| *** osnaya has joined #openstack-ansible | 00:21 | |
| *** cjloader_ has quit IRC | 00:30 | |
| osnaya | @cloudnull @odyssey4me @logan container creation error during setup-hosts.yml log excerpt @http://paste.openstack.org/show/719520/ Can anyone provide pointers? | 00:31 |
|---|---|---|
| osnaya | <osnaya> @cloudnull @odyssey4me @logan same error occured for infra1_heat_apis_container, heat_engine and gnocchi as well | 00:38 |
| *** yifei1 has joined #openstack-ansible | 00:40 | |
| yifei1 | admin0: sorry I was offline yesterday | 00:41 |
| yifei1 | admin0: I did not try that. I don't know whether you noticed that in Queens OSA deploys all neutron agents on bare metal nodes for both control and compute nodes. | 00:43 |
| yifei1 | admin0: I think in your FAQ you only considered running neutron agents in Linux containers on control nodes. | 00:44 |
| osnaya | Has anyone else seen the issues during containers creation such as @ @http://paste.openstack.org/show/719520/ ?? | 00:47 |
| *** tobberydberg has quit IRC | 00:52 | |
| osnaya | will catch up later... | 00:54 |
| *** osnaya has quit IRC | 00:54 | |
| *** dxiri has quit IRC | 01:11 | |
| *** yifei1 has quit IRC | 01:23 | |
| *** yifei1 has joined #openstack-ansible | 01:27 | |
| *** dave-mccowan has joined #openstack-ansible | 01:37 | |
| *** tobberydberg has joined #openstack-ansible | 01:37 | |
| *** mma has quit IRC | 01:41 | |
| *** tobberydberg has quit IRC | 01:42 | |
| mmercer | cloudnull: sorry, been at a company outing all day -- so trying to execute the hatop as instructed yields -- insufficient permissions for socket path /var/run/haproxy.sock (which is weird because im executing as root) | 01:44 |
| *** tobberydberg has joined #openstack-ansible | 01:44 | |
| mmercer | well, this is peculiar... | 01:52 |
| *** markvoelker_ has joined #openstack-ansible | 01:52 | |
| *** markvoelker has quit IRC | 01:53 | |
| mmercer | haproxy.cfg has the correct stats socket line with proper permissions.... service is running successfully, but none of the groups is binding the listening ports.... and i cant run hatop because of insufficient permissions (this is with tag 17.0.2 btw) | 01:54 |
| *** Leo_m has quit IRC | 01:54 | |
| mmercer | ok, helps to use the stats file not the socket =D | 01:59 |
| mmercer | i have a *lot* of downs right now, but im also assuming that is at minimum because the infrastructure step has yet to actually complete to begin with | 02:06 |
| mmercer | ok, so repo_all-front-1 is open, repo_all-back is up | 02:14 |
| mmercer | but i cannot resolve any pip downloads against it because the actual port is not responsive on the aio instance | 02:14 |
| mmercer | iptables rules missing ? | 02:14 |
| mmercer | hmm, im not seeing haproxy be dependent on any other services, which doesnt make sense. given that there are both ebtables and iptables, there has to be rules defined in one of them to allow the port through | 02:27 |
| mmercer | but even beyond that,i dont see the listening port via ss | grep portnum, which should show the instantiated port | 02:27 |
| *** masuberu has joined #openstack-ansible | 02:27 | |
| *** deng_ has joined #openstack-ansible | 02:28 | |
| *** yifei1 has quit IRC | 02:29 | |
| *** masber has quit IRC | 02:31 | |
| mmercer | definitely not seeing it in firewalld, checked the active zone (currently public), and only ssh and dhcpv6-client are enabled, surprisingly | 02:31 |
| *** udesale has joined #openstack-ansible | 03:08 | |
| *** esberglu has joined #openstack-ansible | 03:18 | |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible master: Remove tempest_git_* overrides https://review.openstack.org/559161 | 03:19 |
| *** masuberu has quit IRC | 03:23 | |
| *** masuberu has joined #openstack-ansible | 03:33 | |
| *** nicolasbock has quit IRC | 03:34 | |
| *** masber has joined #openstack-ansible | 03:35 | |
| *** masuberu has quit IRC | 03:38 | |
| mmercer | ok, came to this page: https://docs.openstack.org/openstack-ansible/latest/admin/maintenance-tasks/firewalls.html and it sounds like you have to manually configure firewalld on the aio machine manually yourself | 04:08 |
| *** poopcat has quit IRC | 04:11 | |
| *** dxiri has joined #openstack-ansible | 04:31 | |
| *** Chris__ has joined #openstack-ansible | 04:41 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_nova master: Convert role to use a common systemd service role https://review.openstack.org/466179 | 04:42 |
| Chris__ | I'm experiencing a bug in openstack ansible when deploying, http://paste.openstack.org/show/719522/, it seems like its looking for a neutron_lbaas version that does not exist within the python package index. I looked online at https://pypi.org, and found that neutron_lbaas is depreciated, and as of right now the latest version available for download dis 12.0.0.0, does anyone know how to bypass this issue? | 04:44 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Update stack and add heartbeat https://review.openstack.org/562116 | 04:48 |
| *** openstack_ansibl has quit IRC | 04:59 | |
| *** epalper has joined #openstack-ansible | 04:59 | |
| *** bhujay has joined #openstack-ansible | 05:05 | |
| *** epalper has quit IRC | 05:05 | |
| *** Chris__ has quit IRC | 05:09 | |
| *** udesale_ has joined #openstack-ansible | 05:19 | |
| *** udesale has quit IRC | 05:19 | |
| *** mwarad has joined #openstack-ansible | 05:20 | |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_ironic master: update for depricated driver definition https://review.openstack.org/561277 | 05:20 |
| *** armaan has joined #openstack-ansible | 05:20 | |
| cloudnull | Chris__ My bet would be that the package is not part of your index at "http://int-dev.atl.sandvine.cloud:8181/os-releases/17.0.3/ubuntu-16.04-x86_64/" | 05:26 |
| odyssey4me | cloudnull looks like it's using the index at http://int-dev.atl.sandvine.cloud:8181/simple which is nginx, which will query pypiserver first, then pypi - and pypiserver should be serving it - if it's in the absolute constraints file then it did build the wheel | 05:28 |
| odyssey4me | this has come up a few times, but I've not been able to replicate it | 05:28 |
| prometheanfire | odyssey4me: if I could get direction on https://review.openstack.org/561277 that'd help (see the last few comments), using set_fact has ordering issues because of the facts usage in vars | 05:29 |
| odyssey4me | it's almost like pypiserver is either serving up the pypi index instead of its own, or it's not finding the package for some reason | 05:29 |
| cloudnull | its finding the other versions of that package | 05:29 |
| cloudnull | (from versions: 11.0.2, 11.0.3, 12.0.0.0b1, 12.0.0.0b2, 12.0.0.0b3, 12.0.0.0rc1, 12.0.0) | 05:29 |
| cloudnull | but from where I have no ide a | 05:30 |
| odyssey4me | cloudnull yeah, that looks like pypi's results to me | 05:30 |
| *** dxiri has quit IRC | 05:31 | |
| cloudnull | maybe pre=true is not enabled in the pip.conf ? | 05:31 |
| cloudnull | so its not finding or allowing a "dev" (pre-release) package ? | 05:31 |
| odyssey4me | it should be, we enable by default | 05:31 |
| *** dxiri has joined #openstack-ansible | 05:31 | |
| cloudnull | id be curious to see what pip search shows and whats in that dir as built packages | 05:32 |
| cloudnull | also if theres anything interesting in the pypyserver logs | 05:33 |
| cloudnull | maybe its skipping files for whatever reason | 05:33 |
| cloudnull | admin0: ^ | 05:33 |
| odyssey4me | prometheanfire that makes my brain hurt | 05:34 |
| prometheanfire | odyssey4me: the indirection/ordering issues? | 05:35 |
| odyssey4me | man, that thing is just a mess of vars - isn't there a way to simplify that into a dict | 05:36 |
| odyssey4me | what's with all the vars of vars? | 05:36 |
| *** dxiri has quit IRC | 05:36 | |
| odyssey4me | and then set facts of the same vars | 05:36 |
| prometheanfire | odyssey4me: hell if I know, I came in trying to fix https://bugs.launchpad.net/openstack-ansible/+bug/1758351 | 05:36 |
| openstack | Launchpad bug 1758351 in openstack-ansible "ironic is using old things" [High,In progress] - Assigned to Matthew Thode (prometheanfire) | 05:36 |
| cloudnull | looks like ps2 works? | 05:37 |
| prometheanfire | ps2 was a good system | 05:37 |
| odyssey4me | works, but just as bad in terms of all the vars all over the place | 05:37 |
| odyssey4me | also, note that works just means that it converges - there's no real test in this role afaik | 05:38 |
| prometheanfire | odyssey4me: also true | 05:38 |
| cloudnull | it also doesn't really change how the role functions | 05:38 |
| cloudnull | however I agree, that role is a mess of vars. | 05:38 |
| odyssey4me | looks like the set fact is just setting a bunch of stuff to use in the templates, so why not just implement most of what's being done with the vars of vars in the template and use simple conditionals for the templates to adapt based on standalone vs whatever the other thing is | 05:39 |
| cloudnull | some cleanup would be nice, though idk if it needs to ride in on that pr. | 05:39 |
| prometheanfire | I'm willing to tackle it in another review | 05:39 |
| odyssey4me | I'd suggest cleaning it up first, then doing the new vars thing on top of that | 05:40 |
| prometheanfire | that'll delay me a couple weeks probably, have to focus on n->q stuff | 05:40 |
| odyssey4me | well, as it stands it's unreadable | 05:40 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use production like interfaces in the gate https://review.openstack.org/553636 | 05:41 |
| prometheanfire | yep | 05:41 |
| prometheanfire | odyssey4me: patchset 2 is my prefered solution for right now, it's a little cleaner (doesn't use set_fact) | 05:41 |
| cloudnull | ok, gotta run. | 05:42 |
| prometheanfire | ya, about to go to bed myself | 05:42 |
| * cloudnull needs sleep, has another whole day of meetings tomorrow | 05:42 | |
| prometheanfire | woo | 05:42 |
| cloudnull | indeed. take care all | 05:43 |
| odyssey4me | right now, there is ironic_openstack_driver_loaded_list which is only used in two places - defaults/main and the set_fact in ironic_post_install, that set fact sets ironic_driver_list which is used in the template - take out the middle man and just build that in the template | 05:43 |
| prometheanfire | odyssey4me: that's how I generally prefer to do things in puppet as well | 05:49 |
| prometheanfire | like https://blog.mthode.org/posts/2017/Jun/gentoo-portage-templates/ | 05:49 |
| odyssey4me | I suggest taking proactive step to improve the role, instead of putting blinkers on and only doing what the bug needs solved. | 05:49 |
| odyssey4me | I'm reviewing what's there and making some suggestions. | 05:50 |
| prometheanfire | odyssey4me: I'd love to, but I can't do it immediately | 05:50 |
| odyssey4me | doing it immediately will get your patch in faster, because it'll be easier to read. | 05:50 |
| odyssey4me | getting reviews is all about making it easy for the reviewer - when it's complicated, no-one wants to take responsibility for it | 05:51 |
| *** armaan has quit IRC | 05:51 | |
| odyssey4me | especially for a service the majority of core reviewers don't understand | 05:51 |
| prometheanfire | I have someone threatening to revoke our business licence, an old bdfl trying to come back and take over, a very long meeting to run/plan RFPs to send out for acocunting, find a lawyer | 05:51 |
| prometheanfire | that's just the gentoo shit | 05:52 |
| prometheanfire | so I don't have time right now | 05:52 |
| *** yolanda has joined #openstack-ansible | 05:52 | |
| prometheanfire | I'd love to basically rewrite the ironic stuff (like you said, it kinda sucks), but can't commit to it | 05:52 |
| odyssey4me | well, I don't know what to say - I've given some advice and am reviewing the patch. The rest is up to you. If this matters to you, make the time. If not, then it'll stagnate. | 05:53 |
| prometheanfire | yep | 05:53 |
| prometheanfire | we agree there :P | 05:54 |
| prometheanfire | still working on the gentoo stuff, which is more important to me, devstack is dragging their feet reviewing patches | 05:55 |
| prometheanfire | anyway, nn | 05:56 |
| *** deng_ has quit IRC | 05:56 | |
| *** TobbeCN has joined #openstack-ansible | 05:57 | |
| * prometheanfire wonders if osa can use libvirt-python-4.2.0 on xenial (broken it seems), anyway, nn for realz yo | 05:58 | |
| *** _mwarad_ has joined #openstack-ansible | 06:03 | |
| *** mwarad has quit IRC | 06:03 | |
| *** bhujay has quit IRC | 06:04 | |
| *** bhujay has joined #openstack-ansible | 06:04 | |
| *** cshen has joined #openstack-ansible | 06:08 | |
| *** cshen has quit IRC | 06:09 | |
| *** viktor_t has joined #openstack-ansible | 06:13 | |
| *** cshen has joined #openstack-ansible | 06:14 | |
| *** armaan has joined #openstack-ansible | 06:16 | |
| *** _mwarad_ has quit IRC | 06:16 | |
| *** sar has joined #openstack-ansible | 06:16 | |
| *** bhujay has quit IRC | 06:20 | |
| *** bhujay has joined #openstack-ansible | 06:20 | |
| *** pcaruana has joined #openstack-ansible | 06:21 | |
| *** Taseer has quit IRC | 06:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic master: Do not log passwords https://review.openstack.org/562111 | 06:35 |
| evrardjp | prometheanfire: let me check at that | 06:36 |
| *** Taseer has joined #openstack-ansible | 06:38 | |
| odyssey4me | evrardjp if you could work over all the no password reviews I'd be grateful - they're blocking the virtualenv/pip fixes | 06:39 |
| evrardjp | yup I have them open, I got connectivity issues yesterday. | 06:40 |
| evrardjp | hard to do proper reviews on a phone. | 06:40 |
| odyssey4me | ah, I didn't realise you were afk - no worries | 06:41 |
| odyssey4me | I thought you were just offline to concentrate on other work. | 06:41 |
| jrosser | odyssey4me: upstream pypi repos seem to hace changed cdn arrangement in the last few days, pretty much everything is now a redirect | 06:43 |
| odyssey4me | jrosser hmm, that's a good point - perhaps pypiserver needs an update | 06:43 |
| jrosser | that broke our internal proxy/mirror in unusual ways, one to watch out for | 06:43 |
| evrardjp | odyssey4me: I was, then I tried to get back online. | 06:44 |
| evrardjp | that's true I read an article about it! | 06:45 |
| evrardjp | yeah we should do things. | 06:45 |
| evrardjp | odyssey4me: I am very sad, I thought these trivial patches would be merged by other ppl by now. | 06:46 |
| odyssey4me | apparently everyone is busy :/ | 06:47 |
| *** blinkiz has quit IRC | 06:48 | |
| evrardjp | odyssey4me: for trove should we make centos nv ? | 06:49 |
| openstackgerrit | Merged openstack/openstack-ansible-os_searchlight master: Do not log passwords https://review.openstack.org/562226 | 06:51 |
| evrardjp | http://paste.openstack.org/show/719524/ | 06:51 |
| evrardjp | I guess cryptography version isn't aligned | 06:52 |
| *** deng_ has joined #openstack-ansible | 06:58 | |
| evrardjp | oh I guess chicken and egg | 06:59 |
| *** osnaya has joined #openstack-ansible | 07:00 | |
| evrardjp | odyssey4me: you're gonna kill me, but do we REALLY need the constraints mechanism in the pip_install role? | 07:02 |
| evrardjp | (for a later improvement) | 07:02 |
| evrardjp | I cannot understand why we don't simplify this now in master (keeping the rest the same) | 07:02 |
| odyssey4me | simplify how? | 07:03 |
| *** yifei1 has joined #openstack-ansible | 07:03 | |
| osnaya | @admin0 @evrardjp container creation error during setup-hosts.yml log excerpt @http://paste.openstack.org/show/719520/ Can anyone provide pointers? exact same error seen for rabbitmq, nova_api_placement, heat_engine, heat_api and gnocchi.... | 07:04 |
| evrardjp | odyssey4me: just use pip from packages and pip install --upgrade pip==x setuptools==y wheels==z virtualenv==t (which versions x,y,z,t in the role defaults as requirements, not constraints) | 07:04 |
| odyssey4me | osnaya I have no freaking idea - perhaps your hosts are dirty from a previous configuration? | 07:05 |
| osnaya | @admin0 @evrardjp @odyssey4me ... with the error during container creation on setup-hosts.yml I am blocked and need help please | 07:05 |
| odyssey4me | looks like the networking setup failed, so it's most likely something to do with that | 07:05 |
| evrardjp | osnaya: yeah what odyssey4me is true | 07:06 |
| odyssey4me | evrardjp for what purpose? what does it help? | 07:06 |
| evrardjp | you have a dangling veth pair for example | 07:06 |
| osnaya | @odyssey4me Is there a way to clean up other than reimage with Ubuntu? | 07:06 |
| odyssey4me | osnaya given that we have no idea what you did, you're the only one who can answer that | 07:06 |
| evrardjp | odyssey4me: I will check if simply installing virtualenv from packages wouldn't be enough for all we need on the hosts, as part of the plan | 07:07 |
| evrardjp | it would be an intermediate step | 07:07 |
| odyssey4me | evrardjp hwoarang has a few patches up to try something like that | 07:07 |
| evrardjp | if the os doesn't provide the proper package, that would install the proper version | 07:07 |
| *** mbuil has joined #openstack-ansible | 07:07 | |
| evrardjp | I will spin up recent centos to see where the packaging is. | 07:07 |
| evrardjp | because it was the only one holding us back | 07:08 |
| *** shardy has joined #openstack-ansible | 07:10 | |
| odyssey4me | evrardjp as mentioned before, if we can progress the python build simplification, then pip_install will only have to execute on a much smaller amount of hosts anyway - so at this point I'd rather have you try and figure out how to help me progress that - I seem to be getting stuck and going in circles | 07:13 |
| osnaya | @odyssey4me @evrardjp ... Ok in that case re-imaging the only option for now, right? OK, will re-install tomorrow... Just curious what pointed out that there is dangling veth pair? | 07:14 |
| odyssey4me | osnaya it'll certainly be simpler | 07:14 |
| osnaya | @odyssey4me @evrardjp @admin0 Wanted to check before I re-image.... since someone else reported similar issue on nova_api_.... wanted to understand if there is anything I should look out for... | 07:16 |
| odyssey4me | osnaya then that's between you and admin0 | 07:17 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Do not log passwords https://review.openstack.org/562225 | 07:18 |
| evrardjp | odyssey4me: how can I help on the circle thing? | 07:21 |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat master: Do not log passwords https://review.openstack.org/562223 | 07:24 |
| hwoarang | evrardjp: odyssey4me the patches i have posted for lxc_hosts and galera friends seem to be enough for infra to be pip clean | 07:24 |
| evrardjp | hwoarang: that's cool | 07:25 |
| evrardjp | how do you ensure pip clealiness? | 07:25 |
| evrardjp | should we build a test for that? | 07:25 |
| odyssey4me | awesome :) | 07:25 |
| odyssey4me | that means that we can ditch the pip_install execution from there if we want to | 07:25 |
| evrardjp | hwoarang: I have done a check experimental to see if there is no problem in the general build | 07:26 |
| hwoarang | ok | 07:26 |
| *** osnaya has quit IRC | 07:26 | |
| evrardjp | odyssey4me: they are already ditched | 07:26 |
| evrardjp | you mean the pip_install meta dependency? | 07:26 |
| evrardjp | now everything is in openstack_hosts | 07:26 |
| evrardjp | but I mean we can do a generic flip in openstack_hosts | 07:27 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: Do not log passwords https://review.openstack.org/562221 | 07:27 |
| evrardjp | put it only in containers | 07:27 |
| odyssey4me | yes, that's what I mean - openstack_hosts still installs it everywhere - we may be able to reduce that footprint | 07:27 |
| evrardjp | yes if it becomes something NOT to do across the board, it has no place in openstack_hosts. | 07:27 |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: Do not log passwords https://review.openstack.org/562229 | 07:28 |
| *** pmannidi has quit IRC | 07:29 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate master: Do not log passwords https://review.openstack.org/562222 | 07:32 |
| *** mma has joined #openstack-ansible | 07:32 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Do not log passwords https://review.openstack.org/562230 | 07:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_panko master: Do not log passwords https://review.openstack.org/562246 | 07:36 |
| *** electrofelix has joined #openstack-ansible | 07:40 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Do not log passwords https://review.openstack.org/562231 | 07:40 |
| *** epalper has joined #openstack-ansible | 07:41 | |
| evrardjp | or if it's still to do accross the board EXCEPTED some groups, then I guess we can avoid the run when: '<exception>' in groupnames | 07:44 |
| *** pcaruana has quit IRC | 07:45 | |
| *** pcaruana has joined #openstack-ansible | 07:46 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh master: Do not log passwords https://review.openstack.org/562227 | 07:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Do not log passwords https://review.openstack.org/562236 | 07:48 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Do not log passwords https://review.openstack.org/562233 | 07:50 |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: Do not log passwords https://review.openstack.org/562232 | 07:55 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara master: Do not log passwords https://review.openstack.org/562234 | 07:59 |
| *** DanyC has joined #openstack-ansible | 08:00 | |
| *** tosky has joined #openstack-ansible | 08:03 | |
| *** DanyC has quit IRC | 08:04 | |
| *** admin0 has quit IRC | 08:05 | |
| hwoarang | odyssey4me: evrardjp passing '-c http://git.openstack.org/cgit/openstack/requirements/plain/upper-constraints.txt?h=master' to pip it means 'use the packages listed there are upper constrains' right? so if a lower version is installed it shouldn't be upgraded | 08:09 |
| hwoarang | correct? | 08:09 |
| *** bhujay has quit IRC | 08:10 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Do not log passwords https://review.openstack.org/562022 | 08:10 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/queens: Do not log passwords https://review.openstack.org/562021 | 08:10 |
| odyssey4me | hwoarang nope - whatever's in the constraints will be installed | 08:12 |
| odyssey4me | requirements are ignored | 08:12 |
| hwoarang | so what's the different with the -r parameter/ | 08:12 |
| odyssey4me | -r tells it what to install from a package standpoint, constraints tells it what to install from a version standpoint | 08:13 |
| hwoarang | ok | 08:13 |
| odyssey4me | you can have hundreds of packages in constraints, but only the packages in -r will be installed | 08:13 |
| odyssey4me | it's a little counterintuitive, I know | 08:13 |
| hwoarang | i guess there is a good reason to request specific versions? | 08:13 |
| *** cshen has quit IRC | 08:14 | |
| odyssey4me | yep, it's the only way we can pin things to ensure that the versions are always the same for every tag | 08:14 |
| hwoarang | ok then i hit a deadend | 08:14 |
| odyssey4me | otherwise your experience today is different to your experience tomorrow | 08:14 |
| hwoarang | because the contrains in os_tempest always pulls the latest cryptography | 08:14 |
| hwoarang | overriding the distro package | 08:14 |
| hwoarang | so the tempest role has to be converted to distro/source as well for tempest to actually work | 08:14 |
| hwoarang | with distro keystone | 08:15 |
| odyssey4me | can the role not have entirely seperate code paths for distro vs non distro? | 08:15 |
| evrardjp | you can install a specific version with the -r | 08:15 |
| hwoarang | you mean os_tempest? | 08:15 |
| hwoarang | sure | 08:15 |
| hwoarang | i was just hoping to use the existing code in os_tempest to make keystone work | 08:15 |
| *** gkadam has joined #openstack-ansible | 08:16 | |
| odyssey4me | yeah, I think it's going to require some work in the tempest role too | 08:16 |
| odyssey4me | although tempest is installed into a venv, so what's the issue with cryptography? is it installed on the host? | 08:17 |
| hwoarang | yeah os_tempest also pulls packages on host | 08:19 |
| odyssey4me | ah, we should try to eliminate that | 08:19 |
| hwoarang | https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/tempest_install.yml#L36 | 08:19 |
| hwoarang | yeah... | 08:19 |
| hwoarang | that messes things up | 08:19 |
| *** mwarad has joined #openstack-ansible | 08:20 | |
| odyssey4me | hmm, https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L152-L158 | 08:20 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest master: Do not log passwords https://review.openstack.org/562239 | 08:20 |
| hwoarang | yeah the pip packages look rather strange :p | 08:20 |
| hwoarang | no clue why they need to be on the host | 08:20 |
| *** mwarad has quit IRC | 08:21 | |
| odyssey4me | so virtualenv is there for obvious reasons, shade is there and the others are there because we're using the ansible modules to set things up in the cloud | 08:21 |
| *** mwarad has joined #openstack-ansible | 08:22 | |
| odyssey4me | we could either move those into the venv, and use the venv when executing those tasks with ansible_python_interpreter | 08:22 |
| odyssey4me | or we could move those into the ansible-runtime venv and make those tasks delegate to localhost | 08:22 |
| odyssey4me | if we get rid of the use of our custom modules, we could probably also get rid of several of those packages | 08:23 |
| hwoarang | the bit i dont understand is why the various openstack clients have to be one the host *and* in the tempest venv | 08:23 |
| odyssey4me | they don't - it's just lazy | 08:24 |
| hwoarang | they are used by ansible i suppose so it's easier ot have them on the host than using them from venv | 08:24 |
| hwoarang | that's my understanding of what you said | 08:24 |
| odyssey4me | nope, just set the python interpreter - that's it | 08:24 |
| hwoarang | ok | 08:24 |
| odyssey4me | perhaps when this was first done that wasn't an option - or it wasn't known | 08:24 |
| hwoarang | ok i will have a look then | 08:25 |
| odyssey4me | I'd go for removing those and setting the interpreter on all the tasks. | 08:25 |
| *** radeks has joined #openstack-ansible | 08:25 | |
| odyssey4me | virtualenv should ideally be the only one to stay - and we can use a distro package for that. | 08:25 |
| odyssey4me | we should be able to do much the same in most roles | 08:26 |
| hwoarang | why do we need virtualenv there? other os_* roles do not explicitly pull virtualenv | 08:26 |
| hwoarang | at least keystone doesn't | 08:26 |
| hwoarang | since that's in front of me | 08:27 |
| odyssey4me | probably an oversight - it probably gets installed by something else... | 08:27 |
| hwoarang | maybe we assume that something pulled that in already at an earlier stage | 08:27 |
| hwoarang | yeah | 08:27 |
| *** DanyC has joined #openstack-ansible | 08:29 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Do not log passwords https://review.openstack.org/562111 | 08:32 |
| *** mma has quit IRC | 08:33 | |
| Taseer | evrardjp: can we close congress once and for all ? => https://review.openstack.org/#/c/560839/ | 08:34 |
| Taseer | or do I still need to have some functional testing ? | 08:36 |
| *** jwitko has quit IRC | 08:53 | |
| *** bhujay has joined #openstack-ansible | 09:00 | |
| evrardjp | Taseer: it's not killing once and for all, it's just the start :) | 09:04 |
| evrardjp | but yes, for me it means the work is done | 09:04 |
| Taseer | evrardjp: I see I got your approval. But again, do you still want functional tests in the future ? so I had prepare myself for further work | 09:06 |
| *** hamza21 has joined #openstack-ansible | 09:27 | |
| *** mwarad has quit IRC | 09:34 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: templates: keystone-uwsgi: Allow uwsgi to autoload required plugins. https://review.openstack.org/560307 | 09:38 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: [RFC] Add support for using distribution packages for OpenStack services https://review.openstack.org/560308 | 09:38 |
| *** vnogin has joined #openstack-ansible | 09:39 | |
| *** foutatoro has joined #openstack-ansible | 09:44 | |
| *** armaan has quit IRC | 09:55 | |
| *** cshen has joined #openstack-ansible | 09:55 | |
| *** armaan has joined #openstack-ansible | 09:55 | |
| *** vnogin has quit IRC | 09:58 | |
| *** vnogin has joined #openstack-ansible | 10:02 | |
| *** cshen has quit IRC | 10:02 | |
| *** bhujay has quit IRC | 10:33 | |
| *** dxiri has joined #openstack-ansible | 10:38 | |
| *** sar has quit IRC | 10:40 | |
| *** dxiri has quit IRC | 10:42 | |
| ivve | hey guys, would it be considered a deployerror to deploy the openstack with defaults (self signed cert) and not have aodh confed not to verify ssl cert? | 10:44 |
| evrardjp | Taseer: no I think those functional tests are enough | 10:44 |
| evrardjp | Taseer: just maintain them as working :) | 10:44 |
| ivve | i mean deployer must be aware of it & use an override | 10:44 |
| evrardjp | let's move them to voting when we had multiple runs okay without timeouts | 10:44 |
| ivve | feels like it should default to, aodh.conf rest_notifier_ssl_verify = false | 10:45 |
| ivve | in the [DEFAULT] stanza | 10:45 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server master: Set the pypiserver root to the correct directory https://review.openstack.org/562544 | 10:45 |
| odyssey4me | evrardjp hwoarang https://review.openstack.org/562544 solves an inconsistent bug which has been popping up - it'd be good to get that resolved asap | 10:46 |
| evrardjp | odyssey4me: depends on ^ for https://review.openstack.org/#/c/549012/2 ? | 10:51 |
| odyssey4me | evrardjp then I lose my votes ;) | 10:51 |
| evrardjp | I will apply them. | 10:51 |
| odyssey4me | but yeah, will do that now | 10:51 |
| evrardjp | Except if you want to keep mhayden vote. | 10:51 |
| evrardjp | that's collector now | 10:51 |
| evrardjp | :) | 10:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove default pip_links value https://review.openstack.org/549012 | 10:52 |
| evrardjp | andymccr: could you quickly vote on https://review.openstack.org/#/c/562544/1 please? | 10:52 |
| andymccr | si si | 10:53 |
| odyssey4me | tyvm all | 10:53 |
| andymccr | nice catch. damn | 10:53 |
| *** nicolasbock has joined #openstack-ansible | 10:57 | |
| *** armaan has quit IRC | 10:59 | |
| *** armaan has joined #openstack-ansible | 10:59 | |
| *** sar has joined #openstack-ansible | 11:00 | |
| *** cshen has joined #openstack-ansible | 11:00 | |
| *** ansmith has quit IRC | 11:13 | |
| *** mwarad has joined #openstack-ansible | 11:19 | |
| *** deng_ has quit IRC | 11:20 | |
| *** armaan has quit IRC | 11:37 | |
| evrardjp | hwoarang: andymccr cloudnull mgariepy we have plenty of reviews pending, for example Jesse's patches are queuing up: https://review.openstack.org/#/q/owner:jesse.pretorius%2540rackspace.co.uk+status:open++label:%22-Workflow%22+is:mergeable | 11:39 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Add IP addresses to infra node br-vxlan in the examples https://review.openstack.org/562329 | 11:40 |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca-agent master: Replace virtualenv-tools with a script https://review.openstack.org/562208 | 11:42 |
| *** mugsie has quit IRC | 11:42 | |
| *** mugsie has joined #openstack-ansible | 11:42 | |
| *** mugsie has quit IRC | 11:42 | |
| *** mugsie has joined #openstack-ansible | 11:42 | |
| *** vnogin has quit IRC | 11:43 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_monasca stable/ocata: Initial OSA zuul v3 role jobs https://review.openstack.org/516646 | 11:45 |
| *** vnogin has joined #openstack-ansible | 11:46 | |
| *** vnogin has quit IRC | 11:47 | |
| *** viktor_t has quit IRC | 11:51 | |
| *** sar has quit IRC | 11:54 | |
| *** sar has joined #openstack-ansible | 11:58 | |
| *** armaan has joined #openstack-ansible | 12:00 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/queens: Remove perl library from Ubuntu installs https://review.openstack.org/562317 | 12:00 |
| *** mwarad has quit IRC | 12:01 | |
| *** vnogin has joined #openstack-ansible | 12:03 | |
| *** vnogin has quit IRC | 12:07 | |
| *** vnogin has joined #openstack-ansible | 12:07 | |
| *** deng_ has joined #openstack-ansible | 12:10 | |
| *** armaan has quit IRC | 12:11 | |
| *** vnogin has quit IRC | 12:17 | |
| *** vnogin has joined #openstack-ansible | 12:17 | |
| *** armaan has joined #openstack-ansible | 12:17 | |
| *** armaan has quit IRC | 12:22 | |
| *** vnogin has quit IRC | 12:31 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/pike: Extend linters time out https://review.openstack.org/562027 | 12:32 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/queens: Re-order database migrations https://review.openstack.org/562023 | 12:37 |
| mgariepy | morning everyone | 12:37 |
| * mgariepy reviewing | 12:38 | |
| andymccr | mgariepy: hey! | 12:43 |
| mgariepy | what's up andymccr | 12:43 |
| *** sar has quit IRC | 12:45 | |
| *** vnogin has joined #openstack-ansible | 12:45 | |
| *** ansmith has joined #openstack-ansible | 12:47 | |
| *** yifei1 has quit IRC | 12:47 | |
| *** vnogin has quit IRC | 12:51 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_searchlight master: Only replace python when re-initializing the venv https://review.openstack.org/562196 | 12:51 |
| *** cmart has joined #openstack-ansible | 12:55 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Remove spurious VxLAN bridge IP from ceph osd hosts https://review.openstack.org/560878 | 13:01 |
| *** armaan has joined #openstack-ansible | 13:02 | |
| *** esberglu has quit IRC | 13:05 | |
| *** markvoelker_ has quit IRC | 13:16 | |
| *** mwarad has joined #openstack-ansible | 13:18 | |
| *** markvoelker has joined #openstack-ansible | 13:18 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Only replace python when re-initializing the venv https://review.openstack.org/562163 | 13:19 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Remove Octavia from translations scenario https://review.openstack.org/562578 | 13:19 |
| evrardjp | could I have votes on ^ | 13:19 |
| evrardjp | that would unblock translations | 13:19 |
| evrardjp | andymccr: maybe ? | 13:19 |
| andymccr | looking! | 13:20 |
| andymccr | LGTM | 13:20 |
| *** esberglu has joined #openstack-ansible | 13:24 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Do not log passwords https://review.openstack.org/562582 | 13:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 13:27 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/ocata: Fix repo_build_venv_rebuild toggle effectiveness https://review.openstack.org/562337 | 13:27 |
| evrardjp | that one is preventing success in horizon and for the next sha bump: https://review.openstack.org/#/c/562582/ | 13:28 |
| evrardjp | thanks everyone! | 13:28 |
| mgariepy | hwoarang, http://logs.openstack.org/33/562033/1/gate/openstack-ansible-functional-opensuse-423/0bd6742/job-output.txt.gz#_2018-04-19_13_04_13_775635 | 13:28 |
| mgariepy | seems like the image got renamed.. | 13:29 |
| mgariepy | https://github.com/openSUSE/docker-containers-build/commit/81638d69fffe52e00863e83e9949a0c5d2cfdba3 | 13:29 |
| evrardjp | that's gonna be painful | 13:32 |
| evrardjp | haters are gonna hate! | 13:32 |
| odyssey4me | :/ and the old image is no longer present? that's not good practise | 13:32 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Only replace python when re-initializing the venv https://review.openstack.org/562177 | 13:33 |
| odyssey4me | thankfully that's master only, but if this continues happening then we're going to have trouble with these oci image sources - we need a stable source | 13:33 |
| hwoarang | mgariepy: i am on it | 13:34 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/ocata: Extend linters time out https://review.openstack.org/562028 | 13:34 |
| hwoarang | mgariepy: https://github.com/openSUSE/docker-containers-build/issues/29 | 13:34 |
| hwoarang | i am working on some ansible code to determine the build number on the fly | 13:34 |
| evrardjp | hwoarang: you're the deal | 13:35 |
| *** TobbeCN has quit IRC | 13:35 | |
| hwoarang | odyssey4me: from what i can tell these images are just autogenerated from time to time (for all distros) so it may not be the same image all the time. The rest of the distros do not carry any build info on them as well. they use the same tarball for all releases | 13:35 |
| *** TobbeCN has joined #openstack-ansible | 13:36 | |
| hwoarang | i have to ask cloudnull why we moved to these images instead of those provided by images.linuxcontainers.org | 13:36 |
| idlemind | so the successerator (did i spell that right) that you guys use for retries ... can that be harvested for use in other playbooks? in particular i'm looking for a solution to deal w/intermittent failures of the new "network_cli" connection while connected via VPN the hosts aren't really down, just for some reason I occasionally see failures connecting to some of the devices that way. | 13:36 |
| odyssey4me | idlemind we haven't used the successerator for some time | 13:37 |
| evrardjp | hwoarang: I think it's not something to only ask. I think it should be documented. | 13:37 |
| evrardjp | Part of the reference architecture if possible | 13:37 |
| evrardjp | a new section here: https://docs.openstack.org/openstack-ansible/latest/reference/architecture/index.html , like container sources? | 13:37 |
| *** jwitko has joined #openstack-ansible | 13:38 | |
| odyssey4me | or perhaps we should just build our own instead? | 13:38 |
| hwoarang | maybe yes | 13:38 |
| evrardjp | odyssey4me: what do you mean there? | 13:38 |
| evrardjp | I am scared. | 13:38 |
| evrardjp | hwoarang: I am pretty sure it's cloudnull 's crusage to make lxc BE THE SAME as nspawn. | 13:39 |
| odyssey4me | well, there's a role for diskimage-builder - and the base image does not need very much | 13:39 |
| idlemind | odyssey4me shame, ya i didn't get a response in #ansible-network on the retry ability of network_cli but it doesn't seem to follow the normal ansible_ssh_retries ... i'll have to keep looking or i'll look through the repo for osa to have a glance at the successerator and see if i can do something similar ... i want to switch more of my customer facing real job work to ansible driven but a lot of that is done over vpn | 13:39 |
| idlemind | and it appears that sometimes network_cli fails when it shouldn't (immediately check via ping and manual ssh and the host is there) | 13:39 |
| odyssey4me | evrardjp well, the justification was that we had to remove a bunch of stuff from it and that it was too heavy | 13:39 |
| evrardjp | odyssey4me: for me, we should just say: Use that tar. | 13:39 |
| evrardjp | that tar is from a website | 13:40 |
| odyssey4me | evrardjp sure, but what tar -we have no stable source for it | 13:40 |
| *** TobbeCN has quit IRC | 13:40 | |
| evrardjp | I agree but it's the same story as everyone else | 13:40 |
| *** mwarad has quit IRC | 13:40 | |
| evrardjp | can't we suck it up from dockerhub? | 13:40 |
| odyssey4me | idlemind using retry + delay + until | success with a register? | 13:40 |
| evrardjp | https://hub.docker.com/r/opensuse/amd64/ ? | 13:42 |
| odyssey4me | evrardjp we already have implementations for using diskimage-builder inside trove and octavia, it seems sensible to me to unify those and the base prep | 13:42 |
| odyssey4me | and, at the same time, pull the base prep out of the {lxc,nspawn}_hosts roles | 13:42 |
| evrardjp | odyssey4me: don't get me wrong, I am not against it | 13:42 |
| evrardjp | I'd simply like us to say "Use that image" | 13:42 |
| evrardjp | that image can even be a group var | 13:43 |
| odyssey4me | that's already an option afaik | 13:43 |
| hwoarang | seems i opened the usual can of worms | 13:43 |
| evrardjp | hwoarang: hahahahah | 13:43 |
| evrardjp | hwoarang: I love it very much when you do it :) | 13:43 |
| hwoarang | you go fix one thing and ten more things pop up on the side | 13:44 |
| evrardjp | odyssey4me: I think we all agree | 13:44 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/pike: Fix repo_build_venv_rebuild toggle effectiveness https://review.openstack.org/562336 | 13:44 |
| evrardjp | we should spin that out of the role, the role should just consume images | 13:44 |
| odyssey4me | given that I have more familiarity with how dib works, perhaps I should just go ahead and do it - I've had enough of walking in this circle | 13:44 |
| evrardjp | the images can be built with DIB at the beggining of the process | 13:44 |
| evrardjp | odyssey4me: well | 13:44 |
| evrardjp | it's gonna be a problem with OOMing IIRC. | 13:44 |
| evrardjp | by adding that in the process of the deploy | 13:45 |
| evrardjp | and definitely lengthen the build | 13:45 |
| evrardjp | so we need a way to have a separate repo doing it | 13:45 |
| evrardjp | and the others consuming it | 13:45 |
| hwoarang | fyi we are doing that in OPNFV and dib is not cheap. you are looking at 10-15+ extra time for deployments | 13:45 |
| evrardjp | which is the cause of issues IIRC. | 13:45 |
| odyssey4me | I doubt it will cause an OOM, but yes it may lengthen the build a bit. | 13:45 |
| evrardjp | odyssey4me: that's how we got the first OOMing in octavia | 13:46 |
| evrardjp | IIRC | 13:46 |
| odyssey4me | well, we only need a very small base - no kernel, etc - bear that in mind | 13:46 |
| *** deng_ has quit IRC | 13:46 | |
| odyssey4me | this is not a dib for a whole host base like bifrost needs, or a VM base like octavia builds | 13:47 |
| odyssey4me | we could, for now, flip the default back to the 'legacy' lxc pull | 13:49 |
| evrardjp | odyssey4me: could you tell me how dib can leverage debootstrap for example? | 13:49 |
| evrardjp | Because for me the lightest image is debootstrap + dbus | 13:50 |
| *** dxiri has joined #openstack-ansible | 13:50 | |
| odyssey4me | evrardjp dib does that, but also caches and has a ton of optimisation it'd be a waste for us to repeat | 13:50 |
| evrardjp | or pacstrap for arch :) | 13:50 |
| odyssey4me | also dib uses the best system on each distro for the job | 13:50 |
| evrardjp | that's cool | 13:50 |
| evrardjp | that's very nice | 13:51 |
| evrardjp | it proves I really have 0 knowledge of how cool dib is | 13:51 |
| evrardjp | we have too many things going on | 13:51 |
| evrardjp | let's not do that right now | 13:51 |
| odyssey4me | debootstrap is debian/ubuntu only IIRC | 13:51 |
| evrardjp | odyssey4me: yes. | 13:51 |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: Only replace python when re-initializing the venv https://review.openstack.org/562179 | 13:51 |
| evrardjp | let's focus on python simplification it's a big chunk | 13:52 |
| *** dxiri has quit IRC | 13:52 | |
| evrardjp | and hwoarang mnaser focusing on packages | 13:52 |
| evrardjp | while I focus on the inventory | 13:52 |
| odyssey4me | sure, but we also need a stable source for these base images - otherwise the rocky release is at risk of being useless | 13:52 |
| hwoarang | -ESOMANYTHINGSTODO | 13:53 |
| evrardjp | :) | 13:53 |
| evrardjp | hwoarang: the plan is not complete, there is Tahvok cleaning up stuff too | 13:53 |
| evrardjp | and new features! | 13:53 |
| evrardjp | I am just worried we are stretched thin | 13:54 |
| evrardjp | but indeed making sure it's stable is coming first | 13:54 |
| evrardjp | but I think python/packages are postive improvements for stability | 13:54 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-lxc_hosts master: SUSE: Determine the build number on the fly https://review.openstack.org/562599 | 13:54 |
| hwoarang | evrardjp: odyssey4me what do you think^ | 13:55 |
| evrardjp | maybe I should focus more on switching images then | 13:55 |
| evrardjp | hwoarang: is that the same version number for all arches? | 13:55 |
| hwoarang | it's not the same but you retrieve that on the fly | 13:56 |
| hwoarang | parsing the appropriate Dockerfile | 13:56 |
| odyssey4me | seemse reasonable enough | 13:57 |
| odyssey4me | I guess another option is to use a base docker image from docker hub and get the image details by doing an api query and getting the response? I'm not sure how stable that interface is. | 13:57 |
| evrardjp | odyssey4me: there is probably a module for that | 13:58 |
| hwoarang | no clue... i still want to know what was wrong with images.linuxcontainers.org | 13:58 |
| hwoarang | :) | 13:58 |
| evrardjp | http://docs.ansible.com/ansible/2.5/modules/docker_image_facts_module.html#docker-image-facts-module | 13:58 |
| odyssey4me | the only trouble with this mechanism is that it reliaes on this two step process, instead of just downloading a tarball like it does now | 13:58 |
| evrardjp | http://docs.ansible.com/ansible/2.5/modules/docker_login_module.html#docker-login-module | 13:58 |
| hwoarang | oh that module needs a bunch of docker deps | 13:58 |
| evrardjp | fetching it would be http://docs.ansible.com/ansible/2.5/modules/docker_image_module.html#docker-image-module | 13:59 |
| evrardjp | yeah and we've never been using docker | 13:59 |
| evrardjp | so now we have all the API issues we've never seen before | 13:59 |
| hwoarang | use docker to configure lxc | 13:59 |
| hwoarang | oh the irony | 13:59 |
| evrardjp | the question is where is the irony... :) | 13:59 |
| *** niraj_singh has joined #openstack-ansible | 14:00 | |
| jrosser | has anyone deployed fwaas? i follow the steps in the os_neutron role docs and get "neutron.manager [req-1133af1e-a39e-4004-9ee2-026ed69516bc - - - - -] Plugin 'firewall' not found." | 14:00 |
| *** cmart has quit IRC | 14:00 | |
| hwoarang | i think that was recently fixed | 14:00 |
| hwoarang | mbuil: ^ ? | 14:00 |
| hwoarang | and it should already be backported in queens | 14:01 |
| evrardjp | maybe a bump was not yet done? Not so sure to have seen that in Q | 14:01 |
| evrardjp | or it might be long ago | 14:01 |
| odyssey4me | machinectl no longer does pull-dkr "because of the forever breaking api of docker" | 14:02 |
| odyssey4me | so yeah, no go on that idea | 14:02 |
| evrardjp | yeah I remember that PR :) | 14:03 |
| mbuil | hwoarang, jrosser: it was fixed last week | 14:05 |
| mbuil | for Queens --> https://review.openstack.org/#/c/560020/ | 14:05 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/queens: Allow public network name to be changed https://review.openstack.org/562237 | 14:06 |
| *** nicolasbock has quit IRC | 14:10 | |
| jrosser | mbuil: ah interesting, would you expect fwaas v1 to fail without that too? | 14:10 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/pike: Allow public network name to be changed https://review.openstack.org/562238 | 14:11 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tempest master: Avoid mixing pip and distribution packages. https://review.openstack.org/562606 | 14:12 |
| evrardjp | jrosser: v1 is supposed to be dead IIRC | 14:12 |
| evrardjp | or maybe I am confused? | 14:12 |
| jrosser | me too :) | 14:12 |
| evrardjp | confused I think | 14:12 |
| *** Adri2000 has quit IRC | 14:13 | |
| *** cmart has joined #openstack-ansible | 14:13 | |
| jrosser | i mean, it's a valid quesstion which i *should* be using | 14:13 |
| evrardjp | I think it's better to find info more upstream | 14:14 |
| jrosser | either way i can't make the install if either work, v2 seems to have a patch ^^^ but v1 i don't know why | 14:14 |
| * jrosser digs a bit more | 14:14 | |
| mbuil | evrardjp, jrosser: correct, fwaas v1 is not supported anymore | 14:15 |
| evrardjp | ok | 14:15 |
| evrardjp | that's what I thought. | 14:15 |
| evrardjp | cool | 14:15 |
| evrardjp | we should remove all references to v1 then | 14:15 |
| jrosser | ah right ok - configure_network_services.rst needs a tweek then | 14:15 |
| mbuil | the official statement is: "fwaas v1 is deprecated. it is still in the tree, but new usage is highly discouraged." | 14:16 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Re-order database migrations https://review.openstack.org/562024 | 14:17 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: [RFC] Add support for using distribution packages for OpenStack services https://review.openstack.org/560308 | 14:17 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: Avoid mixing pip and distribution packages https://review.openstack.org/562607 | 14:17 |
| hwoarang | evrardjp: odyssey4me do you have a list of stuff that needs revewing? i am done coding for the day | 14:17 |
| hwoarang | otherwise i can simply go through the osa dashboard | 14:17 |
| evrardjp | yup go through the dashbaord :) | 14:18 |
| hwoarang | ok | 14:18 |
| odyssey4me | hwoarang if we can get something merged to unblock suse tests for master then that'd be a good start... much of what merged today (or just got voted on) I need to port back to queens at least | 14:18 |
| hwoarang | odyssey4me: i proposed https://review.openstack.org/#/c/562599/ to unblock the gates | 14:19 |
| *** epalper has quit IRC | 14:26 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_cinder master: (WIP) Update to use oslo.messaging service for RPC and Notify https://review.openstack.org/560574 | 14:28 |
| *** armaan has quit IRC | 14:28 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_keystone master: (WIP) Update to use oslo.messaging service for RPC and Notify https://review.openstack.org/562026 | 14:33 |
| *** vnogin has joined #openstack-ansible | 14:34 | |
| *** vnogin has quit IRC | 14:35 | |
| evrardjp | ansmith: thanks for the continued effort! :) | 14:37 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/queens: Fix repo_build_venv_rebuild toggle effectiveness https://review.openstack.org/562335 | 14:39 |
| spotz | Hey don't forget my little patch! https://review.openstack.org/#/c/551316/ :) | 14:39 |
| ansmith | evrardjp: wanted to clean up CI a bit before asking for folks to take a look at the reviews | 14:40 |
| ansmith | evrardjp: probably getting close to doing a sanity check on the approach knowing there is still more to do | 14:41 |
| *** hamza21 has quit IRC | 14:44 | |
| tosky | evrardjp: hi again - we discussed about adding a cross-repo job today in the sahara meeting and it was agreed that it's good | 14:46 |
| evrardjp | great! | 14:46 |
| *** markvoelker has quit IRC | 14:47 | |
| evrardjp | odyssey4me and I have worked on the infra testing, so we can help you on the jobs | 14:47 |
| evrardjp | keystone was a first example | 14:47 |
| tosky | so I will work on this; I guess I would need to start with one job to not overload the infrastructure too much | 14:47 |
| tosky | I think I can manage the job part in general; I suspect I will ask more questions about the specific policies for your jobs | 14:47 |
| evrardjp | yes, just one job is fine: The deploy of sahara with the os_sahara role | 14:47 |
| tosky | sure, but there are many platforms :) | 14:48 |
| evrardjp | so latest sahara changes can be properly tested | 14:48 |
| evrardjp | oh sure :) | 14:48 |
| evrardjp | starting with ubuntu is the easiest, because that's what we've done already | 14:48 |
| evrardjp | but extending should be easy | 14:48 |
| *** markvoelker has joined #openstack-ansible | 14:48 | |
| *** SimAloo has joined #openstack-ansible | 14:49 | |
| tosky | I've also seen a centos 7 job, if I'm not mistaken | 14:49 |
| evrardjp | but I'd like we go for another test later: to add a full OSA deploy scenario with sahara | 14:49 |
| tosky | but let's see, changing that is easy | 14:49 |
| evrardjp | tosky: indeed, and SUSE | 14:49 |
| evrardjp | yeah, and ubuntu has been our history for a long while, so it's the most stable and the fastest. | 14:49 |
| tosky | oh, you didn't mean a full OSA deploy scenario? I though you mean a job which is equivalent to the devstack deployment | 14:49 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Fix osa_toolkit dictutils import https://review.openstack.org/560008 | 14:50 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Add nova/ironic/glance cross-service vars to group_vars/all https://review.openstack.org/560024 | 14:50 |
| evrardjp | you don't need the full things if you're just looking for a devstack equivalent | 14:50 |
| odyssey4me | evrardjp let's get it in to stop the bleeding: https://review.openstack.org/#/c/562599/1 | 14:50 |
| evrardjp | tosky: that's the good thing :) | 14:50 |
| evrardjp | odyssey4me: adding a bug | 14:50 |
| tosky | technically I'd like to add a full deployment + scenario tests with the fake plugin instead or in addition to the tempest tests | 14:50 |
| tosky | like the sahara-tests-scenario jobs defined in the sahara-tests repository | 14:51 |
| tosky | that's a better measurement that everything is fine | 14:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: Revert "Configure plugins and policies after RabbitMQ" https://review.openstack.org/562621 | 14:52 |
| evrardjp | odyssey4me: https://bugs.launchpad.net/openstack-ansible/+bug/1765432 | 14:52 |
| openstack | Launchpad bug 1765432 in openstack-ansible "Decide the source of the base images" [Undecided,New] | 14:52 |
| evrardjp | mnaser: if https://review.openstack.org/#/c/562621/1 merges before you got the chance to look at it, please be aware of the revert. | 14:53 |
| odyssey4me | tosky yeah, our ideal is some sort of canary test rather than a full function api test or something like that - we're not wanting to test that the API is conforming to the spec, we're wanting to verify that we have a functioning system | 14:53 |
| *** vnogin has joined #openstack-ansible | 14:54 | |
| mnaser | evrardjp: sorry :( | 14:55 |
| *** vnogin has quit IRC | 14:55 | |
| mnaser | evrardjp: would it be okay to share some of those breakages | 14:55 |
| mnaser | odyssey4me: ^^ | 14:55 |
| tosky | odyssey4me: sure, that's why I suggest a scenario test | 14:55 |
| tosky | with the fake plugin | 14:56 |
| tosky | that's our canary test, sadly, due to the interaction between sahara and any possible openstack project under the sun | 14:56 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Remove python3 job https://review.openstack.org/562666 | 14:58 |
| odyssey4me | tosky ah, if you have a fake plugin then you could shorten the test time quite a bit on that role - right now we deploy all the services! | 14:58 |
| evrardjp | andymccr: we should use the same topic | 14:59 |
| odyssey4me | mnaser sure, lemme register a bug | 14:59 |
| evrardjp | let's name it osa_python3? | 14:59 |
| tosky | odyssey4me: and we would need all the services | 14:59 |
| tosky | I think I'm not explaining enough how sahara need a full deployment if you want to avoid surprises | 14:59 |
| andymccr | evrardjp: ok cool | 15:00 |
| *** cjloader_ has joined #openstack-ansible | 15:01 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Remove python3 job https://review.openstack.org/562666 | 15:02 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests master: Add python3 functional voting and non-voting jobs https://review.openstack.org/562667 | 15:02 |
| evrardjp | andymccr: for your patch to pass, we need to remove the job completely first, so my patch should be included first. | 15:02 |
| andymccr | ok i can put a depends on | 15:02 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests master: Add python3 functional voting and non-voting jobs https://review.openstack.org/562667 | 15:03 |
| andymccr | that look better? | 15:03 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Use a sensible vlan range in the example configs https://review.openstack.org/560877 | 15:03 |
| evrardjp | andymccr: you mind if I edit it? | 15:04 |
| *** TobbeCN has joined #openstack-ansible | 15:04 | |
| evrardjp | not sure the depends on is a good idea for that | 15:04 |
| andymccr | go for it | 15:04 |
| evrardjp | but let's try | 15:04 |
| andymccr | i'll check the diff so i know what i didnt do :D | 15:05 |
| evrardjp | yeah depends on doesn't seem to work | 15:05 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: Add python3 functional voting and non-voting jobs https://review.openstack.org/562667 | 15:05 |
| evrardjp | you did good | 15:05 |
| evrardjp | just simplifying | 15:05 |
| andymccr | ahh i see | 15:05 |
| andymccr | cool | 15:05 |
| evrardjp | (doing like the other jobs) | 15:05 |
| andymccr | good tip | 15:05 |
| evrardjp | so now you can use them in the roles, the way you want | 15:06 |
| hwoarang | evrardjp: odyssey4me sadly the opensuse tarballs are going away https://github.com/openSUSE/docker-containers-build/issues/29 | 15:07 |
| odyssey4me | mnaser https://bugs.launchpad.net/openstack-ansible/+bug/1765438 | 15:07 |
| openstack | Launchpad bug 1765438 in openstack-ansible "RabbitMQ policy application not working for AIO" [Undecided,New] | 15:07 |
| hwoarang | so we must come with an alternative soonish... | 15:07 |
| hwoarang | possibly go back to images.linuxcontainers.org | 15:08 |
| *** kstev has joined #openstack-ansible | 15:08 | |
| odyssey4me | yeah | 15:08 |
| *** TobbeCN has quit IRC | 15:08 | |
| odyssey4me | or perhaps the docker api has stabilised enough now that we can patch machinectl ;) | 15:09 |
| evrardjp | lol | 15:09 |
| evrardjp | odyssey4me: ask hwoarang how easy is it to include a patch into systemd | 15:09 |
| evrardjp | :p | 15:09 |
| hwoarang | :( | 15:10 |
| evrardjp | sorry I had to :( | 15:10 |
| *** vnogin has joined #openstack-ansible | 15:10 | |
| evrardjp | did you see my bug? | 15:10 |
| evrardjp | Could you add this behavior change into the bug? | 15:10 |
| *** zerick has joined #openstack-ansible | 15:10 | |
| evrardjp | https://bugs.launchpad.net/openstack-ansible/+bug/1765432 | 15:10 |
| openstack | Launchpad bug 1765432 in openstack-ansible "Decide the source of the base images" [Undecided,New] | 15:10 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: SUSE: Determine the build number on the fly https://review.openstack.org/562599 | 15:10 |
| evrardjp | hwoarang: it would be nice to vote on https://review.openstack.org/#/c/562621/1 | 15:11 |
| *** DanyC has quit IRC | 15:11 | |
| *** DanyC has joined #openstack-ansible | 15:12 | |
| evrardjp | FYI, I've asked for the inclusion of masakari into the openstack ansible namespace, and I will work with niraj_singh on making this happen. Thanks niraj_singh ! | 15:12 |
| * odyssey4me has no idea what masakari is, but thanks niraj_singh :) | 15:13 | |
| evrardjp | hwoarang: woot it's not a hard no | 15:15 |
| evrardjp | if we have official light images from openSUSE team, that would be very nice. | 15:15 |
| *** DanyC has quit IRC | 15:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_panko master: Replace virtualenv-tools with a script https://review.openstack.org/562190 | 15:17 |
| evrardjp | we got a few patches flowing it's very cool | 15:17 |
| *** mbuil has quit IRC | 15:17 | |
| odyssey4me | yeah, thanks for putting that workaround together so quickly hwoarang | 15:17 |
| evrardjp | odyssey4me: thanks for translating my words, I realize it was not really clear, once again. | 15:18 |
| evrardjp | :D | 15:19 |
| evrardjp | good teaming! | 15:19 |
| *** mma has joined #openstack-ansible | 15:22 | |
| evrardjp | hwoarang: to what I see https://github.com/openSUSE/docker-containers-build/issues/29 , we can still use it for a certain time | 15:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562673 | 15:22 |
| evrardjp | but that's indeed a ticking time bomb | 15:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562674 | 15:23 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562675 | 15:23 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_searchlight stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562676 | 15:23 |
| *** vnogin has quit IRC | 15:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_monasca-agent stable/queens: Replace virtualenv-tools with a script https://review.openstack.org/562677 | 15:24 |
| xdfil | Hi, folks. | 15:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic stable/queens: Do not log passwords https://review.openstack.org/562678 | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic stable/pike: Do not log passwords https://review.openstack.org/562679 | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/queens: Do not log passwords https://review.openstack.org/562680 | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara stable/queens: Do not log passwords https://review.openstack.org/562681 | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara stable/pike: Do not log passwords https://review.openstack.org/562682 | 15:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum stable/queens: Do not log passwords https://review.openstack.org/562683 | 15:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum stable/pike: Do not log passwords https://review.openstack.org/562684 | 15:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron stable/queens: Do not log passwords https://review.openstack.org/562685 | 15:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron stable/pike: Do not log passwords https://review.openstack.org/562686 | 15:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/queens: Do not log passwords https://review.openstack.org/562688 | 15:29 |
| *** vnogin has joined #openstack-ansible | 15:29 | |
| evrardjp | hey xdfil | 15:29 |
| xdfil | say i'm using 17.0.3, where the default group_var setting for neutron_agents, is_metal: true | 15:29 |
| xdfil | But, when the play runs it makes a neutron_agents container still. | 15:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/queens: Do not log passwords https://review.openstack.org/562689 | 15:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/pike: Do not log passwords https://review.openstack.org/562690 | 15:29 |
| evrardjp | xdfil: check the inventory maybe? | 15:29 |
| xdfil | I have to copy the env.d to my /etc/openstack_deploy for it to build as metal | 15:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_panko stable/queens: Do not log passwords https://review.openstack.org/562691 | 15:29 |
| evrardjp | xdfil: and you have an env.d already about that group? | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_panko stable/pike: Do not log passwords https://review.openstack.org/562693 | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican stable/queens: Do not log passwords https://review.openstack.org/562694 | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican stable/pike: Do not log passwords https://review.openstack.org/562695 | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate stable/queens: Do not log passwords https://review.openstack.org/562696 | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate stable/pike: Do not log passwords https://review.openstack.org/562697 | 15:30 |
| evrardjp | env.d would have precedence to group var as being host var | 15:30 |
| xdfil | so it's "fixed" but I'm seeing a pattern emerging that I am bad at predicting which variables take precedence | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/queens: Do not log passwords https://review.openstack.org/562699 | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/pike: Do not log passwords https://review.openstack.org/562700 | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi stable/queens: Do not log passwords https://review.openstack.org/562701 | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi stable/pike: Do not log passwords https://review.openstack.org/562702 | 15:31 |
| xdfil | no I didn't have a neutron.yml in my /etc | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/queens: Do not log passwords https://review.openstack.org/562703 | 15:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/pike: Do not log passwords https://review.openstack.org/562704 | 15:31 |
| xdfil | I had a cinder.yml that's all | 15:32 |
| pabelanger | morning, has OSA started testing on ansible 2.5? | 15:32 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer stable/queens: Do not log passwords https://review.openstack.org/562705 | 15:32 |
| evrardjp | pabelanger: yes | 15:32 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer stable/pike: Do not log passwords https://review.openstack.org/562706 | 15:32 |
| evrardjp | pabelanger: we have a few bugs pending | 15:32 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_searchlight stable/queens: Do not log passwords https://review.openstack.org/562707 | 15:32 |
| evrardjp | we want to introduce a job for devel in the future | 15:32 |
| pabelanger | evrardjp: ansible bugs or issues with playbooks? | 15:32 |
| evrardjp | we have, but we want to make that more interesting | 15:32 |
| evrardjp | right now it's issues with plugins | 15:32 |
| pabelanger | do you mind sharing the location of your bugs? We're discussing updating zuul directly to 2.5 and skipping 2.4 | 15:34 |
| pabelanger | currently 2.3 | 15:34 |
| evrardjp | wow | 15:34 |
| evrardjp | that's a big change | 15:34 |
| pabelanger | right, trying to see if we should step to 2.4 first | 15:34 |
| pabelanger | then into 2.5 | 15:34 |
| evrardjp | well if you can afford the 2.5 move go for it | 15:34 |
| evrardjp | 2.5.1 is just out | 15:35 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562709 | 15:35 |
| pabelanger | fear is breakage of all jobs in openstack obviously, so would need to see the breaking changes from 2.3 to 2.5 | 15:35 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562710 | 15:35 |
| evrardjp | pabelanger: jmccrory: is working on those 2.5 bump | 15:35 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562711 | 15:35 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_molteniron stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562717 | 15:36 |
| evrardjp | we have this one: https://review.openstack.org/#/c/550665/ | 15:36 |
| evrardjp | pabelanger: ^ | 15:36 |
| evrardjp | check also the check experimental there | 15:36 |
| evrardjp | for 2.4 we have a few changes too. | 15:37 |
| evrardjp | I can't remember. Let me check if I put all of that into a topic | 15:37 |
| pabelanger | k, thanks | 15:38 |
| odyssey4me | apologies for that patch storm - that's most of the merged changes relating to the password linting and the py3 issues in venvs done | 15:39 |
| odyssey4me | there'll be some more tomorrow once the outstanding ones merge, and I got some merge conflicts when trying to port some back - so I'll have to resolve those manually | 15:39 |
| odyssey4me | assuming those all pass tests, I'd appreciate it if we could get those out of the queue asap | 15:39 |
| evrardjp | pabelanger: https://review.openstack.org/#/q/topic:ansible-2.4 for 2.4 | 15:39 |
| *** dxiri has joined #openstack-ansible | 15:40 | |
| pabelanger | evrardjp: thanks! looking | 15:40 |
| evrardjp | odyssey4me: will review them | 15:40 |
| evrardjp | pabelanger: there are probably other patches, so let's just say we'll do better this time, and mark everything as ansible_update. | 15:41 |
| evrardjp | ansible-update | 15:41 |
| pabelanger | wfm | 15:41 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_panko stable/queens: Replace virtualenv-tools with a script https://review.openstack.org/562720 | 15:42 |
| evrardjp | oh sorry we already had this one merged too: https://review.openstack.org/#/c/543576/ | 15:42 |
| *** armaan has joined #openstack-ansible | 15:42 | |
| evrardjp | it will stay ansible2.5 I guess | 15:42 |
| evrardjp | else we'll lose track | 15:42 |
| evrardjp | good luck anyway | 15:43 |
| evrardjp | if you need help, just shoot | 15:43 |
| evrardjp | :) | 15:43 |
| *** mma has quit IRC | 15:43 | |
| odyssey4me | alright everyone - I've had a long day so I'm going to go off early - have a good day/evening! | 15:44 |
| evrardjp | odyssey4me: really thank you for the work! | 15:45 |
| evrardjp | tty tomorrow! | 15:45 |
| *** cshen has quit IRC | 15:45 | |
| *** armaan has quit IRC | 15:47 | |
| spotz | seeya odyssey4me | 15:48 |
| *** armaan has joined #openstack-ansible | 15:48 | |
| xdfil | what happens if I don't specify a repo_host? | 15:49 |
| xdfil | should things still work? | 15:49 |
| *** RandomTech has joined #openstack-ansible | 15:50 | |
| RandomTech | Hello everyone, does anyone happen to know were osa stores its playbooks run log? | 15:52 |
| jrosser | RandomTech: /openstack/log/ansible-logging/ansible.log on your deploy host | 15:53 |
| RandomTech | thank you | 15:54 |
| RandomTech | seems my setup openstack failed hm | 15:56 |
| RandomTech | ill try rerunning it | 15:56 |
| evrardjp | xdfil: good question. It's not planned for now. but I'd be happy if someone could work on that. | 15:59 |
| xdfil | So then... at this moment if I don't define one I shouldn't expect things to work? | 15:59 |
| xdfil | evrardjp: to your point though that does seem like something I'd work on. the repo container and I have never gotten along that well | 16:01 |
| *** armaan has quit IRC | 16:02 | |
| mmercer | do any of the devs actually use centos 7 or is everyone running ubuntu for the aio and whatnot ? | 16:02 |
| *** nicolasbock has joined #openstack-ansible | 16:02 | |
| evrardjp | xdfil: odyssey4me is working on the simplification there. But our role tests are basically not using the repo server, so it's possible to do it. It will just be very slow. | 16:02 |
| *** pcaruana has quit IRC | 16:03 | |
| evrardjp | (because you remove the caching layers) | 16:03 |
| xdfil | I hate all the caching hehe | 16:04 |
| evrardjp | mmercer: most of the ppl are using ubuntu | 16:04 |
| mmercer | thats what i figured. | 16:04 |
| mmercer | guess ill do a ubuntu install, maybe ill finally be able to get this darned thing working | 16:04 |
| evrardjp | then less people, but still some ppl, on suse and centos | 16:04 |
| evrardjp | mmercer: centos has the less love | 16:04 |
| xdfil | caching in regards to OSA has wasted far more of my time than it has saved :) | 16:04 |
| evrardjp | and is the slowest. | 16:04 |
| evrardjp | mmercer: darn thing? | 16:05 |
| evrardjp | it should work fine | 16:05 |
| evrardjp | ofc master is a moving target | 16:05 |
| evrardjp | stable/ should work | 16:05 |
| mmercer | evrardjp: ive done 5 bare metal installs following the quickstart aio guide, and it *does not* get past the pip part (tag 17.0.2) | 16:05 |
| mmercer | specifically on setup-infrastructure.yml | 16:06 |
| evrardjp | I guess the issue you have is with pip10? | 16:06 |
| mmercer | mmm, on centos 7, were running pip 7 or 8 | 16:06 |
| evrardjp | mmercer: so if you start clean, and then post a log of the issue, we can help | 16:06 |
| mmercer | it gets past modern pip in some of the cases, but then after modern pip, it fails to reach the repo container | 16:06 |
| evrardjp | network issue then | 16:07 |
| evrardjp | ? | 16:07 |
| evrardjp | please paste a full log it's easier | 16:07 |
| evrardjp | this way we can track things | 16:07 |
| mmercer | will go dig it out in a few minutes | 16:07 |
| evrardjp | cool | 16:07 |
| evrardjp | hwoarang: a few last things I should do before I leave for today? | 16:08 |
| *** electrofelix has quit IRC | 16:08 | |
| *** radeks has quit IRC | 16:10 | |
| odyssey4me | xdfil the package caching you can disable, if you have a pre-built repo you can use it too - otherwise you'll need the repo | 16:11 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: Do not log passwords https://review.openstack.org/562733 | 16:11 |
| hwoarang | evrardjp: not really | 16:11 |
| evrardjp | hwoarang: cool | 16:11 |
| evrardjp | odyssey4me: could you quickly vote on https://review.openstack.org/#/c/562733/ ? | 16:11 |
| evrardjp | it's in the chain of patches for unbreaking translations | 16:11 |
| evrardjp | same hwoarang :) | 16:12 |
| evrardjp | thank you! | 16:12 |
| evrardjp | I am glad I broke things so well. | 16:12 |
| evrardjp | with the do not log passwords. | 16:12 |
| evrardjp | it was elegant :) | 16:12 |
| odyssey4me | evrardjp :) now we try to merge all the backports of those patches before we merge the lint again | 16:13 |
| odyssey4me | maybe best to mark those lint changes -w for now | 16:14 |
| *** vnogin has quit IRC | 16:14 | |
| evrardjp | odyssey4me: fair :) | 16:14 |
| evrardjp | It proves the usefulness of making it a lint test:D | 16:14 |
| odyssey4me | very useful, and very good - great way to maintain a standard | 16:15 |
| odyssey4me | we should do more of that | 16:15 |
| evrardjp | If we do much more of that, we should refactor the job runner | 16:15 |
| evrardjp | but that's fine for now | 16:15 |
| evrardjp | basically we have one runner per test, which is expensive | 16:15 |
| *** cshen has joined #openstack-ansible | 16:24 | |
| *** yifei has quit IRC | 16:25 | |
| *** yifei has joined #openstack-ansible | 16:26 | |
| *** chyka has joined #openstack-ansible | 16:27 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: Avoid mixing pip and distribution packages https://review.openstack.org/562607 | 16:30 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: [RFC] Add support for using distribution packages for OpenStack services https://review.openstack.org/560308 | 16:30 |
| openstackgerrit | Merged openstack/openstack-ansible-plugins master: Remove python3 job https://review.openstack.org/562666 | 16:45 |
| RandomTech | hey when running the setup-openstack playbook i encountered this error: http://paste.openstack.org/show/719574/ | 16:46 |
| RandomTech | do you think i should rerun the first two playbooks? | 16:47 |
| *** jiaopengju has quit IRC | 16:47 | |
| *** cshen has quit IRC | 16:47 | |
| *** Leo_m has joined #openstack-ansible | 16:49 | |
| RandomTech | wait i think something really weird happened with the lxc containers its saying this container is on the wrong host possibly | 16:52 |
| RandomTech | evrardjp: if you have a moment does that error message seem odd to you as well? | 16:53 |
| *** Leo_m has quit IRC | 16:54 | |
| *** Leo_m has joined #openstack-ansible | 17:02 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: Set the pypiserver root to the correct directory https://review.openstack.org/562544 | 17:02 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible master: Adds the certificate client key password to user_secrets https://review.openstack.org/558244 | 17:09 |
| *** gkadam has quit IRC | 17:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server stable/queens: Set the pypiserver root to the correct directory https://review.openstack.org/562759 | 17:11 |
| openstackgerrit | Merged openstack/ansible-role-python_venv_build master: Only replace python when re-initializing the venv https://review.openstack.org/562033 | 17:13 |
| *** yifei1 has joined #openstack-ansible | 17:17 | |
| *** yifei has quit IRC | 17:18 | |
| *** yifei1 is now known as yifei | 17:18 | |
| *** armaan has joined #openstack-ansible | 17:24 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker master: Only replace python when re-initializing the venv https://review.openstack.org/562199 | 17:24 |
| *** udesale_ has quit IRC | 17:24 | |
| *** shardy has quit IRC | 17:35 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_rally master: Only replace python when re-initializing the venv https://review.openstack.org/562191 | 17:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate master: Only replace python when re-initializing the venv https://review.openstack.org/562168 | 17:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562766 | 17:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_rally stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562767 | 17:43 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562769 | 17:43 |
| *** hamza21 has joined #openstack-ansible | 17:44 | |
| *** chyka_ has joined #openstack-ansible | 17:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh master: Only replace python when re-initializing the venv https://review.openstack.org/562158 | 17:48 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: Only replace python when re-initializing the venv https://review.openstack.org/562171 | 17:49 |
| *** chyka has quit IRC | 17:50 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: Only replace python when re-initializing the venv https://review.openstack.org/562157 | 17:54 |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat master: Only replace python when re-initializing the venv https://review.openstack.org/562173 | 17:54 |
| *** armaan has quit IRC | 17:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Only replace python when re-initializing the venv https://review.openstack.org/562164 | 17:58 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara master: Only replace python when re-initializing the venv https://review.openstack.org/562195 | 18:01 |
| *** poopcat has joined #openstack-ansible | 18:06 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Only replace python when re-initializing the venv https://review.openstack.org/562161 | 18:12 |
| *** pabelanger has quit IRC | 18:17 | |
| *** pabelanger has joined #openstack-ansible | 18:17 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Only replace python when re-initializing the venv https://review.openstack.org/562197 | 18:23 |
| *** vakuznet has joined #openstack-ansible | 18:24 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Only replace python when re-initializing the venv https://review.openstack.org/562183 | 18:36 |
| *** TobbeCN has joined #openstack-ansible | 18:36 | |
| *** TobbeCN has quit IRC | 18:41 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: Avoid mixing pip and distribution packages https://review.openstack.org/562607 | 18:47 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: [RFC] Add support for using distribution packages for OpenStack services https://review.openstack.org/560308 | 18:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562791 | 18:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562792 | 18:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562794 | 18:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562795 | 18:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562796 | 18:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562797 | 18:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562798 | 18:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron stable/queens: Only replace python when re-initializing the venv https://review.openstack.org/562799 | 18:52 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tempest master: Avoid mixing pip and distribution packages. https://review.openstack.org/562606 | 19:00 |
| *** DanyC has joined #openstack-ansible | 19:00 | |
| *** DanyC has quit IRC | 19:05 | |
| *** sar has joined #openstack-ansible | 19:23 | |
| *** foutatoro has quit IRC | 19:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Do not log passwords https://review.openstack.org/562733 | 19:47 |
| *** guilhermesp has joined #openstack-ansible | 19:51 | |
| guilhermesp | hey there! Does anyone knows the best way to apply a cherry pick in a openstack-ansible production environment? In my case, is just a bugfix in horizon https://review.openstack.org/#/c/510188/6 | 19:52 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tempest master: Avoid mixing pip and distribution packages. https://review.openstack.org/562606 | 19:57 |
| *** hamza21 has quit IRC | 20:01 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: Avoid mixing pip and distribution packages https://review.openstack.org/562607 | 20:04 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_keystone master: [RFC] Add support for using distribution packages for OpenStack services https://review.openstack.org/560308 | 20:04 |
| jrosser | guilhermesp: i think you would make your own git repo, cherry pick the patch onto it, and then override these vars to suit https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml#L95-L98 | 20:04 |
| *** RandomTech has quit IRC | 20:04 | |
| *** Adri2000 has joined #openstack-ansible | 20:06 | |
| guilhermesp | jrosser: i see.. is it possible to only change the commit in "horizon_git_install_branch" with the commit id (here's the commit id https://git.openstack.org/cgit/openstack/horizon/commit/?h=stable/pike&id=780961ac1c470d3997367e6c7d335b460e162286) and then run the os-horizon-install.yml OR are there hidden steps that I'm not understanding clearly? | 20:11 |
| *** tosky has quit IRC | 20:15 | |
| *** tosky has joined #openstack-ansible | 20:16 | |
| pabelanger | evrardjp: good news, ansible 2.5.1 just worked with my playbooks / roles. That's the first time I've upgrades without having to change anything | 20:17 |
| pabelanger | ofcourse I get some deprecated warnings now, but I can fix them | 20:17 |
| spotz | pabelanger: Always good news | 20:36 |
| pabelanger | yah, a first | 20:38 |
| *** ansmith has quit IRC | 20:42 | |
| *** DanyC has joined #openstack-ansible | 20:48 | |
| *** sar has quit IRC | 20:49 | |
| *** DanyC_ has joined #openstack-ansible | 20:50 | |
| *** DanyC has quit IRC | 20:53 | |
| *** jwitko has quit IRC | 21:02 | |
| *** armaan has joined #openstack-ansible | 21:14 | |
| *** vnogin has joined #openstack-ansible | 21:15 | |
| *** kstev has quit IRC | 21:19 | |
| *** vnogin has quit IRC | 21:19 | |
| *** vakuznet has quit IRC | 21:20 | |
| *** gyee has joined #openstack-ansible | 21:22 | |
| gyee | lbragstad, on the fernet token rotation feature, by looking at the code, looks like the rotation won't happen if the first node is not online for whatever reason. That right? | 21:24 |
| lbragstad | i was under the impression that it would | 21:24 |
| gyee | how? as the cron job is schedule on the first node only | 21:25 |
| * lbragstad goes to dig | 21:25 | |
| lbragstad | gyee: oh - weird | 21:27 |
| lbragstad | https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tasks/keystone_fernet_keys_autorotate.yml#L40-L47 | 21:27 |
| lbragstad | maybe i imagined that | 21:28 |
| lbragstad | gyee: yeah - you're right | 21:28 |
| gyee | that make sure the cron job is on the first node | 21:28 |
| gyee | in a HA cluster, if the first node goes down, we will miss a rotation or two | 21:28 |
| lbragstad | sure | 21:29 |
| lbragstad | do you have monitoring on your hosts? | 21:29 |
| gyee | yes we do, maybe not that big of an issue | 21:29 |
| gyee | anyway, just making sure I am not reading the code wrong :-) | 21:30 |
| *** ansmith has joined #openstack-ansible | 21:30 | |
| lbragstad | i guess if you can detect the first host goes down you could drop the auto rotation script on another host or resolve it later | 21:30 |
| *** esberglu has quit IRC | 21:30 | |
| lbragstad | gyee: yeah - no, you're right. sorry for the bum direction | 21:30 |
| gyee | yeah, I have some logic to do these kind of dance for galera | 21:30 |
| gyee | maybe we can do the same for fernet? it's much easier for fernet as the files are numbers | 21:31 |
| gyee | we can find the one with the highest name and do the sync | 21:32 |
| gyee | highest number | 21:32 |
| lbragstad | well - all the keys should be consistent across the cluster, right? | 21:32 |
| lbragstad | so you could just put the auto rotation script on any host that is !down | 21:32 |
| lbragstad | or the first host that passes a health check? | 21:33 |
| gyee | I am worried about split brain situation | 21:34 |
| gyee | say node #1 is down, while node #2 and #3 had rotated | 21:34 |
| lbragstad | oh - if the first node comes back up and assume responsibility for the key repository? | 21:34 |
| gyee | right | 21:34 |
| gyee | worst case is that the new key got sync'ed to node 2 and 3 again, which may invalid tokens issued by those nodes | 21:35 |
| gyee | same file name, different key | 21:36 |
| lbragstad | yeah - maybe the process of recovering the node removes the autorotation bit from node 1 | 21:36 |
| lbragstad | and pull a fresh copy of the keys from 2 or 3 right before tying it back into the load balancer? | 21:36 |
| gyee | right | 21:36 |
| gyee | let me play around with it and see | 21:37 |
| lbragstad | cool - let me know what you find out | 21:37 |
| *** DanyC_ has quit IRC | 21:50 | |
| *** tomtom001 has quit IRC | 21:52 | |
| *** tomtom001 has joined #openstack-ansible | 21:52 | |
| evrardjp | guilhermesp: https://docs.openstack.org/openstack-ansible/latest/user/source-overrides/index.html | 21:53 |
| evrardjp | particularily https://docs.openstack.org/openstack-ansible/latest/user/source-overrides/index.html#overriding-other-upstream-projects-source-code | 21:53 |
| xdfil | pip_configure: false | 21:53 |
| evrardjp | it's in the docs... | 21:53 |
| evrardjp | pabelanger: generally following the migration guide is enough :) Except if you have weird fancy things :p | 21:54 |
| evrardjp | glad it was fine | 21:54 |
| evrardjp | has anyone tried a master build recently? | 21:55 |
| xdfil | I can't imagine the total amount of time I have spent chasing my tail through defaults/vars/group_vars trying to stop things from happening with pip | 21:55 |
| xdfil | I just found this lol | 21:55 |
| xdfil | pip_configure: false | 21:55 |
| xdfil | say it with me now kids! | 21:55 |
| openstackgerrit | Paul Belanger proposed openstack/ansible-hardening master: Switch from debian-jessie to debian-stable (stretch) https://review.openstack.org/562842 | 21:55 |
| openstackgerrit | Paul Belanger proposed openstack/ansible-hardening master: Switch from debian-jessie to debian-stable (stretch) https://review.openstack.org/562842 | 22:12 |
| *** cjloader_ has quit IRC | 22:20 | |
| openstackgerrit | Jean-Philippe Evrard (evrardjp) proposed openstack/openstack-ansible-tests stable/queens: Do not log passwords https://review.openstack.org/562853 | 22:22 |
| openstackgerrit | Jean-Philippe Evrard (evrardjp) proposed openstack/openstack-ansible-tests stable/pike: Do not log passwords https://review.openstack.org/562854 | 22:22 |
| openstackgerrit | Jean-Philippe Evrard (evrardjp) proposed openstack/openstack-ansible-plugins stable/queens: Remove python3 job https://review.openstack.org/562855 | 22:23 |
| openstackgerrit | Jean-Philippe Evrard (evrardjp) proposed openstack/ansible-config_template master: Temporarily disable python3 job https://review.openstack.org/562860 | 22:33 |
| *** chyka has joined #openstack-ansible | 22:34 | |
| *** chyka_ has quit IRC | 22:37 | |
| *** TobbeCN has joined #openstack-ansible | 22:37 | |
| *** TobbeCN has quit IRC | 22:41 | |
| *** armaan has quit IRC | 22:50 | |
| *** armaan has joined #openstack-ansible | 22:50 | |
| *** tosky has quit IRC | 23:10 | |
| *** chyka_ has joined #openstack-ansible | 23:15 | |
| openstackgerrit | Merged openstack/ansible-config_template master: Temporarily disable python3 job https://review.openstack.org/562860 | 23:17 |
| *** chyka has quit IRC | 23:19 | |
| *** chyka_ has quit IRC | 23:20 | |
| *** cmart has quit IRC | 23:21 | |
| *** chyka has joined #openstack-ansible | 23:41 | |
| *** chyka has quit IRC | 23:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!