| *** elico has joined #openstack-ansible | 00:43 | |
| elico | @: Any mariadb version recommended??? | 01:13 |
|---|---|---|
| *** gyee has quit IRC | 01:44 | |
| *** owalsh has joined #openstack-ansible | 02:29 | |
| *** owalsh_ has quit IRC | 02:33 | |
| *** dave-mccowan has quit IRC | 03:14 | |
| *** d34dh0r53 has joined #openstack-ansible | 03:14 | |
| *** rh-jelabarre has quit IRC | 03:52 | |
| *** d34dh0r53 has quit IRC | 04:04 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-ansible | 04:33 | |
| *** redrobot has quit IRC | 05:17 | |
| *** markvoelker has joined #openstack-ansible | 05:19 | |
| *** markvoelker has quit IRC | 05:23 | |
| *** markvoelker has joined #openstack-ansible | 05:26 | |
| *** udesale has joined #openstack-ansible | 05:27 | |
| *** markvoelker has quit IRC | 05:30 | |
| *** udesale has quit IRC | 05:50 | |
| *** elico has quit IRC | 05:52 | |
| *** udesale has joined #openstack-ansible | 06:04 | |
| *** arkan has quit IRC | 06:53 | |
| *** tosky has joined #openstack-ansible | 07:19 | |
| *** arkan has joined #openstack-ansible | 07:59 | |
| *** andrewbonney has joined #openstack-ansible | 08:00 | |
| *** arkan has quit IRC | 08:13 | |
| *** jbadiapa has joined #openstack-ansible | 08:28 | |
| *** stingrayza has quit IRC | 08:45 | |
| *** stingrayza has joined #openstack-ansible | 08:46 | |
| *** yolanda has quit IRC | 09:06 | |
| *** yolanda has joined #openstack-ansible | 09:16 | |
| arxcruz | noonedeadpunk: hey, regarding my patch the python_venv_build is called because of stackviz https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/tempest_install.yml#L38 | 09:28 |
| arxcruz | noonedeadpunk: and this package is missing on tripleo, because we have the undercloud and overcloud, upstream these two are only one machine with standalone jobs, however, when we run downstream, we can have a different machine running both, and os_tempest is executed through undercloud | 09:30 |
| arxcruz | it's a particular way, not sure other installers do the same | 09:30 |
| arxcruz | would be enough just add the virtualenv package on the other distros ? | 09:30 |
| noonedeadpunk | um, I think ut's not package but a role, that handles virtualenv creation and instaal required packages into it | 09:32 |
| arxcruz | jrosser: ^ | 09:32 |
| arxcruz | noonedeadpunk: the python_venv_build requires virtualenv package | 09:32 |
| arxcruz | which we don't have installed by default in undercloud | 09:32 |
| noonedeadpunk | Oh, wait, I missed the patch itself | 09:32 |
| arxcruz | noonedeadpunk: sorry, it was jrosser who -1 I made a mistake :D | 09:33 |
| arxcruz | anyway, you too need to review :P | 09:33 |
| noonedeadpunk | arxcruz: I think jrosser meant that package that needs to be installed should be placed for tempest_distro_packages not tempest_service_distro_packages | 09:34 |
| noonedeadpunk | as tempest_service_distro_packages got installed only for distro installations type, so it will miss for source one | 09:35 |
| noonedeadpunk | however, I think that probably instead we should patch python_venv_build role | 09:35 |
| noonedeadpunk | so that it got virtualenv isntalled once it really needs it | 09:35 |
| arxcruz | noonedeadpunk: i'm not familiar with python_venv_build, let me take a look where it install | 09:37 |
| noonedeadpunk | I think it should be added somewhere here? https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/vars/main.yml#L16 | 09:38 |
| arxcruz | noonedeadpunk: hmmm, it's not working | 09:43 |
| arxcruz | noonedeadpunk: https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/tasks/python_venv_install.yml#L28-L29 | 09:43 |
| noonedeadpunk | odyssey4me: do you think we shouldn't bother with carrying virtualenv into role? I mean, that installing virtualenv with os_tempest is not really appropiate as well... | 09:43 |
| arxcruz | noonedeadpunk: it will not install because those are empty by default | 09:43 |
| noonedeadpunk | yeah I see. | 09:43 |
| arxcruz | https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/defaults/main.yml#L35-L39 | 09:44 |
| noonedeadpunk | I mean, that anyway we either should put venv installation into python_venv_build role itself or keep it out... But os_tempest is not really appropriate place then as well... | 09:45 |
| noonedeadpunk | I think we install virtualenv with some pre-step playbook | 09:45 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible-openstack_hosts/src/branch/master/vars/debian.yml#L53 | 09:45 |
| arxcruz | noonedeadpunk: yeah, but if the role requires the package, that means the role should ensure the package is installed, not other pre-step playbook right ? | 09:46 |
| noonedeadpunk | which generally prepares hosts | 09:46 |
| arxcruz | noonedeadpunk: in case of osa installer, in the case of tripleo or other installer might not be true | 09:46 |
| noonedeadpunk | have you seen odyssey4me comment under your patch? | 09:46 |
| arxcruz | i understand this was initially created for osa | 09:46 |
| noonedeadpunk | I think I'm ok to move venv installation to python_venv_build role | 09:47 |
| arxcruz | brb, quickly | 09:47 |
| noonedeadpunk | arxcruz: but I think that tha point to keep this out of the role, so save some time, as it's heavily re-used, and triggering package installation every time is not as rational as can be | 09:49 |
| noonedeadpunk | In the meanwhile, this can be a pre-task in the playbook which runs os_tempest (or undercloud or whatever), which will install virtualenv package. I think you still have host preparation playbook? | 09:50 |
| noonedeadpunk | But I'm really pretty sure that os_tempest shouldn't bother with venv installation for another role | 09:54 |
| jrosser | seems like osa & tripleo (and any other user of python_venv_build) should carry their own opinion about how virtualenv gets installed because that can happen all sorts of ways | 10:03 |
| jrosser | it’s an indicator that there is no single right answer for this that the role makes an assumption that it’s been pre-prepared by something defining a particular deploy/install style | 10:05 |
| odyssey4me | jrosser: yes, that was always the state - the python_venv_create role was created with the assumption that virtualenv must already be there... which is why it did the check to see if it was and failed early | 10:07 |
| odyssey4me | jrosser noonedeadpunk something I've been thinking of doing is patching it to make it use python -m venv if it's py3 so that the requirement is lifted | 10:07 |
| odyssey4me | an alternative is to use virtualenv if it's there, and python -m venv if it's not | 10:08 |
| odyssey4me | noonedeadpunk: the reason we wanted to make sure it stays out was so that OSA only installs virtualenv in one place... otherwise things can get complicated | 10:10 |
| jrosser | can we assume always py3 these days? | 10:11 |
| arxcruz | jrosser: not on tripleo | 10:12 |
| arxcruz | xD | 10:12 |
| jrosser | the virtualenv version check got completely out of hand recently (different version string format with every new release nearly) and had to be simplified to present/not present | 10:12 |
| odyssey4me | jrosser: in the python_venv_build role, no - because it's generic to be used across multiple projects | 10:12 |
| arxcruz | we are moving some old jobs to centos8 but for now we can't assume py3 | 10:13 |
| odyssey4me | lemme push up a change to the role to use python -m venv as an alternative | 10:13 |
| jrosser | arxcruz: even for master branch you need py2? Victoria doesn’t support anything py2 does it? | 10:15 |
| arxcruz | jrosser: we still have queens and stein jobs | 10:16 |
| arxcruz | running on centos 7 | 10:16 |
| jrosser | but this role is branched, you need to respect those branches | 10:16 |
| arxcruz | we don't generate package for that, so we normally use latest version, tripleo-quickstart that is the installer, is branchless | 10:18 |
| jrosser | odyssey4me: I guess our position for master branch (V release) from an osa perspective was we can finally remove all support for py2 and centos7 | 10:19 |
| jrosser | U would be the last place we keep that | 10:19 |
| jrosser | this is gigantic tech debt for us right now | 10:19 |
| odyssey4me | jrosser: I understand that position, but consider that this is a utility role meant for broad consumption - while OSA may be using it branched, the original purpose was to let it be for general use and it did not need branching | 10:20 |
| odyssey4me | the cost on this role in particular shouldn't be much tech debt - but the position is understandable | 10:21 |
| odyssey4me | I'm guessing there's likely a big cost coming with pip changes (the new dep resolver) though... that's going to make life complicated | 10:22 |
| jrosser | yes it will, that’s going to require some big work | 10:22 |
| jrosser | also soon when osa drops centos7 the CI jobs for that will go as well | 10:22 |
| jrosser | odyssey4me: I think you are right tbh, python_venv_build master is quite portable anyway and we should keep it like that, but this new resolver is actually the biggest challenge we need to deal with | 10:28 |
| jrosser | I wonder if we should start early on that as it’s already there just not default in a current pip | 10:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Use 'python3 -m venv' instead of virtualenv for python3 https://review.opendev.org/744120 | 10:38 |
| *** markvoelker has joined #openstack-ansible | 10:41 | |
| odyssey4me | arxcruz: for the tripleo tests - the playbook should just add the ensure-virtualenv zuul role ahead of time... lemme get you an example | 10:42 |
| arxcruz | odyssey4me: i'm already working on it on tempest playbook | 10:43 |
| arxcruz | that calls os_tempest | 10:43 |
| arxcruz | odyssey4me: in fact, i made a mistake when I was cleaning up that playbook, and remove the virtualenv installation | 10:43 |
| arxcruz | that's why we are getting this error now xD | 10:43 |
| arxcruz | shame on me | 10:43 |
| odyssey4me | arxcruz: ah - I see you've done something there... I dunno if you might want to do https://review.opendev.org/744001 instead - I guess it depends on whether you want to use pypi or packages... and packages would be more stable | 10:44 |
| arxcruz | odyssey4me: yatin just reverted my patch, i think we are good, i just want to get rid of this bug, it's been haunting me for a week now | 10:45 |
| *** markvoelker has quit IRC | 10:46 | |
| *** mgariepy has quit IRC | 11:52 | |
| *** mgariepy has joined #openstack-ansible | 11:53 | |
| *** rgogunskiy has joined #openstack-ansible | 11:53 | |
| *** mgariepy has quit IRC | 12:01 | |
| *** rh-jelabarre has joined #openstack-ansible | 12:14 | |
| *** mgariepy has joined #openstack-ansible | 12:15 | |
| *** dave-mccowan has joined #openstack-ansible | 12:57 | |
| openstackgerrit | Jan Marquardt proposed openstack/openstack-ansible-os_designate master: Adds handler for copying policy.json to the right place https://review.opendev.org/744149 | 12:59 |
| *** blue_asni has joined #openstack-ansible | 13:16 | |
| *** jbadiapa has quit IRC | 13:17 | |
| *** spatel has joined #openstack-ansible | 13:26 | |
| *** blue_asni has quit IRC | 13:27 | |
| *** d34dh0r53 has joined #openstack-ansible | 13:51 | |
| *** udesale_ has joined #openstack-ansible | 14:05 | |
| *** udesale has quit IRC | 14:08 | |
| openstackgerrit | James Gibson proposed openstack/openstack-ansible-ops master: Make ELK Journalbeat seek configuration a variable https://review.opendev.org/744180 | 14:15 |
| *** spatel has quit IRC | 14:40 | |
| *** spatel has joined #openstack-ansible | 14:42 | |
| *** jbadiapa has joined #openstack-ansible | 14:45 | |
| openstackgerrit | James Gibson proposed openstack/openstack-ansible-ops master: Make ELK Journalbeat seek configuration a variable https://review.opendev.org/744180 | 14:50 |
| *** blue_asni has joined #openstack-ansible | 15:06 | |
| blue_asni | hello openstackers, please where can i get openstack-ansible chat archieve. | 15:09 |
| *** blue_asni has quit IRC | 15:12 | |
| *** chkumar|rover is now known as raukadah | 15:12 | |
| *** redrobot has joined #openstack-ansible | 15:38 | |
| *** udesale_ has quit IRC | 15:47 | |
| admin0 | in this process /bin/sh /etc/keepalived/pingable_check.sh 1 x.y.z --- how is this IP determined/computed ? | 15:48 |
| *** idlemind has joined #openstack-ansible | 15:49 | |
| admin0 | its on vrrp_script_pingable_check_script | 15:50 |
| *** irclogbot_1 has quit IRC | 15:56 | |
| *** irclogbot_2 has joined #openstack-ansible | 15:57 | |
| *** irclogbot_2 has quit IRC | 16:15 | |
| *** irclogbot_3 has joined #openstack-ansible | 16:15 | |
| jrosser | admin0: https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/haproxy/keepalived.yml#L44 | 16:42 |
| spatel | blue_asni - http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/ | 16:56 |
| *** andrewbonney has quit IRC | 16:58 | |
| *** kleini_ has quit IRC | 17:14 | |
| *** alanmeadows has quit IRC | 17:15 | |
| *** alanmeadows has joined #openstack-ansible | 17:15 | |
| admin0 | this is the 2nd deployment .. i am now using 20.1.4 .. it has 20.1.3 where it failed on setup-infrastrucutre on utility because it was not able to find the openstack command | 17:29 |
| spatel | jrosser: look like we short of stuck here with some of out standing issue. https://review.opendev.org/#/q/topic:osa/el8+(status:open+OR+status:merged) | 17:29 |
| *** gyee has joined #openstack-ansible | 17:45 | |
| admin0 | hi ..the playbook is stuck on python_venv_build: install python packages into the venv .. its on the keystone plaubook .. stuck for a very long time | 18:08 |
| admin0 | how long does it usually take ? 20.1.4 | 18:10 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Add nova_management_address to defaults https://review.opendev.org/741146 | 18:32 |
| *** spatel has quit IRC | 18:36 | |
| *** dave-mccowan has quit IRC | 18:46 | |
| *** spatel has joined #openstack-ansible | 18:57 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Deprecate os_congress role https://review.opendev.org/742521 | 19:26 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Deprecate os_congress role https://review.opendev.org/742521 | 19:27 |
| noonedeadpunk | folks, can we merge this one? https://review.opendev.org/#/c/742520/ | 19:33 |
| *** CeeMac has quit IRC | 19:34 | |
| *** CeeMac has joined #openstack-ansible | 19:45 | |
| jrosser | noonedeadpunk: have you ever seen error like this? http://paste.openstack.org/show/796504/ | 20:13 |
| jrosser | root@network1:~# sysctl net.ipv4.igmp_max_memberships | 20:24 |
| jrosser | net.ipv4.igmp_max_memberships = 20 | 20:24 |
| jrosser | oh dear | 20:24 |
| *** owalsh has quit IRC | 20:24 | |
| *** owalsh has joined #openstack-ansible | 20:45 | |
| *** spatel has quit IRC | 21:04 | |
| *** owalsh has quit IRC | 21:17 | |
| *** owalsh has joined #openstack-ansible | 21:32 | |
| noonedeadpunk | jrosser: no, never ever | 21:35 |
| noonedeadpunk | while I also have the same value for igmp_max_memberships (which is apperently too low for some cases) | 21:37 |
| *** CeeMac has quit IRC | 22:04 | |
| *** tosky has quit IRC | 22:10 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!