*** rh-jelabarre has quit IRC | 01:09 | |
*** also_stingrayza has joined #openstack-ansible | 01:49 | |
*** stingrayza has quit IRC | 01:51 | |
*** redrobot has quit IRC | 02:40 | |
*** Guest52018 has joined #openstack-ansible | 03:19 | |
*** Guest52018 is now known as redrobot | 03:20 | |
*** gyee has quit IRC | 03:27 | |
*** cheng1 has quit IRC | 03:28 | |
*** cheng1 has joined #openstack-ansible | 03:33 | |
*** evrardjp has joined #openstack-ansible | 04:33 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/train: Use Train RDO packages on stable/train branch https://review.opendev.org/734418 | 05:41 |
---|---|---|
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Always use the utility host for service setup tasks https://review.opendev.org/733854 | 05:46 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/ussuri: Use ansible openstack collection https://review.opendev.org/734419 | 05:47 |
*** jawad_axd has joined #openstack-ansible | 05:48 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/ussuri: Synchronise move to openstack ansible collection to all roles https://review.opendev.org/734272 | 06:04 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_designate stable/ussuri: Use ansible openstack collection https://review.opendev.org/734293 | 06:07 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_octavia stable/ussuri: Use ansible openstack collection https://review.opendev.org/734278 | 06:07 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_gnocchi stable/ussuri: Use ansible openstack collection https://review.opendev.org/734277 | 06:07 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_trove stable/ussuri: Use ansible openstack collection https://review.opendev.org/734291 | 06:08 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_magnum stable/ussuri: Use ansible openstack collection https://review.opendev.org/734303 | 06:08 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone stable/ussuri: Use ansible openstack collection https://review.opendev.org/734274 | 06:08 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_swift stable/ussuri: Use ansible openstack collection https://review.opendev.org/734276 | 06:08 |
*** cheng1 has quit IRC | 06:10 | |
*** cheng1 has joined #openstack-ansible | 06:13 | |
jrosser | noonedeadpunk: my rebuilt galera 10.4.12 for focal does not work | 06:15 |
jrosser | well, mariadb to be specific | 06:16 |
*** pcaruana has joined #openstack-ansible | 06:25 | |
noonedeadpunk | jrosser: maybe let's copy their 10.4.14 focal build? | 06:26 |
noonedeadpunk | It will be daily and not stable, but we will be able to move forward | 06:26 |
jrosser | yeah, i had quite a look at the broken one and the wsrep is all not working | 06:27 |
jrosser | and it's very unclear how to debug it | 06:27 |
jrosser | do you know where their daily build is? | 06:27 |
*** udesale has joined #openstack-ansible | 06:28 | |
jrosser | else it's just tip of https://github.com/MariaDB/server/tree/10.4 and I can compile that? | 06:28 |
jrosser | they have a quite messy repo so it's hard to know what is what | 06:28 |
jrosser | also idk if we need a rebuilt version of https://github.com/MariaDB/galera or not | 06:29 |
*** dpaclt has joined #openstack-ansible | 06:39 | |
*** jbadiapa has joined #openstack-ansible | 06:55 | |
noonedeadpunk | Let me try to browse history... | 07:11 |
noonedeadpunk | I pasted it another day | 07:11 |
* noonedeadpunk can't find the link:( | 07:20 | |
jrosser | i am having one of those days yoo | 07:28 |
jrosser | too | 07:28 |
jrosser | just trying a focal AIO | 07:28 |
jrosser | and i have no idea why but this https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/main.yml#L62-L72 | 07:28 |
jrosser | includes debian.yml | 07:28 |
*** tosky has joined #openstack-ansible | 07:32 | |
*** fghaas has joined #openstack-ansible | 07:34 | |
jrosser | noonedeadpunk: oh hahahahah https://review.opendev.org/#/c/729274/ | 07:35 |
jrosser | i've been trying "master" | 07:36 |
jrosser | which of course isnt | 07:36 |
jrosser | thats not good, we've been merging stuff against that | 07:36 |
openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Updated from OpenStack Ansible Tests https://review.opendev.org/734190 | 07:38 |
openstackgerrit | Merged openstack/openstack-ansible-os_masakari master: Updated from OpenStack Ansible Tests https://review.opendev.org/734200 | 07:38 |
openstackgerrit | Merged openstack/openstack-ansible-os_placement master: Updated from OpenStack Ansible Tests https://review.opendev.org/734207 | 07:39 |
openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Updated from OpenStack Ansible Tests https://review.opendev.org/734209 | 07:41 |
noonedeadpunk | yeah... we should have merged that before anything else:( | 07:43 |
openstackgerrit | Merged openstack/openstack-ansible-os_blazar master: Updated from OpenStack Ansible Tests https://review.opendev.org/734188 | 07:45 |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Updated from OpenStack Ansible Tests https://review.opendev.org/734203 | 07:46 |
openstackgerrit | Merged openstack/openstack-ansible-os_sahara master: Updated from OpenStack Ansible Tests https://review.opendev.org/734208 | 07:46 |
openstackgerrit | Merged openstack/openstack-ansible-os_panko master: Updated from OpenStack Ansible Tests https://review.opendev.org/734206 | 07:47 |
openstackgerrit | Merged openstack/openstack-ansible-os_heat master: Updated from OpenStack Ansible Tests https://review.opendev.org/734196 | 07:48 |
openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Updated from OpenStack Ansible Tests https://review.opendev.org/734204 | 07:52 |
openstackgerrit | Merged openstack/openstack-ansible-os_trove master: Updated from OpenStack Ansible Tests https://review.opendev.org/734212 | 07:52 |
openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: Updated from OpenStack Ansible Tests https://review.opendev.org/734198 | 07:53 |
openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Updated from OpenStack Ansible Tests https://review.opendev.org/734197 | 07:53 |
openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Updated from OpenStack Ansible Tests https://review.opendev.org/734205 | 07:54 |
openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Updated from OpenStack Ansible Tests https://review.opendev.org/734187 | 07:54 |
*** rpittau|afk is now known as rpittau | 07:54 | |
openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Updated from OpenStack Ansible Tests https://review.opendev.org/734189 | 07:59 |
openstackgerrit | Merged openstack/openstack-ansible-os_mistral master: Updated from OpenStack Ansible Tests https://review.opendev.org/734201 | 07:59 |
*** wpp has quit IRC | 08:01 | |
*** wpp has joined #openstack-ansible | 08:01 | |
openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: Updated from OpenStack Ansible Tests https://review.opendev.org/734199 | 08:03 |
*** maharg101 has quit IRC | 08:10 | |
*** kaiokmo has joined #openstack-ansible | 08:21 | |
*** sshnaidm|afk is now known as sshnaidm | 08:22 | |
*** cshen has joined #openstack-ansible | 08:33 | |
noonedeadpunk | jrosser: found the link :p http://hasky.askmonty.org/archive/10.4/build-33293/kvm-deb-focal-amd64/debs/binary/ | 08:41 |
cheng1 | jrosser, I finally get there by deploying 2 nodes. Most of the services are deployed on node1, haproxy is deployed on node2 | 08:42 |
openstackgerrit | Hervé Beraud proposed openstack/openstack-ansible master: Use unittest.mock instead of mock https://review.opendev.org/734452 | 08:51 |
noonedeadpunk | I'll copy packages to the repo | 08:51 |
noonedeadpunk | jrosser: mariadb in repo updated with packages for 10.4.14 | 08:58 |
*** cheng1 has quit IRC | 08:58 | |
*** cheng1 has joined #openstack-ansible | 08:59 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: WIP - Add Ubuntu Focal support https://review.opendev.org/723597 | 09:09 |
jrosser | noonedeadpunk: ^ i've added the victoria cycle patch as depends on to that so lets see what happens | 09:09 |
jrosser | also trying locally too | 09:09 |
ioni | i saw OSA being branched and then noticed at boostrap is still tracking master :( | 09:09 |
jrosser | branching != releasing :) | 09:10 |
ioni | yes | 09:10 |
jrosser | this needs to merge https://review.opendev.org/#/c/729274/ | 09:10 |
jrosser | than master is master again | 09:10 |
openstackgerrit | amolkahat proposed openstack/openstack-ansible-os_tempest master: Added tempest ironic resources setup. https://review.opendev.org/720705 | 09:20 |
noonedeadpunk | jrosser: so, should I return your 10.4.12? Like I was using reprepro which do not store several versions of the same package... | 09:34 |
jrosser | hmm? | 09:34 |
jrosser | my 10.4.12 looks broken i think | 09:34 |
jrosser | unless it never was there? | 09:35 |
noonedeadpunk | it was... | 09:35 |
jrosser | ok so i think that this should be testing your 10.4.14 packages now https://review.opendev.org/#/c/723597/ | 09:35 |
*** jcath has joined #openstack-ansible | 09:39 | |
noonedeadpunk | Oh, I thought we do package pinning for galera but seems we doesn't | 09:41 |
jcath | hello,friends. does osad has to install with lxc container? Can I just install openstack services with openstack-ansible to a physical host environment? centos7, for example | 09:42 |
jrosser | jcath: OSA default install is with LXC, and whilst you can deploy without (bare metal deployment) that does make some things more complicated | 09:43 |
jrosser | jcath: and centos7 is an interesting choice for deploying anything later than Train | 09:43 |
jrosser | where interesting == more problems | 09:43 |
jcath | thanks jrosser, I want to deploy stein with centos 7.7, in an offline LAN environment | 09:45 |
jrosser | noonedeadpunk: did you see that horizon is broken on ussuri / centos7, there is no apache mod_wsgi for python3 | 09:45 |
noonedeadpunk | uh, no:( | 09:45 |
jcath | I deploy a cluster with packstack weeks before, but packstack is not well enough for a prod deployment, i think | 09:46 |
noonedeadpunk | jrosser: we can probably use uwsgi + nginx instread of apache... | 09:46 |
noonedeadpunk | but this is annoying ofc... | 09:47 |
jrosser | noonedeadpunk: yes i was looking at os_keystone becasue that looks similar | 09:47 |
noonedeadpunk | uh, keystone if broken as well.... | 09:47 |
jcath | and kolla need lot's of downloading traffic and space for docker images, so I turn to OSA .... | 09:47 |
noonedeadpunk | *is | 09:48 |
jrosser | noonedeadpunk: i was wondering if mnaser had found this yet on ussuri (maybe thats not centos though?) | 09:48 |
*** cheng1 has quit IRC | 09:48 | |
noonedeadpunk | We were moving to debian... to no new machines deployed on centos | 09:48 |
jrosser | jcath: OSA still needs lots of stuff downloading, git repos, distro packages, pypi stuff | 09:49 |
noonedeadpunk | but we probably had _reason_ why we did apache for some deployments?Like things do not work with uwsgi? | 09:49 |
jrosser | this is horizon? | 09:49 |
noonedeadpunk | for keystone... | 09:49 |
jrosser | oh federation | 09:49 |
jrosser | totally needs apache | 09:50 |
noonedeadpunk | so centos7 is kinda totally broken for U.... | 09:50 |
jrosser | mod_oidc blah blah | 09:50 |
jrosser | from horizon perspective yes, so i was wondering if we should lift/shift the nginx setup from os_keystone into os_horizon | 09:51 |
jcath | jrosser: yes, I notice that ... distro packages is easy to setup a local repo, git&pypi... :( | 09:51 |
jrosser | jcath: do you have no external connectivity at all? | 09:51 |
noonedeadpunk | jrosser: considering, that keystone federation won't work... | 09:51 |
noonedeadpunk | Which is pretty popular deployments imo | 09:51 |
*** cheng1 has joined #openstack-ansible | 09:52 | |
jcath | yes, at the deployment environment, just LAN access | 09:52 |
jcath | so last time, I pack all the packages to a local repo, install Openstack with packstack | 09:53 |
noonedeadpunk | jcath: so even no proxy or smth? | 09:53 |
jcath | no. it is a totally internal network | 09:54 |
jrosser | noonedeadpunk: i wonder if keystone is different for wsgi https://github.com/openstack/openstack-ansible-os_keystone/blob/master/vars/debian.yml#L49 | 09:54 |
jcath | but packstack come with lot's of limitation. you can only deploy a simple controller/computes cluster with it | 09:54 |
jrosser | jcath: you would have to set up internal distro mirror and copies of the git repos | 09:55 |
jrosser | but the most complicated thing by far would be what to do about no access to pypi | 09:55 |
noonedeadpunk | eventually you can probably clone like pypi and things specified in https://opendev.org/openstack/openstack-ansible/src/branch/master/ansible-role-requirements.yml and https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/defaults/repo_packages | 09:55 |
jrosser | python dependancies are recurisve so it's really difficuly | 09:56 |
jrosser | a pypi mirror is *gigantic* | 09:57 |
jcath | that is a problem :( from this point, packstack and kolla do things better | 09:57 |
noonedeadpunk | I guess https://github.com/pypa/bandersnatch may create pypi clone, but it's also tons of diskspace... | 09:57 |
noonedeadpunk | Like https://pypi.org/project/python-pypi-mirror/ can create clone of required things, but you really may spend tons of time for cloning everything... | 09:58 |
jrosser | we used bandersnatch in the past and i think it's needing 2Tbyte or thereabouts these days | 09:58 |
jcath | I dont like lot's of docker images running in a host, just for the openstack application. take lot's of spaces and lot's of container-network/iptables stuff | 09:59 |
jcath | jrosser: 2TB? OMG | 10:00 |
noonedeadpunk | jrosser: yeah, you;re right, keystone is different. and it's really a solution we might use. | 10:00 |
*** cshen has quit IRC | 10:01 | |
noonedeadpunk | Just leave apache and use uwsgi | 10:01 |
jrosser | jcath: i think there are machine learning projects that now publish their datasets to pypi which has pretty much runined anyone who wants to use a mirror | 10:02 |
jcath | another way I think about is, try to install all the packages by rpm, create the database, override the config files of services, do some bootstrap work for specific service like keystone, then try to system up everything | 10:04 |
jcath | OSA come with many sub-projects for different services, is there a way I may easy to collect all the well-design service config files from OSA? | 10:06 |
*** rpittau is now known as rpittau|bbl | 10:11 | |
noonedeadpunk | jcath: eventually, you can probably try to use distro install path.... but there might be caveats as well... | 10:13 |
jcath | with is_metal set true to skip lxc, right? in OSA | 10:13 |
noonedeadpunk | install_method: distro https://opendev.org/openstack/openstack-ansible/src/branch/master/etc/openstack_deploy/user_variables.yml#L36-L40 | 10:15 |
jcath | thanks! I read this before, change from source to distro to use distribution's packges. | 10:17 |
noonedeadpunk | So this will minimize usage of python packages for sure, so probably you won't need pypi... | 10:17 |
noonedeadpunk | However not all projects deployable with osa support distro method | 10:18 |
jcath | I only deploy basic things, nova, keystone, glance, cinder, horizon, ceph, neutron | 10:18 |
*** soren has quit IRC | 10:25 | |
*** soren has joined #openstack-ansible | 10:26 | |
*** cshen has joined #openstack-ansible | 10:50 | |
jamesdenton | mornin | 10:52 |
noonedeadpunk | o/ | 10:54 |
jamesdenton | yo | 10:54 |
mgariepy | jrosser, noonedeadpunk anyone had talked to the mariadb guys ? | 10:54 |
mgariepy | they could use our galera repo to tests the deployment and that the clustering is actually working .. | 10:55 |
noonedeadpunk | I was talking to montywi | 10:55 |
jamesdenton | anyone here doing much Ironic? | 10:55 |
noonedeadpunk | I think they have pretty wide testing at the moment... | 10:56 |
*** udesale_ has joined #openstack-ansible | 10:57 | |
noonedeadpunk | an they cover with tests every found bug not to allow regression in the futrure | 10:57 |
*** udesale has quit IRC | 10:59 | |
mgariepy | http://paste.openstack.org/show/794512/ jrosser that's the error with mariabackup (rsync goes further) | 11:00 |
mgariepy | but fail at creating the user. | 11:01 |
*** shyamb has joined #openstack-ansible | 11:18 | |
*** mbuil_ has quit IRC | 11:21 | |
*** mbuil has joined #openstack-ansible | 11:23 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Remove git repo haproxy backend https://review.opendev.org/734587 | 11:24 |
*** threestrands has quit IRC | 11:27 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: WIP - Add Ubuntu Focal support https://review.opendev.org/723597 | 11:27 |
jrosser | i think we should consider merging this sooner rather than later https://review.opendev.org/#/c/731457/ | 11:35 |
jrosser | we are still installing py2 into all the lxc containers everywhere and thats masking other problems | 11:35 |
jrosser | there is the potential to break stuff though | 11:36 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Test patch for 731457 https://review.opendev.org/734590 | 11:38 |
jrosser | mgariepy: i am at the same point, focal AIO with galera failing at create users | 11:42 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Use temporary repository for MariaDB 10.4.12 for Ubuntu focal https://review.opendev.org/733946 | 11:44 |
jrosser | ahha | 11:45 |
openstackgerrit | Merged openstack/openstack-ansible master: Start Victoria cycle https://review.opendev.org/729274 | 11:51 |
*** amalrajgenocidex has joined #openstack-ansible | 11:51 | |
jrosser | noonedeadpunk: my focal AIO has managed to do keystone db setup with the 10.4.14 packages | 11:52 |
amalrajgenocidex | I was trying to upgrade openstack magnum to master and got below error. | 12:02 |
amalrajgenocidex | I did this because magnum in v20.1.2 has a bug that prevents clusters from creating. | 12:02 |
amalrajgenocidex | http://paste.openstack.org/show/794514/ | 12:02 |
amalrajgenocidex | Can anyone take a look. | 12:02 |
*** rh-jelabarre has joined #openstack-ansible | 12:03 | |
amalrajgenocidex | I'm receiving similar errors for all the services when I try to update. | 12:03 |
amalrajgenocidex | is there a stable version of openstack-ansible in which magnum works properly.? | 12:04 |
amalrajgenocidex | I tried stable/train last day which was going to v20.1.2dev and also had similar issues | 12:06 |
noonedeadpunk | eventually you can override magnum version for any osa release | 12:06 |
amalrajgenocidex | noonedeadpunk Sorry, I didn't understand. Can you explain a bit more | 12:09 |
noonedeadpunk | So we bump services versions this way https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/defaults/repo_packages/openstack_services.yml#L224 | 12:10 |
noonedeadpunk | So you can specify magnum version you want to be deployed with setting magnum_git_install_branch into your user_variables | 12:10 |
noonedeadpunk | You can use either tag or sha (or maybe even branch) | 12:11 |
*** rpittau|bbl is now known as rpittau | 12:11 | |
*** iurygregory has joined #openstack-ansible | 12:14 | |
amalrajgenocidex | noonedeadpunk Thank you, I understand. I will give it a try. | 12:20 |
amalrajgenocidex | Another issue I had is , in deploying latest stable/train. it was not installing the repo packages on all my controllers. It was only getting installed in 1 of 3 controllers. | 12:22 |
noonedeadpunk | amalrajgenocidex: ah, yes, I guess there was a bug on stable branch these days. | 12:22 |
noonedeadpunk | We generally recommend using tagged version instead of stable/* branches | 12:22 |
amalrajgenocidex | Thank you. So which branch do you recommend ? | 12:23 |
noonedeadpunk | here's the fix https://review.opendev.org/#/c/731779/ | 12:23 |
noonedeadpunk | "current" stable/train is already fixed | 12:24 |
amalrajgenocidex | Thank you very much noonedeadpunk Let me give a try again :) | 12:32 |
janno | regarding the check if all cinder services are up: I just tested the task and everything works as intended, but I am still wondering how to handle disabled services, which are down | 12:35 |
*** amalrajgenocidex has quit IRC | 12:37 | |
*** amalrajgenocidex has joined #openstack-ansible | 12:40 | |
amalrajgenocidex | noonedeadpunk It looks like the double quotes is missing in master branch too. | 12:40 |
noonedeadpunk | amalrajgenocidex: I think this should be also fixed at the moment... | 12:41 |
noonedeadpunk | af the same patch has been merged for master as well | 12:41 |
noonedeadpunk | the problem in checking out to branch is that branch got updated from time to time | 12:42 |
amalrajgenocidex | Ok noonedeadpunk . looks like master is also patched. My /etc/ansible directory had files from earlier even though I ran bootstrap for master now. So I thought it wasn't patched. | 12:48 |
amalrajgenocidex | Thanks for your help | 12:48 |
openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-systemd_service master: Remove fail: tasks https://review.opendev.org/734597 | 12:50 |
openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-systemd_service master: Revert "Build out the PrivateNetwork function for services" https://review.opendev.org/734598 | 12:50 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: DNM test task removal from systemd_service https://review.opendev.org/734601 | 12:52 |
mgariepy | noonedeadpunk, jrosser can we set the no_log to false except when it's zuul ? | 12:56 |
mgariepy | for tasks like.. mysql_user ? | 12:57 |
jrosser | that would be ++ helpful | 12:57 |
noonedeadpunk | I'm not sure it's good idea tbh... | 12:57 |
jrosser | mgariepy: btw look at this https://zuul.opendev.org/t/openstack/stream/99b79f1fbe4047cda200bf50aa473d7f?logfile=console.log | 12:57 |
mgariepy | the issue with galera and focal is the socker var not defined | 12:57 |
jrosser | i have fixedit | 12:57 |
noonedeadpunk | Like it will really disclose creds into log and stdout... | 12:57 |
mgariepy | lol jrosser nice :) | 12:58 |
jrosser | in here https://review.opendev.org/#/c/733946/7/vars/ubuntu-20.04.yml | 12:58 |
noonedeadpunk | we can disclose it only for zuul for instance | 12:58 |
mgariepy | i didn't refresh my patch this morning haha | 12:58 |
jrosser | i am up to neutron locally here, it's still going | 12:59 |
noonedeadpunk | and just drop our ansible-lint custom rule which checks that. | 12:59 |
jrosser | right now we have a really big mess of patches around focal | 12:59 |
jrosser | and i am thinking that the fixes for focal are breaking some other things | 12:59 |
mgariepy | the mysql_user error would have been very trivial with the no_log false. | 12:59 |
mgariepy | hmm well i mis spoke, i want it to log in zuul but not when it's not zull | 13:00 |
mgariepy | zuul** | 13:00 |
noonedeadpunk | jrosser: btw https://stackoverflow.com/questions/47650902/httpd-mod-wsgi-for-python3-4-not-working-in-centos7 seems a working and pretty trivial thing to fix for centos7 py3 and horizon. | 13:00 |
noonedeadpunk | But maybe using uwsgi is better... | 13:00 |
jrosser | i just want centos-7 to go away :) | 13:01 |
noonedeadpunk | with centos 8 lol | 13:02 |
jrosser | but yeah, i dunno, i'm not really knowing what the best answer is, copy keystone nginx or hack in some 3rd party repo | 13:02 |
noonedeadpunk | But eventually there's no need in 3rd party repo | 13:03 |
noonedeadpunk | just pip and httpd-devel | 13:03 |
*** shyamb has quit IRC | 13:03 | |
jrosser | doesnt it have to compile the module | 13:03 |
noonedeadpunk | yeah, it compiles it indeed | 13:03 |
mgariepy | jrosser, so focal is all fixed now ? | 13:03 |
mgariepy | centos8 is really painful. | 13:04 |
noonedeadpunk | but not sure if compiling is a big problem... yeah, it's pretty hacky.... | 13:04 |
*** shyamb has joined #openstack-ansible | 13:04 | |
jrosser | i would guess my AIO locally here is going to pass tempest | 13:04 |
jrosser | and i think that chances are https://review.opendev.org/#/c/723597/ is going to pass for focal too | 13:04 |
jrosser | downside is i see buster and centos broken on 723597 | 13:06 |
jrosser | so we've got to start working through the patches needed to make focal work and merge/fix those in a way that keeps everything else OK | 13:06 |
*** shyamb has quit IRC | 13:07 | |
jrosser | and i see from the logs on https://review.opendev.org/#/c/733946/ that the temporary packages for 10.4.14 have worked in the focal functional tests | 13:07 |
jrosser | so thats showing a cluster can form | 13:07 |
noonedeadpunk | I'd wish maria got released sooner :( | 13:08 |
jrosser | noonedeadpunk: i don't know what we'd do if there was one nginx on the host having to do horizon and keystone | 13:09 |
jrosser | or if thats just ok | 13:09 |
noonedeadpunk | jrosser: but we can do apache and mod_uwsgi | 13:09 |
noonedeadpunk | isntead of nginx | 13:09 |
noonedeadpunk | the same as keystone do... | 13:09 |
noonedeadpunk | but horizon might be different... | 13:09 |
noonedeadpunk | I'm playing with it now.. | 13:10 |
jrosser | ok, i need to not try to fix all the things in my head at once :) | 13:10 |
noonedeadpunk | yeah | 13:10 |
jamesdenton | QQ: will config_template module create the dest directory? | 13:17 |
jrosser | i'm not sure - its derived from the copy module i think.... | 13:19 |
jrosser | mgariepy: it's passed tempest locally here on LXC / AIO / focal | 13:20 |
mgariepy | great success ! :D | 13:20 |
openstackgerrit | Merged openstack/openstack-ansible master: Only use zuul cloned git repos for code hosted at opendev.org https://review.opendev.org/731451 | 13:21 |
jamesdenton | copy says "If dest is a non-existent path and if either dest ends with "/" or src is a directory, dest is created.", so maybe we're good | 13:24 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: [WIP] Compile mod-wsgi module for CentOS 7 https://review.opendev.org/734608 | 13:24 |
jrosser | jamesdenton: you seen anything like this? http://paste.openstack.org/show/794519/ | 13:25 |
jamesdenton | never | 13:26 |
jrosser | mgariepy: so i think it passes but the debugging sort of starts now, i see tracebacks from neutron | 13:30 |
*** spatel has joined #openstack-ansible | 13:30 | |
*** cshen has quit IRC | 13:34 | |
jamesdenton | this is your focal AIO? | 13:39 |
*** cshen has joined #openstack-ansible | 13:41 | |
jrosser | yes it is | 13:51 |
mgariepy | jrosser, debian on hub.com/openstack/openstack-ansible is due to bad interpreter for the lxc playbook it uses /usr/bin/python which point to python2 and doesn't have the lxc module. | 13:53 |
mgariepy | on https://review.opendev.org/#/c/723597 | 13:53 |
mgariepy | damn. my pasting skills are bad this morning. | 13:54 |
jrosser | yeah, so interesting question there is if it is actually a debian problem | 13:54 |
jrosser | or if it is the debian CI image | 13:54 |
jrosser | if stock buster does not come with py2 then we would not see this outside openstack CI | 13:55 |
jrosser | mgariepy: i have this as a potential mitigation for that https://review.opendev.org/#/c/731476/ | 13:56 |
mgariepy | failing on metal test. "module_stderr": "/bin/sh: 1: /usr/bin/python: not found | 13:58 |
mgariepy | for bionic | 13:58 |
jrosser | yep | 13:58 |
jrosser | it's all related to ansible interpreter discovery | 13:59 |
jrosser | i've got a patch to openstack-ansible-tests for similar trouble https://review.opendev.org/#/c/733948/ | 14:00 |
mgariepy | sotck image have pyton2.7 and python3.7 installed | 14:00 |
mgariepy | for buster. | 14:00 |
jrosser | ansible will prefer /usr/bin/python always if it finds it, and you don't tell it otherwise | 14:01 |
jrosser | and we can't just universally set it to /usr/bin/python3 becasue that breaks centos-7 iirc | 14:03 |
jrosser | and as i found in the tests repo you can't use ansible facts to determine the value, becasue that needs, ummm, python :) | 14:04 |
mgariepy | lol | 14:04 |
mgariepy | can we try setting the interpreter discovery to ayuto ? https://docs.ansible.com/ansible/latest/reference_appendices/interpreter_discovery.html#interpreter-discovery | 14:05 |
mgariepy | ha, not doing the correct stuff | 14:06 |
jrosser | i looked at this https://github.com/ansible/ansible/blob/c987c0c8a1c18be7a35157e9c75dcd2cdbdad003/lib/ansible/config/base.yml#L1503-L1518 | 14:07 |
jrosser | and i was confused | 14:08 |
mgariepy | i do indertand why i guess.. | 14:16 |
*** itandops has joined #openstack-ansible | 14:21 | |
*** ioni has quit IRC | 14:23 | |
*** ioni has joined #openstack-ansible | 14:23 | |
*** fghaas has quit IRC | 14:24 | |
*** ioni has quit IRC | 14:28 | |
*** ioni has joined #openstack-ansible | 14:30 | |
*** ioni has quit IRC | 14:34 | |
*** cshen has quit IRC | 14:36 | |
*** ioni has joined #openstack-ansible | 14:37 | |
*** dpaclt has quit IRC | 14:40 | |
*** cshen has joined #openstack-ansible | 14:42 | |
*** cshen has quit IRC | 14:46 | |
*** jawad_axd has quit IRC | 14:48 | |
vakuznet | hi, I'm seing a lot of "2020/06/02 05:56:37 [info] 96532#0: *7085 recv() failed (104: Connection reset by peer) while sending to client, client: 10.207.253.5, server: , request: "POST /v3/auth/tokens HTTP/1.1", upstream: "uwsgi://127.0.0.1:35358", host: "10.207.253.29:5000" in keystone-wsgi-public-error.log | 14:55 |
vakuznet | anyone have a clue? | 14:55 |
vakuznet | also this: "[error] 426#0: *29197 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 10.207.253.4, server: , request: "POST /v3/auth/tokens HTTP/1.1", upstream: "uwsgi://127.0.0.1:35358" | 14:59 |
mgariepy | jrosser, debian10 python interpreter stuff is not set in the base_config, | 15:35 |
mgariepy | it seems | 15:36 |
jrosser | the interesting place to look is in the nodebuilder log | 15:36 |
jrosser | becasue for focal py2 got pulled in as an accidental dependacy of something else | 15:36 |
jrosser | which is why the bionic CI image has py2, but a bionic cloud image does not | 15:36 |
mgariepy | setting the config to auto and having the mapping for debian seems to fix the issue | 15:39 |
mgariepy | at least for the lxc python3 stuff. | 15:40 |
mgariepy | and buster | 15:40 |
*** amalrajgenocidex has quit IRC | 15:46 | |
rh-jelabarre | I'm wondering if the "set_fact" module can set a value from the sontents of a URL. For example, the URL https://download.cirros-cloud.net/version/released shows me the current released version number for Cirros OS images, and I would like to use it in a playbook for downloading the current version, as well as using it in the OpenStack command for uploading images to the stack. | 15:48 |
*** gyee has joined #openstack-ansible | 15:49 | |
*** jcath has quit IRC | 16:08 | |
*** rpittau is now known as rpittau|afk | 16:09 | |
noonedeadpunk | #startmeeting openstack_ansible_meeting | 16:14 |
openstack | Meeting started Tue Jun 9 16:14:27 2020 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:14 |
openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:14 |
*** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)" | 16:14 | |
openstack | The meeting name has been set to 'openstack_ansible_meeting' | 16:14 |
noonedeadpunk | #topic office hours | 16:14 |
*** openstack changes topic to "office hours (Meeting topic: openstack_ansible_meeting)" | 16:14 | |
noonedeadpunk | sorry for being pretty late with that) | 16:14 |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: Add integrated test for Ubuntu Bionic using ML2/OVN driver https://review.opendev.org/733017 | 16:30 |
*** cshen has joined #openstack-ansible | 16:32 | |
*** cshen has quit IRC | 16:37 | |
jrosser | hmm hello just been in other meetings | 16:46 |
noonedeadpunk | I just don't have much to say since ptg tbh | 16:47 |
jrosser | if possible i'd like to have a bit more help with the work on merging focal in particular | 16:47 |
noonedeadpunk | like really had no time on working on stuff... | 16:47 |
jrosser | i've got a ton of patches and the end result is that focal passes but along the way other things break | 16:47 |
jrosser | so we need to consider the whole set and make some decisions on approach | 16:48 |
mgariepy | can we organise stuff in an etherpad ? or do you have a topic ? | 16:48 |
jrosser | i started putting a few things in here https://etherpad.opendev.org/p/openstack-ansible-ussuri | 16:48 |
jrosser | but it's not very organised yet - there are so many patches and things i've tried its now hard to keep on top of it all | 16:49 |
jrosser | this is not to do now (i'm tired) but maybe we can try to do a little every day to work out what is good/bad and start merging and fixing stuff | 16:50 |
noonedeadpunk | I'm for it | 16:52 |
* jrosser heads off to eat, need a break | 16:52 | |
mgariepy | i can help on the focal stuff and i'd like to also do some centos8 one but i would prioritize focal over centos. | 16:52 |
jrosser | it's all got so complicated now that i'd like to use an etherpad or something to all be understanding the same thing | 16:52 |
mgariepy | yes, i'm also for organising to work together instead of doing bits each in our side. | 16:53 |
noonedeadpunk | #endmeeting | 17:05 |
*** 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" | 17:05 | |
openstack | Meeting ended Tue Jun 9 17:05:38 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 17:05 |
openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-06-09-16.14.html | 17:05 |
openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-06-09-16.14.txt | 17:05 |
openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-06-09-16.14.log.html | 17:05 |
noonedeadpunk | :( | 17:05 |
*** jawad_axd has joined #openstack-ansible | 17:11 | |
*** maharg101 has joined #openstack-ansible | 17:13 | |
*** udesale_ has quit IRC | 17:13 | |
*** maharg101 has quit IRC | 17:14 | |
*** maharg101 has joined #openstack-ansible | 17:14 | |
*** jeh has quit IRC | 17:19 | |
openstackgerrit | Hervé Beraud proposed openstack/openstack-ansible master: Use unittest.mock instead of mock https://review.opendev.org/734452 | 17:22 |
*** jeh has joined #openstack-ansible | 17:40 | |
*** jawad_axd has quit IRC | 17:40 | |
mgariepy | https://github.com/ansible/ansible/blob/stable-2.9/lib/ansible/config/base.yml#L1451 fun | 17:41 |
mgariepy | 2.9 doesn't define debian 10 or ubuntu other than 14 and 16.. | 17:42 |
mgariepy | if it would define ubuntu 18.20 and debian we could set the interpreter_python to auto and it would select the correct one depending on the OS. | 17:43 |
*** fghaas has joined #openstack-ansible | 17:45 | |
*** sshnaidm is now known as sshnaidm|afk | 18:07 | |
*** AJaeger has joined #openstack-ansible | 18:22 | |
*** AJaeger has left #openstack-ansible | 18:23 | |
*** AJaeger has joined #openstack-ansible | 18:23 | |
AJaeger | could I get another review on ansible-hardening ussuri change to unbreak the sync job, please? https://review.opendev.org/733571 | 18:24 |
mgariepy | AJaeger, done | 18:25 |
AJaeger | thanks, mgariepy | 18:25 |
*** AJaeger has left #openstack-ansible | 18:25 | |
*** cshen has joined #openstack-ansible | 18:33 | |
*** cshen has quit IRC | 18:37 | |
*** jawad_axd has joined #openstack-ansible | 18:38 | |
*** iurygregory has quit IRC | 18:43 | |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: auto-detect interpreter. https://review.opendev.org/734672 | 18:53 |
*** jawad_axd has quit IRC | 19:02 | |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-tests master: auto-detect interpreter https://review.opendev.org/734676 | 19:09 |
*** ianychoi__ has joined #openstack-ansible | 19:16 | |
*** ianychoi_ has quit IRC | 19:19 | |
*** cshen has joined #openstack-ansible | 19:29 | |
*** jamesdenton has quit IRC | 19:32 | |
*** jamesdenton has joined #openstack-ansible | 19:38 | |
openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-systemd_service master: Remove fail: tasks https://review.opendev.org/734597 | 19:38 |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-tests master: auto-detect interpreter https://review.opendev.org/734676 | 19:39 |
*** jawad_axd has joined #openstack-ansible | 19:40 | |
*** jawad_axd has quit IRC | 19:45 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Use temporary repository for MariaDB 10.4.14 for Ubuntu focal https://review.opendev.org/733946 | 19:52 |
openstackgerrit | Merged openstack/ansible-hardening stable/ussuri: Enable syncing of docs https://review.opendev.org/733571 | 20:01 |
*** stingrayza has joined #openstack-ansible | 20:02 | |
*** also_stingrayza has quit IRC | 20:05 | |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-tests master: auto-detect interpreter https://review.opendev.org/734676 | 20:07 |
mgariepy | works a bit better when all the refs to /usr/bin/python are removed.. :/ | 20:09 |
jrosser | mgariepy: with 734676 you'd have to edit every repo tox.ini i think | 20:11 |
mgariepy | maybe, i want to see if it works correclty first tho :) | 20:12 |
*** cshen has quit IRC | 20:12 | |
mgariepy | for debian we need a PR on ansible. | 20:12 |
jrosser | there is another problem that you see described in the commit message of https://review.opendev.org/#/c/733948/ | 20:13 |
jrosser | the interpreter is *sometimes* in the tox virtualenv | 20:13 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: WIP - Use appropriate python interpreter for functional tests https://review.opendev.org/733948 | 20:16 |
mgariepy | the code in ansible to select the correct interpreter seems to works correctly, when does the interpreter is taken from the tox venv ? | 20:18 |
jrosser | well sort of like this https://zuul.opendev.org/t/openstack/build/06f2dbe525384484b8f0dbfde3a3c5a8/log/job-output.txt#2238 | 20:20 |
jrosser | but it's tangled up with the interpreter detection, because /usr/bin/python is python2 on the focal ci image | 20:20 |
jrosser | and there is no python-apt for python2 on focal | 20:21 |
jrosser | sorry | 20:21 |
jrosser | /usr/bin/python is not present on focal | 20:21 |
mgariepy | yep but with the python_interpreter set to auto_legacy (which is the default) it will use /usr/bin/pythnon first. but if set to auto it will use the distro info in the base.yml to select python3. | 20:23 |
mgariepy | i migth not fully understand what you are saying tho. | 20:23 |
jrosser | is setting it to auto going to work everywhere? | 20:24 |
mgariepy | i would assume so. | 20:24 |
mgariepy | it's going to be default at some point. | 20:24 |
mgariepy | and auto would take care of the centos7 stuff directly. | 20:25 |
jrosser | it looks like its working, thats great | 20:26 |
mgariepy | but it will not work for debian right now.. | 20:26 |
mgariepy | sadly it's not possible to ovewrite the list of python right now either. | 20:27 |
jrosser | right, so what do we think about https://review.opendev.org/#/c/733948/ instead | 20:28 |
jrosser | becasue we can put whatever logic we like in the test-ansible-functional.sh to set the interpreter per distro | 20:29 |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-tests master: auto-detect interpreter https://review.opendev.org/734676 | 20:29 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: WIP - Use appropriate python interpreter for functional tests https://review.opendev.org/733948 | 20:31 |
mgariepy | on bionic the /usr/bin/python is set to py2 or py3 ? | 20:31 |
jrosser | in a bionic cloud image it's not there, in openstack CI it's py2 | 20:32 |
mgariepy | so for lxc stuff you would have issue with pyhton_lxc module. | 20:32 |
jrosser | and we also mess up in our containers right now https://review.opendev.org/#/c/731457/ | 20:32 |
jrosser | yes | 20:32 |
mgariepy | if we use the auto stuff for bionic it would use py3. | 20:33 |
mgariepy | hmm when just updating the commit message zuul is not rechecking :( | 20:39 |
jrosser | buster is in the same sort of place as bionic for python lxc module https://zuul.opendev.org/t/openstack/build/e5b98191448946d095456e8bf2588af4/log/job-output.txt#10600-10612 | 20:42 |
*** cshen has joined #openstack-ansible | 20:44 | |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-tests master: auto-detect interpreter https://review.opendev.org/734676 | 20:46 |
mgariepy | looks like the tox.ini stuff is not needed | 20:46 |
*** cshen has quit IRC | 20:48 | |
*** jbadiapa has quit IRC | 20:48 | |
mgariepy | Yep, this is good, but will have to be in >= 2.10 only. I think it's too late to change the behavior for 2.9 (ie should not be backported). grrr.. https://github.com/ansible/ansible/pull/63097 | 20:49 |
jrosser | right - so you could do a hybrid | 20:50 |
mgariepy | so. we keep your patch for 2.9 and move to the new stuff on 2.10 for V ? | 20:50 |
mgariepy | i guess ? | 20:50 |
jrosser | take the env var / lookup that i did and set it to auto except for buster where it's explicitly /usr/bin/python3 | 20:50 |
jrosser | maybe theres a best-of-both approach | 20:50 |
mgariepy | or ... we lineinfile - add - buster to base.yml ahha | 20:51 |
mgariepy | it would be cleaner i think | 20:51 |
mgariepy | and if it fails we can always fallback to the buster approach. | 20:52 |
mgariepy | i need to leave for tonight i'll continue on that tomorrow. | 20:53 |
mgariepy | have a nice evening. | 20:55 |
jrosser | you too | 20:57 |
*** rh-jelabarre has quit IRC | 20:59 | |
*** rh-jelabarre has joined #openstack-ansible | 20:59 | |
*** fghaas has quit IRC | 21:19 | |
*** jawad_axd has joined #openstack-ansible | 21:29 | |
mgariepy | \o/ https://review.opendev.org/#/c/734676 focal passed. | 21:33 |
*** jawad_axd has quit IRC | 22:02 | |
*** kaiokmo has quit IRC | 22:13 | |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/ussuri: Update OVN metadata agent binary name https://review.opendev.org/733849 | 22:40 |
*** cshen has joined #openstack-ansible | 22:44 | |
*** cshen has quit IRC | 22:48 | |
*** tosky has quit IRC | 22:59 | |
*** spatel has quit IRC | 23:02 | |
*** gyee has quit IRC | 23:13 | |
*** jawad_axd has joined #openstack-ansible | 23:20 | |
*** itandops has quit IRC | 23:22 | |
*** trident has quit IRC | 23:27 | |
*** trident has joined #openstack-ansible | 23:29 | |
*** LuckyClover has joined #openstack-ansible | 23:35 | |
*** jawad_axd has quit IRC | 23:54 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!