| *** fawadkhaliq has quit IRC | 00:15 | |
| *** fawadkhaliq has joined #openstack-ansible | 00:17 | |
| *** fawadkhaliq has quit IRC | 00:28 | |
| *** thorst has joined #openstack-ansible | 00:29 | |
| *** thorst has quit IRC | 00:30 | |
| *** fawadkhaliq has joined #openstack-ansible | 00:30 | |
| *** thorst has joined #openstack-ansible | 00:31 | |
| *** fawadkhaliq has quit IRC | 00:34 | |
| *** fawadkhaliq has joined #openstack-ansible | 00:35 | |
| *** thorst has quit IRC | 00:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder: Cinder functional testing https://review.openstack.org/295265 | 00:37 |
|---|---|---|
| *** sdake has quit IRC | 00:41 | |
| *** fawadkhaliq has quit IRC | 00:41 | |
| *** galstrom is now known as galstrom_zzz | 00:42 | |
| cloudnull | if anyone is around to review https://review.openstack.org/#/c/296840/ itd be much appreciated. | 00:44 |
| cloudnull | the test is now building an indepth set of packages which was taken from one of my multi-node builds using mitaka | 00:45 |
| cloudnull | -cc automagically, jmccrory | 00:45 |
| jmccrory | cloudnull nice...12 minutes on ovh! | 00:51 |
| jmccrory | about 10 minutes there too | 00:51 |
| cloudnull | yea should be good all around. | 00:52 |
| cloudnull | im hoping for a ~10 min saving in the integrated gate too | 00:52 |
| *** klamath has quit IRC | 00:52 | |
| jmccrory | need to go back and find logan's averages, but repo_build took 23 minutes on the an ovh log i have saved | 00:54 |
| *** Mudpuppy has joined #openstack-ansible | 00:54 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-os_cinder: OpenStack services should reach Glance via the internal LB VIP https://review.openstack.org/290856 | 00:56 |
| cloudnull | jmccrory: http://paste.openstack.org/show/491537/ -- these were fairly common timeframes in my testing old vs new | 00:58 |
| *** Mudpuppy has quit IRC | 00:58 | |
| *** galstrom_zzz is now known as galstrom | 00:58 | |
| *** fawadkhaliq has joined #openstack-ansible | 01:08 | |
| *** fawadkhaliq has quit IRC | 01:11 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-repo_build: Updated repo_build process to be faster https://review.openstack.org/296840 | 01:27 |
| cloudnull | jmccrory: updated for your comments in the scripts | 01:28 |
| *** thorst has joined #openstack-ansible | 01:46 | |
| *** thorst has quit IRC | 01:47 | |
| *** iceyao has joined #openstack-ansible | 01:53 | |
| *** fawadkhaliq has joined #openstack-ansible | 01:56 | |
| *** weezS has joined #openstack-ansible | 01:56 | |
| *** d9k_ has quit IRC | 01:57 | |
| *** sdake has joined #openstack-ansible | 01:57 | |
| *** d9k_ has joined #openstack-ansible | 01:59 | |
| *** neilus has quit IRC | 02:03 | |
| *** jorge_munoz has quit IRC | 02:13 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-os_swift: Add ability to set up Swift container-sync realms https://review.openstack.org/286568 | 02:19 |
| *** jthorne has quit IRC | 02:19 | |
| *** jthorne has joined #openstack-ansible | 02:19 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone: Update role for stable/mitaka testing https://review.openstack.org/296838 | 02:23 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron: Update role for stable/mitaka testing https://review.openstack.org/296795 | 02:25 |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate: Update role for stable/mitaka testing https://review.openstack.org/296832 | 02:25 |
| *** galstrom is now known as galstrom_zzz | 02:26 | |
| *** fawadkhaliq has quit IRC | 02:27 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh: Update role for stable/mitaka testing https://review.openstack.org/296787 | 02:28 |
| *** jthorne has quit IRC | 02:28 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance: Update role for stable/mitaka testing https://review.openstack.org/296792 | 02:29 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer: Update role for stable/mitaka testing https://review.openstack.org/296789 | 02:31 |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat: Update role for stable/mitaka testing https://review.openstack.org/296794 | 02:34 |
| *** cemason has quit IRC | 02:38 | |
| * neillc heads out to be a taxi | 02:40 | |
| neillc | In case In don't get back in time, have a great Easter everyone. | 02:40 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova: Update role for stable/mitaka testing https://review.openstack.org/296796 | 02:40 |
| *** retreved has quit IRC | 02:47 | |
| *** galstrom_zzz is now known as galstrom | 02:47 | |
| *** cemason has joined #openstack-ansible | 03:02 | |
| *** iceyao has quit IRC | 03:08 | |
| *** sdake_ has joined #openstack-ansible | 03:21 | |
| *** sdake has quit IRC | 03:22 | |
| *** fawadkhaliq has joined #openstack-ansible | 04:03 | |
| jwitko | hey neillc, have you ever seen a case where the neutron dhcp agents weren't pingable and were not responding to DHCPDISCOVER requests? | 04:04 |
| jwitko | in my new cluster I just spun up (Kilo) I'm having this issue. | 04:04 |
| jwitko | the dnsmasq logs on the neutron agent container show the DHCPDISCOVER request coming through and its greeted with an error | 04:05 |
| jwitko | Mar 23 23:39:53 dnsmasq-dhcp[9113]: DHCPDISCOVER(ns-c4af496d-ab) fa:16:3e:b1:f7:6a no address available | 04:05 |
| *** galstrom is now known as galstrom_zzz | 04:06 | |
| *** hybridpollo has quit IRC | 04:23 | |
| *** fawadkhaliq has quit IRC | 04:27 | |
| *** fawadkhaliq has joined #openstack-ansible | 04:27 | |
| *** sdake_ is now known as sdake | 04:29 | |
| *** bryan_att has quit IRC | 04:42 | |
| openstackgerrit | Merged openstack/openstack-ansible: Remove old MariaDB apt sources https://review.openstack.org/289977 | 04:46 |
| *** chhavi has joined #openstack-ansible | 05:03 | |
| *** saneax is now known as saneax_AFK | 05:04 | |
| *** saneax_AFK is now known as saneax | 05:06 | |
| *** sdake has quit IRC | 05:16 | |
| *** sdake has joined #openstack-ansible | 05:18 | |
| *** javeriak has joined #openstack-ansible | 05:19 | |
| *** sdake_ has joined #openstack-ansible | 05:20 | |
| *** sdake has quit IRC | 05:23 | |
| *** markvoelker has quit IRC | 05:28 | |
| jwitko | So the VM can't see the DHCPOFFER coming from dnsmas on the neutron agent container | 05:31 |
| *** fawadkhaliq has quit IRC | 05:32 | |
| *** fawadkhaliq has joined #openstack-ansible | 05:32 | |
| *** weezS has quit IRC | 05:43 | |
| *** asettle has quit IRC | 05:44 | |
| *** asettle has joined #openstack-ansible | 05:44 | |
| *** asettle has quit IRC | 05:45 | |
| *** sdake_ is now known as sdake | 05:45 | |
| *** pcaruana has quit IRC | 05:48 | |
| *** markvoelker has joined #openstack-ansible | 06:29 | |
| *** daneyon_ has quit IRC | 06:34 | |
| *** markvoelker has quit IRC | 06:34 | |
| *** daneyon has joined #openstack-ansible | 06:34 | |
| *** finchd has quit IRC | 06:49 | |
| *** fawadkhaliq has quit IRC | 06:53 | |
| *** chhavi has quit IRC | 07:10 | |
| *** mikelk has joined #openstack-ansible | 07:11 | |
| *** jiteka has joined #openstack-ansible | 07:15 | |
| *** cristicalin has joined #openstack-ansible | 07:19 | |
| *** chhavi has joined #openstack-ansible | 07:27 | |
| *** pcaruana has joined #openstack-ansible | 07:32 | |
| *** markvoelker has joined #openstack-ansible | 07:56 | |
| *** finchd has joined #openstack-ansible | 08:00 | |
| *** markvoelker has quit IRC | 08:02 | |
| *** sdake has quit IRC | 08:02 | |
| *** sdake has joined #openstack-ansible | 08:06 | |
| *** cristicalin has quit IRC | 08:07 | |
| *** winggundamth has joined #openstack-ansible | 08:11 | |
| winggundamth | got this error on 12.0.8 while running openstack-ansible repo-install.yml http://paste.openstack.org/show/491670/ | 08:12 |
| winggundamth | Maybe repo container missing something? ImportError: No module named M2Crypto | 08:13 |
| winggundamth | oh maybe related to this https://bugs.launchpad.net/openstack-ansible/+bug/1559380 | 08:15 |
| openstack | Launchpad bug 1559380 in openstack-ansible trunk "repo build broken due to pywbem 0.8.1" [High,Fix committed] - Assigned to Kevin Carter (kevin-carter) | 08:15 |
| *** sdake has quit IRC | 08:18 | |
| *** jiteka has quit IRC | 08:19 | |
| winggundamth | will try to remove pywbem from cinder and swift and run playbook again | 08:20 |
| winggundamth | confirm that after remove pywbem as PR fixes the problem | 08:28 |
| *** neilus has joined #openstack-ansible | 08:48 | |
| odyssey4me | jwitko are you sure you're looking at the right log, and is that mac matching the instance mac? dnsmasq is basically saying that there are no addresses available to that mac - so either your virtual net is out of addresses (check the dhcp range), or that mac is not known to openstack and is therefore being ignored | 09:09 |
| odyssey4me | winggundamth well found :) | 09:09 |
| *** neilus has quit IRC | 09:12 | |
| *** neilus has joined #openstack-ansible | 09:12 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on Mitaka RC's https://review.openstack.org/296799 | 09:24 |
| mattt | odyssey4me: so what are we tackling today? | 09:25 |
| mattt | i kind of want to wait for https://review.openstack.org/#/c/296598/ to merge to see the fallout from that change | 09:25 |
| odyssey4me | mattt so it looks like https://review.openstack.org/296799 is failing on telemetry and then some basic ops too: http://logs.openstack.org/99/296799/3/check/gate-openstack-ansible-dsvm-commit/15e060a/console.html#_2016-03-24_03_54_14_078 | 09:27 |
| mattt | odyssey4me: shouldn't the roles be bumped first or has that been done already ? | 09:27 |
| odyssey4me | done last night | 09:28 |
| odyssey4me | jmccrory did a slight tqweak to your cinder api checks and it worked perfectly - apparently it's no longer a 200 response, it could be 200 (OK) or 300 (options) | 09:28 |
| odyssey4me | so the cinder checks are in | 09:29 |
| mattt | yeah i noticed, that's cool | 09:29 |
| mattt | ok let me see if i can replicate that telemetry stuff locally | 09:29 |
| mattt | unless you're already on it | 09:29 |
| odyssey4me | not on it atm | 09:29 |
| mattt | k | 09:29 |
| odyssey4me | I've promised to spend a bit more time on getting Ironic to a point of working, so I'm going to have to focus there. | 09:30 |
| odyssey4me | There is a strong drive to include it in the Mitaka release. | 09:30 |
| odyssey4me | (integrated) | 09:30 |
| mattt | sure | 09:30 |
| mattt | i actually have to switch gears for a sec today too, to look at that pbr wheels issue that we reported upstream a few months ago | 09:31 |
| mattt | because in mitaka keystone bundles wsgi scripts which we should be building, but at present our wheels don't include them | 09:31 |
| odyssey4me | ah yes - to try and have a more dynamic inclusion for the wsgi stuff | 09:31 |
| mattt | so sigmavirus24_awa actually fixed the issue, i just need to see if we can get that version of pbr into openstack/osa and then make the relevant changes to os_keystone | 09:32 |
| odyssey4me | well, the roles all build from stable/mitaka now instead of a SHA | 09:32 |
| odyssey4me | once requirements creates a stable/mitaka branch then we'll build from that too | 09:32 |
| odyssey4me | I expect that if we want that version of pbr used in openstack, then pbr will just have to tag | 09:33 |
| odyssey4me | we'll then inherit it from upper-constraints | 09:33 |
| mattt | yeah, i'll have a poke and see | 09:33 |
| *** admin0 has joined #openstack-ansible | 09:44 | |
| admin0 | good morning | 09:45 |
| *** d9k_ has quit IRC | 09:49 | |
| mattt | admin0: still waiting to see your docs PR :) | 09:55 |
| admin0 | today will start on it | 09:56 |
| admin0 | yesterday worked on something :) | 09:56 |
| admin0 | do curl gw.am :) | 09:56 |
| admin0 | from console | 09:56 |
| admin0 | the best is to do a file transfer . curl -T file gw.am | 09:56 |
| admin0 | and you can wget it back | 09:56 |
| admin0 | i was having a hassle doing rdp, console, vpn, console where copy/paste does not work | 09:57 |
| admin0 | hopefully people can find it useful as quick way to transfer files between machines | 09:57 |
| *** markvoelker has joined #openstack-ansible | 09:58 | |
| *** winggundamth has quit IRC | 10:00 | |
| *** jiteka has joined #openstack-ansible | 10:02 | |
| *** markvoelker has quit IRC | 10:03 | |
| *** moorryan has joined #openstack-ansible | 10:11 | |
| *** asettle has joined #openstack-ansible | 10:16 | |
| *** electrofelix has joined #openstack-ansible | 10:22 | |
| *** asettle has quit IRC | 10:24 | |
| *** saneax is now known as saneax_AFK | 10:30 | |
| *** saneax_AFK is now known as saneax | 10:45 | |
| *** admin0 has quit IRC | 10:53 | |
| *** admin0 has joined #openstack-ansible | 10:55 | |
| *** iceyao has joined #openstack-ansible | 11:01 | |
| *** moorryan has quit IRC | 11:08 | |
| *** brad[] has quit IRC | 11:08 | |
| *** brad[] has joined #openstack-ansible | 11:09 | |
| *** chhavi has quit IRC | 11:13 | |
| *** admin0 has quit IRC | 11:16 | |
| *** admin0 has joined #openstack-ansible | 11:16 | |
| *** chhavi has joined #openstack-ansible | 11:25 | |
| *** electrofelix has quit IRC | 11:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Ensure that wheel version used matches upper-constraints https://review.openstack.org/297037 | 11:29 |
| odyssey4me | mattt ^ please review when you have a moment - it's pretty crucial that we get that fix into this tag | 11:30 |
| odyssey4me | I'm going to have to figure out how to grab the upper-constraints version of wheel and use it in the specs. | 11:31 |
| mattt | k | 11:31 |
| *** weshay has joined #openstack-ansible | 11:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Always rebuild venvs when rebuilding wheels https://review.openstack.org/296618 | 11:38 |
| mattt | odyssey4me: how did that version of wheel slip into liberty? | 11:39 |
| mattt | thought the process of bumping was to manually check upper constraints file? | 11:39 |
| odyssey4me | https://github.com/openstack/openstack-ansible/commit/6ea8e986f1bc826ffd54477eddc1ea6f04b8ea53#diff-b4ef698db8ca845e5845c4618278f29a | 11:40 |
| odyssey4me | we didn't validate as we should have | 11:40 |
| mattt | k, +2d | 11:41 |
| odyssey4me | thanks mattt | 11:42 |
| *** thorst has joined #openstack-ansible | 11:43 | |
| mhayden | morning, folks | 11:57 |
| *** markvoelker has joined #openstack-ansible | 11:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add release note for pywbem issue https://review.openstack.org/297060 | 11:59 |
| odyssey4me | o/ mhayden how're your load balancers doing? are they actually running some sort of load or are they just present? | 12:00 |
| odyssey4me | mhayden I noticed that the Mitaka updated for Ceilometer include some meters for LBaaSv2 | 12:00 |
| mhayden | odyssey4me: i'm about to check -- just sat down with my coffee | 12:01 |
| *** neilus1 has joined #openstack-ansible | 12:01 | |
| mhayden | i wasn't running them under load overnight | 12:01 |
| odyssey4me | mattt https://review.openstack.org/297060 is another crucial one for this tag | 12:01 |
| mhayden | i forgot to run the script :| | 12:01 |
| mhayden | some meters would be nice, but there are some basic stats pulled into the database about the LB's | 12:01 |
| mhayden | like bytes in/out and requests | 12:01 |
| mhayden | i'm not sure if those can be queried via API, though | 12:01 |
| odyssey4me | and hey, mattt https://review.openstack.org/#/c/296598/ has merged! | 12:02 |
| odyssey4me | I don't know if it takes a period of time to actually happen on the images though. | 12:02 |
| odyssey4me | lemme ask | 12:02 |
| *** neilus1 has quit IRC | 12:02 | |
| *** neilus has quit IRC | 12:04 | |
| *** electrofelix has joined #openstack-ansible | 12:04 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create: Remove dependency on python2_lxc git source https://review.openstack.org/279788 | 12:07 |
| mhayden | whoa, made my first legit non-backport kilo commit | 12:09 |
| * mhayden thinks he should be proud of that, but he isn't sure | 12:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server: Remove dependency on python2_lxc git source https://review.openstack.org/279789 | 12:11 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server: Remove dependency on python2_lxc git source https://review.openstack.org/279790 | 12:13 |
| odyssey4me | congrats mhayden :) | 12:14 |
| *** neilus has joined #openstack-ansible | 12:14 | |
| odyssey4me | mhayden spotz_zzz please santise my terribad english in the release notes https://review.openstack.org/297060 and https://review.openstack.org/297037 | 12:16 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift: Remove dependency on python2_lxc git source https://review.openstack.org/297070 | 12:19 |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible-os_cinder: [WIP] Do not merge https://review.openstack.org/297071 | 12:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift: Update role for stable/mitaka testing https://review.openstack.org/297074 | 12:22 |
| odyssey4me | mattt it seems like it doesn't just work - I may need to pus some of that magic into the roles to make the python2_lxc git source removal work | 12:24 |
| odyssey4me | see how you go | 12:24 |
| mattt | with this change we should be able to drop use of ansible_python_interpreter and drop any cruft we were installing like lxc-python2 | 12:25 |
| mattt | so far my test review is still running, which is slightly concerning | 12:26 |
| odyssey4me | mattt give the gremlins a little time to do their thing :p | 12:27 |
| *** woodard has joined #openstack-ansible | 12:32 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Update source-branch-updater to work with IRR's https://review.openstack.org/297083 | 12:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on Mitaka RC's https://review.openstack.org/296799 | 12:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on Mitaka RC's https://review.openstack.org/296799 | 12:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on Mitaka RC's https://review.openstack.org/296799 | 12:34 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on Mitaka RC's https://review.openstack.org/296799 | 12:35 |
| odyssey4me | sorry for the spam | 12:35 |
| odyssey4me | <---- being a bit stupid today | 12:35 |
| odyssey4me | mattt so https://review.openstack.org/297071 passed, which is good - the question is whether it'll pass after removing all the cruft | 12:36 |
| odyssey4me | I have the sneaky suspicion that the change to sudoers may only take effect once the current nodepool's prepped nodes are used up - so we may have to wait for this all to take effect until later, or tomorrow. | 12:37 |
| *** pjm_ has joined #openstack-ansible | 12:38 | |
| pjm_ | Hi there | 12:38 |
| odyssey4me | andymccr mattt we need to get https://review.openstack.org/297074 approved in order to unblock the swift role's builds | 12:39 |
| *** electrofelix has quit IRC | 12:39 | |
| mattt | looking | 12:39 |
| mattt | odyssey4me: https://github.com/openstack/openstack-ansible-os_swift/blob/master/tests/test.yml#L350 | 12:42 |
| mattt | should that not be changed also ? | 12:42 |
| automagically | morning | 12:43 |
| pjm_ | can I make a suggestion ? | 12:44 |
| mattt | pjm_: go for it :) | 12:45 |
| mattt | morning automagically | 12:45 |
| pjm_ | in the file https://github.com/openstack/openstack-ansible/blob/36791049643a566edb4fcf11829f9bf005333949/doc/source/install-guide/configure-fwaas.rst | 12:45 |
| pjm_ | it wouldn't be more generic if was configure-network-services.rst | 12:45 |
| pjm_ | and chaning the title for | 12:45 |
| pjm_ | “Configuring the Network Firewall Service (Optional)” -> “Configuring the Network Services (Optional)” | 12:46 |
| mattt | pjm_: sensible sounding suggestion | 12:46 |
| mattt | pjm_: is that a change you yourself can propose on gerrit ? | 12:46 |
| pjm_ | because if I understand well, this readme tells how configure the network services related | 12:46 |
| pjm_ | sure =) if its not absurd | 12:47 |
| mattt | it's very easy! the part i forget is how to sign the openstack contributors agreement :) | 12:49 |
| mattt | pjm_: http://docs.openstack.org/infra/manual/developers.html#getting-started | 12:49 |
| pjm_ | matt: thanks :) | 12:49 |
| pjm_ | another thing, the file mentioned in the readme | 12:50 |
| pjm_ | don't have the options | 12:50 |
| *** chhavi_ has joined #openstack-ansible | 12:50 | |
| pjm_ | neutron_plugin_base: | 12:50 |
| *** neilus has quit IRC | 12:50 | |
| mattt | pjm_: yeah, the file in the README is basically the place where you add all the overrides for your environment | 12:50 |
| mattt | so unless you have a specific need to override something the default will be applied | 12:51 |
| pjm_ | I tried that steps enabling all the services and neutron stops working, commented that line and works again. I will try find it =) | 12:51 |
| mattt | pjm_: openstack-ansible is composed of a bunch of 'modules', each with their own set of defaults | 12:51 |
| *** chhavi has quit IRC | 12:51 | |
| pjm_ | right, and that file will overwrite any defaults existing in the playbooks, right? | 12:52 |
| mattt | pjm_: https://github.com/openstack/openstack-ansible-os_neutron/blob/master/defaults/main.yml#L107 | 12:52 |
| pjm_ | ok, that makes sense =) thanks | 12:53 |
| mattt | pjm_: np, happy to help you get that review in when you've done all the necessary bits | 12:54 |
| *** galstrom_zzz is now known as galstrom | 12:54 | |
| *** galstrom is now known as galstrom_zzz | 12:55 | |
| odyssey4me | mattt - swift doesn't have a stable/mitaka branch yet - so master is fine | 12:59 |
| odyssey4me | we'll switch that once they cut thr branch | 12:59 |
| odyssey4me | *the | 12:59 |
| *** johnmilton has joined #openstack-ansible | 13:00 | |
| odyssey4me | pjm_ note that if you wish to change the title of http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-fwaas.html then you may have to merge it with http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-lbaas.html too which makes it a little unwieldy - have a go at it though, we'd like to see your proposed change | 13:01 |
| pjm_ | Thanks odyssey4me I didn't see that file | 13:02 |
| pjm_ | i'm reading the website that mattt give to me to complete the proposed change =) | 13:02 |
| odyssey4me | awesome :) | 13:03 |
| *** neilus has joined #openstack-ansible | 13:06 | |
| odyssey4me | o/ automagically how're you doing? | 13:09 |
| automagically | Just great, thanks | 13:09 |
| mattt | does the US break for easter ? | 13:09 |
| automagically | Some folks have good friday off, but its rare | 13:10 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Include security role in setup-hosts.yml https://review.openstack.org/290526 | 13:11 |
| odyssey4me | cloudnull with regards to https://review.openstack.org/296840 - there may be some more optimisations we can use - see http://stackoverflow.com/questions/26228136/pip-build-option-to-use-multicore / http://stackoverflow.com/questions/11021130/parallel-pip-install (maybe max-procs works with wheel?) | 13:11 |
| cloudnull | sadly pip wheel does not have the --install-options flag | 13:13 |
| odyssey4me | cloudnull typey typer make-the-fix :) | 13:13 |
| * cloudnull working on it | 13:13 | |
| odyssey4me | there are two issues that would help us a lot to be solved - one is improving the use of parallel processing (for downloads, and wheel building), the other is being able to actually resolve a full tree of dependencies | 13:14 |
| odyssey4me | if we could resolve deps properly, we could drop carrying arbitrary wheels and simply have the resolving work from our requirements and from the services being deployed... instead of building wheels for every single thing in global-requirements | 13:15 |
| odyssey4me | it's obviously not crucial for now, but would be highly beneficial - we can always target tackling that in the next cycle | 13:17 |
| cloudnull | I agree | 13:18 |
| cloudnull | I'd love to work on that and maybe sigmavirus24_awa could help us out with some of that too . | 13:19 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Improved search for unlocked system accounts https://review.openstack.org/290059 | 13:19 |
| odyssey4me | cloudnull ++ | 13:19 |
| odyssey4me | both would be great additions to pip which so many projects depend on, I'm sure the project would appreciate the submission | 13:19 |
| odyssey4me | it looks like dstufft was working on something at one time and did some improvements - see https://github.com/pypa/pip/issues/988 | 13:20 |
| cloudnull | as an aside adding install-options to the build/install process causes "UserWarning: Disabling all use of wheels due to the use of --build-options / --global-options / --install-options.cmdoptions.check_install_build_global(options)" | 13:21 |
| odyssey4me | it may be worth approaching him in #pypa-dev to see where help is needed | 13:21 |
| odyssey4me | yuk :( | 13:21 |
| *** javeriak has quit IRC | 13:26 | |
| *** b3rnard0_away is now known as b3rnard0 | 13:27 | |
| cloudnull | bummer. its the same thing if i use build-options in the wheel command. | 13:28 |
| cloudnull | pip becomes very unhappy | 13:28 |
| odyssey4me | :/ it was worth a try | 13:30 |
| cloudnull | im trying the xarg bits now | 13:33 |
| *** cemason has quit IRC | 13:34 | |
| odyssey4me | would you mind making this go? https://review.openstack.org/297074 | 13:34 |
| odyssey4me | also https://review.openstack.org/297037 / https://review.openstack.org/297060 | 13:35 |
| cloudnull | xargs is a fail w/ wheel too. | 13:38 |
| cloudnull | if 2 packages attempt to build at the same time, due to dependencies because we dont have them listed twice in the requirements file, it causes a collision and the process dies. | 13:39 |
| odyssey4me | hrm | 13:41 |
| *** jthorne has joined #openstack-ansible | 13:41 | |
| mattt | odyssey4me: do we want to explore this any further or shall i kill it? https://review.openstack.org/#/c/292911/ | 13:41 |
| odyssey4me | mattt I think we should kill it | 13:42 |
| mattt | done! | 13:43 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-repo_build: Updated repo_build process to be faster https://review.openstack.org/296840 | 13:51 |
| mhayden | hughsaunders: i'm not opposed to your suggestion here, but i cannot find the right incantation to make it work: https://review.openstack.org/#/c/290059/3/tasks/auth.yml | 13:51 |
| mhayden | i can't get it to work via shell or raw | 13:51 |
| cloudnull | odyssey4me automagically jmccrory ^ updated because it was in merge conflict | 13:51 |
| mhayden | either the YAML parser and/or bash complains | 13:51 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift: Update role for stable/mitaka testing https://review.openstack.org/297074 | 13:52 |
| hughsaunders | mhayden: ahh, I tried it in terminal, but didn't try embedding in yaml, the result is the same as your current tasks, its just a bit shorter, so if it doesn't embed well we can got with the current istuation | 13:52 |
| mhayden | hughsaunders: i might do that | 13:52 |
| cloudnull | odyssey4me: idk if we're going to get the build process to be much faster at this point (given the limitations with parallel building). | 13:53 |
| cloudnull | fairly good improvement overall IMO, before: ~25 min, after: ~10 min | 13:53 |
| cloudnull | maybe we can make it better with some upstream pip work | 13:54 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift: Add ability to set up Swift container-sync realms https://review.openstack.org/286568 | 13:54 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift: Remove dependency on python2_lxc git source https://review.openstack.org/297070 | 13:54 |
| odyssey4me | yup, sounds good - happy to have some optimisation and I think working on improving this upstream (with pip and Ansible) would be good | 13:55 |
| cloudnull | ++ | 13:56 |
| odyssey4me | cloudnull with regards to https://review.openstack.org/296839 - why don't we switch to using a pip install for installing ansible? it would allow us to set it as a pin in requirements.txt and the source branch updater can update the pins appropriately | 13:56 |
| cloudnull | idk if that works . it hadnt in the past | 13:57 |
| cloudnull | pip install was not bringing in the core / extra modules. | 13:57 |
| *** jaypipes has joined #openstack-ansible | 13:57 | |
| cloudnull | if it is fixed we should do that | 13:58 |
| odyssey4me | the reason for using the git source in the first place was to ensure that we pull the extras in - those are now included in the 'ansible' pip package | 13:58 |
| odyssey4me | they have been for quite some time | 13:58 |
| cloudnull | well, it would seem im stuck in the past. | 14:01 |
| * cloudnull yells at some kids to get off his lawn | 14:01 | |
| odyssey4me | lol | 14:01 |
| *** mgoddard_ has joined #openstack-ansible | 14:01 | |
| mattt | going to start on the nova functional tests | 14:03 |
| mattt | wish me luck | 14:03 |
| *** mgoddard has quit IRC | 14:04 | |
| cloudnull | godspeed mattt | 14:06 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:09 | |
| *** KLevenstein has joined #openstack-ansible | 14:11 | |
| *** retreved has joined #openstack-ansible | 14:11 | |
| *** asettle has joined #openstack-ansible | 14:12 | |
| sigmavirus24 | cloudnull: I had ideas for parallel processing of dependencies | 14:12 |
| sigmavirus24 | it's plausible but we could run into some non-unique versions possibly | 14:12 |
| cloudnull | sigmavirus24: assuming the project requirements are setup properly, i'd imagine that should be ok. | 14:13 |
| sigmavirus24 | cloudnull: right. We are using upper-constraints now too, right? | 14:14 |
| cloudnull | we are | 14:14 |
| jwitko | Hey guys, I'm having some trouble with DHCP in my enivronment. I can not ping the DHCP servers that openstack sets up when I create a new subnet and with further analysis of the dnsmasq logs I see that the VMs within my openstack cluster can not see DHCP packets coming from the DHCP agents | 14:14 |
| cloudnull | jwitko: liberty ? | 14:15 |
| *** spotz_zzz is now known as spotz | 14:15 | |
| jwitko | cloudnull, sorry, Kilo | 14:15 |
| jwitko | OSA release 11.2.9 | 14:15 |
| odyssey4me | jwitko check out your backscroll -did you see my feedback from earlier? | 14:16 |
| *** asettle has quit IRC | 14:16 | |
| jwitko | sorry just woke up, let me check (that was late last night) | 14:16 |
| openstackgerrit | Pedro Magalhães proposed openstack/openstack-ansible: Joining documentation of network services in one file https://review.openstack.org/297179 | 14:17 |
| mattt | ^^^ pjm_ nice :) | 14:17 |
| jwitko | so odyssey4me, thanks for that feedback. i investigated more after I posted that last night and basically you were correct. DNSMasq did not know about the MAC address in its host file. However restarting neutron-dhcp-agent made it recognize the MAC. and Yes I confirmed the MAC matches the interface on my VM (Although to note this is a brand new installation and therefore only | 14:17 |
| jwitko | has this 1 test VM i created) | 14:17 |
| pjm_ | thanks for the help matt and odyssey4me =) | 14:18 |
| pjm_ | hope its help | 14:18 |
| jwitko | so what I did to get to this point, where I knew the MAC existed, was create the VM and then restart the DHCP agent on neutron. The MAC and IP mapping now exists in the dnsmasq host file | 14:18 |
| jwitko | That is when I actually got dnsmasq to respond to the DHCPDISCOVER with a DHCPOFFER, however a tcpdump on the VM shows that only outgoing DHCP packets are even being seen | 14:18 |
| jwitko | the VM sees absolutely no incoming traffic on port 67 or 68 UDP | 14:19 |
| jwitko | I do see the DHCPDISCOVER packets leaving the VM, and I see dnsmasq picking up the discover packet and returning an OFFER | 14:19 |
| jwitko | but thats it | 14:19 |
| odyssey4me | automagically as far as I recall, either you or jmccrory work behind a proxy - can you validate that https://review.openstack.org/296840 works properly behind a proxy? | 14:19 |
| automagically | odyssey4me: Yeah, I saw your note on that. I will attempt to validate later today | 14:20 |
| odyssey4me | jwitko hrm, I'd love to help but I need to get some other stuff working on a deadline - hopefully someone else can assist | 14:20 |
| *** mgoddard_ has quit IRC | 14:23 | |
| *** mgoddard has joined #openstack-ansible | 14:23 | |
| openstackgerrit | Pedro Magalhães proposed openstack/openstack-ansible: Joining documentation of network services in one file https://review.openstack.org/297179 | 14:23 |
| odyssey4me | mattt I've checking out a build using the updated SHA's - plenty of nova brokenness - are you going to figure those out? | 14:24 |
| mattt | odyssey4me: yeah, my initial full build failed because of cloud networking i think, got another build which is just about done | 14:25 |
| odyssey4me | ok, I'll leave nova to you | 14:25 |
| odyssey4me | other brokenness is in glance and ceilometer - lemme take a crack at glance quickly | 14:26 |
| mattt | alright, looking at nova then ceilometer | 14:27 |
| spotz | cloudnull the comment in https://review.openstack.org/#/c/296490/ says it's abandon and the review is now in the same review FYI | 14:27 |
| mattt | odyssey4me: presumably nova is busted because glance is busted | 14:27 |
| mattt | having a peek at glance also | 14:28 |
| jwitko | odyssey4me, no problem. maybe later if you're free | 14:28 |
| *** d9k_ has joined #openstack-ansible | 14:30 | |
| cloudnull | spotz: ah. yea cp error | 14:31 |
| cloudnull | https://review.openstack.org/#/c/296840/ | 14:31 |
| cloudnull | thats the new review | 14:31 |
| odyssey4me | mattt BackendException: Cannot find swift service endpoint : Expecting to find domain in project - the server could not comply with the request since it is either malformed or otherwise incorrect. The client is assumed to be in error. (HTTP 400) | 14:31 |
| Nepoc | Greetings! I have some "issues" with overriding the galera my.cnf. I have to add in skip-name-resolve and skip-host-cache but these are not in the format of variable = setting. How can I correctly format the override to just have the text skip-name-resolve entered into the mysqld section of the config (with out an = on the line) | 14:31 |
| spotz | Thanks cloudnull in my pile from channel and email notices | 14:32 |
| *** ametts has joined #openstack-ansible | 14:33 | |
| cloudnull | if any cores are around to review and potentially make this go https://review.openstack.org/#/c/296840/ It'd be much appreciated. | 14:33 |
| odyssey4me | cloudnull yep, waiting for a behind-the-proxy test by automagically | 14:34 |
| *** Mudpuppy has joined #openstack-ansible | 14:36 | |
| *** sdake has joined #openstack-ansible | 14:37 | |
| *** elopez has quit IRC | 14:38 | |
| Nepoc | I'm guessing the overrides can't do what I need.... or I just don't know the format | 14:41 |
| Nepoc | I either end up with and = in the line or I get a new [DEFAULT] section at the top | 14:42 |
| *** pjm_ has quit IRC | 14:42 | |
| *** rohanp_ has joined #openstack-ansible | 14:42 | |
| openstackgerrit | Merged openstack/openstack-ansible: Ensure that wheel version used matches upper-constraints https://review.openstack.org/297037 | 14:45 |
| Nepoc | galera_my_cnf_overrides: | 14:46 |
| Nepoc | mysqld: | 14:46 |
| Nepoc | skip-host-cache: <- anytime there is a : in this line you get an = in the line. If you don't then you end up with the [DEFAULT] section and a mysqld = [skip-host-cache,skip-name-resolve] | 14:46 |
| cloudnull | Nepoc: this will work for that | 14:47 |
| cloudnull | http://cdn.pasteraw.com/cyysyej9xdnuvx37fey85vpuedk9qh | 14:47 |
| cloudnull | whille most docs say to just set the name, the run time config with just the name is NAME=1 | 14:47 |
| cloudnull | so if you set that using the config template the galera setup will do the right thing | 14:48 |
| Nepoc | cloudnull: trying it out :) | 14:48 |
| cloudnull | let me know how it goes | 14:48 |
| Nepoc | [ERROR] /usr/sbin/mysqld: option '--skip-host-cache' cannot take an argument | 14:49 |
| Nepoc | :( | 14:49 |
| *** electrofelix has joined #openstack-ansible | 14:49 | |
| *** electrofelix has quit IRC | 14:50 | |
| cloudnull | skip-name-resolve is working right ? | 14:50 |
| spotz | mhayden LBaas v2 is Octavia and not Neutron correct? | 14:50 |
| cloudnull | im looking at skip-host-cache | 14:50 |
| Nepoc | Yes looks like skip-name-resolve is working | 14:51 |
| *** jorge_munoz has joined #openstack-ansible | 14:51 | |
| automagically | cloudnull: Starting the behind the proxy test now | 14:51 |
| Nepoc | I'll try running without skip-host-cache and see if I have success | 14:52 |
| cloudnull | Nepoc: http://cdn.pasteraw.com/ik2s23afhfavff1mg1foy4zg70940n6 -- it doesnt seem that an option in galera | 14:52 |
| cloudnull | rather mariadb | 14:52 |
| cloudnull | maybe it reads it in as a cli arg ? | 14:52 |
| cloudnull | i did a look up using: show variables like "%skip%"; | 14:53 |
| Nepoc | Interesting, well I'll just go without. I assumed it was "working" since it didn't complain with it set manually in the file :) | 14:54 |
| *** jorge_munoz_ has joined #openstack-ansible | 14:55 | |
| Nepoc | cloudnull: thank you for the help again | 14:55 |
| cloudnull | Nepoc: are you testing kilo ? | 14:56 |
| Nepoc | liberty | 14:56 |
| *** jorge_munoz has quit IRC | 14:56 | |
| *** jorge_munoz_ is now known as jorge_munoz | 14:56 | |
| cloudnull | ok, so https://dev.mysql.com/doc/refman/5.5/en/host-cache.html that option was there is 5.5 but it seems it was removed in 5.6 (mariadb 10) | 14:57 |
| cloudnull | it seems skip-name-resolve is all you need going forward. | 14:58 |
| *** lkoranda has quit IRC | 15:02 | |
| *** lkoranda has joined #openstack-ansible | 15:02 | |
| *** mgoddard_ has joined #openstack-ansible | 15:05 | |
| *** chhavi_ has quit IRC | 15:06 | |
| Nepoc | Excellent :) | 15:07 |
| *** mgoddard has quit IRC | 15:09 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Improved search for unlocked system accounts https://review.openstack.org/290059 | 15:10 |
| *** hughsaunders_ has joined #openstack-ansible | 15:11 | |
| *** adreznec_ has joined #openstack-ansible | 15:12 | |
| mhayden | spotz: LBaaSv2 only is a specification for the *API* | 15:12 |
| mhayden | the implementation can be haproxy via agent or octavia | 15:12 |
| mhayden | that confused me for a while, too | 15:12 |
| spotz | mhayden Ok just wanted to be sure for a doc review we've got | 15:12 |
| *** admin0 has quit IRC | 15:13 | |
| *** iceyao has quit IRC | 15:13 | |
| *** scarlisle has quit IRC | 15:13 | |
| *** tiagogomes_ has quit IRC | 15:13 | |
| *** fignew has quit IRC | 15:13 | |
| *** kysse has quit IRC | 15:13 | |
| *** hughsaunders has quit IRC | 15:13 | |
| *** admiralboom has quit IRC | 15:13 | |
| *** adreznec has quit IRC | 15:13 | |
| *** dmsimard has quit IRC | 15:13 | |
| *** hughsaunders_ is now known as hughsaunders | 15:13 | |
| Nepoc | Is this a known error? sysctl: separators should not be repeated: ..rp_filter sysctl: cannot stat /proc/sys/net/ipv4/conf//rp_filter: No such file or directory | 15:14 |
| Nepoc | Looks like it's missing the interface | 15:14 |
| *** mgoddard_ has quit IRC | 15:15 | |
| *** mgoddard has joined #openstack-ansible | 15:16 | |
| *** adreznec_ is now known as adreznec | 15:19 | |
| openstackgerrit | Pedro Magalhães proposed openstack/openstack-ansible: Joining documentation of network services in one file https://review.openstack.org/297179 | 15:22 |
| *** pjm_ has joined #openstack-ansible | 15:23 | |
| *** scarlisle has joined #openstack-ansible | 15:25 | |
| *** admin0 has joined #openstack-ansible | 15:25 | |
| *** dmsimard has joined #openstack-ansible | 15:25 | |
| *** sdake_ has joined #openstack-ansible | 15:25 | |
| *** sdake has quit IRC | 15:25 | |
| *** fignew has joined #openstack-ansible | 15:25 | |
| *** weezS has joined #openstack-ansible | 15:27 | |
| *** tiagogomes_ has joined #openstack-ansible | 15:29 | |
| *** kysse has joined #openstack-ansible | 15:29 | |
| *** admiralboom has joined #openstack-ansible | 15:29 | |
| jwitko | cloudnull, hey can I ask for your help on a DHCP issue within the neutron dhcp agent? I just spun up a new stack on Kilo 11.2.9 release and my VMs can't get DHCP packets from neutron-dhcp-agent. | 15:31 |
| cloudnull | do you see the request leaving the neutron-agent container ? | 15:32 |
| jwitko | Yes, dnsmasq logs and tcpdump show DHCPOFFER being sent | 15:32 |
| cloudnull | on the compute host do you see it as well ? | 15:32 |
| jwitko | you know i actually didn't check that, let me go find out now | 15:32 |
| *** oneswig has joined #openstack-ansible | 15:33 | |
| cloudnull | also is the vm being built using a provider network or tenant network? | 15:33 |
| *** asettle has joined #openstack-ansible | 15:34 | |
| jwitko | cloudnull, the VM is using provider network. the compute host sees both DHCP coming in and going Out | 15:38 |
| jwitko | 11:37:47.303369 B fa:16:3e:b1:f7:6a ethertype IPv4 (0x0800), length 348: 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from fa:16:3e:b1:f7:6a, length 304 | 15:38 |
| openstackgerrit | Merged openstack/openstack-ansible: Add release note for pywbem issue https://review.openstack.org/297060 | 15:38 |
| jwitko | 11:37:47.303466 Out fa:16:3e:b1:f7:6a ethertype IPv4 (0x0800), length 348: 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from fa:16:3e:b1:f7:6a, length 304 | 15:38 |
| *** asettle has quit IRC | 15:38 | |
| cloudnull | so the request leaves the compute host, hits dnsmasq within the neutron container, returns to compute node? | 15:40 |
| odyssey4me | mattt I ended up in a meeting. Have you made progress with the failing master integrated build? | 15:40 |
| cloudnull | does the neutron linuxbridge agent show anything in the logs indicating there is an issue ? | 15:41 |
| mattt | odyssey4me: negative | 15:45 |
| odyssey4me | mattt okieodokey | 15:45 |
| *** mgoddard_ has joined #openstack-ansible | 15:45 | |
| mattt | odyssey4me: all i know is it's related to using keystone v3 for the swift glance_storage, and whatever the issue is was introduced between 0.11.0 and 0.12.0 (feb 12 -> marc 1) | 15:45 |
| odyssey4me | mattt I thought we had this one licked. A regression? | 15:46 |
| mattt | the issues are all starting to merge into 1 for me :) | 15:47 |
| jwitko | cloudnull, so i think the path is currently: VM --> Compute node --> Controller node --> neutron agent container --> Controller node --> Compute node --Death-- | 15:48 |
| *** mgoddard has quit IRC | 15:49 | |
| jwitko | cloudnull, the linuxbridge-agent logs show nothing at all unhappy | 15:49 |
| jwitko | just gave em' a restart for good measure though | 15:49 |
| cloudnull | do you see the same issue if you use a tenant network instead of a provider network ? | 15:49 |
| *** skamithi has joined #openstack-ansible | 15:50 | |
| jwitko | I havent tried, never done that before. | 15:52 |
| mattt | odyssey4me: ok i /think/ this is the one that broke it https://github.com/openstack/glance_store/commit/1b782cee8552ec02f7303ee6f9ba9d1f2c180d07 | 15:52 |
| odyssey4me | sigmavirus24 ^ potential keystone v3 API regression for glance_store | 15:53 |
| mattt | may not be a regression, still looking at it :) | 15:54 |
| sigmavirus24 | hah | 15:54 |
| odyssey4me | it looks, at a glance, like it has all the right things to pass into the ksclient - are we setting all the needful things? | 15:54 |
| sigmavirus24 | has no one ever learned how to not use the versioned portions of libraries like that? | 15:55 |
| sigmavirus24 | those libraries go to so much trouble to provide abstractions | 15:55 |
| sigmavirus24 | and no one ever bothers using them | 15:55 |
| sigmavirus24 | </rant> | 15:55 |
| cloudnull | jwitko: are you running with l2population ? | 15:56 |
| odyssey4me | at first blush, it seems we're doing the right things: https://github.com/openstack/openstack-ansible-os_glance/blob/master/templates/glance-api.conf.j2#L98-L113 | 15:57 |
| jwitko | cloudnull, I'm not sure how can I tell? | 15:57 |
| jwitko | cloudnull, I ran dhcpdump on the tap interface for the VM on the compute node. that interface only ever sees DHCPDISCOVER, never anything in return | 15:57 |
| cloudnull | in kilo its on by default https://github.com/openstack/openstack-ansible/blob/kilo/playbooks/roles/os_neutron/defaults/main.yml#L192 | 15:58 |
| cloudnull | and you see the discover in the agents contianer ? | 15:58 |
| jwitko | so, at this point i'm using a dhcpdiscover at the compute node vlan bridge | 15:59 |
| odyssey4me | mattt sigmavirus24 it does seem like we're supplying 'swift_store_user', for example, and the library is looking for 'username' | 15:59 |
| jwitko | and I don't see any DHCPOFFER reply | 15:59 |
| jwitko | so I need to go see exactly where the offer stops | 15:59 |
| *** neilus has quit IRC | 15:59 | |
| sigmavirus24 | odyssey4me: that should be the correct thing to specify | 16:00 |
| odyssey4me | I'm not sure if there's some sort of key munging that happens between glance and glance_store | 16:00 |
| *** mikelk has quit IRC | 16:00 | |
| cloudnull | jwitko: would you mind also doing an ``lxc-stop -n $NEUTRON_AGENT_CONTAINER; lxc-start -n $NEUTRON_AGENT_CONTAINER;`` just to rule out namespace and tap device tom foolery | 16:00 |
| *** jmccrory_ has joined #openstack-ansible | 16:00 | |
| cloudnull | rather: ``lxc-stop -n $NEUTRON_AGENT_CONTAINER; lxc-start -dn $NEUTRON_AGENT_CONTAINER;`` | 16:01 |
| mattt | are we using single or multi-tenant store ? | 16:01 |
| sigmavirus24 | odyssey4me: https://github.com/openstack/glance_store/blob/master/glance_store/_drivers/swift/store.py#L59 | 16:01 |
| sigmavirus24 | mattt: single-tenant iirc | 16:01 |
| sigmavirus24 | we're storing everything as the glance service user iirc which would be single-tenant | 16:01 |
| mattt | yeah that's it | 16:01 |
| automagically | odyssey4me: Meeting? | 16:01 |
| *** KLevenstein has quit IRC | 16:01 | |
| sigmavirus24 | I think multi-tenant store would require the swift driver to get a trust per-user and would then use the user's tenant to store the image data in swift | 16:01 |
| odyssey4me | ah yes - time crept up on me | 16:01 |
| odyssey4me | meeting in #openstack-meeting-4 cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden, scarlisle, luckyinva, ntt, javeriak, automagically, spotz, vdo, jmccrory, alextricity25, jasondotstar, KLevenstein, admin0, michaelgugino, | 16:02 |
| odyssey4me | ametts, v1k0d3n, severion, bgmccollum | 16:02 |
| mattt | sigmavirus24: i may pop into #openstack-glance to see if someone wants to have a peep | 16:05 |
| sigmavirus24 | mattt: can you give me more details | 16:05 |
| sigmavirus24 | I only got pinged regarding a possible regression in glance_store | 16:05 |
| sigmavirus24 | Idk what we're talking about otherwise | 16:06 |
| *** admin0 has quit IRC | 16:06 | |
| jwitko | cloudnull, so it looks like the cut-off is between the controller and the compute | 16:08 |
| mattt | sigmavirus24: so we have a review to bump shas to stable/mitaka | 16:08 |
| mattt | sigmavirus24: that broke glance, i'm seeing this in the logs https://gist.githubusercontent.com/mattt416/badf92f9636121945552/raw/92b16388d238bf75102d95e5fb2a132d41e8c345/gistfile1.txt | 16:08 |
| mattt | sigmavirus24: i tried going through minor versions of the glance store, and issue arises between 0.11.0 and 0.12.0 | 16:09 |
| mattt | sigmavirus24: believe it's https://github.com/openstack/glance_store/commit/1b782cee8552ec02f7303ee6f9ba9d1f2c180d07 | 16:09 |
| jwitko | cloudnull, the path is currently: VM --> Compute node --> Controller node --> neutron agent container --> Controller node --> --Death-- | 16:09 |
| mattt | sigmavirus24: that's all i know at the moment | 16:09 |
| jwitko | i will try to restart the container now like you asked | 16:09 |
| sigmavirus24 | mattt: I'm reading some codes for you | 16:09 |
| mattt | sigmavirus24: u da bayst | 16:10 |
| mattt | sigmavirus24: also flipping to keystone v2 in glance store also makes it work | 16:10 |
| *** mgoddard_ has quit IRC | 16:14 | |
| *** mgoddard has joined #openstack-ansible | 16:14 | |
| sigmavirus24 | mattt: I think it's related to https://github.com/openstack/glance_store/blob/master/glance_store/_drivers/swift/store.py#L808 and https://github.com/openstack/glance_store/blob/master/glance_store/_drivers/swift/store.py#L951 | 16:16 |
| sigmavirus24 | but I haven't figured out exactly how yet ;) | 16:16 |
| prometheanfire | is OSA following the mitaka branch of projects that have that yet? | 16:17 |
| cloudnull | prometheanfire: in the projects yes, the intgrated gate is still pending | 16:17 |
| cloudnull | i believe | 16:17 |
| sigmavirus24 | Found it https://github.com/openstack/glance_store/blob/master/glance_store/_drivers/swift/utils.py#L79 | 16:18 |
| *** mgoddard_ has joined #openstack-ansible | 16:18 | |
| odyssey4me | prometheanfire yes, in the roles - there are some kinks being worked out in the integrated build | 16:19 |
| prometheanfire | cloudnull: k, was just curious, as not doing that could make it hard to cut a proper mitaka release | 16:19 |
| mattt | sigmavirus24: what is it, the fact that the default project is being set to None ? | 16:19 |
| prometheanfire | odyssey4me: thanks | 16:19 |
| mattt | oh return {} also | 16:20 |
| mattt | weird | 16:20 |
| jwitko | cloudnull, so 2/3 restarted without issue. the third one is not stopping | 16:21 |
| jwitko | i tried --kill but no luck | 16:21 |
| *** mgoddard has quit IRC | 16:21 | |
| cloudnull | lxc-stop -kn NAME | 16:21 |
| cloudnull | no go ? | 16:22 |
| sigmavirus24 | mattt: yeah, I'm puzzling through this minefield | 16:22 |
| sigmavirus24 | don't have a full explanation yet | 16:22 |
| sigmavirus24 | but I'm documenting what I find for others playing along at home | 16:22 |
| mattt | hehe ok thanks sigmavirus24 | 16:22 |
| mattt | shall i put any of my findings into a glance bug or just leave it for now ? | 16:22 |
| jwitko | cloudnull, no that does not work. BUT! you were right, restarting the containers fixed it :) | 16:24 |
| jwitko | the other two came up and are serving dhcp | 16:24 |
| jwitko | VM now got IP | 16:25 |
| cloudnull | woot! | 16:25 |
| cloudnull | I wonder if the issue was a tap device in the container or if the container had a broken veth | 16:25 |
| sigmavirus24 | mattt: leave it for now | 16:30 |
| sigmavirus24 | mattt: can confirm that is the code to the SingleTenant store | 16:31 |
| mattt | cool, shame there's no way to turn off this connection manager | 16:31 |
| mattt | odyssey4me: shall i update the sha bump to use the last good SHA ? | 16:31 |
| mattt | actually derp, glance_store is pulled in implicitly | 16:31 |
| odyssey4me | mattt yep, glance_store comes in from the requirements | 16:32 |
| mattt | well that's a problem, it's glance_store>=0.13.0 | 16:33 |
| mattt | busted version is 0.12.0 | 16:33 |
| *** sdake_ is now known as sdake | 16:34 | |
| *** Bjoern has joined #openstack-ansible | 16:34 | |
| *** yaya has joined #openstack-ansible | 16:36 | |
| stevelle | automagically: were you going to look at https://review.openstack.org/#/c/290834 | 16:37 |
| *** saneax is now known as saneax_AFK | 16:37 | |
| automagically | I can, its been on my backburner for a bit | 16:39 |
| palendae | Has anyone else experience this? I can reproduce the behavior, but I have not been able to track down the cause https://bugs.launchpad.net/openstack-ansible/+bug/1540531 | 16:41 |
| openstack | Launchpad bug 1540531 in openstack-ansible trunk "Upgrade issues with 'Create log aggregation links'" [Medium,Confirmed] - Assigned to Nolan Brubaker (nolan-brubaker) | 16:41 |
| palendae | Additional eyes might help; I *think* an rsyslog variable value is being remembered somewhere, but I can't prove it | 16:42 |
| stevelle | palendae: what I experience is that is always marked as changed, and that bugs me | 16:42 |
| palendae | stevelle: That task? | 16:42 |
| stevelle | haven't looked for that issue though | 16:42 |
| stevelle | yes, the task | 16:42 |
| palendae | This seems to happen specifically when you teardown | 16:42 |
| palendae | At least, dumping the haproxy stuff | 16:43 |
| stevelle | hmm, so don't teardown? :) | 16:43 |
| palendae | Well, yes :) | 16:43 |
| palendae | Teardown does take that file with it though | 16:43 |
| stevelle | a good clue at least | 16:43 |
| palendae | Which confused me a lot the last I looked at this | 16:43 |
| palendae | Oh, it likely always says changed cause (in liberty, at least) it's set to force: "yes" | 16:44 |
| stevelle | automagically: we can figure out what to do with 290834 a bit later | 16:45 |
| *** mgoddard_ has quit IRC | 16:47 | |
| *** mgoddard has joined #openstack-ansible | 16:47 | |
| *** yaya has quit IRC | 16:48 | |
| *** jwagner is now known as jwagner_lunvh | 16:48 | |
| *** yaya has joined #openstack-ansible | 16:54 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible: Add group_vars for swift_remote_all hosts https://review.openstack.org/297257 | 16:54 |
| *** odyssey4me has quit IRC | 16:55 | |
| *** odyssey4me has joined #openstack-ansible | 16:55 | |
| *** oneswig has quit IRC | 16:58 | |
| spotz | gerrit be slow.... | 16:58 |
| *** openstackgerrit has quit IRC | 17:01 | |
| *** openstackgerrit has joined #openstack-ansible | 17:02 | |
| *** jorge_munoz has quit IRC | 17:02 | |
| *** KLevenstein has joined #openstack-ansible | 17:05 | |
| odyssey4me | ++ | 17:09 |
| *** pabelanger has quit IRC | 17:11 | |
| *** pabelanger has joined #openstack-ansible | 17:11 | |
| spotz | gerrit is back | 17:14 |
| *** d9k_ is now known as d9k | 17:15 | |
| spotz | andymccr: odyssey4me: I thought we were trying to eliminate calls to rpc-repo? | 17:16 |
| andymccr | spotz: happy to change it to anything really, those vars just need to be defined so i copied whatever was in the hosts.yml group_vars | 17:17 |
| *** eil397 has joined #openstack-ansible | 17:17 | |
| spotz | andymccr: The code looks good just want to make sure it's still good to refer to it. It had been removed from the docs already | 17:18 |
| odyssey4me | spotz trying to eliminate it as a *requirement* for now - we can do more clean up next cycle | 17:18 |
| odyssey4me | and yes, ideally no references in the docs | 17:19 |
| andymccr | spotz: honestly, i think it should be fixed in upstream ansible - you shouldn't need to specify vars that aren't required by a role just because they are expanded through filters, and you want to reference another hosts hostvars | 17:19 |
| andymccr | but that is the quickest fix for now, and since we'd need to move to ansible 2.1 and are unlikely to get it fixed in 1.9.4, it seems like the quickest path to success | 17:19 |
| spotz | andymccr: Has anyone bugged it there? | 17:19 |
| andymccr | spotz: im going to be testing it on 2.1 next week :) | 17:20 |
| spotz | Ok sweet, I'll go ahead and +1 the review in the meantime:) | 17:20 |
| andymccr | thanks! | 17:20 |
| *** asettle has joined #openstack-ansible | 17:22 | |
| *** jaypipes has quit IRC | 17:23 | |
| *** weezS has quit IRC | 17:25 | |
| *** asettle has quit IRC | 17:27 | |
| odyssey4me | andymccr hmm, those vars are already defined https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/hosts.yml#L29 / https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/hosts.yml#L38 | 17:28 |
| andymccr | odyssey4me: yeah but swift_remote_all are not in the hosts group, because they shouldn't have tasks targetted at hosts run against them | 17:28 |
| odyssey4me | ah, I see https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/swift-remote.yml | 17:29 |
| andymccr | yip! | 17:29 |
| *** eil397 has quit IRC | 17:30 | |
| andymccr | they shouldn't need those vars either, except that ansible expands the local hosts (which do have those vars) vars that are passed through filters (netloc/netorigin) and then bombs because the var doesnt exist. | 17:30 |
| odyssey4me | ok, technically repo_pip_default_index is not the right name anyway - so that's pretty weird | 17:36 |
| odyssey4me | it's supposed to be repo_build_pip_default_index | 17:36 |
| odyssey4me | but yeah, if it's causing a fail then I'll revisit that when I circle back to https://review.openstack.org/265867 | 17:37 |
| *** jiteka1 has joined #openstack-ansible | 17:37 | |
| andymccr | https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/hosts.yml#L29 | 17:38 |
| *** jiteka has quit IRC | 17:38 | |
| andymccr | its really not a used var, its just any var that gets filtered through netloc or netorigin. (https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/hosts.yml#L67-L69) | 17:38 |
| andymccr | it just happens to be those 2 vars only | 17:38 |
| odyssey4me | yep, that'll no longer be needed when https://review.openstack.org/296594 merges | 17:39 |
| andymccr | ok there is one other occurence | 17:39 |
| andymccr | but if thats going in thats awesome - but can that be backported? | 17:39 |
| andymccr | https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/hosts.yml#L43-L47 | 17:39 |
| odyssey4me | yes, I'm intending to as https://review.openstack.org/296594 solves an important issue when working behind proxies | 17:40 |
| andymccr | ok cool | 17:40 |
| odyssey4me | and yeah, I'm intending to fix that last ref in https://review.openstack.org/265867 as that's not right | 17:40 |
| *** pcaruana has quit IRC | 17:40 | |
| andymccr | ok sweet, as long as they aren't filtered through then we shouldn't need to do anything for the sync roles | 17:41 |
| andymccr | although somewhat interestingly, it may be best to avoid using var expansion inside group_vars, because if you reference a different host's hostvars it will take it's local expansion of the var and not the other host's. | 17:42 |
| odyssey4me | I think my brain exploded. :p | 17:42 |
| andymccr | which im not sure would be expected behaviour, for e.g. if a group_var was : "container_name2: "{{ container_name }}2" and you referenced hostvars[notmyhost]['container_name2'] you would get "your own container_name"2 back and not "other hosts container_name"2 | 17:43 |
| andymccr | at least in 1.9.4 | 17:43 |
| *** eil397 has joined #openstack-ansible | 17:44 | |
| andymccr | because it only performs the expansion based on your local vars. | 17:44 |
| andymccr | or so this seems. | 17:44 |
| andymccr | while you piece your brain back together - im heading out :) have fun everybody! | 17:45 |
| odyssey4me | have a great weekend andymccr ! | 17:45 |
| *** neilus has joined #openstack-ansible | 17:50 | |
| *** neilus has quit IRC | 17:55 | |
| *** chhavi has joined #openstack-ansible | 18:01 | |
| *** jorge_munoz has joined #openstack-ansible | 18:03 | |
| *** sdake_ has joined #openstack-ansible | 18:07 | |
| openstackgerrit | Rohan Parulekar proposed openstack/openstack-ansible-os_neutron: Nuage neutron plugin ansible changes https://review.openstack.org/296525 | 18:10 |
| *** sdake has quit IRC | 18:10 | |
| odyssey4me | mattt sigmavirus24 any progress on confirming whether we've hit a bug? | 18:10 |
| *** neilus has joined #openstack-ansible | 18:11 | |
| sigmavirus24 | odyssey4me: chatting in #openstack-glance with jokke about it | 18:11 |
| odyssey4me | sigmavirus24 great, thanks - I'll lurk there | 18:12 |
| *** weezS has joined #openstack-ansible | 18:13 | |
| *** yaya has quit IRC | 18:15 | |
| *** Bjoern is now known as Bjoern_zZzZzZzZ | 18:18 | |
| prometheanfire | odyssey4me: you have a link for the neutron fix in osa liberty? | 18:26 |
| prometheanfire | the sha bump | 18:26 |
| *** tiagogomes_ has quit IRC | 18:29 | |
| *** jiteka1 has quit IRC | 18:32 | |
| logan- | this one? https://review.openstack.org/#/c/291810/ | 18:36 |
| prometheanfire | ya, so we can include that | 18:37 |
| prometheanfire | don't know when the fast tags/sha bumps are suposed to happen | 18:37 |
| prometheanfire | just know that we agreed that they would :D | 18:37 |
| *** pjm_ has quit IRC | 18:40 | |
| *** admin0 has joined #openstack-ansible | 18:41 | |
| *** Bjoern_zZzZzZzZ is now known as Bjoern | 18:42 | |
| *** asettle has joined #openstack-ansible | 18:43 | |
| *** jwagner_lunvh is now known as jwagner | 18:48 | |
| *** asettle has quit IRC | 18:48 | |
| odyssey4me | prometheanfire you're asking about the tags for OSA? | 18:49 |
| odyssey4me | FYI - the release for Kilo is in the queue: https://review.openstack.org/297274 | 18:49 |
| odyssey4me | I'll submit SHA bumps shortly. | 18:49 |
| prometheanfire | ah, cool | 18:51 |
| admin0 | any reason why this comes : SSL exception connecting to https://domain.com:5000/v3/auth/tokens: [SSL: UNKNOWN_PROTOCOL] unknown protocol (_ssl.c:590) | 19:03 |
| *** hybridpollo has joined #openstack-ansible | 19:07 | |
| stevelle | cloudnull: on https://review.openstack.org/#/c/290834/ is there a diff between host / port and api_servers such that we can't just specify a default value for api_servers to get rid of the deprecated configs? | 19:11 |
| stevelle | the comment makes it seem like there is something I don't understand there | 19:12 |
| automagically | Leaving the case statement in place seems to resolve the issue | 19:13 |
| cloudnull | no difference other than there can be more than one. | 19:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on 12.0.10 https://review.openstack.org/297318 | 19:13 |
| automagically | I do like cloudnull ’s suggestion of adopting a nova_glance_api_servers var to properly namespace, but that can be a follow-on change | 19:14 |
| odyssey4me | admin0 more than likely trying to SSL to a non SSL service, or trying to HTTP to an SSL service | 19:14 |
| openstackgerrit | Hector I Gonzalez Mendoza proposed openstack/openstack-ansible-os_glance: Updated role using Multi-Distro framework https://review.openstack.org/297320 | 19:16 |
| odyssey4me | d34dh0r53 prometheanfire Please verify that https://review.openstack.org/297318 includes the SHA's you need. | 19:16 |
| prometheanfire | yep, already showed him :P | 19:17 |
| stevelle | agreed automagically, best to add a follow-on for it though I think that should happen too | 19:18 |
| stevelle | I'll try to refresh that review today | 19:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Updates all repo SHAs to open up work on 11.2.13 https://review.openstack.org/297325 | 19:22 |
| prometheanfire | odyssey4me: thanks | 19:24 |
| admin0 | odyssey4me: the UR~L is in the rc file .. | 19:27 |
| admin0 | its https: | 19:27 |
| admin0 | trying to use the openstack api | 19:27 |
| logan- | can you curl https://url | 19:28 |
| admin0 | curl: (35) Unknown SSL protocol error in connection to | 19:29 |
| logan- | what about curl http://url | 19:29 |
| admin0 | http works fine | 19:29 |
| *** mgoddard has quit IRC | 19:29 | |
| *** zhangjn has quit IRC | 19:31 | |
| odyssey4me | lol, you can't serve two protocols from the same port | 19:31 |
| odyssey4me | so clearly you've misconfigured something | 19:32 |
| *** zhangjn has joined #openstack-ansible | 19:32 | |
| admin0 | :D | 19:32 |
| admin0 | hmm | 19:32 |
| odyssey4me | openrc normally has the internal endpoint URL, but somehow you appear to have what I assume is your public URL in openrc? | 19:33 |
| jthorne | it'll have the publicURL if you downloaded the openrc from Horizon | 19:33 |
| odyssey4me | or you changed up the config, but didn't run all the plays? | 19:33 |
| odyssey4me | ^yep | 19:33 |
| admin0 | https://gist.github.com/a1git/8456c8a8d7bfbd57644d | 19:38 |
| admin0 | that is what I have so far .. | 19:38 |
| admin0 | what playbook might i have missed | 19:38 |
| admin0 | horizon is working fine .. same ssl | 19:38 |
| admin0 | internal stuff works | 19:38 |
| *** jmccrory_ has quit IRC | 19:38 | |
| admin0 | just public is not | 19:38 |
| *** asettle has joined #openstack-ansible | 19:40 | |
| admin0 | does haproxy also read : openstack_service_publicuri_proto: https ? | 19:42 |
| *** asettle has quit IRC | 19:45 | |
| jwitko | hey guys is it possible to define multiple subnets for use in "br-mgmt" | 19:45 |
| odyssey4me | horizon doesn't, by default, use the public endpoint | 19:45 |
| odyssey4me | admin0 it seems likely that whatever you're using to do ssl offloading isn't correctly configured | 19:46 |
| admin0 | its default haproxy that ansible installs | 19:46 |
| admin0 | or let me rephrase | 19:46 |
| odyssey4me | admin0 note that if you're expecting haproxy to do it, I think it needs more config than just saying you want https on your endpoint | 19:46 |
| admin0 | oh | 19:46 |
| admin0 | :D | 19:46 |
| odyssey4me | I think you need to set haproxy to do it. | 19:47 |
| odyssey4me | The option caters for other LB's to handle it. | 19:47 |
| admin0 | so how to tell haproxy .. “hey .. keystone is set for https:” ? | 19:47 |
| mhayden | cloudnull: i might have to hug you for https://review.openstack.org/#/c/296840/ | 19:48 |
| odyssey4me | admin0 add 'haproxy_ssl: true' to user_variables.yml | 19:48 |
| admin0 | doing that :) | 19:48 |
| cloudnull | mhayden: so should i abandon it? | 19:48 |
| cloudnull | ;) | 19:48 |
| spotz | heheh. Bugs cloudnull he has stickers at his desk | 19:49 |
| odyssey4me | admin0 and keystone_service_publicuri_proto: "https" | 19:49 |
| spotz | Bugs=But, not sure about that typo:) | 19:49 |
| odyssey4me | that should be all | 19:49 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible-repo_build: Updated repo_build process to be faster https://review.openstack.org/296840 | 19:50 |
| cloudnull | automagically: ^ updated per your feedback | 19:50 |
| mhayden | #makerepobuildgreatagain | 19:50 |
| cloudnull | automagically: did the the shell module work behind the proxy? | 19:51 |
| cloudnull | just to confirm | 19:51 |
| cloudnull | i updated the script executions to all use shell based on my understand of that. | 19:51 |
| cloudnull | mhayden: hahahaha | 19:51 |
| odyssey4me | d34dh0r53 prometheanfire FYI the tags will likely only get processed tomorrow in the US daytime | 19:52 |
| d34dh0r53 | odyssey4me: tys | 19:52 |
| odyssey4me | I've submitted the requests, so they're just waiting for formalised processing. | 19:53 |
| automagically | cloudnull: Yes, because of https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/openstack_proxy_settings.yml | 19:53 |
| prometheanfire | odyssey4me: are we tagging after the sha bump? | 19:53 |
| odyssey4me | meanwhile the SHA bumps can go ahead as soon as they gate - the tags will be at the current SHA's | 19:53 |
| cloudnull | automagically: cool | 19:53 |
| admin0 | ran haprpxy playbook .. running openstack playbook again | 19:53 |
| prometheanfire | or before (for both kilo and liberty) | 19:53 |
| openstackgerrit | Hector I Gonzalez Mendoza proposed openstack/openstack-ansible-os_glance: Updated role using Multi-Distro framework https://review.openstack.org/297320 | 19:54 |
| odyssey4me | prometheanfire tag before the OpenStack SHA bump - but the tags will be done at the requested OSA SHA's | 19:54 |
| odyssey4me | prometheanfire but the SHA bump patches I've submitted can proceed regardless | 19:55 |
| prometheanfire | ok | 19:55 |
| prometheanfire | don't think we care about the kilo sha, just the tag | 19:55 |
| prometheanfire | and we don't use liberty tags, just the shas | 19:55 |
| prometheanfire | fun distiction | 19:55 |
| odyssey4me | sure | 19:56 |
| admin0 | odyssey4me: so seeing https:// on all public and being happy that “hey its https” is not true :) | 20:02 |
| odyssey4me | admin0 yep, it only actually works if something is handling https | 20:03 |
| prometheanfire | wooo | 20:04 |
| prometheanfire | put nginx in front of it | 20:05 |
| prometheanfire | :D | 20:05 |
| prometheanfire | made a couple of blog posts about it, but it's a one-off setup type thing (for home) | 20:05 |
| admin0 | odyssey4me: https://domain.com:5000/v3 is in the public endpint . when I do that url, it throws me back to links": [{"href": "http://domain.com:5000/v3/", | 20:06 |
| admin0 | so why add the endpoints as https:// in the first place and have then return http instead of http | 20:06 |
| *** johnmilton has quit IRC | 20:06 | |
| automagically | admin0: That’s due to a missing X-Forwarded-Proto header | 20:06 |
| automagically | The app isn’t seeing SSL traffic, its seeing HTTP since the SSL is terminated at HAProxy | 20:07 |
| automagically | Therefore it thinks it should produce links to the HTTP version of itself | 20:07 |
| *** Bjoern has quit IRC | 20:07 | |
| admin0 | http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-haproxy.html — this page needs a lot of work :D | 20:07 |
| *** sdake has joined #openstack-ansible | 20:07 | |
| admin0 | i can do that if i can make this work now | 20:07 |
| *** sdake_ has quit IRC | 20:07 | |
| admin0 | can X-Forwarded-Proto header be also added/overwritten on the user_variables ? | 20:08 |
| automagically | admin0: This should show you the pattern: https://review.openstack.org/277199 | 20:08 |
| admin0 | i cannot +1 vote right :D | 20:09 |
| *** sdake_ has joined #openstack-ansible | 20:10 | |
| *** KLevenstein has quit IRC | 20:12 | |
| *** sdake has quit IRC | 20:12 | |
| admin0 | hmm.. so now for me to fix this, i need to check somehow to add reqadd X-Forwarded-Proto:\ https ? | 20:14 |
| admin0 | isnt there smoething that can be instered into “ haproxy_ssl: "{% if haproxy_ssl | bool and keystone_service_internaluri_proto == 'https' %}true{% else %}false{% endif %}”” ? | 20:16 |
| *** johnmilton has joined #openstack-ansible | 20:17 | |
| *** yaya has joined #openstack-ansible | 20:18 | |
| *** rETROpunK has quit IRC | 20:18 | |
| *** rETROpunK has joined #openstack-ansible | 20:18 | |
| *** admin0 has left #openstack-ansible | 20:19 | |
| odyssey4me | hmm, I thought we did have a header insertion there | 20:19 |
| odyssey4me | hang on | 20:19 |
| *** rohanp_ has quit IRC | 20:20 | |
| odyssey4me | it is already: https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/haproxy_server/templates/service.j2#L18 | 20:21 |
| odyssey4me | and admin0 you can vote +1 | 20:22 |
| odyssey4me | anyone can vote +1/-1 | 20:22 |
| *** admin0 has joined #openstack-ansible | 20:23 | |
| *** chhavi has quit IRC | 20:24 | |
| odyssey4me | and that gets activated based on the public proto: https://github.com/openstack/openstack-ansible/blob/master/playbooks/vars/configs/haproxy_config.yml#L92 | 20:25 |
| jwitko | hey guys is it possible to define multiple subnets for use in the container/management ip range? like multiple cidr's in openstack_user_config.yaml ? | 20:25 |
| odyssey4me | admin0 are your internal_lb_address and external_lb_address the same address? | 20:25 |
| odyssey4me | jwitko it's a bit of a hack, but essentially you define a CIDR that covers them both, then set the ranges you don't want containers to have to get 'used' | 20:26 |
| jwitko | odyssey4me, thanks. | 20:27 |
| jwitko | odyssey4me, wouldn't that mean they all had to share the same gateway then? | 20:27 |
| jwitko | and netmask ? | 20:27 |
| *** johnmilton has quit IRC | 20:28 | |
| odyssey4me | jwitko hmm, I never thought that far personally - I immediately found a bucket to get ill into | 20:28 |
| odyssey4me | that's something we need to resolve in Newton in a far better way | 20:29 |
| odyssey4me | maybe cloudnull has a better hack? | 20:29 |
| *** sdake has joined #openstack-ansible | 20:30 | |
| *** sdake_ has quit IRC | 20:33 | |
| cloudnull | i've got nothing | 20:33 |
| admin0 | odyssey4me: they are different | 20:34 |
| *** pjm6 has joined #openstack-ansible | 20:38 | |
| admin0 | lb_internal is cloud101int.domain.com lb_external is cloud101.domain.com | 20:39 |
| *** jduhamel has joined #openstack-ansible | 20:43 | |
| *** jduhamel has quit IRC | 20:43 | |
| *** Mudpuppy has quit IRC | 20:44 | |
| odyssey4me | admin0 and on the haproxy nodes, in the keystone_service config, do you see the X-Forwarded-Proto config there? | 20:46 |
| odyssey4me | jwitko we don't do the network config on the hosts, so there is that - and you can inject routes into the containers, so there's that | 20:46 |
| odyssey4me | but generally speaking having multiple CIDRs for your management network as a story sucks right now | 20:47 |
| *** openstackgerrit has quit IRC | 20:48 | |
| admin0 | i see for keystone_service: reqadd X-Forwarded-Proto:\ https | 20:48 |
| admin0 | hmm | 20:48 |
| admin0 | its there | 20:48 |
| *** woodard_ has joined #openstack-ansible | 20:49 | |
| *** openstackgerrit has joined #openstack-ansible | 20:49 | |
| odyssey4me | admin0 ok, so when you curl to the endpoint the info it gives back is the http location, right? | 20:52 |
| *** woodard has quit IRC | 20:52 | |
| admin0 | yep | 20:52 |
| *** woodard_ has quit IRC | 20:53 | |
| odyssey4me | admin0 as I recall, you'll also have to set https://github.com/openstack/openstack-ansible-os_keystone/blob/master/templates/keystone.conf.j2#L6-L8 | 20:54 |
| odyssey4me | keystone_public_endpoint | 20:54 |
| odyssey4me | that's the one that shows up in the API response | 20:54 |
| odyssey4me | although logan- yourself and cloudnull did some digging and recently added https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/os_keystone/templates/keystone.conf.j2#L14-L16 I think? | 20:55 |
| odyssey4me | heh, no - that was me | 20:56 |
| *** skamithi has left #openstack-ansible | 20:58 | |
| odyssey4me | that might work instead, actually - set https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/os_keystone/defaults/main.yml#L105-L106 | 20:58 |
| odyssey4me | and set 'keystone_ssl: true' | 20:59 |
| odyssey4me | that combo should achieve the goal too | 21:00 |
| admin0 | trying :) | 21:00 |
| odyssey4me | I think that keystone_public_endpoint is specifically to advertise something like a name instead of an IP, and a different protocol - often used when a proxy or something is in front of keystone and you want to obscure it | 21:01 |
| *** sdake has quit IRC | 21:01 | |
| odyssey4me | whereas the configuration of the expected header for traffic coming from an LB doing SSL offloading will provide the same name, but acknowledge the protocol difference | 21:02 |
| admin0 | so i enabled keystone_secure_proxy_ssl_header: X-Forwarded-For .. i just need to run the keystone playbook right ? | 21:02 |
| openstackgerrit | Pedro Magalhães proposed openstack/openstack-ansible: Joining documentation of network services in one file https://review.openstack.org/297179 | 21:02 |
| odyssey4me | admin0 and 'keystone_ssl: true' | 21:02 |
| admin0 | that i already have | 21:02 |
| odyssey4me | ok, then keystone playbook - yeah | 21:03 |
| *** rETROpunK has quit IRC | 21:06 | |
| *** rETROpunK has joined #openstack-ansible | 21:06 | |
| *** mrda has quit IRC | 21:07 | |
| *** mrda has joined #openstack-ansible | 21:07 | |
| admin0 | do i also need to do setup-openstack ? | 21:09 |
| *** johnmilton has joined #openstack-ansible | 21:10 | |
| *** spotz is now known as spotz_zzz | 21:11 | |
| *** thorst has quit IRC | 21:12 | |
| admin0 | hmm.. still returns http .. what i have so far is enabled keystone_secure_proxy_ssl_header: X-Forwarded-For and slo set keystone_ssl: true on the main.yml inside keystone, keystone playbook re-run . openstack_service_publicuri_proto: https is set .. i also see keystone_service: reqadd X-Forwarded-Proto:\ https | 21:13 |
| *** fawadkhaliq has joined #openstack-ansible | 21:13 | |
| *** retreved has quit IRC | 21:14 | |
| *** thorst has joined #openstack-ansible | 21:14 | |
| automagically | See you all tomorrow | 21:15 |
| *** rromans has quit IRC | 21:17 | |
| odyssey4me | admin0 you shouldn't be changing things in the code - everything must be set in user_variables | 21:18 |
| odyssey4me | otherwise other things will override them | 21:18 |
| *** thorst has quit IRC | 21:18 | |
| odyssey4me | cheers automagically | 21:19 |
| *** retreved has joined #openstack-ansible | 21:19 | |
| openstackgerrit | Robb Romans proposed openstack/openstack-ansible: Fix doc build warnings https://review.openstack.org/297370 | 21:28 |
| *** rromans has joined #openstack-ansible | 21:28 | |
| mattt | odyssey4me: was it identified as a bug then ? | 21:28 |
| cloudnull | anyone want to punch this through https://review.openstack.org/#/c/296840/ ? | 21:31 |
| rromans | punch it! | 21:32 |
| *** asettle has joined #openstack-ansible | 21:33 | |
| *** homerp_ has quit IRC | 21:34 | |
| *** weshay has quit IRC | 21:35 | |
| *** zhangjn has quit IRC | 21:37 | |
| admin0 | well odyssey4me .. i am kind of lost :) | 21:38 |
| *** homerp has joined #openstack-ansible | 21:39 | |
| *** openstack has joined #openstack-ansible | 21:51 | |
| *** fawadkhaliq has joined #openstack-ansible | 21:52 | |
| *** ametts has quit IRC | 21:57 | |
| *** kysse has joined #openstack-ansible | 22:00 | |
| *** admiralboom has joined #openstack-ansible | 22:00 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Updated repo_build process to be faster https://review.openstack.org/296840 | 22:00 |
| stevelle | woot | 22:04 |
| *** johnmilton has quit IRC | 22:08 | |
| *** thorst has joined #openstack-ansible | 22:11 | |
| *** fawadkhaliq has quit IRC | 22:11 | |
| *** fawadkhaliq has joined #openstack-ansible | 22:12 | |
| *** asettle has quit IRC | 22:12 | |
| *** thorst has quit IRC | 22:16 | |
| *** yaya has quit IRC | 22:16 | |
| *** thorst has joined #openstack-ansible | 22:16 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 22:18 |
| cloudnull | awesome jmccrory | 22:19 |
| *** thorst has quit IRC | 22:21 | |
| *** b3rnard0 is now known as b3rnard0_away | 22:23 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:31 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Add rootwraps, implement config_template https://review.openstack.org/297398 | 22:43 |
| *** jthorne has quit IRC | 22:45 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 22:54 |
| *** johnmilton has joined #openstack-ansible | 22:56 | |
| openstackgerrit | Merged openstack/openstack-ansible: Updates all repo SHAs to open up work on 12.0.10 https://review.openstack.org/297318 | 22:57 |
| *** sdake has joined #openstack-ansible | 22:58 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 22:59 |
| *** cfarquhar has quit IRC | 23:00 | |
| *** saneax_AFK is now known as saneax | 23:00 | |
| *** cfarquhar has joined #openstack-ansible | 23:00 | |
| *** cfarquhar has quit IRC | 23:00 | |
| *** cfarquhar has joined #openstack-ansible | 23:00 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:02 |
| *** thorst has joined #openstack-ansible | 23:02 | |
| *** thorst has quit IRC | 23:05 | |
| *** thorst has joined #openstack-ansible | 23:05 | |
| *** fawadkhaliq has quit IRC | 23:06 | |
| *** fawadkhaliq has joined #openstack-ansible | 23:07 | |
| *** pjm6 has quit IRC | 23:10 | |
| *** thorst has quit IRC | 23:10 | |
| *** weezS has quit IRC | 23:10 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:11 |
| *** thorst has joined #openstack-ansible | 23:14 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:16 |
| *** thorst_ has joined #openstack-ansible | 23:17 | |
| *** sdake_ has joined #openstack-ansible | 23:17 | |
| *** sdake has quit IRC | 23:18 | |
| *** thorst has quit IRC | 23:19 | |
| *** thorst has joined #openstack-ansible | 23:21 | |
| *** thorst_ has quit IRC | 23:22 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:23 |
| *** admin0 has quit IRC | 23:26 | |
| *** admin0 has joined #openstack-ansible | 23:26 | |
| *** admin0 has quit IRC | 23:26 | |
| jmccrory | odyssey4me would it be possible to rename the repo to os_ironic to be in line with the other IRRs? | 23:28 |
| *** sdake has joined #openstack-ansible | 23:28 | |
| odyssey4me | jmccrory apparently it's an expensive exercise - they have to take gerrit down to do it | 23:28 |
| *** sdake_ has quit IRC | 23:28 | |
| jmccrory | ah wow | 23:29 |
| odyssey4me | so it'll take somewhere around amonth to do it | 23:29 |
| *** jwitko has quit IRC | 23:30 | |
| odyssey4me | unfortunately it seems like the change to sudoers has done nothing to resolve the issues we were hoping for it to solve | 23:30 |
| jmccrory | ocd's been gnawing at me seeing it come through in chat the last week or two, can probably hold out a month though | 23:31 |
| odyssey4me | we'll have to revisit that next week | 23:31 |
| odyssey4me | yeah, I'd like to go through with it before we integrate it - I'll re-discuss it next week | 23:31 |
| odyssey4me | as it is, our repo-build process assumes that it only builds venvs for roles starting with os_ | 23:32 |
| odyssey4me | haha! it works! | 23:40 |
| odyssey4me | jmccrory if you have a gap, and the inclination, please take a look through that patch | 23:41 |
| jmccrory | sure, taking a look | 23:41 |
| odyssey4me | jmccrory automagically cloudnull d34dh0r53 stevelle mattt hughsaunders andymccr https://review.openstack.org/296077 is passing a convergence test - it'd be great to get some reviews so that the Aus crew (mcarden, neillc, mrda) can get on with implementing some functional tests and work on making it more functional | 23:42 |
| odyssey4me | aha, the requirements repo has cut a stable branch | 23:47 |
| *** elopez has joined #openstack-ansible | 23:49 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:50 |
| *** jcrst has joined #openstack-ansible | 23:54 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ironic: Implement initial test inventory/plays https://review.openstack.org/296077 | 23:56 |
| odyssey4me | jmccrory just added the switch to using stable/mitaka in defaults and tests | 23:56 |
| jmccrory | odyssey4me probably wouldn't matter since this is just convergence test, but should ironic_standalone_auth_strategy be overriden to keystone? | 23:58 |
| jmccrory | defaults to noauth | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!