| *** cjloader has joined #openstack-ansible | 00:14 | |
| *** cjloader has quit IRC | 00:17 | |
| *** dxiri has quit IRC | 00:22 | |
| openstackgerrit | Leopard Ma proposed openstack/openstack-ansible-os_horizon master: Add source code repository notes link to README https://review.openstack.org/579421 | 00:35 |
|---|---|---|
| openstackgerrit | Leopard Ma proposed openstack/openstack-ansible-os_heat master: Add source code repository notes link to README https://review.openstack.org/579419 | 00:39 |
| openstackgerrit | Leopard Ma proposed openstack/openstack-ansible-os_gnocchi master: Add source code repository notes link to README https://review.openstack.org/579417 | 00:44 |
| openstackgerrit | Leopard Ma proposed openstack/openstack-ansible-os_glance master: Add source code repository notes link to README https://review.openstack.org/579416 | 00:47 |
| *** tux_ has joined #openstack-ansible | 00:50 | |
| *** chyka has joined #openstack-ansible | 00:55 | |
| *** chyka has quit IRC | 01:00 | |
| cloudnull | I like ponies | 01:17 |
| guilhermesp | cloudnull really like them, that's cute | 01:27 |
| guilhermesp | be happy: https://i.imgur.com/ottKmhq.jpg | 01:33 |
| *** dxiri has joined #openstack-ansible | 01:43 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_tempest master: fix tox python3 overrides https://review.openstack.org/580069 | 02:26 |
| *** dxiri has quit IRC | 02:40 | |
| *** udesale has joined #openstack-ansible | 03:44 | |
| *** dxiri has joined #openstack-ansible | 03:52 | |
| *** tux_ has quit IRC | 03:55 | |
| *** dxiri has quit IRC | 03:59 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_sahara master: fix tox python3 overrides https://review.openstack.org/580064 | 04:12 |
| *** Bhujay has joined #openstack-ansible | 04:21 | |
| *** vnogin has joined #openstack-ansible | 04:32 | |
| *** vnogin has quit IRC | 04:36 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_neutron master: fix tox python3 overrides https://review.openstack.org/580058 | 04:41 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-repo_server master: fix tox python3 overrides https://review.openstack.org/580079 | 04:44 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-rabbitmq_server master: fix tox python3 overrides https://review.openstack.org/580076 | 04:48 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-plugins master: fix tox python3 overrides https://review.openstack.org/580075 | 04:53 |
| *** jbadiapa has quit IRC | 05:24 | |
| *** yolanda has joined #openstack-ansible | 05:30 | |
| *** dxiri has joined #openstack-ansible | 05:33 | |
| *** ianychoi has joined #openstack-ansible | 05:34 | |
| *** nicolasbock has joined #openstack-ansible | 05:49 | |
| *** dxiri has quit IRC | 06:06 | |
| *** armaan has joined #openstack-ansible | 06:13 | |
| *** armaan has quit IRC | 06:15 | |
| *** armaan has joined #openstack-ansible | 06:16 | |
| *** armaan has quit IRC | 06:21 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_neutron master: fix tox python3 overrides https://review.openstack.org/580058 | 06:22 |
| *** cshen has joined #openstack-ansible | 06:24 | |
| *** dxiri has joined #openstack-ansible | 06:27 | |
| *** dxiri has quit IRC | 06:31 | |
| *** cjloader has joined #openstack-ansible | 06:32 | |
| *** lvdombrkr has joined #openstack-ansible | 06:32 | |
| *** cjloader has quit IRC | 06:36 | |
| openstackgerrit | Vu Cong Tuan proposed openstack/ansible-config_template master: Replace unicode with six.text_type https://review.openstack.org/580299 | 06:41 |
| hwoarang | good morning | 06:42 |
| hwoarang | i have a question about the dynamic inventory | 06:42 |
| hwoarang | my understsanding is that group/hosts are defined in inventory/env.d/* . How can an operator disable/remove some of these groups? | 06:42 |
| *** maybebuggy has joined #openstack-ansible | 06:42 | |
| hwoarang | i am specifically looking to remove the repo_all group when deploying with distribution packages | 06:42 |
| openstackgerrit | Leopard Ma proposed openstack/openstack-ansible-memcached_server master: Add source code repository notes link to README https://review.openstack.org/579715 | 06:43 |
| *** jbadiapa has joined #openstack-ansible | 06:43 | |
| openstackgerrit | Vu Cong Tuan proposed openstack/openstack-ansible-plugins master: Replace unicode with six.text_type https://review.openstack.org/580301 | 06:49 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_neutron master: fix tox python3 overrides https://review.openstack.org/580058 | 06:59 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-repo_build master: fix tox python3 overrides https://review.openstack.org/580077 | 07:04 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-pip_install master: fix tox python3 overrides https://review.openstack.org/580074 | 07:08 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_rally master: fix tox python3 overrides https://review.openstack.org/580060 | 07:10 |
| *** gkadam has joined #openstack-ansible | 07:11 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_ironic master: fix tox python3 overrides https://review.openstack.org/580024 | 07:12 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_glance master: fix tox python3 overrides https://review.openstack.org/580020 | 07:15 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_designate master: fix tox python3 overrides https://review.openstack.org/580018 | 07:18 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_barbican master: fix tox python3 overrides https://review.openstack.org/580016 | 07:21 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-galera_server master: fix tox python3 overrides https://review.openstack.org/580012 | 07:23 |
| *** vnogin has joined #openstack-ansible | 07:23 | |
| *** fghaas has joined #openstack-ansible | 07:25 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-lxc_container_create master: fix tox python3 overrides https://review.openstack.org/580013 | 07:26 |
| *** epalper has joined #openstack-ansible | 07:26 | |
| *** vnogin has quit IRC | 07:27 | |
| *** fghaas has quit IRC | 07:29 | |
| *** Bhujay has quit IRC | 07:31 | |
| *** tosky has joined #openstack-ansible | 07:32 | |
| *** cjloader has joined #openstack-ansible | 07:32 | |
| *** cjloader has quit IRC | 07:36 | |
| *** DanyC has joined #openstack-ansible | 07:43 | |
| *** DanyC has quit IRC | 07:44 | |
| *** DanyC has joined #openstack-ansible | 07:44 | |
| *** fghaas has joined #openstack-ansible | 07:47 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: Add support for using distribution packages for OpenStack services https://review.openstack.org/579770 | 07:53 |
| *** suggestable has joined #openstack-ansible | 08:01 | |
| evrardjp | gppd morning | 08:06 |
| evrardjp | hwoarang: removal is hard | 08:07 |
| evrardjp | hwoarang: making sure the group is empty is easier | 08:07 |
| *** vnogin has joined #openstack-ansible | 08:07 | |
| evrardjp | two ways to do that | 08:07 |
| evrardjp | 1) do not use the global "infra_hosts" groups, that are spawning entries in subgroups | 08:08 |
| evrardjp | 2) use an affinity of 0 into a container type | 08:08 |
| hwoarang | hmm | 08:11 |
| hwoarang | do you have an example about the 2) option? | 08:12 |
| evrardjp | https://docs.openstack.org/openstack-ansible/latest/reference/inventory/configure-inventory.html#deploying-0-or-more-than-one-of-component-type-per-host | 08:13 |
| evrardjp | hwoarang: ^ | 08:13 |
| hwoarang | thank you. let me try that... | 08:14 |
| *** Bhujay has joined #openstack-ansible | 08:15 | |
| *** fghaas has quit IRC | 08:17 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: Add support for using distribution packages for OpenStack services https://review.openstack.org/579770 | 08:31 |
| *** fghaas has joined #openstack-ansible | 08:38 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_neutron master: fix tox python3 overrides https://review.openstack.org/580058 | 08:45 |
| *** fghaas has quit IRC | 08:58 | |
| *** fghaas has joined #openstack-ansible | 09:15 | |
| odyssey4me | hwoarang: perhaps better to separate out the repo group into its own conf.d file, then only apply it to the scenarios which use it | 09:28 |
| hwoarang | possibly... i am still experimenting :/ | 09:30 |
| *** cjloader has joined #openstack-ansible | 09:32 | |
| odyssey4me | ideally I'd prefer to move away from using the super-group infra_hosts anyway, and instead ensure the use of specific conf.d files for components we want to include | 09:34 |
| odyssey4me | later after M3 that'll be something I work on to allow us to use the integrated repo for role tests | 09:34 |
| *** cjloader has quit IRC | 09:37 | |
| openstackgerrit | Tuan Do Anh proposed openstack/ansible-hardening master: fix tox python3 overrides https://review.openstack.org/579778 | 09:38 |
| hwoarang | odyssey4me: right now i am looking at the simplest and less disruptive way to decouple AIO from repo_container so i can have basekit tests for distro packages | 09:40 |
| hwoarang | i haven't put much thought to make it pretty :/ | 09:41 |
| odyssey4me | in my mind, that is the simplest and least disruptive way ;) but yeah, it'll be interesting to see what you come up with | 09:43 |
| hwoarang | yeah prob because for me the dynamic inventory is complex ;p | 09:44 |
| *** armaan has joined #openstack-ansible | 09:44 | |
| hwoarang | but i think what you say makes sense | 09:45 |
| openstackgerrit | Robert Donovan proposed openstack/openstack-ansible-os_horizon master: Add support for the default_dns_nameservers variable https://review.openstack.org/574227 | 09:46 |
| *** spsurya_ has quit IRC | 09:49 | |
| *** fghaas has quit IRC | 09:50 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/openstack-ansible-os_neutron master: fix tox python3 overrides https://review.openstack.org/580058 | 09:51 |
| *** fghaas has joined #openstack-ansible | 09:54 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Allow inventories with no "properties" https://review.openstack.org/580335 | 10:05 |
| *** armaan has quit IRC | 10:06 | |
| *** armaan has joined #openstack-ansible | 10:07 | |
| *** pingfrog has joined #openstack-ansible | 10:10 | |
| hwoarang | or if one could decouple repo from infra-hosts i can build on top of it... | 10:10 |
| *** armaan has quit IRC | 10:11 | |
| *** jbadiapa has quit IRC | 10:15 | |
| openstackgerrit | Robert Donovan proposed openstack/openstack-ansible-os_nova master: Adds support for the cpu_model customisation in nova.conf https://review.openstack.org/580337 | 10:16 |
| *** Bhujay has quit IRC | 10:19 | |
| evrardjp | well I am building a static inventory, so it would all become simpler. | 10:26 |
| evrardjp | we could then for gate testing use ini_file for patching the file. or config template. | 10:26 |
| evrardjp | no need to care about the dynamic inventory anymore. | 10:26 |
| evrardjp | I could spend cycles on the separation, but there is no point to it in the future, so for now, I'd say go for the past with the least resistence | 10:27 |
| *** cjloader has joined #openstack-ansible | 10:32 | |
| *** cjloader has quit IRC | 10:36 | |
| evrardjp | hwoarang: for your case there is another solution | 10:50 |
| evrardjp | as repo is already kinda conditional | 10:50 |
| evrardjp | you could have an extra condition when building repo_masters group | 10:50 |
| evrardjp | that depends on the global flag install_packages state | 10:51 |
| evrardjp | when install_packages (the global boolean flip) is set to true, you can skip that task, and therefore no hosts will be added to the group, and no repo will be built | 10:51 |
| evrardjp | we only then need to document that people should not BOTH set install_packages to True and a subpart like nova_install_packages to False. | 10:52 |
| evrardjp | that would apply to all users which makes it quite elegant at the end | 10:53 |
| *** Bhujay has joined #openstack-ansible | 10:55 | |
| *** radeks has joined #openstack-ansible | 11:01 | |
| *** radeks has quit IRC | 11:05 | |
| *** radeks has joined #openstack-ansible | 11:06 | |
| *** MB_Adriaan has joined #openstack-ansible | 11:06 | |
| *** suggestable has quit IRC | 11:07 | |
| *** suggestable has joined #openstack-ansible | 11:08 | |
| *** cshen has quit IRC | 11:08 | |
| *** cshen has joined #openstack-ansible | 11:21 | |
| *** armaan has joined #openstack-ansible | 11:22 | |
| *** pingfrog has quit IRC | 11:25 | |
| *** chyka has joined #openstack-ansible | 11:36 | |
| hwoarang | evrardjp: i would need time to parse and implement that . this area of aio is not very clear to me | 11:37 |
| hwoarang | anyway lets make basekit pass first and we will see | 11:38 |
| *** hachi has joined #openstack-ansible | 11:38 | |
| *** radeks has quit IRC | 11:39 | |
| *** chyka has quit IRC | 11:40 | |
| evrardjp | that's the interesting thing: this would not become an AIO thing or a testing thing. With what I proposed a few lines above, it works for all the deployers whatever their env is | 11:41 |
| *** radeks has joined #openstack-ansible | 11:48 | |
| hwoarang | ok | 11:48 |
| mgariepy | morning everyone | 11:54 |
| *** mardim has quit IRC | 11:57 | |
| evrardjp | morning mgariepy | 11:57 |
| *** lvdombrkr89 has joined #openstack-ansible | 11:58 | |
| *** mardim has joined #openstack-ansible | 11:59 | |
| *** lvdombrkr has quit IRC | 11:59 | |
| hwoarang | morning | 12:00 |
| mgariepy | how are you guys doing ? | 12:00 |
| hwoarang | buuuuuusy | 12:09 |
| hwoarang | evrardjp: "[13:50:49] <evrardjp> you could have an extra condition when building repo_masters group" | 12:09 |
| hwoarang | where does that happen? | 12:09 |
| *** admin0 has quit IRC | 12:12 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Be consistent with haproxy group name https://review.openstack.org/580359 | 12:13 |
| evrardjp | hwoarang: https://github.com/openstack/openstack-ansible/blob/master/playbooks/repo-build.yml#L40 | 12:13 |
| evrardjp | just add a when: :) | 12:13 |
| hwoarang | evrardjp: i skip this playbook entirely | 12:14 |
| hwoarang | it's not enough | 12:14 |
| evrardjp | yeah | 12:14 |
| evrardjp | but how do you skip it? | 12:14 |
| hwoarang | when: install_method == source in setup-infrastructure | 12:14 |
| hwoarang | or wherever it's called | 12:14 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Ensure container name doesn't need to be defined https://review.openstack.org/580363 | 12:19 |
| evrardjp | oh ok | 12:19 |
| evrardjp | true | 12:19 |
| evrardjp | and that works? | 12:19 |
| evrardjp | why do you need the group to be not defined/empty ? | 12:20 |
| *** jbadiapa has joined #openstack-ansible | 12:20 | |
| hwoarang | because some things depend on repo container be there. like haproxy configuration. | 12:21 |
| odyssey4me | evrardjp: because the repo container will be built, although that shouldn't be a problem given that the repo server container is used for a package cacher | 12:21 |
| hwoarang | i mean, it's ok to leave repo container build, populate stuff etc. but nothing will use it so here is 20 minutes of wasted time | 12:21 |
| odyssey4me | hwoarang: but if you skip the repo build, does that not work? | 12:22 |
| hwoarang | nope becuase repo_all group is present | 12:22 |
| andymccr | if anybody is about: https://review.openstack.org/#/c/580140/ :D | 12:22 |
| hwoarang | and contains repo_container | 12:22 |
| hwoarang | and haproxy checks the length of that group and tries to do stuff | 12:22 |
| hwoarang | and i hit undefined variables | 12:22 |
| andymccr | also (less urgent from my PoV): https://review.openstack.org/#/c/580128/ | 12:22 |
| hwoarang | a bit of a mess | 12:22 |
| hwoarang | ;p | 12:22 |
| suggestable | Hey guys, question for you. We would prefer to go entirely on metal. If that means having to split out haproxy onto dedicated metal, that's potentially fine, too. Alternatively, what should we look for in a dedicated hardware load balancer? | 12:23 |
| *** hachi has quit IRC | 12:24 | |
| *** hachi has joined #openstack-ansible | 12:25 | |
| odyssey4me | cores - could we get reviews for https://review.openstack.org/#/q/I6aa297ee6e1ca3a76f016a6a0c3d4a735433a9b2 and https://review.openstack.org/#/q/I5a4d4096457d84fcf3a9e21a6873dd61f2a456fc please | 12:26 |
| mgariepy | odyssey4me, done. | 12:28 |
| evrardjp | hwoarang: I am cleaning up stuff, maybe it's in one of my fixes | 12:28 |
| odyssey4me | cores, also could we get eyes on https://review.openstack.org/#/q/topic:move-mq-setup-into-roles+(status:open+OR+status:merged) please | 12:29 |
| odyssey4me | thanks mgariepy | 12:29 |
| evrardjp | done | 12:29 |
| odyssey4me | thanks evrardjp | 12:29 |
| hwoarang | we shall see :) | 12:30 |
| *** radeks has quit IRC | 12:37 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Do not run tasks for containers when on metal https://review.openstack.org/580365 | 12:38 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/pike: Pin get-pip.py to 3.2 https://review.openstack.org/580117 | 12:43 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/ocata: Pin get-pip.py to 3.2 https://review.openstack.org/580118 | 12:43 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/queens: Pin get-pip.py to 3.2 https://review.openstack.org/580116 | 12:43 |
| *** armaan has quit IRC | 12:46 | |
| *** armaan has joined #openstack-ansible | 12:47 | |
| *** radeks has joined #openstack-ansible | 12:47 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: Add support for using distribution packages for OpenStack services https://review.openstack.org/579770 | 12:49 |
| *** mardim_ has joined #openstack-ansible | 12:50 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use a static inventory skeleton https://review.openstack.org/580368 | 12:50 |
| *** mardim has quit IRC | 12:55 | |
| openstackgerrit | Matt Riedemann proposed openstack/openstack-ansible-os_nova master: Run "nova-status upgrade check" during install and upgrade https://review.openstack.org/575125 | 12:56 |
| suggestable | Anyone? evrardjp odyssey4me | 13:11 |
| hwoarang | andymccr: evrardjp re https://review.openstack.org/#/c/580128/ could you tell how does docstring affect the actual code? or tell me what to google for :) | 13:13 |
| evrardjp | suggestable: you really have something against lxc apparently : p | 13:13 |
| evrardjp | anyway if you have two nodes for the haproxy it should be fine | 13:13 |
| evrardjp | you can use these nodes for network nodes if you like | 13:13 |
| andymccr | hwoarang: i pasted teh error in there - im sure we could dig into the ansible internals to find out where it is functionally using that | 13:13 |
| andymccr | *pasted | 13:13 |
| hwoarang | ah it's an ansible thing | 13:14 |
| hwoarang | not a python | 13:14 |
| suggestable | evrardjp: So do you, it seems, from your recent code proposals. Static inventory sounds great to me! | 13:14 |
| *** cmart has joined #openstack-ansible | 13:14 | |
| hwoarang | ok then | 13:14 |
| hwoarang | thank you andymccr | 13:14 |
| evrardjp | suggestable: don't get me wrong: it's possible to do it with existing code, probably submitting a few patches. | 13:14 |
| evrardjp | suggestable: you can for example add extra ips on your management nodes, and make sure the services listen to the right IPs | 13:15 |
| evrardjp | we have a bug open for that if you're ready to tackle it | 13:15 |
| suggestable | evrardjp: What sort of hardware would need to be in the nodes to support the environment I'm deploying? We're running 20Gbps LACP bonds on the controller hosts and 10Gbps links on the compute hosts. | 13:15 |
| andymccr | hwoarang: the exception happens here: https://github.com/ansible/ansible/blob/devel/lib/ansible/config/manager.py#L310 | 13:15 |
| andymccr | but i didn't dig too much further than that | 13:15 |
| hwoarang | ok thank you i was just curious | 13:15 |
| andymccr | no problem, definitely worth asking those questions | 13:15 |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible stable/pike: Install gnocchi before ceilometer https://review.openstack.org/580375 | 13:16 |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible stable/queens: Install gnocchi before ceilometer https://review.openstack.org/580376 | 13:16 |
| *** armaan has quit IRC | 13:16 | |
| *** armaan has joined #openstack-ansible | 13:17 | |
| evrardjp | suggestable: I cannot be the architect of your environement, I don't know your workloads or anything ... | 13:17 |
| evrardjp | only you can now that. | 13:17 |
| evrardjp | know* | 13:17 |
| evrardjp | what I can tell, is that osa has no limitations whatsoever in regards to what you want to use | 13:17 |
| suggestable | evrardjp: OK. How would you suggest I go about getting haproxy to use a specific IP address (or two of them, if necessary) so we can use the three controller hosts as LB hosts? | 13:19 |
| *** maybebuggy has quit IRC | 13:19 | |
| *** throwsb1 has joined #openstack-ansible | 13:19 | |
| suggestable | Simply add the extra IPs to the management interface and set those IPs in the openstack_user_config.yml? | 13:20 |
| suggestable | (This being a purely on-metal deployment) | 13:21 |
| *** armaan has quit IRC | 13:21 | |
| evrardjp | suggestable: well, again, I don't know your deployment. If it's small, yes, I'd patch OSA to make sure everything runs on metal with haproxy on the same node on metal too. | 13:22 |
| *** kouroshvivan has joined #openstack-ansible | 13:23 | |
| suggestable | Our current Icehouse cluster is about 30 nova nodes, running about 300 instances across them. If that's small, then yes, it's a small deployment. | 13:23 |
| evrardjp | suggestable: please also see: https://bugs.launchpad.net/openstack-ansible/+bug/1778914 | 13:24 |
| openstack | Launchpad bug 1778914 in openstack-ansible "galera server binds to 0.0.0.0 address" [Low,Confirmed] | 13:24 |
| *** tux___ has joined #openstack-ansible | 13:24 | |
| evrardjp | suggestable: basically nobody has cared about your use case, because everybody prefer using containers, so isolation is properly done. | 13:24 |
| *** armaan has joined #openstack-ansible | 13:24 | |
| evrardjp | if you're willing to step in and help for your use case, we can help you. | 13:24 |
| evrardjp | I still don't understand why you don't want lxc ocntainers, as you'd be alone in that story, so alone also for the upgrades, which will probably be harder than with containers. | 13:25 |
| evrardjp | but I respect your opinion | 13:26 |
| ioni | is anyone running network nodes on compute nodes? | 13:26 |
| evrardjp | I just want to pinpoint what the reference architecture is | 13:26 |
| evrardjp | and there was reasons for it | 13:26 |
| evrardjp | suggestable: you might also want to read: https://docs.openstack.org/openstack-ansible/latest/reference/architecture/index.html | 13:26 |
| evrardjp | ioni: I guess ppl using DVR : p | 13:27 |
| kouroshvivan | Hello, we have a patch who brake gnocchi installation in master (see: https://bugs.launchpad.net/openstack-ansible/+bug/1778195) fix need some +2, my question is : If I want to backport in stable/pike and stable/queens, i need to cherry-pick the first commit then the fix ? | 13:29 |
| openstack | Launchpad bug 1778195 in openstack-ansible "[ceilometer on gnocchi] install failed" [Undecided,In progress] - Assigned to kourosh vivan (kourosh-vivan) | 13:29 |
| ioni | evrardjp for DVR is openvswitch required or works with linuxbridge as well? | 13:30 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: Disable keepalived ping tests by default https://review.openstack.org/580379 | 13:30 |
| kouroshvivan | dvr use ovs | 13:30 |
| evrardjp | ioni: dvr is an ovs thing. Distributed routers works with lxb too, it's just a different thing. | 13:31 |
| evrardjp | ioni: dvr is... meh. | 13:31 |
| evrardjp | if you can do it differently using lxb, do it. | 13:31 |
| *** ansmith has joined #openstack-ansible | 13:32 | |
| ioni | i'll dig into the docs to see how to enable it with openstack-ansible | 13:32 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: Disable keepalived ping tests by default https://review.openstack.org/580379 | 13:33 |
| evrardjp | ioni: you could also do linux bridge on some of your compute nodes? | 13:34 |
| evrardjp | ioni: dvr is hard to debug. | 13:35 |
| ioni | evrardjp i'm currently all in linux bridge | 13:35 |
| evrardjp | ioni: ok so you can keep it that way : ) | 13:35 |
| evrardjp | ioni: what's the issue? | 13:35 |
| ioni | just currious to see if there are issues with having network nodes on compute nodes | 13:35 |
| evrardjp | ioni: not per se | 13:35 |
| ioni | i just want to have routers mainly on them | 13:36 |
| evrardjp | ioni: but I am not sure it's a good idea -- bad neighbours on one compute node would have an impact on all your cloud networking | 13:36 |
| ioni | dhcp and bgp agents can still be on lxc | 13:36 |
| evrardjp | that sounds like a bad design to me. | 13:36 |
| evrardjp | ioni: I guess it all depends on your nic setup and everything | 13:37 |
| kouroshvivan | I have network node = compute node, no issues | 13:37 |
| evrardjp | there is no technical issue for running compute and network nodes at the same place. | 13:37 |
| evrardjp | for OSA at least | 13:37 |
| evrardjp | after that, what you do, is architecting for your solution : ) | 13:37 |
| evrardjp | kouroshvivan: that's fine for me to do a combined backport if you prefer. But you can also backport separately. | 13:38 |
| evrardjp | Make sure your patches have the information on the cherry pick. | 13:38 |
| evrardjp | if you do a combined backport it also helps on explaining WHY you are doing it. | 13:39 |
| evrardjp | else you can maybe file a bug | 13:39 |
| evrardjp | and reference to it | 13:39 |
| evrardjp | kouroshvivan: it seems you already have your bug open,so you can just reference to it :p | 13:39 |
| kouroshvivan | evardjp: maybe my question is stupid but can i propose a change directly to stable branch ? | 13:42 |
| evrardjp | I am not sure to understand what you imply | 13:42 |
| evrardjp | if a fix need backporting, go from upper branch (master) and then propose the cherry picks on lower branches. | 13:43 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: Disable keepalived ping tests by default https://review.openstack.org/580379 | 13:43 |
| evrardjp | if the patch doesn't apply to master, skip that branch, but apply the same reasoning. | 13:43 |
| tosky | and never merge in 'branch-1' if it's not merged in 'branch' is not merg | 13:43 |
| tosky | or stable maintainers will complain :) | 13:44 |
| kouroshvivan | so only cherry pick from master can go to stable branch ? | 13:44 |
| evrardjp | kouroshvivan: we accept re-implementations if it is well explained why it needed re-implementing. | 13:44 |
| evrardjp | (impossible to cherry pick without bringing features for example) | 13:45 |
| evrardjp | I guess it all depends on the patch : ) | 13:45 |
| evrardjp | but the idea is there | 13:45 |
| evrardjp | we respect standard openstack behaviours. | 13:45 |
| kouroshvivan | ok nicest way is combined backport, how to do it ? | 13:45 |
| mnaser | odyssey4me (and other cores): https://review.openstack.org/#/c/579941 and https://review.openstack.org/#/c/579944 are ready and passing and should clean things up in the role and unbreak centos gate | 13:46 |
| evrardjp | on your environement, you apply both patches into the same commir | 13:46 |
| evrardjp | commit | 13:46 |
| *** cshen has quit IRC | 13:46 | |
| evrardjp | making sure the cherry-picked from ... message is there. | 13:46 |
| odyssey4me | mnaser: I'll trade you :) https://review.openstack.org/#/q/topic:move-mq-setup-into-roles+(status:open+OR+status:merged) | 13:47 |
| evrardjp | mnaser: isn't cinder moving to privsep? | 13:47 |
| *** tux___ has quit IRC | 13:47 | |
| mnaser | evrardjp: all projects will eventually move to that but until they do we need to make sure we have the right rootwrap config files | 13:48 |
| mnaser | if we don't, then tests dont pass :< | 13:48 |
| mnaser | odyssey4me: let me see | 13:48 |
| evrardjp | mnaser: FYI, we are carrying those files because 1) we modify/modified them 2) Jesse's patches to ship them inside packages weren't merged | 13:48 |
| evrardjp | mnaser: if they are unmodified vs upstream, I am fine to not carry those : ) | 13:48 |
| odyssey4me | evrardjp: we don't modify the rootwrap filters at all - we used to | 13:49 |
| evrardjp | odyssey4me: "modified" | 13:49 |
| evrardjp | :p | 13:49 |
| odyssey4me | also, almost every single project now has them built into the venv :) | 13:49 |
| evrardjp | odyssey4me: that's point 2 | 13:49 |
| evrardjp | ppl should read my sentences until the end : p | 13:50 |
| evrardjp | or maybe I should just be clearer, I have to think about it. | 13:50 |
| mnaser | odyssey4me: i like the idea of syncing that file | 13:51 |
| mnaser | i really don't want to add more work but is there any reason why we didnt go the role path? | 13:51 |
| odyssey4me | "go to the role path" ? | 13:52 |
| mnaser | as in, create a role called.. "openstack_mq" or "oslo_messaging" | 13:52 |
| mnaser | and do an include_role rather than include a list of tasks | 13:52 |
| odyssey4me | mnaser: because it's not worth the extra weight - with this we can use the proposal bot to keep them in sync | 13:53 |
| odyssey4me | if that task list grows, which is might, then it might be worth splitting out | 13:53 |
| evrardjp | mnaser: if there is no variables defaults, no need for a role | 13:53 |
| mnaser | yeah i'm totally okay with that too. plus include_role is slowwwww | 13:53 |
| evrardjp | mnaser: and buggy. | 13:53 |
| *** cshen has joined #openstack-ansible | 13:53 | |
| mnaser | yeah cloudnull apparently found out that `private` didnt really mean what it is | 13:54 |
| evrardjp | mnaser: well it never really worked | 13:54 |
| evrardjp | as far as I am aware | 13:54 |
| mnaser | it was just a prank | 13:54 |
| evrardjp | mnaser: what did cloudnull discover though? | 13:54 |
| evrardjp | I am not aware of this | 13:54 |
| evrardjp | mnaser: I reviewed https://review.openstack.org/#/c/579944 could you explain the WHY of the change to use synchronize and delegation? | 13:55 |
| odyssey4me | evrardjp: looks to me like this is the only tweak being done to a rootwrap filter, which may not even be relevant any more: https://github.com/openstack/openstack-ansible/blob/master/scripts/sources-branch-updater-lib.sh#L172-L176 | 13:55 |
| evrardjp | odyssey4me: thanks for checking. | 13:55 |
| mnaser | i dont know the details but they found some ansible bug related use of prifvate | 13:55 |
| evrardjp | odyssey4me: we should clean up that script then | 13:55 |
| evrardjp | odyssey4me: yeah that's what I remember with the path : p | 13:55 |
| odyssey4me | evrardjp mnaser basically the private implementation was merged into docs, but not into code - and is still not in code | 13:56 |
| evrardjp | odyssey4me: but I think there was some with the exec dir though? | 13:56 |
| evrardjp | odyssey4me: haha | 13:56 |
| evrardjp | ok | 13:56 |
| odyssey4me | include_role is currently "private: yes" as it stands, whereas "roles: foo" if "private: no" as it stands | 13:56 |
| evrardjp | I don't understand that sentence. | 13:57 |
| odyssey4me | ie if you use include_role, any facts set in the role do not leave the role | 13:57 |
| evrardjp | mmm | 13:57 |
| odyssey4me | if you use the older roles mechanism, the facts set in the role are available after the role execution in the same play | 13:57 |
| evrardjp | not so sure about that : p | 13:57 |
| evrardjp | but ok | 13:57 |
| evrardjp | if you say so | 13:57 |
| evrardjp | I will try just to be sure | 13:57 |
| odyssey4me | well, that's what the bug said | 13:57 |
| evrardjp | ok | 13:58 |
| mnaser | evrardjp: copy with remote_src does not work because i cant "copy all files inside folder" and i dont want to add a list of files there | 13:58 |
| odyssey4me | I haven't validated | 13:58 |
| mnaser | cause if i add a list | 13:58 |
| mnaser | then we might run into the same silly issue | 13:58 |
| guilhermesp | that's why I spent a lot of time adapting the tests becasue the fake feature of include_role | 13:58 |
| mnaser | if a project decides to add new file | 13:58 |
| odyssey4me | synchronize is better because it will also remove files ;) | 13:58 |
| mnaser | yup so effectively cleaning things up | 13:58 |
| mnaser | so the state will be clean | 13:58 |
| guilhermesp | unfortunatly the os_monasca will not use it as the other roles | 13:59 |
| guilhermesp | mnaser: can you show me the ansible bug? I think it was the one that I commented | 13:59 |
| evrardjp | mnaser: I am not against the change (while not really FOR it) -- I am against the fact it's not documented we've changed minds. | 13:59 |
| mnaser | guilhermesp: i dont know the details much, i just saw talk here in the channel | 13:59 |
| mnaser | evrardjp: can you tell me what you need to get a +2 | 14:00 |
| mnaser | i want to fix centos. | 14:00 |
| evrardjp | mnaser: we should say it's because we want to ensure folder contents, not only copying the files. | 14:00 |
| mnaser | OK. so change the commit message?! | 14:00 |
| evrardjp | (including deletion if need be) | 14:00 |
| evrardjp | mnaser: I can change it if you like. | 14:00 |
| evrardjp | you prefer that way? | 14:00 |
| mnaser | no, ill change it, ask for odyssey4me to +2 again and wait for the whole check/gate dance | 14:00 |
| mnaser | evrardjp: if you want to edit it now | 14:01 |
| mnaser | i'd appreciate it | 14:01 |
| mnaser | and then i'll try to get more reviews and chase it through the gate. | 14:01 |
| evrardjp | trust me 6 months from now when ansible synchronize and delegation will fail, we'll check at this and all the others. | 14:01 |
| evrardjp | hahaha | 14:01 |
| guilhermesp | mnaser: https://github.com/ansible/ansible/issues/21890 | 14:02 |
| guilhermesp | FYI | 14:02 |
| mnaser | pretty stable api, dont think it's going to be breaking seriously anytime soon on such a common usecase | 14:02 |
| evrardjp | mnaser: ok | 14:02 |
| evrardjp | I am editing | 14:02 |
| odyssey4me | evrardjp: by then I'm hoping we won't need to bother with all that, because source builds will just use a symlink from /etc/<service> to /openstack/venv/<service>-<tag>/etc/<service> | 14:02 |
| odyssey4me | or something to that effect | 14:02 |
| odyssey4me | instead of copying things around, we just use symlinks to the venv contents - this'll enable roll-back for minor upgrades or other changes | 14:03 |
| odyssey4me | but hey, I'm getting ahead of myself | 14:03 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_cinder master: Use rootwrap configuration files from the virtualenv https://review.openstack.org/579944 | 14:04 |
| mnaser | odyssey4me: ^ | 14:05 |
| kouroshvivan | can I put two #bug in Closes-Bug ? If yes what is the separator ? | 14:05 |
| evrardjp | kouroshvivan: yes | 14:05 |
| evrardjp | two lines | 14:05 |
| mnaser | kouroshvivan: you put two Closes-Bug lines | 14:05 |
| kouroshvivan | thanks | 14:05 |
| mnaser | Closes-Bug: #1 <enter> Closes-Bug: #2 | 14:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: Enable CentOS 7 distro jobs https://review.openstack.org/579945 | 14:05 |
| *** vnogin has quit IRC | 14:06 | |
| *** vnogin has joined #openstack-ansible | 14:06 | |
| mnaser | thanks odyssey4me | 14:06 |
| evrardjp | mnaser: if you still believe the ansible api and implementation is stable accross versions, I will give you a euro next time we meet. | 14:06 |
| mnaser | and evrardjp | 14:06 |
| evrardjp | (or a beverage of your choice) | 14:06 |
| evrardjp | mnaser: the problem is that synchronize and delegation was poorly tested. | 14:06 |
| odyssey4me | hahaha | 14:06 |
| odyssey4me | it's getting better | 14:06 |
| mnaser | btw | 14:06 |
| evrardjp | odyssey4me: yeah due to massive testing, starting by gundalow and co | 14:07 |
| mnaser | yeah i agree i think its a progress | 14:07 |
| mnaser | i dont have time to look into this right now | 14:07 |
| mnaser | but bionic + rabbitmq seems to fail | 14:07 |
| evrardjp | I just can't give 100% certainty that in 6 months it's still gonna work like it is now : p | 14:07 |
| evrardjp | anyway. | 14:07 |
| mnaser | so if someone can dig into that it will help with some of the other rabbitmq project efforts | 14:07 |
| evrardjp | I am done fighting. | 14:07 |
| evrardjp | mnaser: link? | 14:08 |
| evrardjp | also I thought the copy module was recursive | 14:09 |
| odyssey4me | https://media.giphy.com/media/3xC2k2iGO40Io/giphy.gif | 14:09 |
| evrardjp | I will double check. | 14:09 |
| mnaser | http://logs.openstack.org/54/575954/5/check/openstack-ansible-functional-ubuntu-bionic/f194719/logs/ara-report/ | 14:09 |
| mnaser | any bionic job | 14:09 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/pike: Pin get-pip.py to 3.2 https://review.openstack.org/580138 | 14:09 |
| *** Bhujay has quit IRC | 14:09 | |
| evrardjp | odyssey4me: I don't get it , the pythons have hurt your head too much : p | 14:10 |
| *** tux_ has joined #openstack-ansible | 14:10 | |
| evrardjp | odyssey4me: what's the reference? | 14:10 |
| evrardjp | formy culture | 14:10 |
| odyssey4me | evrardjp: the ppl stare at him because he doesn't look like a newt... so he says 'I got better.' | 14:10 |
| odyssey4me | it's like ansible :p | 14:10 |
| evrardjp | mnaser: thanks. | 14:10 |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible-os_ceilometer stable/pike: Adds ceilometer-upgrade task https://review.openstack.org/580390 | 14:11 |
| evrardjp | odyssey4me: hahaha | 14:11 |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible-os_ceilometer stable/queens: Adds ceilometer-upgrade task https://review.openstack.org/580391 | 14:12 |
| cloudnull | mornings | 14:16 |
| *** cjloader has joined #openstack-ansible | 14:20 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Be consistent with memcached group name https://review.openstack.org/580394 | 14:23 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/ocata: Pin get-pip.py to 3.2 https://review.openstack.org/580139 | 14:24 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/queens: Pin get-pip.py to 3.2 https://review.openstack.org/580137 | 14:24 |
| evrardjp | morning cloudnull | 14:24 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Pin get-pip.py to 3.2 https://review.openstack.org/580398 | 14:29 |
| *** armaan has quit IRC | 14:29 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Pin get-pip.py to 3.2 https://review.openstack.org/580399 | 14:32 |
| jrosser | cloudnull: with the elk stuff do the container journals somehow end up on the host? | 14:41 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Copy rootwrap.d config files in source installs only https://review.openstack.org/579941 | 14:45 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Add example host confd file for barbican https://review.openstack.org/579281 | 14:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Pin get-pip.py to 3.2 https://review.openstack.org/580402 | 14:46 |
| *** ansmith has quit IRC | 14:48 | |
| odyssey4me | mnaser evrardjp hwoarang oh, by the way - it seems that xenial is the officially tested platform for openstack in general for ubuntu for rocky, and it's most likely that they'll only start removing xenial in stein: http://lists.openstack.org/pipermail/openstack-dev/2018-June/131194.html | 14:52 |
| odyssey4me | so we can add bionic support during rocky and we should be fine | 14:53 |
| evrardjp | odyssey4me: that's what we are doing, and what we decided? : ) | 14:55 |
| odyssey4me | evrardjp: yep, but it's nice to know that we aren't an outlier this time | 14:55 |
| evrardjp | haha | 14:55 |
| evrardjp | yeah | 14:55 |
| evrardjp | I'd need your help figuring something out odyssey4me | 14:55 |
| odyssey4me | evrardjp sure, I have a meeting for an hour in 5 ins, then can help out | 14:56 |
| odyssey4me | *mins | 14:56 |
| evrardjp | ok | 14:56 |
| cloudnull | jrosser they should end up on the host however that may need tweaking. | 15:06 |
| jrosser | cloudnull: hmm yes i think i'm missing journal logs on the host from in the containers | 15:08 |
| *** armaan has joined #openstack-ansible | 15:08 | |
| hwoarang | odyssey4me: but but will there be packages? | 15:09 |
| jrosser | cloudnull: i recall you looking at journal forwarding, but what do you think the best approach is | 15:10 |
| * hwoarang afk | 15:10 | |
| cloudnull | we probably just need to see what's missing in the base setup. | 15:11 |
| cloudnull | I suspect I simply overlooked a step | 15:12 |
| jrosser | yes, i wasnt sure if it was expected that they all went to the host, or if journalbeat needs to land everywhere | 15:12 |
| cloudnull | I think in an older deployment (starting with newton) it might make some sense to install journalbeat everywhere. | 15:16 |
| cloudnull | I'd have to go look but I believe the container journals are on the host in queens + | 15:16 |
| *** chyka has joined #openstack-ansible | 15:17 | |
| *** fghaas has quit IRC | 15:20 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: Move MQ vhost/user creation into role https://review.openstack.org/568517 | 15:21 |
| *** fghaas has joined #openstack-ansible | 15:21 | |
| jrosser | cloudnull: if i use lsof on the journalbeat process i see it only has journals open for the host machine-id | 15:23 |
| cloudnull | do you see the containers machine-id's in /var/log/journal ? | 15:24 |
| jrosser | yes i've got a whole bunch of dirs in there | 15:24 |
| jrosser | although they dont appear to line up with /etc/machine-id from inside the containers | 15:26 |
| openstackgerrit | jacky06 proposed openstack/openstack-ansible-ops master: Pin get-pip.py to 3.2 https://review.openstack.org/580407 | 15:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_keystone master: Ensure __pycache__ folder is deleted https://review.openstack.org/580408 | 15:30 |
| *** gyee has joined #openstack-ansible | 15:31 | |
| openstackgerrit | jacky06 proposed openstack/openstack-ansible-repo_build master: Pin get-pip.py to 3.2 https://review.openstack.org/580409 | 15:31 |
| *** epalper has quit IRC | 15:35 | |
| *** radeks has quit IRC | 15:35 | |
| *** srf_ has joined #openstack-ansible | 15:36 | |
| srf_ | hello everyone, can I ask, when I install packstack I have some problem in "Availability Zone", the error say :"There are no Availability Zones." | 15:36 |
| srf_ | whether this person has experienced like this and how to overcome it, thx | 15:36 |
| srf_ | the error i found in packstack | 15:37 |
| *** kouroshvivan has quit IRC | 15:38 | |
| evrardjp | srf_: well , we don't use packstack to deploy so I doubt we can be very helpful | 15:47 |
| evrardjp | maybe someone in #openstack or #openstack-dev can help | 15:47 |
| evrardjp | I can't at least. | 15:47 |
| *** dxiri has joined #openstack-ansible | 15:51 | |
| *** fghaas has quit IRC | 15:51 | |
| srf_ | evrardjp : okay thx you, but ya, i ask too in #openstack and #openstack-dev, no problem :) | 15:52 |
| *** radeks has joined #openstack-ansible | 15:55 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Collect journals from containers as well as the host https://review.openstack.org/580433 | 15:57 |
| jrosser | cloudnull: ^^ becasue of https://github.com/mheese/journalbeat/blob/master/beater/journalbeat.go#L59-L83 | 15:58 |
| cloudnull | jrosser nice! | 15:59 |
| jrosser | slightly eccentric behaviour in the config file there | 16:00 |
| openstackgerrit | jacky06 proposed openstack/openstack-ansible master: Pin get-pip.py to 3.2 https://review.openstack.org/580436 | 16:00 |
| *** cjloader_ has joined #openstack-ansible | 16:02 | |
| *** cjloader has quit IRC | 16:02 | |
| *** udesale has quit IRC | 16:04 | |
| *** cshen has quit IRC | 16:05 | |
| *** fghaas has joined #openstack-ansible | 16:06 | |
| *** cjloader_ has quit IRC | 16:08 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Update all SHAs for milestone 3 https://review.openstack.org/574006 | 16:08 |
| *** srf_ has quit IRC | 16:08 | |
| *** yolanda_ has joined #openstack-ansible | 16:13 | |
| evrardjp | cloudnull: I am happy if you submit patches for https://review.openstack.org/#/c/574006/7/playbooks/defaults/repo_packages/openstack_services.yml but you're breaking the things at the same time | 16:14 |
| evrardjp | I don't know how you bump things | 16:14 |
| *** lvdombrkr89 has quit IRC | 16:14 | |
| evrardjp | but it doesn't respect the comments # FROZEN | 16:15 |
| cloudnull | what? | 16:15 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs for milestone 3 https://review.openstack.org/574006 | 16:15 |
| evrardjp | cloudnull: I've fixed it ^ | 16:15 |
| evrardjp | you can check the diff | 16:15 |
| cloudnull | oh, what is #FROZEN ? | 16:16 |
| *** cjloader has joined #openstack-ansible | 16:16 | |
| cloudnull | I just ran the script | 16:16 |
| evrardjp | networking-bgpvpn and networking-odl | 16:16 |
| *** suggestable has quit IRC | 16:16 | |
| evrardjp | which script? | 16:16 |
| *** yolanda has quit IRC | 16:17 | |
| evrardjp | I haven't merged my ugly spaghetti with what's inside openstack/openstack-ansible (and in fact, we should remove that code) | 16:17 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/scripts/sources-branch-updater.sh | 16:18 |
| evrardjp | we have to refactor releases tooling -- as we now release differently. | 16:18 |
| cloudnull | ah. | 16:18 |
| evrardjp | and there are these cases that are not handled in that script. | 16:18 |
| cloudnull | ok. | 16:18 |
| evrardjp | we could also fix things upstream, but I haven't got the chance. I hope mbuil would eventually do it | 16:20 |
| evrardjp | but I need to remember him | 16:20 |
| *** radeks has quit IRC | 16:22 | |
| cloudnull | ok | 16:23 |
| evrardjp | so cloudnull who will you support in the World cup now? | 16:24 |
| evrardjp | stopped watching it? | 16:24 |
| cloudnull | france and belgium! | 16:24 |
| evrardjp | friday! | 16:24 |
| cloudnull | indeed! | 16:25 |
| evrardjp | let's hope for the miracle for the 6PM UTC match, if you see what I mean ;) | 16:26 |
| *** cjloader_ has joined #openstack-ansible | 16:27 | |
| *** cjloader has quit IRC | 16:27 | |
| *** radeks has joined #openstack-ansible | 16:30 | |
| guilhermesp | noooo brazil on friday evrardjp cloudnull you two against me hahaha | 16:39 |
| evrardjp | haha guilhermesp :) | 16:40 |
| openstackgerrit | Merged openstack/openstack-ansible master: Install gnocchi before ceilometer https://review.openstack.org/577379 | 16:45 |
| *** armaan_ has joined #openstack-ansible | 16:45 | |
| *** armaan has quit IRC | 16:45 | |
| jrosser | cloudnull: whilst that patch opens all the journal files something is still wrong | 16:46 |
| jrosser | the host journal dir names are not matching all of my container machine-id's | 16:47 |
| jrosser | and the majority of them are stale in some way | 16:47 |
| *** armaan_ has quit IRC | 16:50 | |
| *** cjloader_ has quit IRC | 16:51 | |
| *** cjloader has joined #openstack-ansible | 16:53 | |
| *** cjloader has quit IRC | 16:53 | |
| *** yolanda__ has joined #openstack-ansible | 16:53 | |
| *** cjloader has joined #openstack-ansible | 16:53 | |
| *** yolanda_ has quit IRC | 16:56 | |
| *** suggestable has joined #openstack-ansible | 16:59 | |
| *** suggestable has quit IRC | 17:07 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Collect journals from containers as well as the host https://review.openstack.org/580433 | 17:09 |
| *** yolanda_ has joined #openstack-ansible | 17:12 | |
| *** yolanda__ has quit IRC | 17:15 | |
| *** ansmith has joined #openstack-ansible | 17:28 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tempest master: tasks: tempest_install: Add conditional for PIP package installation https://review.openstack.org/580455 | 17:28 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 17:40 |
| *** cshen has joined #openstack-ansible | 17:58 | |
| *** mattinator has joined #openstack-ansible | 18:01 | |
| *** mattinator has quit IRC | 18:05 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Sync task files to roles https://review.openstack.org/579527 | 18:06 |
| odyssey4me | hwoarang could you take a peek at https://review.openstack.org/579527 to validate whether it will work and do the right thing? | 18:07 |
| *** poopcat has joined #openstack-ansible | 18:09 | |
| *** yolanda__ has joined #openstack-ansible | 18:13 | |
| *** mattinator has joined #openstack-ansible | 18:15 | |
| *** yolanda_ has quit IRC | 18:15 | |
| *** vnogin has quit IRC | 18:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Move MQ vhost/user creation into role (glance) https://review.openstack.org/568522 | 18:21 |
| *** armaan has joined #openstack-ansible | 18:32 | |
| *** yolanda_ has joined #openstack-ansible | 18:35 | |
| *** DanyC has quit IRC | 18:38 | |
| *** yolanda__ has quit IRC | 18:39 | |
| *** chris__ has joined #openstack-ansible | 18:39 | |
| *** dxiri has quit IRC | 18:39 | |
| *** yolanda__ has joined #openstack-ansible | 18:39 | |
| chris__ | Hey Guys, does anyone know how to modify the console_token_ttl expiration time in openstack-ansible queens? | 18:40 |
| *** yolanda_ has quit IRC | 18:42 | |
| *** armaan has quit IRC | 18:44 | |
| *** armaan has joined #openstack-ansible | 18:45 | |
| *** vnogin has joined #openstack-ansible | 18:47 | |
| *** armaan has quit IRC | 18:49 | |
| *** vnogin has quit IRC | 18:52 | |
| *** cshen has quit IRC | 19:00 | |
| *** tomtom002 has joined #openstack-ansible | 19:05 | |
| *** dxiri has joined #openstack-ansible | 19:07 | |
| *** tomtom001 has quit IRC | 19:08 | |
| *** cmart has quit IRC | 19:21 | |
| *** Chealion has quit IRC | 19:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: On container destroy remove container journal https://review.openstack.org/580474 | 19:23 |
| cloudnull | jrosser ^ | 19:24 |
| *** Chealion has joined #openstack-ansible | 19:24 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Move MQ vhost/user creation into role (glance) https://review.openstack.org/568531 | 19:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Ensure remote_user and become are consistently implemented https://review.openstack.org/580156 | 19:27 |
| *** radeks has quit IRC | 19:27 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cloudkitty master: Execute service setup against a delegated host using Ansible built-in modules https://review.openstack.org/579977 | 19:29 |
| *** gkadam has quit IRC | 19:32 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_congress master: Execute service setup against a delegated host using Ansible built-in modules https://review.openstack.org/579981 | 19:34 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Use openstack_version to determine Journalbeat hosts https://review.openstack.org/580476 | 19:34 |
| jrosser | cloudnull: nice, will try that. one back for you ^ | 19:34 |
| *** cshen has joined #openstack-ansible | 19:35 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican master: Execute service setup against a delegated host using Ansible built-in modules https://review.openstack.org/579947 | 19:36 |
| cloudnull | jrosser giving that a spin now | 19:39 |
| jrosser | same, mines just running now | 19:40 |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_monasca master: [WIP] Fix monasca https://review.openstack.org/578721 | 19:40 |
| *** cshen has quit IRC | 19:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: Execute service setup against a delegated host using Ansible built-in modules https://review.openstack.org/579991 | 19:41 |
| cloudnull | jrosser it works! however I think we'll need to create a pre-playbook to dynamically create the group | 19:46 |
| cloudnull | I had to run `openstack-ansible /tmp/test.yml -e openstack_release=17` | 19:46 |
| cloudnull | to make it detect the openstack_release variable | 19:46 |
| *** cmart has joined #openstack-ansible | 19:47 | |
| jrosser | ah did you see what i put in the readme? | 19:51 |
| *** poopcat has quit IRC | 19:51 | |
| jrosser | if you can think of a better way than that then cool | 19:51 |
| *** poopcat has joined #openstack-ansible | 19:52 | |
| *** poopcat has quit IRC | 19:54 | |
| cloudnull | http://paste.openstack.org/show/725142/ | 20:00 |
| cloudnull | That picks up the embedded variable when coupled along with osa | 20:01 |
| cloudnull | then i changed teh hosts: section in the playbook to use all_journalbeat_deployments | 20:01 |
| jrosser | cloudnull: i'll give that a go | 20:03 |
| cloudnull | http://paste.openstack.org/show/725143/ | 20:04 |
| cloudnull | that was the whole playbook | 20:04 |
| jrosser | how does it get openstack_release? | 20:05 |
| *** vnogin has joined #openstack-ansible | 20:07 | |
| cloudnull | The group vars are exposed to the tasks of a playbook. | 20:08 |
| jrosser | cloudnull: the version is in here /opt/openstack-ansible/inventory/group_vars/all/all.yml and i dont think that is included? | 20:08 |
| cloudnull | It was when running with `openstack-ansible` | 20:08 |
| cloudnull | With the embedded ansible it would not be. | 20:09 |
| cloudnull | So the option would still be required. | 20:09 |
| cloudnull | In that case. | 20:09 |
| jrosser | ah hmm right because for me the embedded is the way to get 2.5 | 20:10 |
| cloudnull | ++ I'm running with Rocky at the moment. | 20:11 |
| cloudnull | So i think we still need the blerb in the readme to ensure all edge cases are accounted for. | 20:12 |
| *** poopcat has joined #openstack-ansible | 20:12 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fix quota for Octavia https://review.openstack.org/571368 | 20:12 |
| odyssey4me | cloudnull: https://media.giphy.com/media/DwIdasRkFKsMg/giphy.gif | 20:12 |
| cloudnull | ? Lol | 20:13 |
| odyssey4me | :) running rocky, not really something that jrosser can afford to do :p | 20:13 |
| odyssey4me | right, time for me to !computer - cheers all | 20:14 |
| cloudnull | take care odyssey4me | 20:25 |
| *** poopcat1 has joined #openstack-ansible | 20:40 | |
| *** throwsb1 has quit IRC | 20:41 | |
| *** poopcat has quit IRC | 20:42 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Use openstack_version to determine Journalbeat hosts https://review.openstack.org/580476 | 20:48 |
| jrosser | cloudnull: ^ your change works for me | 20:49 |
| jrosser | i do notice that TASK [Enable and restart journalbeat] is quick on my containers and exceedingly slow on the hosts | 20:50 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Use rootwrap configuration files from the virtualenv https://review.openstack.org/579944 | 21:06 |
| *** albertcard2 has joined #openstack-ansible | 21:10 | |
| *** poopcat1 has quit IRC | 21:13 | |
| *** albertcard2 is now known as poopcat | 21:16 | |
| *** poopcat is now known as Guest72776 | 21:17 | |
| *** cjloader_ has joined #openstack-ansible | 21:18 | |
| *** cjloader has quit IRC | 21:18 | |
| *** cjloader_ has quit IRC | 21:18 | |
| *** cjloader has joined #openstack-ansible | 21:18 | |
| *** Guest72776 has quit IRC | 21:19 | |
| *** poopcat1 has joined #openstack-ansible | 21:19 | |
| mnaser | yay | 21:20 |
| mnaser | that merged | 21:20 |
| mnaser | i think this will finally make centos metal GREEN | 21:21 |
| *** armaan has joined #openstack-ansible | 21:23 | |
| jrosser | cloudnull: this journal collection stuff is all missing tags | 21:25 |
| jrosser | i wonder if taking the first part of {foo}.slice as the tag is enough | 21:26 |
| *** tux_ has quit IRC | 21:28 | |
| cloudnull | jrosser ? odd that container would be quick and slow on the host ? | 21:28 |
| jrosser | yes i don't understand that | 21:29 |
| cloudnull | are you seeing missing field tags or just tags in general ? | 21:29 |
| jrosser | although we've done little with ansible 2.5 and i'm expecting the usual surprises | 21:29 |
| jrosser | for things that are collected from the journal they're not tagged in the same way as the log files are | 21:30 |
| jrosser | so no "tags openstack, nova, beats_input_codec_plain_applied" | 21:31 |
| cloudnull | ah yes. thats largely coming from the logstash groks | 21:31 |
| jrosser | so it's sort of there | 21:31 |
| jrosser | in the systemd slice name | 21:32 |
| cloudnull | I wonder if we could extend those groks to do that ? | 21:32 |
| jrosser | i was wondering if that was an easy way to do the same | 21:32 |
| jrosser | but it does highlight some missing slice names | 21:32 |
| *** cjloader has quit IRC | 21:33 | |
| *** cshen has joined #openstack-ansible | 21:36 | |
| *** cshen has quit IRC | 21:41 | |
| cloudnull | hum | 21:41 |
| cloudnull | looking at the journalbeat data and the config file IDK if we can apply that same logic there. | 21:42 |
| *** ivve has quit IRC | 21:55 | |
| *** vnogin has quit IRC | 21:55 | |
| *** vnogin has joined #openstack-ansible | 22:04 | |
| *** chris__ has quit IRC | 22:04 | |
| *** cmart has quit IRC | 22:09 | |
| *** albertcard2 has joined #openstack-ansible | 22:10 | |
| *** poopcat1 has quit IRC | 22:13 | |
| *** nicolasbock has quit IRC | 22:17 | |
| *** fghaas has quit IRC | 22:23 | |
| *** hachi has quit IRC | 22:30 | |
| cloudnull | jrosser looks like we need to change the service units to have `--use-journal` | 22:35 |
| cloudnull | http://paste.openstack.org/show/725160/ | 22:36 |
| *** armaan has quit IRC | 22:36 | |
| cloudnull | which requires `apt install pkg-config` and `/openstack/venvs/nova-17.0.5/bin/pip install systemd-python` | 22:36 |
| *** armaan has joined #openstack-ansible | 22:37 | |
| cloudnull | I think I can make this all part of the repo server, its something that I will need to look into regardless, especially if the services are not actually loogging to the journal | 22:37 |
| *** armaan has quit IRC | 22:41 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Pin get-pip.py to 3.2 https://review.openstack.org/580398 | 22:47 |
| *** chyka has quit IRC | 22:57 | |
| *** tosky has quit IRC | 23:01 | |
| *** ianychoi_ has joined #openstack-ansible | 23:01 | |
| *** ianychoi has quit IRC | 23:04 | |
| *** yolanda_ has joined #openstack-ansible | 23:07 | |
| *** yolanda__ has quit IRC | 23:08 | |
| *** yolanda__ has joined #openstack-ansible | 23:11 | |
| *** yolanda_ has quit IRC | 23:13 | |
| *** vnogin has quit IRC | 23:15 | |
| *** cshen has joined #openstack-ansible | 23:37 | |
| *** yolanda_ has joined #openstack-ansible | 23:40 | |
| *** poopcat1 has joined #openstack-ansible | 23:40 | |
| *** cshen has quit IRC | 23:42 | |
| *** yolanda__ has quit IRC | 23:42 | |
| *** albertcard2 has quit IRC | 23:43 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
