Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-11-27

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

All times shown according to UTC.

Time Nick Message
00:34 _ilbot joined #puppet-openstack
00:34 Topic for #puppet-openstack is now Place to collaborate on Puppet/OpenStack tools: logs at http://irclog.perlgeek.de/puppet-openstack/today
00:50 rmoe joined #puppet-openstack
00:56 otherwiseguy joined #puppet-openstack
01:26 otherwiseguy joined #puppet-openstack
01:26 rongze joined #puppet-openstack
01:46 rharrison joined #puppet-openstack
02:02 sputnik13 joined #puppet-openstack
02:07 xingchao joined #puppet-openstack
02:09 rongze joined #puppet-openstack
02:38 brdude joined #puppet-openstack
02:49 sputnik13 joined #puppet-openstack
03:08 xarses joined #puppet-openstack
03:47 badiane_ka joined #puppet-openstack
05:18 rongze joined #puppet-openstack
05:34 michchap joined #puppet-openstack
05:51 mjblack joined #puppet-openstack
06:13 rongze joined #puppet-openstack
06:22 comptona joined #puppet-openstack
06:31 otherwiseguy joined #puppet-openstack
06:42 comptona joined #puppet-openstack
06:56 comptona joined #puppet-openstack
07:40 dvanders joined #puppet-openstack
07:42 EmilienM michchap: o/
07:44 dvanders_ joined #puppet-openstack
07:50 dyerm joined #puppet-openstack
08:38 michchap EmilienM: ola
08:38 EmilienM michchap: hey
08:38 EmilienM thanks for your doc
08:38 EmilienM michchap: are you using puppet-openstack module ?
08:39 michchap It's ok, I'm going to keep building it up until I have enough conttent to go through all the parts and do a screencast
08:39 michchap I think it's used, but not for control
08:40 michchap it might be used for the openstack::compute wrapper, which is still kinda useful. I will check
08:40 michchap ah! It's used for provision
08:41 michchap and for swift
08:41 EmilienM michchap: i was wondering where are manifests
08:41 michchap no manifests, just yaml
08:41 * EmilienM does not know yera
08:41 EmilienM hiera*
08:41 michchap it's not hiera replacing the manifests
08:41 michchap the scenario_node_terminus does that
08:42 michchap so it will compile the provided data so that nodes are including the correct classes
08:42 michchap and that's determined by mapping hostnames to roles, and then mapping roles to classes+class_groups
08:43 michchap hiera comes in after that, to provide parameters to the classes
08:44 michchap so the site.pp just looks like this:
08:44 michchap node default {}
08:44 michchap ie. it doesn't actually do anything.
09:03 giorgiodinapoli joined #puppet-openstack
09:04 toMeloos joined #puppet-openstack
09:05 derekh joined #puppet-openstack
09:13 bogdando joined #puppet-openstack
10:31 rongze_ joined #puppet-openstack
11:15 rongze joined #puppet-openstack
11:57 dmsimard joined #puppet-openstack
12:37 rharrison joined #puppet-openstack
12:44 markvoelker1 joined #puppet-openstack
13:15 marun joined #puppet-openstack
13:31 giorgiodinapoli joined #puppet-openstack
13:38 seiflotfy joined #puppet-openstack
13:38 mkoderer joined #puppet-openstack
13:43 dprince joined #puppet-openstack
13:43 dalgaaf_ joined #puppet-openstack
14:39 marun joined #puppet-openstack
14:50 mjblack joined #puppet-openstack
15:04 dachary dmsimard: ping ?
15:07 dmsimard dachary: o/
15:07 dachary \o
15:08 djschaap joined #puppet-openstack
15:09 dachary dmsimard: I would like to know how you feel about https://review.openstack.org/#/c/56770/ . I'm not asking that you review it right away, just if the proposed solution is agreeable to you. If not I'll try to figure something else.
15:10 dmsimard Yeah I haven't had time, been busy with production - my first concern about the review was not addressed, I'm not sure about adding instructions that do not (yet) work or will end up working differently, it's confusing for the end user
15:11 dmsimard I can settle with a usecase file
15:14 dachary cool :-) We have no user yet, not much risk for confusion.
15:16 dachary dmsimard: you are using centos right ?
15:20 dmsimard nope, Ubuntu or Debian
15:23 dachary ah, xarses does use centos, sorry for the confusion dmsimard :-)
15:23 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-cinder: Improve Cinder Backup support  https://review.openstack.org/58792
15:24 openstackgerrit Sebastien Badia proposed a change to stackforge/puppet-swift: Bugfix: Fix fragment name for Quota middleware  https://review.openstack.org/58793
15:26 toMeloos joined #puppet-openstack
15:41 xarses dachary: centos, ubuntu and rohs (probably less and less)
15:41 xarses erm rhos
15:42 mattymo xarses, by the way, RHOS is still solid on our manifests. just tested a few mins ago
15:46 dmsimard EmilienM: So many nova's in the README of ironic :)
15:47 EmilienM dmsimard:  ?
15:47 dmsimard EmilienM: Do a "find" for nova in the readme, you'll understand
15:47 EmilienM dmsimard: oh yeah i took nova example
15:48 EmilienM but i need to write doc indeed
15:48 EmilienM dmsimard: again, it's a WIP
15:48 dmsimard EmilienM: Not criticizing, just feedback :p
15:48 EmilienM sbadia: ^
15:48 dmsimard EmilienM: that gemfile.lock file too
15:48 EmilienM it's your commit ;)
15:49 sbadia :-D oups indeed …
15:50 sbadia readme is also a wip ;-)
15:50 EmilienM ahah
15:56 dmsimard It's all good, looking forward to ironic
16:00 otherwiseguy joined #puppet-openstack
16:04 EmilienM dmsimard: coming soon, stay tuned
16:04 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-cinder: Improve Cinder Backup support  https://review.openstack.org/58792
16:04 EmilienM sbadia: should be better ^
16:04 EmilienM \o/
16:05 sbadia EmilienM: syntastic changed my life :-D
16:05 EmilienM sbadia: actually jenkins is not happy
16:05 EmilienM sbadia: damn, it is on my laptop
16:08 michchap joined #puppet-openstack
16:11 EmilienM L47 of spec
16:12 sbadia EmilienM: http://pub.sebian.fr/pub/syntastic.png
16:12 sbadia EmilienM: for syntastic (vi plugin)
16:14 dachary What is the recommended way to run a 2 lines shell script from within a puppet module ?
16:15 dachary exec { command => '/bin/cmd foo bar' } is ok
16:15 EmilienM sbadia: nice
16:15 giorgiodinapoli joined #puppet-openstack
16:17 dachary exec { command => '/bin/cmd foo bar ; other arg arg ' } ?
16:20 dachary command     => '/sbin/ifdown -a ; /sbin/ifup -a',
16:21 dachary says puppet-nova/manifests/init.pp
16:22 mgagne joined #puppet-openstack
16:22 fvollero EmilienM: you saw my comment? :)
16:22 EmilienM fvollero: where ? sorr
16:22 EmilienM sorry
16:23 fvollero EmilienM: on your review
16:23 sbadia dachary: hum, i think it's not a good idea
16:23 EmilienM fvollero: oh thank you :)
16:23 EmilienM get it
16:23 sbadia because if the first command fail, the return code is not fetched
16:23 fvollero EmilienM: yeah, that's just a silly mistake, but i still don't understand what's wrong with jenkins
16:23 sbadia false;true;echo $?
16:23 sbadia 0
16:23 sbadia :)
16:24 fvollero sbadia: in a way, it's a right behavior :)
16:24 fvollero s/a/the/
16:25 sbadia indeed, in this example it's ok
16:27 dachary sbadia: do you mean that it's ok as long as the exit code accurately reflects the script success / failure ?
16:29 sbadia maybe this is heresy indeed :-)
16:32 dachary how do you mean ?
16:35 mgagne dachary: http://thingsyoudidntknowaboutjenkins.tumblr.com/
16:35 mgagne dachary: first post
16:38 dachary I'm absolutely in favor of having an exit code that means something, indeed ;-) Bottom line is : it's ok to have multi command exec as long as the exit code is not hidden from puppet. Right ?
16:39 mgagne dachary: sorry, thought you talked about jenkins. with puppet, yes, it's the same behaviour, you can only have ONE exit code.
16:39 mgagne dachary: I got confused when fvollero talked about jenkins
16:39 dachary :-)
16:40 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-cinder: Improve Cinder Backup support  https://review.openstack.org/58792
16:40 EmilienM fvollero: feel free to review ^
16:40 EmilienM if jenkins is happy this time
16:40 mgagne dachary: use && if you want to make sure ALL commands are executed with success
16:41 dachary great, thanks. I was not sure if multi command exec was frowned upon or not.
16:42 EmilienM fvollero: seem yes, nice. Looking forward for review
16:42 mgagne dachary: it depends. If you move everything to an external script and call it from puppet, what's the difference? I mean, a script is a multi command exec hidden in a script. It's all about debugging and having a clear error on failure
16:43 mgagne dachary: if "cmd_a && cmd_b" fails, puppet will probably returns "cmd_a && cmd_b" failed which is kind of useless.
16:43 dachary mgagne: my question was more oriented toward having less than 5 commands in a row and absolutely no logic in the script. Just to avoid repeating the command with 5 exec
16:44 mgagne dachary: as long as it's readable and easy to understand ;)
16:44 dachary and debuggable
16:44 mgagne dachary: yep
16:47 dachary I drives me nuts to have an error message such as "fail error, failure fatal error" ;-)
16:47 EmilienM mgagne: sorry to spam you with reviews, i had a lot of patch on the last days
16:48 mgagne EmilienM: it's ok, sorry if i'm not active this weeks, lot of work to catch up here ^^'
16:48 EmilienM mgagne: and i have a lot of new ones coming soon
16:48 EmilienM mgagne: good luck with that, looking forward having -1 from you
16:48 EmilienM :)
16:48 mgagne haha
16:48 * EmilienM miss -1 from mgagne
16:53 xarses joined #puppet-openstack
16:55 brdude joined #puppet-openstack
17:02 dachary how do you run only a designated test ?
17:04 dachary nvm
17:07 tnoor1 joined #puppet-openstack
17:13 sputnik13 joined #puppet-openstack
17:32 bogdando joined #puppet-openstack
17:34 angdraug joined #puppet-openstack
17:38 rmoe joined #puppet-openstack
18:18 Gwyxx joined #puppet-openstack
18:27 toMeloos anybody around with experience with the enovance/telekomcloud puppet-ceph module?
18:28 toMeloos joined #puppet-openstack
18:33 tnoor1 joined #puppet-openstack
18:34 tnoor1 joined #puppet-openstack
18:36 tnoor1 joined #puppet-openstack
18:37 tnoor2 joined #puppet-openstack
18:42 tnoor1 joined #puppet-openstack
18:43 xarses toMeloos: you probably want to poke dmsimard
18:45 toMeloos thanks yes I will. but I see they are on the official ceph channel so I won't pollute this channel
18:46 sputnik13 joined #puppet-openstack
18:49 mjblack_ joined #puppet-openstack
18:53 rmoe_ joined #puppet-openstack
19:17 tnoor2 joined #puppet-openstack
19:24 tnoor1 joined #puppet-openstack
19:24 xarses puppet lint question
19:27 xarses I have a string in which bash var sub occurs, puppet accepts line single quoted, or double quoted (with the the $ escaped). In single quote lint errors on the string that there is a single quoted string. in the second case it warn's that double quotes are used with out a variable in the string
19:27 xarses any thoughts on how best to avoid this?
19:31 rongze joined #puppet-openstack
19:34 mgagne xarses: so far, I found no way but to disable the check
19:34 dmsimard xarses: For ceph instead of setting something like /var/lib/ceph/mon/mon.$id, I set /var/lib/ceph/mon/mon._id
19:34 dmsimard and tested accordingly
19:34 mgagne xarses: I found one one trick and it's ugly as hell
19:35 xarses hmm, aparently if i escape both, \$\{ it shuts up
19:35 xarses and leave it single quoted
19:35 mgagne xarses: really??
19:35 xarses need to make sure it dosn't fobber the string
19:35 xarses but lint did shut up
19:35 dmsimard xarses: https://github.com/stackforge/puppet-ceph/bl​ob/master/spec/classes/ceph_osd_spec.rb#L44
19:39 xarses and it mangled the string
19:41 mgagne 2013-08-05T22:13:26  <mgagne> clarkb: check is looking for "${"
19:41 mgagne 2013-08-05T22:13:45  <mgagne> clarkb: https://github.com/sbryant/puppet-lint/blob/mast​er/lib/puppet-lint/plugins/check_strings.rb#L64
19:41 mgagne 2013-08-05T22:14:02  <mgagne> clarkb: but removing {} would trigger "variable not enclosed in {}"
19:42 mgagne 2013-08-05T22:33:19  <mgagne> clarkb: I think I had the worst idea ever
19:42 mgagne 2013-08-05T22:35:57  <mgagne> clarkb: chomp("file:///var/lib/git/\${name}.git\n")
19:42 mgagne that's what I found last time I checked
19:42 mgagne http://eavesdrop.openstack.org/irclogs/%23open​stack-infra/%23openstack-infra.2013-08-05.log
20:01 rongze joined #puppet-openstack
20:06 xarses uggly as sin
20:07 xarses but it works if you use join(['some string with ', '$', '{bash} sub'], '')
20:11 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-ceph: Add ceph::repo RedHat support  https://review.openstack.org/57063
20:11 mgagne xarses: try to grep that now
20:11 xarses ?
20:11 mgagne xarses: try to grep stuff splited in a join()
20:12 xarses ya, i know
20:12 xarses hopefully, no one needs to
20:12 mgagne xarses: haha, you wish
20:12 xarses otherwise, it would probably have to come from a template file
20:13 xarses alternaly, if an actual puppet interpolation occurrs and there is a escaped one, it tricks lint into being ok with it
20:15 xarses "${pupp_var}/\${bash_var}" works fine
20:16 xarses lint doesn't  care then
20:16 mgagne fun ^^'
20:17 xarses seriously
20:23 otherwiseguy joined #puppet-openstack
20:31 rongze joined #puppet-openstack
20:37 Jean-Roger joined #puppet-openstack
20:44 bkero joined #puppet-openstack
20:52 dalgaaf_ toMeloos: did you try to start all Mons in parallel or only one of them?
21:03 toMeloos one at a time
21:05 toMeloos but I'm thinking if "ceph --name 'mon.' --keyring '/var/lib/ceph/tmp/keyring.mon.1' -s" times out then isn't this an authentication issue?
21:09 sputnik13 joined #puppet-openstack
21:23 dalgaaf_ try to start all 3 mons in parallel
21:34 rongze joined #puppet-openstack
22:04 EmilienM mgagne: it fix a blocking bug: https://review.openstack.org/#/c/58793/
22:04 EmilienM mgagne: if you could have a quick review that would be awesome
22:06 mgagne EmilienM: fragments are realized with pipeline in swift::proxy right?
22:09 mgagne EmilienM: AFAIK, it isn't a blocker. I see no manifest depending on the concat::fragment title
22:10 mgagne EmilienM: pipeline items are used to computed the classes /etc/swift/proxy-server.conf depends on.
22:14 EmilienM mgagne: i agree, but i did an error in the first patch when i added support for quota middlewares, they need to be fixed to seem like others
22:14 mgagne EmilienM: it's not a blocker ;)
22:14 EmilienM mgagne: you're right
22:14 mgagne EmilienM: a blocker: you can't install swift anymore
22:14 EmilienM indeed
22:15 EmilienM my sentence was too high :)
22:15 EmilienM mgagne: i know you are busy, i just try to catch your attention :P
22:15 mgagne ok
22:16 openstackgerrit A change was merged to stackforge/puppet-swift: Bugfix: Fix fragment name for Quota middleware  https://review.openstack.org/58793
22:17 EmilienM mgagne: thanks
22:32 xarses mgagne: https://github.com/rodjek/puppet-lint/pull/234
22:33 mgagne xarses: awesome!
22:33 michchap joined #puppet-openstack
22:34 xarses mgagne: fixes the issue we where talking about and some complex escaping stuff that a coworker did
22:34 mgagne xarses: yes! can't wait for the fix to merge :D
22:35 rongze joined #puppet-openstack
22:50 otherwiseguy joined #puppet-openstack
22:58 dachary xarses: Warning: Scope(Class[Apt::Update]): Could not look up qualified variable 'apt::update_timeout'; class apt has not been evaluated
22:59 mgagne dachary: make sure apt class is included
22:59 dachary in ceph::repo ( as it currently is ) which is fixed if include apt instead of include apt::update
22:59 mgagne dachary: they refactored the module I guess...
23:00 dachary :-)
23:00 dachary mgagne: I'm not sure I understand why this error shows
23:01 mgagne dachary: apt::update is now depending on a variable initialized in apt. You need to include apt to have it declared and initialized.
23:02 dachary hum
23:02 dachary ok
23:02 xarses dachary: is this related to https://review.openstack.org/#/c/57063/ ?
23:02 dachary xarses: no
23:03 dachary xarses: but I figured you may have run into the issue while working on it, hence the ping ;-)
23:03 xarses dachary: ah, ok
23:05 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-ceph: Add ceph::repo RedHat support  https://review.openstack.org/57063
23:09 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Ensure ml2 is set in core_plugin  https://review.openstack.org/58881
23:09 EmilienM michchap: ^
23:10 dachary We did not discuss --cluster versus --conf yet. --cluster implies /etc/ceph/$cluster.conf and is used consistenly in the upstart scripts. --conf is not used. In other words, the upstart scripts can be leveraged to start the mon / osd of a given cluster with ceph-mon cluster=foo for instance. But there is no way to specify an alternate ceph.conf path. It follows that the puppet module should use --cluster over --conf .
23:11 EmilienM michchap: how could i backport that ?
23:11 EmilienM regarding to my ML2 general backport
23:11 EmilienM michchap: need i keep 2 backports (the current + the last fix of now) ?
23:13 michchap I have no idea what the standard practice is
23:13 EmilienM michchap: ok, i think the best way is to backport the 2 patches
23:13 michchap I would just add that change to your current patch because it's easier
23:13 michchap haha ok
23:13 EmilienM michchap: oh ok
23:13 EmilienM let's do it know
23:13 EmilienM now*
23:19 rharrison joined #puppet-openstack
23:23 dachary when puppet runs with  puppet apply --debug --trace /tmp/puppetapply.921831018  it exits with code 0
23:24 dachary when I add the  --detailed-exitcodes
23:24 dachary it exits with 2
23:24 EmilienM michchap: backport ready
23:24 dachary have you faced something similar already ?
23:24 rharrison dachary, that means it ran successfully but made changes.
23:24 dachary hrmms
23:25 dachary rharrison: thanks :-)
23:25 xarses dachary: if you intend to use puppet control a upstart service that relies on instance, puppet will freak out
23:25 EmilienM michchap: oops, spec file has not been cherry picked
23:25 rharrison dachary, yeah the detailed-exitcodes is something I use most of the time.
23:26 xarses dachary: we already found that the radosgw upstart scripts cant be used with puppet
23:26 michchap EmilienM: jenkins is not happy
23:26 dachary xarses: ack. It's probably wise to use them when they are good and not when they are bad ;-)
23:27 EmilienM michchap: he does not like me i know
23:27 EmilienM fixing it
23:27 dachary rharrison: rspec-system fails because of that, I'll check it out. Thanks for the info ;-)
23:28 xarses hmm i guess jenkins wedged again
23:29 rharrison dachary, it fails because of a non-zero?  That's sort of the point of detailed-exitcodes is that it provides more information in the exit code than "I worked" or "something went wrong" with it on I'd expect puppet to return 2 on the first run of puppet apply and hopefully 0 on follow up runs.
23:29 EmilienM michchap: working !
23:29 EmilienM michchap: i'm done for today
23:29 dachary no, it fails because I require it to be zero, my mistake
23:29 dachary rharrison: ^
23:30 xarses yay https://review.openstack.org/#/c/57063/ passed jenkins
23:32 EmilienM cheers guys
23:32 michchap EmilienM: night
23:36 rongze joined #puppet-openstack
23:39 rharrison dachary, yeah you either want to have 0 or 2 be passing or drop the detailed-exitcodes from the run.
23:39 dachary xarses: :-)
23:41 rharrison dachary, actually these are jenkins tests?  The environment should be clean so I'd expect 0 should be a fail as well
23:42 dachary rharrison: these are rspec-system tests
23:42 dachary rspec-system-puppet
23:48 sputnik13 joined #puppet-openstack
23:48 rharrison dachary, I've been meaning to play with that.  Does it depend on rspec-puppet or is totally independent?
23:49 dachary I'm not sure. It probably does. The syntax is very similar.
23:52 rharrison I'm still waiting for rojek to release the next version of rspec-puppet so I can finally get it packaged for Fedora.  The current version depends on puppet-spec_helper which depends on rspec-puppet...
23:55 dachary ah :-)

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