Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2016-10-11

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

All times shown according to UTC.

Time Nick Message
00:04 openstackgerrit Matt Fischer proposed openstack/puppet-keystone: set 0600 permissions on fernet keys & folder  https://review.openstack.org/382136
00:36 markvoelker joined #puppet-openstack
00:37 openstackgerrit Keith Schincke proposed openstack/puppet-openstack-integration: Introduce scenario004  https://review.openstack.org/339106
01:02 ducttape_ joined #puppet-openstack
01:08 ducttape_ joined #puppet-openstack
01:08 xiaohua joined #puppet-openstack
01:12 markvoelker joined #puppet-openstack
01:20 openstackgerrit Keith Schincke proposed openstack/puppet-openstack-integration: Introduce scenario004  https://review.openstack.org/339106
01:35 xiaohua joined #puppet-openstack
01:47 ilbot3 joined #puppet-openstack
01:47 Topic for #puppet-openstack is now CI status: green - Puppet OpenStack modules http://docs.openstack.org/deve​loper/puppet-openstack-guide/
02:07 xiaohua joined #puppet-openstack
02:10 markvoelker joined #puppet-openstack
02:42 xiaohua joined #puppet-openstack
02:49 openstackstatus NOTICE: Jobs running on osic nodes are failing due to network issues with the mirror. We are temporarily disabling the cloud.
03:02 ducttape_ joined #puppet-openstack
03:11 xiaohua joined #puppet-openstack
04:01 xiaohua joined #puppet-openstack
04:11 openstackgerrit Keith Schincke proposed openstack/puppet-openstack-integration: Introduce scenario004  https://review.openstack.org/339106
04:15 ducttape_ joined #puppet-openstack
04:17 vern joined #puppet-openstack
05:08 xiaohua joined #puppet-openstack
05:15 markvoelker joined #puppet-openstack
05:17 ducttape_ joined #puppet-openstack
05:18 markvoelker joined #puppet-openstack
05:19 markvoelker joined #puppet-openstack
05:20 ayoung joined #puppet-openstack
05:25 markvoelker_ joined #puppet-openstack
05:27 markvoelker joined #puppet-openstack
05:28 markvoelker joined #puppet-openstack
05:45 markvoelker joined #puppet-openstack
06:07 rcernin joined #puppet-openstack
06:14 xiaohua joined #puppet-openstack
06:18 vern joined #puppet-openstack
06:18 ducttape_ joined #puppet-openstack
06:35 xiaohua joined #puppet-openstack
06:53 vern joined #puppet-openstack
06:56 dgurtner joined #puppet-openstack
06:58 hjensas joined #puppet-openstack
07:00 tdb joined #puppet-openstack
07:03 jaosorior joined #puppet-openstack
07:03 Vinsh joined #puppet-openstack
07:06 openstackgerrit Christian Schwede proposed openstack/puppet-tripleo: Add versioned_writes to Swift proxy config  https://review.openstack.org/384704
07:10 ccamacho joined #puppet-openstack
07:50 dgurtner joined #puppet-openstack
07:56 aarefiev joined #puppet-openstack
07:58 hogepodge joined #puppet-openstack
08:05 jesusaur joined #puppet-openstack
08:27 openstackgerrit Athlan-Guyot sofer proposed openstack/puppet-tripleo: Ensure presence of pacemaker restart directory.  https://review.openstack.org/384870
08:29 openstackgerrit Athlan-Guyot sofer proposed openstack/puppet-tripleo: Ensure presence of pacemaker restart directory.  https://review.openstack.org/384870
08:32 xarses_ joined #puppet-openstack
08:32 pester joined #puppet-openstack
08:34 derekh joined #puppet-openstack
08:35 openstackgerrit Javier Peña proposed openstack/puppet-openstacklib: [TEST DO NOT MERGE] Test for Packstack debugging on Puppet 4  https://review.openstack.org/384637
08:39 llua` joined #puppet-openstack
08:40 amoralej_ joined #puppet-openstack
08:40 mjblack- joined #puppet-openstack
08:44 dgurtner joined #puppet-openstack
08:44 dgurtner joined #puppet-openstack
08:44 openstackgerrit Merged openstack/puppet-ec2api: Fix documentation for log_dir parameter  https://review.openstack.org/384593
08:44 openstackgerrit Merged openstack/puppet-murano: Fix documentation for log_dir parameter  https://review.openstack.org/384575
08:45 openstackgerrit Merged openstack/puppet-vitrage: Change worker defaults to ::os_workers  https://review.openstack.org/383959
08:45 cgfbee joined #puppet-openstack
08:49 davidmichaelkarr joined #puppet-openstack
08:49 vern joined #puppet-openstack
08:49 guimaluf joined #puppet-openstack
08:49 khappone joined #puppet-openstack
08:49 holser_ joined #puppet-openstack
08:49 mwhahaha joined #puppet-openstack
08:49 mgagne joined #puppet-openstack
08:49 mjb` joined #puppet-openstack
08:50 davidmichaelkarr joined #puppet-openstack
08:50 openstackgerrit Merged openstack/puppet-octavia: Fix documentation for log_dir parameter  https://review.openstack.org/384591
08:52 gfidente joined #puppet-openstack
08:53 mwhahaha joined #puppet-openstack
08:57 markvoelker joined #puppet-openstack
08:59 openstackgerrit Merged openstack/puppet-watcher: Fix documentation for log_dir parameter  https://review.openstack.org/384560
09:00 tdb joined #puppet-openstack
09:14 openstackgerrit Merged openstack/puppet-panko: Fix documentation for log_dir parameter  https://review.openstack.org/384502
09:28 paramite joined #puppet-openstack
09:30 openstackgerrit Juan Antonio Osorio Robles proposed openstack/puppet-aodh: Add option to enable Aodh's SSL middleware  https://review.openstack.org/384371
09:31 openstack joined #puppet-openstack
09:39 jaosorior joined #puppet-openstack
09:41 jpich joined #puppet-openstack
09:47 shardy joined #puppet-openstack
09:59 markvoelker joined #puppet-openstack
09:59 jaosorior iberezovskiy, degorenko: are you around?
10:00 degorenko jaosorior, yes
10:00 iberezovskiy +
10:01 jaosorior degorenko, iberezovskiy: So, I've put out this commit https://review.openstack.org/#/c/384371/ but haven't quite figured out why it's failing the unit tests http://logs.openstack.org/71/384371/​2/check/gate-puppet-aodh-puppet-unit​-3.8-centos-7/545dc8d/console.html
10:02 jaosorior SystemStackError: stack level too deep
10:02 jaosorior what's that about?
10:03 degorenko strange error
10:04 degorenko haven't seen this before
10:04 Vinsh joined #puppet-openstack
10:05 openstackgerrit Juan Antonio Osorio Robles proposed openstack/puppet-aodh: Add option to enable Aodh's SSL middleware  https://review.openstack.org/384371
10:08 openstackgerrit Juan Antonio Osorio Robles proposed openstack/puppet-gnocchi: Add option to enable Gnocchi's SSL middleware  https://review.openstack.org/384358
10:11 openstackgerrit Juan Antonio Osorio Robles proposed openstack/puppet-ceilometer: Add option to enable Ceilometer's SSL middleware  https://review.openstack.org/384366
10:11 jaosorior degorenko: I think I figured it out
10:11 jaosorior came from a copy/paste error
10:12 degorenko jaosorior, it happens :)
10:13 richm joined #puppet-openstack
10:21 openstackgerrit Maksym Yatsenko proposed openstack/puppet-manila: Change worker defaults to ::os_workers  https://review.openstack.org/384620
10:23 ducttape_ joined #puppet-openstack
10:31 ntpttr joined #puppet-openstack
10:44 openstackgerrit Merged openstack/puppet-congress: Fix documentation for log_dir parameter  https://review.openstack.org/384449
10:47 fultonj joined #puppet-openstack
10:57 hjensas joined #puppet-openstack
10:59 markvoelker joined #puppet-openstack
11:08 xiaohua joined #puppet-openstack
11:11 basilAB joined #puppet-openstack
11:24 ducttape_ joined #puppet-openstack
11:36 xiaohua joined #puppet-openstack
11:37 xiaohua joined #puppet-openstack
11:41 johnmilton joined #puppet-openstack
11:42 openstackgerrit Javier Peña proposed openstack/puppet-nova: [TEST DO NOT MERGE] Test for Packstack debugging on Puppet 4  https://review.openstack.org/384946
11:47 dprince joined #puppet-openstack
11:48 ansiwen EmilienM: regarding neutron_vpnaas_available, why did you set it to false, and do you have any further information how to make it work?
11:49 openstackgerrit joined #puppet-openstack
11:49 ansiwen EmilienM: enabling results in this error: http://logs.openstack.org/42/356442/29/ch​eck/gate-puppet-openstack-integration-4-s​cenario001-tempest-centos-7/bef0e08/logs/​tempest.txt.gz#_2016-10-10_14_55_07_900
11:49 ansiwen EmilienM: but the ec2-api tempest tests include VPN tests... :-/
11:54 xiaohua joined #puppet-openstack
11:59 EduNeto joined #puppet-openstack
12:00 markvoelker joined #puppet-openstack
12:06 EmilienM hello
12:07 mjblack hello
12:14 EmilienM ansiwen: I haven't made it work
12:15 pradk joined #puppet-openstack
12:17 ducttape_ joined #puppet-openstack
12:18 openstackgerrit joined #puppet-openstack
12:19 dgurtner joined #puppet-openstack
12:19 dgurtner joined #puppet-openstack
12:30 ducttape_ joined #puppet-openstack
12:32 dgurtner_ joined #puppet-openstack
12:32 dgurtner_ joined #puppet-openstack
12:38 dmsimard|pto mwhahaha: re: puppet-designate & monasca: https://bugs.launchpad.net/designate/+bug/1630666 :)
12:38 openstack Launchpad bug 1630666 in Designate "monasca-statsd is a non-optional import even when disabled" [Low,Triaged]
12:39 dtantsur hey folks, do you have work session for the summit planned already? I've just realized I have a window at 11:50 on Thu, and there is a work session there.
12:43 markvoelker joined #puppet-openstack
12:58 jpich joined #puppet-openstack
13:03 chem joined #puppet-openstack
13:04 fultonj_ joined #puppet-openstack
13:05 jtriley joined #puppet-openstack
13:12 xiaohua joined #puppet-openstack
13:14 basilAB joined #puppet-openstack
13:14 jperry joined #puppet-openstack
13:23 morazi joined #puppet-openstack
13:24 StraubTW joined #puppet-openstack
13:30 jaosorior Hey guys, are there currently issues with gate-puppet-ceilometer-puppe​t-beaker-rspec-ubuntu-xenial or gate-puppet-openstack-integration-4​-scenario001-tempest-ubuntu-xenial ?
13:30 mwhahaha ugh, why does rdo use *-dist.conf files :/
13:30 mwhahaha jaosorior: link?
13:31 openstackgerrit Alexey Deryugin proposed openstack/puppet-trove: Make swift_service_type configurable in guestagent.conf  https://review.openstack.org/385000
13:31 jaosorior mwhahaha: https://review.openstack.org/#/c/384371/  http://logs.openstack.org/71/384371/3/che​ck/gate-puppet-openstack-integration-4-sc​enario001-tempest-ubuntu-xenial/1111f14/
13:32 mwhahaha http://logs.openstack.org/71/384371/3/che​ck/gate-puppet-openstack-integration-4-sc​enario001-tempest-ubuntu-xenial/1111f14/c​onsole.html#_2016-10-11_11_47_37_872735
13:32 mwhahaha looks like connection issues with the apt configuration
13:32 jaosorior funky
13:33 jaosorior mwhahaha: and http://logs.openstack.org/66/384366/2/che​ck/gate-puppet-ceilometer-puppet-beaker-r​spec-ubuntu-xenial/a9d1f78/console.html
13:34 mwhahaha same deal
13:34 mwhahaha looks like it hung at the same spot doing the same thing
13:35 jaosorior mwhahaha: I'll recheck them then. Thanks
13:38 EmilienM jaosorior: yeah http://logs.openstack.org/71/384371/3/che​ck/gate-puppet-openstack-integration-4-sc​enario001-tempest-ubuntu-xenial/1111f14/c​onsole.html#_2016-10-11_11_47_37_880317
13:38 EmilienM it sounds like a transient issue to reach ceph repo
13:43 morazi joined #puppet-openstack
13:44 EmilienM I see a lot of ssl issues
13:44 EmilienM http://logs.openstack.org/63/384463/1/check/gate-p​uppet-ceilometer-puppet-beaker-rspec-ubuntu-xenial​/8b271ea/console.html#_2016-10-11_10_32_28_253776
13:46 StraubTW_ joined #puppet-openstack
13:46 EmilienM https://download.ceph.com/keys/release.asc looks down
13:47 EmilienM well, http://download.ceph.com/ looks down
13:47 dprince joined #puppet-openstack
13:49 EmilienM which causes all CI jobs running on Xenial to fail
13:52 dmsimard mwhahaha: what's the problem with -dist.conf files ?
13:53 EmilienM dmsimard: have you seen the thread "[openstack-dev] [rpm-packaging][chef][puppet]​[salt][openstack-ansible][HA] Schema proposal for config file handling for services" ?
13:53 mwhahaha dmsimard: our assumptions around using os_service_default is broken if it's defined in a dist file
13:53 mwhahaha I didn't know the dist file was a thing and I am sad
13:53 EmilienM imho, it's too opinionated to add parameters in a dist file
13:53 EmilienM we should rely on what upstream provides
13:53 EmilienM and fix upstream if something is not correct
13:55 EmilienM ok http://ceph.com/ is down too
13:56 agrebennikov joined #puppet-openstack
13:57 dmsimard mwhahaha, EmilienM: not aware of the historical reasons of providing those, I'm not closed to the idea of getting them out
13:57 dmsimard in most cases, the dist conf files don't really make sense anyway
13:57 dmsimard EmilienM: no, I have not
13:57 mwhahaha they are the 'defaults'
13:57 clayton mwhahaha: I don't think it's really invalid, it's just different
13:57 mwhahaha they actually define like db connections and stuff
13:57 moravec joined #puppet-openstack
13:57 mwhahaha so they map to usually what are the required params anyway
13:57 mwhahaha which is probably why we've never noticed
13:58 clayton anyone that doens't want that behavior can remove the -dist.conf files
13:58 mwhahaha clayton: by default our puppet stance has been to fall back to the python defaults
13:59 mwhahaha we shouldn't have to clean out packager's files to get that
13:59 mwhahaha it's already an annoyance when we have to do it for the apache files or deal with autostarting crap from deb
13:59 mwhahaha now for redhat we also need to remember to nuke some random dist.conf?
13:59 mwhahaha meh
13:59 clayton I think if an operator is using a distribution that does that, then they've implicitly accepted that the vendor knows better than upstream
13:59 dfisher joined #puppet-openstack
14:00 mwhahaha yea that's totally not the case
14:00 mwhahaha i don't think anyone asctually knows better
14:00 mwhahaha not even upstream
14:00 moravec joined #puppet-openstack
14:00 clayton mwhahaha: the real point of os_service_default was to get puppet out of the business of hard coding defaults
14:00 clayton the other parts of it was a side-effect
14:00 mwhahaha right but packaging is doing that
14:00 mwhahaha and they aren't doing it smartly either
14:00 clayton sure, but that's a different issue, I don't see how the puppet modules can solve that issue
14:01 ducttape_ joined #puppet-openstack
14:01 mwhahaha i'm not trying to solve it with puppet modules
14:01 mwhahaha i'm trying to prevent packagers from doign this
14:01 mwhahaha just fyi we can't clear out log_dir
14:01 mwhahaha because it's in the dist file
14:01 clayton I agree with you that they shouldn't, but fundamentally it's kind of their and the operators problem
14:02 clayton why would removing log_dir and defaulting to the value in the dist file cause an issue?
14:02 mwhahaha the default value in dist is /var/log/heat
14:02 mwhahaha for example in heat-dist.conf
14:02 clayton we already have a similar issue on ubuntu because it hard codes the log file paths in the startup scripts
14:02 mwhahaha so to not log to a dir at all you need to completely remove it
14:03 mwhahaha right this is just one example as to why these files are not a good idea
14:03 clayton is that something people actually want to do, or just an example?
14:03 mwhahaha people actually want to do it (and it's an example)
14:03 clayton so, personally I think the dist files are an improvement over the vendor shipping a huge config file full of settings, even if I have to override it
14:04 clayton I'd rather the file that puppet manages only contain things that it put there
14:04 mwhahaha i think my issue is ordering around this
14:04 clayton if I were deploying a distribution that did put dist files in place, I'd probably want a way to disable reading the dist files, but I'd leave them in place for reference.
14:04 mwhahaha we've operated under the assumption that puppet is the source of truth for configuration
14:04 mwhahaha but it's not when distros drop files in
14:04 clayton I don't think that's true
14:05 clayton The way I think of it is that puppet is only managing things that the operator has explicitly asked it to, directly or indirectly.
14:06 clayton the purge config items pushes the puppet modules more in the direction you said, but it isn't on by default
14:06 mwhahaha well here's the case where an operator has told it to configure something and it doesn't take because of the dist file
14:06 mwhahaha not a complete purge but we are purging some values
14:06 mwhahaha via os_service_default
14:06 clayton so using your  example, how does an operator tell a module to remove log_dir from the config?
14:07 mwhahaha which breaks when the dist files actually contain it
14:07 mwhahaha sets it to os_service_default
14:07 mwhahaha ie remove it completely from the config files
14:07 clayton I think that's a bad example, because log_dir isn't set to os_service_default
14:07 mwhahaha it used to be set it to false but that got broken a while ago
14:07 mwhahaha it is
14:08 clayton https://github.com/openstack/puppet-n​ova/blob/e8a4765c624cc8b69ad9b0a7bd11​202c76df4c34/manifests/logging.pp#L99
14:08 mwhahaha it's not by default but if a user wants to remove it they can't
14:08 mwhahaha ah you're right we define it, but we tell the user to use os_service_default to disbale it
14:08 clayton I guess a user could set it to os_service_default
14:08 mwhahaha so if they do explictly tell it to do something, it wouldn't work as designed
14:08 mwhahaha https://github.com/openstack/puppet-n​ova/blob/e8a4765c624cc8b69ad9b0a7bd11​202c76df4c34/manifests/logging.pp#L25
14:09 clayton I don't know if it's still the case, it used to be explicitly set to not os_service_default because services required it
14:09 clayton I assume that's changed.
14:09 mwhahaha we used to say false but oslo broke that at somepoint so false gets interpreted as a file or dir name
14:09 mwhahaha it's not required anymore as far as i know
14:09 mwhahaha anwyay
14:10 clayton my take on this is that the dist files are roughly equivalent to the packager changing the defaults in the python code, just slightly cleaner
14:10 clayton it's them saying "we know better than upstream, you need this set to something different"
14:10 mwhahaha my concern is that feature and settings will creep into the dist files
14:11 mwhahaha if you're relying on a configuration management tool, you'd expect it would do that for you
14:11 mwhahaha you're not relying on the packaging for that info
14:11 clayton maybe so, but packagers pretty regularly do dumb things, at least in my experience.
14:11 mwhahaha which is even more of a reason why i don't want them playing in the config files
14:11 mwhahaha i'm not against a smart schema to allow people to do better things like a conf.d directory
14:12 mwhahaha i'm against packagers injecting configuration defaults unecessarily
14:12 clayton sure, I could see where people not using the puppet modules might find that useful
14:12 mwhahaha it would be better to fall back to the dist file if no service.conf file exists
14:12 clayton sure, I just don't see how much can be done for it.
14:13 mwhahaha otherwise if say heat.conf exists don't use heat-dist.conf
14:21 iurygregory_ joined #puppet-openstack
14:36 openstackgerrit Javier Peña proposed openstack/puppet-openstacklib: [TEST DO NOT MERGE] Test for Packstack debugging on Puppet 4  https://review.openstack.org/384637
14:41 xiaohua joined #puppet-openstack
14:46 iurygregory_ meeting in 15 minutes
14:46 iurygregory_ :D
14:46 Vinsh will there be food served?
14:48 iurygregory_ Vinsh, only in Barcelona =P
14:48 Vinsh -2
14:48 iurygregory_ :O
14:51 jperry joined #puppet-openstack
14:59 llua joined #puppet-openstack
15:01 mwhahaha shall we do a quick meeting or should we skip it
15:01 mwhahaha no real topics, just some past stuff
15:01 iurygregory_ i was wondering this XD
15:01 iurygregory_ nothing to discuss about summit or PTG?
15:02 mwhahaha i don't think so
15:02 mwhahaha we'll just do a quick meeting
15:04 openstackgerrit Maksym Yatsenko proposed openstack/puppet-magnum: Change worker defaults to ::os_workers  https://review.openstack.org/384619
15:15 openstackgerrit Iury Gregory Melo Ferreira proposed openstack/puppet-openstack-guide: Meeting #93  https://review.openstack.org/385054
15:19 moravec joined #puppet-openstack
15:24 EmilienM http://www.dreamhoststatus.com/2016/10/11/drea​mcompute-us-east-1-cluster-service-disruption/
15:24 EmilienM so this is why ceph.com is down
15:25 EmilienM I propose we disable ceph/ceph repos this afternoon if no change
15:25 EmilienM it's blocking our CI
15:25 EmilienM or make xenial jobs non voting
15:25 _ducttape_ joined #puppet-openstack
15:26 Vinsh so, is ceph.com down because ceph that it's hosted on is down!? :-0
15:28 EmilienM most likely
15:33 jpena anyone wants to help me debug the most awkward puppet issue I've ever seen? :)
15:34 jpena (spoiler alert: it's puppet4-related)
15:34 mwhahaha jpena: what's the error?
15:35 jpena mwhahaha, if I understood it correctly, the puppet-nova nova_config provider is trying to use openstackconfig's ini_setting as parent, instead of ruby as it should according to https://github.com/openstack/puppet-​nova/blob/master/lib/puppet/provider​/nova_config/openstackconfig.rb#L3
15:36 jpena what's worse, it only happens in the gate, I've seen 3 different people trying to reproduce it locally without success
15:36 jpena I can prepare an etherpad with a long explanation if it helps
15:37 mwhahaha yea that'd help alternatively do you have the raw logs somewhere?
15:37 jpena mwhahaha, http://logs.openstack.org/87/383787/12/check/​gate-packstack-integration-scenario001-tempes​t-centos-7/5a09c54/logs/packstack/latest/mani​fests/15.184.65.66_controller.pp.log.txt.gz is the main log
15:38 jpena search for Nova_config[DEFAULT/debug], and you'll see a debug marker I added as a test with https://review.openstack.org/384637
15:40 mwhahaha not sure how does nova_config have anything to do with the nova_flavor not working?
15:41 jpena mwhahaha, that's because nova.conf is incorrectly configured. Nova fails to start, then nova_flavor can't contact any API server
15:41 jpena http://logs.openstack.org/87/383787/12/check/ga​te-packstack-integration-scenario001-tempest-ce​ntos-7/5a09c54/logs/etc/nova/nova.conf.txt.gz is the generated nova.conf file, see how all configured options looks like arrays
15:41 mwhahaha that's a new one
15:41 EmilienM the question is, how does it work in p-o-i and not in packstack?
15:41 jpena I'd love to know that
15:42 jpena also, why does puppet-neutron (which does exactly the same as puppet-nova) work fine?
15:43 moravec joined #puppet-openstack
15:43 mwhahaha what version of ini_file is it using?
15:44 mjblack EmilienM,mwhahaha,jpena: was the nova_config updated recently for multiline?
15:44 mwhahaha wonder if puppet-nova needs https://github.com/openstack/puppet-neu​tron/blob/master/lib/puppet/provider/ne​utron_config/openstackconfig.rb#L10-L13
15:45 mwhahaha mjblack: no it's been there for a while
15:45 mjblack well
15:45 jpena mjblack, the change merged late July
15:45 mjblack hmmm
15:46 mjblack btw who is the maintainer of packstack?
15:46 mwhahaha jpena: what verison of inifile are you using in packstack?
15:46 jpena mwhahaha, it's using inifile from master
15:46 jpena mjblack, I'm one of the packstack maintainers
15:46 jpena imcsk8 (say hi!) is another one
15:47 imcsk8 hi!
15:47 mjblack :D
15:47 openstackgerrit Alexey Deryugin proposed openstack/puppet-murano: Make Murano application provider use prefetch  https://review.openstack.org/378741
15:47 mjblack so with the arrays, if the provider wasnt written right it will do that
15:47 mwhahaha jpena: we're using 1.6.0
15:47 jpena mwhahaha, the weirdest thing is that nova_config shouldn't ever use inifile, it should use the ruby provider of openstack_config
15:48 mjblack or rather the type
15:48 jpena mwhahaha, I can test inifile 1.6.0, let me try
15:48 mwhahaha just to rule out inifile changes
15:49 jpena let's give it 20-30 mins, brb
15:50 mwhahaha though i'm not seeing much changes bettwen 1.6.0 and master which would affect functionality
15:50 EmilienM jpena: we are using latest tag of dependencies
15:50 EmilienM jpena: our CI automatically bump Puppetfile to latest tags
15:50 EmilienM https://github.com/puppetlabs​/puppetlabs-inifile/releases
15:54 mwhahaha jpena: https://github.com/openstack/puppet-op​enstacklib/blob/master/lib/puppet/prov​ider/openstack_config/ruby.rb#L59-L66 it might be using the right provider
15:59 mjblack EmilienM: want a good laugh? http://logs.openstack.org/79/382379/4/check​/gate-puppet-openstack-integration-4-scenar​io002-tempest-ubuntu-xenial/e4c68e4/logs/ba​rbican/main.txt.gz#_2016-10-10_18_28_57_416
15:59 EmilienM can I cry?
16:00 mjblack depends
16:00 mjblack have you found the issue on that line? :P
16:01 mjblack once you do, you'll understand why it fails the volume encryption
16:01 mjblack 00000000-0000-0000-0000-000000000000 appears to be returned by barbican on error
16:02 mjblack request: 404 Not Found - GET http://127.0.0.1:9311/v1/secrets/00​000000-0000-0000-0000-000000000000
16:10 ducttape_ joined #puppet-openstack
16:10 jpena mwhahaha, I don't get it. That bit is ensuring we're passing an array, but https://github.com/openstack/puppet-nova/blo​b/master/lib/puppet/type/nova_config.rb#L10 also sets the value to be an array. Or am I getting it wrong?
16:21 _ducttape_ joined #puppet-openstack
16:56 mwhahaha jpena: that might be it but like you said neutron supports that and it worked :D
16:58 jpena mwhahaha, yeah well... :). BTW, inifile 1.6.0 didn't do the trick. I'm checking the log files, but the issue stays
16:58 mwhahaha so weird
16:59 jpena is there any good way I could trace the puppet provider, so I can see where it comes from? Something like traceback in python
16:59 mwhahaha no idea, chem would probably know better
17:00 moravec joined #puppet-openstack
17:00 mwhahaha jpena: packstack captures all the modules right?
17:01 mwhahaha in the ci logs
17:01 jpena mwhahaha, yes
17:01 mwhahaha i wonder if you have two nova_configs
17:01 mwhahaha and the older one is the one loading
17:01 chem jpena: do you mean interactive debugging session or something else ?
17:02 mwhahaha i really need to write something to pull down all the logs from openstack ci so you can just grep locally
17:02 jpena chem; I mean a full backtrace
17:02 EmilienM mwhahaha: like a single tarball?
17:03 mwhahaha EmilienM: that would be ideal, but if not something you can just paste a url and it pulls all the files down
17:03 EmilienM well, a tarball can be implemented in one line
17:03 EmilienM in the copy logs scripts
17:03 mwhahaha EmilienM: to keep storage down, rather than duplicating the stuff in both exploded and unexploded forms
17:03 mwhahaha it's handy to be able to pass urls
17:03 mwhahaha but sometimes I just want to grep all the files
17:04 mwhahaha also across projects :D
17:04 chem jpena: hum ... I don't see the use case.  Do you have a example where you would like the provider to dump a full stacktrace ?
17:05 mwhahaha chem: he's trying to figure out why nova_config appears to be writing out using the ini_setting rather than the ruby provider
17:05 jpena that ^^ :)
17:05 mwhahaha since the options in the config file are being written out as arrays
17:05 mwhahaha which makes it look like it's using the ini_settings rather than ruby
17:06 chem jpena: ah, I see, you want the code path.  But show me the problem, I have some familiarity with this ini/ruby provider.  Just need a refresher.
17:07 jpena chem: https://etherpad.openstack​.org/p/Packstack-Puppet-4 is a summary
17:07 chem jpena: (the only option here is to add Puppet.info("blah ${var}" inside the provider)
17:07 jpena chem: ok, that's more or less what I'm trying :)
17:07 xarses_ joined #puppet-openstack
17:09 chem jpena: you can add 'caller.join("\n")' to have the stacktrace inside your message, but I doubt it will be useful
17:10 chem jpena: so you have a array output while you would like a string output, right ?
17:10 ayoung joined #puppet-openstack
17:11 jpena chem: I've just tried that. No, it shows ruby functions, but not the provider calls. Something like https://paste.fedoraproject.org/448689/20587014/
17:13 chem jpena: the best way is to use 'gem "pry"' to the Gemfile, bundle install, and add dynamic breakpoint inside the provider with "binding.pry"
17:13 ayoung joined #puppet-openstack
17:13 chem jpena: but I think I can sort this out ...
17:13 chem jpena: I've got this problem ... a long time ago, I believe
17:14 jpena chem: if you do, I'll owe you a few beers in Barcelona. I've been scratching my head for two days
17:14 chem jpena: well for the trace, that what I though, it's kind of useless as it's so tangled with puppet that it's undecipherable.
17:15 chem jpena: give me 10
17:17 ayoung joined #puppet-openstack
17:20 chem jpena: so well, I did something on this : https://review.openstack.org/#/c/354094/1/README.md
17:21 chem jpena: I try to associate that with your current problem
17:21 chem jpena: (note the binding.pry stuff is not doable at the gate)
17:23 dgurtner joined #puppet-openstack
17:25 fultonj joined #puppet-openstack
17:27 AnarchyAo joined #puppet-openstack
17:29 iurygregory joined #puppet-openstack
17:47 jpena chem: gotta go now, thanks for your help and I'll keep on troubleshooting ;)
17:48 chem jpena: this is an interesting problem :)  I'm still spending some time on it, keep you updated
17:48 jpena great, thanks!
17:51 ducttape_ joined #puppet-openstack
17:52 ducttape_ joined #puppet-openstack
18:16 EmilienM this think with ceph.com makes me think we should stop relying on ceph.com by default
18:17 EmilienM and maybe use openstack mirrors or officials repos in ubuntu
18:18 iurygregory we have thoose options?
18:21 ducttape_ joined #puppet-openstack
18:27 EmilienM iurygregory: not yet ;-) but I'm on it
18:27 iurygregory EmilienM, you rock o/
18:32 dmsimard What is ::neutron::memcache_servers for ? Looks like overlap with keystone authtoken ? https://github.com/openstack/puppet-n​eutron/blob/master/manifests/init.pp
18:33 dmsimard Forgot to deprecate it ?
18:34 dmsimard The one from ::nova was properly deprecated: https://github.com/openstack/puppet-no​va/blob/master/manifests/init.pp#L428
18:34 dmsimard iurygregory: ^
18:35 iurygregory dmsimard, looking
18:36 * dmsimard working on upgrading from mitaka to newton
18:36 dmsimard so far so good
18:36 iurygregory dmsimard,  memcache_servers vs memcached_servers
18:37 iurygregory they should do something different?
18:37 dmsimard iurygregory: https://github.com/openstack/puppet-neu​tron/blob/master/manifests/init.pp#L578
18:37 dmsimard parameter is memcache_servers, config is memcached_servers :)
18:37 iurygregory damm it LOL
18:38 dmsimard iurygregory: worth fixing and backporting that to stable/newton I think
18:38 iurygregory dmsimard, yep we missed in the review https://review.openstack.org/#/c/351272/
18:39 iurygregory doing now
18:39 iurygregory if you don't mind =)
18:40 EmilienM dmsimard: no
18:40 EmilienM dmsimard: it's for servicegroup
18:40 EmilienM like in Nova
18:40 EmilienM it's used to store the status of Neutron services
18:40 dmsimard EmilienM: I don't understand
18:42 iurygregory humm in nova the memcached_servers was used in keystone_authtoken not in DEFAULT
18:42 iurygregory that's why we deprecated
18:42 iurygregory https://github.com/openstack/puppet-nova/com​mit/94a6343b57d9f6ddc4f580a781449857983e9441​#diff-60ae41fd0a31977447947f59940ee9a4L646
18:42 iurygregory so this option in neutron is used for something else like dmsimard
18:42 dmsimard but we can agree that if DEFAULT/memcached_servers is set, it'll be used as default if the one in keystone_authtoken is not yet, right ?
18:43 EmilienM I'm not sure
18:43 dmsimard let me look at the neutron config
18:43 EmilienM let me find in code
18:43 * dmsimard cloning neutron to generate sample config file T_T
18:43 ducttape_ joined #puppet-openstack
18:44 EmilienM it seems like the option is not used anymore
18:45 iurygregory http://git.openstack.org/cgit/openstack/op​enstack-manuals/tree/doc/config-reference/​source/samples/neutron/neutron.conf.sample
18:46 iurygregory so maybe we should deprecate =D
18:46 dmsimard EmilienM, iurygregory: up to date sample http://paste.openstack.org/show/585381/
18:47 dmsimard only see it in keystone_authtoken
18:47 iurygregory yep
18:47 iurygregory and the code in init is for DEFAULT
18:48 EmilienM let's deprecate it!
18:48 dmsimard sorry :P
18:57 iurygregory we will remove after Ocata?
18:57 EmilienM yes
18:57 iurygregory ack
19:03 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
19:08 iurygregory waiting for local tests before push to gerrit =)
19:16 xarses I thought ensure_packages() was supposed to avoid http://logs.openstack.org/36/384736/1/che​ck/gate-puppet-openstack-integration-4-sc​enario002-tempest-ubuntu-xenial/eb478a3/c​onsole.html#_2016-10-10_23_19_16_950531 regarding https://review.openstack.org/#/c/​384736/1/manifests/backend/swift.pp
19:16 xarses I guess I need to make it so they have the same name then?
19:20 iurygregory xarses, maybe use the same name and tags
19:20 xarses I need to add the new tag
19:20 iurygregory you can add
19:20 iurygregory but also use openstack i think
19:20 xarses it has openstack already
19:21 xarses https://github.com/openstack/puppet-s​wift/blob/master/manifests/client.pp
19:21 xarses if it's declared in swift
19:22 iurygregory but you are in glance no? so maybe  tag ['openstack', 'glance-support-package']
19:24 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
19:35 moravec joined #puppet-openstack
19:40 johnmilton joined #puppet-openstack
19:45 johnmilton joined #puppet-openstack
19:48 moravec joined #puppet-openstack
19:52 openstackgerrit Tim Rozet proposed openstack/puppet-neutron: Changes OpenDaylight OVS 'provider_mappings' to allow an array type  https://review.openstack.org/382551
19:53 openstackgerrit Iury Gregory Melo Ferreira proposed openstack/puppet-neutron: Deprecate memcache_servers  https://review.openstack.org/385135
19:54 jperry joined #puppet-openstack
19:56 openstackgerrit Andrew Woodward proposed openstack/puppet-glance: Ensure that swiftclient is installed  https://review.openstack.org/384736
19:58 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
19:59 openstackgerrit Tim Rozet proposed openstack/puppet-neutron: Changes OpenDaylight OVS 'provider_mappings' to allow an array type  https://review.openstack.org/382551
20:11 dmsimard Upgraded mitaka puppet recipes to successfully deploy newton with minimal effort, gj :)
20:20 mwhahaha yay?
20:20 mwhahaha :D
20:25 dmsimard There was only one deprecation that was brutal
20:25 dmsimard the one with the ec2 removal from puppet-nova
20:25 dmsimard unless mistaken that was outright removed without deprecation
20:25 dmsimard I forgot to check though
20:28 dmsimard or maybe it was something else, I forget now
20:33 mwhahaha yea we kinda missed the window on that one i think
20:33 mwhahaha kinda got everyone at the last minute
20:36 xarses anything we can do to reduce the cloning errors?
20:36 xarses http://logs.openstack.org/36/384736/2/check/gat​e-puppet-glance-puppet-beaker-rspec-centos-7/71​c4441/console.html#_2016-10-11_20_03_51_593732
20:36 mwhahaha not really
20:37 dmsimard yeah ultimately even if we carried them in /opt/git cache in the images, we need to fetch and that'd fail
20:38 dmsimard and I don't know if r10k even supports fetching from git cache
20:54 ducttape_ joined #puppet-openstack
20:59 michapma_alt joined #puppet-openstack
21:17 openstackgerrit David Moreau Simard proposed openstack/puppet-tempest: Fix config parameter name for volume-feature-enabled  https://review.openstack.org/385152
21:17 dmsimard colonwq: wanna rebase your patches with a Depends-On: ^ ? :)
21:18 dmsimard or I can do it, hang on
21:18 colonwq dmsimard, I can do that for all 3.
21:19 openstackgerrit David Moreau Simard proposed openstack/puppet-openstack-integration: Add cinder backup to scenario002  https://review.openstack.org/381083
21:19 moravec joined #puppet-openstack
21:19 openstackgerrit David Moreau Simard proposed openstack/puppet-openstack-integration: Add cinder::backup test to scenario0001  https://review.openstack.org/377158
21:20 colonwq dmsimard, I am building a truth table for 004 to see what minimum conbination of p-o-i::tempest options will generate a working tepest config.
21:20 dmsimard colonwq: ^
21:20 colonwq dmsimard, thanks.
21:20 dmsimard colonwq: all 3 ? did you also have that issue with scenario004 ?
21:21 colonwq dmsimard, for 004, there is a combo of p-o-i::tempest arguments which will update part of the tempest packaging to a point where it runs.
21:21 colonwq It is a diff tempeset problem.
21:23 dmsimard colonwq: ok haven't had the chance to look at 004 yet.. got the low hanging cinder fruit :)
21:23 dmsimard pretty confident that'll fix it
21:25 colonwq dmsimard, thanks. I am wash/rinse/repeat my test with diff settings. I'll update the patch with a minimal set soon.
21:27 dmsimard colonwq: looks like all the centos7 errors in 004 are due to endpoint not found on cinder (???)
21:27 dmsimard tempest::cinder_available is set to true in fixture but you don't deploy it
21:29 dmsimard the rest seems like it works relatively ok ?
21:29 dmsimard I left a comment
21:31 colonwq dmsimard, thanks for the comment. I have other config options where the tempest run fails. That run ran with a bunch of extra stuff turned on.
21:32 dmsimard Ah, I don't have historical context - my comment is on the latest patchset of course :)
21:32 ducttape_ joined #puppet-openstack
21:32 colonwq I'll be back in a few hours. I have a boy scout meeting to take care of.
21:32 dmsimard see you later
21:37 dmsimard mwhahaha: I'm really itching to refactor puppet-tempest, wish I had more time.. I'll probably do it bit by bit throughout ocata ..
21:38 mwhahaha k
21:39 dmsimard do you typically write specs for significant refactors ?
21:44 mwhahaha not in puppet
21:44 mwhahaha unless you're really going to hose backwards compatibility
21:50 dmsimard ok
21:51 dmsimard man the mirror issues today all over the place :(
21:53 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
21:56 EmilienM CI looks broken http://logs.openstack.org/24/385124/3/ch​eck/gate-puppet-openstack-integration-4-​scenario001-tempest-centos-7/a5855db/con​sole.html#_2016-10-11_20_07_56_018957
21:56 EmilienM dmsimard: ^
21:57 dmsimard There's a ton of network errors on OSIC nodes recently
21:57 dmsimard #openstack-infra
21:58 EmilienM dmsimard: ack, thanks
22:05 zigo joined #puppet-openstack
22:14 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
22:14 * EmilienM still trying to bring our CI back alive
22:16 dmsimard EmilienM: clarkb is working on isolating the issue
22:16 EmilienM dmsimard: I'm dealing with ceph.com down
22:16 EmilienM ceph.com has been down for 12 hours
22:17 EmilienM and of course, they host bug tracker, website, repo on the same node
22:17 openstackgerrit Ben Nemec proposed openstack/puppet-heat: Make default domain id match actual default  https://review.openstack.org/385179
22:38 openstackgerrit Ben Nemec proposed openstack/puppet-heat: Make default domain id match actual default  https://review.openstack.org/385179
22:55 iurygregory_ joined #puppet-openstack
23:08 iurygregory_ CI still red right?
23:14 EmilienM iurygregory_: yes
23:14 EmilienM iurygregory_: updated ^
23:14 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
23:34 xarses joined #puppet-openstack
23:41 iurygregory_ joined #puppet-openstack
23:52 openstackgerrit Emilien Macchi proposed openstack/puppet-ceph: repo: deploy ceph.com apt key only when required  https://review.openstack.org/385124
23:56 ducttape_ joined #puppet-openstack
23:56 jtriley joined #puppet-openstack

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