Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-09-13

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

All times shown according to UTC.

Time Nick Message
00:14 prad joined #puppet-openstack
00:35 openstackgerrit Ian Wienand proposed a change to stackforge/puppet-nova: Deprecate sql_* for database_*  https://review.openstack.org/46172
00:36 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Add missing dependency on neutron-server  https://review.openstack.org/46395
00:39 ianw bodepd: ping
00:42 bodepd ianw: what up?
00:42 bodepd ianw: sorry for slightly hi-jacking your patch comments :)
00:43 ianw bodepd: np!
00:43 mgagne1 testing havana, ironically in grizzly :D
00:43 ianw bodepd: re the database_connection in init.pp; i'd agree but would like to maintain consistency too
00:44 bodepd no worries. I was hoping to get that patch in and saw your patch as the perfect time to try to get it in
00:44 bodepd epic fail :)
00:44 bodepd ianw: sorry, what part did you agree with?
00:44 bodepd ianw: wanting to split it out?
00:45 ianw yeah, it doesn't really belong in init.pp i think
00:45 ianw same with heat, it really belongs in heat::engine as that's the only thing that wants the db
00:45 mgagne ianw: at least it's optional
00:45 bodepd ianw: I did intend for my comments to mean that we should do it across all of the modules (not just nova)
00:46 bodepd ianw: I'm not even sure what services from nova need access to the db
00:46 bodepd ianw: I don't think it's a blocker for getting that patch merged
00:46 bodepd ianw: but, I do think it's a great time for that refactor
00:47 bodepd mgagne: having it be optional has it's own set of problems
00:47 bodepd mgagne: way to easy for people to accidentally create local sqlite dbs
00:47 mgagne bodepd: fun
00:48 ianw bodepd: yeah, i was thinking how to refactor in general.  has anyone ever put together a wiki with a matrix of the classes for the projects?
00:48 bodepd I'm looking at the code to make sure I actually know what I'm talking about
00:48 ianw because there seem to be a lot of commonality and duplicated code
00:48 bodepd mchapman wrote something
00:48 mgagne bodepd: I don't think I have that problem on my side
00:48 bodepd ianw: I totally agree
00:48 bodepd ianw: there should be an openstack-common module to remove some of the duplicate
00:48 mgagne bodepd: havana is coming fast and that kind of refactor need better planning as maintaing backward compatibility will be hard (IMO)
00:48 bodepd coe
00:49 bodepd moving things to a common module would not break backwards compat
00:50 bodepd and if we did move where data_connection is set as a part of the current proposed patches
00:50 bodepd ah
00:50 bodepd gotcha
00:50 bodepd it would be hard to maintain backwards compat
00:50 mgagne bodepd: yes and I think it's better to wait for icehouse for that kind of change
00:52 mgagne bodepd: and it could be mixed with a broader refactor of all modules while we are here
00:54 ianw mgagne bodepd: is this worth a blueprint?
00:54 mgagne ianw: or etherpad =)
00:55 ianw can etherpad do decent tables?
00:58 digicloseup joined #puppet-openstack
00:59 mgagne ianw: do we need tables?
00:59 mgagne ianw: wiki otherwise
00:59 digicloseup joined #puppet-openstack
01:00 ianw mgagne: i was thinking a list of modules and classes would maybe reveal cross-overs.  let me see how that might look on a wiki page
01:02 bodepd mgagne: https://github.com/bodepd/openstack-installer/blob/data_refactor/data/data_mappings/scenario/2_role.yaml
01:02 bodepd ianw: that link is kind of relevant for what you are asking about
01:02 mgagne need moar data
01:02 bodepd ianw: it shows how class parameters map to the same value for the 2_role deployment
01:03 mgagne bodepd: oh god, my nightmare
01:04 ianw yeah :)
01:07 ianw bodepd mgagne: hmm, maybe trying to take a cohesive view and fix everything isn't going to work then :)
01:07 bodepd ianw: we've been pretty sensitive to breakages in backwards compat
01:07 mgagne we scared hunner
01:07 ianw what about working on one specific component that is duplicated a lot, like sql/database_connection or rpc_backend?
01:08 ianw if a theoretical openstack-puppet-common was providing rpc_backend that would get rid of a lot of duplicate code?
01:08 mgagne ianw: those are the big chunks and I feel finding the solution for one will make fixing the other easier.
01:08 bodepd ianw: if you can do it without breaking backwards compat, please do
01:08 bodepd mgagne: why is that your nightmare? the inconsistencies?
01:09 ianw is it at the point you just declare "v2" :) ?
01:09 bodepd ianw: that is a good question for the list
01:09 mgagne ianw: yes and no. Can't OpenStack projects upgrade oslo components at their own speed?
01:10 bodepd mgagne: https://github.com/bodepd/openstack-installer/blob/data_refactor/data/data_mappings/common.yaml#L23
01:10 mgagne bodepd: your stubborn idea of assigning left value to the right one =)
01:10 bodepd mgagne: get over that. I will have to change it anyways to port it to hiera-2
01:10 bodepd mgagne: focus on the other parts of it :)
01:11 bodepd mgagne: I was trying to get you to look at the other parts.
01:11 bodepd mgagne: how data flows to populate classes, and how it highlights inconsistencies in naming
01:12 mgagne bodepd: I know inconsistencies exist =)
01:12 bodepd mgagne: one nice thing, is that I am trying to use as many default values as possible
01:13 bodepd mgagne: to limit the total number of params, which has led to all of my patches this week about default inconsistencies
01:13 mgagne bodepd: I think I kind of mitigated the problem by rolling out my own openstack module and forgot about it
01:13 bodepd mgagne: yeah, so did I
01:14 mgagne bodepd: now I sleep in peace
01:14 mgagne bodepd: but I agree with you that
01:14 mgagne bodepd: a cleanup has to be done in that aspect
01:14 bodepd one more thing for you
01:14 bodepd https://github.com/bodepd/openstack-installer/blob/data_refactor/data/hiera_data/grizzly_hack.yaml
01:15 bodepd a hiera file to backport all of my defaul changing patches for havana to grizzly :)
01:15 bodepd (I haven't submitted all of the patches yet, but those shoudl be all of them)
01:16 mgagne :D
01:16 mgagne about enabling nova services by default
01:17 mgagne should we do it?
01:17 bodepd mgagne: yes
01:17 bodepd mgagne: things should work by default
01:17 bodepd and the end user should have to supply as few parameters as possible
01:18 bodepd https://github.com/bodepd/openstack-installer/blob/data_refactor/data/data_mappings/advanced.yaml#L2
01:18 mgagne bodepd: yes
01:18 bodepd package_ensure wins the award for most inconsistent
01:18 mgagne :D
01:18 bodepd with keystone_url related things taking a close second
01:18 mgagne oh god
01:18 bodepd keystone vs. auth vs. ...
01:19 mgagne bodepd: tell me about it
01:19 bodepd I can show you :)
01:21 mgagne if a class is defined with class {}, can it be included after?
01:21 mgagne bodepd: I'm trying to see how we could design our modules with hiera in mine while still making it optional
01:22 bodepd yes.
01:22 bodepd it can. it's really confusing
01:22 xingchao joined #puppet-openstack
01:22 mgagne bodepd: the main problem is with mandatory values
01:23 bodepd yeah. I don't know how to do that
01:23 bodepd redhat update to 3.x
01:23 bodepd so, maybe just make hiera mandatory
01:23 * mgagne is celebrating
01:24 bodepd it would simplify things
01:24 bodepd maybe that is a target for ice-house
01:24 mgagne bodepd: also, I have a difficult situation here where some people I know want to rsync the stuff onto each servers and run puppet apply from there. My concern is with hiera data.
01:25 mgagne bodepd: hiera data which might contain sensitive information. (of course)
01:25 mgagne bodepd: but that shouldn't land on some servers for security purposes.
01:25 mgagne bodepd: my making hiera mandatory, I don't know how I should handle that use case.
01:25 bodepd mgagne: you can do that with hiera.
01:25 bodepd mgagne: keep the hiera data in a stupic simple couchdb
01:26 bodepd mgagne: then let them run puppet apply and target that with a custom backend?
01:26 mgagne bodepd: hmmm. how to handle acl?
01:27 mgagne bodepd: with puppetmaster, the node only has access to the value it needs and nothing else.
01:27 bodepd maintain variables server side that are assigned to each server based on certname?
01:27 bodepd and then, just send the data that is compiled from that authorized list?
01:27 mgagne bodepd: hmmm
01:27 bodepd you would of coarse, have to build that :)
01:28 mgagne bodepd: maybe my question is out of scope for the discussion, although somehow related to the use of hiera
01:28 blkperl joined #puppet-openstack
01:28 mgagne bodepd: some sort of hieradata builder?
01:29 fandikurnia01 joined #puppet-openstack
01:34 mgagne bodepd: https://etherpad.openstack.org/puppet-openstack-refactor
01:36 mgagne bodepd: trying to write-down the general outline of what a refactor would mean
01:42 ianw mgagne: mind if i add some (probably wrong) specifics on the rpc backend bit?
01:42 mgagne ianw: go ahead, etherpad is there for that =)
01:42 bodepd mgagne: I'm out for now. I can look at it tomorrow
01:43 mgagne bodepd: ok, thanks for the chat. lot of work ahead of us =)
01:50 mgagne ianw: people are moving to oslo components right? are people copying a version into their projects or is a global/common lib used?
01:51 ianw from what i've seen there's a lot of copying happening
01:51 mgagne ianw: could there be incompatible changes between versions? Which would require us to handle difference between projects using different versions?
01:52 ianw mgagne: almost certainly :)
01:52 mgagne ianw: I'll put that in the category "challenges"
01:52 ianw i just mean, even if not intended it's going to happen when projects async copy in components i guess
01:55 mgagne pardon my later of understanding: current idiom means current situation?
01:56 mgagne lack*
01:56 mgagne ianw: ^
01:57 ianw mgagne: yes, i'm just trying to explain a bit how the config files are currently generated
01:58 ianw s/idiom/pattern i guess
02:03 mgagne ianw: who the hell ninja commented our document lol
02:04 fandikurnia01 joined #puppet-openstack
02:07 mgagne ianw: I have to drive and get my girlfriend. I might not be back for a while. Thanks your help! I'll continue later.
02:08 ianw mgagne: np, it's a good start :)
02:08 mgagne ianw: it's taking shape, I like that. less talking, more writing. =)
02:17 ari joined #puppet-openstack
03:02 e1mer joined #puppet-openstack
03:43 mgagne joined #puppet-openstack
03:43 mgagne joined #puppet-openstack
03:45 mgagne1 joined #puppet-openstack
03:45 mgagne1 joined #puppet-openstack
03:58 fandikurnia01 joined #puppet-openstack
04:16 fandikurnia01 joined #puppet-openstack
05:32 michchap joined #puppet-openstack
07:06 bauzas hello folks
07:08 bauzas yesterday, I spotted a bug in heat module
07:08 bauzas ianw: you there ?
07:24 ianw bauzas: hi
07:25 bauzas hi ianw
07:25 bauzas :)
07:25 ianw that should have been fixed (i'm not sure how it was ever working ... )
07:25 bauzas could you please look at : https://bugs.launchpad.net/puppet-heat/+bug/1224827
07:26 bauzas the trunk is still having the heat::db class called : https://github.com/stackforge/puppet-heat/blob/master/manifests/engine.pp#L38
07:26 ianw yeah, it hasn't merged, but i added the review to the bug
07:26 bauzas ?
07:26 bauzas weird
07:27 bauzas let me glance the review
07:27 bauzas nah, can't see
07:29 bauzas ok, will rollback to the closest previous merge
07:30 bauzas and cherry-pick the heat::keystone patch
07:30 ianw yeah, or git review -d will get that change in for you too
07:30 bauzas don't know this
07:30 bauzas ah
07:31 bauzas for cherry-picking you mean ?
07:31 bauzas yup, i know then :)
07:31 ianw it just grabs the change from geritt into a branch, then you can cherry-pick it
07:33 bauzas gotcha
07:36 e1mer joined #puppet-openstack
07:42 mmagr joined #puppet-openstack
07:47 e1mer joined #puppet-openstack
08:00 e1mer joined #puppet-openstack
11:44 fvollero Hi fellas, did someone here tried rspec-system ?
12:25 michchap joined #puppet-openstack
12:33 dprince joined #puppet-openstack
12:39 michchap joined #puppet-openstack
12:59 morazi joined #puppet-openstack
13:23 openstackgerrit Sandro Mathys proposed a change to stackforge/puppet-horizon: Catch up with upstream's changes to local_settings.py  https://review.openstack.org/46491
13:44 prad joined #puppet-openstack
13:55 dmsimard joined #puppet-openstack
13:58 dmsimard1 joined #puppet-openstack
14:25 lnxnut_ joined #puppet-openstack
14:32 openstackgerrit Sandro Mathys proposed a change to stackforge/puppet-horizon: Catch up with upstream's changes to local_settings.py  https://review.openstack.org/46491
14:34 otherwiseguy joined #puppet-openstack
14:38 openstackgerrit Sandro Mathys proposed a change to stackforge/puppet-horizon: Catch up with upstream's changes to local_settings.py  https://review.openstack.org/46491
14:42 openstackgerrit Sandro Mathys proposed a change to stackforge/puppet-horizon: Catch up with upstream's changes to local_settings.py  https://review.openstack.org/46491
15:50 ari_ joined #puppet-openstack
15:59 otherwiseguy joined #puppet-openstack
16:11 marun joined #puppet-openstack
17:13 openstackgerrit François Charlier proposed a change to stackforge/puppet-keystone: Enable serving keystone from apache mod_wsgi  https://review.openstack.org/29059
17:13 ari joined #puppet-openstack
17:15 openstackgerrit A change was merged to stackforge/puppet-openstack: Add OVS provider network params to all.pp  https://review.openstack.org/46345
17:22 openstackgerrit François Charlier proposed a change to stackforge/puppet-keystone: Enable serving keystone from apache mod_wsgi  https://review.openstack.org/29059
17:38 mmagr joined #puppet-openstack
17:58 bodepd fvollero: I've been meaning to.
17:59 bodepd fvollero: I know the author really well :)
17:59 bodepd fvollero: but I wasnt sure how useful it was for my general use case of installing openstack and running tempest
17:59 bodepd fvollero: I would be extremely interested in seeing your progress if you start to put something together
18:01 fvollero bodepd: I was thinking about it in conjunction to jenkins to test the modules in different environments/systems and i discovered it just today thank to the puppetlabs podcast
18:02 bodepd fvollero: I'm definitely interested
18:02 bodepd fvollero: It would be really nice to be able to validate individual success criteria as well
18:03 fvollero fvollero: me too, very curious to see how it work :)
18:03 bodepd fvollero: ie: these ports are up, these logs have no errors, puppet has no failures
18:03 fvollero rofl
18:03 fvollero bodepd: me too, very curious to see how it work :)
18:03 bodepd fvollero: I do that all of the time. My fingers and brain don't always agree on what to do
18:03 fvollero bodepd: exactly for that, i've to run all the time a bash script to check such things!
18:04 fvollero bodepd: our work would be way more lighter!
18:05 fvollero bodepd: i'll investigate further in the weekend :) hopefully on monday i'll be able to give you some feedback :)
18:05 fvollero|gone have a great weekend everyone!
18:36 morazi joined #puppet-openstack
19:21 otherwiseguy joined #puppet-openstack
19:32 technolo-g joined #puppet-openstack
19:46 lnxnut joined #puppet-openstack
20:21 otherwiseguy joined #puppet-openstack
20:26 morazi joined #puppet-openstack
20:33 ari_ joined #puppet-openstack
21:00 ari joined #puppet-openstack
21:08 ari_ joined #puppet-openstack
23:53 badiane_ka joined #puppet-openstack
23:59 lnxnut joined #puppet-openstack

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