| *** sacharya has joined #openstack-ansible | 00:12 | |
| *** rebase has quit IRC | 00:15 | |
| *** sacharya has quit IRC | 00:17 | |
| *** mss has quit IRC | 00:21 | |
| *** openstackstatus has quit IRC | 00:24 | |
| *** openstackstatus has joined #openstack-ansible | 00:24 | |
| *** ChanServ sets mode: +v openstackstatus | 00:24 | |
| *** KLevenstein has quit IRC | 00:26 | |
| *** rebase has joined #openstack-ansible | 00:34 | |
| *** eil397 has quit IRC | 00:43 | |
| *** sdake has joined #openstack-ansible | 00:46 | |
| *** rebase has quit IRC | 00:51 | |
| *** galstrom is now known as galstrom_zzz | 00:55 | |
| *** sdake has quit IRC | 01:06 | |
| *** gtt116__ has joined #openstack-ansible | 01:06 | |
| *** sdake has joined #openstack-ansible | 01:19 | |
| *** karimb has quit IRC | 01:20 | |
| *** openstackgerrit has quit IRC | 01:22 | |
| *** openstackgerrit has joined #openstack-ansible | 01:22 | |
| *** sdake has quit IRC | 01:32 | |
| *** markvoelker has quit IRC | 01:46 | |
| *** coolj has joined #openstack-ansible | 01:52 | |
| coolj | cloudnull: do we use older python-openstackclient versions for a reason? http://rpc-slushee.rackspace.com/os-releases/11.0.0/python_openstackclient-1.0.2-py2.py3-none-any.whl | 01:55 |
|---|---|---|
| coolj | i need at least 1.6.0 for OS_ENDPOINT_TYPE support https://github.com/openstack/python-openstackclient/commit/ee64c2fa6b9eb6f8d628e8899df39a4ee01c2bb3 | 01:55 |
| cloudnull | kysse: So far the only other way that I can cause this issue, is to have a misconfiguration in neutron or to to remove a nic from a network or compute node | 01:58 |
| kysse | arhm. So I only use two infra nodes, and those nodes are also network nodes | 01:58 |
| cloudnull | so i was curious if you have all of the br-vlan, br-vxlan, br-mgmt interfaces on both your network and compute nodes | 01:59 |
| kysse | neutron containers run on those two machines | 01:59 |
| cloudnull | coolj: it looks like openstackclient it pinned for kilo | 02:00 |
| cloudnull | https://github.com/openstack/requirements/blob/stable/kilo/global-requirements.txt#L133 | 02:00 |
| cloudnull | and we have http://rpc-repo.rackspace.com/os-releases/11.2.5/python_openstackclient-1.0.4-py2.py3-none-any.whl | 02:00 |
| cloudnull | kysse: so those bridges are on your infra boxes, but are they also on your compute node? | 02:00 |
| kysse | http://paste.nerv.fi/45555345.txt | 02:01 |
| cloudnull | ok | 02:04 |
| kysse | I'm worried about interfaces 'inside' lxc containers. | 02:05 |
| cloudnull | whys that ? | 02:05 |
| kysse | is it possible that those interfaces are not specified correctly? | 02:05 |
| coolj | cloudnull: yuck, but okay thanks | 02:05 |
| kysse | just saying, cuz I dont fully understand that eth11..12 naming inside them. | 02:06 |
| cloudnull | I mean in the grand scheme of things everything is possible. | 02:06 |
| cloudnull | you could destroy the neutron containers and recreate them | 02:06 |
| cloudnull | just to be sure | 02:07 |
| kysse | and what host_bind_override is supposed to do? | 02:07 |
| cloudnull | openstack-ansible lxc-container-destory.yml --limit neutron_agent; openstack-ansible lxc-container-create.yml --limit neutron_agent | 02:08 |
| cloudnull | the host_bind_override is to give a network a specific interface | 02:08 |
| kysse | omg. destroyer. | 02:08 |
| kysse | hmh. | 02:09 |
| cloudnull | but only a specific interface on a given machine | 02:09 |
| kysse | http://paste.nerv.fi/91658287.txt If I have specified things like that | 02:09 |
| cloudnull | however you've removed the flat network from the openstack_user_config.yml, correct? | 02:09 |
| kysse | yes | 02:09 |
| cloudnull | if you have that entry | 02:10 |
| cloudnull | you would need to create the eth12 device on your compute nodes as well as your network nodes. | 02:10 |
| kysse | ah sorry, looked at aio file.. | 02:10 |
| cloudnull | and neutron would bind to the specified interface instead of the known bridge | 02:11 |
| kysse | I'm pretty much trying to avoid physical interface name mappings | 02:11 |
| kysse | soo let's destroy those containers.. | 02:12 |
| cloudnull | lol | 02:13 |
| cloudnull | once done you'll need to rerun the os-neutron-install.yml play to resetup neutron within the containers. | 02:13 |
| kysse | yep | 02:15 |
| openstackgerrit | Merged openstack/openstack-ansible: Add container-reconciler and object-expirer to os-swift https://review.openstack.org/247290 | 02:16 |
| openstackgerrit | Merged openstack/openstack-ansible: Adjust default nova_force_config_drive to be False https://review.openstack.org/251232 | 02:16 |
| openstackgerrit | Merged openstack/openstack-ansible: Remove ssh args to ensure pipelining is working https://review.openstack.org/250088 | 02:16 |
| cloudnull | kysse: also open up the /etc/openstack_deploy/openstack_inventory.json file and see if you have an entry like http://cdn.pasteraw.com/dfgl9qlvs4l02tup39qdgnyld6xsf6f | 02:18 |
| cloudnull | note, i have eth12 for a flat network | 02:18 |
| cloudnull | however you should not | 02:18 |
| kysse | ok | 02:19 |
| kysse | http://sprunge.us/jUPc | 02:28 |
| *** sacharya has joined #openstack-ansible | 02:36 | |
| kysse | so.. | 02:37 |
| kysse | same error again, I did everything. Hmm. | 02:38 |
| kysse | PortBindingFailed: Binding failed for port d408b88e-eb5f-4757-be94-28f8e05750a1, please check neutron logs for more information. | 02:38 |
| kysse | what is that UID? | 02:38 |
| kysse | I tried to search, but I could not find it from mysql | 02:38 |
| kysse | or neutron any-list etc. | 02:38 |
| cloudnull | that looks right | 02:42 |
| *** sacharya_ has joined #openstack-ansible | 02:43 | |
| cloudnull | what about the /var/log/neutron/neutron-linuxbridge-agent.log file from within the agent containers | 02:45 |
| cloudnull | do you have a logging node | 02:45 |
| cloudnull | if so you should be able to easily get all logs as they're shipped | 02:45 |
| *** sacharya has quit IRC | 02:46 | |
| kysse | yep | 02:47 |
| openstackgerrit | Merged openstack/openstack-ansible: Fix logrotate config for horizon and keystone. https://review.openstack.org/246327 | 02:49 |
| kysse | lets see | 02:52 |
| kysse | http://sprunge.us/DDBN | 02:57 |
| kysse | http://sprunge.us/LQdR | 02:57 |
| kysse | http://sprunge.us/QMKL | 02:57 |
| kysse | nothing interesting.. | 02:57 |
| *** fawadkhaliq has joined #openstack-ansible | 02:57 | |
| kysse | oh gosh. from within the agent containers? HMM | 02:58 |
| kysse | there are no such log! | 02:58 |
| kysse | only compute1(su9d1), infra1,infra2 | 02:59 |
| *** sacharya_ has quit IRC | 03:28 | |
| *** sacharya has joined #openstack-ansible | 03:29 | |
| *** baker has joined #openstack-ansible | 03:36 | |
| *** baker has quit IRC | 03:51 | |
| *** sacharya has quit IRC | 04:37 | |
| *** sacharya has joined #openstack-ansible | 04:38 | |
| *** jbweber has joined #openstack-ansible | 04:44 | |
| *** tlian2 has quit IRC | 04:53 | |
| *** sdake has joined #openstack-ansible | 05:00 | |
| *** sdake_ has joined #openstack-ansible | 05:03 | |
| *** sdake has quit IRC | 05:04 | |
| *** sdake_ has quit IRC | 05:10 | |
| iceyao_ | hi | 05:30 |
| *** fawadkhaliq has quit IRC | 05:36 | |
| *** sacharya_ has joined #openstack-ansible | 05:38 | |
| *** sacharya has quit IRC | 05:40 | |
| *** alkari has joined #openstack-ansible | 05:49 | |
| *** fawadkhaliq has joined #openstack-ansible | 05:58 | |
| *** fawadkhaliq has quit IRC | 05:58 | |
| *** fawadkhaliq has joined #openstack-ansible | 05:59 | |
| *** sdake has joined #openstack-ansible | 06:12 | |
| *** rminmin has joined #openstack-ansible | 06:12 | |
| *** javeriak has joined #openstack-ansible | 06:14 | |
| *** sacharya_ has quit IRC | 06:28 | |
| *** alkari has quit IRC | 06:33 | |
| *** phiche has joined #openstack-ansible | 06:34 | |
| *** javeriak has quit IRC | 06:36 | |
| *** rminmin has quit IRC | 06:37 | |
| *** sdake has quit IRC | 06:38 | |
| *** mss has joined #openstack-ansible | 06:43 | |
| *** mss has quit IRC | 06:52 | |
| *** phiche has quit IRC | 07:02 | |
| *** phiche has joined #openstack-ansible | 07:15 | |
| *** fawadkhaliq has quit IRC | 07:52 | |
| *** fawadkhaliq has joined #openstack-ansible | 07:53 | |
| *** fawadkhaliq has quit IRC | 07:53 | |
| *** fawadkhaliq has joined #openstack-ansible | 07:54 | |
| *** fawadkhaliq has quit IRC | 07:57 | |
| *** fawadkhaliq has joined #openstack-ansible | 07:57 | |
| *** fawadkhaliq has quit IRC | 07:57 | |
| *** fawadkhaliq has joined #openstack-ansible | 07:58 | |
| *** fawadkhaliq has quit IRC | 08:02 | |
| *** fawadkhaliq has joined #openstack-ansible | 08:03 | |
| *** fawadkhaliq has quit IRC | 08:03 | |
| *** fawadkhaliq has joined #openstack-ansible | 08:04 | |
| *** javeriak has joined #openstack-ansible | 08:19 | |
| *** adac has joined #openstack-ansible | 08:20 | |
| adac | I'm trying to recover the galera cluster (All machines have 'seqno: -1' ) with "mysqld --wsrep-recover" however the sequence number is exact the same as also 'ansible galera_container -m shell -a "cat /var/lib/mysql/grastate.dat"' shows on all three nodes | 08:21 |
| adac | still the procedure described in http://docs.openstack.org/developer/openstack-ansible/install-guide/ops-galera-start.html cannot start up the cluster anymore | 08:21 |
| adac | any ideas what I might can do about this? | 08:21 |
| *** permalac has joined #openstack-ansible | 08:25 | |
| adac | It seems there are two warnings when starting the recovery: https://gist.github.com/anonymous/314d8cbcf8b57e981f11 | 08:27 |
| *** mpavone has joined #openstack-ansible | 08:30 | |
| *** tiagogomes has joined #openstack-ansible | 08:39 | |
| odyssey4me | kysse cloudnull the keystone_authtoken settings are right as they are, the admin_ options were all deprecated in kilo - the user-name deprecation error is incorrect as user-name is an inappropriate option for any ini files... I think you had a red herring there | 08:43 |
| odyssey4me | cloudnull also, 'auth_plugin = password' is the new default in most services for liberty | 08:44 |
| odyssey4me | adac, so all your nodes have seq -1 ? | 08:45 |
| adac | odyssey4me, exactly yes | 08:46 |
| odyssey4me | adac from /opt/openstack-ansible/playbooks execute: openstack-ansible -e galera_ignore_cluster_state=true galera-install.yml | 08:47 |
| *** javeriak has quit IRC | 08:47 | |
| odyssey4me | evrardjp morning - can you please backport https://review.openstack.org/246503 and https://review.openstack.org/248819 to the liberty branch? | 08:48 |
| odyssey4me | morning iceyao_ :) | 08:48 |
| *** karimb has joined #openstack-ansible | 08:48 | |
| adac | odyssey4me, would this delete the data in the mysql database? I mean in my testing case it is ok, just asking not that I execute this command at some point in production as well if its harmful ;-) | 08:53 |
| odyssey4me | adac ideally this should not be run in production, but when your whole cluster is at seq -1 then I don't think there's harm in it - all it's going to do is initiate the cluster on one node and then start the service on the others | 08:56 |
| odyssey4me | it's be the equivalent of running: ansible galera_all[0] -m shell -a '/etc/init.d/mysql start --wsrep-new-cluster' | 08:57 |
| adac | odyssey4me, I see. thanks a lot! | 08:58 |
| hughsaunders | odyssey4me: galera:!galera[0] | 08:59 |
| odyssey4me | then ansible galera_all:!galera_all[0] -m shell -a '/etc/init.d/mysql start' | 09:00 |
| odyssey4me | thanks hughsaunders :) | 09:00 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Allow ramdisk_id, kernel_id to be null on schema https://review.openstack.org/251720 | 09:01 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Add container-reconciler and object-expirer to os-swift https://review.openstack.org/251721 | 09:01 |
| *** mpavone has quit IRC | 09:02 | |
| adac | odyssey4me, I now did "openstack-ansible -e galera_ignore_cluster_state=true galera-install.yml" but "ansible galera_container -m shell -a "cat /var/lib/mysql/grastate.dat"" still shows "seqno -1" Do I still have to restart them manually? | 09:02 |
| *** fawadkhaliq has quit IRC | 09:03 | |
| *** fawadkhaliq has joined #openstack-ansible | 09:03 | |
| *** karimb_ has joined #openstack-ansible | 09:05 | |
| odyssey4me | evrardjp lol, never mind - I see that the horizon branding one is already done | 09:05 |
| odyssey4me | adac it sounds to me like your cluster is utterly broken - you'll have to dig into what went wrong using mariadb/galera documentation | 09:05 |
| odyssey4me | if this is a test setup, then you can simply blow away the containers and rebuild them | 09:06 |
| *** iceyao_ has quit IRC | 09:06 | |
| odyssey4me | the data will still be there anyway | 09:06 |
| *** karimb has quit IRC | 09:07 | |
| *** xek has joined #openstack-ansible | 09:08 | |
| adac | odyssey4me, exactly make everything new | 09:10 |
| adac | odyssey4me, may I ask which command/playbook I have to run for to re-setup the galera cluster? | 09:11 |
| *** javeriak has joined #openstack-ansible | 09:16 | |
| *** mpavone has joined #openstack-ansible | 09:16 | |
| odyssey4me | adac I think it's along the lines of: openstack-ansible lxc-containers-destroy.yml --limit galera_all | 09:16 |
| odyssey4me | adac then you'll need to re-run: openstack-ansible setup-hosts.yml | 09:20 |
| odyssey4me | (to recreate the containers) | 09:20 |
| odyssey4me | then openstack-ansible galera-install.yml | 09:20 |
| odyssey4me | (ro reinstall galera) | 09:20 |
| adac | ok thank you! | 09:21 |
| *** iceyao has joined #openstack-ansible | 09:21 | |
| *** javeriak_ has joined #openstack-ansible | 09:35 | |
| *** javeriak has quit IRC | 09:38 | |
| *** javeriak has joined #openstack-ansible | 09:39 | |
| *** javeriak_ has quit IRC | 09:40 | |
| *** javeriak_ has joined #openstack-ansible | 09:46 | |
| *** javeriak has quit IRC | 09:49 | |
| *** javeriak has joined #openstack-ansible | 09:57 | |
| *** javeriak_ has quit IRC | 09:58 | |
| permalac | if I would like to deploy an openstack instance with ansible, where I should start reading? | 10:15 |
| andymccr | permalac: there is an install guide available here http://docs.openstack.org/developer/openstack-ansible/install-guide/index.html | 10:19 |
| andymccr | oh wait instance | 10:19 |
| andymccr | you mean just deploy an instance on an existing openstack cloud using ansible? | 10:19 |
| permalac | sorry, my bad. I mean prepare an openstack installation from scratch. | 10:20 |
| andymccr | ahh excellent yes, tehn the above link will work :) | 10:20 |
| permalac | the install guide is good. | 10:20 |
| permalac | thanks | 10:20 |
| andymccr | *then | 10:20 |
| andymccr | no problem! | 10:20 |
| andymccr | if you run into any problems feel free to drop questions in here - theres usually somebody about who will be able to help | 10:21 |
| adac | odyssey4me, When reinstalling galera, I get the following error: https://gist.github.com/anonymous/d6ea8af57ebd78cf9ee5 I have no luck with galera it seems (: | 10:33 |
| odyssey4me | permalac you can kick the tyres with a single VM too, using http://docs.openstack.org/developer/openstack-ansible/developer-docs/quickstart-aio.html | 10:33 |
| odyssey4me | adac if you do openstack-ansible -e galera_ignore_cluster_state=true galera-install.yml now? | 10:34 |
| adac | odyssey4me, trying that out | 10:35 |
| *** timrc has quit IRC | 10:47 | |
| *** timrc has joined #openstack-ansible | 10:49 | |
| adac | odyssey4me, it just finished, it says still the following: https://gist.github.com/anonymous/51cf4702ee80ef0b2ba6 | 11:00 |
| adac | howeer this installation seems to be fine, there was no error in it | 11:00 |
| adac | *however | 11:00 |
| *** marekd has joined #openstack-ansible | 11:04 | |
| *** mpavone has quit IRC | 11:09 | |
| adac | I destry everything now and rebuild it | 11:12 |
| adac | :) | 11:13 |
| odyssey4me | :) | 11:19 |
| *** mpavone has joined #openstack-ansible | 11:21 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-specs: Added spec for policy-files-distribution This explains/opens the discussion to component policy file distribution with the help of openstack-ansible. https://review.openstack.org/249227 | 11:33 |
| *** misc has quit IRC | 11:34 | |
| *** mss has joined #openstack-ansible | 11:52 | |
| *** misc has joined #openstack-ansible | 11:54 | |
| *** mss has quit IRC | 11:56 | |
| *** javeriak has quit IRC | 11:57 | |
| odyssey4me | hughsaunders mattt did you see cloudnull's response in https://review.openstack.org/241483 ? | 12:05 |
| mattt | odyssey4me: updated | 12:08 |
| *** javeriak has joined #openstack-ansible | 12:21 | |
| *** jaypipes has joined #openstack-ansible | 12:22 | |
| *** jaypipes has quit IRC | 12:49 | |
| mhayden | morning | 12:49 |
| *** jaypipes has joined #openstack-ansible | 12:54 | |
| *** fawadkhaliq has quit IRC | 13:00 | |
| Sam-I-Am | mornings | 13:00 |
| *** fawadkhaliq has joined #openstack-ansible | 13:01 | |
| *** fawadkhaliq has quit IRC | 13:02 | |
| mgariepy | good morning everyone | 13:02 |
| *** fawadkhaliq has joined #openstack-ansible | 13:03 | |
| *** fawadkhaliq has quit IRC | 13:03 | |
| odyssey4me | o/ all | 13:03 |
| Sam-I-Am | yo | 13:03 |
| *** javeriak has quit IRC | 13:04 | |
| *** fawadkhaliq has joined #openstack-ansible | 13:04 | |
| *** gparaskevas has joined #openstack-ansible | 13:06 | |
| gparaskevas | hello you guys(cartman voice) | 13:10 |
| mattt | hehe | 13:10 |
| gparaskevas | can i have the same variable name in more than one ansible role for the same playbook? | 13:11 |
| gparaskevas | i have a playbooks that uses 3 roles lets say and each role has a variable named game_name: | 13:11 |
| gparaskevas | every role has the same variable name with diferent value | 13:12 |
| gparaskevas | the variable file i use is in defaults | 13:12 |
| gparaskevas | on every roles | 13:12 |
| gparaskevas | rol* | 13:12 |
| gparaskevas | role* | 13:12 |
| mattt | gparaskevas: is each role getting called at the same time ? | 13:14 |
| mattt | sounds like a bad idea personally, but i'm not an ansible expert | 13:15 |
| gparaskevas | basicaly they ar called like this | 13:15 |
| gparaskevas | http://paste.openstack.org/show/480494/ | 13:17 |
| odyssey4me | gparaskevas if you want to use the same name in multiple role defaults or role vars you can, but if you don't namespace them individually then you'll end up with confusion if you ever run two roles together in a playbook with a playbook var or an override var | 13:17 |
| odyssey4me | that's why the best practise is to namespace them | 13:17 |
| *** persia has joined #openstack-ansible | 13:18 | |
| gparaskevas | how do i namespace them? | 13:18 |
| *** fawadkhaliq has quit IRC | 13:18 | |
| gparaskevas | let me check | 13:18 |
| *** fawadkhaliq has joined #openstack-ansible | 13:19 | |
| gparaskevas | so what i do is wrong? | 13:19 |
| mattt | gparaskevas: you gonna have a bad time right there | 13:20 |
| mattt | gparaskevas: do role1_game_name, role2_game_name, etc. | 13:20 |
| odyssey4me | gparaskevas if you have a role called 'role1' and a var called 'var1', then a role called 'role2' and a var in role2 called 'var1' then if you call both roles with group_vars/playbook vars you can only set the var in group_vars/playbook vars for both to the same value - never different | 13:20 |
| odyssey4me | so to namespace them is to simply call them, in each role, role1_var1 and role2_var1 | 13:21 |
| gparaskevas | OK this is what i am doing now. let make a meme about that as well | 13:22 |
| mattt | gparaskevas: https://cessnachick.files.wordpress.com/2015/08/youre-going-to-have-a-bad-time.png | 13:23 |
| gparaskevas | http://cdn.meme.am/instances2/500x/3125581.jpg | 13:23 |
| gparaskevas | #proper | 13:24 |
| gparaskevas | :P | 13:24 |
| gparaskevas | hahah | 13:24 |
| mattt | :D | 13:24 |
| gparaskevas | thanks both of you guys! | 13:25 |
| gparaskevas | ski instructors* | 13:25 |
| mattt | gparaskevas: have you been drinking? | 13:27 |
| mattt | :) | 13:27 |
| gparaskevas | loll because of the happiness? nah | 13:28 |
| gparaskevas | :P | 13:28 |
| gparaskevas | maybe its the cafeine effect :P | 13:28 |
| *** sdake has joined #openstack-ansible | 13:34 | |
| *** adac has quit IRC | 13:34 | |
| *** sdake has quit IRC | 13:42 | |
| *** permalac has quit IRC | 13:44 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Updating tests for openstack-ansible-security https://review.openstack.org/251430 | 13:45 |
| cloudnull | Morning | 13:48 |
| kysse | good morning | 13:49 |
| *** markvoelker_ has joined #openstack-ansible | 13:49 | |
| mhayden | odyssey4me: thanks for the tips on https://review.openstack.org/#/c/251430/ -- i've got it working now | 13:49 |
| *** sdake has joined #openstack-ansible | 13:50 | |
| *** tlian has joined #openstack-ansible | 13:51 | |
| cloudnull | Kysse did you make it all happy yet? | 13:51 |
| cloudnull | Or is it still mad? | 13:52 |
| *** mss has joined #openstack-ansible | 13:52 | |
| kysse | she is angry. | 13:52 |
| cloudnull | Gparaskevas rofl | 13:53 |
| *** sdake_ has joined #openstack-ansible | 13:53 | |
| gparaskevas | cloudnull: about the meme :P hehe | 13:53 |
| cloudnull | Kysse same error? Vif binding? | 13:54 |
| *** sdake has quit IRC | 13:54 | |
| cloudnull | Yup the meme made me lol | 13:54 |
| gparaskevas | cloudnull: hah thanks | 13:54 |
| *** mss has quit IRC | 13:57 | |
| kysse | same error | 13:57 |
| cloudnull | ... :'( | 13:59 |
| cloudnull | And in all we've not seen anything of merit in the neuron agent logs, right? | 14:00 |
| cloudnull | Linuxbridge,l3,server on computer node and neuron containers | 14:01 |
| kysse | nothing. Linux-bridge logs are empty @neutron containers | 14:05 |
| cloudnull | Neutron_agent containers right? | 14:06 |
| cloudnull | I'd the neutron-linuxbridge-agent running? | 14:06 |
| cloudnull | *is | 14:07 |
| mattt | kysse: do a nova service-list | 14:15 |
| kysse | I got 2 jump on the bus, I'm traveling from Turku Finland to Pori finland now, so lets check those things in 2h. | 14:17 |
| *** targon has joined #openstack-ansible | 14:18 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Updating tests for openstack-ansible-security https://review.openstack.org/251430 | 14:19 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Check mode compatibility for security role https://review.openstack.org/245813 | 14:19 |
| cloudnull | Kysse have a nice trip | 14:22 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Check mode compatibility for security role https://review.openstack.org/245813 | 14:24 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Updating tests for openstack-ansible-security https://review.openstack.org/251430 | 14:24 |
| mhayden | gerrit dependencies are wild sometimes | 14:24 |
| *** ctina has joined #openstack-ansible | 14:26 | |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Add new alarm for the galera checks https://review.openstack.org/251890 | 14:27 |
| *** KLevenstein has joined #openstack-ansible | 14:31 | |
| *** markvoelker has joined #openstack-ansible | 14:34 | |
| *** markvoelker_ has quit IRC | 14:35 | |
| *** daneyon has joined #openstack-ansible | 14:53 | |
| *** daneyon has quit IRC | 14:55 | |
| *** targon has quit IRC | 14:57 | |
| mattt | cloudnull: quick question for you | 15:04 |
| *** linggao has joined #openstack-ansible | 15:05 | |
| linggao | Good morning. | 15:05 |
| *** spotz_zzz is now known as spotz | 15:06 | |
| linggao | I have installed all-in-on using ansible. But somehow I could not find the container for nova_compute. And nova hypervisor-list shows no hypervisors. | 15:06 |
| mattt | linggao: nova-compute will run on the host itself | 15:07 |
| linggao | mattt, thanks. Good to know. But why there is no hypervisor listed? | 15:07 |
| linggao | nova hypervisor-list | 15:09 |
| linggao | +----+---------------------+-------+--------+ | 15:09 |
| linggao | | ID | Hypervisor hostname | State | Status | | 15:09 |
| linggao | +----+---------------------+-------+--------+ | 15:09 |
| linggao | +----+---------------------+-------+--------+ | 15:09 |
| bgmccollum | well, seems nova-compute isn't running... | 15:09 |
| bgmccollum | or isn't checking in | 15:09 |
| bgmccollum | check /var/log/nova/nova-compute.log for clues | 15:10 |
| *** daneyon has joined #openstack-ansible | 15:10 | |
| cloudnull | Mattt what's up? | 15:14 |
| mattt | cloudnull: hey! https://github.com/openstack/openstack-ansible/blob/master/playbooks/os-nova-install.yml#L79-L93 | 15:14 |
| mattt | cloudnull: we reference nova_virt_type there but that var isn't available in the pre_tasks from what i can tell ... any way around that short of moving the task into the os_nova role? | 15:15 |
| *** linggao_ has joined #openstack-ansible | 15:15 | |
| *** linggao has quit IRC | 15:15 | |
| *** linggao_ has quit IRC | 15:16 | |
| *** linggao has joined #openstack-ansible | 15:17 | |
| cloudnull | mattt the shell task ? | 15:17 |
| mattt | cloudnull: yah, the when: clause references nova_virt_type | 15:18 |
| mattt | but unless you explicitly set that it will cause os-nova-install.yml to fail | 15:18 |
| mattt | or do we mandate that you always set that in your env ? | 15:18 |
| *** Mudpuppy has joined #openstack-ansible | 15:18 | |
| mattt | linggao: did you see bgmccollum's response ? | 15:19 |
| cloudnull | ah i see | 15:19 |
| cloudnull | i think we could change https://github.com/openstack/openstack-ansible/blob/master/playbooks/os-nova-install.yml#L90 to - 'nova_virt_type is defined and nova_virt_type==kvm' | 15:19 |
| *** baker has joined #openstack-ansible | 15:20 | |
| mattt | cloudnull: but that would skip the task when someone is actually using kvm but didn't set an override :) | 15:20 |
| cloudnull | rather - "nova_virt_type is defined and nova_virt_type=='kvm'" | 15:20 |
| cloudnull | correct. | 15:20 |
| cloudnull | but that task is only used when nova is in a container | 15:20 |
| mattt | which is what i'm testing now :P | 15:21 |
| linggao | mattt, yes. I am trying to bring up the nova-compute by hand on host. Got the following error. http://pastebin.com/77HEEsQs | 15:21 |
| cloudnull | so then I would say that the variable would need to be defined when using kvm and running nova in a container | 15:21 |
| cloudnull | because the role has no clue what a container is | 15:21 |
| mattt | cloudnull: k | 15:23 |
| mattt | cloudnull: i was hoping you could vars_files os_nova's defaults file but that doesn't even work | 15:23 |
| mattt | at least not for pre_tasks | 15:23 |
| cloudnull | to that end, I believe in the liberty release we've removed everything container releated from the various roles and put it all in the tasks. | 15:24 |
| cloudnull | for the plays that is | 15:24 |
| cloudnull | mattt: i think you can use straight "vars" in the play | 15:24 |
| mattt | cloudnull: i see, let me have a look ... i'm actually fiddling w/ kilo here specifically | 15:25 |
| *** fawadkhaliq has quit IRC | 15:25 | |
| cloudnull | so vars: "{{ _virt_type | default('kvm') }}" | 15:25 |
| cloudnull | then in the task use _virt_type as the check variable | 15:25 |
| mattt | cloudnull: ah i see what you mean, not ideal because we now decouple from defaults in os_nova ... but yeah it's either that or we bail if someone doesn't specify it | 15:26 |
| mattt | what is worse do you think? | 15:26 |
| *** fawadkhaliq has joined #openstack-ansible | 15:26 | |
| *** markvoelker_ has joined #openstack-ansible | 15:28 | |
| cloudnull | i think its safe to assume that KVM will be used when running nova-compute in an LXC container | 15:30 |
| cloudnull | maybe add a check before hand to see if "/dev/kvm /dev/kvm" are available | 15:31 |
| *** markvoelker has quit IRC | 15:32 | |
| cloudnull | however this begs the larger question, should nova compute run in a container? but Ill leave that to the philosophers. | 15:32 |
| *** markvoelker_ has quit IRC | 15:33 | |
| *** gtt116_ has joined #openstack-ansible | 15:34 | |
| *** sdake_ has quit IRC | 15:36 | |
| *** gtt116__ has quit IRC | 15:38 | |
| linggao | mattt, bgmccollum I got hte following error when starting nova-compute. 2015-12-01 09:36:27.571 4038 WARNING nova.virt.libvirt.driver [req-f7b2861f-dba5-40b3-b489-fe11794f6bcf - - - - -] Cannot update service status on host "ctl2b" since it is not registered. | 15:39 |
| mattt | linggao: yeah probably due to the libvirt error in your paste | 15:41 |
| mattt | linggao: what are you running this on ? | 15:41 |
| linggao | mattt, I am running it on ubuntu 14.04.3 LTS | 15:41 |
| mattt | linggao: you running kvm in kvm or something? | 15:42 |
| linggao | mattt, the host is a vm. I believe it is xen. | 15:43 |
| mattt | linggao: try setting virt_type = qemu in /etc/nova/nova.conf and restart nova-compute | 15:44 |
| linggao | ok | 15:44 |
| mattt | linggao: if that works, you can add an override to /etc/openstack_deploy/user_variables.yml (nova_virt_type: qemu) on your host where you run ansible | 15:45 |
| bgmccollum | linggao: can you paste /var/log/libvirt/libvirtd.log too...? | 15:46 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Convert AIO bootstrap from bash to Ansible https://review.openstack.org/239525 | 15:46 |
| linggao | mattt the virt_type is already qumu in nova.conf. It is default from ansible. | 15:46 |
| mattt | bummer :( | 15:46 |
| linggao | bgmccollum, http://pastebin.com/GY35KKKz | 15:47 |
| bgmccollum | linggao: i know how to fix this... | 15:48 |
| bgmccollum | odyssey4me: we have a live example of an unidentifiable cpu like i talked about last week...! | 15:48 |
| mattt | bgmccollum: yay ! | 15:48 |
| odyssey4me | heh bgmccollum | 15:49 |
| bgmccollum | linggao: ok, you need to replace your /usr/share/libvirt/cpu_map.xml with this -- https://raw.githubusercontent.com/cloud-training/rpc-heat-ansible/master/playbooks/files/usr/share/libvirt/cpu_map.xml | 15:49 |
| odyssey4me | bgmccollum I think the right solution for this is to get it upstream into the libvirt codebase - but we may also have to work around it in some way... | 15:49 |
| bgmccollum | then restart libvirt-bin, then restart nova-compute | 15:50 |
| linggao | bgmccollum, ok. Let me try..... | 15:50 |
| odyssey4me | mhayden alright - let's see if the updated https://review.openstack.org/239525 passes - if it does, do you have some time to work through it tomorrow? | 15:50 |
| mhayden | i should | 15:50 |
| bgmccollum | odyssey4me: can we also grab /var/log/libvirt/libvirtd.log in the gate? | 15:51 |
| odyssey4me | mhayden see if you can review it today already and have any questions/comments | 15:51 |
| mhayden | odyssey4me: it seems like there are a bunch of deletions in there, but i think that's from rebasing | 15:52 |
| odyssey4me | bgmccollum we should - I think we'd need a patch to collect it though - I take it that it needs to come from the compute host? | 15:52 |
| mhayden | yeah, rebasing | 15:52 |
| mhayden | okay nevermind -- got spooked there at first when i diff'ed 63/64 | 15:52 |
| bgmccollum | odyssey4me: yes compute host... | 15:52 |
| bgmccollum | odyssey4me: im guess thats lives somewhere in infra? | 15:53 |
| odyssey4me | mhayden it's quite a large restructure. but a lot of what you did was used... I just managed to cut down a lot of the shell stuff by using modules | 15:53 |
| mhayden | thanks for hacking through that, odyssey4me | 15:53 |
| odyssey4me | bgmccollum the log collection has to be set by us - right now I don't think we have anything that tells us to collect that log | 15:53 |
| mhayden | could anyone give sharma some advice on the openstack-dev thread? | 15:54 |
| mhayden | i have some ideas but someone a little more experienced might be better | 15:54 |
| odyssey4me | bgmccollum I'll look into that once we have this updated bootstrap and gate check tested to be working. It's in my TODO list. :) | 15:54 |
| linggao | bgmccollum, it does not work. same error. | 15:55 |
| mattt | derp. | 15:55 |
| *** javeriak has joined #openstack-ansible | 15:55 | |
| bgmccollum | linggao: is this your own xen host? | 15:56 |
| linggao | no, I rented this from softlayer. | 15:56 |
| odyssey4me | mhayden the good news is that both the voting and non voting checks have gone through the bootstrap already :) | 15:56 |
| *** javeriak has quit IRC | 15:57 | |
| *** javeriak_ has joined #openstack-ansible | 15:57 | |
| bgmccollum | linggao: that cpu_map.xml was specifically crafted to match the cpu flags passed down to rackspace public cloud instances on xen hosts... | 15:57 |
| bgmccollum | linggao: what you're going to need to do is customize it to match the CPU flags your host passes to your guests | 15:57 |
| odyssey4me | bgmccollum are you sure that this issue can't be worked around rather simply using the cpu_type nova.conf setting? | 15:57 |
| bgmccollum | i was hoping it would work out of the box... | 15:58 |
| bgmccollum | odyssey4me: maybe...can you specify a make / model? | 15:58 |
| linggao | bgmccollum, thanks. I will give it a try. | 15:58 |
| odyssey4me | bgmccollum yes, I think so - it's been a while since I checked | 15:58 |
| odyssey4me | any volunteers to run the bug triage? my brain is fried | 15:58 |
| odyssey4me | cloudnull ^ | 15:59 |
| cloudnull | I can | 15:59 |
| odyssey4me | thanks cloudnull - I need to step outside for a few minutes, brb | 16:00 |
| -cloudnull- bug triage time cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, dolphm, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden, scarlisle, luckyinva, ntt, javeriak | 16:00 | |
| mattt | \o | 16:00 |
| *** markvoelker has joined #openstack-ansible | 16:00 | |
| cloudnull | opps. i think i sent that twice ... | 16:00 |
| cloudnull | either way o/ | 16:00 |
| cloudnull | should be faily quick , we only have two new issues | 16:01 |
| cloudnull | ill let people filter in for a min or two | 16:02 |
| * mhayden woots | 16:03 | |
| * andymccr filters | 16:04 | |
| palendae | o/ | 16:04 |
| cloudnull | so first up | 16:05 |
| cloudnull | https://bugs.launchpad.net/openstack-ansible/+bug/1506285 | 16:05 |
| openstack | Launchpad bug 1506285 in python-openstackclient "11.2.1 : openstack client with V3 auth causes usability issues" [Undecided,New] | 16:05 |
| Sam-I-Am | i gotta sit this one out | 16:06 |
| cloudnull | which leads to https://bugs.launchpad.net/openstack-ansible/+bug/1518351 and this patch | 16:08 |
| openstack | Launchpad bug 1518351 in openstack-ansible trunk "Playbooks stop when enabling keystone ldap backend " [Medium,In progress] - Assigned to Kevin Carter (kevin-carter) | 16:08 |
| cloudnull | https://review.openstack.org/#/c/248900/ | 16:08 |
| cloudnull | which seems to need a reabase | 16:08 |
| cloudnull | and this patch to OSC https://review.openstack.org/#/c/251597/ | 16:09 |
| cloudnull | the last patch is interesting . it essentially fixes the client by assuming the domain is "default" | 16:10 |
| *** sdake has joined #openstack-ansible | 16:11 | |
| cloudnull | but Keystone PTL is against that, so id assume that would not go in | 16:11 |
| stevelle | is https://review.openstack.org/#/c/238509/ worth mentioning as related too | 16:12 |
| cloudnull | for the bug itself, i think https://review.openstack.org/#/c/248900/ resolve it for us to make sure specific ldap commands are not run when the service users are in LDAP. | 16:12 |
| cloudnull | stevelle: ++ | 16:12 |
| *** mss has joined #openstack-ansible | 16:12 | |
| cloudnull | https://review.openstack.org/#/c/238509/ is also related to the various ldap fixes. | 16:12 |
| mattt | didn't we triage these same issues last week? :) | 16:13 |
| cloudnull | now if we only had an LDAP gate... wink wink, nudge nudge (somebody) | 16:13 |
| cloudnull | mattt: likely | 16:13 |
| cloudnull | idk though | 16:13 |
| odyssey4me | cloudnull yeah, working on that soon | 16:14 |
| cloudnull | i wasnt mentioning names, :p | 16:14 |
| mattt | cloudnull: i could be mistaken! | 16:14 |
| cloudnull | so for this issue https://bugs.launchpad.net/openstack-ansible/+bug/1506285 | 16:14 |
| openstack | Launchpad bug 1506285 in python-openstackclient "11.2.1 : openstack client with V3 auth causes usability issues" [Undecided,New] | 16:14 |
| cloudnull | it was marked invalid, I tend to agree that the specific client issue is invalid for our case. | 16:15 |
| cloudnull | and that the other two PRs in flight are fixing the core problem for us | 16:15 |
| *** linggao has quit IRC | 16:15 | |
| *** linggao has joined #openstack-ansible | 16:15 | |
| odyssey4me | agreed - the source issue is being dealt with in the ldap patches... the usability issue has to be dealt with upstream | 16:15 |
| cloudnull | tiagogomes: worked on https://review.openstack.org/#/c/238509/ and it would be nice to land that | 16:16 |
| cloudnull | and i added a check to skip tasks when ldap is used https://review.openstack.org/#/c/248900/ | 16:16 |
| *** phiche has quit IRC | 16:18 | |
| cloudnull | done | 16:18 |
| cloudnull | next | 16:18 |
| cloudnull | https://bugs.launchpad.net/openstack-ansible/+bug/1521233 | 16:18 |
| openstack | Launchpad bug 1521233 in openstack-ansible "Adjust commonly failing tasks in openstack-ansible-security" [Undecided,New] - Assigned to Major Hayden (rackerhacker) | 16:18 |
| cloudnull | mhayden: what say you? | 16:19 |
| * mhayden winks | 16:19 | |
| mattt | notbug | 16:19 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Update pinned lxc version https://review.openstack.org/251945 | 16:19 |
| mhayden | this one blocks proper gate checks | 16:19 |
| mhayden | i'm almost done with the dev work on 1521233 | 16:19 |
| mhayden | i'm gradually turning those fail: module uses into actual changes with variables to toggle them on/off | 16:20 |
| cloudnull | wishlist, triaged, assigned . done | 16:20 |
| mhayden | thanks sir ;) | 16:20 |
| cloudnull | :) | 16:20 |
| cloudnull | so thats the end of the new ones. | 16:20 |
| cloudnull | anything that we need to address otherwise? | 16:20 |
| * mhayden has one, searching | 16:21 | |
| mhayden | https://bugs.launchpad.net/openstack-ansible/+bug/1516142 | 16:22 |
| openstack | Launchpad bug 1516142 in openstack-ansible "openstack-ansible-security needs a check mode for auditing" [Wishlist,In progress] - Assigned to Major Hayden (rackerhacker) | 16:22 |
| mhayden | that one is blocking some of my future work there, especially gating | 16:22 |
| mhayden | if anyone has time to put eyeballs on it, i'd be much obliged | 16:22 |
| mhayden | odyssey4me gave some feedback already | 16:22 |
| * cloudnull looking | 16:22 | |
| *** jwagner has joined #openstack-ansible | 16:23 | |
| cloudnull | i can look that over in a bit | 16:23 |
| mhayden | relevant review -> https://review.openstack.org/#/c/245813/ | 16:23 |
| *** jwagner has quit IRC | 16:23 | |
| mhayden | it didn't get linked to the ticket for some reason :/ | 16:23 |
| *** jwagner has joined #openstack-ansible | 16:24 | |
| cloudnull | its all good. | 16:24 |
| cloudnull | it happens | 16:24 |
| *** daneyon has quit IRC | 16:24 | |
| mhayden | once the gate checks are ready for check/functional, i should be able to get this thing looking more like a finisheed product | 16:24 |
| mhayden | it might even impress mattt... just maybe | 16:24 |
| mhayden | ;) | 16:25 |
| odyssey4me | mhayden fyi it didn't get linked because your first commit didn't have the closes-bug line in the commit message :) | 16:25 |
| mhayden | i figured it was my error :( | 16:25 |
| mhayden | i added to the bug ticket | 16:25 |
| cloudnull | what else we got? | 16:26 |
| palendae | mhayden: Yeah, has to be the first commit. Subsequent ones won't do it | 16:26 |
| odyssey4me | subsequent additions to the commit message will only reflect in the bug when it merges | 16:26 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Updated the py_pkgs lookup plugin for multi source https://review.openstack.org/241483 | 16:29 |
| cloudnull | mattt hughsaunders ^ | 16:29 |
| cloudnull | hughsaunders: i made the changes on your box , you should be able to clean that one up now | 16:30 |
| cloudnull | if you want that is | 16:30 |
| *** jwagner has quit IRC | 16:30 | |
| *** phiche has joined #openstack-ansible | 16:30 | |
| *** sdake has quit IRC | 16:30 | |
| *** mpavone has quit IRC | 16:31 | |
| hughsaunders | cloudnull: I'll reuse it for testing your next patchset :) | 16:31 |
| cloudnull | its currently checked out on your box | 16:31 |
| bgmccollum | anyone know the difference between libvirt cpu_modes `host-model` and `host-passthrought`...? | 16:32 |
| bgmccollum | for reference -- https://github.com/openstack/nova/blob/78d07cce136467328eeabeeba90a8e914635ec85/nova/virt/libvirt/driver.py#L219-L230 | 16:32 |
| *** jwagner has joined #openstack-ansible | 16:33 | |
| *** jwagner has quit IRC | 16:33 | |
| cloudnull | bgmccollum: https://wiki.openstack.org/wiki/LibvirtXMLCPUModel | 16:34 |
| bgmccollum | yeah stumbled upon that...nova | 16:35 |
| bgmccollum | nova's default is host-model...but it says..."Beware, due to the way libvirt detects host CPU and due to the fact libvirt does not talk to QEMU/KVM when creating the CPU model, CPU configuration created using host-model may not work as expected." | 16:35 |
| bgmccollum | great | 16:35 |
| odyssey4me | yeah, host-model is meant to be the safer, more compatible one for situations where you want to live migrate between compute hosts that are different | 16:36 |
| odyssey4me | the most safe way is to specify a custom model and keep that consistent across all hosts, or at least across a host aggregate | 16:37 |
| cloudnull | bgmccollum: i think you can blanket OpenStack with that statement. . . "Beware, due to the way OpenStack detects bits and due to the fact OpenStack does not talk to some bits when creating things, configurations using somethings may not work as expected." | 16:37 |
| * cloudnull a helper | 16:37 | |
| bgmccollum | lulz | 16:38 |
| bgmccollum | cpu_mode: custom | 16:39 |
| bgmccollum | cpu_model: 486 | 16:39 |
| bgmccollum | enjoy your cloud | 16:39 |
| odyssey4me | lol | 16:39 |
| bgmccollum | that shit will work everywhere | 16:39 |
| cloudnull | cpu_model: 486dx2 # for performance | 16:39 |
| *** sacharya has joined #openstack-ansible | 16:39 | |
| bgmccollum | cpu_model: ppc64 | 16:40 |
| *** gparaskevas has quit IRC | 16:40 | |
| bgmccollum | cpu_model: 8088 | 16:40 |
| bgmccollum | blinkenlights | 16:40 |
| *** jwagner has joined #openstack-ansible | 16:41 | |
| mhayden | odyssey4me: did you want me to whip out a scalpel and pull out unneeded stuff from the scripts library? | 16:41 |
| odyssey4me | mhayden I think for now a thorough review is more important... let's find all the bugs and clean them up | 16:42 |
| mhayden | ah, so did you want me to take action on your comment for now? | 16:42 |
| * mhayden is waiting on jenkins to finish | 16:43 | |
| odyssey4me | mhayden nope - just test it, review how it's done, see if you find any issues | 16:43 |
| mhayden | can do | 16:43 |
| odyssey4me | let's mark all the issues in this review, then work through the patch in detail tomorrow (I need a fresh head) | 16:43 |
| mhayden | hah i can imagine | 16:44 |
| mhayden | i think i lost brain cells working through these patches | 16:44 |
| mhayden | also, the ssh errors randomly popping up along the way didn't help :P | 16:45 |
| odyssey4me | even without the scripts library removal, that's over 500 lines of bash code removed :) | 16:45 |
| mhayden | woot | 16:45 |
| mhayden | not that i dislike bash or anything, but i like seeing more ansible than bash :P | 16:45 |
| odyssey4me | go on mhayden - do us a meme for that | 16:45 |
| mhayden | WAT | 16:45 |
| *** mgoddard_ has joined #openstack-ansible | 16:46 | |
| odyssey4me | well, I suppose I can take you through it now if you like? | 16:46 |
| odyssey4me | my brain seems to be enjoying some caffiene | 16:46 |
| bgmccollum | anyone noticing a difference in SSH failure? | 16:46 |
| mhayden | odyssey4me: i've gotta scoot to meet my wife for lunch | 16:46 |
| bgmccollum | hopefully a reduction | 16:46 |
| odyssey4me | bgmccollum haven't seen one in a while now | 16:46 |
| bgmccollum | nod...disk prep for HP too right? | 16:47 |
| odyssey4me | bgmccollum yeah, the combo of both has made master/liberty work nicely - https://review.openstack.org/239525 has converted the bash prep scripts to ansible, which is a lot cleaner and easier to read | 16:48 |
| odyssey4me | including a pre-flight check :) https://review.openstack.org/#/c/239525/64/tests/roles/bootstrap-host/tasks/check-requirements.yml,cm | 16:48 |
| bgmccollum | nice... | 16:48 |
| *** mgoddard has quit IRC | 16:49 | |
| *** fawadkhaliq has quit IRC | 16:50 | |
| spotz | bgmccollum Are you talking recently? They were awful during the handon session at summit:( | 16:51 |
| bgmccollum | spotz: yes, specifically SSH failure in gate checks | 16:52 |
| bgmccollum | not SSH failures in general ;) | 16:52 |
| *** daneyon has joined #openstack-ansible | 16:53 | |
| spotz | ahh yeah I haven't submitted anything for you guys in a bit and that was a doc fix. So.. no comment:) | 16:55 |
| *** sdake has joined #openstack-ansible | 16:57 | |
| odyssey4me | mhayden are you ok with me pulling out the red hat support for now with the view that it gets added in a subsequent patch that can properly test it? | 17:01 |
| *** mgoddard has joined #openstack-ansible | 17:01 | |
| *** mgoddard_ has quit IRC | 17:01 | |
| bgmccollum | linggao: glad you got it working...! | 17:02 |
| linggao | thanks bgmccollum! | 17:03 |
| bgmccollum | linggao: do you mind trying something else? | 17:04 |
| *** daneyon_ has joined #openstack-ansible | 17:05 | |
| bgmccollum | linggao: instead of letting libvirt try to identiy the CPU make / model, i want to try specifying it explicitly... | 17:05 |
| linggao | bgmccollum, yes. Please let me know how. | 17:06 |
| *** daneyon has quit IRC | 17:08 | |
| bgmccollum | linggao: in your /etc/nova/nova.conf, under the [libvirt] section...set cpu_mode to custom, and cpu_model to Haswell...restart libvirt-bin and nova-compute...check your nova-compute.log and libvirtd.log for errors... | 17:08 |
| bgmccollum | if it doesnt like that model...then you can step back one generation at a time...SandyBridge, Westmere, Nehalem, Penryn, Conroe etc... | 17:11 |
| linggao | bgmccollum, should I change the cpu_map.xml to original or use the one I just modified? | 17:13 |
| bgmccollum | to be scientific, you should backup the working cpu_map.xml, and revert to the original... | 17:13 |
| linggao | ok | 17:14 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Convert AIO bootstrap from bash to Ansible https://review.openstack.org/239525 | 17:15 |
| *** linggao has quit IRC | 17:16 | |
| *** linggao has joined #openstack-ansible | 17:16 | |
| bgmccollum | linggao: if none of that works, we have 1 more thing to try... | 17:18 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Convert AIO bootstrap from bash to Ansible https://review.openstack.org/239525 | 17:20 |
| *** fawadkhaliq has joined #openstack-ansible | 17:22 | |
| *** phiche has quit IRC | 17:24 | |
| *** phiche has joined #openstack-ansible | 17:25 | |
| *** baker has quit IRC | 17:31 | |
| *** alkari has joined #openstack-ansible | 17:34 | |
| linggao | bgmccollum, I have tried a few. Haswell, SandyBridge, Westmere, Nehalem, Penryn, Conroe. None of them worked. | 17:36 |
| bgmccollum | linggao: ok...one last thing to try... | 17:36 |
| bgmccollum | linggao: can you comment out the cpu_model, and set cpu_mode to host-passthrough? | 17:37 |
| *** karimb_ has quit IRC | 17:37 | |
| bgmccollum | restart and tail logs as before... | 17:37 |
| bgmccollum | and if that doesnt work either...then its back to cpu_mode host-model, and your custom cpu_map.xml... :/ | 17:38 |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Add new alarm for the galera checks https://review.openstack.org/251890 | 17:38 |
| linggao | bgmccollum, I did not work. Same errors in both logs. | 17:40 |
| *** baker has joined #openstack-ansible | 17:40 | |
| bgmccollum | dang...ok. thanks for checking. sounds like those options won't work for us either...cc odyssey4me | 17:40 |
| *** gtt116__ has joined #openstack-ansible | 17:42 | |
| bgmccollum | linggao: so for now...put your custom cpu_map.xml back in place, set cpu_mode = host-model, remove cpu_model option, and restart the bits... | 17:42 |
| *** gtt116_ has quit IRC | 17:45 | |
| bgmccollum | food... | 17:50 |
| mhayden | odyssey4me: sounds okay with me | 17:51 |
| linggao | bgmccollum, sure. thanks a lot! | 17:51 |
| linggao | food too.... | 17:51 |
| odyssey4me | mhayden I found some leftover bits, and cleaned up the scripts library a bit too :) | 17:51 |
| mhayden | gracias | 17:52 |
| odyssey4me | mhayden there does appear to be an issue with some of the new providers in the gate check - the updated patch has failed in the voting gate already, but is waiting for the RAX check to complete. | 17:52 |
| mhayden | darn | 17:53 |
| odyssey4me | it would seem that this new provider either doesn't have a second disk, or the bash to find the largest disk is wonky | 17:53 |
| odyssey4me | this rendered a blank result: lsblk -brndo NAME,TYPE,RO,SIZE | awk '/d[b-z]+ disk 0/{ if ($4>m){m=$4; d=$1}}; END{print d}' | 17:54 |
| cloudnull | any chance i can get some love on https://review.openstack.org/#/c/241483/ ? | 17:54 |
| odyssey4me | cloudnull I'll ask mattt and hughsaunders to re-test/re-review tomorrow, and I'll also re-test. | 17:56 |
| prometheanfire | have people been seeing can't connect ssh issues since https://github.com/openstack/openstack-ansible/commit/fa0772aa1b0d036e8b7e490a37cd37a65d6ee50a ? | 17:57 |
| odyssey4me | it'd be nice to get some other reviews though, perhaps sigmavirus24_awa or stevelle have some time to look through it? | 17:57 |
| odyssey4me | prometheanfire nope, that was committed to resolve ongoing ssh connection issues | 17:57 |
| prometheanfire | cool | 17:57 |
| * mhayden is in git branching hell | 17:58 | |
| prometheanfire | just told qe to do that to fix that | 17:58 |
| prometheanfire | mhayden: why? | 17:58 |
| mhayden | just shuffling some things around and got my branches mixed up :) | 17:58 |
| mhayden | trying to bring in odyssey4me's new stuff | 17:58 |
| prometheanfire | fail | 17:58 |
| *** dmsimard is now known as dmsimard|afk | 18:00 | |
| mhayden | odyssey4me: i think bootstrap_host_data_disk_device comes back as defined, but empty | 18:01 |
| *** sdake has quit IRC | 18:01 | |
| odyssey4me | mhayden which frighteningly may mean there is only one disk - we'll have to wait for the whole job to complete to review all the logs | 18:01 |
| mhayden | odyssey4me: the builds on raxcloud won't have a data disk | 18:02 |
| odyssey4me | mhayden they do ;) | 18:02 |
| mhayden | really? | 18:02 |
| odyssey4me | mhayden the flavor used by infra, yes | 18:02 |
| mhayden | they didn't when i was hacking through the bootstrap stuff, unless i totally missed it | 18:03 |
| odyssey4me | mhayden see this example: http://logs.openstack.org/90/247290/4/check/gate-openstack-ansible-commit-nv/70d7c0b/logs/instance-info/host_info_1448450817.log | 18:03 |
| odyssey4me | notice xvde | 18:03 |
| mhayden | weird | 18:04 |
| mhayden | i wonder if that device has a partition table when the instance first comes up | 18:04 |
| odyssey4me | mhayden nope: http://logs.openstack.org/90/247290/4/check/gate-openstack-ansible-commit-nv/70d7c0b/logs/instance-info/host_info_1448447420.log | 18:04 |
| mhayden | imma wander back to the security stuff until jenkins finishes doing its thing ;) | 18:04 |
| mhayden | odyssey4me: wait, block devices section shows xvde + xvde1 | 18:05 |
| odyssey4me | yeah, so initially it's partitioned but not mounted | 18:05 |
| mhayden | oh i was talking about the dsvm, not the nv | 18:05 |
| odyssey4me | we trash and make it however we want it | 18:05 |
| odyssey4me | the nv build is always rax right now :) | 18:06 |
| odyssey4me | the dsvm build is all over the place | 18:06 |
| odyssey4me | here's a rax dsvm build: http://logs.openstack.org/90/247290/4/check/gate-openstack-ansible-dsvm-commit/6550631/logs/instance-info/host_info_1448447570.log | 18:06 |
| *** sdake has joined #openstack-ansible | 18:06 | |
| odyssey4me | see under block devices - xvde is still there | 18:07 |
| odyssey4me | the builds are using the same flavor, just a different image | 18:07 |
| mhayden | weird, okay | 18:07 |
| *** alkari1 has joined #openstack-ansible | 18:09 | |
| *** alkari has quit IRC | 18:10 | |
| *** eil397 has joined #openstack-ansible | 18:14 | |
| *** linggao_ has joined #openstack-ansible | 18:16 | |
| *** linggao has quit IRC | 18:16 | |
| *** iceyao has quit IRC | 18:16 | |
| *** sdake has quit IRC | 18:16 | |
| *** iceyao has joined #openstack-ansible | 18:22 | |
| *** alkari has joined #openstack-ansible | 18:22 | |
| *** phiche has quit IRC | 18:23 | |
| *** alkari1 has quit IRC | 18:23 | |
| *** mgoddard has quit IRC | 18:24 | |
| *** mancdaz has quit IRC | 18:26 | |
| odyssey4me | OMG, the blocker to LB going voting in the gate is finally voted through: https://review.openstack.org/193485 | 18:27 |
| *** mancdaz has joined #openstack-ansible | 18:27 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Adjusting commonly failing tasks https://review.openstack.org/252009 | 18:27 |
| *** galstrom_zzz is now known as galstrom | 18:41 | |
| *** baker has quit IRC | 18:42 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Functional backport for the new repo-build process https://review.openstack.org/244215 | 18:45 |
| kysse | mancdaz: all running and up | 18:47 |
| cloudnull | stevelle: logan- mgariepy https://review.openstack.org/#/c/244215/ was updated to rebase off the head of kilo and to pull in the changes from master. if you all can revote on that it would be appreciated. | 18:47 |
| cloudnull | kysse: working ? | 18:47 |
| kysse | nope | 18:48 |
| kysse | 16:15 mattt| kysse: do a nova service-list | 18:48 |
| kysse | I'm thinking about downgrading ansible playbook | 18:48 |
| cloudnull | what about `neutron agent-list` | 18:48 |
| d9k | :-) | 18:49 |
| kysse | all true and :-) | 18:49 |
| cloudnull | kysse: so i did an install last night of 10 nodes off the head of liberty and I cant recreate the issue ... | 18:49 |
| kysse | hngh. | 18:49 |
| kysse | and compute node should not have any lxc containers? | 18:49 |
| cloudnull | and these were 10 physical nodes | 18:49 |
| cloudnull | correct | 18:50 |
| kysse | this is killing me | 18:50 |
| logan- | cloudnull: ill do a build an revote in a few thanks! | 18:50 |
| cloudnull | thank you ! | 18:50 |
| cloudnull | on the neutron agent container we should be seeing traffic and logs indicating bridges and namespaces are being created. | 18:50 |
| cloudnull | kysse: your saying that the neutron agent container has no logs | 18:51 |
| kysse | well, it does now. But no errors. | 18:51 |
| cloudnull | do you mind clearing the log file, restarting the service, building a vm and posting that data ? | 18:51 |
| kysse | ok | 18:52 |
| kysse | cont1 http://sprunge.us/KMVg cont2 http://sprunge.us/KMVg | 18:55 |
| kysse | plugin rpc error appears when restarting service | 18:56 |
| *** phiche has joined #openstack-ansible | 18:56 | |
| kysse | arght. cont1 //sprunge.us/QcVZ cont2 http://sprunge.us/KMVg | 18:56 |
| cloudnull | so that error brings up https://bugs.launchpad.net/neutron/+bug/1470584 | 18:59 |
| openstack | Launchpad bug 1470584 in neutron "Linux bridge agent fails badly with physical_interface_mappings issue" [Undecided,Fix released] - Assigned to Kevin Benton (kevinbenton) | 18:59 |
| cloudnull | which was raised by Sam-I-Am | 18:59 |
| cloudnull | and references at the bottom "on Liberty for specifying a wrong interface." | 18:59 |
| *** baker has joined #openstack-ansible | 19:00 | |
| cloudnull | all of this leads me to believe the interface mapping is still wrong | 19:03 |
| cloudnull | kysse: whats the `physical_interface_mappings` say in the neutron agent container and the compute node ? | 19:05 |
| cloudnull | prometheanfire: looks like you ran into this a bit ago too http://eavesdrop.openstack.org/irclogs/%23openstack-neutron/%23openstack-neutron.2015-10-22.log.html ? | 19:07 |
| prometheanfire | which line? | 19:08 |
| cloudnull | RE: CRITICAL neutron [-] AttributeError: 'LinuxBridgeNeutronAgentRPC' object has no attribute 'plugin_rpc' | 19:09 |
| d9k | vlan:br-vlan : compute, vlan:eth11 : infra1_agentcont | 19:10 |
| prometheanfire | oh | 19:10 |
| openstackgerrit | Merged openstack/openstack-ansible: Update pinned lxc version https://review.openstack.org/251945 | 19:10 |
| prometheanfire | it db_upgrade didn't make an entry into a table | 19:10 |
| cloudnull | was that for security group related ? | 19:11 |
| prometheanfire | http://eavesdrop.openstack.org/irclogs/%23openstack-neutron/%23openstack-neutron.2015-10-22.log.html#t2015-10-22T19:30:19 | 19:11 |
| prometheanfire | insert into networksecuritybindings values ('34d13f35-d93d-4baf-b493-4b079518c789', 't'); | 19:12 |
| cloudnull | d9k: on compute br-vlan exists, on infra1_agentcont eth11 exists ? | 19:12 |
| prometheanfire | postgres syntax | 19:12 |
| d9k | yes | 19:12 |
| *** erikmwilson has quit IRC | 19:15 | |
| *** erikmwilson has joined #openstack-ansible | 19:16 | |
| *** linggao_ has quit IRC | 19:16 | |
| *** linggao_ has joined #openstack-ansible | 19:16 | |
| cloudnull | d9k: can you check your DB to see if the "networksecuritybindings" table exists? and that it has entries for your networks in it ? | 19:17 |
| cloudnull | d9k: kysse you shouls see something like http://cdn.pasteraw.com/lopivmwhlukzkwkirgrz1jzscfal7qr | 19:17 |
| cloudnull | if not we can add it like prometheanfire had to and see if theres some random issue causing that to fail in upstream neutron | 19:18 |
| d9k | http://sprunge.us/AfcM | 19:21 |
| prometheanfire | that looks right to me | 19:22 |
| cloudnull | ok well thats not it , it would seem | 19:22 |
| cloudnull | d9k: do you have any istances still in error mode ? | 19:26 |
| cloudnull | can you do a nova show <ID> | 19:26 |
| d9k | http://sprunge.us/WTid | 19:27 |
| cloudnull | also do you have the commands used when the networks were created ? | 19:27 |
| cloudnull | oh, nevermind i think they were done in horizon, right? | 19:28 |
| d9k | y | 19:28 |
| cloudnull | well thats showing "no valid host" do you have a lot of error mode VMs ? | 19:29 |
| kysse | 1 or two | 19:29 |
| cloudnull | Sam-I-Am: ^ is that still what nova will show even if theres a vif issue? | 19:29 |
| kysse | it always says no valid host | 19:31 |
| kysse | I thought it was cuz the "no binding plaaplaa" error | 19:31 |
| kysse | or what it was.. port binding failed or something | 19:31 |
| *** sdake has joined #openstack-ansible | 19:36 | |
| *** baker has quit IRC | 19:42 | |
| *** dmsimard|afk is now known as dmsimard | 19:45 | |
| *** karimb has joined #openstack-ansible | 19:46 | |
| Sam-I-Am | cloudnull: yeah thats the generic error message | 19:48 |
| *** karimb has quit IRC | 19:51 | |
| cloudnull | maybe this is all related to the way the networks were created at this point ? | 19:52 |
| cloudnull | kysse, Sam-I-Am, d9k ^ | 19:52 |
| kysse | maybe | 19:53 |
| *** sdake_ has joined #openstack-ansible | 19:55 | |
| *** sdake has quit IRC | 19:55 | |
| kysse | so using neutron cli then.. | 20:00 |
| *** sdake_ is now known as sdake | 20:03 | |
| Sam-I-Am | is the network available on the compute hosts? | 20:06 |
| kysse | .. how? | 20:07 |
| *** mss_ has joined #openstack-ansible | 20:08 | |
| *** fawadkhaliq has quit IRC | 20:09 | |
| *** mss has quit IRC | 20:11 | |
| Sam-I-Am | kysse: what kind of network are you trying to use? | 20:12 |
| kysse | vlan | 20:12 |
| Sam-I-Am | provider or private? | 20:12 |
| kysse | provider | 20:14 |
| Sam-I-Am | is the interface (or bridge) defined in interface_mappings available on your infra and compute nodes? | 20:16 |
| *** linggao_ has quit IRC | 20:16 | |
| *** linggao_ has joined #openstack-ansible | 20:16 | |
| kysse | yes it is. | 20:16 |
| kysse | hmh. | 20:16 |
| kysse | I created a network with neutron cli | 20:16 |
| kysse | network's ports show status down | 20:16 |
| Sam-I-Am | thats normal until you plug it into something | 20:18 |
| *** phiche1 has joined #openstack-ansible | 20:19 | |
| *** phiche has quit IRC | 20:20 | |
| linggao_ | Hi all, I got all-in-one installed and one instance deployed. But cannot ping the instance. I can get into the instance from the console on horizon. | 20:31 |
| linggao_ | any hint? | 20:31 |
| *** errr_ is now known as errr | 20:32 | |
| linggao_ | I am using the image from installed by os-tempest-install.yaml | 20:32 |
| bgmccollum | linggao: security groups? | 20:33 |
| linggao_ | bgmccollum, I did not specify security groups with nova boot command. | 20:35 |
| bgmccollum | you can change them after booting... | 20:35 |
| linggao_ | ok. let me try... | 20:36 |
| cloudnull | kysse: anything when creating the network by hand ? | 20:45 |
| linggao_ | bgmccollum, it works now. thanks again :-) | 20:57 |
| bgmccollum | linggao_ great...! | 20:57 |
| kysse | nope not really | 21:00 |
| *** sdake has quit IRC | 21:06 | |
| *** sdake has joined #openstack-ansible | 21:08 | |
| *** phiche has joined #openstack-ansible | 21:09 | |
| *** phiche1 has quit IRC | 21:09 | |
| cloudnull | kysse: i really dont know what the blocker is at this point. | 21:10 |
| kysse | yep | 21:12 |
| kysse | what version of openstack-ansible you suggest? | 21:12 |
| *** sdake has quit IRC | 21:12 | |
| *** jbweber has quit IRC | 21:13 | |
| *** daneyon_ has quit IRC | 21:13 | |
| *** sdake has joined #openstack-ansible | 21:14 | |
| *** jbweber has joined #openstack-ansible | 21:15 | |
| *** linggao_ has quit IRC | 21:16 | |
| *** linggao_ has joined #openstack-ansible | 21:17 | |
| cloudnull | kysse: liberty is working for me off the head of the liberty branch. you could give 11.2.5 a go which'll deploy kilo | 21:20 |
| *** javeriak_ has quit IRC | 21:21 | |
| kysse | I'm just wordless cuz I dont get _any_ good log error message for this problem. I'd hate to run something old. | 21:21 |
| kysse | maybe I just have to create virtual enviroment for this deployment stuff. It's not fun to reainstall all those machines and configure network every time | 21:22 |
| kysse | :p | 21:22 |
| *** ctina_ has joined #openstack-ansible | 21:22 | |
| kysse | what you had to do to repeat that error message? | 21:25 |
| *** ctina has quit IRC | 21:26 | |
| *** ctina_ has quit IRC | 21:27 | |
| *** harlowja has quit IRC | 21:27 | |
| *** harlowja has joined #openstack-ansible | 21:28 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Fixes playbook runtime issues with ldap https://review.openstack.org/248900 | 21:28 |
| cloudnull | kysse: the only way I could get the vif error was to either take a network away from one of my compute nodes | 21:29 |
| kysse | ok | 21:29 |
| cloudnull | or to mess with the provider networks such that it created a broken physical map | 21:30 |
| cloudnull | like having the flat network bound to the same interface | 21:30 |
| cloudnull | etc | 21:30 |
| *** phiche has quit IRC | 21:30 | |
| *** sdake has quit IRC | 21:30 | |
| kysse | yepyep | 21:31 |
| *** mancdaz has quit IRC | 21:31 | |
| *** mancdaz has joined #openstack-ansible | 21:32 | |
| *** sdake has joined #openstack-ansible | 21:34 | |
| cloudnull | so im once again rekicking my lab will try to reproduce the issue | 21:35 |
| kysse | heh | 21:36 |
| mhayden | odyssey4me: i did a build of the ansible bootstrap stuff on rax with an instance that has extra storage -- it worked just fine | 21:36 |
| mhayden | it picked up xvde | 21:36 |
| *** sdake has quit IRC | 21:41 | |
| cloudnull | kysse: its quite perplexing why you're env is not happy but im intent on trying to get to the bottom of i t | 21:50 |
| kysse | indeed | 21:50 |
| kysse | we are doing hard debugging also. | 21:50 |
| *** baker has joined #openstack-ansible | 22:06 | |
| *** Mudpuppy has quit IRC | 22:06 | |
| *** ysm has joined #openstack-ansible | 22:13 | |
| *** linggao__ has joined #openstack-ansible | 22:16 | |
| *** linggao_ has quit IRC | 22:16 | |
| *** sacharya has quit IRC | 22:18 | |
| *** harlowja has quit IRC | 22:19 | |
| *** harlowja has joined #openstack-ansible | 22:21 | |
| *** galstrom is now known as galstrom_zzz | 22:31 | |
| *** daneyon has joined #openstack-ansible | 22:37 | |
| *** mcarden_ has joined #openstack-ansible | 22:45 | |
| *** mcarden has quit IRC | 22:45 | |
| *** mcarden_ is now known as mcarden | 22:46 | |
| *** daneyon has quit IRC | 22:50 | |
| *** harlowja has quit IRC | 22:57 | |
| cloudnull | kysse: still around ? | 22:58 |
| kysse | yes | 22:59 |
| kysse | I'm running it again now, so I cant test. | 23:00 |
| cloudnull | when your run is complete can you set "neutron_l2_population: true" in your user variables file and rerun the os-neutron-install.yml play | 23:00 |
| kysse | sure but whe deviced to run kilo first, then we can upgrade and test that. | 23:01 |
| kysse | we decided* | 23:01 |
| *** harlowja has joined #openstack-ansible | 23:01 | |
| cloudnull | I'm attempting to recreate your exact env w/ the config details you've provided and in my tests so far that is one thing that has differed between my installs and yours | 23:02 |
| kysse | this is good training for us too. | 23:02 |
| kysse | mmh ok | 23:02 |
| cloudnull | i have it set, you do not, because the playbook assumes the use of l3ha when more than one network node is present | 23:02 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/tasks/neutron_post_install.yml#L16-L34 | 23:02 |
| kysse | btw, is it possible to set flat network point to a bridge? | 23:03 |
| cloudnull | l2 population enabled is the default in kilo | 23:03 |
| kysse | ok | 23:03 |
| cloudnull | yes | 23:03 |
| cloudnull | you can | 23:03 |
| kysse | cool. | 23:03 |
| cloudnull | however you should hang a veth off of the bridge | 23:03 |
| cloudnull | and set the host bind override to the name of the veth interface | 23:03 |
| kysse | oh I see. | 23:04 |
| cloudnull | which is what we do here | 23:04 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/etc/network/interfaces.d/aio_interfaces.cfg#L52-L59 | 23:04 |
| cloudnull | then vlan, flat, even vxlan could all go off of the same bridge if you wanted it to | 23:04 |
| *** ysm has quit IRC | 23:06 | |
| cloudnull | FYI here are my logs for the neutron lxb agent on the compute node: http://cdn.pasteraw.com/9bv8qknv72x4so83ih1wg0rdsha9ljs | 23:07 |
| cloudnull | im not seeing that stacktrace in the logs when i restart the service | 23:08 |
| kysse | hmh. yep yep | 23:10 |
| kysse | thanks | 23:10 |
| cloudnull | so at this moment im thinking that the root cause is that l2 pop needs to be enabled. | 23:10 |
| cloudnull | but im testing more | 23:10 |
| cloudnull | so more soon | 23:10 |
| cloudnull | he might not be around however mattt i think ran into something similar a while back | 23:13 |
| cloudnull | if i remember right | 23:14 |
| *** daneyon has joined #openstack-ansible | 23:15 | |
| *** sdake has joined #openstack-ansible | 23:16 | |
| *** linggao__ has quit IRC | 23:16 | |
| *** linggao__ has joined #openstack-ansible | 23:16 | |
| *** linggao__ has quit IRC | 23:16 | |
| cloudnull | kysse: w/ l2pop disabled http://cdn.pasteraw.com/dg3bmy773jz1n7lfxl1wgge3aml6xt | 23:18 |
| kysse | hah.. | 23:19 |
| kysse | so you solved it, I think :-) | 23:19 |
| cloudnull | maybe | 23:19 |
| cloudnull | re-enabling and will rerun the test | 23:19 |
| cloudnull | kysse: are you on launchpad ? | 23:20 |
| kysse | I think I'll stop installing kilo, and test that with liberty | 23:20 |
| kysse | no | 23:20 |
| cloudnull | ok | 23:20 |
| kysse | I think I can register. | 23:20 |
| *** KLevenstein has quit IRC | 23:22 | |
| kysse | ville.vuorinen@multim.fi | 23:22 |
| cloudnull | confirmed http://cdn.pasteraw.com/nbz6csr3wf3xjh154b9p29uwwm28bob | 23:25 |
| cloudnull | the 2 active tests are w/ l2pop enabled | 23:25 |
| cloudnull | the 5 errors are w/ l2pop disabled | 23:25 |
| kysse | cool, cool cool cool. | 23:25 |
| *** darrenc is now known as darrenc_afk | 23:31 | |
| cloudnull | kysse https://bugs.launchpad.net/openstack-ansible/+bug/1521793 | 23:32 |
| openstack | Launchpad bug 1521793 in openstack-ansible "Master/Liberty w/ L2pop disabled breaks neutron" [High,Triaged] - Assigned to Kevin Carter (kevin-carter) | 23:32 |
| *** spotz is now known as spotz_zzz | 23:34 | |
| kysse | thanks | 23:34 |
| *** baker has quit IRC | 23:34 | |
| *** sdake has quit IRC | 23:36 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Fix neutron issue w/ l2pop https://review.openstack.org/252100 | 23:37 |
| cloudnull | ^ fix for the issue | 23:37 |
| *** networkninja has joined #openstack-ansible | 23:38 | |
| cloudnull | kysse: in your env when you reconfigure neutron you'll need to restart the neutron lxb service on your compute node and likely restart the neutron agent containers after the play has been completed. | 23:39 |
| *** sdake has joined #openstack-ansible | 23:39 | |
| kysse | ok | 23:39 |
| cloudnull | im off for a bit , but let me know how it goes | 23:41 |
| *** dstanek has quit IRC | 23:41 | |
| *** dstanek has joined #openstack-ansible | 23:41 | |
| *** daneyon has quit IRC | 23:42 | |
| *** daneyon has joined #openstack-ansible | 23:47 | |
| *** daneyon has quit IRC | 23:48 | |
| *** darrenc_afk is now known as darrenc | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
