*** markvoelker has quit IRC | 00:00 | |
*** rady has quit IRC | 00:00 | |
openstackgerrit | Matthew Kassawara proposed openstack/openstack-ansible-specs: Improve installation guide https://review.openstack.org/241037 | 00:02 |
---|---|---|
*** woodard has quit IRC | 00:08 | |
*** Mudpuppy has quit IRC | 00:32 | |
*** elo1 has joined #openstack-ansible | 00:37 | |
*** elo has joined #openstack-ansible | 00:47 | |
*** elo1 has quit IRC | 00:48 | |
*** phalmos has quit IRC | 00:55 | |
*** ybabenko has joined #openstack-ansible | 00:57 | |
*** markvoelker has joined #openstack-ansible | 01:00 | |
*** ybabenko has quit IRC | 01:01 | |
mss | I've hit a different issue http://pastebin.com/c8CrQUWx openstack-ansible keeps looping during install and failing at this point. Installing these packages manually doesn't seem to fix the problem | 01:02 |
mss | Though they do install successfully whet I try manually | 01:03 |
*** markvoelker has quit IRC | 01:05 | |
*** sdake has quit IRC | 01:07 | |
*** sdake has joined #openstack-ansible | 01:10 | |
*** ysm has joined #openstack-ansible | 01:13 | |
mss | nvm, possibly a mysqldb vs mariadb library conflict | 01:19 |
*** sdake has quit IRC | 01:19 | |
*** dolpher has joined #openstack-ansible | 01:34 | |
*** elo has quit IRC | 01:51 | |
*** markvoelker has joined #openstack-ansible | 02:01 | |
*** gouthamr_ has quit IRC | 02:03 | |
mss | It looks like the galera containers was trying to install the modules without an index. | 02:03 |
mss | Commenting out that line in the pip.conf of that container allowed me to install MySQL-python and python-memcached without issue | 02:04 |
mss | Is there a reason openstack-ansible sets the no-index flag in pip to true? | 02:06 |
*** markvoelker has quit IRC | 02:06 | |
*** tnarg has quit IRC | 02:12 | |
*** sdake has joined #openstack-ansible | 02:13 | |
*** tnarg has joined #openstack-ansible | 02:14 | |
*** mss has quit IRC | 02:22 | |
*** Bjoern_ is now known as Bjoern_zZzZzZzZ | 02:24 | |
*** mss has joined #openstack-ansible | 02:30 | |
*** sdake has quit IRC | 02:35 | |
*** daneyon has joined #openstack-ansible | 02:37 | |
*** Bjoern_zZzZzZzZ is now known as Bjoern_ | 02:38 | |
*** daneyon has quit IRC | 02:38 | |
*** ysm has quit IRC | 02:45 | |
*** Bjoern_ has quit IRC | 02:46 | |
*** markvoelker has joined #openstack-ansible | 03:02 | |
*** markvoelker has quit IRC | 03:07 | |
*** luckyinva has joined #openstack-ansible | 03:24 | |
*** elgertam has joined #openstack-ansible | 03:44 | |
*** sdake has joined #openstack-ansible | 03:47 | |
*** ybabenko has joined #openstack-ansible | 03:57 | |
*** ybabenko has quit IRC | 04:02 | |
*** elgertam1 has joined #openstack-ansible | 04:19 | |
*** elgertam has quit IRC | 04:21 | |
elgertam1 | I'm a total n00b here, but I'm wondering what benefits I get from installing Ceilometer | 04:23 |
elgertam1 | I'm trying the all-in-one in my home lab server | 04:23 |
*** sdake has quit IRC | 04:57 | |
*** sdake has joined #openstack-ansible | 04:58 | |
*** markvoelker has joined #openstack-ansible | 05:03 | |
*** markvoelker has quit IRC | 05:07 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 05:18 | |
cloudnull | elgertam1: ceilometer will give you metering as a service, which has a few nice metrics however to use it you'll need mongodb which is not generally supported by the project at this time, however if you use the all in one gate testing script it will create a functional mongo stack, though scale will be limited. | 05:22 |
cloudnull | if you dont need metering I'd suggest skipping it, unless your familiar with mongodb | 05:23 |
*** sigmavirus24 is now known as sigmavirus24_awa | 05:24 | |
elgertam1 | thanks! | 05:25 |
elgertam1 | @cloudnull: could I add it later using OSAD if I wanted to? | 05:29 |
elgertam1 | without too much weeping and gnashing of teeth | 05:29 |
cloudnull | absolutely | 05:29 |
cloudnull | its a simple change in the user config file and a matter of running the plays | 05:30 |
elgertam1 | excellent | 05:30 |
elgertam1 | I really appreciate the help | 05:30 |
cloudnull | OSAD was built to upgrade/extend so most, if not everything that we support, can be implemented in pieces as need. | 05:31 |
cloudnull | anytime elgertam1. ping out if you have any questions. | 05:31 |
cloudnull | also if you've not already seen it http://docs.openstack.org/developer/openstack-ansible/ | 05:31 |
cloudnull | online, living docs. | 05:32 |
elgertam1 | :D | 05:32 |
cloudnull | notes about configuring ceilometer http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-ceilometer.html | 05:32 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - OpenStack Hosts https://review.openstack.org/241102 | 05:34 |
cloudnull | im off have a good one. | 05:36 |
*** ashishb has joined #openstack-ansible | 06:14 | |
*** tnarg has quit IRC | 06:15 | |
*** tnarg has joined #openstack-ansible | 06:16 | |
*** cfarquhar has quit IRC | 06:19 | |
*** cfarquhar has joined #openstack-ansible | 06:21 | |
*** cfarquhar has quit IRC | 06:21 | |
*** cfarquhar has joined #openstack-ansible | 06:21 | |
*** tnarg has quit IRC | 06:21 | |
*** ybabenko has joined #openstack-ansible | 06:24 | |
*** ybabenko has quit IRC | 06:29 | |
*** alextric_ has joined #openstack-ansible | 06:30 | |
*** alextricity has quit IRC | 06:32 | |
*** rev_dr has quit IRC | 06:35 | |
*** rev_dr has joined #openstack-ansible | 06:38 | |
*** javeriak has joined #openstack-ansible | 06:41 | |
*** cfarquhar has quit IRC | 06:43 | |
*** openstackstatus has quit IRC | 06:49 | |
*** javeriak has quit IRC | 06:49 | |
*** openstackstatus has joined #openstack-ansible | 06:49 | |
*** ChanServ sets mode: +v openstackstatus | 06:49 | |
*** javeriak has joined #openstack-ansible | 06:49 | |
*** markvoelker has joined #openstack-ansible | 06:50 | |
*** cfarquhar has joined #openstack-ansible | 06:51 | |
*** cfarquhar has quit IRC | 06:51 | |
*** cfarquhar has joined #openstack-ansible | 06:51 | |
*** javeriak has quit IRC | 06:53 | |
*** mss has quit IRC | 06:54 | |
*** javeriak has joined #openstack-ansible | 06:55 | |
*** markvoelker has quit IRC | 06:55 | |
*** phiche has joined #openstack-ansible | 07:13 | |
*** shausy has joined #openstack-ansible | 07:20 | |
*** luckyinva has quit IRC | 07:28 | |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - pip install https://review.openstack.org/241125 | 07:42 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - pip install https://review.openstack.org/241125 | 07:43 |
*** mss has joined #openstack-ansible | 07:49 | |
*** mss has quit IRC | 07:54 | |
odyssey4me | mss make sure that you 'apt-get purge -y libmysqlclient18 mysql-common' (this is done in the AIO bootstrap script) - to get rid of any mysql libraries that may be on your host before installing | 07:57 |
*** slotti has joined #openstack-ansible | 08:19 | |
*** ybabenko has joined #openstack-ansible | 08:31 | |
*** ybabenko has quit IRC | 08:32 | |
*** subscope has joined #openstack-ansible | 08:39 | |
*** ybabenko has joined #openstack-ansible | 08:40 | |
*** shausy has quit IRC | 08:47 | |
*** javeriak has quit IRC | 08:48 | |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - rsyslog_role https://review.openstack.org/241138 | 08:48 |
*** shausy has joined #openstack-ansible | 08:48 | |
*** markvoelker has joined #openstack-ansible | 08:52 | |
mattt | cloudnull: bro | 08:52 |
cloudnull | o/ | 08:52 |
cloudnull | is 17:30 in my head | 08:53 |
mattt | cloudnull: :) | 08:53 |
cloudnull | i cant sleep, at all... i think i may need to stay up for like 24 hours. | 08:53 |
cloudnull | then just crash | 08:53 |
*** markvoelker has quit IRC | 08:56 | |
*** subscope has quit IRC | 09:01 | |
*** sdake has quit IRC | 09:02 | |
*** metral is now known as metral_zzz | 09:03 | |
*** metral_zzz is now known as metral | 09:03 | |
*** sdake has joined #openstack-ansible | 09:03 | |
*** subscope has joined #openstack-ansible | 09:04 | |
*** tiagogomes_ has quit IRC | 09:05 | |
*** tiagogomes has joined #openstack-ansible | 09:05 | |
*** shausy has quit IRC | 09:20 | |
*** karimb has joined #openstack-ansible | 09:22 | |
*** shausy has joined #openstack-ansible | 09:22 | |
*** dolpher has quit IRC | 09:27 | |
*** shausy has quit IRC | 09:30 | |
*** shausy has joined #openstack-ansible | 09:31 | |
*** javeriak has joined #openstack-ansible | 09:32 | |
*** subscope has quit IRC | 09:40 | |
*** mss has joined #openstack-ansible | 09:50 | |
*** markvoelker has joined #openstack-ansible | 09:53 | |
*** mss has quit IRC | 09:55 | |
*** markvoelker has quit IRC | 09:58 | |
*** ybabenko has quit IRC | 09:59 | |
*** subscope has joined #openstack-ansible | 09:59 | |
*** jhesketh has quit IRC | 10:09 | |
*** tiagogomes has quit IRC | 10:09 | |
*** tiagogomes has joined #openstack-ansible | 10:10 | |
*** jhesketh has joined #openstack-ansible | 10:10 | |
*** ybabenko has joined #openstack-ansible | 10:16 | |
*** ybabenko has quit IRC | 10:19 | |
*** subscope has quit IRC | 10:20 | |
*** subscope has joined #openstack-ansible | 10:20 | |
evrardjp | hello everyone | 10:25 |
evrardjp | because I missed the last weeks, I'd like to know the current status of liberty release... is 12.0.0 worth testing (is it stable to put customers on it?) or should I wait for 12.0.1? | 10:27 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - LXC Host https://review.openstack.org/241159 | 10:28 |
evrardjp | cloudnull: is os-cloud an organization with your repos? | 10:31 |
cloudnull | it is | 10:31 |
cloudnull | evrardjp: i'd wait for 12.0.1 there were two issues fixed right after release that I'd hold out for . | 10:32 |
evrardjp | ok thanks | 10:32 |
evrardjp | do you know what's the timeframe? | 10:32 |
cloudnull | evrardjp: if you we're going to deploy liberty for testing , the branch should be good to go | 10:32 |
cloudnull | odyssey4me: ^ time for liberty tag ? | 10:32 |
evrardjp | could you explain what are the two issues? | 10:35 |
odyssey4me | evrardjp I would say that liberty is fine for fresh deployments, but we haven't had anyone do thorough testing for upgrades yet... | 10:35 |
odyssey4me | this needs a review: https://review.openstack.org/240560 | 10:36 |
evrardjp | odyssey4me: yeah, I'm talking about fresh right now | 10:36 |
evrardjp | I'll check on the review | 10:37 |
odyssey4me | and there's a different patch for liberty that needs to go in which is functionally the same, but implemented differently as the mitaka/liberty upstream code is different | 10:37 |
evrardjp | I see only one bug for target release 12.0.1 that's why I'm surprised with the 2 bugs | 10:37 |
cloudnull | https://github.com/openstack/openstack-ansible/commit/7d4d7d586e5f9ca55147ee51feae8f960753cda5 https://github.com/openstack/openstack-ansible/commit/fe3f6c8588ff64d5b8ff1931ddcf61f7e9d36f7f | 10:38 |
cloudnull | and then there was an update to the SHAs being used | 10:38 |
cloudnull | https://github.com/openstack/openstack-ansible/commit/0242bbf664a95f8c7ea0b35d9ebc4750f35823ad | 10:38 |
odyssey4me | I can tag any time, but I thought it might be good to discuss it on Thu and do it then | 10:38 |
odyssey4me | a lot of people are still waking up from Tokyo | 10:38 |
* cloudnull still have not slept | 10:39 | |
evrardjp | yeah cloudnull, it's a little "different" to see you at this time of the day... did you change timezone? ;) | 10:39 |
cloudnull | im back home. | 10:43 |
cloudnull | however the body thinks is JP time still | 10:43 |
openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible: Update juno for new dev work - 2 Nov 2015 https://review.openstack.org/240911 | 10:44 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - pip install https://review.openstack.org/241163 | 10:44 |
evrardjp | for this one: https://review.openstack.org/#/c/240560/ I'm not sure I can help... I'm not a neutron expert... | 10:45 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - py from git https://review.openstack.org/241163 | 10:46 |
odyssey4me | evrardjp as with most patches, you can check syntax, style and test it :) | 10:47 |
cloudnull | so https://review.openstack.org/#/q/owner:%22Kevin+Carter%22+status:open+project:openstack/openstack-ansible-specs,n,z the specs are online to be approved. I think I'm going to hold off a bit before doing too much more so that I dont have to go into rebase hell. | 10:51 |
cloudnull | the current irr implementations that I would be migrating over are in os-cloud which can be used to test the role for now. | 10:51 |
cloudnull | os-cloud https://github.com/os-cloud | 10:52 |
*** subscope has quit IRC | 10:53 | |
odyssey4me | cloudnull one small spelling error for https://review.openstack.org/241163 | 10:54 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - py from git https://review.openstack.org/241163 | 10:55 |
cloudnull | odyssey4me: ^ updated | 10:56 |
odyssey4me | cloudnull small nit on https://review.openstack.org/241159 | 10:59 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - LXC Host https://review.openstack.org/241159 | 11:01 |
*** markvoelker has joined #openstack-ansible | 11:09 | |
*** markvoelker has quit IRC | 11:13 | |
odyssey4me | meteorfox ping | 11:21 |
odyssey4me | alextric_ ping | 11:21 |
*** mss has joined #openstack-ansible | 11:35 | |
*** mss has quit IRC | 11:39 | |
*** mgoddard_ has joined #openstack-ansible | 11:44 | |
*** mgoddard has quit IRC | 11:44 | |
*** ybabenko has joined #openstack-ansible | 11:45 | |
*** ybabenko has quit IRC | 11:50 | |
openstackgerrit | Matt Thompson proposed openstack/openstack-ansible: Reference correct group in haproxy_config.yml https://review.openstack.org/241173 | 11:50 |
odyssey4me | mattt can you review https://review.openstack.org/240911 please? | 11:52 |
mattt | odyssey4me: sure | 11:52 |
*** ybabenko has joined #openstack-ansible | 11:54 | |
*** wtie has joined #openstack-ansible | 11:55 | |
*** javeriak has quit IRC | 11:58 | |
*** javeriak has joined #openstack-ansible | 12:08 | |
*** markvoelker has joined #openstack-ansible | 12:09 | |
*** subscope has joined #openstack-ansible | 12:10 | |
openstackgerrit | Matt Thompson proposed openstack/openstack-ansible: Update neutron_migrations_facts https://review.openstack.org/240560 | 12:12 |
*** markvoelker has quit IRC | 12:14 | |
*** shausy has quit IRC | 12:16 | |
*** shausy has joined #openstack-ansible | 12:17 | |
*** lkoranda has quit IRC | 12:21 | |
*** lkoranda has joined #openstack-ansible | 12:22 | |
*** subscope has quit IRC | 12:24 | |
*** subscope has joined #openstack-ansible | 12:26 | |
*** javeriak has quit IRC | 12:29 | |
*** lkoranda has quit IRC | 12:33 | |
*** lkoranda has joined #openstack-ansible | 12:34 | |
*** lkoranda has quit IRC | 12:35 | |
*** lkoranda has joined #openstack-ansible | 12:37 | |
mattt | mhayden: http://hardening.io/ | 12:39 |
*** lkoranda has quit IRC | 12:39 | |
*** lkoranda has joined #openstack-ansible | 12:41 | |
*** lkoranda has quit IRC | 12:45 | |
*** mgoddard__ has joined #openstack-ansible | 12:45 | |
*** mgoddard__ has quit IRC | 12:46 | |
*** lkoranda has joined #openstack-ansible | 12:46 | |
*** mgoddard__ has joined #openstack-ansible | 12:46 | |
*** mgoddard_ has quit IRC | 12:48 | |
*** ybabenko has quit IRC | 12:49 | |
*** lkoranda has quit IRC | 12:49 | |
*** subscope has quit IRC | 12:50 | |
*** lkoranda has joined #openstack-ansible | 12:51 | |
*** lkoranda has quit IRC | 12:54 | |
*** shausy has quit IRC | 12:55 | |
*** lkoranda has joined #openstack-ansible | 12:56 | |
mhayden | morning@! | 12:57 |
mhayden | mattt: saw that, but it's really incomplete | 12:58 |
*** subscope has joined #openstack-ansible | 12:58 | |
*** subscope has quit IRC | 13:00 | |
mhayden | mattt: i may have gotten somewhere on the random aio failures -> http://logs.openstack.org/25/239525/34/check/gate-openstack-ansible-dsvm-commit/ada8f24/console.html#_2015-11-02_22_41_39_591 | 13:01 |
mhayden | i'm not sure what ends up in ~/.ssh/known_hosts | 13:02 |
mattt | mhayden: weird :-/ | 13:02 |
mhayden | i'm wondering if i ought to clear out ~/.ssh/known_hosts | 13:02 |
*** ybabenko has joined #openstack-ansible | 13:10 | |
*** ashishb has quit IRC | 13:12 | |
odyssey4me | it may have to do with the switch in ssh keys within the bootstrap | 13:14 |
odyssey4me | the keys shouldn't be changed if they already exist, ideally - they should only be changed if a component is missing | 13:15 |
odyssey4me | so if the private key or the public key is missing, recreate them both | 13:15 |
odyssey4me | but yeah, it may then require also removing the known_hosts | 13:16 |
mhayden | odyssey4me: i'll give that a go | 13:20 |
odyssey4me | I'd say that the removal of known_hosts should only happen if the keys change | 13:21 |
mhayden | which keys, specifically? | 13:21 |
mhayden | id_rsa / id_rsa.pub? | 13:21 |
mhayden | or host keys | 13:21 |
*** ybabenko has quit IRC | 13:24 | |
odyssey4me | what does known_hosts get populated with? is it known targets when sshing from this host? | 13:24 |
*** ybabenko has joined #openstack-ansible | 13:24 | |
mhayden | i'm about to push a patch to cat it out | 13:24 |
mhayden | i'm curious to know what's in it | 13:24 |
odyssey4me | go for it | 13:24 |
odyssey4me | to save yourself time you can make it exit somewhere in run_playbooks... then you don't have to wait for too long | 13:25 |
*** woodard has joined #openstack-ansible | 13:26 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 13:28 |
*** ybabenko has quit IRC | 13:34 | |
*** mgoddard has joined #openstack-ansible | 13:35 | |
*** mss has joined #openstack-ansible | 13:35 | |
*** mgoddard__ has quit IRC | 13:36 | |
*** subscope has joined #openstack-ansible | 13:38 | |
*** ybabenko has joined #openstack-ansible | 13:39 | |
*** ybabenko has quit IRC | 13:41 | |
*** mss has quit IRC | 13:41 | |
*** luckyinva has joined #openstack-ansible | 13:42 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 13:42 |
*** cristicalin has joined #openstack-ansible | 13:44 | |
*** ybabenko has joined #openstack-ansible | 13:45 | |
*** ybabenko has quit IRC | 13:47 | |
*** ybabenko has joined #openstack-ansible | 13:47 | |
mhayden | ah, so known_hosts doesn't even exist | 13:47 |
mhayden | but the private/public keys are being set correctly | 13:48 |
*** woodard has quit IRC | 13:49 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 13:50 |
mattt | mhayden: throw your key in, ssh in and debug | 13:52 |
mattt | it'll save you days of troubleshooting | 13:52 |
*** alejandrito has joined #openstack-ansible | 13:52 | |
mhayden | yeah, i removed it before when things began working | 13:53 |
mhayden | but it looks like it needs to come back :| | 13:53 |
*** subscope has quit IRC | 13:55 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 13:56 |
mhayden | key in :| | 13:56 |
*** gouthamr has joined #openstack-ansible | 13:56 | |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - LXC Container Create https://review.openstack.org/241209 | 14:02 |
*** sdake has quit IRC | 14:06 | |
*** sdake has joined #openstack-ansible | 14:06 | |
cloudnull | so now we need to get some reviewers on those IRR specs, and a few repos created to move some code into :) | 14:06 |
cloudnull | odyssey4me: what i have done so far gets the setup-hosts.yml play complete as independent roles | 14:07 |
cloudnull | IE : https://github.com/os-cloud/os-ansible-deployment/blob/master/playbooks/setup-hosts.yml | 14:07 |
cloudnull | which will also serve as the basis for most of the work that will need to tackle https://github.com/os-cloud/os-ansible-deployment/blob/master/playbooks/setup-infrastructure.yml | 14:08 |
*** sdake has quit IRC | 14:08 | |
odyssey4me | cloudnull yep, once the specs are reviewed and merged I'll sort out the requests to infra for the repo creations and their tests. | 14:08 |
cloudnull | yup yup . | 14:08 |
cloudnull | for now, im done with that (awaiting reviews) . | 14:09 |
*** markvoelker has joined #openstack-ansible | 14:10 | |
cloudnull | as for testing we should be able to use the exact same job for all of the roles now that its all integrated into tox | 14:10 |
*** subscope has joined #openstack-ansible | 14:12 | |
cloudnull | odyssey4me: RE: https://bugs.launchpad.net/openstack-ansible/+bug/1488315 | 14:15 |
openstack | Launchpad bug 1488315 in openstack-ansible trunk "The python-requests package is pulled in by apt via dependency" [High,Confirmed] - Assigned to Jesse Pretorius (jesse-pretorius) | 14:15 |
cloudnull | do we BP the venv bits to kilo ? | 14:15 |
*** markvoelker has quit IRC | 14:15 | |
cloudnull | nova in a venv would ensure that issue is not possible anylonger . | 14:17 |
*** tlian has joined #openstack-ansible | 14:17 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:18 |
mhayden | bootstrap-ansible.sh had some ssh-key related stuff in it | 14:20 |
mhayden | i didn't realize that | 14:20 |
*** jmckind has joined #openstack-ansible | 14:21 | |
odyssey4me | cloudnull nit https://review.openstack.org/241209 | 14:24 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-specs: IRR - LXC Container Create https://review.openstack.org/241209 | 14:25 |
odyssey4me | cloudnull that issue has not yet appeared in kilo afaik - only juno - it may appear later in kilo, but backporting venvs to kilo is up to the downstream consumers - as that's a pretty radical feature addition to a stable branch | 14:25 |
cloudnull | it is a change. | 14:26 |
odyssey4me | personally I'd rather see those consumers move as quickly towards Liberty as possible to get that improvment | 14:26 |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:26 |
cloudnull | I guess it'd be nice to have RAX and other deployers weigh in on it. | 14:27 |
cloudnull | I think we should think about setting up stable branch maintainers. | 14:27 |
odyssey4me | yup, we can discuss it in the meeting on thu | 14:27 |
cloudnull | so vested parties can make sure that the released branch is meeting their deployment expectations. | 14:28 |
odyssey4me | exactly | 14:28 |
odyssey4me | those very same parties are also able to cherry-pick the appropriate patches into their kilo branches if they want to - there's no reason why the project needs to backport it unless enough of those using the branch want it | 14:29 |
*** mgoddard_ has joined #openstack-ansible | 14:31 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:33 |
*** mgoddard has quit IRC | 14:34 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:36 |
*** sigmavirus24_awa is now known as sigmavirus24 | 14:36 | |
*** ybabenko has quit IRC | 14:41 | |
*** ybabenko has joined #openstack-ansible | 14:41 | |
*** spotz_zzz is now known as spotz | 14:44 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:44 |
*** ybabenko has quit IRC | 14:45 | |
mattt | odyssey4me: PortBindingFailed: Binding failed for port 0a2de302-c215-414c-a599-caf348cf686c, please check neutron logs for more information. | 14:47 |
*** ybabenko_ has joined #openstack-ansible | 14:49 | |
mhayden | interesting error -> https://gist.github.com/major/b30a9aa49f01e96625a1 | 14:50 |
mhayden | perhaps it's an issue with UTF-8 chars? | 14:50 |
mhayden | looks like it might be some letters from a Polish word | 14:51 |
*** jcannava has quit IRC | 14:52 | |
mhayden | https://github.com/ansible/ansible/issues/11348 | 14:52 |
*** jcannava has joined #openstack-ansible | 14:52 | |
*** ybabenko_ has quit IRC | 14:53 | |
*** ybabenko has joined #openstack-ansible | 14:53 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 14:54 |
mhayden | opened a bug on that unarchive problem https://bugs.launchpad.net/openstack-ansible/+bug/1512736 | 14:57 |
openstack | Launchpad bug 1512736 in openstack-ansible "Unarchiving LXC contained throws an error due to UTF-8 characters" [Undecided,New] | 14:57 |
*** phalmos has joined #openstack-ansible | 14:59 | |
*** phalmos has quit IRC | 14:59 | |
cloudnull | ohh , thats exciting :) | 15:00 |
*** jmckind has quit IRC | 15:04 | |
mhayden | we could switch to tar via the command module in the meantime if nobody objects | 15:04 |
mhayden | for some reason, the unarchive works on the second try | 15:05 |
mhayden | but it causes the lxc-hosts playbook to run two times | 15:05 |
cloudnull | mhayden: yes we can | 15:05 |
mhayden | i'll go propose a patch | 15:06 |
*** jmckind has joined #openstack-ansible | 15:06 | |
cloudnull | make sure in the tag for the command your changing to you add "skip_ansible_lint" so that the ansible lint doesn't hate you | 15:06 |
cloudnull | sigmavirus24: is this still an issue: https://bugs.launchpad.net/openstack-ansible/+bug/1506285 ? | 15:08 |
openstack | Launchpad bug 1506285 in openstack-ansible "11.2.1 : openstack client with V3 auth causes usability issues" [Undecided,New] - Assigned to Ian Cordasco (icordasc) | 15:08 |
sigmavirus24 | cloudnull: I was never given an environment to test it in | 15:09 |
cloudnull | ok, can we mark it incomplete / invalid ? | 15:09 |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Workaround for UTF-8 unarchive bug https://review.openstack.org/241235 | 15:10 |
sigmavirus24 | cloudnull: you can mark it incomplete | 15:15 |
cloudnull | done | 15:15 |
cloudnull | tyvm sir | 15:15 |
odyssey4me | cloudnull mattt's discovered an issue and we need some clarity | 15:16 |
cloudnull | yes? | 15:16 |
odyssey4me | l2population in neutron needs to be disabled when l3ha is enabled, right? | 15:17 |
cloudnull | yes | 15:17 |
mattt | cloudnull: so l2_pop is enabled in AIO, my heat multi node had it disabled and all builds were failing | 15:17 |
mattt | cloudnull: the moment i flipped it to enabled in my multi node builds started working | 15:17 |
cloudnull | oh, i wonder if they fixed it ? | 15:18 |
cloudnull | mattt did you disable l3ha ? | 15:18 |
mattt | cloudnull: so i guess i need to see why the gate AIO has it set to True, while my heat multi node got False | 15:18 |
mattt | cloudnull: i didn't explicitly do that, no | 15:18 |
mattt | is it disabled in gate ? | 15:18 |
tiagogomes | l2_pop needs to be disabled when the tenant networks are vxlan no? | 15:18 |
tiagogomes | (for l3ha) | 15:18 |
mattt | tiagogomes: that is possible, as i'm using vxlan | 15:19 |
mattt | tiagogomes: but the moment i flipped it to enabled my builds worked :( | 15:19 |
odyssey4me | yeah, and according to https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/os_neutron/tasks/neutron_post_install.yml#L16-L25 I would guess that the AIO should have it disabled by default too | 15:19 |
cloudnull | if you set neutron_l2_population if overrides everything | 15:20 |
cloudnull | else if all of those conditions are met then its set accordingly. | 15:20 |
mattt | cloudnull: how do i disable l3ha? want to see if this is being done in gate | 15:20 |
*** jaypipes has quit IRC | 15:21 | |
mattt | cloudnull: n/m i see it | 15:22 |
mattt | having 2 or more neutron agents automatically disables l2_pop | 15:22 |
mattt | which is the case on my multi node | 15:22 |
*** skamithi has joined #openstack-ansible | 15:23 | |
cloudnull | ah , so thats the choice, l3ha or l2pop and it seems that your nodes need l2pop to function. | 15:23 |
cloudnull | so set neutron_l2_population in the heat template and you should be good to go . | 15:24 |
mattt | cloudnull: ok cool | 15:24 |
cloudnull | doing so will overide everything forcing l3ha off | 15:24 |
cloudnull | the template logic wont allow l3ha by default if neutron_l2_population is enabled. | 15:24 |
mattt | cloudnull: ok cool, i'll make that adjustment then | 15:26 |
*** CheKoLyN has joined #openstack-ansible | 15:26 | |
*** markvoelker has joined #openstack-ansible | 15:26 | |
odyssey4me | so I guess the question is whether this is an edge case, or a common case | 15:26 |
cloudnull | idk for sure. | 15:26 |
*** jimchou has joined #openstack-ansible | 15:29 | |
*** markvoelker has quit IRC | 15:31 | |
*** subscope has quit IRC | 15:33 | |
*** mgoddard_ has quit IRC | 15:33 | |
*** mgoddard has joined #openstack-ansible | 15:33 | |
*** mss has joined #openstack-ansible | 15:37 | |
*** subscope has joined #openstack-ansible | 15:41 | |
*** mss has quit IRC | 15:42 | |
*** tiagogomes has quit IRC | 15:43 | |
*** tiagogomes_ has joined #openstack-ansible | 15:43 | |
*** tiagogomes_ has quit IRC | 15:45 | |
*** tiagogomes_ has joined #openstack-ansible | 15:45 | |
*** daneyon has joined #openstack-ansible | 15:46 | |
*** Bjoern_ has joined #openstack-ansible | 15:47 | |
*** mss has joined #openstack-ansible | 15:48 | |
openstackgerrit | Matt Thompson proposed openstack/openstack-ansible: Reference correct group in haproxy_config.yml https://review.openstack.org/241173 | 15:52 |
*** tnarg has joined #openstack-ansible | 15:55 | |
*** tnarg has quit IRC | 15:55 | |
*** tnarg has joined #openstack-ansible | 15:56 | |
*** ybabenko has quit IRC | 15:58 | |
odyssey4me | bug triage in this room shortly cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, mancdaz, dolphm, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden, scarlisle, luckyinva | 15:59 |
*** mss has quit IRC | 15:59 | |
cloudnull | o/ | 15:59 |
*** phalmos has joined #openstack-ansible | 15:59 | |
* mhayden woots | 15:59 | |
prometheanfire | hi | 15:59 |
*** daneyon has quit IRC | 15:59 | |
palendae | o/ | 15:59 |
d34dh0r53 | o/ | 16:00 |
mattt | o/ | 16:00 |
*** ybabenko has joined #openstack-ansible | 16:01 | |
*** mss has joined #openstack-ansible | 16:01 | |
odyssey4me | first up https://bugs.launchpad.net/openstack-ansible/+bug/1508411 | 16:02 |
openstack | Launchpad bug 1508411 in openstack-ansible "defaults-file not being handled properly." [Undecided,New] | 16:02 |
*** jimchou_ has joined #openstack-ansible | 16:02 | |
odyssey4me | ah yes - cloudnull this was one of a series where you added 'related-bug' instead of closes-bug... is there a reason to keep these open? | 16:03 |
cloudnull | we pinned it | 16:03 |
cloudnull | we to a lower version | 16:03 |
cloudnull | so we can close it | 16:03 |
*** phalmos has quit IRC | 16:04 | |
*** jimchou has quit IRC | 16:04 | |
cloudnull | done. | 16:04 |
odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1508448 is the same? | 16:04 |
openstack | Launchpad bug 1508448 in Percona XtraBackup 2.3 "--default-file requires datadir to be set." [Medium,Triaged] | 16:04 |
Sam-I-Am | moo. | 16:04 |
odyssey4me | and https://bugs.launchpad.net/openstack-ansible/+bug/1508450 ? | 16:04 |
openstack | Launchpad bug 1508450 in Percona XtraBackup 2.3 "secure_auth forced client side in 2.3.2" [Wishlist,Triaged] | 16:04 |
cloudnull | same thing | 16:05 |
odyssey4me | ok, so you'll take care of those cloudnull ? | 16:05 |
cloudnull | done | 16:05 |
odyssey4me | cool, next: https://bugs.launchpad.net/openstack-ansible/+bug/1508517 | 16:05 |
openstack | Launchpad bug 1508517 in openstack-ansible "nova fails to deploy on hosts without container_networks set" [Undecided,New] | 16:05 |
*** ybabenko has quit IRC | 16:06 | |
*** mss has quit IRC | 16:06 | |
cloudnull | so i saw this when doing the port work earlier this morning | 16:06 |
*** elo has joined #openstack-ansible | 16:07 | |
cloudnull | its an issue where the iterator attempts to load the items in the object but its undefined. | 16:07 |
cloudnull | the fix is https://github.com/os-cloud/openstack-ansible-lxc_container_create/blob/master/tasks/container_create.yml#L110 | 16:07 |
odyssey4me | ok, any volunteers for porting that in asap? | 16:08 |
*** cristicalin has quit IRC | 16:08 | |
*** ysm has joined #openstack-ansible | 16:08 | |
cloudnull | yes ill go bang it out | 16:08 |
odyssey4me | ok, great | 16:09 |
odyssey4me | next: https://bugs.launchpad.net/openstack-ansible/+bug/1510244 | 16:09 |
openstack | Launchpad bug 1510244 in openstack-ansible "Issues when using fernet tokens in Keystone" [Undecided,New] | 16:09 |
*** jimchou_ has quit IRC | 16:09 | |
*** jmckind has quit IRC | 16:10 | |
odyssey4me | logan2 ping? | 16:10 |
logan2 | i will try to get that addl. info in there this week. haven't had a chance to look into it further yet | 16:10 |
odyssey4me | logan2 it sounds like perhaps the cron job to sync the keys isn't working between the keystone containers | 16:11 |
odyssey4me | assuming this is the kilo branch or higher this is important to get more info on | 16:11 |
logan2 | the keys contents of /etc/keystone/fernet-keys were identical in the keystone containers if that is what the cronjob is syncing | 16:12 |
odyssey4me | hmm, yes - that's odd then | 16:12 |
odyssey4me | anyone else seen anything like this? | 16:12 |
*** galstrom_zzz is now known as galstrom | 16:12 | |
odyssey4me | we have multiple containers in the gate check, so it would seem like this may be an edge case of some sort | 16:12 |
lbragstad | logan2 interestings, did the key repository from oscontroller5 have the same permissions as the ones on oscontroller4? | 16:13 |
lbragstad | interesting* | 16:13 |
logan2 | not sure. can you comment that on the bug so I don't forget to check when I do a deploy with fernet again | 16:13 |
lbragstad | logan2 sure | 16:13 |
*** woodard has joined #openstack-ansible | 16:14 | |
*** nosleep77 has left #openstack-ansible | 16:14 | |
*** erikmwilson has joined #openstack-ansible | 16:14 | |
odyssey4me | ok, so for now I'll leave this as a new bug to give us a chance to chat about it again next week or on thu | 16:15 |
odyssey4me | everyone happy with that? | 16:15 |
cloudnull | +1 | 16:15 |
logan2 | yep thanks | 16:15 |
mattt | +1 | 16:15 |
odyssey4me | alright, next: https://bugs.launchpad.net/openstack-ansible/+bug/1510596 | 16:15 |
openstack | Launchpad bug 1510596 in openstack-ansible "Don't send galera root credentials file to compute hosts" [Undecided,New] | 16:15 |
cloudnull | yea thats a thing | 16:16 |
cloudnull | if someone has cycles to put into that it'd be appreciated. | 16:16 |
odyssey4me | yeah, as noted in the bug we have a spec for that which missed the Liberty release, so it'd be great if someone could put some time into getting it done | 16:19 |
odyssey4me | mhayden perhaps you'd like to get into it as part of the security work you're doing? | 16:19 |
mhayden | i'd gladly take that one on | 16:19 |
mhayden | i need a break from AIO work :P | 16:19 |
mattt | this isn't specific to compute hosts tho is it ? | 16:20 |
mattt | it's like every host/container has a .my.cnf in it | 16:20 |
odyssey4me | yeah, the spec was to remove it wherever it wasn't needed | 16:21 |
mattt | ok cool | 16:21 |
odyssey4me | basically it should only be in the utility container - I think | 16:21 |
odyssey4me | perhaps in the galera containers too | 16:21 |
mattt | i know we used to drop it on the hosts for monitoring purposes | 16:21 |
odyssey4me | http://specs.openstack.org/openstack/openstack-ansible-specs/specs/mitaka/limit-mysql-config-distribution.html | 16:21 |
mattt | but that monitoring requirement should be handled externally now, so it should be fine to remove | 16:21 |
odyssey4me | happy for this work to be done in master and backported to liberty, assuming it makes it in before 12.1.0 | 16:22 |
*** ybabenko has joined #openstack-ansible | 16:22 | |
odyssey4me | alright, next: https://bugs.launchpad.net/openstack-ansible/+bug/1510990 | 16:22 |
openstack | Launchpad bug 1510990 in openstack-ansible ""Add in cinder devices types" task failing" [Undecided,New] | 16:22 |
odyssey4me | tiagogomes_ ping? | 16:23 |
logan2 | ahh didn't see the bug for that but i am experiencing it too. extending the health check should fix it. i have about 6-8 backends and it times out for the first 3-4 usually then starts working after that | 16:23 |
logan2 | or not times out, but returns 503 | 16:24 |
*** woodard has quit IRC | 16:24 | |
*** ybabenko has quit IRC | 16:27 | |
odyssey4me | so it seems that we need a better service up check, and perhaps a few more retries? | 16:27 |
logan2 | http://paste.gentoolinux.info/wayacutavu.py from last week when we were discussing it in here | 16:27 |
odyssey4me | any volunteers to pick that one up? | 16:29 |
*** ashishb has joined #openstack-ansible | 16:29 | |
odyssey4me | it would seem that this is an issue from kilo onwards? | 16:29 |
*** woodard has joined #openstack-ansible | 16:29 | |
openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible: Implement Neutron LBAAS using haproxy https://review.openstack.org/241282 | 16:30 |
logan2 | <- kilo only atm so i can't speak on liberty | 16:30 |
*** ybabenko has joined #openstack-ansible | 16:31 | |
odyssey4me | ok cool - as this is specifically in the ansible tasks I expect that this will carry forward | 16:31 |
odyssey4me | no volunteers? | 16:31 |
mattt | odyssey4me: i can have a peek, but i have other things i need to look at over the next few days | 16:32 |
mattt | so if someone can smash it sooner then by all means | 16:32 |
odyssey4me | ok, I'll leave it unassigned for now | 16:32 |
odyssey4me | anyone can pick it up when they're ready | 16:32 |
odyssey4me | next: https://bugs.launchpad.net/openstack-ansible/+bug/1512533 | 16:32 |
openstack | Launchpad bug 1512533 in openstack-ansible "Only neutron_server group should contain ml2_conf.ini" [Undecided,New] | 16:32 |
mattt | yep that sounds reasonable | 16:32 |
odyssey4me | Sam-I-Am ping? | 16:32 |
Sam-I-Am | yo? | 16:33 |
Sam-I-Am | looks like cloudnull covered that bug with the comment in https://bugs.launchpad.net/openstack-ansible/+bug/1512531 | 16:33 |
openstack | Launchpad bug 1512531 in openstack-ansible "Remove ml2_conf.ini from Linux bridge agent init script" [Undecided,Incomplete] | 16:33 |
odyssey4me | I might be wrong, but when I tested this implementation it seemed to be that the agents also needed the ml2_conf.ini ? | 16:33 |
Sam-I-Am | but in liberty i have not seen the issue cloudnull mentions | 16:33 |
Sam-I-Am | and the liberty install guide uses linuxbridge | 16:33 |
Sam-I-Am | according to the neutron folks, the ml2 stuff needs to only reside on the server | 16:34 |
Sam-I-Am | supposedly its been this way for a while | 16:34 |
cloudnull | Sam-I-Am: yea what odyssey4me said, without it being sourced it was ignoring options causing issues with metadata | 16:35 |
odyssey4me | ok, so I've marked the two as duplicates | 16:35 |
Sam-I-Am | they're not really dupes, they're just related | 16:35 |
Sam-I-Am | cloudnull: which is odd, because the ml2 config has nothing about metadata | 16:36 |
prometheanfire | Sam-I-Am: so the client linuxbridge configs aren't needed for the lb-agent to run on the compute hosts? | 16:36 |
*** jaypipes has joined #openstack-ansible | 16:36 | |
Sam-I-Am | prometheanfire: no, the ml2 config file isnt needed for the lb agent. just neutron.conf and lb_agent.ini | 16:36 |
prometheanfire | ah, neat | 16:36 |
Sam-I-Am | the ml2 plugin only resides on the server | 16:36 |
prometheanfire | the linuxbridge agent config is in the ml2 dir now btw | 16:37 |
odyssey4me | is someone keen to verify this and submit a patch for review? | 16:37 |
Sam-I-Am | it just so happens that people were sticking agent config into ml2_conf.ini because there was no better place | 16:37 |
prometheanfire | that's where upstream sticks it now instead of it's own dir like kilo had | 16:37 |
logan2 | odyssey4me: https://bugs.launchpad.net/openstack-ansible/+bug/1510244 I think accidentally got marked as fixed :) | 16:38 |
openstack | Launchpad bug 1510244 in openstack-ansible "Issues when using fernet tokens in Keystone" [Undecided,Fix released] | 16:38 |
Sam-I-Am | odyssey4me: i'll try to put something together | 16:38 |
odyssey4me | thanks Sam-I-Am any optimisation of this sort is welcome :) | 16:39 |
cloudnull | Sam-I-Am: when odyssey4me was testing, if https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/os_neutron/templates/plugins/ml2/ml2_conf.ini.j2#L8 was there and the file was not being directly sourced it caused gate failure which looked like metadata issues. | 16:39 |
cloudnull | maybe we have config that shouldn't be there still. | 16:39 |
cloudnull | idk ? | 16:39 |
Sam-I-Am | hmm, you'd think that only applies to ml2 | 16:40 |
odyssey4me | yeah, the Neutron config update for Liberty was a bit tricky for me... considering I know nothing about Networking or Neutron :p | 16:40 |
odyssey4me | so hey, I'm happy to be proved wrong :p | 16:40 |
odyssey4me | logan2 fixed - sorry about that | 16:40 |
logan2 | no worries, thanks | 16:41 |
*** wtie has quit IRC | 16:41 | |
Sam-I-Am | odyssey4me: i'll poke around and also poke neutron folks again | 16:41 |
odyssey4me | thanks Sam-I-Am | 16:41 |
Sam-I-Am | the install guide uses the suggested config and seems to work, so idk :) | 16:41 |
Sam-I-Am | its neutron! | 16:42 |
odyssey4me | that's it for new bugs - does anyone have other bugs that they want to discuss? | 16:42 |
prometheanfire | does openstack itself count? | 16:42 |
odyssey4me | haha | 16:42 |
odyssey4me | openstack is not a bug, but it may perhaps be considered a horde of bugs :p | 16:43 |
odyssey4me | a swarm? | 16:43 |
Sam-I-Am | features | 16:43 |
prometheanfire | pestilence of bugs? | 16:43 |
prometheanfire | infestation? | 16:43 |
odyssey4me | ok, thank you all for participating :) | 16:44 |
cloudnull | o/ tyvm odyssey4me | 16:45 |
*** markvoelker has joined #openstack-ansible | 16:45 | |
Sam-I-Am | cloudnull: its been a while since i've used teardown.sh on an aio. seems like after reboot i need to run bootstrap-aio again? | 16:47 |
*** mgoddard_ has joined #openstack-ansible | 16:47 | |
*** ybabenko has quit IRC | 16:48 | |
palendae | Sam-I-Am: Yeah. I'm not sure if rackertom's fix for cgmanager holding on to volume groups was integrated yet | 16:49 |
palendae | But after teardown.sh/a reboot you need to re-run bootstrap-aio | 16:49 |
openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Deployment fix when container_networks are not set https://review.openstack.org/241293 | 16:49 |
*** mgoddard has quit IRC | 16:50 | |
Sam-I-Am | for a while there it was deleting /opt/o-a | 16:50 |
palendae | Believe that's been fixed | 16:50 |
palendae | At least in amster | 16:50 |
palendae | It was deleting /opt/openstack* | 16:51 |
palendae | Because that's what the old build dirs were | 16:51 |
openstackgerrit | Merged openstack/openstack-ansible: Update juno for new dev work - 2 Nov 2015 https://review.openstack.org/240911 | 16:51 |
cloudnull | logan2: odyssey4me: https://review.openstack.org/#/c/241293/1 fix for the nova management network issues. | 16:52 |
Sam-I-Am | cloudnull: has there been any work on ovs in o-a? | 16:53 |
cloudnull | no, we need it OSA to work | 16:53 |
mhayden | WRT https://bugs.launchpad.net/openstack-ansible/+bug/1512736 -- it looks like my patch doesn't break anything but does fix the issue temporarily | 16:53 |
openstack | Launchpad bug 1512736 in openstack-ansible "Unarchiving LXC contained throws an error due to UTF-8 characters" [Undecided,In progress] - Assigned to Major Hayden (rackerhacker) | 16:53 |
odyssey4me | logan2 if you can test that fix it'd be great to see your feedback in the review! | 16:53 |
palendae | cloudnull: Rephrase, cannot parse | 16:54 |
Sam-I-Am | cloudnull: lulz | 16:54 |
palendae | Is it further ovs snark? | 16:54 |
Sam-I-Am | cloudnull: throwing around the idea as using o-a as a dev platform for ovn rather than devstack | 16:54 |
Sam-I-Am | because devstack sucks | 16:54 |
odyssey4me | Sam-I-Am in revising the neutron configs to split out the linuxbridge, I noticed that it'd probably be pretty trivial to do - but we still need someone to care about it :) | 16:54 |
palendae | Sam-I-Am: Oh, miguelgrinberg did a talk about that :p | 16:55 |
Sam-I-Am | palendae: ovn? | 16:55 |
palendae | o-a as a dev platform | 16:55 |
* cloudnull rephrasing | 16:55 | |
palendae | I didn't go to that one, but I hear it was really well attended | 16:55 |
Sam-I-Am | odyssey4me: ehhh, yeah sort of | 16:56 |
cloudnull | no, OvS is a pit of failures and should be abandoned as a technology. but if someone wants to build support for it PRs are more than welcome :) | 16:56 |
skamithi | +1 | 16:57 |
*** jimchou has joined #openstack-ansible | 16:57 | |
skamithi | ovs sucks | 16:57 |
cloudnull | ^ my thoughts exactly | 16:57 |
tiagogomes_ | Open vSwitch? | 16:57 |
logan2 | i will give it a try cloudnull, odyssey4me. thanks | 16:58 |
*** bangfoo has joined #openstack-ansible | 16:58 | |
Sam-I-Am | i'm leading up to ovn | 16:58 |
Sam-I-Am | which actually makes better use of ovs | 16:58 |
cloudnull | OvN is the lesser evil of OvS. | 16:58 |
cloudnull | thats like saying you have Ford pinto and a AC gremlin but the gremlin has new tires... | 16:59 |
cloudnull | OvN is the gremlin w/ new tires | 16:59 |
cloudnull | =P | 16:59 |
odyssey4me | lol, well - I do think that some pretty interesting neutron agents have been introduced in the last cycle - so if we have parties interested in making them work and maintaining them then we can definitely make that go :) | 16:59 |
Sam-I-Am | my goal is to write usable docs for ovn prior to release, so it doesnt get up in a storm of confusion and fud, or the only docs that exist are "heres how you use in devstack" | 16:59 |
Sam-I-Am | <- trying to fix neutron | 16:59 |
Sam-I-Am | it would be nice if o-a could support its bits | 17:00 |
odyssey4me | Sam-I-Am try and rustle up some more interested parties :) | 17:00 |
cloudnull | ^ ++ | 17:00 |
* cloudnull looks at Sam-I-Am | 17:00 | |
odyssey4me | it does seem like there were some interested neutron devs after miguelgrinberg's talk | 17:00 |
odyssey4me | I had a chat with one who was keen to make it go | 17:00 |
*** mgoddard_ has quit IRC | 17:00 | |
cloudnull | Sam-I-Am: I'm sure there will be a blog post out there somewhere on how to use OvN ;) | 17:02 |
Sam-I-Am | well, there is a blog post | 17:03 |
cloudnull | shipit ! | 17:04 |
prometheanfire | adding services? | 17:04 |
Sam-I-Am | and some stuff in neutron devref.... and the ovs docs | 17:04 |
*** ybabenko has joined #openstack-ansible | 17:05 | |
Sam-I-Am | the unique feature of o-a vs all other deployment tools is ability to deploy from source | 17:07 |
Sam-I-Am | most of the other ones use packages of some sort | 17:07 |
*** ysm has quit IRC | 17:08 | |
*** bangfoo has left #openstack-ansible | 17:09 | |
*** subscope has quit IRC | 17:10 | |
*** ybabenko has quit IRC | 17:10 | |
*** gouthamr has quit IRC | 17:13 | |
cloudnull | Sam-I-Am: ok i've now watched the ovn talk from vancouver / read docs from the openstack wiki and i now believe ovn is a solution looking for a problem. it still a controller based solution, it still relies on ovs, and they're still not integrated with openflow... | 17:13 |
*** javeriak has joined #openstack-ansible | 17:13 | |
cloudnull | Sam-I-Am: so why is it better than before, in your oppinion ? | 17:14 |
*** ysm has joined #openstack-ansible | 17:14 | |
Sam-I-Am | cloudnull: ultimately it moves routing and acls/secgroups into ovs | 17:14 |
Sam-I-Am | it'll dvr, but with less suck (hopefully) | 17:15 |
Sam-I-Am | no more dhcp or l3 agents | 17:15 |
cloudnull | and thats a good thing ? -- asking for a friend | 17:15 |
cloudnull | :D | 17:15 |
cloudnull | no more dhcp or l3 agents , replaced with new shinny ovn agent (when they simply call a deamon now) . | 17:16 |
Sam-I-Am | heh | 17:16 |
cloudnull | will it drive dvr without all of the nonsense ? | 17:17 |
*** mgoddard has joined #openstack-ansible | 17:17 | |
Sam-I-Am | i sure hope so | 17:18 |
*** mss has joined #openstack-ansible | 17:18 | |
*** tiagogomes_ has quit IRC | 17:18 | |
cloudnull | thats better than definitely maybe +1 typie typie make it go | 17:18 |
Sam-I-Am | you should chat with russellb about it for more info, but in my feeble neutron mind it sounds better than plain old ovs | 17:18 |
cloudnull | but still no openflow integration ? | 17:19 |
Sam-I-Am | it uses flows | 17:19 |
cloudnull | its in the name :) | 17:19 |
cloudnull | I've not run it however I like what opendaylight is doing -- https://wiki.opendaylight.org/view/OpenStack_and_OpenDaylight -- but sadly they're an OvS based solution. | 17:21 |
palendae | Is OVS in systemd yet? | 17:21 |
cloudnull | thatll be OvSd | 17:21 |
palendae | switchd | 17:22 |
cloudnull | +1 i like it | 17:22 |
Bjoern_ | yeah I looking at opendaylight too, any new SDN uses OVS. Don't think we can sustain just offering linux bridge for quite long | 17:26 |
*** Bjoern_ is now known as BjoernT | 17:26 | |
*** karimb has quit IRC | 17:26 | |
odyssey4me | well, SDN is also offered by the non OSS providers - and they don't need OVS | 17:27 |
BjoernT | not many, I don't know any other than plumgrid | 17:28 |
BjoernT | Project callico or opendaylight or floodlight all need OVS | 17:29 |
palendae | There were a lot of providers at summit | 17:29 |
odyssey4me | Cisco's Nexus platform supports the use of vxlan | 17:29 |
odyssey4me | and SDN all the way through | 17:29 |
palendae | Hot area | 17:29 |
BjoernT | the "others" might be pure commercial setups, open source tends to use OVS | 17:29 |
BjoernT | yes "commercial" | 17:29 |
odyssey4me | sure - but if you want to do SDN properly, you probably want to use a commercial platform | 17:30 |
BjoernT | just getting rid of iptables is worth looking at OVS | 17:30 |
Sam-I-Am | well, ovn :) | 17:30 |
palendae | Not sure why commercial's necessarily a dirty word | 17:30 |
BjoernT | that's your opinion, OVS is not so bad | 17:30 |
odyssey4me | hah, so then the question would be how many linux engineers understand SDN flows - how to work with them, troubleshoot them, etc | 17:30 |
*** phalmos has joined #openstack-ansible | 17:30 | |
odyssey4me | the issue is not one of like or dislike - it's more a matter of supportability | 17:31 |
BjoernT | so? And the alternative would be closed source where you don't even see flows ? | 17:31 |
odyssey4me | there are a lot more Cisco-familiar engineers around than OVS capable engineers | 17:31 |
BjoernT | I don't know it is worth looking at OVS | 17:31 |
BjoernT | people tend to argue with a OVS 0.9 or older in mind that's not where they are today | 17:32 |
odyssey4me | but hey, that's a deployer choice - as far as I've seen most DC providers have a sunk investment into some sort of commercial networking technology | 17:32 |
odyssey4me | whatever that is is probably what they should add support for | 17:32 |
palendae | I admittedly know very little about OVS. My impression is that it's going to be like systemd - new, and yes, people are not as familiar, but it's coming | 17:32 |
Sam-I-Am | most of the problem with ovs was general lack of docs aimed at admin/operator types. so it was a black box. | 17:33 |
*** ybabenko has joined #openstack-ansible | 17:33 | |
Sam-I-Am | that and, by itself, adds extra complexity to openstack networking (still needs a linuxbridge for secgrous) | 17:33 |
BjoernT | Correct. But's that's true for any open source even openstack | 17:33 |
odyssey4me | that, and scale, and kernel panics on every kernel upgrade unless it was lock stepped with the right OVS kernel module | 17:33 |
odyssey4me | general pain-in-the-buttness | 17:34 |
*** th720t has joined #openstack-ansible | 17:34 | |
palendae | odyssey4me: Welome to computers :( | 17:34 |
BjoernT | Well OVS in now part of the kernel not bolted on | 17:34 |
odyssey4me | well, the fact of the matter is that from the project standpoint - until someone cares enough to add the code and maintain it, it will not be supported | 17:35 |
*** jmckind has joined #openstack-ansible | 17:35 | |
BjoernT | IMHO, as a open source project we can't rule out any possible solution. OVS is coming and it came to stay there. More or less bugs it not even an argument. Without bugs we would not have our jobs we have today. And because it's hard, it's not an argument to look at it | 17:35 |
*** jmckind has quit IRC | 17:36 | |
BjoernT | Yeah I know someone has to make the first move to integrate ovs as a alternative solution to linuxbridge | 17:36 |
*** slotti has quit IRC | 17:36 | |
*** jmckind has joined #openstack-ansible | 17:36 | |
odyssey4me | we don't rule anything out - all we do is ask that there be enough people to maintain the code going forward | 17:37 |
*** ybabenko has quit IRC | 17:37 | |
*** jmckind has quit IRC | 17:37 | |
BjoernT | That's fine and a legit concern | 17:38 |
*** jmckind has joined #openstack-ansible | 17:38 | |
*** jmckind has quit IRC | 17:39 | |
logan2 | it seems user_*.yml in /etc/openstack_deploy cannot be used to override playbooks/defaults/repo_packages/openstack_services.yml *_git_repo and *_git_install_branch settings. is there a good way to do that without having to patch openstack_services.yml directly? | 17:39 |
*** jmckind has joined #openstack-ansible | 17:39 | |
*** greg_a has joined #openstack-ansible | 17:40 | |
*** jmckind has quit IRC | 17:40 | |
*** jmckind has joined #openstack-ansible | 17:41 | |
th720t | Hello. When I execute setup-hosts.yml, the following error is produced: "RuntimeError: maximum recursion depth exceeded while calling a Python object." The file that this occurs in is "../playbooks/inventory/dynamic_inventory.py". We have tried multiple openstack_user_config.yml configs (all of which are syntactically valid), and the problem still persists. Any idea on what could be causing this? | 17:41 |
*** jmckind has quit IRC | 17:42 | |
cloudnull | logan2: you should be able to override anything in the https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml file by setting it on the user_* files | 17:46 |
cloudnull | is that now working ? or returning the same data? | 17:46 |
logan2 | i thought the same but it was completley ignoring the git repos I had set in user_variables.yml | 17:46 |
logan2 | completely* | 17:47 |
cloudnull | hum... going to go look | 17:47 |
*** javeriak_ has joined #openstack-ansible | 17:48 | |
*** javeriak has quit IRC | 17:48 | |
odyssey4me | argh, it's probably that ansible bug regarding precedence | 17:50 |
*** sdake has joined #openstack-ansible | 17:51 | |
logan2 | it seems like everything else is overridable in user_variables so I figured the git repos were. then I was looking at things like _get_files in py_pkgs.py and tried setting the overrides a file named playbooks/defualts/repo_packages/openstack_services_overrides.yml which it loaded but screwed things up even more. seemed like it started duplicating repos and doing all kinds of crazy | 17:58 |
logan2 | stuff | 17:58 |
*** phalmos has quit IRC | 17:59 | |
*** ybabenko has joined #openstack-ansible | 18:00 | |
*** abitha has joined #openstack-ansible | 18:00 | |
*** gouthamr has joined #openstack-ansible | 18:00 | |
*** greg_a has quit IRC | 18:01 | |
cloudnull | logan2: digging into it , the plugin py_pkgs does ignore the overrides. and from the looks of it , its been that way forever. | 18:01 |
cloudnull | we should fix that :) | 18:01 |
cloudnull | logan2: oddly enough this was raised regarding a similar use case https://bugs.launchpad.net/openstack-ansible/+bug/1510575 | 18:02 |
openstack | Launchpad bug 1510575 in openstack-ansible "Adding additional Python wheels to the repo" [Undecided,Incomplete] | 18:02 |
*** ybabenko has quit IRC | 18:04 | |
logan2 | yeah if it would let me override using a file that _get_files picks up without breaking the build duping repos as it loads both files etc that would work fine too. i just want to override nova, horizon, cinder, and neutron without having to modify openstack_services.yml so I don't have to manually merge every time upstream changes git revisions ;) | 18:04 |
*** ashishb has quit IRC | 18:11 | |
*** alkari has joined #openstack-ansible | 18:12 | |
alkari | Anyone got OSA installed on RHEL/CentOS? | 18:12 |
odyssey4me | alkari it is possible, with some work - we'd like to actually get that done formally in the project, but no-one has cared enough to step up to actually do it | 18:13 |
odyssey4me | we're ready to help get it done, but we need enough people to care enough to maintain it too - otherwise it'll bit rot | 18:14 |
*** woodard has quit IRC | 18:15 | |
*** BjoernT has quit IRC | 18:15 | |
alkari | odyssey4me: thanks. where can I find whatever work that's been done, if any? | 18:16 |
*** daneyon has joined #openstack-ansible | 18:17 | |
odyssey4me | alkari most who have done it have kept it in private repositories | 18:17 |
odyssey4me | so if you're keen to do it then I would ask that you add it to the Thu meeting agenda | 18:17 |
odyssey4me | we can discuss it, add the spec and get it done | 18:18 |
odyssey4me | I would encourage you, meanwhile, to familiarise yourself with what is in the project now as that will be important for the work | 18:18 |
alkari | Where do I find Thu meeting info please? | 18:19 |
odyssey4me | alkari in the channel topic title ;) | 18:19 |
odyssey4me | https://wiki.openstack.org/wiki/Meetings/openstack-ansible#Agenda_for_next_meeting | 18:19 |
alkari | awesome. thank you odyssey4me: cheers | 18:21 |
*** sdake has quit IRC | 18:23 | |
*** ashishb has joined #openstack-ansible | 18:23 | |
*** daneyon has quit IRC | 18:24 | |
*** trunyon has joined #openstack-ansible | 18:24 | |
*** phiche has quit IRC | 18:24 | |
*** ybabenko has joined #openstack-ansible | 18:27 | |
*** ybabenko has quit IRC | 18:31 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 18:33 | |
*** abitha has quit IRC | 18:35 | |
*** phalmos has joined #openstack-ansible | 18:36 | |
*** elo has quit IRC | 18:37 | |
*** elo has joined #openstack-ansible | 18:37 | |
*** woodard has joined #openstack-ansible | 18:39 | |
*** woodard has quit IRC | 18:42 | |
*** woodard has joined #openstack-ansible | 18:42 | |
*** rebase has joined #openstack-ansible | 18:49 | |
*** sdake has joined #openstack-ansible | 18:49 | |
*** ybabenko has joined #openstack-ansible | 18:54 | |
*** th720t has quit IRC | 18:54 | |
*** jaybeers has joined #openstack-ansible | 18:55 | |
*** ybabenko has quit IRC | 19:00 | |
mhayden | i've made it past all of the AIO ssh shenanigans, but now i'm getting an error in tempest around "No valid host was found" | 19:05 |
mhayden | http://logs.openstack.org/25/239525/44/check/gate-openstack-ansible-dsvm-commit/6909a9d/console.html#_2015-11-03_18_04_38_013 | 19:05 |
mhayden | i know what the error means, but i was wondering if anyone can spot why it's showing up | 19:05 |
*** woodard has quit IRC | 19:06 | |
mhayden | ah, nevermind -- i think i'm not limiting certain container counts | 19:06 |
*** skamithi has quit IRC | 19:07 | |
*** alkari has quit IRC | 19:08 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 19:12 | |
*** ashishb has quit IRC | 19:16 | |
*** javeriak_ has quit IRC | 19:16 | |
*** javeriak has joined #openstack-ansible | 19:16 | |
*** phalmos has quit IRC | 19:18 | |
*** jmckind has joined #openstack-ansible | 19:20 | |
*** alejandrito has quit IRC | 19:21 | |
*** skamithi has joined #openstack-ansible | 19:25 | |
*** phalmos has joined #openstack-ansible | 19:26 | |
*** greg_a has joined #openstack-ansible | 19:26 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 19:27 |
*** jmckind has quit IRC | 19:35 | |
openstackgerrit | Merged openstack/openstack-ansible: Workaround for UTF-8 unarchive bug https://review.openstack.org/241235 | 19:36 |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Switch -xf to xf for tar command https://review.openstack.org/241347 | 19:41 |
*** jmckind has joined #openstack-ansible | 19:43 | |
*** greg_a has quit IRC | 19:48 | |
*** gouthamr has quit IRC | 19:56 | |
*** gouthamr has joined #openstack-ansible | 19:56 | |
*** mgoddard has quit IRC | 19:58 | |
*** mgoddard has joined #openstack-ansible | 19:58 | |
*** jmckind has quit IRC | 20:07 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Remove dashes from short tar params https://review.openstack.org/241347 | 20:35 |
mhayden | so i'm looking over https://bugs.launchpad.net/openstack-ansible/+bug/1510596 | 20:38 |
openstack | Launchpad bug 1510596 in openstack-ansible trunk "Don't send galera root credentials file to compute hosts" [Wishlist,Confirmed] - Assigned to Major Hayden (rackerhacker) | 20:38 |
mhayden | which nodes actually *need* that file? i could see where the galera server containers would need it | 20:39 |
mhayden | are there any others? | 20:39 |
spotz | mhayden No fair uploading new stuff just as I'm about to +1 what was already there! | 20:41 |
mhayden | oopsies | 20:44 |
mhayden | i realized there was another spot | 20:44 |
spotz | heheh | 20:44 |
mhayden | odyssey4me: OMG the aio finished a dvsm test | 20:44 |
mhayden | those ssh and container count issues were the worst | 20:44 |
*** rebase has quit IRC | 20:50 | |
*** th720 has joined #openstack-ansible | 20:50 | |
th720 | I believe there may be a bug regarding haproxy host names in openstack_user_config.yml. When defining an haproxy_hosts machine with a name of "haproxy", it appears to cause a recursion error when running setup-hosts.yml | 20:55 |
th720 | Could this be related to the recent change in env.d/haproxy? | 20:55 |
th720 | I notice that the container_skel also references "haproxy" | 20:56 |
*** phiche has joined #openstack-ansible | 21:01 | |
*** ysm has quit IRC | 21:03 | |
*** sdake has quit IRC | 21:15 | |
*** sdake has joined #openstack-ansible | 21:16 | |
*** elo1 has joined #openstack-ansible | 21:23 | |
*** elo has quit IRC | 21:24 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 21:31 |
mhayden | hopefully that's the last push on AIO before it needs a review | 21:32 |
mhayden | just removed the debug info | 21:33 |
mhayden | th720: you may have hit a bug -- i haven't tested that but i could see how it might happen | 21:33 |
*** trunyon has quit IRC | 21:34 | |
*** ysm has joined #openstack-ansible | 21:48 | |
*** gouthamr has quit IRC | 21:50 | |
*** sdake has quit IRC | 21:50 | |
*** sdake has joined #openstack-ansible | 21:50 | |
th720 | Alright, I wrote it up: https://bugs.launchpad.net/openstack-ansible/+bug/1512883 | 21:55 |
openstack | Launchpad bug 1512883 in openstack-ansible "Host name of haproxy in openstack_user_config.yml causes recursion error" [Undecided,New] | 21:55 |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Only deploy .my.cnf file on galera servers https://review.openstack.org/241382 | 22:08 |
*** daneyon has joined #openstack-ansible | 22:11 | |
*** britthouser has joined #openstack-ansible | 22:15 | |
*** tlian has quit IRC | 22:18 | |
*** phiche has quit IRC | 22:19 | |
*** daneyon_ has joined #openstack-ansible | 22:21 | |
*** skamithi has quit IRC | 22:24 | |
*** daneyon has quit IRC | 22:25 | |
*** gouthamr has joined #openstack-ansible | 22:26 | |
openstackgerrit | Major Hayden proposed openstack/openstack-ansible: AIO bootstrap in Ansible https://review.openstack.org/239525 | 22:32 |
mhayden | odyssey4me / mattt: just talked to d34dh0r53 and he suggested that sshd might not be restarting in time when the AIO playbook finishes... sure enough, i found that there's maybe a 0.6 sec gap between the first restart and the first ssh | 22:33 |
mhayden | that could be the source of the ssh issues with the bootstrap-aio role | 22:33 |
palendae | lawl | 22:34 |
palendae | Build a sleep in? | 22:34 |
*** jmckind has joined #openstack-ansible | 22:34 | |
mhayden | if that fixes it, i plan to do things in this order: 1) hug d34dh0r53 2) cry into a beer 3) move to a shack by the river with no internet access | 22:34 |
mhayden | :P | 22:34 |
mhayden | wow | 22:35 |
mhayden | it worked | 22:35 |
coolj | put in 2 sheps for good measure | 22:35 |
mhayden | d34dh0r53: you're my hero there, sir | 22:35 |
mhayden | that was 2.5 days of head-scratching | 22:36 |
*** galstrom is now known as galstrom_zzz | 22:37 | |
*** jmckind has quit IRC | 22:38 | |
*** jmckind has joined #openstack-ansible | 22:39 | |
*** sdake has quit IRC | 22:43 | |
*** sdake has joined #openstack-ansible | 22:43 | |
*** spotz is now known as spotz_zzz | 22:44 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 22:55 | |
*** ysm has quit IRC | 22:56 | |
*** jhesketh has quit IRC | 23:00 | |
*** jhesketh has joined #openstack-ansible | 23:01 | |
*** jaypipes has quit IRC | 23:03 | |
*** jmckind has quit IRC | 23:14 | |
*** jimchou has quit IRC | 23:15 | |
*** sdake_ has joined #openstack-ansible | 23:16 | |
*** th720 has quit IRC | 23:17 | |
*** sdake has quit IRC | 23:18 | |
*** javeriak has quit IRC | 23:20 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!