| *** openstack has joined #openstack-ansible | 07:32 | |
| *** ChanServ sets mode: +o openstack | 07:32 | |
| *** shyamb has joined #openstack-ansible | 07:34 | |
| *** tosky has joined #openstack-ansible | 07:37 | |
| *** also_stingrayza is now known as stingrayza | 07:44 | |
| *** shyamb has quit IRC | 08:02 | |
| *** shyamb has joined #openstack-ansible | 08:09 | |
| *** shyamb has quit IRC | 08:17 | |
| admin0 | morning | 08:19 |
|---|---|---|
| admin0 | stable/train .. containers are created but they only have 172.29.236.x mgmt ips .. no resolver.conf so dns resolution is failing .. there is no default gateway or route either | 08:20 |
| admin0 | i.e they are missing the lxcbr0 ip range of the 10.0.3.x | 08:21 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove package cache proxy cleanup https://review.opendev.org/742098 | 08:29 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove fastest-infra-wheel-mirror.py https://review.opendev.org/742099 | 08:29 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove Centos-7 support https://review.opendev.org/742100 | 08:29 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Remove Centos-7 support https://review.opendev.org/742103 | 08:43 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Update CI tests for currently supported operating systems https://review.opendev.org/737982 | 08:52 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Remove Centos-7 support https://review.opendev.org/742104 | 08:52 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Update galera to 10.5.4 https://review.opendev.org/742105 | 08:54 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Remove Centos-7 support https://review.opendev.org/742103 | 08:55 |
| admin0 | what could be the reason that lxc containers are missing the 10.x ip and only mgmt ip is present ? | 08:56 |
| jrosser | do they have an eth0? | 08:57 |
| jrosser | they get the 10.x IP from the lxc dnsmasq process on the host by dhcp | 08:58 |
| jrosser | so you should check that is all running OK and nothing bad in the logs of lxc-dnsmasq | 08:58 |
| jrosser | and you can try to re-dhcp from in the container with 'networkctl reconfigure eth0' | 08:58 |
| admin0 | thanks jrosser .... | 08:59 |
| admin0 | they do have eth0 .. maybe something wrong with the dnsmasq | 08:59 |
| admin0 | hmm.. lxc-dnsmasq dies for some reason | 09:01 |
| admin0 | jrosser, thanks .. solved | 09:14 |
| *** markvoelker has joined #openstack-ansible | 09:16 | |
| *** markvoelker has quit IRC | 09:21 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Remove workaround for Centos-7 log socket https://review.opendev.org/742111 | 09:28 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update ansible metadata to indicate support for centos-8 https://review.opendev.org/742112 | 09:28 |
| *** shyamb has joined #openstack-ansible | 09:36 | |
| *** arkan has quit IRC | 09:47 | |
| *** mmethot_ has joined #openstack-ansible | 10:08 | |
| *** shyam89 has joined #openstack-ansible | 10:08 | |
| *** mmethot has quit IRC | 10:10 | |
| *** shyamb has quit IRC | 10:12 | |
| *** markvoelker has joined #openstack-ansible | 10:27 | |
| *** noonedeadpunk has joined #openstack-ansible | 10:28 | |
| noonedeadpunk | hi everyone! | 10:29 |
| noonedeadpunk | what have I missed for the last week? Seems I've got channel accidentally closed for some reason, so missed all tags:( | 10:30 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible stable/ussuri: Add Centos-8 support https://review.opendev.org/740289 | 10:31 |
| noonedeadpunk | jrosser: anything that needs attention?:) | 10:31 |
| *** shyamb has joined #openstack-ansible | 10:31 | |
| *** markvoelker has quit IRC | 10:31 | |
| noonedeadpunk | I guess focal is broken because of openstacksdk or smth? | 10:32 |
| *** shyam89 has quit IRC | 10:32 | |
| watersj | what do you all suggest for corosync install, I have 3 controller nodes in HA. and will have close 200 compute nodes. Subset of those nodes I'll have my pets on. 16 or less node corosync cluster , or pacemaker_remote on compute nodes and 2 corosync cluster? | 10:45 |
| noonedeadpunk | yeah, that's a problem as corosync can handle up to 64 nodes iirc? | 10:51 |
| noonedeadpunk | so yeah, multiple clusters is the only thing here | 10:51 |
| noonedeadpunk | iirc you can set clustername or id, to distinguish compute nodes | 10:52 |
| *** udesale_ has joined #openstack-ansible | 10:56 | |
| jrosser | noonedeadpunk: o/ hello! | 10:56 |
| jrosser | focal is broken? :) i've not been paying too much attention this week - been a bit manic with $dayjob | 10:56 |
| noonedeadpunk | according to https://review.opendev.org/#/c/740592/ it is:) | 10:57 |
| jrosser | i think getting centos8 merged on ussuri is pretty much super-important | 10:57 |
| noonedeadpunk | but this means I have no worries that I missed something cool :p | 10:57 |
| noonedeadpunk | yeah that's true. as we need to release. and I guess I will take a look on bind-to-mgmt thing at last | 10:58 |
| jrosser | noonedeadpunk: i did some work on Zun too, it would be nice to get that finally working properly | 10:58 |
| *** udesale has quit IRC | 10:58 | |
| noonedeadpunk | oh! | 10:59 |
| noonedeadpunk | it was super close | 10:59 |
| jrosser | i have an AIO with it *almost* working | 10:59 |
| jrosser | no its not :) our role is massively out of date for modern Zun | 10:59 |
| jrosser | there is a bunch of zun-cni stuff thats needed now | 10:59 |
| jrosser | anyway - perhaps thats something we can both look at once releasing ussuri is done | 11:00 |
| noonedeadpunk | at lest I thought it was:( but yeah, I see. I guess there wasn't much use of it | 11:00 |
| noonedeadpunk | yeah, totally | 11:00 |
| *** shyamb has quit IRC | 11:03 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Remove support for Centos-7 https://review.opendev.org/742120 | 11:04 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Remove support for LXC2 configuration keys https://review.opendev.org/742121 | 11:04 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Migrate LXC config keys to LXC3 https://review.opendev.org/742122 | 11:06 |
| *** shyamb has joined #openstack-ansible | 11:13 | |
| kleini | https://docs.openstack.org/project-deploy-guide/openstack-ansible/train/targethosts.html <- The preparation of target hosts mentions, that chrony for a properly synchronized time is necessary. Now I noticed, that OSA deploys chrony, too. Is I automated the target host preparation, this is a little bit confusing. | 11:15 |
| kleini | Are there any reasons, why chrony should be deployed on target hosts before OSA is used? | 11:16 |
| *** markvoelker has joined #openstack-ansible | 11:18 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Remove support for Centos-7 https://review.opendev.org/742124 | 11:19 |
| *** markvoelker has quit IRC | 11:22 | |
| *** mjwales has quit IRC | 11:24 | |
| noonedeadpunk | kleini: I think that hardening role deploys chrony | 11:26 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove Centos-7 support https://review.opendev.org/742100 | 11:27 |
| noonedeadpunk | and there should be a variable to disable this behaviour | 11:27 |
| noonedeadpunk | but no, you can avoid isntalling chrony before osa and rely on role for it's installation | 11:27 |
| *** shyamb has quit IRC | 11:31 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Update galera to 10.5.4 https://review.opendev.org/742105 | 11:32 |
| noonedeadpunk | jrosser: I think before doing that https://review.opendev.org/#/c/742105/2 we should fix upgrade jobs at last... | 11:35 |
| noonedeadpunk | I didn't look at what's wrong with tox, but I tried to basically write down things https://review.opendev.org/#/c/738910/ | 11:37 |
| noonedeadpunk | and I still think that https://review.opendev.org/#/c/710849/ is a good start | 11:37 |
| noonedeadpunk | which can be reworked later and simplified | 11:37 |
| *** sshnaidm|afk is now known as sshnaidm | 11:41 | |
| *** jbadiapa has joined #openstack-ansible | 11:43 | |
| *** shyamb has joined #openstack-ansible | 11:48 | |
| *** markvoelker has joined #openstack-ansible | 11:50 | |
| mgariepy | for the openstack complete that fails on focal. can we just make it so the bash completion is not that important ? in the past i've seen it fails a few times depending on the clients lib installed in the venv. ex. deploying mistral will pull mostly all the cli tool and some were failing to be loaded for the generation of the bash_completion | 11:55 |
| noonedeadpunk | tbh I'd prefer autocompletion to be preset out of the box :p it's so handy... | 12:05 |
| noonedeadpunk | eventually if autocompletion fails there's no warranties other parts of cli tooling works... | 12:06 |
| mgariepy | yes but having it working doesn't validate the other part either. | 12:06 |
| mgariepy | yes completion is handy why do we need to install all the clients in the venv by default anyway ? | 12:07 |
| mgariepy | https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_4a2/740592/1/check/openstack-ansible-deploy-aio_lxc-ubuntu-focal/4a2c895/logs/ara-report/result/c78b1a4a-40ad-4476-840f-39604642a680/ | 12:08 |
| mgariepy | for the list of services that are deployed it's kinda a bit too much for what's installed no ? | 12:08 |
| noonedeadpunk | and that's a great question | 12:09 |
| noonedeadpunk | I think we probably should really install only required clients from what we have in our inventory | 12:09 |
| *** rh-jelabarre has joined #openstack-ansible | 12:09 | |
| mgariepy | like i said when i deployed mistral it was pulling all the clients lib. and that broke the complete command. | 12:10 |
| mgariepy | maybe we can display a warning in case it fails ? | 12:10 |
| mgariepy | but not block on it. | 12:10 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Fix centos-8 functional tests https://review.opendev.org/742166 | 12:11 |
| noonedeadpunk | but um, do we need mistral client on aio without mistral? | 12:12 |
| noonedeadpunk | however yes, we can set failed_when: false here as well | 12:13 |
| noonedeadpunk | (as well as everything else) | 12:13 |
| mgariepy | focal install did have extra pacakge that are upgraded during the install of the libs. | 12:26 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Remove support for LXC2 configuration keys https://review.opendev.org/742121 | 12:29 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Remove Centos-7 support https://review.opendev.org/742103 | 12:36 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Remove Centos-7 support https://review.opendev.org/742104 | 12:38 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Remove support for Centos-7 https://review.opendev.org/742124 | 12:39 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Remove support for Centos-7 https://review.opendev.org/742120 | 12:40 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update log socket workaround to be for all Centos versions https://review.opendev.org/742111 | 12:43 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update log socket workaround to be for all Centos versions https://review.opendev.org/742111 | 12:44 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update ansible metadata to indicate support for centos-8 https://review.opendev.org/742112 | 12:44 |
| *** shyamb has quit IRC | 12:54 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_glance stable/ussuri: Cleanup glance-registry service https://review.opendev.org/742179 | 12:58 |
| *** spatel has joined #openstack-ansible | 12:59 | |
| spatel | jrosser: good morning! | 13:01 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/ussuri: Add Centos-8 support https://review.opendev.org/740289 | 13:01 |
| jrosser | spatel: hello! | 13:01 |
| jrosser | i am having another bash at centos-8 | 13:02 |
| spatel | Yesterday i split distro file as we talked https://review.opendev.org/#/c/739400/ | 13:02 |
| spatel | But i can't see redhat-7.yml file getting checked-in (even that file is in git but i can't see them here https://review.opendev.org/#/c/739400/) | 13:02 |
| jrosser | errrr what is happening there ! | 13:03 |
| spatel | I can see redhat-7.yml file in (git ls-files ) | 13:03 |
| spatel | i am not seeing any error or anything. | 13:03 |
| spatel | can i delete this patch and re-add fresh (i think i poked it a lot) | 13:04 |
| jrosser | i have no idea what is happening there | 13:05 |
| jrosser | ARM64 pipeline - what?! | 13:05 |
| spatel | i have seen that and not sure from where ARM64 coming from | 13:05 |
| spatel | if you ok then i can delete that patch and start fresh one :) | 13:06 |
| jrosser | yes no problem | 13:10 |
| spatel | jrosser: thanks will do that | 13:13 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Use Ussuri RDO repo on ussuri branch https://review.opendev.org/742188 | 13:17 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/ussuri: Add Centos-8 support https://review.opendev.org/740289 | 13:18 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Remove support for LXC2 configuration keys https://review.opendev.org/742121 | 13:21 |
| jrosser | looks like we need https://review.opendev.org/742188 and https://review.opendev.org/742075 and we should be good for centos-8 support to merge for U | 13:24 |
| *** pcaruana has quit IRC | 13:35 | |
| *** d34dh0r53 has quit IRC | 13:55 | |
| *** pcaruana has joined #openstack-ansible | 14:01 | |
| *** d34dh0r53 has joined #openstack-ansible | 14:04 | |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server master: Update log socket workaround to be for all Centos versions https://review.opendev.org/742111 | 14:19 |
| openstackgerrit | Georgina Shippey proposed openstack/openstack-ansible-ops master: Install virtualenv https://review.opendev.org/741997 | 14:24 |
| * jrosser sighs at centos | 14:26 | |
| *** arkan_ has joined #openstack-ansible | 14:29 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Use Ussuri RDO repo on ussuri branch https://review.opendev.org/742188 | 14:34 |
| *** cloudnull has joined #openstack-ansible | 14:44 | |
| cloudnull | o/ - mnaser - would we be able to get a new tag cut for https://github.com/openstack/ansible-config_template maybe a 1.1.1 ? | 14:45 |
| cloudnull | there are some fixes there that we'd like to pull into our next sync but we need a tag to make that happen | 14:46 |
| *** arkan_ has quit IRC | 14:46 | |
| *** arkan has joined #openstack-ansible | 14:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone stable/ussuri: Add Centos-8 support https://review.opendev.org/742075 | 14:48 |
| mnaser | noonedeadpunk is our release liasion, they can help with that | 14:49 |
| noonedeadpunk | cloudnull: tag on top of the master? | 14:50 |
| cloudnull | yesir | 14:50 |
| cloudnull | unless you intend to branch that repo at some point | 14:50 |
| cloudnull | we could use a branch as well. but a simple tag would give us everything we needed | 14:51 |
| noonedeadpunk | tbh I was thinking about branching it, but not so many changes there to start doing that.. | 14:52 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Migrate LXC config keys to LXC3 https://review.opendev.org/742122 | 14:58 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Bump ansible version to 2.9.10 https://review.opendev.org/737935 | 15:04 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Bump ansible version to 2.9.10 https://review.opendev.org/737935 | 15:04 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Bump ansible version to 2.9.11 https://review.opendev.org/737935 | 15:04 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible version to 2.9.11 https://review.opendev.org/737936 | 15:05 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible version to 2.9.11 https://review.opendev.org/737936 | 15:05 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/train: Bump ansible version to 2.8.13 https://review.opendev.org/736465 | 15:05 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/train: Bump ansible version to 2.8.13 https://review.opendev.org/736465 | 15:06 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/train: Bump ansible version to 2.8.13 https://review.opendev.org/736467 | 15:06 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/train: Bump ansible version to 2.8.13 https://review.opendev.org/736467 | 15:06 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Use Ussuri RDO repo on ussuri branch https://review.opendev.org/742188 | 15:08 |
| *** mjwales has joined #openstack-ansible | 15:13 | |
| spatel | jrosser: before i submit new patch i found one error in my lab during rebuilding ceilometer - http://paste.openstack.org/show/796167/ | 15:22 |
| spatel | if i run /openstack/venvs/ceilometer-21.1.0.dev35/bin/ceilometer-upgrade on lxc container then not throwing any error but during build it does | 15:22 |
| spatel | This is the task its trying to execute - https://opendev.org/openstack/openstack-ansible-os_ceilometer/src/branch/master/tasks/ceilometer_db_setup.yml | 15:25 |
| jrosser | if you run /openstack/venvs/ceilometer-21.1.0.dev35/bin/ceilometer-upgrade by hand in the container does it give you any useful output? | 15:25 |
| spatel | no output but echo $? status code is 1 | 15:27 |
| spatel | they means its error out somewhere | 15:27 |
| jrosser | does it have --debug or something (guessing?) | 15:29 |
| spatel | This is what inside that upgrade script - http://paste.openstack.org/show/796168/ | 15:29 |
| spatel | i tried --debug and no luck let me see what its trying to do | 15:29 |
| jrosser | noonedeadpunk: fatal: [infra1]: FAILED! => {"msg": "'hostvars' is undefined"} | 15:29 |
| jrosser | from https://github.com/openstack/openstack-ansible-galera_server/blob/4819f088b0d2309a6258ce16c480302b829f22a6/handlers/main.yml#L61-L73 | 15:29 |
| jrosser | thats, interesting, new failure for ansibles > 2.9.9 | 15:30 |
| noonedeadpunk | hostvars are still present in ansible doc as a valid thing to reffer | 15:32 |
| jrosser | i wonder if its like this https://github.com/ansible/ansible/issues/36575 | 15:33 |
| jrosser | "Using hostvars to smuggle values into play context is not recommended, and we have made changes that actually make using hostvars in this context fail for normal playbook runs as well as syntax checks" | 15:33 |
| *** gyee has joined #openstack-ansible | 15:41 | |
| *** yolanda has quit IRC | 15:42 | |
| *** yolanda has joined #openstack-ansible | 15:44 | |
| *** arkan has quit IRC | 15:46 | |
| *** arkan has joined #openstack-ansible | 15:52 | |
| noonedeadpunk | but this change should be at least one year old... | 15:53 |
| noonedeadpunk | and eventually we use it as regular var... | 15:54 |
| arkan | hi guys | 15:54 |
| arkan | can somebody decipher this http://paste.openstack.org/show/796171/ | 15:55 |
| arkan | it's from inside Fedora Atomic instance | 15:55 |
| arkan | I'm creating kubernetes cluster with magnum | 15:56 |
| arkan | and I'm getting the above messages from inside master node | 15:56 |
| arkan | but I don't know what is the problem | 15:56 |
| arkan | also I'm using this image Fedora-Atomic 27-1.6.x86_64 | 15:57 |
| *** udesale_ has quit IRC | 15:59 | |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 16:11 |
| openstack | Meeting started Tue Jul 21 16:11:13 2020 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:11 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:11 |
| *** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)" | 16:11 | |
| openstack | The meeting name has been set to 'openstack_ansible_meeting' | 16:11 |
| noonedeadpunk | #topic office hours | 16:11 |
| *** openstack changes topic to "office hours (Meeting topic: openstack_ansible_meeting)" | 16:11 | |
| jrosser | o/ hello | 16:11 |
| noonedeadpunk | \o/ | 16:12 |
| noonedeadpunk | Sorry for missing last week meeting | 16:12 |
| noonedeadpunk | not sure if it even happen :( | 16:12 |
| noonedeadpunk | so, from good things, centos8 is pretty close to be packported? | 16:12 |
| noonedeadpunk | *backported | 16:13 |
| noonedeadpunk | from bad ones, focal on master has issues and we're no good with new ansible? | 16:13 |
| jrosser | sort of | 16:13 |
| jrosser | i hope that https://review.opendev.org/#/c/742188/ is the last thing to need before we merge centos-8 into ussuri | 16:13 |
| jrosser | but there are still things like telemetry on master iirc | 16:13 |
| * noonedeadpunk needs to look on them | 16:14 | |
| jrosser | perhaps we have to go back to an etherpad and checkist off all the roles | 16:14 |
| jrosser | *checklist | 16:14 |
| jrosser | and yes new ansible really does not like what we do in the galera_server role | 16:15 |
| arkan | I recopied all the logs from the vm http://paste.openstack.org/show/796174/ | 16:15 |
| noonedeadpunk | I'm pretty sure we use hostvars not only in galera role | 16:15 |
| jrosser | i was wondering if we should just set_fact the thing we need in a previous task | 16:15 |
| jrosser | but that task is odd, delegate_to and also override ansible_host | 16:16 |
| jrosser | i am not sure why that is like that | 16:16 |
| noonedeadpunk | https://github.com/openstack/openstack-ansible-galera_server/commit/3d405dfd52c0a5059cefd877fd578114bcdd912d | 16:17 |
| noonedeadpunk | but I think we probably should look into connection plugin instead? | 16:17 |
| noonedeadpunk | not sure though... | 16:17 |
| noonedeadpunk | eventually why not to delegate to `hostvars[galera_server_bootstrap_node]['ansible_host']` at once.. | 16:18 |
| noonedeadpunk | as they are not in inventory? | 16:18 |
| *** chandankumar is now known as raukadah | 16:19 | |
| *** mjwales has quit IRC | 16:20 | |
| noonedeadpunk | also, question about nova compute nodes has been raised again in bugs, https://bugs.launchpad.net/openstack-ansible/+bug/1887952 | 16:21 |
| openstack | Launchpad bug 1887952 in openstack-ansible "Running os-nova-install with --limit may not distribute SSH keys correctly" [Undecided,Opinion] | 16:21 |
| noonedeadpunk | And I think, maybe it's time we thought about more advanced way of nodes auth? or at least we can probably stage nodes open keys on the deployment host | 16:22 |
| noonedeadpunk | to be able to cleanly distribute them afterwards | 16:22 |
| jrosser | oh yes that | 16:32 |
| jrosser | i would like the get rid of the key distribution entirely | 16:32 |
| jrosser | i think we should look at sshd TrustedUserCAKeys for this | 16:36 |
| noonedeadpunk | yeah, I totally like that idea | 16:48 |
| jrosser | it would completely simplify it | 16:49 |
| jrosser | because currently every compute node needs the keys from all the others, sort of N*N type of problem | 16:49 |
| jrosser | but with TrustedUserCAKeys we would install a (list?) of trusted CA on each host | 16:50 |
| jrosser | then generate a suitable key once per compute node, and it's done | 16:50 |
| jrosser | i think i already have a patch somewhere which was going to setup a CA for haproxy, so thats already existing | 16:51 |
| jrosser | https://review.opendev.org/#/c/644555/ this could all be refreshed for SSH certificates instead | 16:52 |
| noonedeadpunk | In terms of the role it's more complicated, but from the deployer prespective it's nicer | 16:53 |
| * noonedeadpunk tries to recall why needed to use openssl command everywhere instead of module | 16:55 | |
| noonedeadpunk | ah, it was centos7 issue which is not the case. | 16:55 |
| noonedeadpunk | Btw, I saw you started the cleanup? | 16:55 |
| noonedeadpunk | #endmeeting | 17:06 |
| *** openstack changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: https://bit.ly/2SAcGAn" | 17:06 | |
| openstack | Meeting ended Tue Jul 21 17:06:10 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 17:06 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-07-21-16.11.html | 17:06 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-07-21-16.11.txt | 17:06 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-07-21-16.11.log.html | 17:06 |
| *** yolanda has quit IRC | 17:10 | |
| *** yolanda has joined #openstack-ansible | 17:13 | |
| openstackgerrit | Satish Patel proposed openstack/openstack-ansible-os_ceilometer master: Add centos-8 support https://review.opendev.org/742253 | 17:15 |
| spatel | jrosser: after debug i found its failing on keystone, i think my keystone is busted so i will fix it later but for now i am submitting patch | 17:18 |
| jrosser | spatel: did you see we only backported keystone to ussuri today? https://review.opendev.org/#/c/742075/ | 17:20 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_cinder stable/ussuri: Add Centos-8 support https://review.opendev.org/742254 | 17:23 |
| * jrosser head -> desk | 17:23 | |
| spatel | jrosser: i didn't know that, also my lab is running very old master so i think i need to destroy and re-build it | 17:24 |
| jrosser | spatel: i think also i may be confues if you are doing master or ussuri...... | 17:24 |
| -spatel- [root@aio1 playbooks]# git describe | 17:25 | |
| -spatel- 21.0.0.0rc1-19-g78c86afc5 | 17:25 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Replace yum config manager with dnf alternative https://review.opendev.org/742255 | 17:26 |
| jrosser | noonedeadpunk: what about this https://review.opendev.org/#/c/738913/ | 17:27 |
| jrosser | i am sure i have seen a job fail this week becasue the hash was wrong | 17:28 |
| jrosser | but didnt look too hard | 17:28 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_designate stable/ussuri: Add Centos-8 support https://review.opendev.org/742256 | 17:29 |
| noonedeadpunk | centos8 one? | 17:29 |
| jrosser | yes i think it was | 17:30 |
| jrosser | all the complexity the centos-7 code path goes to to extract the hash and build the repo url | 17:31 |
| jrosser | i was not sure why we do not do the same thing for 8 | 17:31 |
| noonedeadpunk | Because we use hashes in repo by default? | 17:31 |
| noonedeadpunk | I mean rhel8 rdo repo has hashed structure anyway | 17:32 |
| jrosser | oh right yes | 17:32 |
| noonedeadpunk | so we really can't add it via dnf_repository module and just curl the repo they have and place it as a file | 17:32 |
| jrosser | i just wget it and i see that now | 17:32 |
| noonedeadpunk | And we did workaround in CI to have the same thing for rhel7 as far as I get the code | 17:33 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/ussuri: Add Centos-8 support https://review.opendev.org/740289 | 17:33 |
| jrosser | noonedeadpunk: we have quite a pile of unmerged el8 stuff on master | 17:36 |
| jrosser | https://review.opendev.org/#/q/topic:osa/el8+(status:open) | 17:36 |
| *** jbadiapa has quit IRC | 17:36 | |
| noonedeadpunk | spatel: https://review.opendev.org/#/c/739660/1 can you confirm, that this package is present for centos8 without epel repo? Otherwise we should also add epel configuration here for rhel8 | 17:39 |
| spatel | jrosser: in that pile some of them are failing for no reason not sure why, example: https://review.opendev.org/#/c/739653/ | 17:39 |
| spatel | noonedeadpunk: let me check | 17:40 |
| spatel | noonedeadpunk: python3-qpid-proton is part of rdo-deps repo | 17:41 |
| noonedeadpunk | oh, nice | 17:41 |
| spatel | and we do have that repo enable everywhere | 17:41 |
| noonedeadpunk | yeah, fair, just didn't have centos8 box under my hand:) | 17:42 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_zun master: Add centos-8 support https://review.opendev.org/739649 | 17:42 |
| *** d34dh0r53 has quit IRC | 17:42 | |
| jrosser | spatel: i think you may have made the gnocchi / ceilometer / aodh patches all in a circular dependancy | 17:44 |
| jrosser | thats why the jobs won't run | 17:44 |
| jrosser | it's a bit of an impossiblity tbh becasue they all need each other in order to pass | 17:44 |
| spatel | i did but then i remove dependency for ceilometer patch and i am going to do same for gnocchi/aodh (i don't think we need that) | 17:45 |
| noonedeadpunk | yeah, that's true( | 17:45 |
| jrosser | noonedeadpunk: do you think we should just set centos-8 to be nv on the telemetry jobs and push in the changes? | 17:45 |
| jrosser | otherwise we will never see if its working or not | 17:45 |
| noonedeadpunk | yep | 17:45 |
| spatel | let me handle them as soon as my ceilometer patch pass | 17:45 |
| *** d34dh0r53 has joined #openstack-ansible | 17:45 | |
| spatel | noonedeadpunk: can you check - https://review.opendev.org/#/c/739646/ look like something going on here.. | 17:47 |
| spatel | No package matching 'python-pip' found available, installed or updated | 17:47 |
| spatel | we need pip3 for centos-8 right? or ussuri | 17:48 |
| noonedeadpunk | ah, damn | 17:49 |
| noonedeadpunk | but it's centos7 job... | 17:49 |
| noonedeadpunk | spatel: I wrote this workaround for centos7 https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/common-tasks/ceph-server.yml#L47 | 17:50 |
| spatel | yes but ussuri only support python3 right? | 17:50 |
| noonedeadpunk | but they broke it by removing python-pip package... | 17:50 |
| noonedeadpunk | but ansible still discovers ansible_python_interpreter to be py2 by default | 17:51 |
| jrosser | they? centos? | 17:51 |
| spatel | just curious why this specific role need python-pip but not others? | 17:51 |
| noonedeadpunk | yeah lol)) | 17:51 |
| noonedeadpunk | spatel: like it needs ceph | 17:51 |
| spatel | oh got it, its filesystem related role | 17:52 |
| noonedeadpunk | which is either a part of experimental, or manilla | 17:52 |
| noonedeadpunk | ok, let's try to somehow fix that... but for master I'd rather just replaced lol | 17:53 |
| spatel | what if we do something like - name: "{{ (ansible_os_family | lower == 'redhat') | ternary('python-pip', 'python3-pip') }}" | 17:57 |
| spatel | i meant centos-7 vs 8 condition | 17:57 |
| noonedeadpunk | centos7 dropped python-pip package at all | 17:58 |
| spatel | damn it.. so many variable :( | 17:58 |
| noonedeadpunk | but the thing is about how to make ansible use py3 by default for centos7 | 17:58 |
| noonedeadpunk | except setting it in group_vars for all | 17:59 |
| noonedeadpunk | but I'm afraid other things may got broken then... | 18:00 |
| spatel | in my centos-7 desktop when i do yum install python-pip it does install it using eple repo | 18:02 |
| spatel | may be i have over epel repo | 18:02 |
| spatel | older* | 18:02 |
| noonedeadpunk | maybe they left it in epel.... | 18:03 |
| noonedeadpunk | but that means we need epel which also sad and needs patching | 18:04 |
| spatel | noonedeadpunk: i think replacing python2-pip should fix issue for centOS7 | 18:06 |
| spatel | i can see in centos-7 we have /etc/yum.repos.d/epel-lxc_hosts.repo repo | 18:07 |
| noonedeadpunk | the main question for me is how to set ansible_python_interpreter for centos7 only( | 18:07 |
| noonedeadpunk | yeah, but that repo has list of packjages that can be installed from it | 18:07 |
| jrosser | using py3 by default for centos7 will proabably break a ton of other things | 18:09 |
| jrosser | like lxc | 18:09 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: [WIP] Install py3 for centos7 by default https://review.opendev.org/742262 | 18:11 |
| noonedeadpunk | ah, yeah, I guess you're right... | 18:12 |
| noonedeadpunk | so, install epel in addition to that then.... | 18:12 |
| spatel | or do pip install pip :) | 18:13 |
| noonedeadpunk | um... does it work that way? | 18:13 |
| noonedeadpunk | I mean to instally pip2 with pip3? | 18:13 |
| spatel | i never test but i can test it.. let me see.. | 18:14 |
| noonedeadpunk | I guess we can install with curl and bash | 18:14 |
| spatel | This patch should work and safe right - https://review.opendev.org/742262 | 18:15 |
| spatel | not good for if someone trying to install older release | 18:15 |
| noonedeadpunk | no, it's not good( | 18:16 |
| noonedeadpunk | at least that part https://review.opendev.org/#/c/742262/1/inventory/group_vars/all/ceph.yml | 18:16 |
| spatel | pip install pip won't work | 18:18 |
| spatel | i think final solution would be to install epel-release in ceph container | 18:18 |
| spatel | epel-release repo will fix all issue | 18:18 |
| noonedeadpunk | yeah... | 18:19 |
| spatel | or something like -> curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py | 18:19 |
| spatel | i think epel-release would be good (it will provide all decency also ) | 18:19 |
| spatel | yum install epel-release | 18:24 |
| spatel | yum install python-pip | 18:24 |
| spatel | it works on my desktop | 18:24 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: [WIP] Fix ceph deployment for CentOS 7 https://review.opendev.org/742262 | 18:28 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Fix ceph deployment for CentOS 7 https://review.opendev.org/742262 | 18:28 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Fix ceph deployment for CentOS 7 https://review.opendev.org/742262 | 18:29 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Add centos-8 support https://review.opendev.org/739646 | 18:30 |
| noonedeadpunk | let's see.... | 18:30 |
| noonedeadpunk | cloudnull: oh, btw, ansible-config_template tagging patch is here https://review.opendev.org/#/c/742238/ | 18:33 |
| cloudnull | Very cool! | 18:34 |
| cloudnull | tyvm | 18:34 |
| spatel | cloudnull: :) | 18:40 |
| cloudnull | o/ | 18:40 |
| spatel | good to see you after looooong time | 18:41 |
| cloudnull | it’s been a while. | 18:41 |
| spatel | oh yeah! i am sure you must be master of tripleO :) | 18:42 |
| cloudnull | Still very much learning | 18:43 |
| spatel | After you gone from here, i have built 3 more openstack cloud and now building 4th one soon (+ going to use some GPU compute nodes soon) | 18:45 |
| *** arkan has quit IRC | 18:46 | |
| spatel | google driving us crazy ($$$) so thinking to build our own GPU cloud. | 18:46 |
| spatel | noonedeadpunk: do you think we should make your https://review.opendev.org/#/c/742262/ patch as dependency for manila patch for centOS-8? | 18:47 |
| noonedeadpunk | spatel: I think I already did that?:) | 18:48 |
| spatel | sweet! | 18:48 |
| spatel | how long it take to kick CI job in, my ceilometer patch waiting for last 2 hours | 18:49 |
| noonedeadpunk | uh, that depends on CI load, what we can't influence... | 19:01 |
| spatel | noonedeadpunk: how hard its to install zuul CICD in my lab environment? | 19:05 |
| noonedeadpunk | um... in docker it's easy) manually not so | 19:07 |
| noonedeadpunk | It has lot's of nuances tbh | 19:07 |
| noonedeadpunk | I found it very interesting but also a bit challenging) | 19:08 |
| noonedeadpunk | I guess took about 30 hours of my spare time :p but I also deployed it with gerrit | 19:09 |
| noonedeadpunk | but if you need CI/CD tooling and you're about to choose between jenkins or zuul, I'd totally say go on with zuul | 19:10 |
| jrosser | I was always wanting to try https://softwarefactory-project.io/docs/index.html but never found time | 19:10 |
| noonedeadpunk | jrosser: oh, I guess rdo use it? | 19:10 |
| noonedeadpunk | yeah, that's nice | 19:10 |
| spatel | noonedeadpunk: i will give it a try and see which one looks good | 19:35 |
| spatel | One more question for both of you, how do you write quick code and check its going to work or not :) - Example: https://review.opendev.org/#/c/742262/4/playbooks/common-tasks/ceph-server.yml | 19:35 |
| spatel | Do you have something to test on local machine or its all experience :) | 19:36 |
| spatel | noonedeadpunk: jrosser ^^ | 19:36 |
| *** arkan has joined #openstack-ansible | 19:46 | |
| *** arkan has quit IRC | 19:46 | |
| *** arkan has joined #openstack-ansible | 19:46 | |
| noonedeadpunk | I just partialy took code from other place which is known to work :p | 19:53 |
| spatel | ah! easy | 20:27 |
| spatel | i was thinking how to write and test code on laptop where you have no full environment | 20:27 |
| *** arkan has quit IRC | 20:28 | |
| *** poopcat has quit IRC | 20:34 | |
| *** mjwales has joined #openstack-ansible | 20:34 | |
| *** mjwales has quit IRC | 20:49 | |
| *** jbadiapa has joined #openstack-ansible | 21:17 | |
| *** spatel has quit IRC | 22:09 | |
| dmsimard | spatel: +1 for software factory :) | 22:46 |
| dmsimard | oh no, he's gone :( | 22:46 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/ussuri: Add Centos-8 support https://review.opendev.org/742254 | 23:27 |
| *** tosky has quit IRC | 23:47 | |
| dmsimard | jrosser, noonedeadpunk: web frontend is hard but working on a CLI is easier http://paste.openstack.org/show/796193/ | 23:52 |
| dmsimard | still rough around the edges but it works :p | 23:52 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!