| *** ansmith has quit IRC | 00:15 | |
| *** ansmith has joined #openstack-ansible | 00:15 | |
| *** sreejithp_ has quit IRC | 00:16 | |
| *** hamzy has joined #openstack-ansible | 00:20 | |
| *** dave-mccowan has joined #openstack-ansible | 00:32 | |
| *** markvoelker has quit IRC | 00:33 | |
| *** dave-mccowan has quit IRC | 00:35 | |
| *** hwoarang has quit IRC | 00:37 | |
| *** hwoarang has joined #openstack-ansible | 00:37 | |
| *** hwoarang has quit IRC | 00:54 | |
| *** hwoarang has joined #openstack-ansible | 00:54 | |
| *** sreejithp has joined #openstack-ansible | 01:13 | |
| *** sreejithp has quit IRC | 01:39 | |
| *** hwoarang has quit IRC | 01:45 | |
| *** hwoarang has joined #openstack-ansible | 01:45 | |
| *** hwoarang has quit IRC | 01:58 | |
| *** hwoarang has joined #openstack-ansible | 01:59 | |
| *** BjoernT has joined #openstack-ansible | 02:08 | |
| *** BjoernT has quit IRC | 02:08 | |
| *** hwoarang has quit IRC | 02:18 | |
| *** hwoarang has joined #openstack-ansible | 02:18 | |
| *** shyamb has joined #openstack-ansible | 02:21 | |
| *** sreejithp has joined #openstack-ansible | 02:22 | |
| *** BjoernT has joined #openstack-ansible | 02:22 | |
| *** BjoernT_ has joined #openstack-ansible | 02:27 | |
| *** BjoernT has quit IRC | 02:29 | |
| *** markvoelker has joined #openstack-ansible | 02:34 | |
| *** Talion has quit IRC | 02:43 | |
| *** shyamb has quit IRC | 02:44 | |
| *** sreejithp_ has joined #openstack-ansible | 02:50 | |
| *** sreejithp has quit IRC | 02:53 | |
| *** BjoernT_ has quit IRC | 03:00 | |
| *** partlycloudy has joined #openstack-ansible | 03:05 | |
| *** markvoelker has quit IRC | 03:08 | |
| *** threestrands has joined #openstack-ansible | 03:31 | |
| *** hwoarang has quit IRC | 03:38 | |
| *** hwoarang has joined #openstack-ansible | 03:38 | |
| *** shyamb has joined #openstack-ansible | 03:42 | |
| *** altlogbot_2 has quit IRC | 03:44 | |
| *** altlogbot_1 has joined #openstack-ansible | 03:45 | |
| *** partlycloudy has quit IRC | 03:58 | |
| *** partlycloudy has joined #openstack-ansible | 03:59 | |
| *** markvoelker has joined #openstack-ansible | 04:05 | |
| *** shyam89 has joined #openstack-ansible | 04:07 | |
| *** shyamb has quit IRC | 04:08 | |
| *** bradm has joined #openstack-ansible | 04:32 | |
| *** csmart has quit IRC | 04:36 | |
| *** markvoelker has quit IRC | 04:37 | |
| *** raukadah is now known as chandankumar | 04:37 | |
| *** ivve has quit IRC | 04:37 | |
| *** altlogbot_1 has quit IRC | 04:38 | |
| *** altlogbot_1 has joined #openstack-ansible | 04:39 | |
| *** trident has quit IRC | 04:39 | |
| *** altlogbot_1 has quit IRC | 04:40 | |
| *** altlogbot_1 has joined #openstack-ansible | 04:41 | |
| *** csmart has joined #openstack-ansible | 04:41 | |
| *** trident has joined #openstack-ansible | 04:41 | |
| *** sreejithp_ has quit IRC | 04:49 | |
| *** shyam89 has quit IRC | 04:53 | |
| snadge | hmm.. latest aio deployment im getting | 05:09 |
|---|---|---|
| snadge | TASK [systemd_service : Create TEMP run dir] ***************************************************************************************fatal: [localhost]: FAILED! => {"msg": "'dict object' has no attribute 'interface'"} | 05:09 |
| snadge | ubuntu 18.04 with latest updates.. it wasn't doing it before.. but i noticed a python update slipped by | 05:09 |
| snadge | and now i've snapshotted it after that update.. sigh | 05:09 |
| jrosser | snadge: which osa branch and ansible version? | 05:11 |
| snadge | master | 05:11 |
| snadge | and 2.7.10 | 05:11 |
| jrosser | I have also seen that on master, it feels like an ansible bug. Master is fine if you are doing dev work on osa, not so much for a real deploy | 05:14 |
| *** pcaruana has joined #openstack-ansible | 05:15 | |
| *** shyamb has joined #openstack-ansible | 05:22 | |
| snadge | ahh okay i can flip back to rc1 .. but i was told that was kind of crusty | 05:24 |
| snadge | its only an AIO i suppose.. i'd like to deploy on stein, and im guessing its not too far from release? | 05:24 |
| *** hwoarang has quit IRC | 05:28 | |
| *** hwoarang has joined #openstack-ansible | 05:28 | |
| *** markvoelker has joined #openstack-ansible | 05:34 | |
| *** ivve has joined #openstack-ansible | 05:41 | |
| jrosser | Master is the train version, not stein | 05:51 |
| snadge | it is? | 05:51 |
| jrosser | Well it will become train in the future when the branch is created, but it is certainly not stein | 05:52 |
| snadge | okay.. according to the https://releases.openstack.org/stein/ | 05:53 |
| snadge | most recent version is 19.0.0.0rc1 | 05:53 |
| snadge | so there is only a beta or a release candidate version of osa for stein? | 05:54 |
| jrosser | Correct | 05:55 |
| jrosser | Deployment projects by necessity sometimes have their release trail behind the actual Openstack service components | 05:56 |
| snadge | is 19.0.0.0.0 likely to be some time away? | 05:56 |
| snadge | i guess i can go back to a previous release like 18.1.6 | 05:56 |
| jrosser | As I understand it there is still one major thing, moving the placement service out of nova. This is done in osa master but not yet on the stein branch | 05:58 |
| *** shyamb has quit IRC | 06:07 | |
| *** shyamb has joined #openstack-ansible | 06:08 | |
| *** markvoelker has quit IRC | 06:08 | |
| *** cshen has joined #openstack-ansible | 06:23 | |
| *** shyamb has quit IRC | 06:35 | |
| snadge | its failing in the same place with 19.0.0.0rc1 | 06:36 |
| snadge | im suspecting its one of these updates i installed which broke it.. because it was working before.. even on master | 06:36 |
| *** shyamb has joined #openstack-ansible | 06:39 | |
| *** gregoryo has joined #openstack-ansible | 06:41 | |
| snadge | i'd take a look at it myself.. it looks pretty simple this "Create TEMP run dir" rule.. but im admittedly not very knowledgeable with ansible | 06:42 |
| jrosser | snadge: are you starting with a fresh install each time? | 06:43 |
| snadge | fresh install of ubuntu yes.. i've pulled down the osa git.. then snapshotted before running the bootstrap command etc | 06:44 |
| snadge | but i installed a small handful of updates.. deleted the snapshot and created a new one | 06:44 |
| snadge | which i now regret.. but the good news i guess, is that it must have been one of those few updates i installed, which is causing the problem | 06:44 |
| snadge | since it was working prior to installing those updates | 06:44 |
| jrosser | Ubuntu updates? | 06:45 |
| snadge | Ubuntu 18.04.2 LTS yes | 06:45 |
| snadge | the updates were: Upgrade: python3-software-properties:amd64 (0.96.24.32.7, 0.96.24.32.8), python3-distupgrade:amd64 (1:18.04.32, 1:18.04.33), ubuntu-release-upgrader-core:amd64 (1:18.04.32, 1:18.04.33), software-properties-common:amd64 (0.96.24.32.7, 0.96.24.32.8) | 06:46 |
| snadge | according to the apt history log | 06:46 |
| *** hwoarang has quit IRC | 06:47 | |
| *** hwoarang has joined #openstack-ansible | 06:47 | |
| snadge | none of those seem particularly relevant though | 06:47 |
| snadge | then the bootstrap pulls in a bunch more stuff.. so its possible one of those has updated | 06:48 |
| jrosser | The odd thing is that this does not happen in our CI | 06:49 |
| jrosser | But now two of us have the same error in local AIO | 06:49 |
| snadge | the logical part of me says to turn on a crapload of debugging with ansible and see where its tripping up? | 06:52 |
| jrosser | Yes | 06:53 |
| gregoryo | Debugging Ansible has always felt very doable, if clumsy. A given task will fail, and you can find (grep) it in a playbook, then try to do that manually or something. | 06:54 |
| gregoryo | I'm coming in late though, not sure if that's helpful. | 06:54 |
| snadge | im about to bail from work.. but its easy for me to spin up 18.04 in virtualbox or something at home.. and try again | 06:55 |
| snadge | i'll have a look at the boostrap-aio command and see how to turn on extra debugging | 06:56 |
| jrosser | evrardjp: is adding openstacksdk instead of shade acceptable here for master and stein? https://review.opendev.org/#/c/660779/ | 06:56 |
| *** shyamb has quit IRC | 06:56 | |
| *** shyamb has joined #openstack-ansible | 06:56 | |
| *** starborn has joined #openstack-ansible | 06:59 | |
| jrosser | snadge: there is an ANSIBLE_DEBUG env var, as well as regular -vvvv on the playbook | 06:59 |
| *** oyrogerg has joined #openstack-ansible | 07:01 | |
| *** gregoryo has quit IRC | 07:03 | |
| *** markvoelker has joined #openstack-ansible | 07:05 | |
| *** oyrogerg has quit IRC | 07:06 | |
| *** tosky has joined #openstack-ansible | 07:11 | |
| *** shyamb has quit IRC | 07:12 | |
| *** shyamb has joined #openstack-ansible | 07:21 | |
| evrardjp | good morning | 07:29 |
| evrardjp | jrosser: I will have a look | 07:29 |
| evrardjp | as I said, I doubt this is the right thing to do, I would do it a little above | 07:30 |
| evrardjp | utility_distro_openstack_clients_packages contains all the things to really do actions in that container | 07:30 |
| evrardjp | is it's for a place to delegate actions to, then I guess it makes sense to ensure standard python have the right libs, but because it's managed by the distro packages, I would not mix both basically | 07:32 |
| evrardjp | then for the shade/openstacksdk, let me check for our version | 07:32 |
| jrosser | I am confused, do I need a distro package? Not from pip? | 07:32 |
| jrosser | The ansible modules delegated to the utility container need the right libs to be in the venv | 07:33 |
| evrardjp | in the venv? | 07:33 |
| evrardjp | do we point to target node's venv? | 07:33 |
| jrosser | yes | 07:33 |
| evrardjp | where? | 07:34 |
| evrardjp | in that case it totally doesn't make sense to put that package there. | 07:34 |
| evrardjp | and yes, confirmed, for rocky we need shade | 07:34 |
| jrosser | Here https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/user_variables.yml#L34 | 07:35 |
| evrardjp | and for stein and above openstacksdk | 07:35 |
| jrosser | Why doesn’t it make sense to put the package there? | 07:35 |
| evrardjp | ok, so assuming ppl uncomment both | 07:36 |
| jrosser | Yes that is required | 07:36 |
| jrosser | https://github.com/openstack/openstack-ansible/blob/master/playbooks/utility-install.yml#L124 | 07:37 |
| *** markvoelker has quit IRC | 07:37 | |
| evrardjp | yes | 07:37 |
| evrardjp | I am on that line, for rocky :) | 07:37 |
| evrardjp | poorly named thing, sorry, I was confused again | 07:38 |
| evrardjp | it would work just fine if both are defined | 07:38 |
| evrardjp | I would still prefer make that part of the client list | 07:39 |
| *** tosky has quit IRC | 07:40 | |
| evrardjp | I guess I should spin up a rocky to see if we really have the shade in the requirements_absolute_requirements.txt | 07:41 |
| jrosser | Hmmm, that involves a grossly more complicated code oath | 07:41 |
| jrosser | Path | 07:41 |
| evrardjp | (I am just wanting to ensure u-c are respected for this, so I am still evaluating) | 07:42 |
| *** tosky has joined #openstack-ansible | 07:42 | |
| *** aedc has joined #openstack-ansible | 07:42 | |
| *** Emine has joined #openstack-ansible | 07:43 | |
| evrardjp | (I just found a bug in our utility install though... argh) | 07:43 |
| evrardjp | two in fact | 07:43 |
| evrardjp | argh | 07:43 |
| evrardjp | ok I guess it's not perfect anyway, let me revert my vote to a positive one | 07:44 |
| evrardjp | jrosser: ok, re-commented | 07:46 |
| jrosser | evrardjp: thankyou will change it | 07:47 |
| jrosser | evrardjp: i have one more query related to requirements, this i found quite surprising https://review.opendev.org/#/c/638461 | 07:48 |
| evrardjp | I am waiting for the rest of the question :) | 07:49 |
| jrosser | ok - so there is a release note to oslo recommending the use of a version of amqp that lies outside requirements versions | 07:50 |
| evrardjp | well yes | 07:50 |
| jrosser | what do we (as OSA) do about that when faced with potential brokenness in deploys becasuse we stick to the older version | 07:50 |
| evrardjp | it's not because there is a new version of oslo you can bump this inside your project | 07:51 |
| jrosser | someone asked in here about this very thing yesterday | 07:51 |
| jrosser | and i didnt know what the right thing to do was | 07:51 |
| evrardjp | that's a good question | 07:51 |
| evrardjp | so the u-c ensure co-installability. This is why we are using those, it's because they are tested | 07:51 |
| evrardjp | stepping outside and here be dragons | 07:52 |
| evrardjp | however, it could make sense for a user to manually override the oslo.messaging version | 07:52 |
| evrardjp | we can probably in that case add a known issue and explain how to fix it | 07:52 |
| evrardjp | in the renos | 07:52 |
| evrardjp | it's also technically possible to bump constraints everywhere, to benefit from this | 07:55 |
| evrardjp | but that's a lot of effort | 07:55 |
| *** hamzaachi has joined #openstack-ansible | 07:56 | |
| evrardjp | which I am not ready to take, but patches are welcomed I guess? :) | 07:56 |
| *** tstrul has joined #openstack-ansible | 07:56 | |
| *** shyamb has quit IRC | 08:04 | |
| *** hamzaachi has quit IRC | 08:04 | |
| *** hamzaachi_ has joined #openstack-ansible | 08:04 | |
| *** hamzaachi_ has quit IRC | 08:06 | |
| *** hamzaachi_ has joined #openstack-ansible | 08:06 | |
| *** kopecmartin|off is now known as kopecmartin | 08:09 | |
| *** hamzaachi__ has joined #openstack-ansible | 08:11 | |
| *** hamzaachi_ has quit IRC | 08:11 | |
| *** hamzaachi__ has quit IRC | 08:16 | |
| *** hamzaachi__ has joined #openstack-ansible | 08:17 | |
| *** hamzaachi_ has joined #openstack-ansible | 08:23 | |
| *** hamzaachi__ has quit IRC | 08:27 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/rocky: Utility container venv is missing python libs for ansible openstack modules https://review.opendev.org/661812 | 08:28 |
| *** hamzaachi__ has joined #openstack-ansible | 08:28 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Utility container venv is missing python libs for ansible openstack modules https://review.opendev.org/660779 | 08:30 |
| *** hamzaachi_ has quit IRC | 08:31 | |
| *** markvoelker has joined #openstack-ansible | 08:34 | |
| *** hamzaachi__ has quit IRC | 08:51 | |
| *** hamzaachi__ has joined #openstack-ansible | 08:51 | |
| *** hamzaachi_ has joined #openstack-ansible | 08:54 | |
| *** shyamb has joined #openstack-ansible | 08:56 | |
| *** hamzaachi_ has quit IRC | 08:56 | |
| *** hamzaachi__ has quit IRC | 08:57 | |
| *** hamzaachi has joined #openstack-ansible | 08:57 | |
| *** hamzaachi_ has joined #openstack-ansible | 09:00 | |
| *** hamzaachi has quit IRC | 09:02 | |
| *** markvoelker has quit IRC | 09:08 | |
| *** aedc has quit IRC | 09:16 | |
| *** aedc has joined #openstack-ansible | 09:17 | |
| *** threestrands has quit IRC | 09:54 | |
| *** markvoelker has joined #openstack-ansible | 10:05 | |
| *** shyamb has quit IRC | 10:05 | |
| *** shyamb has joined #openstack-ansible | 10:05 | |
| *** electrofelix has joined #openstack-ansible | 10:06 | |
| noonedeadpunk | mornings | 10:28 |
| *** dave-mccowan has joined #openstack-ansible | 10:32 | |
| *** dave-mccowan has quit IRC | 10:37 | |
| *** markvoelker has quit IRC | 10:37 | |
| *** shyamb has quit IRC | 10:38 | |
| *** hwoarang has quit IRC | 10:41 | |
| *** hwoarang has joined #openstack-ansible | 10:42 | |
| *** hwoarang has quit IRC | 10:56 | |
| *** hwoarang has joined #openstack-ansible | 10:57 | |
| *** shyamb has joined #openstack-ansible | 11:03 | |
| CeeMac | afternoon | 11:11 |
| *** fresta has quit IRC | 11:18 | |
| *** kplant has joined #openstack-ansible | 11:18 | |
| *** fresta has joined #openstack-ansible | 11:20 | |
| *** aedc has quit IRC | 11:21 | |
| *** fresta_ has quit IRC | 11:21 | |
| *** mkuf_ has quit IRC | 11:24 | |
| *** mkuf_ has joined #openstack-ansible | 11:25 | |
| *** mkuf_ has quit IRC | 11:28 | |
| *** mkuf has joined #openstack-ansible | 11:28 | |
| *** cshen has quit IRC | 11:31 | |
| *** cshen has joined #openstack-ansible | 11:33 | |
| *** markvoelker has joined #openstack-ansible | 11:34 | |
| *** hwoarang has quit IRC | 11:34 | |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible master: rgw keystone setup fix python interpreter https://review.opendev.org/661006 | 11:36 |
| *** hwoarang has joined #openstack-ansible | 11:37 | |
| *** macza has joined #openstack-ansible | 11:38 | |
| *** redrobot has quit IRC | 11:39 | |
| *** macza has quit IRC | 11:43 | |
| *** markvoelker has quit IRC | 11:54 | |
| *** markvoelker has joined #openstack-ansible | 11:54 | |
| *** fresta has quit IRC | 11:56 | |
| *** fresta has joined #openstack-ansible | 11:56 | |
| *** aedc has joined #openstack-ansible | 12:04 | |
| *** ansmith has quit IRC | 12:05 | |
| *** redrobot has joined #openstack-ansible | 12:23 | |
| *** udesale has joined #openstack-ansible | 12:25 | |
| *** ansmith has joined #openstack-ansible | 12:44 | |
| *** shyamb has quit IRC | 12:59 | |
| openstackgerrit | Chandan Kumar (raukadah) proposed openstack/openstack-ansible-os_tempest master: [wip][redhat] Update the list of python3 packages https://review.opendev.org/661979 | 13:21 |
| chandankumar | jrosser: odyssey4me I need some help on this review https://review.opendev.org/#/c/661979/1/vars/redhat.yml@30 using ternary with in a ternary | 13:22 |
| *** lucashxu has joined #openstack-ansible | 13:24 | |
| jrosser | chandankumar: if it’s not working I’d probably write a tiny local playbook with similar logic to test the behaviour | 13:29 |
| chandankumar | ah, yes trying that | 13:29 |
| *** BjoernT has joined #openstack-ansible | 13:34 | |
| *** BjoernT_ has joined #openstack-ansible | 13:39 | |
| *** BjoernT has quit IRC | 13:39 | |
| jrosser | chandankumar why not just define _tempest_plugin_distro_packages_python2 and _tempest_plugin_distro_packages_python3 then use a single ternary to assign one of those to tempest_plugin_distro_packages | 13:41 |
| jrosser | and before we go too far some consideration to the package naming on suse and ubuntu is probably required | 13:42 |
| *** sreejithp has joined #openstack-ansible | 13:43 | |
| jrosser | so see if there are similar requirements so that the logic can be common | 13:43 |
| *** starborn has quit IRC | 13:59 | |
| jrosser | evrardjp: didnt we fix this? http://logs.openstack.org/79/660779/2/gate/openstack-ansible-deploy-aio_metal-opensuse-150/f6cac8a/job-output.txt.gz#_2019-05-29_11_14_53_050635 | 13:59 |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: [wip][redhat] Update the list of python3 packages https://review.opendev.org/661994 | 14:12 |
| *** lucashxu has quit IRC | 14:25 | |
| *** ivve has quit IRC | 14:29 | |
| *** cshen has quit IRC | 14:30 | |
| *** bline has joined #openstack-ansible | 14:32 | |
| *** hamzaachi__ has joined #openstack-ansible | 14:36 | |
| *** aedc has quit IRC | 14:37 | |
| *** hamzaachi_ has quit IRC | 14:38 | |
| *** chandankumar is now known as raukadah | 14:54 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: Update the list of python3 packages https://review.opendev.org/661994 | 15:04 |
| *** shyamb has joined #openstack-ansible | 15:07 | |
| *** aedc has joined #openstack-ansible | 15:08 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/rocky: Bump SHAs for stable/rocky https://review.opendev.org/659967 | 15:08 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/queens: Bump SHAs for stable/queens https://review.opendev.org/659969 | 15:09 |
| *** spatel has joined #openstack-ansible | 15:24 | |
| *** shyamb has quit IRC | 15:24 | |
| spatel | is stein stable enough to deploy on new production cloud? | 15:29 |
| *** gyee has joined #openstack-ansible | 15:33 | |
| *** altlogbot_1 has quit IRC | 15:35 | |
| *** altlogbot_0 has joined #openstack-ansible | 15:35 | |
| *** irclogbot_0 has quit IRC | 15:36 | |
| *** irclogbot_0 has joined #openstack-ansible | 15:36 | |
| *** macza has joined #openstack-ansible | 15:37 | |
| *** hamzaachi__ has quit IRC | 15:37 | |
| *** tstrul has quit IRC | 15:47 | |
| *** hamzaachi has joined #openstack-ansible | 15:49 | |
| *** shyamb has joined #openstack-ansible | 15:56 | |
| noonedeadpunk | spatel: mnaser actually already migrated to stein, but I'd recommend to wait for upgrade scripts being ready for stein as weel | 15:59 |
| noonedeadpunk | *well | 15:59 |
| spatel | This is new installation ( I am testing AIO on CentOS 7.5 ) | 16:00 |
| spatel | Getting some strange error :( | 16:00 |
| spatel | Do you suggesting i should do rocky? | 16:01 |
| *** ivve has joined #openstack-ansible | 16:02 | |
| noonedeadpunk | YOu can give stein a try:) iirc cloudnull is using stein in prod, so it's stable enough I think | 16:02 |
| noonedeadpunk | probably you should use this patch https://review.opendev.org/#/c/659966/ which will eventually become the next tag | 16:02 |
| * jrosser wonders about placement & stein | 16:03 | |
| noonedeadpunk | it's not required, but yes... good point | 16:03 |
| jrosser | my concern with deploying stein today will be that it comes with non extracted placement, we then get that backported and theres an immediate ugly upgrade situation for early stein installs | 16:04 |
| jrosser | spatel: out of interest what error are you getting? | 16:06 |
| spatel | give me one min i found something... | 16:06 |
| spatel | jrosser: This is the issue - http://paste.openstack.org/show/752257/ | 16:10 |
| spatel | I am using CentOS 7.5 distro | 16:10 |
| spatel | during that command it failed with dependencies | 16:10 |
| spatel | i believe its my repo mixing up somewhere | 16:11 |
| cjloader | are you using nspawn, lxc, or no containers? | 16:12 |
| jrosser | quite likley, a few things broke when 7.6 came out and needed fixing up | 16:12 |
| jrosser | so i expect trying to do 7.5 is going to be tricky | 16:12 |
| *** hwoarang has quit IRC | 16:13 | |
| *** hwoarang has joined #openstack-ansible | 16:13 | |
| spatel | This is CentOS and AIO should setup networking right? | 16:14 |
| spatel | CentOS doesn't like nspawn ( because in my current cloud i had lots of issue and i remove it ) | 16:15 |
| spatel | Let me give it a try with 7.6 | 16:15 |
| *** shyamb has quit IRC | 16:20 | |
| canori01 | Hello, I deployed gnocchi with using osa, but it didn't generate the resource types. So whenever ceilometer passes metrics to it, it fails. Did I miss a step? | 16:22 |
| *** radeks__ has joined #openstack-ansible | 16:27 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Load extra variables correctly in wrapper https://review.opendev.org/662032 | 16:31 |
| *** Emine has quit IRC | 16:31 | |
| *** hwoarang has quit IRC | 16:32 | |
| *** hwoarang has joined #openstack-ansible | 16:37 | |
| *** dave-mccowan has joined #openstack-ansible | 16:40 | |
| prometheanfire | selinux and OSA, I'm guessing no? | 16:43 |
| prometheanfire | iirc there are some policies for openstack, but since OSA has difrent paths and the like... | 16:44 |
| *** hamzaachi has quit IRC | 16:46 | |
| prometheanfire | also, how strong is basic debian support, I imagine it's good because how much ubuntu 'borrows' from debian repo wise | 16:46 |
| jrosser | prometheanfire: afaik its pretty solid | 16:48 |
| * prometheanfire may be using it, so patches if it's not | 16:48 | |
| * prometheanfire shrugs | 16:48 | |
| jrosser | there is pretty complete coverage in the zuul jobs anyway | 16:49 |
| logan- | No selinux. Major wrote some policies but no one maintained them so they were removed a few cycles ago | 16:50 |
| noonedeadpunk | canori01: try running ceilometer-upgrade on your ceilometer-central container/host | 16:50 |
| noonedeadpunk | and make sure that gnocchi resource-type list is not empty | 16:51 |
| * jrosser starts to think there are some interesting gremlins in ansible 2.8.0 | 16:51 | |
| spatel | jrosser: FYI - 7.6 works look like i am not seeing error in that specific step | 16:53 |
| *** Weifan has joined #openstack-ansible | 16:54 | |
| *** hwoarang has quit IRC | 16:54 | |
| *** hwoarang has joined #openstack-ansible | 16:56 | |
| *** Weifan has quit IRC | 16:59 | |
| *** Weifan has joined #openstack-ansible | 17:09 | |
| *** priteau has joined #openstack-ansible | 17:11 | |
| *** jcosmao has left #openstack-ansible | 17:12 | |
| *** aedc has quit IRC | 17:13 | |
| *** hwoarang has quit IRC | 17:15 | |
| *** hwoarang has joined #openstack-ansible | 17:16 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Move hosts file management toggle to all group vars https://review.opendev.org/662050 | 17:25 |
| *** priteau has quit IRC | 17:33 | |
| *** luksky has joined #openstack-ansible | 17:41 | |
| *** kopecmartin is now known as kopecmartin|off | 17:47 | |
| *** tosky__ has joined #openstack-ansible | 17:59 | |
| *** tosky has quit IRC | 18:00 | |
| *** tosky__ is now known as tosky | 18:00 | |
| kplant | jrosser: i've been thinking more about routing fips through a vrouter on 1918 space and i think i've hit a wall... how would you allocate FIPs from one cider (public) but have the router interface have an address in 1918? | 18:02 |
| *** electrofelix has quit IRC | 18:03 | |
| kplant | cidr* ugh | 18:03 |
| kplant | i need some cider | 18:03 |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: Update the list of python3 packages https://review.opendev.org/661994 | 18:03 |
| *** priteau has joined #openstack-ansible | 18:03 | |
| jrosser | Well isn’t it more that the computers all have a bgp speaker with an rfc1918, and so does your next hop hardware | 18:04 |
| jrosser | And bgp propagates /32 routes for the fip from the computers to the hardware router | 18:04 |
| kplant | yes, that part i'm sold on | 18:05 |
| jrosser | Ok.... | 18:05 |
| kplant | what i mean is, when a user of a project creates their vrouter.. it will have an interface with a 1918 address for the external network | 18:05 |
| kplant | how would they also be able to allocate floating IPs on a different network? | 18:05 |
| jrosser | I think that’s ok for the vrouter, because it is the legitimate ip of the next hop for the fip | 18:06 |
| jrosser | The host routes from the fip to the hardware router via its rfc1918 address | 18:07 |
| jrosser | But I also would like to see a worked example of this :) | 18:07 |
| kplant | yeah i have no doubt the routing will work, i guess what i'm more unsure about is neutron trying to not do something it thinks isn't possible | 18:07 |
| kplant | i'm not getting beyond how you're supposed to ensure the vrouter gets an interface with 1918 space and the FIPs are not 1918 | 18:08 |
| jrosser | Yes I understand what you are saying there - there was some stuff about the dvr gateway up which is probably totally relevant here | 18:08 |
| jrosser | Up/ip | 18:09 |
| kplant | yeah maybe with neutron-bgp and dvr there's some traits or additional metadata for this | 18:09 |
| kplant | it sounds like to me there needs to be something that says take router interface addresses from [here] and fips from [overhere] | 18:10 |
| * jrosser just travelling, can’t google so much :) | 18:10 | |
| kplant | ah | 18:10 |
| jrosser | I’m wondering in this case if there is a concept of router interface address | 18:12 |
| jrosser | Because it’s the same, on a given node for all the fip | 18:12 |
| kplant | i'm thinking more on the project level, so a project needs to stitch their tenant network with the provider net | 18:14 |
| kplant | so they need their own router with at least a pair of interfaces | 18:14 |
| spatel | kplant: interesting conversation.. | 18:21 |
| kplant | i'm trying to save the IPs! | 18:21 |
| kplant | heh | 18:21 |
| spatel | when i was testing DVR while ago i had same issue... | 18:21 |
| kplant | lots of nodes with lots of projects and DVR explodes this | 18:21 |
| spatel | in DVR every vrouter need Public IP ( if you have 300 compute nodes then you will have 300 vrouter with 300 public ip) | 18:22 |
| kplant | yep | 18:22 |
| kplant | that's why i'm trying to chuck them on 1918 space and route through them | 18:23 |
| spatel | There is a way to create service IP on DVR which doesn't need public IP on vRouter but i don't know how that work with floating IP ( you have same question ) | 18:23 |
| spatel | I would contact James D. he is very good at DVR stuff... | 18:24 |
| spatel | Finally i decided not to use DVR. currently my cloud using VLAN provider | 18:24 |
| kplant | you're just putting the VMs right on the provider net? | 18:25 |
| openstackgerrit | Merged x/monitorstack master: Fix docs building https://review.opendev.org/661633 | 18:25 |
| openstackgerrit | Merged x/monitorstack master: Update zuul jobs https://review.opendev.org/661614 | 18:26 |
| spatel | Yes, my VMs directly attacked to my physical infrastructure and VLANs | 18:26 |
| spatel | attached* | 18:26 |
| kplant | aren't you afraid of broadcast storms? | 18:26 |
| spatel | I haven't seen any yet.. but you can create multiple VLANs to control broadcast | 18:27 |
| spatel | I have 300 compute nodes in cluster running 800 VMs and so far so good... | 18:27 |
| kplant | i don't trust my users that much :> | 18:27 |
| spatel | Reason i pick VLANs because i need performance | 18:28 |
| kplant | a year or two ago they plopped a PA appliance on the network with a few interfaces all bridging the same domain | 18:28 |
| kplant | it was good night | 18:28 |
| spatel | NATing is not fun when comes to performance.. ( PPS rate 0 | 18:28 |
| spatel | NATing is not fun when comes to performance.. ( PPS rate ) | 18:28 |
| spatel | lol | 18:29 |
| kplant | yeah that's for sure, especially when it's PAT | 18:29 |
| spatel | Yup! are application is based on UDP and need very low latency | 18:30 |
| kplant | voice? | 18:30 |
| spatel | My whole cloud running on SR-IOV :) | 18:30 |
| spatel | Yes voice 90% | 18:30 |
| kplant | voice is the worst | 18:30 |
| kplant | tiny frames | 18:30 |
| spatel | yup.. small packet size cause lots of overhead on CPU.. | 18:30 |
| kplant | oh yeah... we have a bunch of circuits where we offer L2 tunnels over DOCSIS | 18:31 |
| spatel | After SR-IOV i am happy ( Now building next cloud on DPDK ) | 18:31 |
| kplant | and the poor CMs can't deal with the <128b frames of voice | 18:31 |
| kplant | they just start dropping frames | 18:31 |
| spatel | i had same issue ( tap interface are worst when comes to PPS rate ) | 18:32 |
| spatel | From 30000ft clouds looks good but only cloud operator knows what is going inside cloud :) | 18:33 |
| kplant | they look great to the end user | 18:33 |
| kplant | they don't have to know anything | 18:33 |
| spatel | Why don't you try opencontail ? | 18:33 |
| *** BjoernT_ has quit IRC | 18:33 | |
| kplant | tungsten is on the radar | 18:34 |
| kplant | my company has a great relationship with juniper | 18:34 |
| spatel | I would go to tungsten instead of DVR | 18:34 |
| spatel | http://www.jimmdenton.com/contrail-osa/ | 18:35 |
| *** priteau has quit IRC | 18:35 | |
| kplant | i need to translate that to kolla :x | 18:35 |
| spatel | OSA and Kolla are pretty much same.. | 18:36 |
| spatel | Do you like Kolla or have something historical ? | 18:36 |
| kplant | it's all greenfield right now | 18:36 |
| kplant | i'm totally leavnig behind what we have | 18:37 |
| kplant | so yeah, kolla-ansible is probably my personal favorite | 18:37 |
| spatel | yup sounds good | 18:37 |
| spatel | docker vs lxc | 18:38 |
| kplant | i don't really like docker, i feel it's kind of bloated | 18:39 |
| kplant | but k-a is the only deployer i got to work on the first try | 18:39 |
| kplant | it's just stupid simple | 18:39 |
| spatel | same here, (Love LXC) docker is good behind kubernetes | 18:40 |
| dmsimard | prometheanfire: fwiw RDO ships an openstack-selinux package derived from https://github.com/redhat-openstack/openstack-selinux -- packaging is https://github.com/rdo-packages/openstack-selinux-distgit | 18:43 |
| kplant | spatel: are you provisioning the vlans with ml2? | 18:52 |
| spatel | yes ml2 with linuxbridge | 18:53 |
| *** BjoernT has joined #openstack-ansible | 18:54 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_keystone master: Prevent autoremoval of packages on bare metal https://review.opendev.org/661333 | 18:59 |
| *** BjoernT_ has joined #openstack-ansible | 18:59 | |
| *** BjoernT has quit IRC | 18:59 | |
| CeeMac | kplant: spatel: I've been around in circles so many times trying to work out how best to deploy. I've gone OVS with support for DVR but legacy router as default for tenants to leave my options open. My cloud is starting off small but will likely scale fast so I need a midterm plan. Tungsten is one option (we like juniper here too). OVN is another option, especially since the neutron team seems to be looking at ways | 19:01 |
| CeeMac | of bringing OVN into reference architecture for distributed networking. | 19:01 |
| CeeMac | kplant: if I didn't have a deadline I'd definitely try the RFC 1918 router IP proposal | 19:02 |
| *** tuxjohnson has joined #openstack-ansible | 19:05 | |
| *** tuxjohnson has left #openstack-ansible | 19:05 | |
| spatel | https://lists.gt.net/openstack/operators/52081 | 19:13 |
| spatel | kplant: Do you really need Distributed solution? | 19:14 |
| kplant | i don't _need_ a distributed solution no | 19:14 |
| kplant | but moving all of the traffic through the controllers seems like a giant waste to me | 19:14 |
| spatel | kplant: You can use hardware base gateway also like Cisco ASR1K | 19:15 |
| spatel | if you can spend money ( but again they do have limitation to handle NAT tables etc.) | 19:15 |
| spatel | Cisco ASR1k support neutron and spin up L3 VRF on hardware | 19:17 |
| kplant | we're actually using NCS55A1's for something similar on another project | 19:18 |
| spatel | kplant: or you can try these hardware - https://cumulusnetworks.com/networking-solutions/openstack/ | 19:18 |
| CeeMac | I think you can do similar with Juniper MX and QFX | 19:19 |
| spatel | Does NCS55A1 support neutron ? | 19:19 |
| *** joshualyle has joined #openstack-ansible | 19:21 | |
| CeeMac | https://www.juniper.net/documentation/en_US/release-independent/junos/topics/concept/openstack-neutron-plugin-overview.html | 19:21 |
| *** joshualyle has quit IRC | 19:25 | |
| *** joshualyle has joined #openstack-ansible | 19:33 | |
| *** joshualyle has quit IRC | 19:37 | |
| *** kplant has quit IRC | 19:56 | |
| *** joshualyle has joined #openstack-ansible | 20:04 | |
| *** joshualyle has quit IRC | 20:05 | |
| spatel | Finally my AIO deployment completed - DISTRIB_CODENAME="Stein" | 20:10 |
| *** lathap has joined #openstack-ansible | 20:24 | |
| *** spatel has quit IRC | 20:26 | |
| *** ansmith has quit IRC | 20:43 | |
| *** Weifan has quit IRC | 20:53 | |
| *** Weifan has joined #openstack-ansible | 20:54 | |
| *** Weifan has quit IRC | 20:58 | |
| *** radeks__ has quit IRC | 21:02 | |
| *** priteau has joined #openstack-ansible | 21:21 | |
| *** pcaruana has quit IRC | 21:23 | |
| *** priteau has quit IRC | 21:26 | |
| *** sreejithp has quit IRC | 21:28 | |
| *** hwoarang has quit IRC | 21:35 | |
| *** hwoarang has joined #openstack-ansible | 21:35 | |
| *** priteau has joined #openstack-ansible | 21:38 | |
| *** udesale has quit IRC | 21:43 | |
| *** tosky has quit IRC | 21:49 | |
| *** BjoernT_ is now known as BjoernT | 21:51 | |
| *** priteau has quit IRC | 21:51 | |
| *** kplant has joined #openstack-ansible | 22:03 | |
| *** joshualyle has joined #openstack-ansible | 22:06 | |
| *** joshualyle has quit IRC | 22:11 | |
| *** ianychoi has quit IRC | 22:32 | |
| *** ianychoi has joined #openstack-ansible | 22:33 | |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-galera_server stable/pike: Update percona-release package https://review.opendev.org/662107 | 22:34 |
| *** hamzaachi has joined #openstack-ansible | 22:42 | |
| *** BjoernT has quit IRC | 22:43 | |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-tests stable/pike: Fix clone list for opendev changes https://review.opendev.org/662108 | 22:53 |
| *** hwoarang has quit IRC | 22:53 | |
| *** hwoarang has joined #openstack-ansible | 22:55 | |
| *** hamzaachi has quit IRC | 22:56 | |
| *** hamzaachi has joined #openstack-ansible | 22:56 | |
| *** jonher_ has joined #openstack-ansible | 22:57 | |
| *** lathap has quit IRC | 22:57 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Utility container venv is missing python libs for ansible openstack modules https://review.opendev.org/660779 | 22:59 |
| *** jonher has quit IRC | 23:00 | |
| *** jonher_ is now known as jonher | 23:00 | |
| *** jonher has quit IRC | 23:00 | |
| *** Weifan has joined #openstack-ansible | 23:04 | |
| *** mcarden has joined #openstack-ansible | 23:04 | |
| *** Weifan has quit IRC | 23:04 | |
| *** Weifan has joined #openstack-ansible | 23:05 | |
| *** Weifan has quit IRC | 23:05 | |
| *** Weifan has joined #openstack-ansible | 23:05 | |
| *** macza has quit IRC | 23:09 | |
| *** hamzaachi has quit IRC | 23:10 | |
| *** hamzaachi has joined #openstack-ansible | 23:10 | |
| *** Weifan has quit IRC | 23:10 | |
| *** Weifan has joined #openstack-ansible | 23:11 | |
| *** Weifan has quit IRC | 23:12 | |
| *** Weifan has joined #openstack-ansible | 23:12 | |
| *** luksky has quit IRC | 23:13 | |
| *** hwoarang has quit IRC | 23:30 | |
| *** hwoarang has joined #openstack-ansible | 23:31 | |
| *** hwoarang has quit IRC | 23:40 | |
| *** hwoarang has joined #openstack-ansible | 23:40 | |
| *** hamzaachi has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!