Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-07-10

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:27 aimon joined #puppet-openstack
00:34 mgagne I think we got all this ml2 symlink thing wrong...
00:34 mgagne more news tomorrow
00:41 gildub joined #puppet-openstack
01:00 xarses joined #puppet-openstack
01:41 xingchao joined #puppet-openstack
02:43 Madkiss joined #puppet-openstack
02:43 Madkiss joined #puppet-openstack
02:57 mmaglana joined #puppet-openstack
03:35 mmaglana joined #puppet-openstack
04:00 sputnik1_ joined #puppet-openstack
04:17 sputnik1_ joined #puppet-openstack
05:35 dgollub joined #puppet-openstack
05:55 sputnik1_ joined #puppet-openstack
06:11 EmilienM good morning
06:26 mkoderer joined #puppet-openstack
06:30 xarses EmilienM: morning
06:50 ddmitriev joined #puppet-openstack
07:02 daneyon joined #puppet-openstack
07:12 sputnik1_ joined #puppet-openstack
07:32 aimon joined #puppet-openstack
07:42 qba73 joined #puppet-openstack
07:58 aldavud joined #puppet-openstack
08:15 derekh_ joined #puppet-openstack
08:26 saju_m joined #puppet-openstack
09:09 markvoelker joined #puppet-openstack
09:15 arnaud_orange joined #puppet-openstack
09:49 mmagr joined #puppet-openstack
10:13 fc__ joined #puppet-openstack
10:50 ddmitriev joined #puppet-openstack
10:59 berendt joined #puppet-openstack
11:09 gcha mgagne, EmilienM, fvollero: Hi guys, could you please have a look at those reviews ? https://review.openstack.org/103817, https://review.openstack.org/104550, https://review.openstack.org/94504, https://review.openstack.org/103819
11:09 gcha https://review.openstack.org/105685, https://review.openstack.org/105807 and https://review.openstack.org/105809
11:37 EmilienM wow
11:37 EmilienM like a lot
11:40 gcha for the most, there is already a +2 :-)
12:12 Dafna joined #puppet-openstack
12:17 daneyon joined #puppet-openstack
12:19 daneyon joined #puppet-openstack
12:20 gcha sbadia: thanks ;-)
12:23 sbadia no problems ;)
12:23 dprince joined #puppet-openstack
12:34 gcha sbadia: oops, git.openstack.org Temporary failure in name resolution ???
12:36 sbadia huhu :o
12:36 sbadia hum, for this patch https://review.openstack.org/#/c/103817/
12:37 gcha nope this one 106035
12:37 gcha sbadia: I know I know ... :)
12:37 sbadia on debian/ubuntu, target service is provided by lio-utils and not targetcli (but targetcli is a dep of lio-utils)
12:38 sbadia https://packages.debian.org​/sid/all/targetcli/filelist and https://packages.debian.org/​sid/amd64/lio-utils/filelist
12:41 gcha social: ^^^^
12:45 social sbadia: just replied there
12:46 sbadia ok then :) thanks for the quick answer
12:46 social sbadia: so if I understood it correctly it's similar to rhel where we install targetcli which depends on python-rtslib so when targetcli package gets installed the target server from python-rtslib also gets installed
12:46 social sbadia: othervise it's wrong and I have to change the commit
12:47 sbadia yep it's similar (but on debian/ubuntu python-rtslib is only a lib)
12:48 sbadia no, it's ok I think
12:48 social thanks \o/
12:59 tfhartmann joined #puppet-openstack
13:04 nigelk joined #puppet-openstack
13:11 morazi joined #puppet-openstack
13:18 arnaud_orange left #puppet-openstack
13:25 arnaud_orange joined #puppet-openstack
13:38 tfz joined #puppet-openstack
13:48 tfz joined #puppet-openstack
13:53 prad joined #puppet-openstack
14:11 tfz joined #puppet-openstack
14:42 marun joined #puppet-openstack
14:49 mattymo joined #puppet-openstack
14:58 JayJ joined #puppet-openstack
15:27 xarses joined #puppet-openstack
15:28 spredzy Hello guys, I just made a review to have a finer grain about logging options (https://review.openstack.org/#/c/106055/) after discussing with sbadia, we though that it was a perfect use case for openstacklib, however, it was not clear how to handle the different {nova,neutron,cinder,etc...}_config resources
15:29 spredzy the idea would be to create an openstacklib::logging class that would be imported on every module, and configure everything related to modules
15:29 spredzy logging I meant
15:31 EmilienM spredzy: I +2 this idea so far
15:32 EmilienM since the logging code is in Oslo
15:32 openstackgerrit joined #puppet-openstack
16:15 hogepodge joined #puppet-openstack
16:16 hogepodge good morning!
16:17 mgagne hogepodge: good morning
16:18 mgagne hogepodge: I think we got symlink wrong with ml2, I'll test soon my hypothesis
16:18 hogepodge :-/
16:18 mgagne yep, I think it's a misunderstanding of how config files are organized
16:20 aimon joined #puppet-openstack
16:21 JayJ hogepodge: mgagne: Good Morning
16:22 hogepodge HI JayJ
16:23 JayJ hogepodge: I checkedout the latest sources from git. built local cluster. Otherthan a error in Ceilometer it ran succesfully. Will do some testing today
16:23 anonitus joined #puppet-openstack
16:23 hogepodge Good to hear. Thanks
16:24 anonitus left #puppet-openstack
16:24 anonitus joined #puppet-openstack
16:56 aldavud joined #puppet-openstack
16:59 imcsk8 joined #puppet-openstack
17:09 mgagne hogepodge: https://review.openstack.org/#/c/106035/
17:10 daneyon joined #puppet-openstack
17:18 hogepodge Ugh, ok.
17:27 daneyon joined #puppet-openstack
17:30 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: Add db::mysql and db::mysql::host_access to openstacklib  https://review.openstack.org/105280
17:30 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-neutron: Add kombu_reconnect_delay  https://review.openstack.org/106116
17:38 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-neutron: Add rpc_workers for neutron-server  https://review.openstack.org/106117
17:39 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-neutron: Add rpc_workers for neutron-server  https://review.openstack.org/106117
17:41 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: Adapt tests to work with recently changed inifile  https://review.openstack.org/106118
17:45 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: Client profile for roles/profiles pattern  https://review.openstack.org/100522
17:47 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: Adapt tests to work with recently changed inifile  https://review.openstack.org/106118
17:48 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: Client profile for roles/profiles pattern  https://review.openstack.org/100522
17:48 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: OSD profile for roles/profiles pattern  https://review.openstack.org/101575
17:49 aldavud Sorry for the spam ^^^
18:19 openstackgerrit A change was merged to stackforge/puppet-designate: Normalize Gemfile & pin tests to rspec < 2.99  https://review.openstack.org/101491
18:25 xarses aldavud_away: can you look over https://review.openstack.org/#/c/102574/ and https://review.openstack.org/#/c/100480/ and push fixes for them if possible?
18:26 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: Add db::mysql and db::mysql::host_access to openstacklib  https://review.openstack.org/105280
18:37 mgagne hogepodge: alright, I confirmed my hypothesis
18:39 mgagne hogepodge: we do NOT need to symlink /etc/neutron/plugins/openvsw​itch/ovs_neutron_plugin.ini to /etc/neutron/plugin.ini or /etc/neutron/plugins/ml2/ml2_conf.ini
18:39 mgagne hogepodge: we should stop creating or managing such symlinks
18:40 mgagne hogepodge: ovs_neutron_plugin.ini only contains configs used by the openvswitch mechanism (the agent). It is not needed/read by the ML2 plugin on the node running neutron-server
18:40 tfz1 joined #puppet-openstack
18:41 mgagne EmilienM: ^
18:41 mgagne hogepodge: the only symlink we have to still create and manage is /etc/neutron/plugin.ini
18:41 mgagne hogepodge: which should point to ml2_conf.ini
18:50 tfz joined #puppet-openstack
19:04 EmilienM mgagne: neutron-server start with neutron.conf & plugin file
19:05 tfz joined #puppet-openstack
19:06 mgagne EmilienM: ok?
19:06 EmilienM mgagne: sorry I just catch up
19:07 mgagne EmilienM: I'm testing on my side and I'm able to run without the symlink to ml2_conf.ini (from the ovs plugin conf file)
19:07 EmilienM mgagne: which distro?
19:07 mgagne EmilienM: Ubuntu 12.04
19:07 mgagne EmilienM: should be the same for redhat
19:07 mgagne EmilienM: I'm testing vxlan support with linuxbridge and ML2 atm
19:08 mgagne EmilienM: juste need a decent kernel for vxlan support
19:08 EmilienM there is a backport package
19:09 EmilienM mgagne: so, stop me if I'm wrong
19:09 EmilienM mgagne: you say it's useless to symlink /etc/neutron/plugins/openvsw​itch/ovs_neutron_plugin.ini to plugin.ini
19:09 EmilienM the symlink is only for ml2. right?
19:13 mgagne EmilienM: from plugin.ini to ml2, nothing else
19:13 mgagne EmilienM: ovs_neutron_plugin.ini is only used by the agent, not ml2
19:13 EmilienM mgagne: I agree
19:13 mgagne EmilienM: it "works" because it happens the file end up with the appropriate section/config
19:13 EmilienM mgagne: that's why I did the ml2 ovs agent class
19:14 EmilienM neutron::agent::ml2::ovs
19:14 EmilienM to manage the ml2 config file
19:14 EmilienM and not using the OVS config file
19:14 mgagne EmilienM: on the netowrk node only, not neutron-server
19:14 EmilienM personally, i don't use it.
19:14 mgagne EmilienM: no, you don't
19:14 EmilienM and I don't want to
19:14 EmilienM I'm using ML2, I keep ML2 config file only.
19:14 tfz joined #puppet-openstack
19:14 mgagne EmilienM: you should use neutron_plugin_ovs in the agent manifest
19:15 mgagne EmilienM: that's not how packaging and documentation tells you to do it
19:15 EmilienM and I did the symlink stuff for the db_sync
19:15 mgagne EmilienM: dbsync only needs plugin.ini to ml2_conf.ini
19:18 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-neutron: Add more retry logic to neutron resource auth_neutron call  https://review.openstack.org/106141
19:18 EmilienM mgagne: so the agent would load 3 files?
19:18 mgagne EmilienM: the agent doesn't need to load ml2_conf.ini
19:18 EmilienM when using ML2
19:18 mgagne EmilienM: it works without it
19:18 EmilienM mgagne: really?
19:18 mgagne EmilienM: yep
19:19 EmilienM it's because I put the flag I need in ml2 conf file
19:19 mgagne EmilienM: ml2_conf.ini only has configs used by the PLUGIN on neutron-server, not the agent
19:19 EmilienM that's why it works for me
19:19 mgagne EmilienM: it's because you merged everything in one file
19:19 EmilienM mgagne: yes
19:19 mgagne EmilienM: it works because one way or an other, the sections/configs get loaded
19:19 EmilienM yep
19:20 mgagne EmilienM: but that's not how neutron config files and packaging is done
19:20 mgagne EmilienM: so we should respond the spirit of how neutron and package are done
19:20 openstackgerrit Gael Chamoulaud proposed a change to stackforge/puppet-neutron: Uses neutron_plugin_ovs type to setup OVS Neutron Agent  https://review.openstack.org/106144
19:22 mgagne EmilienM: and you end up with ovs sections in ml2_conf.ini which is wrong, ml2 doesn't care about the actual mechanism
19:22 mgagne EmilienM: that's why everything is split up in multiple files
19:24 EmilienM mgagne: maybe, but to be honest, neutron config was/is such a mess... i'm running this config on my setups and everything works fine. However, i'm very open to improve it
19:25 mgagne EmilienM: I'm adding support for linuxbridge with ML2, I read the doc and that's what it told me to do and it works too
19:25 mgagne EmilienM: and looks more sane
19:26 EmilienM mgagne: I did that long time ago, in November if I recall this
19:26 openstackgerrit Ivan Chavero proposed a change to stackforge/puppet-nova: Install ceph client libraries when using rbd  https://review.openstack.org/105649
19:26 EmilienM anyway
19:26 EmilienM you look confident so let's go with better design !
19:27 mgagne EmilienM: and I don't need neutron::plugins::ml2 on the agent =)
19:27 EmilienM cool
19:28 tfz joined #puppet-openstack
19:35 mgagne gcha: that didn't took too long lol
19:35 mgagne gcha: I had those changes in a branch too
19:37 gcha mgagne: oops ;-)
19:37 mgagne gcha: local branch :P
19:38 mgagne gcha: I would add much more context in the commit message though
19:38 EmilienM gcha: hey, maybe you could improve the commit msg
19:38 EmilienM oops sorry mgagne already mentioned it
19:39 gcha mgagne: yes ! and I have on failure in one test that I have to fix too
19:39 gcha s/on/one/
19:40 gcha mgagne, EmilienM: and I have to juggle with a water leak in my bathroom ... in the same time :/
19:41 mgagne gcha: what me to take over the commit?
19:41 gcha mgagne: feel free !
19:46 aldavud joined #puppet-openstack
19:51 EmilienM gcha: good luck
19:54 aldavud dachary: dmsimard: a recent change in ini_file is breaking all gate jobs. one more core review is needed for the fix to go through. can you have a look https://review.openstack.org/#/c/106118/
19:55 dachary aldavud: you have it ! :-)
19:56 openstackgerrit A change was merged to stackforge/puppet-ceph: Adapt tests to work with recently changed inifile  https://review.openstack.org/106118
19:57 mgagne aldavud: so they released a dot version and broke backward compat... good job
19:59 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Configure OVS mechanism agent configs in its config file  https://review.openstack.org/106144
20:00 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Configure OVS mechanism agent configs in its config file  https://review.openstack.org/106144
20:03 aldavud dachary: great thanks :)
20:04 aldavud mgagne: indeed. I was actually wondering if this wouldn't break many of the puppet-openstack modules
20:04 mgagne aldavud: anyone using the module
20:05 aldavud dachary: one more small thing. due to some clumsiness while implementing the patch i lost your +2 on https://review.openstack.org/#/c/100522/
20:06 aldavud dachary: it's now again the same patchset you previsouly reviewed again... thanks for having a look :)
20:07 openstackgerrit Matt Fischer proposed a change to stackforge/puppet-keystone: allow setting the tenant descriptions  https://review.openstack.org/106157
20:07 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Configure OVS mechanism agent configs in its config file  https://review.openstack.org/106144
20:07 aldavud mgagne: yup. but we're relying heavily on it, as every os module has some custom type inheriting it for config afaik
20:07 mgagne aldavud: yes =)
20:08 openstackgerrit David Gurtner proposed a change to stackforge/puppet-ceph: Escape the base64 encoded key in regexp matching  https://review.openstack.org/102211
20:10 mgagne EmilienM: we should still test against redhat as I'm not sure if the reported bug occurs when using packastack (and puppet-neutron) or not: https://bugzilla.redhat.co​m/show_bug.cgi?id=1087647
20:10 openstackgerrit A change was merged to stackforge/puppet-ceph: Escape the base64 encoded key in regexp matching  https://review.openstack.org/102211
20:14 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-ironic: Add support for MySQL module 2.2  https://review.openstack.org/106063
20:14 gcha mgagne: yes we had a bug in packstack with the l2population mechanism ... and social found that neutron::agents::ml2::ovs was broken
20:14 mgagne gcha: how was it broken?
20:15 mgagne gcha: because agent configs went into the "wrong" config file?
20:15 gcha mgagne: yep
20:15 gcha let me check
20:15 mgagne gcha: ok, so it "should" work now but according to original report, it isn't clear if packstack was used or not
20:16 gcha mgagne: was working with the changeset #1 I sent anyway
20:16 mgagne gcha: oh, with a brand new install?
20:17 gcha mgagne: will check one more time with your last changeset
20:17 tfz joined #puppet-openstack
20:18 gcha mgagne: yes with a brand new install ... https://bugzilla.redhat.co​m/show_bug.cgi?id=1118010
20:20 gcha mgagne: Error: /Stage[main]/Neutron::Agents::Ml2::Ovs/Neut​ron_plugin_ml2[agent/l2_population]/ensure: change from absent to present failed:
20:20 gcha Could not set 'present' on ensure: No such file or directory - /etc/neutron/plugins/ml2/ml2_conf.ini at
20:20 gcha 144:/var/tmp/packstack/92770015ac324419b8d1289b77c​206fc/modules/neutron/manifests/agents/ml2/ovs.pp
20:20 mgagne gcha: it shouldn't be Neutron_plugin_ml2
20:20 mgagne gcha: did I miss one?
20:21 gcha mgagne: no don't worry ... was the error we've got before to work on a puppet-neutron patch ;-)
20:30 mgagne super fun, iproute in precise does not support vxlan
20:42 tfz joined #puppet-openstack
20:55 aldavud joined #puppet-openstack
21:05 xarses mgagne: my understand was that most of precise doesn't support vxlan
21:05 mgagne xarses: sad panda =(
21:06 mgagne xarses: what's the point of backporting OpenStack packages if such major feature isn't supported =)
21:06 xarses I recall someone saying that the kernel needs some tlc too
21:13 hogepodge magane I wouldn’t say hiera was “pushed”. It was more of an ill-advised design decision that needs to be undone (I have an outsanding jira ticket for that work)
21:19 hogepodge s/magane/mgagne/ ^^
21:23 mgagne hogepodge: it is to feed a design discussion by making sure people do not make Hiera mandatory
21:24 hogepodge mgagne totally understand. I recognize the mistake I made.
21:24 mgagne hogepodge: there is no mistake to be sorry about =)
21:25 JayJ joined #puppet-openstack
21:26 tfz joined #puppet-openstack
21:29 JayJ hogepodge: There is one error in Ceilometer. How do you want me to report it? Good news is I installed a single node controller and added a single compute. I was able to launch a VM succesfully.
21:29 JayJ hogepodge: Btw, I had to make a manual change to use ML2.
21:30 JayJ hogepodge: Please do let me know when you make a new release so I can retest it
21:30 hogepodge JayJ In the puppet-ceilometer package, or how I’m using it?
21:30 blentz joined #puppet-openstack
21:31 JayJ It was in Ceilometer package. Right not how you are using it
21:31 JayJ I just commented that out for my testing
21:36 hogepodge JayJ Bugs can be filed in Launchpad, and you can submit a patch through the gerrit code review system. https://wiki.openstack.org/wiki/Gerrit_Workflow
21:36 JayJ hogepodge: New to me.. will read through it :)
21:37 JayJ hogepodge: Thanks! You are being very helpful
21:37 hogepodge The launchpad is here. https://launchpad.net/puppet-ceilometer
21:38 hogepodge JayJ It’s my pleasure. It’s a big ecosystem and the process can be a bit daunting. Once you get used to it the benefits are really great and it becomes second nature.
22:05 mfisch Anyone know if puppet-keystone should be able to handle password changes?
22:05 mfisch From the comments in the code I think it should be able to, but it's unhappy
22:05 mgagne mfisch: how unhappy is it?
22:06 mfisch Could not evaluate: Execution of '/usr/bin/keystone --os-auth-url http://foo/v2.0 token-get returned 1: The request you have made requires authentication. (HTTP 401)
22:06 mfisch that error then upsets ansible
22:06 mfisch it is an Error, I didnt paste that portion
22:07 hogepodge It looks like the credentials environment isn’t being set?
22:07 mfisch well the password is wrong
22:08 mfisch for example
22:08 mfisch from my laptop
22:08 mfisch OS_PASSWORD=fred keystone token-get
22:08 mfisch The request you have made requires authentication. (HTTP 401)
22:08 mfisch if the password is wrong, and everything else is set, thats the response you get
22:09 mfisch looks like that raises an exception from the Ruby based on my rudimentary knowledge
22:09 hogepodge Is it trying to use the new password?
22:10 mfisch yeah it should be, it comes from the YAML
22:10 mfisch the comment claims: " if it doesn't then return nil, causing it to be reset"
22:10 mfisch but instead of returning "nil" it raises an exception
22:10 hogepodge That’s the problem I think. For things like mysql we cache the credentials so we can change them later on.
22:10 mfisch if the password is wrong shouldnt it try user-update?
22:11 mfisch user-password-update I mean
22:11 mfisch oh
22:11 mfisch I see
22:11 mfisch the message changed!@
22:11 mfisch return nil if e.message =~ /Not Authorized/
22:11 mfisch keystone client changed
22:11 mfisch I can fix this
22:11 hogepodge That’s a pattern we’re going to have to think about.
22:11 hogepodge crinkle is going to help us fix the underlying problem with aviator and openstacklib
22:12 mfisch that will make my life happier
22:12 mfisch manager was annoyed at deployment issues though, maybe we can fix this quickly for now
22:12 mfisch I'll file a bug
22:12 gcha mgagne: it works like a charm on RHEL ;-) thanks !
22:12 hogepodge with the admin password you shuold be able to change anything, except the admin password.
22:12 mfisch yep
22:13 mfisch let me look at fixing tonight
22:13 mfisch post beers
22:13 mfisch ruby makes more sense then
22:14 hogepodge mfisch just make sure that any fix works for both the old and new new version. It’s not hard, if one check fails try the other. We don’t want to break compatibility.
22:15 mfisch yep
22:15 mfisch a || b
22:15 mfisch was my plan
22:15 mfisch hogepodge: can I get this on your backlog?  https://review.openstack.org/#/c/102025/
22:18 hogepodge mfisch I keep circling back to that one. I’m torn on the test.
22:18 mfisch me too
22:18 mfisch hogepodge: however, horizon is working to redo that file at the BP phase anyway
22:18 mfisch hogepodge: to make it simpler
22:18 hogepodge yeah.
22:19 hogepodge I can +2 it, but you’ll need one more.
22:21 * mfisch looks around for bribe material
22:29 med_ ah discussion was here. I thought you guys were discussing in #openstack-keystone
22:33 rwsu joined #puppet-openstack
22:33 Qu4tre joined #puppet-openstack
22:35 JayJ joined #puppet-openstack
22:36 ikke-t joined #puppet-openstack
22:36 sputnik13 joined #puppet-openstack
22:36 higgins joined #puppet-openstack
22:36 ricardorocha joined #puppet-openstack
22:36 izinovik joined #puppet-openstack
22:36 holser joined #puppet-openstack
22:36 EmilienM joined #puppet-openstack
22:36 beddari joined #puppet-openstack
22:36 jkt joined #puppet-openstack
22:36 sbadia joined #puppet-openstack
22:49 mmaglana joined #puppet-openstack
22:50 mjblack_ joined #puppet-openstack
22:59 hogepodge mfisch poke other core if you need to.
23:17 aldavud joined #puppet-openstack
23:31 mjblack joined #puppet-openstack
23:45 aimon joined #puppet-openstack
23:58 aldavud joined #puppet-openstack

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary