| *** chyka has joined #openstack-ansible | 00:06 | |
| *** chyka has quit IRC | 00:10 | |
| *** weezS has quit IRC | 00:16 | |
| *** acormier has joined #openstack-ansible | 00:28 | |
| *** acormier has quit IRC | 00:31 | |
| *** dxiri has quit IRC | 00:32 | |
| *** dxiri_ has joined #openstack-ansible | 00:32 | |
| *** acormier has joined #openstack-ansible | 00:32 | |
| *** woodard_ has joined #openstack-ansible | 00:47 | |
| *** woodard has quit IRC | 00:47 | |
| *** markvoelker has joined #openstack-ansible | 01:01 | |
| *** chas has joined #openstack-ansible | 01:07 | |
| *** chas has quit IRC | 01:12 | |
| *** gouthamr has quit IRC | 01:37 | |
| *** kenperkins has quit IRC | 02:13 | |
| *** acormier has quit IRC | 02:50 | |
| *** acormier has joined #openstack-ansible | 02:50 | |
| *** acormier has quit IRC | 02:55 | |
| *** MasterOfBugs has quit IRC | 02:56 | |
| *** gfa has joined #openstack-ansible | 03:00 | |
| *** chas has joined #openstack-ansible | 03:08 | |
| *** chas has quit IRC | 03:13 | |
| captaindave | Hey all. I'm trying to set an openstack lab on a sufficiently-beefy single physical machine where the hosts are centos VMs. I believe that I have all the requisite connectivity and I've run setup-hosts without error, but I still have all null values for all the containers' "ansible_host". Here's 'ip addr' from all the vhosts: http://sprunge.us/NMcA Here's my openstack_user_config.yml: | 03:18 |
|---|---|---|
| captaindave | http://sprunge.us/aCcJ And here's my user_variables.yml:http://sprunge.us/KafI Does anyone know what I'm doing wrong? logan- had said it was probably something with my network layout, but I've gone through it with a pretty fine-toothed comb (and plenty of tcpdump), and I'm at a loss. | 03:18 |
| captaindave | This is on Pike and the exact error encountered is "unsupported operand type(s) for +: 'NoneType' and 'str'" in 'Create haproxy service config files', where the '+' in question appears to be, e,g., 'entry.append(hostvars[host_name]['ansible_host'] + \":\" + haproxy_backend_port | string)' | 03:31 |
| *** gkadam has joined #openstack-ansible | 03:39 | |
| cloudnull | evenings | 03:39 |
| cloudnull | captaindave: looking | 03:40 |
| captaindave | And to you. | 03:40 |
| captaindave | Thanks! | 03:40 |
| captaindave | I appreciate any direction you can give. | 03:40 |
| *** udesale has joined #openstack-ansible | 03:44 | |
| cloudnull | everything looks in order. | 03:45 |
| captaindave | Hmm. | 03:45 |
| cloudnull | you might want to add tunnel_bridge: "br-vxlan" to the global_overrides section | 03:46 |
| captaindave | Should the vlans appear tagged on the hosts' bridges? | 03:46 |
| captaindave | Or bare (vlan 0)? | 03:46 |
| cloudnull | have you tried `rm openstack_inventory.json` | 03:46 |
| captaindave | I have not. | 03:46 |
| captaindave | I certainly can. | 03:46 |
| cloudnull | maybe something is wedged in the inventory from previous tries ? | 03:47 |
| cloudnull | if you've not provisioned any containers its safe to nuke it | 03:47 |
| cloudnull | it will regenerate | 03:47 |
| captaindave | I'll try it out. I'll destroy the containers first. | 03:47 |
| cloudnull | from the playbooks directory you can destroy them all with the lxc-container-destory.yml play | 03:48 |
| captaindave | We're well aquainted. ;) | 03:48 |
| cloudnull | ha! | 03:48 |
| captaindave | I've been at this for a while. | 03:48 |
| cloudnull | the stacktrace is when your running the haproxy play ? | 03:49 |
| captaindave | Yep. | 03:49 |
| captaindave | It's coming from one of the jinja templates. | 03:49 |
| *** pmannidi has quit IRC | 03:50 | |
| cloudnull | once you remove the openstack_inventory.json file you can just run the haproxy playbook to see if its still failing | 03:52 |
| captaindave | OK. | 03:52 |
| captaindave | Recreating. . . | 03:52 |
| captaindave | Should I remove the file before or after I recreate the containers? | 03:53 |
| cloudnull | no need to recreate the containers. | 03:53 |
| captaindave | Oh? | 03:53 |
| cloudnull | haproxy normally installs on the host w/out a container . | 03:53 |
| cloudnull | the inventory however will regenerate | 03:54 |
| captaindave | I thought it was going to look for the containers' ansible_host values (the problem at hand). | 03:54 |
| captaindave | Or are they independent: existence and definition? | 03:54 |
| cloudnull | anytime you run the playbook the inventory json file will be regenerated. | 03:55 |
| captaindave | That explains the changing mtime. | 03:55 |
| cloudnull | if you ever want to revert to an older inventory there's a running archive of old version s | 03:56 |
| captaindave | Oh? That's pretty kickin'! Where is that? | 03:57 |
| cloudnull | openstack_inventory.tar | 03:57 |
| cloudnull | in the /etc/openstack_deploy dir | 03:58 |
| captaindave | huh! | 03:58 |
| captaindave | Nope. Same error. | 03:58 |
| captaindave | http://sprunge.us/LQOh | 03:59 |
| cloudnull | hum. | 04:00 |
| * cloudnull looking | 04:01 | |
| captaindave | http://sprunge.us/TLdS | 04:01 |
| logan- | weird that dynamic inventory isnt assigning ips to a bunch of the containers | 04:05 |
| captaindave | Is there any sort of overlap I should be concerned about? | 04:05 |
| captaindave | We're on a 10.0.0.0/8 and I note that the containers are on 10.3.0.0/16(?), but it's NAT, so it shouldn't matter if it's talking just to its subnets and the off-subnet (but non-overlapping) DNS servers. | 04:07 |
| cloudnull | ok its something in the user config, i can reproduce it | 04:07 |
| captaindave | And the world, of course. | 04:07 |
| cloudnull | just looking for what ... | 04:07 |
| logan- | used_ips maybe? | 04:08 |
| captaindave | Yep! | 04:09 |
| captaindave | I left out a digit. | 04:10 |
| logan- | ah yep | 04:10 |
| logan- | 1.x | 04:10 |
| captaindave | I can see that causing issues. | 04:10 |
| captaindave | So with that corrected, rm the inventory and try again, or do I need to re-set up the hosts? | 04:11 |
| cloudnull | just to test you can simply run the haproxy playbook | 04:12 |
| captaindave | Sure thing. | 04:12 |
| captaindave | Same error. inventory still has null for the ansible_host. | 04:15 |
| captaindave | I also have the networks wrong. | 04:16 |
| cloudnull | pabelanger: http://paste.openstack.org/show/622595/ | 04:17 |
| cloudnull | opps captaindave | 04:17 |
| cloudnull | the indentation needs to be adjusted | 04:18 |
| cloudnull | with the indentation updated its working | 04:19 |
| *** dxiri_ has quit IRC | 04:20 | |
| *** dxiri has joined #openstack-ansible | 04:21 | |
| captaindave | Double checking my work . . . . | 04:23 |
| cloudnull | I took your user config and tested with it, all i had to change was here http://paste.openstack.org/show | 04:24 |
| cloudnull | everything else seems to be ok | 04:24 |
| captaindave | I'm rechecking VLANs and networks. | 04:24 |
| captaindave | the numbering. | 04:24 |
| captaindave | Looks like it worked. | 04:28 |
| captaindave | ! | 04:28 |
| cloudnull | woot! | 04:28 |
| captaindave | Thank you so much! This was driving me batty. | 04:30 |
| captaindave | I swear I reconfigured that network (Cisco UCS, so ugly but pretty quick) 10-15 times along the way. | 04:30 |
| cloudnull | sounds like a blast ! | 04:31 |
| captaindave | Thus the messed up network numbers. | 04:31 |
| captaindave | It's been pretty fun, actually. | 04:31 |
| cloudnull | lots of good practice :D | 04:32 |
| cloudnull | had you seen https://github.com/openstack/openstack-ansible-ops/tree/master/multi-node-aio? | 04:32 |
| captaindave | lots of use of parallel and creatitve quoting in bash for the vhost deployments. | 04:32 |
| cloudnull | you mentioned building this on a beefy machine with VMs | 04:32 |
| cloudnull | that only supports ubuntu atm | 04:32 |
| cloudnull | but adding cent shouldn't been too much of an issue | 04:33 |
| captaindave | Yeah. I played around with the single node aio, but I wanted to use centos and to use our real network gear. | 04:33 |
| cloudnull | makes sense. | 04:33 |
| captaindave | I'm planning on putting a second one of these in our other dc as phase 2. | 04:33 |
| cloudnull | cool ! | 04:33 |
| captaindave | And then they can talk back and forth. | 04:33 |
| captaindave | I'm hoping to make this poc/lab into our transition away from vsphere. | 04:34 |
| captaindave | One blade at a time. | 04:34 |
| cloudnull | very cool! | 04:34 |
| captaindave | We shall see... | 04:34 |
| *** galstrom_zzz is now known as galstrom | 04:35 | |
| captaindave | Thanks again, everyone. I'm exhausted and am going to press reset on this and go to bed. | 04:35 |
| cloudnull | take care | 04:35 |
| cloudnull | let us know how it goes ? | 04:36 |
| captaindave | You too. And thank you for the project, btw. | 04:36 |
| captaindave | Will do! | 04:36 |
| cloudnull | thanks! | 04:36 |
| cloudnull | its a pretty great community. | 04:37 |
| captaindave | So I've gathered. | 04:37 |
| captaindave | :) | 04:37 |
| *** mudpuppy has quit IRC | 04:50 | |
| *** mudpuppy has joined #openstack-ansible | 04:50 | |
| *** chas has joined #openstack-ansible | 05:09 | |
| *** chas has quit IRC | 05:13 | |
| *** eumel8 has joined #openstack-ansible | 05:20 | |
| *** markvoelker has quit IRC | 05:20 | |
| *** eumel8 has quit IRC | 05:20 | |
| *** armaan has joined #openstack-ansible | 05:29 | |
| *** armaan has quit IRC | 05:30 | |
| *** armaan has joined #openstack-ansible | 05:30 | |
| *** chas has joined #openstack-ansible | 05:48 | |
| *** portante has quit IRC | 05:53 | |
| *** lhinds has quit IRC | 05:54 | |
| *** pabelanger has quit IRC | 05:56 | |
| *** albertcard1 has quit IRC | 06:01 | |
| *** albertcard1 has joined #openstack-ansible | 06:03 | |
| *** udesale__ has joined #openstack-ansible | 06:05 | |
| *** udesale has quit IRC | 06:06 | |
| *** Oku_OS-away is now known as Oku_OS | 06:08 | |
| *** ajmaidak has quit IRC | 06:15 | |
| *** portante has joined #openstack-ansible | 06:21 | |
| *** pabelanger has joined #openstack-ansible | 06:21 | |
| *** lhinds has joined #openstack-ansible | 06:21 | |
| *** jvidal has joined #openstack-ansible | 06:22 | |
| *** hachi has joined #openstack-ansible | 06:23 | |
| *** udesale__ has quit IRC | 06:25 | |
| *** udesale__ has joined #openstack-ansible | 06:25 | |
| *** ajmaidak has joined #openstack-ansible | 06:34 | |
| *** eumel8 has joined #openstack-ansible | 06:36 | |
| *** chyka has joined #openstack-ansible | 06:45 | |
| *** pcaruana has joined #openstack-ansible | 06:48 | |
| *** chyka has quit IRC | 06:50 | |
| *** udesale has joined #openstack-ansible | 06:51 | |
| *** chyka has joined #openstack-ansible | 06:52 | |
| *** udesale__ has quit IRC | 06:53 | |
| *** drifterza has joined #openstack-ansible | 06:56 | |
| *** chyka has quit IRC | 06:56 | |
| *** taseer1 has joined #openstack-ansible | 06:59 | |
| *** Taseer has quit IRC | 06:59 | |
| *** shardy_afk is now known as shardy | 07:01 | |
| *** udesale__ has joined #openstack-ansible | 07:02 | |
| *** udesale has quit IRC | 07:04 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_build master: Fix typo on conditional https://review.openstack.org/508928 | 07:09 |
| *** OldManTom has joined #openstack-ansible | 07:15 | |
| *** galstrom is now known as galstrom_zzz | 07:16 | |
| *** sdake_ has joined #openstack-ansible | 07:16 | |
| *** sdake_ is now known as Guest52279 | 07:17 | |
| *** Guest52279 has quit IRC | 07:17 | |
| *** taseer1 is now known as Taseer | 07:18 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [Docs] Clarify role development cycle https://review.openstack.org/504279 | 07:19 |
| *** markvoelker has joined #openstack-ansible | 07:21 | |
| *** mbuil has joined #openstack-ansible | 07:23 | |
| *** acormier has joined #openstack-ansible | 07:51 | |
| *** markvoelker has quit IRC | 07:55 | |
| *** acormier has quit IRC | 07:55 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] testing gates g-a-r-r https://review.openstack.org/509368 | 07:58 |
| *** udesale has joined #openstack-ansible | 08:02 | |
| *** udesale__ has quit IRC | 08:03 | |
| *** mrch has joined #openstack-ansible | 08:07 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] testing gates g-a-r-r https://review.openstack.org/509368 | 08:07 |
| *** pbandark has joined #openstack-ansible | 08:10 | |
| *** rodolof has joined #openstack-ansible | 08:10 | |
| *** dasTor has joined #openstack-ansible | 08:21 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-lxc_hosts master: Fix inconsistent cache prep timer https://review.openstack.org/509376 | 08:24 |
| *** ashak has quit IRC | 08:31 | |
| *** ashak has joined #openstack-ansible | 08:33 | |
| *** armaan has quit IRC | 08:36 | |
| *** armaan has joined #openstack-ansible | 08:36 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] testing gates g-a-r-r https://review.openstack.org/509368 | 08:38 |
| evrardjp | most of the builds that don't pass rn are due to IO contention | 08:46 |
| *** armaan has quit IRC | 08:49 | |
| *** armaan has joined #openstack-ansible | 08:49 | |
| *** markvoelker has joined #openstack-ansible | 08:52 | |
| *** hachi has quit IRC | 08:55 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Fix warning in upgrade https://review.openstack.org/509381 | 08:57 |
| *** electrofelix has joined #openstack-ansible | 08:57 | |
| *** OldManTom has quit IRC | 09:01 | |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: Ensure sysstat is running on CentOS https://review.openstack.org/504231 | 09:06 |
| evrardjp | FYI: I will fix the zuul cloner + depends on later today | 09:08 |
| admin0 | morning \o | 09:10 |
| admin0 | evrardjp, from the logs, all i have to do is change the listen address to 0.0.0.0 and LM will work out of the box ? | 09:11 |
| admin0 | i will try .. thanks guys for addressing that | 09:12 |
| *** gunix has joined #openstack-ansible | 09:18 | |
| *** cshen has joined #openstack-ansible | 09:18 | |
| *** cshen has left #openstack-ansible | 09:21 | |
| *** cshen has joined #openstack-ansible | 09:21 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build master: Fix typo on conditional https://review.openstack.org/508928 | 09:25 |
| *** TomMc has joined #openstack-ansible | 09:26 | |
| *** markvoelker has quit IRC | 09:26 | |
| *** hachi has joined #openstack-ansible | 09:27 | |
| evrardjp | admin0: it should work out of the box with qemu+ssh | 09:37 |
| *** bauruine has quit IRC | 09:41 | |
| *** bauruine has joined #openstack-ansible | 09:42 | |
| odyssey4me | evrardjp I'm guessing that https://review.openstack.org/508928 should be ported back ? | 09:43 |
| evrardjp | yes :p | 09:44 |
| evrardjp | You still got the eye | 09:44 |
| evrardjp | of the tiger | 09:44 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_build stable/pike: Fix typo on conditional https://review.openstack.org/509386 | 09:44 |
| evrardjp | I am proposing on all the branches | 09:44 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_build stable/ocata: Fix typo on conditional https://review.openstack.org/509387 | 09:44 |
| odyssey4me | yup | 09:45 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_build stable/newton: Fix typo on conditional https://review.openstack.org/509388 | 09:45 |
| evrardjp | :) | 09:45 |
| *** hachi has quit IRC | 09:46 | |
| *** hachi has joined #openstack-ansible | 09:46 | |
| evrardjp | afk lunch | 09:47 |
| *** DanyC has joined #openstack-ansible | 09:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Fix inconsistent cache prep timer https://review.openstack.org/509376 | 10:19 |
| *** markvoelker has joined #openstack-ansible | 10:23 | |
| odyssey4me | evrardjp ^ probably needs a port back to pike too :) | 10:23 |
| *** chyka has joined #openstack-ansible | 10:29 | |
| *** TomMc has quit IRC | 10:29 | |
| *** udesale has quit IRC | 10:32 | |
| *** chyka has quit IRC | 10:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: Updated from OpenStack Ansible Tests https://review.openstack.org/499656 | 10:34 |
| *** TomMc has joined #openstack-ansible | 10:42 | |
| *** hachi has quit IRC | 10:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/pike: Fix typo on conditional https://review.openstack.org/509386 | 10:51 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/newton: Fix typo on conditional https://review.openstack.org/509388 | 10:51 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/ocata: Fix typo on conditional https://review.openstack.org/509387 | 10:53 |
| *** markvoelker has quit IRC | 10:55 | |
| odyssey4me | nice to see patches moving again | 10:59 |
| *** dave-mccowan has joined #openstack-ansible | 11:01 | |
| *** dave-mcc_ has joined #openstack-ansible | 11:10 | |
| *** TomMc has quit IRC | 11:12 | |
| *** shardy is now known as shardy_afk | 11:12 | |
| *** hamzy has joined #openstack-ansible | 11:12 | |
| *** dave-mccowan has quit IRC | 11:13 | |
| mbuil | if I want to add some new lines to a file, the ansible module 'copy' is not helpful right? If the dst file exists, it will just overwrite it and there is no option to instruct ansible to add those lines to the bottom of the existing file, right? | 11:16 |
| wspthr | You could try with "blockinfile” or “lineinfile” | 11:17 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: scripts: bootstrap-ansible.sh: Do not error on 'alternatives' command https://review.openstack.org/509410 | 11:21 |
| mbuil | wspthr: thanks. Seems like blockinfile is what I was after. Unfortunately, I need to add ~30 lines, so the code might get a bit ugly with so much text after the block parameter, but that's how it is I guess | 11:22 |
| hwoarang | mbuil: is it too much pain to have your own copy of the file constructed as you wish? | 11:22 |
| jrosser | is there a way to run one of the OSA role tests locally, just the functional bit? | 11:25 |
| *** smatzek has joined #openstack-ansible | 11:25 | |
| hwoarang | jrosser: maybe edit the Vagrantfile to run 'run_tests.sh functional' ? | 11:26 |
| hwoarang | and then vagrant up <distro> ? | 11:26 |
| odyssey4me | jrosser or if you're on a host which can just run the stuff, use 'run_tests.sh functional' | 11:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-lxc_hosts stable/pike: Fix inconsistent cache prep timer https://review.openstack.org/509414 | 11:36 |
| *** DanyC has quit IRC | 11:38 | |
| jrosser | odyssey4me: oh wow - just finally found how that works :) | 11:38 |
| *** TomMc has joined #openstack-ansible | 11:38 | |
| odyssey4me | jrosser you can thanks hwoarang for making it work so nicely ;) | 11:39 |
| evrardjp | jrosser: :) | 11:39 |
| evrardjp | Maybe we should stop tox altogether | 11:39 |
| jrosser | you have to read run_tests_common.sh very very carefully | 11:39 |
| odyssey4me | evrardjp given zuul v3 I think that would be fine - we used it previously because we were trying to be less of a snowflake | 11:40 |
| evrardjp | odyssey4me: yes, that's what I thought. | 11:40 |
| odyssey4me | but we should use tox for some standard things, like docs building and all that | 11:40 |
| evrardjp | simplifying things up a little more is good | 11:40 |
| evrardjp | oh yes, sure. | 11:40 |
| odyssey4me | but for all our functional tests, we can kill tox I think | 11:41 |
| evrardjp | but these jobs already exist | 11:41 |
| evrardjp | and IIRC, there was this idea of putting more things directly in docs/ folder | 11:41 |
| evrardjp | but I have to re-read mordred email about it | 11:41 |
| odyssey4me | yes, all docs are moving into the docs/ folder, but there will still likely be multiple tox nvs I think | 11:42 |
| pbandark | any guide which i can refer to manage existing openstack setup using ansible? for ex. creating network/router/instances etc | 11:42 |
| hwoarang | odyssey4me: evrardjp: regarding our discussion about user_*.yml files being passed exclusively using -e, is it ok for deployers to generally drop user_*.yml files there to create different deployment flavors? For example user_scenario1.yml, user_scenario2.yml etc | 11:42 |
| hwoarang | so they can keep a minimal user_variables.yml (since it's last) and then configure the scenarion in the other files | 11:42 |
| evrardjp | well, by default we add all the files in the cli | 11:42 |
| evrardjp | not sure what you mean there | 11:42 |
| evrardjp | why not doing it manually then? | 11:43 |
| odyssey4me | hwoarang absolutely - any number of user_ files can be used and layered as you see fit... they will be applied in alphanumeric order with the last one winning | 11:43 |
| hwoarang | i mean, that any user_* file will be passed to command line | 11:43 |
| *** drifterza has quit IRC | 11:43 | |
| evrardjp | -e @scenario.yml on top of the user_* files should be good enough | 11:43 |
| evrardjp | well you modify it the way you want :) | 11:43 |
| evrardjp | that's standard ansible | 11:43 |
| odyssey4me | so if you define var1 in user_variables.yml, and var1 again in user_zz_variables.yml, the value will come from the one in the last file | 11:43 |
| hwoarang | yeap that's what i meant | 11:43 |
| hwoarang | interesting feature for opnfv | 11:44 |
| evrardjp | keep in mind you're probably gonna have a warning | 11:44 |
| * hwoarang context switches back to opnfv | 11:44 | |
| evrardjp | but it's fine | 11:44 |
| hwoarang | ah i can live with that | 11:44 |
| evrardjp | there are maybe smarter things to do inside yaml | 11:45 |
| evrardjp | but I don't know your use case in details :p | 11:45 |
| hwoarang | evrardjp: just some scenarios want to override some of the user_variables.yml we give them. so we want some sort of layering | 11:47 |
| hwoarang | we provide a good default user_variables.yml for all deployments, and then scenarios may supply their own user_* file to change some stuff | 11:48 |
| evrardjp | fair enough | 11:48 |
| *** drifterza has joined #openstack-ansible | 11:50 | |
| *** markvoelker has joined #openstack-ansible | 11:53 | |
| admin0 | one comment ..not a log of defaults are in user_variables .. so most of the time, deployers like me are not even sure what are the default values .. is it possible to grep/awk all defaults used everywhere in openstack ? | 12:10 |
| admin0 | that can be overwritten using user_variables | 12:10 |
| mhayden | good morning, folks | 12:16 |
| odyssey4me | admin0 any conf options can be set by OSA using the config_overrides... there is no need to ever learn the variable in OSA... it may be helpful sometimes, but is not required | 12:19 |
| odyssey4me | and yes, openstack has a config reference doc to help you find all the possible options that can be set | 12:19 |
| odyssey4me | admin0 as clearly stated at the top of the user_variables.yml default file, it is not meant to contain every option | 12:20 |
| *** udesale has joined #openstack-ansible | 12:23 | |
| mgariepy | morning everyone | 12:23 |
| *** ianychoi has quit IRC | 12:25 | |
| *** ianychoi has joined #openstack-ansible | 12:26 | |
| *** markvoelker has quit IRC | 12:26 | |
| *** markvoelker has joined #openstack-ansible | 12:27 | |
| *** br0k3n has joined #openstack-ansible | 12:27 | |
| *** TomMc has quit IRC | 12:27 | |
| *** smatzek has quit IRC | 12:27 | |
| *** smatzek has joined #openstack-ansible | 12:28 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Fix zuul clonemap https://review.openstack.org/509426 | 12:31 |
| evrardjp | mhayden: ^ more fix up of your things :) | 12:33 |
| *** smatzek has quit IRC | 12:34 | |
| evrardjp | could I have some votes on ^ it should be helpful | 12:38 |
| *** br0k3n has quit IRC | 12:41 | |
| *** kylek3h has quit IRC | 12:43 | |
| *** gkadam has quit IRC | 12:50 | |
| *** esberglu has joined #openstack-ansible | 12:51 | |
| *** smatzek has joined #openstack-ansible | 12:53 | |
| admin0 | odyssey4me, that i know .. but say i want to change the default az from nova to something else.. i have to resort to commands like grep -ri ': nova' /opt/openstack-ansible | grep vars .. and grep -ri ': nova' /etc/ansible | grep defaults to check where that might be coming from and if it can be set via osa | 12:54 |
| odyssey4me | admin0 or just use the config override and the config reference from nova | 12:54 |
| odyssey4me | the OSA vars are only useful if they're applying glue between services, and yes - then you need to grep for them... unless we pull all roles into the same directory that problem won't go away - and even then, you'll still have to grep | 12:55 |
| odyssey4me | unless you have some smart way of auto-documenting variables? | 12:56 |
| *** TomMc has joined #openstack-ansible | 12:56 | |
| *** woodard_ has quit IRC | 12:56 | |
| *** woodard has joined #openstack-ansible | 12:57 | |
| *** pcaruana has quit IRC | 13:02 | |
| *** smatzek has quit IRC | 13:04 | |
| *** lbragstad has joined #openstack-ansible | 13:05 | |
| odyssey4me | evrardjp that only applies to non OSA repositories and roles - what's the problem that's trying to be solved? | 13:05 |
| *** kylek3h has joined #openstack-ansible | 13:08 | |
| evrardjp | it does indeed | 13:08 |
| evrardjp | http://logs.openstack.org/68/509368/3/check/gate-openstack-ansible-linters-ubuntu-xenial/49dc15b/console.html#_2017-10-04_08_42_25_258465 | 13:08 |
| evrardjp | it's just wrong | 13:09 |
| evrardjp | and burps when doing releases | 13:09 |
| evrardjp | which is not osa | 13:09 |
| odyssey4me | fair enough | 13:10 |
| odyssey4me | I don't think we actually need to negate there, because zuul uses the first it finds and ignores the others IIRC... but sure that patch looks fine | 13:10 |
| evrardjp | :) | 13:10 |
| evrardjp | not suer what you meant there | 13:11 |
| evrardjp | sure* | 13:11 |
| evrardjp | oh yes | 13:11 |
| *** smatzek has joined #openstack-ansible | 13:11 | |
| *** pbandark1 has joined #openstack-ansible | 13:11 | |
| evrardjp | although I have just checked the re.sub part of the clone map | 13:12 |
| evrardjp | it may be a dict instead of an OL | 13:12 |
| evrardjp | OD* | 13:12 |
| evrardjp | let me check just a sec | 13:12 |
| *** pbandark has quit IRC | 13:13 | |
| *** pbandark1 is now known as pbandark | 13:13 | |
| evrardjp | yup it's ordered | 13:14 |
| evrardjp | I can change the review if you're more confident that way | 13:14 |
| *** pcaruana has joined #openstack-ansible | 13:15 | |
| *** smatzek has quit IRC | 13:15 | |
| cloudnull | mornings | 13:16 |
| *** hachi__ has joined #openstack-ansible | 13:22 | |
| *** eumel8 has quit IRC | 13:23 | |
| jrosser | cloudnull hi | 13:26 |
| jrosser | i was looking at your machinectl/btrfs patches | 13:26 |
| jrosser | i wasnt clear how it interacted with tasks calling the lxc_container module that have to specify the backing store type | 13:27 |
| cloudnull | the template moves the container store to /var/lib/machinectl | 13:31 |
| cloudnull | so when using that template it overrides the bdir options | 13:32 |
| jrosser | here for example https://github.com/openstack/openstack-ansible-lxc_hosts/blob/de1b45553e856ac618562476095e4023c8f627be/tasks/lxc_cache_create.yml#L79 | 13:32 |
| cloudnull | that would not be run | 13:33 |
| cloudnull | or should not be | 13:33 |
| cloudnull | the container is spawned directly from the cache | 13:34 |
| cloudnull | so there's almost no spin up time | 13:34 |
| jrosser | i was trying to decide what the delta was to get zfs backend going | 13:35 |
| *** TomMc has quit IRC | 13:37 | |
| cloudnull | oh. | 13:37 |
| cloudnull | it looks like zfs is already there. | 13:37 |
| cloudnull | we'd need to set backendstore to zfs | 13:38 |
| jrosser | cloudnull: reason i ask is we've just torn down our POC and are rebuilding it | 13:38 |
| cloudnull | and provide the pool location | 13:38 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: make group membership blocks optional https://review.openstack.org/508269 | 13:38 |
| jrosser | and ZFS backed lxc has been on the to-do for a looong time | 13:38 |
| jrosser | so i have a window to address that now, and it might be hard to change that for a while otherwise | 13:39 |
| cloudnull | yea so --bdir zfs --zfsroot=$PATH | 13:40 |
| cloudnull | and it should work ? | 13:40 |
| cloudnull | sorry bdir --bdev | 13:40 |
| jrosser | and it felt like there were two possible approaches - the one in your patches or extend the existing lxc_hosts to do dir|lvm|zfs | 13:40 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_freezer master: Updated from global requirements https://review.openstack.org/484120 | 13:41 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_horizon master: Updated from global requirements https://review.openstack.org/504683 | 13:41 |
| mgariepy | zfs backend for lxc containers would be really awesome :) | 13:41 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_searchlight master: Updated from global requirements https://review.openstack.org/495238 | 13:42 |
| *** shardy_afk is now known as shardy | 13:42 | |
| *** TomMc has joined #openstack-ansible | 13:42 | |
| *** jwitko has joined #openstack-ansible | 13:44 | |
| cloudnull | jrosser: so http://docs.ansible.com/ansible/latest/lxc_container_module.html | 13:44 |
| cloudnull | you'd need to set zfs_root | 13:45 |
| cloudnull | and the module should do the rest | 13:45 |
| jrosser | ok - and adding a 3rd section to lxc_cache_create.yml to handle zfs? | 13:46 |
| cloudnull | you could probably just add it here https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/container_create.yml#L64-L85 | 13:46 |
| cloudnull | that option is not used unless the backendstore is zfs | 13:47 |
| jrosser | ah yes my brain did bleed a bit with the when: on that | 13:47 |
| jrosser | but if the option is ignores then it doesnt matter | 13:48 |
| *** weezS has joined #openstack-ansible | 13:49 | |
| *** gouthamr has joined #openstack-ansible | 13:53 | |
| *** armaan has quit IRC | 13:55 | |
| *** armaan has joined #openstack-ansible | 13:55 | |
| *** germs has joined #openstack-ansible | 13:57 | |
| odyssey4me | jrosser cloudnull yeah, sorry about that horrible conditional | 14:00 |
| odyssey4me | I was trying to cater for defining a back-end like lvm, but enabling the use of CoW containers (ie lvm snapshots) | 14:00 |
| odyssey4me | but then you get weird ones like overlayfs which is CoW :/ | 14:01 |
| *** germs has quit IRC | 14:01 | |
| odyssey4me | I'd prefer to actually simplify that down to be more consistent and less complex - perhaps something like: backing store = dir, if backing_method = copy-on-write then use overlayfs... for zfs as a backing store, if CoW enabled then use it... etc | 14:02 |
| evrardjp | or maybe backing store, and we assume the capability of it, unless someone overrides it in user vars | 14:07 |
| evrardjp | (capabilities map would be in a vars/ file ) | 14:07 |
| *** yifei has joined #openstack-ansible | 14:12 | |
| odyssey4me | yes, if we do a capabilities map that could be quite nice | 14:12 |
| *** smatzek has joined #openstack-ansible | 14:15 | |
| *** smatzek has quit IRC | 14:15 | |
| *** smatzek has joined #openstack-ansible | 14:15 | |
| *** kenperkins has joined #openstack-ansible | 14:18 | |
| *** acormier has joined #openstack-ansible | 14:19 | |
| *** acormier has quit IRC | 14:20 | |
| *** kenperkins has quit IRC | 14:23 | |
| *** armaan has quit IRC | 14:30 | |
| *** armaan has joined #openstack-ansible | 14:30 | |
| *** udesale has quit IRC | 14:32 | |
| *** pcaruana has quit IRC | 14:33 | |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-repo_server master: Only enable apt-cacher-ng from epel-testing https://review.openstack.org/509493 | 14:33 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_tempest master: Enforce tempest folder only when it doesn't exist https://review.openstack.org/509494 | 14:35 |
| *** armaan has quit IRC | 14:39 | |
| *** fxpester has quit IRC | 14:40 | |
| *** armaan has joined #openstack-ansible | 14:40 | |
| *** galstrom_zzz is now known as galstrom | 14:42 | |
| *** smatzek has quit IRC | 14:42 | |
| *** smatzek has joined #openstack-ansible | 14:43 | |
| *** smatzek has quit IRC | 14:44 | |
| *** pcaruana has joined #openstack-ansible | 14:45 | |
| spotz | cloudnull - did you doc the trusty Newton to Xenial Ocata upgrade? I know if you're using your scripts you stop at Newton to upgrade OS and then I think proceed? | 14:48 |
| spotz | email on the ops list | 14:48 |
| mgariepy | spotz, there is that for the newton - trusty/xenial update https://etherpad.openstack.org/p/osa-newton-xenial-upgrade | 14:53 |
| spotz | Sweet thanks mgariepy I'll reply to the email with that | 14:54 |
| *** smatzek has joined #openstack-ansible | 14:55 | |
| mgariepy | it's not the newton to ocata tho. :) | 14:56 |
| *** eumel8 has joined #openstack-ansible | 14:57 | |
| *** germs has joined #openstack-ansible | 14:57 | |
| *** smatzek_ has joined #openstack-ansible | 14:57 | |
| *** jvidal has quit IRC | 14:57 | |
| *** germs1 has joined #openstack-ansible | 14:58 | |
| *** smatzek has quit IRC | 15:00 | |
| *** smatzek_ is now known as smatzek | 15:00 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: L3 Support for ODL deployments https://review.openstack.org/497458 | 15:00 |
| *** germs has quit IRC | 15:01 | |
| *** yifei has quit IRC | 15:03 | |
| *** Oku_OS is now known as Oku_OS-away | 15:08 | |
| *** armaan_ has joined #openstack-ansible | 15:11 | |
| *** armaan has quit IRC | 15:11 | |
| evrardjp | spotz: it's a two step thing. N Trusty -> Xenial then N Xenial -> O | 15:18 |
| spotz | evrardjp: That part I knew and what they're planning to do. I invited them to join us here:) | 15:19 |
| evrardjp | good choice | 15:19 |
| *** TomMc has quit IRC | 15:20 | |
| spotz | Oh evrardjp horse shoowing this weekend so won't be about tomorrow for the meeting, I"ll be driving the trailer. It does look like there will be onboarding rooms at summit | 15:27 |
| *** pcaruana has quit IRC | 15:28 | |
| *** chas has quit IRC | 15:31 | |
| *** smatzek has quit IRC | 15:38 | |
| *** smatzek has joined #openstack-ansible | 15:39 | |
| *** br0k3n has joined #openstack-ansible | 15:40 | |
| *** smatzek_ has joined #openstack-ansible | 15:43 | |
| *** smatzek has quit IRC | 15:43 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Check proxy status on CentOS https://review.openstack.org/509515 | 15:45 |
| *** chyka has joined #openstack-ansible | 15:45 | |
| *** smatzek_ has quit IRC | 15:46 | |
| evrardjp | spotz: I'll handle the community meeting don't worry | 15:46 |
| spotz | evrardjp: I know you've already cut me back to once a month! | 15:47 |
| evrardjp | haha | 15:47 |
| evrardjp | why? | 15:47 |
| evrardjp | you can handle the bug triages too | 15:47 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Allow for Leapfrogs from other Releases https://review.openstack.org/507645 | 15:47 |
| evrardjp | I see no issue with that | 15:47 |
| *** hachi__ has quit IRC | 15:48 | |
| *** smatzek has joined #openstack-ansible | 15:49 | |
| *** smatzek_ has joined #openstack-ansible | 15:51 | |
| *** smatzek has quit IRC | 15:54 | |
| *** smatzek_ has quit IRC | 15:54 | |
| *** br0k3n has quit IRC | 15:55 | |
| *** armaan_ has quit IRC | 15:56 | |
| *** armaan has joined #openstack-ansible | 15:57 | |
| *** smatzek has joined #openstack-ansible | 15:57 | |
| *** mrch has quit IRC | 16:00 | |
| *** smatzek_ has joined #openstack-ansible | 16:00 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/pike: Fix inconsistent cache prep timer https://review.openstack.org/509414 | 16:01 |
| *** smatzek has quit IRC | 16:01 | |
| evrardjp | so, for ppl reviewing, don't forget to also check the -1 zuuls , because now we have jenkins and zuul | 16:02 |
| *** smatzek_ has quit IRC | 16:03 | |
| evrardjp | example https://review.openstack.org/#/c/509108/ | 16:03 |
| *** smatzek has joined #openstack-ansible | 16:03 | |
| *** smatzek has quit IRC | 16:03 | |
| *** smatzek has joined #openstack-ansible | 16:04 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: vars: Drop development packages https://review.openstack.org/507449 | 16:05 |
| *** woodard has quit IRC | 16:08 | |
| *** woodard_ has joined #openstack-ansible | 16:08 | |
| eumel8 | is there any "big picture" in osa to use ansible roles outside the repo? | 16:09 |
| eumel8 | maybe a spec or blueprint | 16:09 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Update the LXC container create role https://review.openstack.org/509530 | 16:11 |
| cloudnull | jrosser: odyssey4me: first pass at trying to update the lxc create role. | 16:11 |
| cloudnull | hoping to pull it more in line with our other roles and increase general role comprehension. | 16:12 |
| *** armaan has quit IRC | 16:14 | |
| *** armaan has joined #openstack-ansible | 16:14 | |
| odyssey4me | logan- uses that role heavily, so he'll want to peek at it | 16:14 |
| *** armaan has quit IRC | 16:15 | |
| *** armaan has joined #openstack-ansible | 16:15 | |
| nwonknu | anyone know of a way to deploy multiple cinder containers to a single host through osa? | 16:16 |
| nwonknu | use-case is multiple cinder availability zones on a single host (referring to 3par backend storage in different locations, but network-wise available on the same controller node) | 16:16 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Update the LXC container create role https://review.openstack.org/509530 | 16:17 |
| nwonknu | as cinder az's seem to be defined on host-level, not sure how else i'd configure something like this other than having multiple cinders | 16:17 |
| *** rodolof has quit IRC | 16:20 | |
| odyssey4me | cloudnull added a comment in PS1, otherwise looks fine I think at a glance | 16:23 |
| cloudnull | comment makes total sense | 16:26 |
| cloudnull | updating | 16:26 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: templates: Configure galera socket per distribution https://review.openstack.org/508154 | 16:37 |
| *** dxiri has quit IRC | 16:40 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Update the LXC container create role https://review.openstack.org/509530 | 16:46 |
| *** shardy has quit IRC | 16:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Add support for Neutron FWaaS v2 https://review.openstack.org/388548 | 16:52 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron stable/pike: Add support for Neutron FWaaS v2 https://review.openstack.org/509551 | 16:55 |
| *** DanyC has joined #openstack-ansible | 16:58 | |
| jrosser | cloudnull: excellent stuff... boxrick and i were just saing earlier that the container_create role needed splitting into more files like that | 17:01 |
| cloudnull | ++ | 17:01 |
| cloudnull | totally agre | 17:01 |
| cloudnull | *agree | 17:01 |
| *** mrtenio has quit IRC | 17:03 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Update the LXC container create role https://review.openstack.org/509530 | 17:04 |
| cloudnull | updated the release notes ^ | 17:04 |
| jrosser | need some consideration of the copy-on-write flag | 17:04 |
| ivve | nwonknu: you can have multiple backends to one host | 17:05 |
| jrosser | zfs is implicitly cow, but as the code was before you couldnt set the flag otherwise $bad-things looked likley | 17:05 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: adding multiple ovs manager https://review.openstack.org/496580 | 17:06 |
| *** mrtenio has joined #openstack-ansible | 17:06 | |
| ivve | nwonknu: you just put all of them in the cinder stanza in openstack_user_config.yml. would you like some examples? | 17:07 |
| jrosser | ah right i see now copy-on-write flag is specifically for lvm, that makes more sense now | 17:07 |
| *** galstrom is now known as galstrom_zzz | 17:10 | |
| *** mbuil has quit IRC | 17:19 | |
| *** weezS has quit IRC | 17:25 | |
| cloudnull | jrosser: yea with the updated layout we should be able to easily integrated additional stores | 17:27 |
| *** galstrom_zzz is now known as galstrom | 17:27 | |
| cloudnull | I added a zfs stub which just returns "not implemented" | 17:28 |
| cloudnull | but with the new layout adding something zfs specific without having to deal with all of the extra conditional logic should be a lot more simple. | 17:29 |
| *** markvoelker has quit IRC | 17:30 | |
| *** markvoelker has joined #openstack-ansible | 17:30 | |
| *** chas has joined #openstack-ansible | 17:32 | |
| *** dxiri has joined #openstack-ansible | 17:33 | |
| *** lbragstad has quit IRC | 17:34 | |
| *** chas has quit IRC | 17:37 | |
| *** smatzek has quit IRC | 17:37 | |
| *** smatzek has joined #openstack-ansible | 17:38 | |
| *** smatzek_ has joined #openstack-ansible | 17:40 | |
| *** esberglu has quit IRC | 17:41 | |
| *** smatzek has quit IRC | 17:42 | |
| *** esberglu has joined #openstack-ansible | 17:42 | |
| *** esberglu has quit IRC | 17:42 | |
| *** smatzek_ has quit IRC | 17:42 | |
| *** esberglu has joined #openstack-ansible | 17:42 | |
| *** smatzek has joined #openstack-ansible | 17:44 | |
| *** smatzek_ has joined #openstack-ansible | 17:46 | |
| *** smatzek__ has joined #openstack-ansible | 17:48 | |
| *** smatzek has quit IRC | 17:48 | |
| *** acormier has joined #openstack-ansible | 17:50 | |
| *** smatzek_ has quit IRC | 17:50 | |
| *** smatzek__ has quit IRC | 17:51 | |
| *** smatzek has joined #openstack-ansible | 17:53 | |
| *** smatzek has quit IRC | 17:54 | |
| *** smatzek has joined #openstack-ansible | 17:55 | |
| *** weezS has joined #openstack-ansible | 17:56 | |
| *** lbragstad has joined #openstack-ansible | 17:57 | |
| *** AndyWojo has joined #openstack-ansible | 17:58 | |
| *** spikebyt1 has joined #openstack-ansible | 17:59 | |
| *** albertcard2 has joined #openstack-ansible | 18:00 | |
| *** MasterOfBugs has joined #openstack-ansible | 18:00 | |
| *** albertcard2 is now known as poopcat | 18:02 | |
| *** germs1 is now known as germs | 18:02 | |
| *** albertcard1 has quit IRC | 18:03 | |
| *** drifterza has quit IRC | 18:04 | |
| *** weezS has quit IRC | 18:06 | |
| chyka | should it be possible to deploy instances from the provider network, as well as the self service network? | 18:07 |
| chyka | i'm on ocata, and followed the openstack ansible interface configuration, and self service networks work fine | 18:08 |
| *** spikebyt1 is now known as spikebyte | 18:08 | |
| chyka | just using linuxbridge, but instances booted from the provider network can't reach the metadata server | 18:09 |
| *** weezS has joined #openstack-ansible | 18:13 | |
| *** br0k3n has joined #openstack-ansible | 18:22 | |
| cloudnull | chyka: yes that should work fine | 18:22 |
| cloudnull | can you use a self service network with a floating ip from the provider network? | 18:23 |
| mhayden | eumel8: there is a spec for that! | 18:23 |
| *** hachi_ has joined #openstack-ansible | 18:26 | |
| DimGR | cloudnull greetings | 18:42 |
| *** galstrom is now known as galstrom_zzz | 18:45 | |
| *** armaan has quit IRC | 18:45 | |
| openstackgerrit | Merged openstack/openstack-ansible master: scripts: bootstrap-ansible.sh: Do not error on 'alternatives' command https://review.openstack.org/509410 | 18:45 |
| cloudnull | o/ DimGR | 18:46 |
| openstackgerrit | Merged openstack/openstack-ansible master: scripts: sources-branch-updater-lib: Always fetch SHAs when working on master https://review.openstack.org/509285 | 18:46 |
| chyka | yeah i can attach a floating ip from the provider network fine, no issues there | 18:46 |
| *** armaan has joined #openstack-ansible | 18:46 | |
| cloudnull | ok. is DHCP enabled on the provider network ? | 18:46 |
| chyka | yup | 18:48 |
| chyka | oh actually i have the net nodes split off from the infra nodes. they still have the same interface and bridges configured as per the docs though | 18:49 |
| kylek3h | chyka: I've hit the same (although not via OSA install). For a temp work around until I figure it out, I've been using config drive during instance creation...that's worked for me. | 18:51 |
| cloudnull | chyka: ^ that should work every time | 18:53 |
| cloudnull | and its ok to have different network nodes (i prefer having isolated network nodes) | 18:53 |
| kylek3h | cloudnull: missed you at PTG! But you seem to be in a much better place. | 18:53 |
| cloudnull | are you seeing DHCP traffic on your compute hosts? | 18:53 |
| cloudnull | kylek3h: yup, sadly I couldn't make it. | 18:54 |
| cloudnull | i was in a better place, but now im back in TX | 18:54 |
| cloudnull | :P | 18:54 |
| chyka | hmm would there be any issues with config drive using ceph as a back-end, for instance migrations or upgrades? | 18:56 |
| cloudnull | kylek3h: did you get a lot accomplished at the PTG ? | 18:56 |
| cloudnull | chyka: IDK | 18:56 |
| chyka | also which interface would i check for dhcp, the br-vlan-veth? | 18:57 |
| cloudnull | br-vlan and the instance tap device | 18:57 |
| cloudnull | if its making it all the way there then the VM should get DHCP and should be able to hit the metadata service. | 18:58 |
| kylek3h | cloudnull: yeah, although I liked the format of the Altanta one better...where we had the OSA meetings Mon, Tues. | 18:58 |
| cloudnull | were you impacted by the trains ? | 18:58 |
| cloudnull | I heard that was amazing | 18:58 |
| *** gouthamr has quit IRC | 18:58 | |
| *** MasterOfBugs has quit IRC | 18:59 | |
| *** gouthamr_ has joined #openstack-ansible | 18:59 | |
| kylek3h | I scheduled late and couldn't get into that hotel with an approved price...so had to stay elsewhere, which in the end worked out. What trains? lol | 18:59 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Change location of horizon files in Liberty https://review.openstack.org/509591 | 18:59 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Add variable prep for liberty leap https://review.openstack.org/509591 | 19:01 |
| mhayden | oh man, found a tricky bug with centos upgrades | 19:04 |
| mhayden | centos 7.4 + 7.3 directories on the repo server are different | 19:04 |
| mhayden | so 7.3 containers look in the wrong place for their wheels | 19:05 |
| mhayden | not sure how to solve that one | 19:05 |
| chyka | thanks, cloudnull & kylek3h i'll check those out | 19:08 |
| chyka | im also having an issue with the neutron firewall config just staying stuck in pending, i wonder if its related to this | 19:08 |
| *** chas has joined #openstack-ansible | 19:33 | |
| *** lbragstad has quit IRC | 19:35 | |
| *** chas has quit IRC | 19:38 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Add ZFS backing store to the container create role https://review.openstack.org/509608 | 19:38 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening stable/pike: Fix path to useradd config https://review.openstack.org/509609 | 19:40 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Add ZFS backing store to the container create role https://review.openstack.org/509608 | 19:42 |
| *** weezS has quit IRC | 19:46 | |
| *** weezS has joined #openstack-ansible | 19:46 | |
| *** perniciouscaffei has joined #openstack-ansible | 20:00 | |
| *** ianw is now known as ianw|pto | 20:04 | |
| *** armaan has quit IRC | 20:07 | |
| eumel8 | mhayden: got it. thx. We have in our company an internal project with the same structure and I need some arguments to work in the same way like osa :) | 20:10 |
| *** chas has joined #openstack-ansible | 20:14 | |
| *** chas has quit IRC | 20:17 | |
| mhayden | eumel8: you are likely looking for -> https://github.com/openstack/openstack-ansible-specs/blob/master/specs/queens/ansible-roles-reuse.rst | 20:18 |
| mhayden | and i'm sure logan- would love your input | 20:18 |
| eumel8 | thx :) | 20:24 |
| *** lbragstad has joined #openstack-ansible | 20:27 | |
| *** smatzek has quit IRC | 20:28 | |
| *** smatzek has joined #openstack-ansible | 20:28 | |
| *** smatzek has quit IRC | 20:33 | |
| *** jbadiapa_ has quit IRC | 20:39 | |
| eumel8 | btw: this is a typical example how an open community works. problems, specs, proposals and discussions are well document. It's so helpful, easy to follow, in one word: awesome! | 20:44 |
| *** gouthamr_ has quit IRC | 20:46 | |
| cloudnull | I heard that was amazing | 20:55 |
| cloudnull | opps | 20:55 |
| cloudnull | any cores around that might want to give this a review https://review.openstack.org/#/c/509530/ | 20:55 |
| cloudnull | looks like jrosser has a pr up for zfs support | 20:55 |
| jrosser | think i messed up the when: | 20:55 |
| cloudnull | and it'd be nice to see if we can get that going / have a gate around it | 20:56 |
| jrosser | list of one element doesnt seem to pass a local test | 20:56 |
| cloudnull | which one ? | 20:56 |
| jrosser | ahhhhhhh ignore me :) | 20:57 |
| *** smatzek has joined #openstack-ansible | 20:57 | |
| * cloudnull ignoring :D | 20:57 | |
| jrosser | i'm also looking at lxc_hosts | 20:57 |
| jrosser | but thats a bit more mysterious to me :) | 20:58 |
| cloudnull | there's a bit of updating we can do there too | 20:58 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Add variable prep for L->N leap https://review.openstack.org/509591 | 20:58 |
| jrosser | i'm still a bit baffled by the intersection of machinectl and well, anything else! | 20:58 |
| cloudnull | at this point machinctl is just for image caching | 20:58 |
| jrosser | and in this case does that mean the per-container host base image? | 20:59 |
| cloudnull | its just an efficient means to store a cached image. | 20:59 |
| cloudnull | yes | 20:59 |
| jrosser | ok, so for COW stuff we'd want to drop that into the right place | 20:59 |
| cloudnull | youd likely just need to circumvent the systemd/machinctl things | 21:00 |
| *** smatzek has quit IRC | 21:00 | |
| jrosser | cloudnull: i was sort of here http://paste.openstack.org/show/622703/ | 21:01 |
| jrosser | then looking at the rest of the role and figuring i needed some advice | 21:02 |
| jrosser | you know the more i look at all this the more it looks like parallel lxd universe | 21:04 |
| cloudnull | its pretty close | 21:05 |
| *** armaan has joined #openstack-ansible | 21:06 | |
| cloudnull | a lot of this predates LXD however I suspect there's a bit of overlap these days. | 21:06 |
| cloudnull | that diff looks like it'd work fine. | 21:08 |
| jrosser | does anyone run the tests behind a proxy? i'm having to hack up run_tests.sh to make that go | 21:10 |
| *** eumel8 has quit IRC | 21:11 | |
| cloudnull | not i. | 21:12 |
| cloudnull | maybe logan-? | 21:12 |
| cloudnull | I updated my review on https://review.openstack.org/#/c/509608 | 21:16 |
| cloudnull | given the diff it looks like there's going to be a single base image, like overlay and lvm + snap | 21:16 |
| cloudnull | so we should be able to use the lxc-clone command and just create new containers from the known base image. | 21:17 |
| *** tasker has joined #openstack-ansible | 21:17 | |
| tasker | hello. how are things? | 21:17 |
| jrosser | https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/tasks/lxc_cache_preparation.yml#L43 and L45 would be "when not zfs" i think | 21:17 |
| tasker | it's been awhile and I was starting to miss y'all. | 21:17 |
| cloudnull | ohai tasker | 21:18 |
| cloudnull | doing well. how have you been ? | 21:19 |
| tasker | good, good. | 21:20 |
| tasker | I've been complaining about a few things, but I'm the only one in the office. | 21:20 |
| cloudnull | I like yelling at myself in a reflection too . | 21:22 |
| cloudnull | jrosser: looks like lxc-copy already has zfs support | 21:22 |
| cloudnull | so there may be nothing to do. | 21:23 |
| cloudnull | zfs support in the lxc-container-create role may be as simple as https://review.openstack.org/#/c/509530/4/tasks/lxc_container_create_overlayfs.yml | 21:23 |
| jrosser | ah magic | 21:24 |
| cloudnull | ^ that assuming the base image was already in place. | 21:25 |
| jrosser | so my patch would instead include the _cow.yml | 21:25 |
| jrosser | becasue i did like how you drove the file to include off the backing store type | 21:25 |
| cloudnull | yup | 21:25 |
| cloudnull | the zfs file would likely just be "- include: lxc_container_cow.yml" | 21:26 |
| cloudnull | I dont have a zfs env to test with | 21:26 |
| jrosser | i was next thinking about the container_create tests | 21:26 |
| *** br0k3n has quit IRC | 21:26 | |
| jrosser | and how to lash a trivial zfs volume into that | 21:27 |
| cloudnull | yea. we should add this to the test matric | 21:27 |
| cloudnull | **matrix | 21:27 |
| cloudnull | and probably limit zfs support to Ubuntu? | 21:27 |
| jrosser | probably wise | 21:27 |
| cloudnull | gotta run, back in a few | 21:28 |
| *** br0k3n has joined #openstack-ansible | 21:40 | |
| jrosser | cloudnull: i'm done for today - if you have any good ideas about how to do a block device in the test that would be great | 21:47 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Add variable prep for L->N leap https://review.openstack.org/509591 | 21:54 |
| spotz | I have to say I'm happy to see Juenkins +1 again:) | 22:00 |
| spotz | Jenkins even | 22:00 |
| *** kylek3h has quit IRC | 22:02 | |
| cloudnull | jrosser: we should be able to simply use a loop device w/ zfs ? | 22:06 |
| *** threestrands has joined #openstack-ansible | 22:08 | |
| *** acormier has quit IRC | 22:11 | |
| *** pbandark1 has joined #openstack-ansible | 22:11 | |
| *** pbandark has quit IRC | 22:11 | |
| *** pbandark1 is now known as pbandark | 22:11 | |
| *** acormier has joined #openstack-ansible | 22:12 | |
| *** acormier has quit IRC | 22:17 | |
| *** hachi_ has quit IRC | 22:25 | |
| *** br0k3n has quit IRC | 22:25 | |
| *** gouthamr has joined #openstack-ansible | 22:26 | |
| *** lbragstad has quit IRC | 22:38 | |
| *** exodusftw has quit IRC | 22:52 | |
| *** exodusftw has joined #openstack-ansible | 22:58 | |
| *** esberglu has quit IRC | 23:00 | |
| *** esberglu has joined #openstack-ansible | 23:01 | |
| *** esberglu has quit IRC | 23:05 | |
| tasker | why am I waiting so long on "setup-openstack"? is it actually downloading and unpacking all of the venvs from the repo even for containers that haven't changed? I thought OSA tried to be idempotent. | 23:07 |
| *** weezS has quit IRC | 23:15 | |
| *** acormier has joined #openstack-ansible | 23:24 | |
| *** pbandark has quit IRC | 23:25 | |
| *** acormier has quit IRC | 23:28 | |
| *** chyka has quit IRC | 23:40 | |
| tasker | what conditions does OSA use to determine if it needs to reinstall the virtual environments? | 23:40 |
| tasker | is it a timestamp on the repo files? | 23:40 |
| tasker | something in the container? | 23:40 |
| *** chyka has joined #openstack-ansible | 23:41 | |
| *** chyka has quit IRC | 23:45 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!