| *** sdake has joined #openstack-ansible | 00:00 | |
| *** admin0 has quit IRC | 00:00 | |
| *** skamithi13 has quit IRC | 00:18 | |
| *** skamithi14 has joined #openstack-ansible | 00:18 | |
| *** markvoelker has joined #openstack-ansible | 00:43 | |
| *** sdake has quit IRC | 00:45 | |
| *** markvoelker has quit IRC | 00:47 | |
| *** sdake has joined #openstack-ansible | 00:50 | |
| *** sdake has quit IRC | 01:06 | |
| *** sdake has joined #openstack-ansible | 01:08 | |
| *** ShannonM has quit IRC | 01:23 | |
| *** sdake has quit IRC | 01:43 | |
| *** markvoelker has joined #openstack-ansible | 01:44 | |
| *** markvoelker has quit IRC | 01:48 | |
| *** mariongarly_ has quit IRC | 02:09 | |
| *** mariognarly has joined #openstack-ansible | 02:10 | |
| *** sdake has joined #openstack-ansible | 02:31 | |
| *** johnmilton has quit IRC | 02:31 | |
| *** markvoelker has joined #openstack-ansible | 02:44 | |
| *** markvoelker has quit IRC | 02:49 | |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server: Adding ERLANG VM tuning parameters https://review.openstack.org/284899 | 03:12 |
|---|---|---|
| *** sdake has quit IRC | 03:19 | |
| *** gfa is now known as gfa_ | 03:28 | |
| *** sdake has joined #openstack-ansible | 03:47 | |
| *** mc_dachoppa has joined #openstack-ansible | 04:00 | |
| *** shausy has joined #openstack-ansible | 04:11 | |
| *** mc_dachoppa has left #openstack-ansible | 04:32 | |
| *** markvoelker has joined #openstack-ansible | 04:45 | |
| *** markvoelker has quit IRC | 04:50 | |
| *** kstepniewski has joined #openstack-ansible | 04:52 | |
| *** javeriak has joined #openstack-ansible | 05:06 | |
| *** javeriak_ has joined #openstack-ansible | 05:12 | |
| *** javeriak has quit IRC | 05:12 | |
| *** cvstealth has quit IRC | 05:15 | |
| *** zhangjn has quit IRC | 05:38 | |
| *** zhangjn has joined #openstack-ansible | 05:42 | |
| *** fawadkhaliq has joined #openstack-ansible | 05:54 | |
| *** javeriak_ has quit IRC | 05:58 | |
| *** Mudpuppy has quit IRC | 06:01 | |
| *** javeriak has joined #openstack-ansible | 06:15 | |
| *** weezS has joined #openstack-ansible | 06:17 | |
| *** rgogunskiy has joined #openstack-ansible | 06:24 | |
| *** phiche has joined #openstack-ansible | 06:33 | |
| *** phiche has quit IRC | 06:36 | |
| *** phiche has joined #openstack-ansible | 06:40 | |
| *** sdake has quit IRC | 06:43 | |
| *** markvoelker has joined #openstack-ansible | 06:46 | |
| *** zhangjn has quit IRC | 06:47 | |
| *** zhangjn has joined #openstack-ansible | 06:49 | |
| *** weezS has quit IRC | 06:49 | |
| *** markvoelker has quit IRC | 06:50 | |
| *** phiche has quit IRC | 06:54 | |
| *** phiche1 has joined #openstack-ansible | 06:54 | |
| *** cvstealth has joined #openstack-ansible | 06:59 | |
| *** sdake has joined #openstack-ansible | 06:59 | |
| *** fawadk has joined #openstack-ansible | 07:15 | |
| *** fawadkhaliq has quit IRC | 07:15 | |
| *** fawadkhaliq has joined #openstack-ansible | 07:15 | |
| *** sdake has quit IRC | 07:17 | |
| *** fawadk has quit IRC | 07:20 | |
| *** jiteka has joined #openstack-ansible | 07:25 | |
| *** phiche1 has quit IRC | 07:30 | |
| *** zhangjn has quit IRC | 07:33 | |
| *** zhangjn has joined #openstack-ansible | 07:39 | |
| *** pcaruana has joined #openstack-ansible | 07:43 | |
| *** fawadkhaliq has quit IRC | 07:43 | |
| *** TheIntern has quit IRC | 07:45 | |
| *** phiche has joined #openstack-ansible | 07:48 | |
| *** gus has quit IRC | 07:50 | |
| *** gus has joined #openstack-ansible | 07:52 | |
| *** zhangjn has quit IRC | 07:54 | |
| *** zhangjn has joined #openstack-ansible | 08:00 | |
| *** permalac_ has quit IRC | 08:00 | |
| *** permalac_ has joined #openstack-ansible | 08:00 | |
| *** permalac_ has quit IRC | 08:01 | |
| *** Mudpuppy has joined #openstack-ansible | 08:02 | |
| *** Mudpuppy has quit IRC | 08:07 | |
| *** phiche has quit IRC | 08:10 | |
| *** phiche has joined #openstack-ansible | 08:10 | |
| *** javeriak has quit IRC | 08:13 | |
| *** zhangjn has quit IRC | 08:18 | |
| *** fawadkhaliq has joined #openstack-ansible | 08:27 | |
| *** zhangjn has joined #openstack-ansible | 08:29 | |
| *** finchd has quit IRC | 08:29 | |
| *** fawadkhaliq has quit IRC | 08:38 | |
| *** admin0 has joined #openstack-ansible | 08:43 | |
| *** markvoelker has joined #openstack-ansible | 08:47 | |
| *** kstepniewski has quit IRC | 08:48 | |
| *** markvoelker has quit IRC | 08:51 | |
| *** finchd has joined #openstack-ansible | 08:58 | |
| *** javeriak has joined #openstack-ansible | 08:59 | |
| *** javeriak has quit IRC | 09:04 | |
| *** javeriak has joined #openstack-ansible | 09:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Adding release notes for erlang VM tuning parameters https://review.openstack.org/285506 | 09:10 |
| *** finchd has quit IRC | 09:10 | |
| *** evrardjp has joined #openstack-ansible | 09:12 | |
| evrardjp | good morning | 09:12 |
| evrardjp | odyssey4me: are you there? | 09:12 |
| evrardjp | about this: https://review.openstack.org/#/c/280233/ | 09:13 |
| evrardjp | I wanted to test the cache_timeout and see if everything works fine or not | 09:13 |
| odyssey4me | morning evrardjp | 09:13 |
| odyssey4me | and morning everyone else! | 09:13 |
| evrardjp | the behaviour is fine, if I put a really low value it will update the repos | 09:14 |
| evrardjp | but the issue now is to find the correct setting | 09:14 |
| *** finchd has joined #openstack-ansible | 09:14 | |
| evrardjp | because I had issues with some runs running with such a low value of 60 | 09:14 |
| evrardjp | when running on ovh | 09:14 |
| evrardjp | so my question | 09:14 |
| odyssey4me | was that because of the OVH mirror changing or something? | 09:15 |
| evrardjp | or outdated | 09:15 |
| evrardjp | the host thinks he doesn't need to do the apt-get update (because it just started), but he is already outdated | 09:16 |
| evrardjp | so what I thought | 09:16 |
| evrardjp | maybe having some task at the gate that does a first apt-get update | 09:16 |
| evrardjp | and then use the cache_timeout to a larger value | 09:16 |
| evrardjp | I can fix that easily | 09:17 |
| evrardjp | however, this task isn't really needed for something else than ovh | 09:17 |
| evrardjp | do you think I should add this anyway ? | 09:17 |
| odyssey4me | for gating we already have an apt-get update happening on the host in the bootstrapping phase | 09:17 |
| odyssey4me | ah, you mean in this role? | 09:18 |
| evrardjp | yup | 09:18 |
| *** asettle has joined #openstack-ansible | 09:19 | |
| evrardjp | editing this role (in test/ right?) | 09:19 |
| evrardjp | makeing sure the apt-get update is fine | 09:19 |
| evrardjp | or runs at all times | 09:19 |
| evrardjp | don't change anything for the rest | 09:19 |
| evrardjp | OR | 09:19 |
| evrardjp | having a specific task for ovh | 09:20 |
| evrardjp | (I don't like this) | 09:20 |
| odyssey4me | yeah, just add a pre-task into tests/test.yml | 09:20 |
| evrardjp | fine for me | 09:20 |
| odyssey4me | not for OVH specifically - in the role's test playbook we can just do it generally | 09:20 |
| evrardjp | yup, but the issue only triggered on ovh | 09:20 |
| evrardjp | that's why | 09:20 |
| odyssey4me | it doesn't matter - doing a single apt-get update won't hurt | 09:21 |
| odyssey4me | it is interesting though - if we can find what the problem is then perhaps we can better guard against it | 09:21 |
| odyssey4me | infra has switched the mirrors used since you added the patch in, so the problem might actually be gone by now | 09:22 |
| odyssey4me | now all of infra uses the same apt mirror - their own | 09:22 |
| *** permalac has joined #openstack-ansible | 09:23 | |
| *** permalac has quit IRC | 09:23 | |
| *** permalac has joined #openstack-ansible | 09:23 | |
| evrardjp | what do you mean? | 09:23 |
| evrardjp | each infra host is using its own apt-mirror? | 09:23 |
| evrardjp | it makes sense but it will show disparities (correct english term?) accross infra's, I guess? | 09:24 |
| *** shausy has quit IRC | 09:28 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-lxc_hosts: Only update apt cache if necessary https://review.openstack.org/280233 | 09:30 |
| *** shausy has joined #openstack-ansible | 09:30 | |
| evrardjp | let's see how it works on ovh | 09:30 |
| evrardjp | about this: https://review.openstack.org/#/c/279730/ | 09:37 |
| evrardjp | my bad, forget it | 09:38 |
| *** tiagogomes has joined #openstack-ansible | 09:45 | |
| *** woopstar has joined #openstack-ansible | 09:50 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Remove container_cidr check from cidr_networks https://review.openstack.org/277511 | 09:50 |
| woopstar | I've just launched a setup for the openstack-ansible and seems to be launched fine. But i was wondering, if you need DHCP on any of the networks created? I'm not able to launch cinder volumes, nor am i able to launch any instance though there is 2 compute nodes available with 128gb mem each, and 100gb disk each | 09:51 |
| *** sdake has joined #openstack-ansible | 09:51 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add a function to check accuracy of user config https://review.openstack.org/282067 | 09:52 |
| *** admin0 has quit IRC | 09:57 | |
| *** neilus has joined #openstack-ansible | 09:59 | |
| *** admin0 has joined #openstack-ansible | 10:03 | |
| *** neilus1 has joined #openstack-ansible | 10:05 | |
| *** sdake has quit IRC | 10:06 | |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible-pip_install: Test version of pip being installed https://review.openstack.org/285284 | 10:08 |
| *** neilus has quit IRC | 10:08 | |
| odyssey4me | woopstar openstack itself provides dhcp facilities inside the tenant networks - are you creating tenant networks and building instances inside those? | 10:18 |
| *** admin0 has quit IRC | 10:29 | |
| *** admin0 has joined #openstack-ansible | 10:31 | |
| woopstar | Yeah. But i think i found the issue. It seems network related to the neutron agent. Properly the "flat" network. The network description in the install guide is pretty "grey". :( | 10:41 |
| *** neilus1 has quit IRC | 10:43 | |
| *** sdake has joined #openstack-ansible | 10:43 | |
| admin0 | hey a;; | 10:43 |
| admin0 | all :D | 10:43 |
| admin0 | the installer assumes and installs everything on containers(controllers). In real life, operators would like to run neturon network nodes on metal with special network cards for checksum offloading, and then run mysql on special machines that have SSDs, big RAM and also on metal. So how to achieve neutron and mysql on metal? | 10:43 |
| woopstar | I don't think you would do that with MySQL | 10:44 |
| admin0 | got people who aleady have a mysql dba team and due to some iso/stuff, need mysql to be in those specific hardware :( | 10:45 |
| andymccr | you could do that, its not tested really though, but setting "is_metal: true" in the environment for those services. | 10:45 |
| woopstar | I worked quite alot with Percona on the galera project. And 3 vm with 8 cores and 32 memory it runs perfectly with 3000q/sec | 10:45 |
| woopstar | which would be sufficient for Openstack I think | 10:46 |
| admin0 | i am not saying there is a limit on how we do stuff now .. i am just saying that there is also a need ( that came up more than once) for clients requiring it done this way | 10:46 |
| andymccr | admin0: as an example we put nova-compute on metal: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/nova.yml#L68-L75 and you could adjust the galera.yml file in the same location to install on_metal. (results may vary though because it hasnt been tested as a path, so it would probably have a few bugs that would need fixing) | 10:47 |
| *** markvoelker has joined #openstack-ansible | 10:47 | |
| *** markvoelker has quit IRC | 10:52 | |
| odyssey4me | yeah, but then you also need to set the affinity to 0 on the infra group | 10:53 |
| odyssey4me | lemme gist an example up - this gets asked often enough | 10:53 |
| *** sdake has quit IRC | 10:53 | |
| *** neilus has joined #openstack-ansible | 11:00 | |
| admin0 | odyssey4me: awesome | 11:01 |
| odyssey4me | admin0 I think this should work: https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb | 11:01 |
| odyssey4me | andymccr ^ your thoughts? | 11:02 |
| admin0 | odyssey4me: so this line on galeray.yml …. contains: - galera . is what maps to the galeray on openstack_user_config.yml | 11:03 |
| *** gfa_ is now known as gfa | 11:03 | |
| admin0 | if i put contains: - superCOOLssd , then in user_config, it should also have the superCOOLssd: and the servers below that ? | 11:03 |
| odyssey4me | admin0 I'm not entirely sure whether that's the best group name to use, but I think it should work | 11:03 |
| odyssey4me | we really need to pick cloudnull's brain to understand how the groups all map together so that we can document it better | 11:04 |
| admin0 | odyssey4me: http://pastebin.com/L2MRhUnD — this is where i am confused when doing this mapping | 11:06 |
| *** skamithi14 has quit IRC | 11:06 | |
| *** sdake has joined #openstack-ansible | 11:06 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Docs: Ensure a horizontal rule prior to the navigation links https://review.openstack.org/285981 | 11:10 |
| odyssey4me | admin0 why bother trying to change the group name at all? | 11:10 |
| *** xek has joined #openstack-ansible | 11:12 | |
| *** furlongm has quit IRC | 11:14 | |
| *** sdake has quit IRC | 11:16 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Fixing keepalived bug when 2+ backup nodes have the same priority https://review.openstack.org/279730 | 11:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Dynamic inventory: correct the inline commentary https://review.openstack.org/282452 | 11:18 |
| odyssey4me | mattt if you could review https://review.openstack.org/#/q/status:open+topic:repeatable-build please it'd be great to get those done | 11:20 |
| openstackgerrit | Merged openstack/openstack-ansible: Add relnote for openstack_host_systat_ rename https://review.openstack.org/285237 | 11:22 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Fixing keepalived bug when 2+ backup nodes have the same priority https://review.openstack.org/279730 | 11:24 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Information about the keepalived bug when multiple backup nodes have the same priority. https://review.openstack.org/279664 | 11:30 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Information about the keepalived bug when multiple backup nodes have the same priority. https://review.openstack.org/279664 | 11:30 |
| admin0 | odyssey4me: that was just to learn how mapping works :D | 11:31 |
| evrardjp | odyssey4me: here are your reno notes: https://review.openstack.org/279730 and https://review.openstack.org/279664 | 11:31 |
| evrardjp | please tell me how to write them, maybe I was wrong about the usage of reno | 11:31 |
| admin0 | i have this setup .. internet => [firewall / pfsense or vyatta ] => openstack(ansible deployed, as usual) .. now the public IPs only stay on the firewall and do a NAT internally .. now how to put this? external_lb_vip_address .. should this be as the external IP ( that is going to be exposed ) or keep it internal ( but API requests might not be true ) .. | 11:41 |
| odyssey4me | admin0 the external_lb_address will become the public endpoint in the keystone service catalogue - so it can be a DNS name (I think) or the actual public address | 11:43 |
| admin0 | odyssey4me: i willt ry wtih the mysql metal and this (dns name ) and see how far i go | 11:44 |
| andymccr | odyssey4me, admin0: i think that looks good although im not sure you can specify a host_group called "galera"? you would need to add a physical skel "galera_hosts" to the galera conf. (delayed response sorry!) | 11:44 |
| odyssey4me | admin0 in previous deployments I implemented Apache reverse proxies in the DMZ for the API endpoints and did URL rewriting on the way through - but that was a design decision made in that specific case | 11:44 |
| admin0 | andymccr: all I want to know is .. if I have say a group called “my_crappy_hosts” and i can put them under openstack_user_config. , where in the env.d service.yml do I specify those | 11:44 |
| admin0 | so where in galera.yml ( which line ) says it needs to go to those | 11:45 |
| andymccr | admin0: thats the physical_skel section, you don't see it in the galera.yml but for other services it will be there. Keystone is probably the best example of this: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/keystone.yml | 11:45 |
| andymccr | so identity_hosts: in openstack_user_config.yml will be keystone hosts. keystone only | 11:46 |
| odyssey4me | andymccr lemme revise and try again :) | 11:46 |
| admin0 | andymccr: this is why I want to start with c0mpletely_oDD_hosts and _containers, because this example makes me think .. so do I just update the contains: part and replace keystone with new, or also replace belongs_to . and in where .. component, or container or physical | 11:47 |
| admin0 | so if there is a new example on how to specify a new group, and how to map services to this new group that would clarify | 11:47 |
| andymccr | admin0: there is an example file: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/extra_container.yml.example | 11:48 |
| odyssey4me | andymccr better? https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb | 11:50 |
| andymccr | it might not be that clear, in which case it would probably be useful to doc that a bit better :) so if you have any suggestions would be really appreciated! | 11:50 |
| andymccr | odyssey4me: just add "galera_containers" to the belongs_to: section of galera_container | 11:50 |
| andymccr | i kinda think we should have that for each service if im honest, because i may way want to setup a specific service only and not a whole group (this is an example of that) | 11:51 |
| andymccr | and a lot of service you can't do that without editing the env files | 11:51 |
| *** rgogunskiy has quit IRC | 11:52 | |
| odyssey4me | andymccr yeah, I'm kinda thinking that we should have service specific groups, and then these aggregated groups as an optional component | 11:52 |
| andymccr | odyssey4me: yeah i thinkso | 11:52 |
| andymccr | the optional ones are already there so thats easy | 11:52 |
| andymccr | and we can just add a specific host group per service too | 11:52 |
| andymccr | its just a bit more flexible then | 11:52 |
| andymccr | and wont change how things currently work | 11:52 |
| odyssey4me | andymccr typer typey? | 11:52 |
| odyssey4me | consistency would be great | 11:53 |
| andymccr | haha | 11:53 |
| andymccr | yes yes | 11:53 |
| *** javeriak has quit IRC | 11:53 | |
| odyssey4me | andymccr better now? https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb | 11:54 |
| andymccr | odyssey4me: that looks like the one | 11:54 |
| odyssey4me | I'm thinking though - if we remove https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb#file-galera-yml-L12-L13 then we wouldn't have to change affinity - but then what else would break? | 11:54 |
| admin0 | easier would be: galera_service , runs on galera_servers_group .. is_metal =true, or false .. so on galera_servers_group you just define ips .. if its not metal, inside containers, if metal then directly | 12:00 |
| admin0 | this is like multi layer mapping :D | 12:00 |
| admin0 | keystone_service , runs on keystone_servers, or any X servers .. and is_metal makes it define if its direct or on containers :D | 12:01 |
| admin0 | easier | 12:01 |
| odyssey4me | evrardjp are you still around? if so do you have the keepalived bug URL handy? | 12:01 |
| admin0 | odyssey4me: :D http://pastebin.com/wBbJBq1W | 12:03 |
| *** spotz_zzz is now known as spotz | 12:03 | |
| admin0 | so component belong to a container group which further belongs to a physical_skeleton group | 12:05 |
| -openstackstatus- NOTICE: Infra currently has a long backlog. Please be patient and where possible avoid rechecks while it catches up. | 12:05 | |
| admin0 | # This is a meta group of a given component type. — /me is lost | 12:05 |
| admin0 | :D | 12:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Information about the keepalived bug when multiple backup nodes have the same priority. https://review.openstack.org/279664 | 12:06 |
| odyssey4me | evrardjp is ^ my edit there ok? | 12:06 |
| *** rgogunskiy has joined #openstack-ansible | 12:10 | |
| *** rgogunskiy has quit IRC | 12:10 | |
| *** admin0 has quit IRC | 12:16 | |
| *** johnmilton has joined #openstack-ansible | 12:19 | |
| mattt | odyssey4me: sorry, was hacking on a ceph thing, i'll check those reviews in a moment | 12:26 |
| *** admin0 has joined #openstack-ansible | 12:32 | |
| tiagogomes | can't things like `hostvars[inventory_hostname]['container_networks']['container_address']['address'] is defined` be simplified to `container_networks']['container_address']['address'] is defined`? | 12:35 |
| tiagogomes | Sorry: `container_networks['container_address']['address'] is defined` | 12:35 |
| mattt | tiagogomes: not if you're referring to a host that isn't itself | 12:43 |
| tiagogomes | But by writing `hostvars[inventory_hostname]`, aren't you asserting that the host where the variables are get from is itself? | 12:44 |
| mattt | tiagogomes: ah, i see what you mean ... | 12:45 |
| *** shausy has quit IRC | 12:45 | |
| evrardjp | odyssey4me: received thanks | 12:48 |
| *** markvoelker has joined #openstack-ansible | 12:48 | |
| evrardjp | ok I get the gist | 12:48 |
| *** Mudpuppy has joined #openstack-ansible | 12:50 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-lxc_hosts: Only update apt cache if necessary https://review.openstack.org/280233 | 12:50 |
| *** markvoelker has quit IRC | 12:53 | |
| *** pcaruana has quit IRC | 12:53 | |
| *** Mudpuppy has quit IRC | 12:55 | |
| *** spotz is now known as spotz_zzz | 13:00 | |
| *** Oku_OS has joined #openstack-ansible | 13:05 | |
| *** woodard has joined #openstack-ansible | 13:06 | |
| *** woodard has quit IRC | 13:06 | |
| *** woodard has joined #openstack-ansible | 13:07 | |
| *** pcaruana has joined #openstack-ansible | 13:08 | |
| *** evrardjp has quit IRC | 13:12 | |
| *** nwonknu has quit IRC | 13:13 | |
| *** retreved has joined #openstack-ansible | 13:15 | |
| *** evrardjp has joined #openstack-ansible | 13:16 | |
| *** nwonknu has joined #openstack-ansible | 13:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server: Use slurp to collect the nova ssh keys https://review.openstack.org/286042 | 13:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server: Use slurp to collect the repo_server ssh keys https://review.openstack.org/286042 | 13:30 |
| *** mgariepy_ is now known as mgariepy | 13:31 | |
| *** mgariepy has quit IRC | 13:31 | |
| *** mgariepy has joined #openstack-ansible | 13:31 | |
| odyssey4me | mattt ^ that should fix the issue where the repo_server fails to pick up the host key properly | 13:31 |
| *** rgogunskiy has joined #openstack-ansible | 13:33 | |
| *** keedya has joined #openstack-ansible | 13:35 | |
| *** richoid has quit IRC | 13:43 | |
| *** richoid has joined #openstack-ansible | 13:43 | |
| *** fawadkhaliq has joined #openstack-ansible | 13:43 | |
| *** markvoelker has joined #openstack-ansible | 13:47 | |
| *** rromans_afk is now known as rromans | 13:56 | |
| *** admin0 has quit IRC | 14:01 | |
| *** cemmason has joined #openstack-ansible | 14:08 | |
| *** admin0 has joined #openstack-ansible | 14:08 | |
| mhayden | happy monday, folks | 14:11 |
| andymccr | anybody have any ideas how we could make this more aware of changes?: https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/container_create.yml#L101-L112 | 14:14 |
| andymccr | at the moment its just a flat command so it says its "changed" on every run - so we can't do a notify on it. | 14:14 |
| andymccr | but for the case of "static_routes" we would want to restart the container so that the static_routes are in place. | 14:15 |
| andymccr | a notify would just restart the container on every run though :( | 14:15 |
| *** rgogunskiy has quit IRC | 14:18 | |
| *** gfa is now known as gfa_ | 14:19 | |
| *** cemmason has quit IRC | 14:19 | |
| *** metral is now known as metral_zzz | 14:20 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Fixing keepalived bug when 2+ backup nodes have the same priority https://review.openstack.org/279730 | 14:21 |
| mhayden | was there an email summary sent out from the OSA midcycle? i was trying to figure out if i missed it | 14:25 |
| *** woopstar has quit IRC | 14:26 | |
| *** KLevenstein has joined #openstack-ansible | 14:26 | |
| *** rromans has quit IRC | 14:28 | |
| *** rromans has joined #openstack-ansible | 14:28 | |
| evrardjp | good monday morning mhayden | 14:29 |
| mhayden | howdy evrardjp | 14:29 |
| mhayden | oh thanks, Apsu -- your sarcasm got Edgar all worried -> https://review.openstack.org/#/c/278128/19 :P | 14:29 |
| evrardjp | andymccr: maybe adapthing the lxc_container module by adding a template feature? The module may become bloated of features, but that could maybe help you... | 14:30 |
| evrardjp | You should ask cloudnull ;) | 14:30 |
| evrardjp | mhayden: I'm still laughing of the Evan's comment ;) | 14:31 |
| *** sdake has joined #openstack-ansible | 14:35 | |
| *** rgogunskiy has joined #openstack-ansible | 14:35 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-lxc_container_create: Restart containers on network configuration change https://review.openstack.org/286078 | 14:37 |
| *** pcaruana has quit IRC | 14:38 | |
| mhayden | evrardjp: yeah, Apsu will owe me lunch if i'm forced to add another patchset :| | 14:38 |
| evrardjp | :D is 20 your limit? | 14:40 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:42 | |
| *** ametts has joined #openstack-ansible | 14:42 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Adding release notes for erlang VM tuning parameters https://review.openstack.org/285506 | 14:51 |
| *** pcaruana has joined #openstack-ansible | 14:52 | |
| *** ShannonM has joined #openstack-ansible | 14:54 | |
| bgmccollum | andymccr quick (dirty) fix would be to MD5 the file before and after the change and return accordingly if it changed. lineinfile might be an option too... | 14:55 |
| andymccr | bgmccollum: already done that ;) | 14:55 |
| andymccr | https://review.openstack.org/#/c/286078/ | 14:56 |
| andymccr | feel free to review :D | 14:56 |
| bgmccollum | looking :) | 14:56 |
| *** sdake has quit IRC | 14:58 | |
| bgmccollum | andymccr how does notify behave with delegate to? do the ifup / ifdown need to be in the command? | 14:59 |
| andymccr | bgmccollum: unfortunately notify won't work because the "container_command" is a flat "command" so it has no concept of whether it has changed or not | 15:00 |
| andymccr | notify would end up just performing the task every single run. | 15:00 |
| bgmccollum | oh, you have would to also have a notify task for each interface too probably... | 15:00 |
| bgmccollum | hrm...ok | 15:00 |
| andymccr | so the options are: change the module itself to allow that to work, or do what i did there (i think) | 15:00 |
| andymccr | cant think of another option unfortunately | 15:00 |
| bgmccollum | it looks good, just wanted to exhaust all other options | 15:01 |
| andymccr | agree | 15:01 |
| andymccr | if there is a better way im all for it :) | 15:01 |
| odyssey4me | andymccr that's bloody horrible... but I can't think of a better way | 15:01 |
| bgmccollum | i think you can use the "changed when" | 15:01 |
| odyssey4me | except to optimise it a little - see my review | 15:01 |
| bgmccollum | to control when a command should trigger a change | 15:01 |
| andymccr | odyssey4me: you cant do that :( | 15:02 |
| andymccr | because the md5sum command fails if the file doesnt exist | 15:02 |
| andymccr | where as cat | md5sum won't fail | 15:02 |
| andymccr | it'll return the value for a failed cat as the md5sum which will still be viewed as a change (restarting the network as intended) | 15:02 |
| odyssey4me | yeah, I suppose you could stat the file contents - then compare the before and after | 15:02 |
| bgmccollum | andymccr maybe this? -- http://docs.ansible.com/ansible/playbooks_error_handling.html#overriding-the-changed-result | 15:03 |
| andymccr | bgmccollum: yeah you could, but the main problem is that its a "container_command" inside of the lxc_container task, so it becomes a bit complicated | 15:03 |
| bgmccollum | hmm... | 15:03 |
| *** jthorne has joined #openstack-ansible | 15:03 | |
| bgmccollum | headed to the office... | 15:03 |
| *** rgogunskiy has quit IRC | 15:05 | |
| *** Mudpuppy has joined #openstack-ansible | 15:15 | |
| Apsu | mhayden: haha. +1 | 15:16 |
| *** permalac has quit IRC | 15:19 | |
| *** permalac has joined #openstack-ansible | 15:19 | |
| *** permalac has quit IRC | 15:19 | |
| *** permalac has joined #openstack-ansible | 15:20 | |
| *** izaakk has joined #openstack-ansible | 15:21 | |
| *** phalmos has joined #openstack-ansible | 15:24 | |
| *** rohanp has joined #openstack-ansible | 15:25 | |
| rohanp | Hi, i have a query concerning the neutron packages installation in neutron venv? Is it necessary to install any specific neutron packages to be installed on neutron venv? | 15:25 |
| rohanp | I was referring to some parts like this: | 15:25 |
| rohanp | https://github.com/plumgrid/plumgrid-ansible/blob/master/plumgrid_playbooks/post_neutron_plumgrid.yml#L18 | 15:25 |
| rohanp | and this | 15:26 |
| rohanp | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/tasks/plumgrid_config.yml#L35 | 15:26 |
| *** admin0 has quit IRC | 15:26 | |
| rohanp | also the other question i had is what is the significance of making use of inventory hostname like as follows: | 15:27 |
| rohanp | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/tasks/plumgrid_config.yml#L51 | 15:27 |
| lbragstad | cloudnull o/ did something change with os_keystone in the last couple days? This change seems to have been reverted (https://github.com/os-cloud/openstack-ansible-os_keystone/pull/1) | 15:28 |
| *** spotz_zzz is now known as spotz | 15:28 | |
| odyssey4me | rohanp all the openstack bits are installed into venvs by default, so if the neutron agent/server needs access to the library then the library should be installed into the venv as we restrict the services from the host system libraries to prevent conflicts between the venv and the host | 15:30 |
| *** mgoddard__ has joined #openstack-ansible | 15:30 | |
| odyssey4me | rohanp and reference in https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/tasks/plumgrid_config.yml#L51 is to make sure that the package installed only installs on the appropriate group of servers | 15:31 |
| lbragstad | cloudnull yep - I think it was overridden https://github.com/os-cloud/openstack-ansible-os_keystone/commit/7d4cc4a21711328f2b7cb2e104495e5c54ae3aac#diff-29983f0df280e72c9ae0fb45414af972R26 | 15:31 |
| odyssey4me | in this particular case, that will only be installed on the neutron-server group, not the neutron-agent group | 15:31 |
| rohanp | <odyssey4me> makes sense | 15:31 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add notes on running the upgrade script https://review.openstack.org/285419 | 15:32 |
| rohanp | i am in the process of providing a custom neutron plugin like plumgrid have | 15:32 |
| rohanp | so my question was more on lines of whether i should be specifically installing my neutron packages on neutron venv | 15:32 |
| rohanp | or just intsalling them as is shud be good? | 15:32 |
| odyssey4me | rohanp if the neutron service needs access to the package, then it needs to be in the neutron venv | 15:33 |
| *** mgoddard_ has quit IRC | 15:33 | |
| *** woopstar has joined #openstack-ansible | 15:34 | |
| odyssey4me | mhayden no, you didn't miss a summary email because I haven't written one :p | 15:34 |
| mhayden | odyssey4me: ah, okay | 15:35 |
| *** git-harry has left #openstack-ansible | 15:35 | |
| *** raddaoui has joined #openstack-ansible | 15:35 | |
| *** woopstar has quit IRC | 15:35 | |
| *** git-harry has joined #openstack-ansible | 15:36 | |
| rohanp | <odyssey4me>: i am confused with the following approach: | 15:36 |
| rohanp | https://github.com/plumgrid/plumgrid-ansible/blob/master/plumgrid_playbooks/roles/plumgrid_neutron/tasks/main.yml#L53 | 15:36 |
| rohanp | since neutron lxc containers exist only after OSA scripts are run | 15:36 |
| rohanp | can we install neutron packages before that? i assume no; i think it qwould be possible to have neutron packages installed only after the neutron lxc onctainers are in place | 15:37 |
| rohanp | after OSA scripts get run | 15:37 |
| rohanp | but in the above link; those playbooks are run before OSA runs | 15:37 |
| cloudnull | lbragstad: yes I merged the history back into the repo. seems I missed something. | 15:39 |
| odyssey4me | rohanp it would seem that plumgrid is install both on the host and in the venv: https://github.com/plumgrid/plumgrid-ansible/blob/master/plumgrid_playbooks/roles/plumgrid_neutron/tasks/main.yml#L60 | 15:40 |
| cloudnull | lbragstad: you should be good now. | 15:41 |
| lbragstad | cloudnull awesome - I'll re-deploy and test it out | 15:42 |
| *** woopstar has joined #openstack-ansible | 15:43 | |
| *** admin0 has joined #openstack-ansible | 15:43 | |
| *** woopstar has quit IRC | 15:44 | |
| lbragstad | cloudnull fixed - our integration tests passed :) | 15:48 |
| *** woopstar has joined #openstack-ansible | 15:52 | |
| *** woopstar has quit IRC | 15:53 | |
| *** javeriak has joined #openstack-ansible | 15:55 | |
| *** jeh has joined #openstack-ansible | 16:03 | |
| *** phiche has quit IRC | 16:03 | |
| *** pabelanger has joined #openstack-ansible | 16:05 | |
| pabelanger | ohai | 16:05 |
| odyssey4me | pabelanger o/ | 16:06 |
| pabelanger | got stopped by family on way for coffee :) | 16:06 |
| pabelanger | okay, so IIRC you are looking to add depends-on for your ansible roles | 16:06 |
| cloudnull | lbragstad: sweet! | 16:06 |
| pabelanger | which is something I also want to do | 16:06 |
| odyssey4me | alright, pabelanger based on http://docs.openstack.org/infra/zuul/cloner.html#clone-map and https://github.com/openstack/windmill/blob/master/tools/install_roles.sh#L21-L25 can I confirm that you're pulling the roles specified in https://github.com/openstack/windmill/blob/master/tools/install_roles.sh#L30-L38 into the parent directory of the windwill role? | 16:07 |
| cloudnull | odyssey4me: idk if you'd seen this https://github.com/os-cloud?utf8=%E2%9C%93&query=os_ however I've been able to save all of the history for the OS_* roles as they've been broken out. jmccrory gave me a suggestion last week and it seemed to work well enough. | 16:07 |
| odyssey4me | hte mapping seems to say that any repo cloned by the name 'openstack/ansible-role-(.*)' must be placed in 'playbooks/roles/windmill.\1' which equates to 'playbooks/roles/' I think? | 16:08 |
| pabelanger | odyssey4me: almost, they are actually getting cloned into playbooks/roles folder inside of windmill | 16:08 |
| pabelanger | odyssey4me: right | 16:08 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add upgrade config change tools https://review.openstack.org/285564 | 16:08 |
| pabelanger | so, in the case of a role dependency, we can't do that | 16:08 |
| admin0 | time for me to go home, turn ON my test cluster and try the odyssey4me solution :D for mysql on metal | 16:09 |
| odyssey4me | cloudnull friggin awesome! :) maybe you can redo https://github.com/odyssey4me/openstack-ansible-openstack_openrc as you missed that one ? | 16:09 |
| pabelanger | since we expect the current `pwd` from jenkins to be the actual role we are using | 16:09 |
| pabelanger | so, if we are to use clonemap, we should set the dest to .. | 16:09 |
| odyssey4me | cloudnull then https://review.openstack.org/284512 needs an update for your location | 16:09 |
| pabelanger | since we want the directory above a role | 16:09 |
| lbragstad | cloudnull working on tying together a gate for all this stuff :) | 16:09 |
| odyssey4me | pabelanger well, as long as the role is in the expected roles location, then it's all good | 16:10 |
| *** admin0 has left #openstack-ansible | 16:10 | |
| pabelanger | odyssey4me: which 2 roles are you setting up a dependency on? | 16:10 |
| pabelanger | odyssey4me: right, as long as ansible.cfg can find them, we are fine | 16:10 |
| *** admin0 has joined #openstack-ansible | 16:10 | |
| odyssey4me | pabelanger well, I want this facility on them all - for the main repo we use ansible.cfg, but in the independent roles we use the env var | 16:11 |
| odyssey4me | so let's start witht he main repo | 16:11 |
| odyssey4me | I need to figure out how to read https://github.com/openstack/openstack-ansible/blob/master/ansible-role-requirements.yml and use it for zuul cloner - that'll be the first challenge, but I think I can manage that | 16:11 |
| pabelanger | another option, is to setup the ansible JJB to read requirements.yml as a builder, then deal with zuul-cloning the roles. I am not a fan of this approach, as I would rather have tox do all the work | 16:11 |
| *** woopstar has joined #openstack-ansible | 16:12 | |
| cloudnull | odyssey4me: updated https://review.openstack.org/#/c/284512/ | 16:12 |
| pabelanger | odyssey4me: okay, that will be the easier to start with. I've already started work on a requirements.yml parser to zuul-cloner. I should have time to day to finish it | 16:12 |
| pabelanger | and push it up | 16:12 |
| odyssey4me | pabelanger oh? that sounds interesting | 16:13 |
| odyssey4me | cloudnull thanks! | 16:13 |
| odyssey4me | pabelanger I was thinking of implementing an alternative method for pulling down roles into our ansible bootstrap process (ie give the option of using z-c along with ansible-galaxy) | 16:14 |
| pabelanger | odyssey4me: thats the end goal for me, so we don't maintain 2 cloning sources. However, if you want to setup what I did for tools/install_roles.sh to start, that will be the fastest way to use depends-on | 16:14 |
| pabelanger | ya, that works too | 16:14 |
| odyssey4me | pabelanger ok, so if z-c can read the role requirements file - how ill it work then? | 16:14 |
| odyssey4me | do we simply point it at the file and it does the thing? | 16:14 |
| odyssey4me | so basically I don't have to define the map, or the list of repositories - I simply just need to make z-c do the clone instead of ansible-galaxy ? | 16:15 |
| pabelanger | odyssey4me: not directly, I'd update tools/install_roles.sh to read requirements.yml, and convert contents to clonemap. We do that for puppet modules also | 16:15 |
| odyssey4me | oh I see - you're not modifying z-c, you're simply creating an interim converter? | 16:16 |
| pabelanger | right | 16:16 |
| pabelanger | however | 16:16 |
| pabelanger | with zuulv3 and ansible dependency, we could talk to jeblair about adding some sort of mapping into zuul | 16:16 |
| pabelanger | there might be a use case for it for more projects | 16:16 |
| palendae | pabelanger: fwiw, that'd be awesome for https://review.openstack.org/#/c/273771/ | 16:17 |
| *** fawadkhaliq has quit IRC | 16:17 | |
| odyssey4me | pabelanger yeah, that'd be nice | 16:17 |
| *** pcaruana has quit IRC | 16:17 | |
| *** admin0 has quit IRC | 16:17 | |
| pabelanger | odyssey4me: Ya, I think we could bring that up in the future. But to get your team rolling asap, using the install_roles.sh concept I do for windmill is likely the place to start | 16:18 |
| pabelanger | palendae: Ya, I've been meaning to loop into that and try it out | 16:18 |
| *** stevelle_ is now known as stevelle | 16:18 | |
| odyssey4me | pabelanger how does z-c magically know about the dependent patches? | 16:18 |
| *** rromans has quit IRC | 16:19 | |
| pabelanger | odyssey4me: so, there are ZUUL_ environment vars that are expected to be setup. Which zuul does via jenkins | 16:19 |
| pabelanger | https://jenkins04.openstack.org/job/gate-nova-python34/5084/parameters/ for an example | 16:19 |
| odyssey4me | pabelanger so in my testing I have exported all the env vars set and don't see them | 16:20 |
| odyssey4me | I suspect that I may be missing some sort of job setting to pass the vars through to the env? | 16:20 |
| pabelanger | right, if you want to test it locally, then you need to export ZUUL_REF=refs/zuul/master/Zbf503151cebc448da10b600ee946d7f5 for example | 16:20 |
| pabelanger | then zuul-cloner know which project changes to depend on | 16:20 |
| *** rromans has joined #openstack-ansible | 16:21 | |
| odyssey4me | pabelanger alright, let me try and work something up - then I'll ping you for a review if that's ok? | 16:21 |
| pabelanger | odyssey4me: WFM | 16:22 |
| woopstar | Are there any dashboard/gui thingys for the Telemetry service, like CloudWatch on AWS? | 16:22 |
| pabelanger | odyssey4me: should be a matter of search / replace of the script I use. Then telling tox to use it. No other changes needed in the gate | 16:23 |
| woopstar | I was looking a bit on Riemann.io - dunno if there is any better alternatives or something native? | 16:23 |
| *** ametts has quit IRC | 16:24 | |
| odyssey4me | pabelanger our integration gate and role gating work a little differently, but I'll adapt them appropriately - the trick is going to be to interpret the role requirements file... but perhaps I'll base it on a static list to start with | 16:24 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 16:25 | |
| pabelanger | right, static easiest to start with. once you see how it all works, you'll be ready for dynamic | 16:25 |
| odyssey4me | pabelanger one thing I'm thinking through - does the clone map allow a repo to be cloned into a folder with a different name? | 16:25 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 16:25 | |
| pabelanger | odyssey4me: yes. that is what I am doing. | 16:25 |
| pabelanger | odyssey4me: you'll have to read up on the docs to see how regex works | 16:26 |
| odyssey4me | is that what https://github.com/openstack/windmill/blob/master/tools/install_roles.sh#L23 does? | 16:26 |
| pabelanger | odyssey4me: yes, so I take ansible-role-zuul and zuul-clone it into playbooks/roles/windmill.zuul | 16:26 |
| openstackgerrit | Nate Potter proposed openstack/openstack-ansible: Check for AODH host before adding alarm_connection https://review.openstack.org/284392 | 16:27 |
| odyssey4me | pabelanger ah, I see how it maps now - the docs didn't quite get me there (I can be a bit dumb) :p | 16:27 |
| *** abitha has joined #openstack-ansible | 16:27 | |
| odyssey4me | ok cool, got it | 16:27 |
| pabelanger | odyssey4me: I'd suggest looking in openstack/project-config too. There are some other projects doing the same | 16:28 |
| *** mgoddard__ has quit IRC | 16:29 | |
| *** mgoddard has joined #openstack-ansible | 16:29 | |
| *** neilus1 has joined #openstack-ansible | 16:29 | |
| *** gfa_ is now known as gfa | 16:33 | |
| *** neilus has quit IRC | 16:33 | |
| *** electrofelix has joined #openstack-ansible | 16:36 | |
| *** keedya has quit IRC | 16:39 | |
| *** keedya has joined #openstack-ansible | 16:39 | |
| *** weezS has joined #openstack-ansible | 16:47 | |
| *** raddaoui has quit IRC | 16:52 | |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible-pip_install: Test version of pip being installed https://review.openstack.org/285284 | 16:53 |
| *** raddaoui has joined #openstack-ansible | 16:53 | |
| *** sdake has joined #openstack-ansible | 17:06 | |
| *** alextricity25_ is now known as alextricity25 | 17:16 | |
| *** openstackgerrit has quit IRC | 17:18 | |
| *** cemmason has joined #openstack-ansible | 17:18 | |
| *** openstackgerrit has joined #openstack-ansible | 17:18 | |
| *** evrardjp has quit IRC | 17:19 | |
| *** sdake has quit IRC | 17:22 | |
| *** sdake has joined #openstack-ansible | 17:23 | |
| *** asettle has quit IRC | 17:23 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 17:26 | |
| *** gayathri has joined #openstack-ansible | 17:27 | |
| *** admin0 has joined #openstack-ansible | 17:30 | |
| *** woopstar has quit IRC | 17:32 | |
| *** electrofelix has quit IRC | 17:32 | |
| *** electrofelix has joined #openstack-ansible | 17:35 | |
| *** woopstar has joined #openstack-ansible | 17:35 | |
| *** abitha has quit IRC | 17:35 | |
| *** fawadkhaliq has joined #openstack-ansible | 17:36 | |
| *** woopstar has quit IRC | 17:37 | |
| gayathri | Hi | 17:41 |
| gayathri | I am facing an issue while installing swift service using openstack ansible | 17:41 |
| gayathri | playbook ran without throwing any errors | 17:41 |
| gayathri | when i see in dashboard..its throwing error like "unable to retrieve the containers" | 17:42 |
| gayathri | there was no logs also | 17:42 |
| gayathri | Any help is much appreciated | 17:43 |
| *** weezS has quit IRC | 17:43 | |
| *** woopstar has joined #openstack-ansible | 17:45 | |
| *** admin0 has quit IRC | 17:46 | |
| *** admin0 has joined #openstack-ansible | 17:46 | |
| *** woopstar has quit IRC | 17:46 | |
| cloudnull | gayathri: does the swift commands work ? | 17:46 |
| cloudnull | IE can you do a: swift list ? | 17:46 |
| gayathri | no..when i give that command..it shows no service | 17:47 |
| cloudnull | and/or create a container / add an object into the container? | 17:47 |
| cloudnull | like youre getting a 503 ? | 17:48 |
| gayathri | yeah | 17:50 |
| *** woopstar has joined #openstack-ansible | 17:52 | |
| *** woopstar has quit IRC | 17:53 | |
| *** sdake has quit IRC | 17:54 | |
| gayathri | what could be the issue? | 17:55 |
| *** kencjohnston has joined #openstack-ansible | 17:56 | |
| *** sdake has joined #openstack-ansible | 17:57 | |
| cloudnull | that would seem like an issue with the lb | 17:59 |
| cloudnull | are the rest of the services running normally. | 17:59 |
| cloudnull | 's/./?/ | 17:59 |
| cloudnull | did you have an existing deployment and then add swift ? | 18:00 |
| admin0 | time to do the mapping as odyssey4me suggested . and if does not work, disect cloudnull and get that knowledge out :D | 18:00 |
| cloudnull | are you running haproxy? | 18:00 |
| cloudnull | if so, have you rerun the haproxy play? | 18:00 |
| cloudnull | admin0: whats that ? | 18:01 |
| cloudnull | mapping ? | 18:01 |
| *** woopstar has joined #openstack-ansible | 18:01 | |
| *** agireud has quit IRC | 18:01 | |
| admin0 | many people have tried to explain to me how mappings work .. like I want to run neutron servers in a new group of servers called my_awesome_network_servers and mysql and rabbitmq in another group of servers called my_ssd_db_servers .. | 18:02 |
| admin0 | on metal | 18:02 |
| admin0 | so how to map those :D | 18:02 |
| cloudnull | ah. | 18:02 |
| admin0 | at the end, ( after banging the heads on the wall ) everyone said .. wait for cloudnull :D | 18:02 |
| admin0 | you are the guy :D | 18:02 |
| odyssey4me | cloudnull fyi between andymccr and I we came up with https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb | 18:03 |
| *** agireud has joined #openstack-ansible | 18:03 | |
| *** woopstar has quit IRC | 18:03 | |
| *** kencjohnston has quit IRC | 18:04 | |
| admin0 | basically the neutron nodes have special cards that do checksum offloading, so the need, and the db is because the organization has dba team and to cerify/be in complainance, they need to have the DB in specified servers | 18:04 |
| cloudnull | admin0: adding is_metal: true to https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/galera.yml in the properties section should be all thats needed | 18:04 |
| admin0 | cloudnull: i want to run mysql on only mysql servers ( no other infra servers there ) | 18:04 |
| cloudnull | kk. | 18:04 |
| admin0 | which is why the need to come up with a new set of hosts | 18:05 |
| cloudnull | https://gist.github.com/odyssey4me/3e92f90cf6431432b8cb#file-galera-yml that should do it. | 18:05 |
| * cloudnull goes to try | 18:05 | |
| gayathri | ya rest of my services running successfuly.. | 18:05 |
| * admin0 prepares pizza and beer for cloudnull | 18:05 | |
| * cloudnull will work for beer / pizza | 18:06 | |
| * admin0 won’t forget odyssey4me and andymccr as well :D | 18:06 | |
| gayathri | I had added swift in the existing deployment | 18:06 |
| gayathri | i run haproxy..y do i need to rerun haproxy again | 18:06 |
| admin0 | cloudnull: if possible, create 2 set of servers .. new_dba_servers and new_network_servers, and then show a way for example on the new_dba_servers , run mysql on metal and rabbitmq on containers, and on neutron, run the neutron agents on metal | 18:08 |
| admin0 | i think that would be great start and people will finally understand how this whole mapping works | 18:08 |
| odyssey4me | admin0 the trouble is that you'd be adding yet another group for no reason | 18:09 |
| admin0 | ok .. skip the neutron part :) | 18:10 |
| admin0 | new_dba_servers where galera runs on metal, and rabbitmq runs on containers | 18:10 |
| admin0 | ?? | 18:10 |
| odyssey4me | you already have the db on metal | 18:10 |
| odyssey4me | do you want rabbitmq on the same host as the db, but with rabbitmq in a container? | 18:10 |
| admin0 | yes | 18:11 |
| odyssey4me | what's the point - why not just have them both in containers? | 18:11 |
| admin0 | there is nothing wrong with containers, but if client says it needs to be in metal, i can argue .. but he wins .. and I want to learn now how its done | 18:12 |
| odyssey4me | but having the db on metal and rabbit on the same host in a container is rather silly IMO | 18:12 |
| odyssey4me | you gain nothing and, in fact, you're losing flexibility | 18:12 |
| odyssey4me | with both in containers you can replace either at any time - with one on the host and one in a container if you need to re-setup your DB server you risk affecting rabbit | 18:13 |
| odyssey4me | fair enough to do as a learning process - but it is a very, very silly thing to do in production | 18:14 |
| admin0 | i have one case where someone asked that they have a dba team and their mysql needs to be in specific machines on metal . | 18:14 |
| admin0 | and another case is where neutron needs to run on specific hardware with special network cards for checksum offloading | 18:14 |
| cloudnull | admin0: so this is what i did to totally isolate the galera hosts. | 18:14 |
| cloudnull | http://cdn.pasteraw.com/7imvcoqzuz1ty6z2ybfd9g1azxcm9ff | 18:14 |
| odyssey4me | admin0 sure, but then have your rabbitmq on another host | 18:14 |
| cloudnull | in the openstack_user_config file i added | 18:14 |
| cloudnull | http://cdn.pasteraw.com/67ohkezuoijpzp4uc6kmzlxb5mc2ins | 18:14 |
| cloudnull | which generated http://cdn.pasteraw.com/6u1uftc3n9hwziqz424rtizbg2dr3xb | 18:14 |
| admin0 | odyssey4me: that is correct .. i asked here just for exampel to undersand how to learn | 18:15 |
| admin0 | i am not going to have rabbit and galera in the same system | 18:15 |
| *** raddaoui has quit IRC | 18:15 | |
| odyssey4me | ah ok | 18:15 |
| odyssey4me | admin0 also, fyi, those db servers are going to be bored ;) | 18:15 |
| admin0 | :) | 18:16 |
| *** woopstar has joined #openstack-ansible | 18:16 | |
| cloudnull | in that example, i change the name to db_ however galera_ would work too | 18:17 |
| cloudnull | i only did that because it was easy to find the inventory changes | 18:17 |
| *** woopstar has quit IRC | 18:18 | |
| admin0 | cloudnull: so under the container_skel, put arbitary name like belongs_to: new_host_group_containers .. in physical skel also put new_host_group_containers to define it .. and then in user_config, put hosts as new_host_group_hosts: ?? | 18:18 |
| cloudnull | and ansible picked up the changes as i'd expect | 18:18 |
| cloudnull | http://cdn.pasteraw.com/p0zof7s840yv91bp05vz7zj96j9jma | 18:18 |
| cloudnull | yes. | 18:19 |
| cloudnull | in the physical skel a container group must have a corresponding host group | 18:20 |
| admin0 | got it | 18:20 |
| *** woopstar has joined #openstack-ansible | 18:20 | |
| cloudnull | so <NAME>_containers must also have <NAME>_hosts | 18:20 |
| cloudnull | the seperator is _ | 18:20 |
| admin0 | and for any other containers to run here, just do belongs_to: new_container_group .. but not set is_metal: true | 18:20 |
| admin0 | so if is_metal: true = on metal, else containers | 18:20 |
| cloudnull | yup | 18:20 |
| * admin0 gives pizza and beer to cloudnull odyssey4me andymccr | 18:21 | |
| admin0 | \o/ | 18:21 |
| *** neilus1 has quit IRC | 18:21 | |
| admin0 | now i get this :D | 18:21 |
| cloudnull | while the "containers" name is intended for lxc containers if is_metal=true the container becomes the physical host. | 18:21 |
| *** woopstar has quit IRC | 18:21 | |
| * admin0 goes on to create 1000s of cool host_group_names | 18:21 | |
| *** electrofelix has quit IRC | 18:22 | |
| cloudnull | w/ the bits in env.d you can create your own archetecture it super flexable, albeit confusing | 18:23 |
| admin0 | odyssey4me: i will try with this, and using url in the hostname instead of ip .. though I am not sure how haproxy is going to bind it | 18:23 |
| *** Guest63721 is now known as mfisch | 18:24 | |
| logan- | admin0: i use hostnames for internal/external vip | 18:24 |
| logan- | it works ok | 18:24 |
| *** rohanp has quit IRC | 18:25 | |
| *** mfisch is now known as Guest2004 | 18:25 | |
| admin0 | ok | 18:25 |
| logan- | external worked pretty much out of the box. internal had a minor hangup or two but I can't remember what they were | 18:25 |
| admin0 | my setup is internet => firewall/lb/nat => openstack .. so even the public is private and natted from the lb .. | 18:25 |
| admin0 | but will use the hostnames and give it a try | 18:26 |
| admin0 | haproxy is on metal, so if i use hostnames, how do I define the shared ip (for keepalive ) ? | 18:26 |
| logan- | you have to override the keepalive cidr, haproxy_keepalived_external_vip_cidr and haproxy_keepalived_internal_vip_cidr need to get set to actual ip/cidr | 18:28 |
| *** gayathri has quit IRC | 18:28 | |
| *** Guest2004 is now known as mfisch | 18:28 | |
| *** mfisch has quit IRC | 18:28 | |
| *** mfisch has joined #openstack-ansible | 18:28 | |
| admin0 | i will try now :D | 18:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [WIP] Gate: Enable 'Depends-On' cross-repo dependent patching for roles https://review.openstack.org/286195 | 18:31 |
| logan- | if you use internal hostnames make sure all your hosts can resolve them... haproxy etc won't even start unless they can resolve whatever you set the internal/external vip names to | 18:31 |
| *** abitha has joined #openstack-ansible | 18:31 | |
| cloudnull | admin0: word of warning, as you go to redefine your architecture make sure you keep the following groups which are referenced in the plays and roles http://cdn.pasteraw.com/nekofed86qdftw2lr56ish8fnrz54ae | 18:32 |
| admin0 | i will just add 2 new set of physical_hosts .. won’t rename anything else | 18:32 |
| cloudnull | that should be just fine. i didnt want to you run into other issues by breaking those bits so i figured i'd mention it. | 18:33 |
| admin0 | thanks .. | 18:34 |
| odyssey4me | pabelanger if you can peek at https://review.openstack.org/286195 I'd appreciate it... it's very much WIP but some feedback would be good | 18:34 |
| odyssey4me | afk while I relocate | 18:34 |
| *** cloudtrainme has joined #openstack-ansible | 18:37 | |
| *** woopstar has joined #openstack-ansible | 18:39 | |
| *** woopstar has quit IRC | 18:41 | |
| *** sdake has quit IRC | 18:43 | |
| *** jcrubio has joined #openstack-ansible | 18:49 | |
| *** kencjohnston has joined #openstack-ansible | 18:49 | |
| *** rparulek has joined #openstack-ansible | 18:50 | |
| rparulek | Hi, using OSA is there a way we can use "apt" packages install to install packages in say neutron venv rather than using pip install? is there any sample snippet that does this? | 18:51 |
| rparulek | the goal here is to be able to install packages using apt utility within a venv of openstack service say neutron | 18:52 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 18:53 | |
| cloudnull | rparulek: everything is installed, as it pertains to openstack, via pip. | 18:53 |
| cloudnull | is there a specific package you need/want? | 18:54 |
| cloudnull | essentially there's no way to force apt to install python bits into a venv. | 18:54 |
| *** ametts has joined #openstack-ansible | 18:54 | |
| rparulek | <cloudnull>: my goal is to be able to install some custom neutron packages in form of debians on controller targets within a venv | 18:55 |
| cloudnull | you could use https://github.com/blueboxgroup/giftwrap to build apt packages that deploy a venv but thats not really what youre asking for it would seem. | 18:55 |
| rparulek | since neutron-server will need access to these pcakges..i will need to install these on neutron venv right? | 18:56 |
| cloudnull | yes. | 18:56 |
| cloudnull | if you add your custom packages via git sources here https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml | 18:56 |
| *** jthorne_ has joined #openstack-ansible | 18:57 | |
| cloudnull | then add the egg name to https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/defaults/main.yml#L343 | 18:57 |
| *** jthorne has quit IRC | 18:57 | |
| cloudnull | that'll build you custom packages and install them in the neutron venv | 18:57 |
| *** izaakk has quit IRC | 18:57 | |
| cloudnull | eggname == basename | 18:57 |
| cloudnull | so neutron_git_repo: https://git.openstack.org/openstack/neutron-custome-package would have an eggname of "neutron_custom_package" | 18:58 |
| rparulek | <cloudnull>: so what you intend to convey is that to install any custom packages within neutron venv we shud use python pip packages and NOT debians in form of .deb? | 18:58 |
| cloudnull | yes | 18:59 |
| automagically | rparulek: A venv by nature only deals with pip packages | 18:59 |
| cloudnull | they have be git sources that are python installable | 18:59 |
| cloudnull | ^ that | 18:59 |
| cloudnull | i shouldve said that :) | 18:59 |
| rparulek | <cloudnull>: ok, makes sense | 19:00 |
| *** arif-ali has quit IRC | 19:06 | |
| rparulek | <cloudnull>: how is it occuring in case of PLumgrid for that matter? I donot see them adding any entries to files you pointed to? | 19:07 |
| rparulek | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/tasks/plumgrid_config.yml#L35 | 19:07 |
| cloudnull | rparulek: that package is defined here https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/defaults/main.yml#L358 | 19:09 |
| cloudnull | and available on pypi | 19:09 |
| cloudnull | if the package you want to install is already on pypi all you need to do is add it to the list https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/defaults/main.yml#L343 | 19:10 |
| rparulek | <cloudnull>: if i donot have it up there; then steps to follow are 1) add an entry to this file somehwre under neutron here "https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml#L80" with my neutron packages | 19:11 |
| rparulek | 2) add custom package names here https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/defaults/main.yml#L343 | 19:12 |
| cloudnull | yup in 1) 3 lines added something like neutron_custom_thing_git_repo: https://git.openstack.org/openstack/neutron-custom-thing ... https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml#L81-L83 | 19:13 |
| cloudnull | package name would then tranlate to "neutron_custom_thing" | 19:13 |
| alextricity25 | Hey all. If I set the container, tunnel, and storage cidr to be the same network range, will the dynamic inventory script create IPs that overlap? Or will it put the IPS on that range in the same queue for container, tunnel, and storage IPs? | 19:15 |
| *** raddaoui has joined #openstack-ansible | 19:15 | |
| alextricity25 | In other words, is it smart enough to recognize that they are the same CIDR? | 19:16 |
| cloudnull | alextricity25: i've never specifically tried it however i think so | 19:16 |
| cloudnull | i believe each entry is its own queue | 19:16 |
| cloudnull | regardless of the actual cidr value | 19:16 |
| alextricity25 | cloudnull: If that's the case, then there is a possibility of opverlapping IPs | 19:17 |
| cloudnull | yes | 19:17 |
| alextricity25 | How would I go about running all network traffic over "br-mgmt" | 19:17 |
| cloudnull | so no br-vlan, storage, etc | 19:18 |
| *** jcrubio has quit IRC | 19:18 | |
| alextricity25 | cloudnull: Exactly | 19:18 |
| *** jthorne_ has quit IRC | 19:19 | |
| alextricity25 | I'm faced with a use case in which I don't have multiple interfaces, nor have the option to configure VLANs on the ToR switches | 19:19 |
| cloudnull | remove the additional network entries that you dont need from the openstack_user_confgi file and add the elimiated network groups to https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.aio#L25 | 19:19 |
| *** jthorne has joined #openstack-ansible | 19:19 | |
| *** neilus has joined #openstack-ansible | 19:20 | |
| alextricity25 | cloudnull: Ah okay. I see now | 19:20 |
| alextricity25 | Thanks cloudnull :) | 19:20 |
| *** neilus has quit IRC | 19:20 | |
| cloudnull | alextricity25: you'd have something like http://cdn.pasteraw.com/9x42ibwn3gihpjhgza4a4292am62ybc | 19:20 |
| cloudnull | you may have trouble with l3 networks | 19:21 |
| cloudnull | but that should work | 19:21 |
| alextricity25 | I was thinking about keeping the eth12 inside the neutron agent container | 19:22 |
| cloudnull | yea youll have to play with it to make it fit your situation however it should be totally do able. | 19:22 |
| alextricity25 | and just change that bridge to be br-mgmt | 19:22 |
| *** neilus has joined #openstack-ansible | 19:23 | |
| *** gregfaust has joined #openstack-ansible | 19:25 | |
| *** ametts has quit IRC | 19:27 | |
| *** b3rnard0 is now known as b3rnard0_away | 19:28 | |
| *** javeriak_ has joined #openstack-ansible | 19:28 | |
| *** javeriak_ has quit IRC | 19:28 | |
| *** raddaoui has quit IRC | 19:30 | |
| *** jcrubio has joined #openstack-ansible | 19:31 | |
| *** javeriak has quit IRC | 19:32 | |
| *** gfa is now known as gfa_ | 19:33 | |
| *** raddaoui has joined #openstack-ansible | 19:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server: Removed pip requirements from the repo-server role https://review.openstack.org/284894 | 19:34 |
| admin0 | logan-: for the vip address, adding those in /etc/hosts is enough right ? | 19:37 |
| logan- | should be fine as long as you add it in every host and container | 19:38 |
| logan- | it needs to be everywhere in the environment | 19:38 |
| logan- | openstack_hosts does some hosts file management, i wonder if there is a var to do custom hosts entries there | 19:38 |
| admin0 | in hosts i can .. for containers, they get created in runtime by ansible | 19:39 |
| admin0 | hmm.. i will setup dns then :) | 19:39 |
| logan- | yeah I ran into that pretty quick... i am building unbound containers which handle this stuff | 19:40 |
| logan- | https://github.com/Logan2211/ansible-unbound | 19:41 |
| logan- | along with https://github.com/Logan2211/ansible-resolvconf | 19:41 |
| logan- | andthen something like.. http://paste.gentoolinux.info/okidotonam.sm http://paste.gentoolinux.info/yahuzajaho.sm http://paste.gentoolinux.info/zubareyiwe.sm | 19:44 |
| admin0 | my setup is inside a pfsense with pfsense as the gw/dns resolver .. so just adding the 2 entires are easy | 19:45 |
| *** weezS has joined #openstack-ansible | 19:45 | |
| logan- | ah ok, cool | 19:45 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible: Remove exception raised for container_cidr https://review.openstack.org/277511 | 19:48 |
| admin0 | all setup done .. time to run :D | 19:56 |
| *** jcrubio has left #openstack-ansible | 19:57 | |
| admin0 | to test, i have db_hosts, where haproxy and galera are on metal, memcache and rabbitmq on containers, and then rest all like normal . and vip address as hostnames and dns setup properly with proper interace and address override for keealive | 19:57 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add upgrade config change tools https://review.openstack.org/285564 | 19:57 |
| admin0 | how do i scan a file in linux for \t :D | 19:58 |
| admin0 | nvm | 19:59 |
| admin0 | found it | 19:59 |
| bgmccollum | jmccrory odyssey4me Are there still issues with the way this is implemented? -- https://review.openstack.org/#/c/279859/ | 20:01 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible: Allow only vxlan networks to be defined. https://review.openstack.org/276320 | 20:03 |
| *** cemmason1 has joined #openstack-ansible | 20:03 | |
| *** cemmason has quit IRC | 20:04 | |
| *** sdake has joined #openstack-ansible | 20:04 | |
| *** elopez has joined #openstack-ansible | 20:05 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-plugins: Created deprecation filter https://review.openstack.org/286250 | 20:06 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add missing runtime information to upgrade script https://review.openstack.org/286251 | 20:06 |
| admin0 | why does the installer downloads the same rpc-repo on different hosts ? can’t it download to the deploy server and copy locally ? | 20:09 |
| admin0 | 20 hosts = 20 parallel downloads | 20:09 |
| admin0 | suddenly this silence :D .. i asked something wrong ? | 20:11 |
| cloudnull | admin0: sorry been in a different zone. what download are you seeing ? | 20:11 |
| cloudnull | the container image? | 20:12 |
| admin0 | yeah | 20:12 |
| jmccrory | bgmccollum thanks for clearing things up there | 20:12 |
| cloudnull | that could be changed however its not something we've done yet. | 20:12 |
| cloudnull | there is some work to change the lxc create command to use the download template which will do most of this for us. | 20:12 |
| cloudnull | but i've seen nothing come of that just yet. | 20:13 |
| admin0 | i will soon start to contribute .. understanding this is the first process :) | 20:13 |
| *** b3rnard0_away is now known as b3rnard0 | 20:14 | |
| cloudnull | admin0: you can change the download path. | 20:14 |
| cloudnull | to something you already have to make things faster | 20:14 |
| cloudnull | thats defined here https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/defaults/main.yml#L100-L105 | 20:15 |
| *** htruta has joined #openstack-ansible | 20:15 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Security: Check for grub.cfg first https://review.openstack.org/285483 | 20:15 |
| admin0 | :) | 20:16 |
| htruta | hi guys. Is there any known ansible issue related to keystone v3 compatibility? | 20:16 |
| automagically | @admin0 - https://review.openstack.org/#/c/274456/ | 20:17 |
| bgmccollum | htruta are you referring to some of the OpenStack modules available in Ansible? | 20:18 |
| automagically | htruta: I’ve come across this: https://review.openstack.org/#/c/274686/ | 20:18 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add upgrade config change tools https://review.openstack.org/285564 | 20:18 |
| *** phiche has joined #openstack-ansible | 20:19 | |
| admin0 | automagically: nice . .. hopeing someone with +1 +2 it soon :D | 20:20 |
| admin0 | but that still is when fails to download | 20:20 |
| admin0 | still will be 20-40 parallel concurrent downloads | 20:20 |
| admin0 | based on number of hosts deployed | 20:20 |
| htruta | bgmccollum: if you found some, I'm totally interested. but the focus was more ansible related, like some restriction it has come across on setting credentials and stuff like that | 20:20 |
| htruta | automagically: nice! | 20:21 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add missing runtime information to upgrade script https://review.openstack.org/286251 | 20:22 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Upgrades: Cleanup RabbitMQ / vhost https://review.openstack.org/272652 | 20:22 |
| *** sdake has quit IRC | 20:22 | |
| htruta | our effort in keystone team is towards using v3 and keystoneauth where possible | 20:22 |
| admin0 | where do i define the haproxy_keepalived_authentication_password ? | 20:22 |
| *** fawadkhaliq has quit IRC | 20:22 | |
| *** neilus has quit IRC | 20:24 | |
| admin0 | never mind .. somehow its there, but not seeing it | 20:24 |
| admin0 | where to override the dns entries received by the containers ? is it the https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/defaults/main.yml ? | 20:27 |
| logan- | lxc_cache_resolvers | 20:28 |
| cloudnull | admin0: https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/defaults/main.yml#L89-L91 | 20:28 |
| *** jeh has quit IRC | 20:31 | |
| *** raddaoui has quit IRC | 20:34 | |
| *** cemmason1 has quit IRC | 20:35 | |
| *** cemmason has joined #openstack-ansible | 20:35 | |
| admin0 | changed dns in that file, ran setup-hosts again .. i see no changed in the resolv.conf setting | 20:35 |
| *** keedya has quit IRC | 20:35 | |
| admin0 | still see 10.0.3.1 and 8.8.8.8 and 8.8.4.4 | 20:35 |
| logan- | i think it will only change on new containers | 20:35 |
| *** keedya has joined #openstack-ansible | 20:35 | |
| admin0 | oh | 20:36 |
| *** yarkot_ has joined #openstack-ansible | 20:36 | |
| admin0 | i will delete 1 and test | 20:36 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-plugins: Created deprecation filter https://review.openstack.org/286250 | 20:38 |
| *** keedya has quit IRC | 20:38 | |
| *** yarkot_ has quit IRC | 20:38 | |
| *** keedya has joined #openstack-ansible | 20:39 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Replacing LBaaSv1 with v2 configuration https://review.openstack.org/275440 | 20:39 |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Add Ansible bootstrapping to Liberty upgrade https://review.openstack.org/286269 | 20:39 |
| cloudnull | odyssey4me: http://cdn.pasteraw.com/e8ykztoax8b932uudrdl3277c2lxhup <- deprecation warning filter like we talked about at the midcycle. https://review.openstack.org/286250 | 20:40 |
| *** cemmason has quit IRC | 20:40 | |
| *** cemmason has joined #openstack-ansible | 20:40 | |
| admin0 | yep .. delted and recreated and it worked | 20:42 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Replacing LBaaSv1 with v2 configuration https://review.openstack.org/275440 | 20:42 |
| admin0 | hmm.. still the first line there is 10.0.3.1 | 20:43 |
| admin0 | it and 2nd is my custom dns | 20:43 |
| admin0 | the 2nd is the one which has the dns for my vip | 20:43 |
| logan- | you probably dont have lxcbr0 (eth0 in the containers) disabled | 20:43 |
| admin0 | nope .. should it be ? | 20:43 |
| logan- | so the container host dnsmasq is sending resolver thru that | 20:43 |
| logan- | 10.0.3.1 will send requests up to the physical host | 20:44 |
| admin0 | ok | 20:45 |
| admin0 | i will fix the public mapping and dns first i guess :D | 20:45 |
| admin0 | namecheap, 2fa, for dns ..was being lazy | 20:46 |
| *** cemmason has quit IRC | 20:47 | |
| *** cemmason has joined #openstack-ansible | 20:47 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-plugins: Created deprecation filter https://review.openstack.org/286250 | 21:01 |
| cloudnull | odyssey4me: http://cdn.pasteraw.com/7hk2kcuy661c54w4mdaj1df4czl3r1f <- that should do it in a very ansible centric way. | 21:03 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-rabbitmq_server: [WIP] Multi-distro support for rabbitmq-server role https://review.openstack.org/286282 | 21:07 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Replacing LBaaSv1 with v2 configuration https://review.openstack.org/275440 | 21:13 |
| *** cloudtrainme has quit IRC | 21:19 | |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Replacing LBaaSv1 with v2 configuration https://review.openstack.org/275440 | 21:19 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-plugins: Created deprecation filter https://review.openstack.org/286250 | 21:20 |
| *** cloudtrainme has joined #openstack-ansible | 21:21 | |
| *** ametts has joined #openstack-ansible | 21:22 | |
| openstackgerrit | Marcos Simental proposed openstack/openstack-ansible-plugins: define login_project_name key on keystone library. https://review.openstack.org/286287 | 21:24 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Replacing LBaaSv1 with v2 configuration https://review.openstack.org/275440 | 21:25 |
| *** johnmilton has quit IRC | 21:39 | |
| admin0 | cloudnull: it works :D | 21:49 |
| *** fawadkhaliq has joined #openstack-ansible | 21:50 | |
| *** retreved has quit IRC | 21:53 | |
| *** ametts has quit IRC | 21:56 | |
| *** kencjohnston has quit IRC | 22:03 | |
| *** sdake has joined #openstack-ansible | 22:05 | |
| *** woodard has quit IRC | 22:12 | |
| *** phschwartz_ is now known as phschwartz | 22:14 | |
| *** sdake has quit IRC | 22:14 | |
| *** rohanp has joined #openstack-ansible | 22:18 | |
| *** sdake has joined #openstack-ansible | 22:18 | |
| rohanp | Hi, i had a question regarding using local pypi mirror for installing custom made neutron packages; how do we point OSA to use a local pypi mirror hosting custom packages to install neutron packages in neutron venv | 22:19 |
| *** cloudtrainme has quit IRC | 22:20 | |
| *** ametts has joined #openstack-ansible | 22:21 | |
| automagically | rohanp: I believe you would set repo_build_pip_extra_index | 22:22 |
| *** cemmason has quit IRC | 22:23 | |
| *** cemmason1 has joined #openstack-ansible | 22:23 | |
| rohanp | <automagically>: by setting this; OSA will pull packages from this link after it failed to find it using "repo_build_pip_default_index:" right? | 22:24 |
| automagically | So it appears | 22:24 |
| rohanp | so what i need to do is set the extra_index to my local pypi mirror | 22:24 |
| rohanp | makes sense | 22:24 |
| automagically | rohanp: Hope that works out for you | 22:25 |
| rohanp | is this the only place where we can set this? or is there any other option available too that could be used? | 22:25 |
| *** cemmason has joined #openstack-ansible | 22:27 | |
| *** ametts has quit IRC | 22:27 | |
| *** cemmason1 has quit IRC | 22:27 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts: Only update apt cache if necessary https://review.openstack.org/280233 | 22:31 |
| automagically | rohanp: I believe your other option would be to forego the OSA repo altogether, and specify your own repository using openstack_repo_url as indicated here: https://github.com/openstack/openstack-ansible/blob/master/doc/source/install-guide/configure-hostlist.rst | 22:31 |
| rohanp | <automagically>: Thanks for your help and pointers! | 22:38 |
| automagically | no problem, hope it helps | 22:38 |
| pabelanger | odyssey4me: sorry, was busy fighting the spam on wiki.o.o today. Will look later this evening | 22:45 |
| *** sdake has quit IRC | 22:46 | |
| *** phiche has quit IRC | 22:58 | |
| *** KLevenstein has quit IRC | 22:59 | |
| *** sdake has joined #openstack-ansible | 23:03 | |
| *** cloudtrainme has joined #openstack-ansible | 23:05 | |
| *** weezS has quit IRC | 23:09 | |
| *** phalmos has quit IRC | 23:10 | |
| *** admin0 has quit IRC | 23:10 | |
| *** cemmason has quit IRC | 23:14 | |
| *** gregfaust has quit IRC | 23:16 | |
| *** spotz is now known as spotz_zzz | 23:24 | |
| openstackgerrit | Merged openstack/openstack-ansible: Resolve packaging problems in virtualenv(s) https://review.openstack.org/284247 | 23:32 |
| *** Mudpuppy has quit IRC | 23:36 | |
| *** ShannonM has quit IRC | 23:39 | |
| *** weezS has joined #openstack-ansible | 23:43 | |
| *** rohanp has quit IRC | 23:44 | |
| *** jthorne has quit IRC | 23:48 | |
| *** jmckind has joined #openstack-ansible | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!