05-31_03_11_00_668447 | 08:20 |
| odyssey4me | the breaking change is https://review.openstack.org/570881 evrardjp perhaps you can spot the issue quickly so we can solve it | 08:21 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: This patch ensures that the functional tests for Blazar pass. https://review.openstack.org/571388 | 08:21 |
| odyssey4me | otherwise master is blocked for more days | 08:21 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/pike: Fix rabbitmqctl cluster_name regexp extractor https://review.openstack.org/570674 | 08:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/queens: Use the distro package for libvirt-python https://review.openstack.org/571404 | 08:23 |
| openstackgerrit | bhujay kumar proposed openstack/openstack-ansible-lxc_hosts master: making lxc image prestage timeout configurable https://review.openstack.org/571405 | 08:24 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/pike: Use the distro package for libvirt-python https://review.openstack.org/571406 | 08:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Use the distro package for libvirt-python https://review.openstack.org/571407 | 08:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/pike: tests: Pin sahara-tests repo to ~Pike https://review.openstack.org/571137 | 08:29 |
| *** bhujay has quit IRC | 08:30 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: [DNM] Testing openSUSE jobs https://review.openstack.org/571408 | 08:33 |
| *** electrofelix has joined #openstack-ansible | 08:35 | |
| *** radeks__ has joined #openstack-ansible | 08:36 | |
| *** hasseniu_ has quit IRC | 08:36 | |
| odyssey4me | evrardjp we should see in https://review.openstack.org/571137 whether than libvirt-python patch fixes things for centos | 08:37 |
| odyssey4me | evrardjp what's holding us up from moving to ansible 2.5 or perhaps 2.6 for master? | 08:41 |
| *** radeks__ has quit IRC | 08:41 | |
| odyssey4me | evrardjp perhaps you could look at the rest of that sahara-tests pin chain: https://review.openstack.org/#/q/I5da2ffcd4d4f330778707b2edd639089559ed778 | 08:43 |
| *** hassenius has joined #openstack-ansible | 08:44 | |
| andymccr | will look into it now. that is weird since the individual gate passed ok perhaps integrated inventory doesnt get created quite the same way. | 08:44 |
| *** hasseniu_ has joined #openstack-ansible | 08:44 | |
| odyssey4me | andymccr yeah, ideally we should figure out what's different there and make sure the etsts have a similar condition to test against | 08:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/pike: Do not log passwords (part 2) https://review.openstack.org/571150 | 08:45 |
| openstackgerrit | Gaudenz Steinlin proposed openstack/openstack-ansible-os_nova master: Fix nova_console_novnc_ssl.yml include path https://review.openstack.org/571413 | 08:45 |
| *** klamath has joined #openstack-ansible | 08:46 | |
| *** hassenius has quit IRC | 08:48 | |
| *** klamath has quit IRC | 08:51 | |
| evrardjp | how encouraging: http://zuul.openstack.org/builds.html?pipeline=periodic&project=openstack%2Fopenstack-ansible | 08:52 |
| evrardjp | hwoarang: I am sorry if that has been accepted on keystone. this is what I think. I do not say I will block the commit (-2 or -w) if you get 2 +2 and +w. I just think what I commented is the best for the future. | 08:53 |
| evrardjp | We can ofc iterate on it, but I know ourselves: it won't happen soon. | 08:54 |
| evrardjp | odyssey4me: I will have a look at the lxc part. | 08:55 |
| evrardjp | the patch made sense though,so I will spend cycles on it | 08:56 |
| *** mbuil has quit IRC | 08:56 | |
| evrardjp | andymccr: maybe wrap the list with (), like {{ ( whateverdict | list) | sort }} | 08:57 |
| andymccr | evrardjp: yeah testing that out | 08:57 |
| evrardjp | I can help with initial data. Will check ara if you're not on it. In the meantime, I trust you :) | 08:58 |
| hwoarang | evrardjp: that's why i started the distro work with keystone so we can all agree on the implementation details. what you suggest makes sense but that requires quite a bit of a re-design and it's out of scope of the BP so i am not going to do that right now | 08:58 |
| odyssey4me | evrardjp I'm afraid that your suggestion is not at all simple, and a waste of time. Ansible 2.5 will allow us to get rid of the keystoneclient because we can just use shade. | 08:58 |
| hwoarang | i still think the proposed patch is a minor improvement over the current situation | 08:58 |
| evrardjp | odyssey4me: nothing is a problem for 2.5 anymore afaik | 08:59 |
| evrardjp | odyssey4me: it's not a waste of time, because of python3. | 08:59 |
| evrardjp | and because we have a tendency to not remove packages from package lists. | 09:00 |
| odyssey4me | evrardjp also, I tried setting the ansible runtime used to a venv location, and it didn't work | 09:00 |
| hwoarang | 2.5 is beneficial because upstream told me there is not going to be another 2.4 release :( | 09:00 |
| hwoarang | and i had some fixes on that branch which will not make it to a release :( | 09:00 |
| evrardjp | hwoarang: I do not disagree. | 09:00 |
| *** Tahvok has quit IRC | 09:00 | |
| evrardjp | I agree with what you are all stating for distro pacakges, the commit is, I think good. | 09:00 |
| odyssey4me | I've been working on thinning that down in https://review.openstack.org/#/q/topic:cleanup-host-pip-installs+(status:open+OR+status:merged) | 09:00 |
| evrardjp | When doing source installs, I don't think it's a good idea to mix distro packages and venv, for their usage. | 09:01 |
| odyssey4me | I need to figure out the failure in https://review.openstack.org/568142 - I think it's because shade's not on localhost at all. | 09:01 |
| evrardjp | but like I said, I won't block it. Go ahead, let's see what the future will tell us. | 09:01 |
| hwoarang | odyssey4me: possibly but remember ubuntu has a very old shade | 09:02 |
| odyssey4me | hwoarang if stable-2.4 has fixes in it which we need, we can bump to use a SHA | 09:02 |
| hwoarang | you wont be able to use the upstream ansible os_* modules :( | 09:02 |
| *** tux_ has joined #openstack-ansible | 09:02 | |
| evrardjp | let's split the conversation in two shall we? | 09:02 |
| odyssey4me | hwoarang yep, but if you look at those patches, I'm trying to make it use the ansible runtime venv for shad, not a host installed package | 09:02 |
| odyssey4me | the intent is not completely remove using a host installed package | 09:03 |
| evrardjp | it seems in a situation like this: I'd say we should move to 2.5 first, use os_* modules | 09:03 |
| odyssey4me | *to completely remove | 09:03 |
| evrardjp | then do these patches. | 09:03 |
| hwoarang | but those two probs are orthogonal right? | 09:03 |
| evrardjp | probably | 09:03 |
| hwoarang | 2.5 vs os_* modules | 09:03 |
| evrardjp | mmm no | 09:04 |
| odyssey4me | this is why I asked what the hold up is with 2.5 earlier | 09:04 |
| evrardjp | os_* needs 2.5 | 09:04 |
| hwoarang | ah | 09:04 |
| evrardjp | for endpoint mgmt | 09:04 |
| *** markvoelker has joined #openstack-ansible | 09:04 | |
| odyssey4me | yes, in order to do the full set of things with the os_modules instead of our keystone module, we need 2.5 | 09:04 |
| odyssey4me | however, we don't need 2.5 if we can move the use of shade to use the ansible runtime venv and delegate all those tasks to the deploy host | 09:04 |
| evrardjp | odyssey4me: that's exactly my suggestion for source installs. | 09:05 |
| odyssey4me | evrardjp for all installs that will work best | 09:05 |
| *** alefra has joined #openstack-ansible | 09:05 | |
| odyssey4me | that's exactly what https://review.openstack.org/#/c/568142/ does, but it's failing the role tests and I've not yet worked out a solution | 09:06 |
| evrardjp | odyssey4me: I see the biggest problem with distro packages installs and not doing that. But indeed it's a win that can be done accross the board, and reduces code paths. | 09:06 |
| alefra | hi all, i've an issue with iscsi multipath with live migration.... it fails every time on command multipath -f mpathXX | 09:06 |
| evrardjp | odyssey4me: let me check real quick | 09:06 |
| odyssey4me | evrardjp yes, but I don't want to hold hwoarang's work back in the mean time, so I'd like to follow on with this work as an optimisation | 09:07 |
| alefra | it's a driver issue or cinder? | 09:07 |
| *** tux_ has quit IRC | 09:07 | |
| odyssey4me | evrardjp lemme see if I can unblock that patch - I'll fire up a test host and figure out a solution | 09:08 |
| odyssey4me | the issue, I'm quite sure, is that we force localhost not to use the venv so that the python-lxc library can be sourced... but in this case we want to use the runtime venv and not the host's python | 09:09 |
| odyssey4me | the solution here may be to not set the runtime interpreter for the lxc things as broadly, but instead to do it on the task that needs it | 09:10 |
| *** zul_ has joined #openstack-ansible | 09:11 | |
| evrardjp | odyssey4me: that patch can be split in two. | 09:11 |
| *** zul_ is now known as zul | 09:11 | |
| evrardjp | first, do the cleanup of the package list, moving to delegation, with the existing modules | 09:12 |
| odyssey4me | evrardjp yes, but I'd like it to work before I do that | 09:12 |
| evrardjp | then use the shade modules | 09:12 |
| *** Tahvok has joined #openstack-ansible | 09:12 | |
| Taseer | is the discussion above related to keystone's failure to retrieve the conf file from venv ? | 09:12 |
| odyssey4me | I was thinking that perhaps, just for role tests, I could make the playbook that preps the host install shade onto the host | 09:13 |
| odyssey4me | that'd be a very hacky workaround | 09:13 |
| evrardjp | Taseer: nope. | 09:13 |
| odyssey4me | Taseer nope, I've not seen that - where are you seeing that? | 09:13 |
| hwoarang | odyssey4me: is the shade venv thing just to cover the ubuntu bad shade version? | 09:13 |
| hwoarang | i mean, we could talk to uca people and ask them to add a distro shade. that would make things much simpler no/ | 09:13 |
| odyssey4me | hwoarang nope, it's to make sure that we don't need to install shad/keystone/whatever on the host at all | 09:14 |
| evrardjp | odyssey4me: nope I rather not. | 09:14 |
| hwoarang | i see | 09:14 |
| odyssey4me | we totaly get rid of the host install requirements | 09:14 |
| Taseer | odyssey4me: http://logs.openstack.org/88/571388/2/check/openstack-ansible-functional-ubuntu-xenial-nv/3e9a9c5/job-output.txt.gz#_2018-05-31_08_47_27_351844 | 09:14 |
| evrardjp | odyssey4me: I don't see the point, maybe you need to explain that to me very sloooooowly. | 09:14 |
| odyssey4me | hwoarang basically that would mean that for a source-based install, we only ever use a venv on the target host - anything ansible needs to use to register services or whatever can use its runtime venv or optionally the utility container | 09:15 |
| evrardjp | that sounds nice. | 09:15 |
| evrardjp | but that doesn't need a hackjob for it | 09:15 |
| evrardjp | let's just add shade into the bootstrap ansible requirements | 09:15 |
| evrardjp | (which is already done, right?) | 09:15 |
| odyssey4me | evrardjp the issue here is not the integrated build | 09:15 |
| odyssey4me | what I have there will work for that | 09:16 |
| evrardjp | just re-use the same thing :) | 09:16 |
| *** ArchiFleKs has joined #openstack-ansible | 09:16 | |
| odyssey4me | the issue is the role tests where we force the python executable to the host | 09:16 |
| evrardjp | odyssey4me: link? | 09:16 |
| odyssey4me | which means that delegating those tasks to localhost doesn't work, because the host doesn't have shade | 09:16 |
| evrardjp | it's an inventory thing right? | 09:16 |
| odyssey4me | https://github.com/openstack/openstack-ansible-os_glance/blob/master/tests/host_vars/localhost.yml#L20 | 09:17 |
| evrardjp | ok we are hitting the nerve now | 09:17 |
| evrardjp | or the core | 09:17 |
| evrardjp | I don't know how to say that :) | 09:17 |
| odyssey4me | we do that to make the lxc container stuff work, because it needs the lxc python lib (installed on the host) to make the containers | 09:17 |
| evrardjp | ok | 09:17 |
| odyssey4me | so, as I said earlier - I need to figure out a better solution for that, then it'll unblock this | 09:18 |
| evrardjp | that's fine for me | 09:18 |
| odyssey4me | a hack could be, for role tests only, to install shade on the host | 09:18 |
| evrardjp | here is what I think: | 09:18 |
| odyssey4me | but it's probably better to figure out a better solution for the lxc thing | 09:18 |
| *** mbuil has joined #openstack-ansible | 09:18 | |
| evrardjp | the problem is double: | 09:19 |
| evrardjp | the best would be to point to a venv as interpreter at all times, but change it on some specific cases. | 09:19 |
| evrardjp | (like lxc) | 09:19 |
| evrardjp | or the bootstrap | 09:19 |
| evrardjp | the hardest is that it's difficult to change live. | 09:20 |
| evrardjp | what if the ansible_python_interpreter is found from a fact, defaulting to /usr/bin/python2 ? | 09:21 |
| odyssey4me | we don't need to actually point to a venv interpreter if we do things right | 09:21 |
| odyssey4me | it's only in the special cases where that might need to be done | 09:21 |
| *** senk has quit IRC | 09:21 | |
| *** senk has joined #openstack-ansible | 09:21 | |
| evrardjp | Assuming the bootstrap of ansible doesn't require ansible, yes. | 09:21 |
| odyssey4me | every one of our 'required' pip packages across roles is only there for the registering of services or other ansible modules... so if we use delegation to locahost, ansible will automatically use the right interpreter from the runtime | 09:22 |
| odyssey4me | the only other packages required on the hosts are things that we can make ansible put there via distro packages (python-apt, etc) | 09:22 |
| evrardjp | I am not sure of that, but okay. | 09:22 |
| evrardjp | I trust you there | 09:22 |
| evrardjp | python-apt get auto installed | 09:23 |
| evrardjp | so it doesn't matter | 09:23 |
| evrardjp | oh yes it will | 09:23 |
| evrardjp | ok, question | 09:23 |
| evrardjp | should we, for the deploy node only, use site-packages? | 09:24 |
| evrardjp | so, being less isolated... | 09:24 |
| odyssey4me | evrardjp ideally not | 09:24 |
| odyssey4me | I'd like to only get virtualenv on to the deploy node, and everything else is in the venv | 09:24 |
| evrardjp | oh no that won't help on some cases. Will help on others. | 09:24 |
| odyssey4me | I mean, I'd like to be able to one day deploy OSA from my mac | 09:25 |
| evrardjp | docker! | 09:25 |
| evrardjp | :p | 09:25 |
| odyssey4me | that's still a cop-out to me | 09:25 |
| evrardjp | ok so... | 09:26 |
| odyssey4me | I should be able to build a runtime venv (in user space), prep the deploy config (in user space), and execute a deployment. | 09:26 |
| evrardjp | let me try something quick on a machine | 09:28 |
| odyssey4me | so, back to an earlier question - what's holding us back from upgrading to ansible 2.5 ? | 09:28 |
| * hwoarang notices that cinder tempest is broken | 09:28 | |
| hwoarang | the will to do it :D | 09:29 |
| *** cshen has joined #openstack-ansible | 09:29 | |
| hwoarang | i thought jmccrory was working on that | 09:29 |
| hwoarang | https://review.openstack.org/#/c/501814/ ? | 09:29 |
| odyssey4me | yeah, cinder tempest has been broken for a while, but I've not had time to figure out why | 09:29 |
| *** bhujay has joined #openstack-ansible | 09:29 | |
| *** bhujay has quit IRC | 09:31 | |
| evrardjp | jmccrory: said to me he will work on it back from the summit | 09:31 |
| *** bhujay has joined #openstack-ansible | 09:31 | |
| evrardjp | I can take it from there. | 09:32 |
| *** udesale has quit IRC | 09:32 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update Ansible to 2.5.3 https://review.openstack.org/501814 | 09:35 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: Update ansible to 2.5.3 https://review.openstack.org/571423 | 09:37 |
| odyssey4me | we need to get that ARA constraints patch through the door, otherwise master is still blocked for the integrated build | 09:37 |
| *** udesale has joined #openstack-ansible | 09:38 | |
| odyssey4me | I was wondering this morning - is cmd2 a direct dep of ARA, or is it pulled in from something else | 09:38 |
| odyssey4me | if from smoething else, if they release a new version of that with proper constraints then that would also unblock us | 09:39 |
| evrardjp | odyssey4me: cliff | 09:43 |
| *** jwitko has joined #openstack-ansible | 09:43 | |
| evrardjp | iirc | 09:43 |
| evrardjp | so basically all the CLI clients :p | 09:43 |
| *** jwitko_ has quit IRC | 09:44 | |
| evrardjp | http://git.openstack.org/cgit/openstack/cliff/tree/requirements.txt | 09:44 |
| odyssey4me | and there was a tag released with that updated req | 09:45 |
| odyssey4me | https://github.com/openstack/cliff/tree/2.12.0 | 09:45 |
| evrardjp | I planned to do a bump this week-end, and I need to send something on the ML about releases, due to my conversations in the summit. It's pending. | 09:47 |
| evrardjp | Sorry I got so many things to do, so little time :( | 09:47 |
| *** ArchiFleKs has quit IRC | 09:49 | |
| *** ArchiFleKs has joined #openstack-ansible | 09:50 | |
| bhujay | evrardjp: thanks for your suggestion. using lxc_cache_prep_timeout is a good option , but the default value of lxc_cache_prep_timeout is 1200 and /5 will make it lesser than the current default 300 for the prestage job. Do you think /3 shd b ok ? | 09:50 |
| evrardjp | bhujay: please //5 so that it's more aligned with what's ansible is doing | 09:51 |
| evrardjp | and you can up that for your use case | 09:52 |
| evrardjp | or | 09:52 |
| evrardjp | alternatively | 09:52 |
| evrardjp | bump that default value | 09:52 |
| evrardjp | to 1500 | 09:52 |
| evrardjp | odyssey4me: do we have a list of the python c bindings used in OSA on the host directly? So we can take a more systematic approach | 09:54 |
| bhujay | evrardjp: ok , option 1 is better , at least gives a flexibility to handle worst situation. | 09:54 |
| odyssey4me | evrardjp of course we don't :p | 09:54 |
| evrardjp | odyssey4me: and all the packages basically. | 09:54 |
| evrardjp | odyssey4me: should we start with that? | 09:55 |
| evrardjp | I can create a machine, see what's included in a venv with ansible, then do some fun stuff with it, see what happens at the end. | 09:55 |
| evrardjp | diff the two | 09:55 |
| odyssey4me | evrardjp well, my feeling is that if we eliminate any host pip installs we find systematically, then we'll find those as we go | 09:55 |
| odyssey4me | evrardjp no need for you to spend time on that, I'm working on it | 09:56 |
| odyssey4me | it seems likely that this cycle will mostly be me stabilising the existing changes to the python builds (I'm still cleaning up some things in Queens as I find brokeness) and laying the ground work to use the dedicated build role | 09:57 |
| odyssey4me | I'm not sure that it'll all be done in time for m3, but I'll push for that. | 09:58 |
| odyssey4me | The distro-package based installs, though, really helps clean up the line between source builds and non-source builds... so I'd like to help get that done and clean up a bit behind it. | 09:58 |
| evrardjp | I think it's fair to all work on that. | 09:59 |
| evrardjp | that is a big thing. | 09:59 |
| hwoarang | i am doing swift now. then only nova and neutron are left which are somewhat complex | 09:59 |
| hwoarang | odyssey4me: why m3? is that a hard deadline to get stuff in ? | 09:59 |
| odyssey4me | hwoarang we're supposed to be done with features at m3, and then turn focus on bug fixing, upgrade issue solving, etc | 10:00 |
| hwoarang | ok | 10:01 |
| odyssey4me | Taseer the issue you're hitting there is that your tox config in os_blazar is migging the USER env var, like this: https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tox.ini#L17 | 10:01 |
| odyssey4me | hwoarang exceptions can be made, but that's the standard openstack rule around where to focus for each milestone | 10:02 |
| odyssey4me | and given the state of our bugs and upgrade tests, I think we should spend the last milestone focused on fixing what we broke :p | 10:02 |
| jrosser | if anyone has a minute to review these i'd be grateful https://review.openstack.org/#/c/568276/ and https://review.openstack.org/#/c/570456/ | 10:05 |
| *** radeks__ has joined #openstack-ansible | 10:11 | |
| *** cshen has quit IRC | 10:13 | |
| *** radeks__ has quit IRC | 10:17 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server stable/queens: [TEST] do not use ipv6 for nginx reverse proxy https://review.openstack.org/571434 | 10:17 |
| odyssey4me | newton is finally eol, long live newton! | 10:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Ensure that repo checks use the right values https://review.openstack.org/564422 | 10:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Ensure that the repo servers also use the pypi cache https://review.openstack.org/568502 | 10:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Remove default pip_links value https://review.openstack.org/568637 | 10:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: [TEST] do not use ipv6 for nginx reverse proxy https://review.openstack.org/571436 | 10:20 |
| *** radeks__ has joined #openstack-ansible | 10:30 | |
| *** radeks_ has joined #openstack-ansible | 10:32 | |
| *** nicolasbock has joined #openstack-ansible | 10:33 | |
| *** armaan has joined #openstack-ansible | 10:33 | |
| *** radeks__ has quit IRC | 10:35 | |
| openstackgerrit | bhujay kumar proposed openstack/openstack-ansible-lxc_hosts master: making lxc image prestage timeout configurable https://review.openstack.org/571405 | 10:37 |
| evrardjp | odyssey4me: :) | 10:37 |
| evrardjp | the end of an era | 10:38 |
| odyssey4me | evrardjp at the PTG we should probably discuss the LTS model and figure out what we're going to do about it | 10:38 |
| odyssey4me | it might make sense for us to align with what upstream is doing | 10:38 |
| evrardjp | it's up to us | 10:40 |
| evrardjp | I think it's fine doing what upstream is doing | 10:40 |
| evrardjp | let the branch open | 10:40 |
| evrardjp | accept patches, but not actively work on it | 10:40 |
| evrardjp | that's already what we are doing effectively | 10:40 |
| evrardjp | IIRC | 10:41 |
| evrardjp | odyssey4me: FYI I just tested something for delegation and things like that | 10:42 |
| evrardjp | long story short, we don't need an ansible_python_interpreter on localhost. | 10:42 |
| evrardjp | everywhere else, we should point to a venv that gets all the libs symlinked as we do now | 10:43 |
| evrardjp | if you want fully self-contained | 10:43 |
| evrardjp | if we plan to delegate to something else than localhost, that's something we have to think about | 10:44 |
| odyssey4me | evrardjp yes, I'm just reading now that if localhost is in the inventory, it implies that it should use the host's python interpreter | 10:44 |
| odyssey4me | so in this case because I'm delegating, I need to give it the python interpreter path | 10:45 |
| evrardjp | host ansible interpreter | 10:45 |
| evrardjp | even when delegating | 10:45 |
| evrardjp | to what I tested | 10:45 |
| odyssey4me | if you use connection: local, it'll use the venv interpreter | 10:46 |
| odyssey4me | if you use a normal ssh connection, it'll use the host | 10:46 |
| evrardjp | nope | 10:46 |
| odyssey4me | interpreter | 10:46 |
| evrardjp | I just tried it | 10:46 |
| evrardjp | :) | 10:46 |
| evrardjp | can share my results if you like | 10:46 |
| evrardjp | but anyway | 10:46 |
| odyssey4me | wtf, that means it changed yet again | 10:46 |
| evrardjp | it's a slippery slope to not be explicit | 10:46 |
| evrardjp | odyssey4me: yes I remember someone talking about that in #ansible | 10:47 |
| evrardjp | being explicit is good | 10:47 |
| *** klamath has joined #openstack-ansible | 10:47 | |
| evrardjp | but if we do that | 10:47 |
| evrardjp | we lose python-apt | 10:47 |
| evrardjp | so we have to symlink it | 10:47 |
| evrardjp | but that works, I just did it | 10:47 |
| evrardjp | the current grep is not good, but it's generally fine | 10:47 |
| odyssey4me | we used to do that, ugh | 10:48 |
| evrardjp | so it fully depends on how far we want to go | 10:48 |
| Taseer | odyssey4me: thanks | 10:49 |
| odyssey4me | evrardjp so how do I specify the interpreter to use (which is the venv) without knowing the venv's path? | 10:50 |
| odyssey4me | I'm specifically thinking of the tox use-case here, where the path is not fixed | 10:50 |
| evrardjp | what do you mean by "without knowing the venv path" ? | 10:50 |
| odyssey4me | for the integrated build, I know the path because we fix it to /opt/ansible-runtime | 10:51 |
| evrardjp | yes | 10:51 |
| *** klamath has quit IRC | 10:51 | |
| odyssey4me | for role tests it's something like <repo dir>/.tox/<tox env name>/bin/python2 | 10:51 |
| evrardjp | for tox runs it's because we are using localhost in inventory right? | 10:51 |
| evrardjp | if we didn't have localhost in inventory, it would be implicit, and we could know it by default | 10:52 |
| odyssey4me | we have localhost in the inventory because we have settings to apply to it | 10:52 |
| odyssey4me | lxc bridges and such | 10:52 |
| evrardjp | let me check if that's possible with hostvars without setting it | 10:52 |
| odyssey4me | I've tried taking the ansible_python_interpreter out, but that doesn't work | 10:53 |
| odyssey4me | I tried with ansible_python_interpreter: "/usr/bin/env python2" but that doesn't work either | 10:53 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: This patch ensures that the functional tests for Blazar pass. https://review.openstack.org/571388 | 10:54 |
| bhujay | pip install during openstack-setup-host play fails due to pip.conf pointing to repo server which is not yet built. is this change , https://review.openstack.org/568637 , going to address that issue ? | 10:55 |
| evrardjp | odyssey4me: yeah, adding vars for something that happens on localhost is a bad idea | 10:55 |
| evrardjp | maybe we should have a different hostname | 10:55 |
| evrardjp | aio1 | 10:55 |
| evrardjp | ansible_connection=local | 10:55 |
| *** radeks_ has quit IRC | 10:55 | |
| evrardjp | so we do a clean separation of what's on host and what's on ansible venv | 10:55 |
| bhujay | odyssey4me: ^ | 10:55 |
| odyssey4me | bhujay that's already merged in master | 10:55 |
| bhujay | ok , thanks | 10:56 |
| evrardjp | but in the meantime, odyssey4me , we can probably use tox to do that for us | 10:56 |
| odyssey4me | bhujay there are already things in master which verify whether the repo server is there and adjust accordingly | 10:56 |
| evrardjp | odyssey4me: tox knows the location of the venv. He can sed the file to adapt it :p | 10:56 |
| bhujay | ok , thank you | 10:56 |
| odyssey4me | evrardjp yuk | 10:57 |
| odyssey4me | but there is an env var, so perhaps I can use that - lemme have a go | 10:57 |
| evrardjp | how come? | 10:57 |
| odyssey4me | hard setting it to the tox python2 definitely worked | 10:57 |
| evrardjp | yeah I think that would be fine, but... | 10:57 |
| evrardjp | lookup('ENV', ... not gonna work :p | 10:58 |
| evrardjp | that's like chicken and egg thing | 10:58 |
| evrardjp | oh no it's localhost. | 10:58 |
| evrardjp | it could work. | 10:58 |
| evrardjp | interesting. | 10:58 |
| odyssey4me | evrardjp lemme figure out something and propose a patch, then you can faff :p | 10:58 |
| odyssey4me | I was hoping that there was some sort of magic var which we could use on the task | 10:59 |
| odyssey4me | ok, so here we have it | 11:00 |
| odyssey4me | if we universally use the ansible runtime venv, then the apt module fails due to the missing python apt bits... and IIRC those have C bindings so we'd have to symlink.... or we'd have to ensure we always set the interpreter for those tasks | 11:01 |
| *** jwitko has quit IRC | 11:01 | |
| *** markvoelker_ has joined #openstack-ansible | 11:04 | |
| odyssey4me | we could use a lint test to ensure that all package manager tasks always have the python interpreter set, but that's a lot of code churn for very little gain | 11:04 |
| evrardjp | odyssey4me: yes that's what I just told you for the apt thing | 11:04 |
| evrardjp | it works | 11:04 |
| odyssey4me | what works? you're giving me mixed messages here | 11:04 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-lxc_container_create master: Use attribute for sort https://review.openstack.org/571443 | 11:05 |
| andymccr | ^ fixed | 11:05 |
| odyssey4me | I know the symlinking works, but I'd rather avoid that for the ansible runtime venv. | 11:05 |
| Tahvok | Guys | 11:05 |
| andymccr | the disparity is that the role test uses py2, the gate uses py3 | 11:05 |
| Tahvok | Sorry, wrong channel | 11:05 |
| odyssey4me | andymccr orly? | 11:05 |
| andymccr | seems to be the case. the error is basically that in py3 they got strict and wont allow comparisons between dicts for ordering (so you cant just order a list of dicts) | 11:05 |
| andymccr | py2 allows it, although not 100% sure how it decides what comes first | 11:06 |
| odyssey4me | lol, ok, so this is actually better anyway | 11:06 |
| *** markvoelker has quit IRC | 11:06 | |
| andymccr | yeah | 11:06 |
| evrardjp | yeah | 11:06 |
| andymccr | pretty much | 11:06 |
| andymccr | also interface has to be defined based on the template | 11:06 |
| odyssey4me | sort me a dict... OK, YOLO! | 11:06 |
| andymccr | my personal favourite would be to pass a list of container_networks rather than a dict but that would mean doing some larger scale hackery in all sorts of places | 11:07 |
| evrardjp | well... | 11:07 |
| evrardjp | yeah | 11:07 |
| evrardjp | let's just work with this rn | 11:08 |
| evrardjp | odyssey4me: so what do you not like with the symlinking? | 11:08 |
| evrardjp | you'd prefer to use interpreter for package tasks? | 11:08 |
| evrardjp | I think that's more intrusive | 11:09 |
| *** blinkiz has quit IRC | 11:11 | |
| *** markvoelker has joined #openstack-ansible | 11:14 | |
| *** markvoelker_ has quit IRC | 11:15 | |
| odyssey4me | so virtualenv sets VIRTUAL_ENV, and tox has that var as {envdir} | 11:15 |
| odyssey4me | so it could be useful to use that on the applicable tasks | 11:16 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] Test profiling ansible tasks https://review.openstack.org/570993 | 11:17 |
| evrardjp | I am not sure that explicit for those task is very user-friendly | 11:19 |
| evrardjp | let me make sure I understand what you're saying: You're saying that all the tasks doing apt work will need to be set an interpreter separately right? | 11:19 |
| odyssey4me | evrardjp no, I'm saying that we assume using the host interpreter by default - but for tasks that should use the runtime venv, we set those explicitly | 11:20 |
| evrardjp | or you mean the other way around: All the task that won't be using sytem interpreter will require a ansible_python_interpreter (in this case the tasks who gets delegated for shade and others) | 11:21 |
| evrardjp | ok | 11:21 |
| evrardjp | that's somethign I am fine with. | 11:21 |
| odyssey4me | busy testing that now | 11:21 |
| *** bhujay has quit IRC | 11:29 | |
| *** bhujay has joined #openstack-ansible | 11:30 | |
| *** bhujay has quit IRC | 11:31 | |
| *** bhujay has joined #openstack-ansible | 11:32 | |
| *** udesale has quit IRC | 11:33 | |
| *** chyka has joined #openstack-ansible | 11:33 | |
| olivierbourdon38 | Quick question: in https://docs.openstack.org/project-deploy-guide/openstack-ansible/latest/targethosts.html it is said that ntp needs to be installed and configured on target hosts. However, I found out that after deployment is done, ntp is not running anymore but chronyd instead (at least on my Ubuntu Xenial deployment). Is this expected ? what are the reasons to replace one by the other ? many thanks as usual | 11:34 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: This patch ensures that the functional tests for Blazar pass. https://review.openstack.org/571388 | 11:35 |
| *** udesale has joined #openstack-ansible | 11:37 | |
| *** dave-mccowan has joined #openstack-ansible | 11:37 | |
| *** chyka has quit IRC | 11:38 | |
| *** dave-mcc_ has joined #openstack-ansible | 11:40 | |
| *** dave-mccowan has quit IRC | 11:43 | |
| *** ansmith has quit IRC | 11:43 | |
| *** bhujay has quit IRC | 11:46 | |
| *** udesale has quit IRC | 11:47 | |
| *** bhujay has joined #openstack-ansible | 11:48 | |
| *** bhujay has quit IRC | 11:48 | |
| *** armaan has quit IRC | 11:50 | |
| *** armaan has joined #openstack-ansible | 11:50 | |
| *** udesale has joined #openstack-ansible | 11:51 | |
| openstackgerrit | Chason Chan proposed openstack/openstack-ansible-os_aodh master: Remove keystone_authtoken section for aodh https://review.openstack.org/571402 | 11:53 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Use attribute for sort https://review.openstack.org/571443 | 11:54 |
| odyssey4me | evrardjp well, here's some fun | 11:58 |
| odyssey4me | it turns out that ansible_python_interpreter is interpreted literally - you cannot use jinja in the value | 11:58 |
| *** udesale has quit IRC | 12:03 | |
| *** udesale has joined #openstack-ansible | 12:03 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: This patch ensures that the functional tests for Blazar pass. https://review.openstack.org/571388 | 12:06 |
| *** udesale_ has joined #openstack-ansible | 12:07 | |
| odyssey4me | and this is a known issue which apparently was fixed: https://github.com/ansible/ansible/issues/18665 | 12:09 |
| *** udesale has quit IRC | 12:09 | |
| hwoarang | hmm | 12:11 |
| hwoarang | i guess it was fixed in 2.4 but didn't make it to the release? | 12:11 |
| odyssey4me | checking that now | 12:11 |
| hwoarang | "Merged to devel and cherrypicked to stable-2.4 for the 2.4.3beta2 release." | 12:11 |
| hwoarang | so it should be in 2.4.4 :( | 12:11 |
| *** armaan has quit IRC | 12:13 | |
| odyssey4me | hmm, need to find that PR because I'm not seeing it in stable-2.4 | 12:14 |
| *** armaan has joined #openstack-ansible | 12:14 | |
| hwoarang | cores: i'd appreciate reviews on https://review.openstack.org/#/c/566908/ which should make opensuse use the infra mirrors. will speed things up a bit | 12:16 |
| hwoarang | thanks! | 12:16 |
| *** brad[] has quit IRC | 12:16 | |
| odyssey4me | hwoarang question in there for you? | 12:18 |
| odyssey4me | (in review) | 12:18 |
| hwoarang | let me try | 12:22 |
| hwoarang | actually... hold on | 12:22 |
| odyssey4me | yep, it turns out that this is in ansible 2.5 - but not ansible 2.4 | 12:23 |
| evrardjp | ok let's move to 2.5 :) | 12:23 |
| hwoarang | odyssey4me: because of https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/defaults/main.yml#L145 the obs url should inherit the opensuse_mirror right? | 12:24 |
| odyssey4me | Id love to, but I'd like to understand what the hold up is | 12:24 |
| hwoarang | but apparently it doesn't work | 12:24 |
| evrardjp | no it works I think | 12:24 |
| *** prasen has joined #openstack-ansible | 12:24 | |
| evrardjp | I think it was blocked by gate flakiness | 12:25 |
| parasitid | hi | 12:26 |
| parasitid | odyssey4me: i've patched osa-rabbitmq_server | 12:26 |
| parasitid | it's been merged | 12:26 |
| parasitid | now what's the processs to take it into account in osa ? | 12:27 |
| odyssey4me | parasitid when we next do a role sha bump, that patch will come into the integrated build - it then gets tested for two weeks before it's released | 12:27 |
| parasitid | ok | 12:27 |
| odyssey4me | there should be a release shortly for the current batch, so the role sha bump happens immediately after that | 12:27 |
| evrardjp | odyssey4me: I might have missed last release due to summit. | 12:27 |
| parasitid | so i should propose a patch with the new commit id | 12:27 |
| evrardjp | I am still catching up | 12:27 |
| parasitid | great | 12:28 |
| parasitid | thansk | 12:28 |
| *** udesale_ has quit IRC | 12:29 | |
| *** udesale_ has joined #openstack-ansible | 12:30 | |
| odyssey4me | interesting, it looks from https://review.openstack.org/501814 that the master build got past the ara install, so perhaps we're unblocked there | 12:30 |
| odyssey4me | lemme rebase all those patches on top of https://review.openstack.org/570914 | 12:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Use ARA instead of profile_tasks callback https://review.openstack.org/546271 | 12:32 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove rally_git_* overrides https://review.openstack.org/570906 | 12:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Adds the certificate client key password to user_secrets https://review.openstack.org/558244 | 12:34 |
| *** chhagarw has joined #openstack-ansible | 12:34 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Fixed typo in variable oslomsg_notify_transport https://review.openstack.org/571156 | 12:34 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Fixed variable name notify_vhost https://review.openstack.org/571164 | 12:35 |
| evrardjp | odyssey4me: last time we were stuck lxc | 12:35 |
| evrardjp | thanks to andymccr :) | 12:35 |
| evrardjp | hahah | 12:35 |
| odyssey4me | yup, and that's merged now | 12:35 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: common-tasks: nodepool: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/566908 | 12:36 |
| evrardjp | odyssey4me: I know, that's why I rechecked :) | 12:36 |
| mgariepy | morning everyone | 12:37 |
| evrardjp | cool right? | 12:37 |
| evrardjp | good morning mgariepy | 12:37 |
| odyssey4me | evrardjp dammit, you beat me to it :p | 12:37 |
| evrardjp | odyssey4me: lol, for once. | 12:37 |
| evrardjp | hwoarang: why do you add lxc_container_cache_file again? | 12:37 |
| evrardjp | hwoarang: I mean you added it in the commit message, but I don't understand that message. | 12:38 |
| evrardjp | I thought we were using host data there. | 12:38 |
| hwoarang | evrardjp: i added it so we can copy /etc/ci/mirror_info.sh | 12:39 |
| hwoarang | to the container | 12:39 |
| hwoarang | so the container can find out what this the nodepool mirror | 12:39 |
| hwoarang | to use it in the openstack_hosts | 12:39 |
| odyssey4me | no need - commented in review | 12:40 |
| *** schwicht has joined #openstack-ansible | 12:40 | |
| hwoarang | hmm | 12:41 |
| hwoarang | for whatever reason the opensuse_mirror fact is not propagated to containers | 12:41 |
| odyssey4me | oh, heh - Ensure ci data exists before setting variables. is not delegated | 12:41 |
| hwoarang | something is fishy again | 12:41 |
| hwoarang | right | 12:41 |
| hwoarang | catch-22 ;p | 12:42 |
| hwoarang | let me update | 12:42 |
| evrardjp | when is testinfra when you need it :) | 12:42 |
| evrardjp | where* | 12:42 |
| hwoarang | what is testinfra... | 12:43 |
| odyssey4me | evrardjp's new favorite troll | 12:43 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: common-tasks: nodepool: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/566908 | 12:43 |
| evrardjp | something I'd like to see in S and above. And yes my troll :p | 12:43 |
| evrardjp | well | 12:43 |
| evrardjp | not "my" | 12:43 |
| hwoarang | is that a thing? | 12:44 |
| evrardjp | https://testinfra.readthedocs.io/en/latest/ | 12:44 |
| evrardjp | hwoarang: a day I will have time to make it a thing. | 12:44 |
| evrardjp | :D | 12:44 |
| hwoarang | lol unit test for infra | 12:44 |
| *** udesale has joined #openstack-ansible | 12:45 | |
| odyssey4me | somethign we should discuss at the PTG, is whether we persist with using tempest for functional tests, or whether we switch to using something like molecule which brings a test framework that's easier for us to understand and use | 12:45 |
| *** udesale_ has quit IRC | 12:45 | |
| odyssey4me | tempest can still be there, but only be used for refstack testing really - rather than general role tests | 12:46 |
| *** prasen has quit IRC | 12:48 | |
| *** prasen has joined #openstack-ansible | 12:49 | |
| dmsimard | odyssey4me: hey is Hugh still around ? Haven't seen him in a bit | 13:00 |
| *** ansmith has joined #openstack-ansible | 13:03 | |
| odyssey4me | dmsimard you mean hughsaunders? | 13:04 |
| dmsimard | odyssey4me: yeah | 13:04 |
| odyssey4me | he's not in this channel, but I can get him to be - I just don't think he's on IRC much any more as his primary work is different now | 13:05 |
| *** Taseer has quit IRC | 13:05 | |
| dmsimard | odyssey4me: ah, I guess that's why I don't see him around as much -- there's someone who is using the junit export implementation he contributed to ara and had questions :) | 13:05 |
| odyssey4me | ah, is there an ara IRC channel? I'll ask hugh to join there | 13:06 |
| *** Taseer has joined #openstack-ansible | 13:06 | |
| evrardjp | #ara | 13:11 |
| evrardjp | odyssey4me: ^ | 13:11 |
| *** shardy has quit IRC | 13:11 | |
| odyssey4me | ok, I've forwarded the message | 13:11 |
| *** armaan has quit IRC | 13:11 | |
| *** shardy has joined #openstack-ansible | 13:12 | |
| *** armaan has joined #openstack-ansible | 13:12 | |
| *** hughsaunders has joined #openstack-ansible | 13:15 | |
| *** armaan has quit IRC | 13:17 | |
| cloudnull | mornings | 13:19 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: defaults: Do not install the cinderclient package https://review.openstack.org/571474 | 13:20 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Add support for using distribution packages for OpenStack services https://review.openstack.org/571475 | 13:20 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Do not pip install keystoneclient on the host https://review.openstack.org/571476 | 13:20 |
| *** tux_ has joined #openstack-ansible | 13:22 | |
| *** armaan has joined #openstack-ansible | 13:23 | |
| *** tux__ has joined #openstack-ansible | 13:24 | |
| *** tux_ has quit IRC | 13:28 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_nova master: Revert "Switch CentOS to non-voting" https://review.openstack.org/571483 | 13:33 |
| *** udesale_ has joined #openstack-ansible | 13:34 | |
| *** markvoelker_ has joined #openstack-ansible | 13:35 | |
| *** sar has quit IRC | 13:36 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance master: Execute service setup against a delegated host using Ansible built-in modules https://review.openstack.org/568142 | 13:36 |
| *** markvoelker has quit IRC | 13:36 | |
| *** klamath has joined #openstack-ansible | 13:36 | |
| evrardjp | ahah we broke things again andymccr | 13:36 |
| andymccr | hmm | 13:36 |
| andymccr | how so? | 13:37 |
| odyssey4me | evrardjp ^ that works, but only with ansible 2.5 | 13:37 |
| evrardjp | http://logs.openstack.org/23/571423/1/check/openstack-ansible-functional-opensuse-423/79f2851/job-output.txt.gz#_2018-05-31_10_07_22_152731 | 13:37 |
| *** udesale has quit IRC | 13:37 | |
| evrardjp | isn't that the same task as before? | 13:38 |
| *** markvoelker has joined #openstack-ansible | 13:38 | |
| *** armaan has quit IRC | 13:38 | |
| andymccr | nope | 13:39 |
| andymccr | before was network interfaces | 13:39 |
| andymccr | this is bind mounts | 13:39 |
| *** nurdie has joined #openstack-ansible | 13:39 | |
| *** markvoelker_ has quit IRC | 13:40 | |
| *** sep has quit IRC | 13:43 | |
| evrardjp | fun stuff | 13:43 |
| hwoarang | <opnfv hat on> the "ERROR! Unexpected templating type error occurred on ({{ lxc_container_networks_combined.values() | list | sort }}): unorderable types: dict() < dict()" is not fixed yet right? We are looking to move to OSA rocky from queens </opnfv hat off> | 13:43 |
| evrardjp | hwoarang: it just merged | 13:44 |
| hwoarang | ah ok | 13:44 |
| evrardjp | andymccr: my bad, so that's probably an ansible 2.5 issue then | 13:44 |
| evrardjp | I will dig into it. | 13:44 |
| andymccr | hwoarang: should be fixed | 13:44 |
| hwoarang | ok i will try gain | 13:45 |
| andymccr | ahh that is yeah - it'd say sort(attribute='interface') if it was the newer version | 13:45 |
| *** alefra has quit IRC | 13:46 | |
| *** esberglu has joined #openstack-ansible | 13:50 | |
| *** throwsb1 has joined #openstack-ansible | 13:50 | |
| *** markvoelker has quit IRC | 13:52 | |
| *** admin0 has quit IRC | 13:52 | |
| *** tux__ has quit IRC | 13:53 | |
| *** tux_ has joined #openstack-ansible | 13:53 | |
| *** lemouchon has quit IRC | 13:54 | |
| *** tux_ has quit IRC | 13:58 | |
| *** nurdie has quit IRC | 13:58 | |
| *** admin0 has joined #openstack-ansible | 14:00 | |
| *** brad[] has joined #openstack-ansible | 14:01 | |
| *** kstev has joined #openstack-ansible | 14:06 | |
| *** sep has joined #openstack-ansible | 14:08 | |
| *** openstacking_123 has joined #openstack-ansible | 14:09 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Add support for using distribution packages for OpenStack services https://review.openstack.org/571475 | 14:18 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Do not pip install keystoneclient on the host https://review.openstack.org/571476 | 14:18 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: common-tasks: nodepool: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/566908 | 14:20 |
| hwoarang | evrardjp: ^ changed it to physical_host. can i have your +2 now? :) | 14:20 |
| hwoarang | (assuming that it works :/) | 14:20 |
| odyssey4me | lol, yeah - I'm not sure that all the inventories in the roles have that populated - hopefully they do | 14:21 |
| hwoarang | if that fails i can change that to localhost. i dont like changing too many things at once :) | 14:26 |
| *** tux_ has joined #openstack-ansible | 14:26 | |
| *** prasen has quit IRC | 14:27 | |
| *** hughsaunders has quit IRC | 14:28 | |
| evrardjp | did I not give the +2 already? Darn it was a nit :) | 14:31 |
| *** vadim has joined #openstack-ansible | 14:37 | |
| *** nurdie has joined #openstack-ansible | 14:39 | |
| *** tux_ has quit IRC | 14:41 | |
| *** tux_ has joined #openstack-ansible | 14:44 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/queens: Use the distro package for libvirt-python https://review.openstack.org/571404 | 14:45 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/ocata: Use the distro package for libvirt-python https://review.openstack.org/571407 | 14:45 |
| hwoarang | evrardjp: didn't work, can i switch to localhost now? | 14:45 |
| *** armaan has joined #openstack-ansible | 14:46 | |
| *** senk has quit IRC | 14:46 | |
| hwoarang | ok switching ;p | 14:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Use the distro package for libvirt-python https://review.openstack.org/571406 | 14:47 |
| evrardjp | didn't work? | 14:48 |
| evrardjp | interesting | 14:48 |
| *** nurdie_ has joined #openstack-ansible | 14:50 | |
| *** nurdie has quit IRC | 14:50 | |
| hwoarang | yep | 14:50 |
| evrardjp | hwoarang: want me to revert to ps3? | 14:50 |
| hwoarang | yeah i tried to do that in command line but it doesn't allow me | 14:51 |
| *** nurdie_ is now known as nurdie | 14:51 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: common-tasks: nodepool: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/566908 | 14:51 |
| evrardjp | it worked just fine :) | 14:51 |
| evrardjp | git review -d blabla,3 ; git commit --amend; git review -f | 14:52 |
| hwoarang | i tried to do it the 'git way' | 14:53 |
| hwoarang | git reset --hard bla bla | 14:53 |
| *** kstev has quit IRC | 14:53 | |
| * hwoarang needs to learn gerrit cmdline | 14:53 | |
| hwoarang | wow | 14:54 |
| hwoarang | it even restores votes | 14:54 |
| evrardjp | lol | 14:54 |
| evrardjp | no | 14:54 |
| evrardjp | but I love your joke :) | 14:54 |
| hwoarang | but | 14:54 |
| hwoarang | it even has +2 from jesse | 14:54 |
| evrardjp | because jesse is blazing fast? | 14:54 |
| hwoarang | logs say he didn't vote again | 14:54 |
| evrardjp | for real? | 14:55 |
| evrardjp | wow that's cool then | 14:55 |
| hwoarang | well i can't see his vote after you restored p3 | 14:55 |
| evrardjp | TIL :) | 14:55 |
| * hwoarang pats gerrit | 14:55 | |
| evrardjp | that's interesting | 14:56 |
| evrardjp | more gerrit everywhere! | 14:56 |
| *** SimAloo has joined #openstack-ansible | 14:59 | |
| *** kstev has joined #openstack-ansible | 15:17 | |
| *** openstacking_123 has quit IRC | 15:22 | |
| *** gyee has joined #openstack-ansible | 15:26 | |
| *** pcaruana has quit IRC | 15:28 | |
| *** tux_ has quit IRC | 15:31 | |
| *** tux_ has joined #openstack-ansible | 15:33 | |
| *** chyka has joined #openstack-ansible | 15:34 | |
| *** udesale_ has quit IRC | 15:40 | |
| *** jeferobles has joined #openstack-ansible | 15:40 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Use upper constraints when installing ARA https://review.openstack.org/570914 | 15:42 |
| odyssey4me | hwoarang evrardjp yeah, the vote stays if the patch is the same as a previous patch set (other than changes in the commit for author and some other parts) | 15:43 |
| odyssey4me | yay, looks like master is finally unwedged | 15:44 |
| *** hasseniu_ has quit IRC | 15:48 | |
| *** hassenius has joined #openstack-ansible | 15:48 | |
| *** jeferobles has quit IRC | 15:49 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Add support for using distribution packages for OpenStack services https://review.openstack.org/571475 | 15:49 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_swift master: Do not pip install keystoneclient on the host https://review.openstack.org/571476 | 15:49 |
| *** pvradu has quit IRC | 15:51 | |
| *** hassenius has quit IRC | 16:01 | |
| *** hassenius has joined #openstack-ansible | 16:03 | |
| *** chyka_ has joined #openstack-ansible | 16:04 | |
| *** chyka has quit IRC | 16:07 | |
| *** markvoelker has joined #openstack-ansible | 16:11 | |
| hwoarang | would anyone mind if we do an sha bump fro ansible 2.4 until we get 2.5 in ? | 16:12 |
| odyssey4me | hwoarang nope, I think we should do one if there's something you need in there | 16:13 |
| odyssey4me | and for master especially that's perfectly fine | 16:13 |
| hwoarang | ok | 16:14 |
| *** openstacking_123 has joined #openstack-ansible | 16:14 | |
| *** hassenius has quit IRC | 16:16 | |
| *** hughsaunders has joined #openstack-ansible | 16:17 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: bootstrap-ansible.sh: Bump SHA for Ansible 2.4 https://review.openstack.org/571516 | 16:17 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: test-ansible-deps: Bump SHA for Ansible 2.4 https://review.openstack.org/571518 | 16:20 |
| *** openstacking_123 has quit IRC | 16:21 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: test-ansible-deps: Bump SHA for Ansible 2.4 https://review.openstack.org/571518 | 16:21 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: bootstrap-ansible.sh: Bump SHA for Ansible 2.4 https://review.openstack.org/571516 | 16:22 |
| *** schwicht has quit IRC | 16:24 | |
| *** markvoelker has quit IRC | 16:24 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: test-ansible-deps: Bump SHA for Ansible 2.4 https://review.openstack.org/571518 | 16:30 |
| *** jra has quit IRC | 16:31 | |
| *** jra has joined #openstack-ansible | 16:31 | |
| *** pvradu has joined #openstack-ansible | 16:35 | |
| *** rpittau has quit IRC | 16:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Public n/w access for vxlan tenant n/w in ml2.ovs https://review.openstack.org/567809 | 16:37 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: common-tasks: nodepool: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/566908 | 16:39 |
| *** schwicht has joined #openstack-ansible | 16:41 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: tests: bootstrap-host: Set openSUSE OBS mirror for OpenStack CI https://review.openstack.org/571523 | 16:42 |
| *** armaan has quit IRC | 16:47 | |
| *** jra has quit IRC | 16:47 | |
| *** armaan has joined #openstack-ansible | 16:48 | |
| *** jra has joined #openstack-ansible | 16:48 | |
| *** pvradu has quit IRC | 16:52 | |
| *** spine55 has joined #openstack-ansible | 16:57 | |
| *** spine55 has quit IRC | 16:58 | |
| *** spine55 has joined #openstack-ansible | 16:58 | |
| *** prasen has joined #openstack-ansible | 17:02 | |
| *** pathcl has joined #openstack-ansible | 17:09 | |
| *** cjloader has joined #openstack-ansible | 17:09 | |
| pathcl | hello out here ! I'm facing issues regarding AIO: https://pastebin.com/AFNAwDHP | 17:10 |
| pathcl | any help will be appreciate | 17:10 |
| pathcl | s/appreciate/appreciated | 17:10 |
| pathcl | I followed this doc: https://docs.openstack.org/openstack-ansible/newton/developer-docs/quickstart-aio.html | 17:11 |
| *** klamath_ has joined #openstack-ansible | 17:11 | |
| *** klamath__ has joined #openstack-ansible | 17:12 | |
| *** hassenius has joined #openstack-ansible | 17:12 | |
| odyssey4me | pathcl newton is EOL and broken | 17:13 |
| odyssey4me | use queens: https://docs.openstack.org/openstack-ansible/queens/user/aio/quickstart.html | 17:14 |
| odyssey4me | newton needs several workarounds due to more recent releases in various libraries | 17:14 |
| *** klamath has quit IRC | 17:14 | |
| odyssey4me | some are documented in the release notes for newton | 17:14 |
| *** klamath_ has quit IRC | 17:15 | |
| *** jonher has joined #openstack-ansible | 17:16 | |
| pathcl | odyssey4me: thank you ! I happened to noticed the same and right now Im using latest :) | 17:16 |
| *** hassenius has quit IRC | 17:16 | |
| *** jonher_ has quit IRC | 17:17 | |
| *** klamath__ has quit IRC | 17:18 | |
| pathcl | s/noticed/notice | 17:20 |
| *** klamath has joined #openstack-ansible | 17:22 | |
| *** shardy has quit IRC | 17:23 | |
| *** pvradu has joined #openstack-ansible | 17:24 | |
| *** klamath_ has joined #openstack-ansible | 17:24 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Adds the certificate client key password to user_secrets https://review.openstack.org/558244 | 17:26 |
| openstackgerrit | Merged openstack/openstack-ansible master: Use ARA instead of profile_tasks callback https://review.openstack.org/546271 | 17:26 |
| *** klamath has quit IRC | 17:28 | |
| *** klamath has joined #openstack-ansible | 17:29 | |
| *** klamath_ has quit IRC | 17:31 | |
| *** klamath_ has joined #openstack-ansible | 17:32 | |
| *** klamath has quit IRC | 17:35 | |
| *** admin0 has quit IRC | 17:36 | |
| *** admin0 has joined #openstack-ansible | 17:36 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Allow special purpose container interfaces https://review.openstack.org/570456 | 17:37 |
| *** markvoelker has joined #openstack-ansible | 17:39 | |
| mgariepy | fun : Memory cgroup out of memory: Kill process 26338 (ansible-playboo) score 273 | 17:45 |
| cloudnull | mgariepy? | 17:47 |
| cloudnull | OOM killing the run | 17:47 |
| cloudnull | ? | 17:47 |
| mgariepy | yep | 17:48 |
| mgariepy | -f 20 on 4g of ram | 17:48 |
| cloudnull | that's a bummer | 17:50 |
| mgariepy | TASK [lxc_container_create : Drop veth cleanup script] ************************************************** | 17:50 |
| mgariepy | dead also at 8G | 17:50 |
| *** klamath_ has quit IRC | 17:51 | |
| *** markvoelker has quit IRC | 17:51 | |
| *** klamath has joined #openstack-ansible | 17:51 | |
| mgariepy | on queens :) | 17:52 |
| jrosser | cloudnull: this was trivial for lxc, and looked less trivial for nspawn https://review.openstack.org/570456, any ideas? | 17:52 |
| odyssey4me | mgariepy yeah, Tahvok's been to chip away at some of the variables we really don't need to carry in group_vars to cut down on memory usage | 17:52 |
| odyssey4me | I'm also wanting to pull the db creation and mq creation back into the roles for the same reason. | 17:53 |
| *** electrofelix has quit IRC | 17:54 | |
| cloudnull | jrosser we should be able to implement combine(container_extra_networks | default({})) the same in the nspawn hosts. | 17:56 |
| cloudnull | **container_create | 17:56 |
| cloudnull | https://github.com/os-cloud/openstack-ansible-nspawn_container_create/blob/master/tasks/main.yml#L199 | 17:57 |
| mgariepy | yeah was running setup-hosts. only container-deploy don't seems to cause an issue.. | 17:57 |
| *** electrofelix has joined #openstack-ansible | 17:57 | |
| cloudnull | we could also do the same with a default and put that into the tasks | 17:58 |
| *** electrofelix has quit IRC | 17:58 | |
| jrosser | ok i'll take another look | 18:00 |
| *** poopcat has joined #openstack-ansible | 18:00 | |
| odyssey4me | I wouldn't be surprised if most of our playbooks could convert to using include_role, with private: yes - and that'd probably cut down mem usage quite a bit. | 18:01 |
| cloudnull | ++ | 18:01 |
| *** pvradu has quit IRC | 18:03 | |
| *** chhagarw has quit IRC | 18:05 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Add a single variable for combined networks https://review.openstack.org/571542 | 18:05 |
| cloudnull | jrosser ^ | 18:06 |
| cloudnull | I will need to do the same in container create too | 18:06 |
| *** admin0 has quit IRC | 18:07 | |
| spotz | Y'all are chatty and active today:) | 18:07 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_container_create master: Add a single variable for combined networks https://review.openstack.org/571543 | 18:10 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_container_create master: Add a single variable for combined networks https://review.openstack.org/571543 | 18:11 |
| cloudnull | jrosser i think that should cover it, if you have a moment to review those to make sure they'd work as you intend it'd be appreciated. | 18:12 |
| *** admin0 has joined #openstack-ansible | 18:13 | |
| *** DanyC has joined #openstack-ansible | 18:15 | |
| *** DanyC has quit IRC | 18:17 | |
| *** DanyC has joined #openstack-ansible | 18:18 | |
| *** pvradu has joined #openstack-ansible | 18:18 | |
| *** armaan has quit IRC | 18:21 | |
| *** armaan has joined #openstack-ansible | 18:21 | |
| *** pvradu has quit IRC | 18:22 | |
| *** DanyC has quit IRC | 18:22 | |
| *** radeks_ has joined #openstack-ansible | 18:25 | |
| odyssey4me | cores - could we get some eyes on https://review.openstack.org/#/q/topic:bug/1753790+(status:open+OR+status:merged) please? those patches appear to have largely resolved the issue in AIO tests with disk space issues | 18:31 |
| odyssey4me | there does still appear to be a gremlin somewhere, but it's far less frequent now | 18:32 |
| spotz | odyssey4me: those the ones youe posted yesterday? | 18:33 |
| odyssey4me | spotz I posted many yesterday :p | 18:34 |
| spotz | odyssey4me: hehe, there was one list and I think they were all done but hadn't gone through zuul yet. Lookinging at the one on this list still needs review | 18:34 |
| odyssey4me | tyvm mgariepy spotz | 18:36 |
| mgariepy | you are welcome odyssey4me | 18:37 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: Only bind to internal IP. https://review.openstack.org/569141 | 18:38 |
| spotz | odyssey4me: Any time:) | 18:38 |
| *** DanyC has joined #openstack-ansible | 18:45 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/queens: Enable quota system and set qgroups https://review.openstack.org/568867 | 18:47 |
| *** DanyC has quit IRC | 18:49 | |
| *** radeks_ has quit IRC | 18:54 | |
| *** radeks_ has joined #openstack-ansible | 18:55 | |
| *** armaan has quit IRC | 19:04 | |
| *** radeks_ has quit IRC | 19:04 | |
| odyssey4me | Night all - time for me to !computer. | 19:08 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/queens: Enable quota system and set qgroups https://review.openstack.org/568868 | 19:17 |
| *** prasen has quit IRC | 19:18 | |
| *** jra has quit IRC | 19:19 | |
| *** jra has joined #openstack-ansible | 19:19 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Allow special purpose container interfaces https://review.openstack.org/570456 | 19:20 |
| *** DanyC has joined #openstack-ansible | 19:22 | |
| *** radeks_ has joined #openstack-ansible | 19:24 | |
| jrosser | cloudnull: ah i see part in nspawn_hosts and part in nspawn_create.... | 19:26 |
| jrosser | does the nspawn_hosts role take responsibility for some of the work that the deployer used to do? | 19:27 |
| *** radeks__ has joined #openstack-ansible | 19:28 | |
| *** radeks_ has quit IRC | 19:30 | |
| *** mmercer has quit IRC | 19:31 | |
| *** jra has quit IRC | 19:35 | |
| *** jra has joined #openstack-ansible | 19:35 | |
| *** armaan has joined #openstack-ansible | 19:37 | |
| *** armaan has quit IRC | 19:43 | |
| *** tosky has joined #openstack-ansible | 19:44 | |
| *** DanyC has quit IRC | 19:44 | |
| *** DanyC has joined #openstack-ansible | 19:45 | |
| *** DanyC has quit IRC | 19:49 | |
| *** armaan has joined #openstack-ansible | 19:56 | |
| *** mbuil has quit IRC | 20:02 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 20:02 |
| *** armaan has quit IRC | 20:03 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 20:20 |
| cloudnull | jrosser sorry was in the DC, yes | 20:30 |
| cloudnull | the nspawn role(s) take care of all of the required glue. | 20:30 |
| cloudnull | it'll do everything needed to plugin to bridges if present in the provider_networks section or attach directly to a given interface | 20:31 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-nspawn_container_create master: Add test for container specific extra network interface https://review.openstack.org/571567 | 20:32 |
| jrosser | cloudnull: i left some notes on the nspawn_hosts patch, i'm very much in favour of adding the ability to manually specify host bridges in standard group/host vars | 20:33 |
| *** ansmith has quit IRC | 20:34 | |
| jrosser | and i just added a test for the nspawn_container_create patch to check my understanding of it | 20:34 |
| *** radeks__ has quit IRC | 20:39 | |
| cloudnull | ++ very nice | 20:45 |
| *** pvradu has joined #openstack-ansible | 20:50 | |
| *** markvoelker has joined #openstack-ansible | 20:51 | |
| jrosser | cloudnull: do you see anything obvious wrong with the extra network i added, the error is systemd-nspawn[3032]: Failed to resolve interface mv-extra: No such device | 20:54 |
| *** throwsb1 has quit IRC | 21:03 | |
| *** schwicht has quit IRC | 21:05 | |
| *** jeferobles has joined #openstack-ansible | 21:07 | |
| *** jeferobles has quit IRC | 21:10 | |
| *** schwicht has joined #openstack-ansible | 21:11 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-nspawn_container_create master: Add test for container specific extra network interface https://review.openstack.org/571567 | 21:13 |
| *** jeferobles has joined #openstack-ansible | 21:14 | |
| *** jeferobles has quit IRC | 21:20 | |
| *** pvradu has quit IRC | 21:25 | |
| *** jeferobles has joined #openstack-ansible | 21:25 | |
| spotz | mhayden and asettle - this is for you both! ♥️ 🦄 | 21:26 |
| *** markvoelker has quit IRC | 21:26 | |
| *** markvoelker has joined #openstack-ansible | 21:26 | |
| *** markvoelker has quit IRC | 21:31 | |
| *** ansmith has joined #openstack-ansible | 21:33 | |
| *** jeferobles has quit IRC | 21:37 | |
| *** jeferobles has joined #openstack-ansible | 21:39 | |
| *** harryhuang_ has quit IRC | 21:41 | |
| *** jeferobles has quit IRC | 21:42 | |
| *** sar has joined #openstack-ansible | 21:43 | |
| *** jeferobles has joined #openstack-ansible | 21:47 | |
| *** schwicht has quit IRC | 21:49 | |
| *** jeferobles has quit IRC | 21:52 | |
| *** SimAloo has quit IRC | 21:52 | |
| *** sar has quit IRC | 21:52 | |
| *** schwicht has joined #openstack-ansible | 21:52 | |
| *** jeferobles has joined #openstack-ansible | 21:54 | |
| *** nurdie has quit IRC | 21:57 | |
| *** jeferobles has quit IRC | 21:58 | |
| *** nurdie has joined #openstack-ansible | 21:58 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_container_create master: Add test for container specific extra network interface https://review.openstack.org/571567 | 22:00 |
| cloudnull | jrosser sorry, was afk | 22:00 |
| cloudnull | in that test the br-extra interface would need to exist | 22:01 |
| *** nurdie has quit IRC | 22:02 | |
| *** tosky has quit IRC | 22:03 | |
| *** kstev has quit IRC | 22:05 | |
| *** jeferobles has joined #openstack-ansible | 22:06 | |
| cloudnull | the "bridge" in the case of nspawn is any existing interface on the host | 22:07 |
| cloudnull | that could be a bridge, or an ethernet deveice | 22:08 |
| *** jeferobles has quit IRC | 22:09 | |
| *** jeferobles has joined #openstack-ansible | 22:10 | |
| *** tux_ has quit IRC | 22:12 | |
| *** esberglu has quit IRC | 22:14 | |
| *** jeferobles has quit IRC | 22:14 | |
| *** esberglu has joined #openstack-ansible | 22:14 | |
| *** lbragstad has quit IRC | 22:18 | |
| *** esberglu has quit IRC | 22:19 | |
| *** jra has quit IRC | 22:19 | |
| *** jra has joined #openstack-ansible | 22:19 | |
| *** jeferobles has joined #openstack-ansible | 22:21 | |
| *** lbragstad has joined #openstack-ansible | 22:23 | |
| *** jeferobles has quit IRC | 22:24 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 22:32 |
| *** jra has quit IRC | 22:39 | |
| *** jra has joined #openstack-ansible | 22:39 | |
| *** jra has quit IRC | 22:44 | |
| *** jra has joined #openstack-ansible | 22:44 | |
| *** jeferobles has joined #openstack-ansible | 22:49 | |
| *** jeferobles has quit IRC | 22:50 | |
| *** cjloader has quit IRC | 22:51 | |
| *** cjloader_ has joined #openstack-ansible | 22:51 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_octavia master: An OSA password needs to end with _password https://review.openstack.org/570023 | 22:53 |
| *** cjloader_ has quit IRC | 22:55 | |
| *** tux_ has joined #openstack-ansible | 22:57 | |
| *** jeferobles has joined #openstack-ansible | 22:57 | |
| *** tux_ has quit IRC | 23:01 | |
| *** chyka_ has quit IRC | 23:03 | |
| *** schwicht has quit IRC | 23:03 | |
| *** jra has quit IRC | 23:11 | |
| *** jra has joined #openstack-ansible | 23:12 | |
| *** jeferobles has quit IRC | 23:15 | |
| *** jra has quit IRC | 23:22 | |
| *** jra has joined #openstack-ansible | 23:22 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-lxc_container_create master: Don't recurse when creating bind mount directories https://review.openstack.org/571593 | 23:29 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-tests master: Update ansible to 2.5.3 https://review.openstack.org/571423 | 23:31 |
| *** schwicht has joined #openstack-ansible | 23:40 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: Correct cluster name check https://review.openstack.org/571597 | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!