17:01:07 #startmeeting service_chaining 17:01:08 Meeting started Thu Sep 24 17:01:07 2015 UTC and is due to finish in 60 minutes. The chair is LouisF. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:12 The meeting name has been set to 'service_chaining' 17:01:17 hello 17:01:18 o/ 17:01:20 hi 17:01:31 hi all 17:01:32 hi 17:01:36 hi 17:02:02 cathy is on a business trip, I will chair today 17:02:03 Hi 17:02:20 I put together an agenda wiki page 17:02:23 #link https://wiki.openstack.org/wiki/Meetings/ServiceFunctionChainingMeeting 17:02:31 LouisF: yes, she also mentioned it last time as well as sending out an email on ML 17:02:32 pcarver: thanks 17:02:53 I think that'll be a good place for people to put reminders of things to discuss that occur to them between meetings 17:03:27 pcarver: agree 17:03:33 +1 17:04:03 +1 17:04:40 I have split https://review.openstack.org/#/c/207251 into separate patches 17:05:18 See https://wiki.openstack.org/wiki/Meetings/ServiceFunctionChainingMeeting 17:05:33 there are 6 patches 17:05:38 o/ 17:06:09 vikram_: thanks for adding the list to the wiki 17:06:20 Thanks Louis, that'll be much more convenient to review in chunks 17:06:57 LouisF: ;) 17:07:06 i have made some updates to the patches based on vikram's comments and will continue to do so 17:07:30 please review and comment on these 17:07:49 LouisF: question, are you running tox before uploading patches? I've noticed a lot of Jenkins -1s 17:08:08 I'm wondering if Jenkins is catching issues that you could catch before uploading 17:08:43 pcarver: i have run tox on some but not all 17:09:34 Zuul/Jenkins seems better this week, but last week was running upwards of 13 hours backlog, so it's best to try to avoid uploading patches that are going to get -1'd by automated tests 17:10:41 vikram_: you mentioned adding dependencies in the patches to avoid unit test failures 17:11:15 LouisF: Yes, We need to do that for fixing failures. 17:11:22 vikram_: do you mean using Depends-on: ? 17:11:25 LouisF: only if the dependent patch is related to the child patch. 17:11:47 LouisF: yeah, I also noticed that for a chain of patches, none of them have DependsOn set, which looks a bit strange 17:12:09 s3wong: will add that 17:12:24 LouisF: We need to merge the changes on top of the parent patch 17:12:34 LouisF: If you want I can do it 17:12:43 vikram_: ok thx 17:12:44 LouisF: But for the unit test to pass you don't need "Depends on". This would be for merging. 17:13:21 Swami: +1 17:13:25 Swami: +1, We need to merge on top of the parent 17:13:43 vikram_: can you do the merge on top of the parent 17:13:56 LouisF: Yes... 17:14:04 vikram_: thanks 17:14:37 LouisF: I will fix all the patches which needs dependencies 17:14:49 vikram_: great! 17:15:27 LouisF: Did you raised patch for all the changes? 17:15:59 vikram_: what do you mean exactly? 17:16:55 LouisF: All the changes are up for review now? 17:17:21 LouisF: "https://review.openstack.org/#/c/207251" completely broken down? 17:17:32 LouisF: on something is left? 17:17:46 vikram_: The above shown link is broken 17:18:03 Swami: Ok.. 17:19:00 vikram_: all of the patch has been split into the 6 sub-patches 17:19:32 LouisF: Ok.. 17:20:11 mohan_: you have updated the CLI patch https://review.openstack.org/#/c/210008 17:20:36 yes 17:20:46 we spend most of the time this week in client cli testing with server changes and we posted tested client patch requsting everyone to review and share comments 17:20:47 vikram_, LouisF: so we should abandon https://review.openstack.org/#/c/207251? 17:21:07 s3wong: yes 17:21:24 +1 17:21:31 +1 17:21:33 mohan_: thanks, everyone please review that 17:22:04 LouisF, All: It's working well .. We have tested :) 17:22:18 vikram_: glad to hear it 17:23:18 pcarver: you made a comment on updating the spec so it is consistent wit the code 17:23:43 LouisF: yes 17:24:09 We can discuss if there are opinions, but I favor updating the .rst in the same commit as the code where you discover you need to alter the spec 17:24:24 pcarver: agree need to ensure consistency 17:24:43 i.e., if you're writing Python and notice something wrong with the spec, fix the spec and add it to your same commit as the correct code you're writing 17:24:54 +1 17:24:59 rather than creating a separate commit to correct the spec 17:25:43 If reviewers feel the spec is correct and the "correction" is wrong, that feedback can be handled in comments on the single review 17:26:01 pcarver: make sense.. 17:26:07 pcarver: +1 17:26:53 pcarver: +1 , yes thts the best way 17:27:36 mohan_: please ensure this for CLI changes 17:28:10 #agreed if there is a needed spec change, update spec and code in same patch update 17:28:31 vikram: yes i have few updates , will do that 17:29:07 there was also an update to the Horizon patch https://review.openstack.org/#/c/197056/ 17:29:40 but it has some pep8 errors 17:30:03 LouisF: Horizon i am not updated any code .. 17:30:13 we are in middle of integration testing with client code , we wanted to post patch after some basic testing with client and server cli code . 17:30:34 mohan_: https://review.openstack.org/#/c/197056/ has a recent post 17:31:01 mohan_: ps7 17:31:02 mohan_: https://review.openstack.org/#/c/197056/ can you please confirm why pep8 issues are there for thi 17:31:09 some commit messages get updated ! 17:31:27 mohan_: How is the horizon progress 17:31:59 mohan_: When you are planning to post the changes? 17:32:41 some integration testing going on , planning next week tentatively 17:32:58 mohan_: thanks 17:33:06 mohan_: Is the testing smooth or need any help? 17:34:14 vikram_ : as of now , no issues , will update if any help needs :) 17:34:30 mohan_: good to hear 17:35:00 mohan_: great! 17:35:05 any other items for discussion? 17:35:28 I need a couple of quick reviews of a doc change https://review.openstack.org/#/c/225254/ 17:35:53 Just adding some introductory paragraphs now that the documentation is linked to from docs.openstack.org 17:36:33 The other topic I have is on the Zuul/Jenkins config 17:36:47 specifically, do we want Python 3 jobs to run? 17:36:52 pcarver: go ahead 17:37:49 pcarver: where exactly is that configured? 17:37:53 And the other job that's not currently configured is translations. I assume it isn't the top priority, but we'll need to look into internationalization at some point 17:38:27 Refer to https://review.openstack.org/#/c/221589/ 17:38:39 That's the review where I added docs and PyPI publishing 17:39:25 but if you take a look at http://docs.openstack.org/infra/manual/creators.html#configure-zuul-to-run-jobs you'll see an example zuul/layout.yaml snipped 17:39:48 snippet that contains python3-jobs and translation-jobs 17:40:00 I left those out 17:40:21 We need to decide whether we want Jenkins/Zuul to enforce Python 3 17:40:51 Jenkins will post -1s for changes that aren't Python 3 compliant if I add that job. 17:41:00 pcarver: i'm not sure about that - anyone have guidance? 17:41:23 My thought is that if we want to aim for Python 3 compliance, better to do it from the start 17:41:29 pcarver: I feel we must have python 3 support 17:41:40 pcarver: +1.. 17:41:44 otherwise we'd possibly have to do a bunch of fixes later 17:41:49 pcarver: It's mandate in neutron now 17:42:06 vikram_: then we should do it 17:42:10 pcarver: All the sub projects are fixing issues 17:42:12 If we're agreed I'll put in a change to add it 17:42:19 +1 17:42:23 +1 17:42:24 pcarver: +1 17:42:51 +1 17:42:52 #agreed add python3-jobs 17:43:15 As far as translation-job is concerned, I assume we probably need to talk to someone about how we engage people with the language skills to create translations 17:43:53 At least, I'm assuming translation-jobs has to do with translating English language messages into the corresponding strings in all other supported languages 17:44:27 how is that handled elsewhere in neutron? 17:45:19 I'm not entirely sure, but I guess there must be a bunch of multilingual people who work on translating messages. I don't think this is software at all. 17:45:39 pcarver: I think we can leave that for later 17:45:41 I think there must be files of message strings that the underscore function uses 17:46:15 LouisF: agreed, not at all urgent 17:46:28 just something to keep in mind that we're going to need eventually 17:46:42 pcarver: agree 17:47:01 any other topics for discussion? 17:47:13 LouisF: I have one 17:47:22 vikram_: go ahead 17:47:26 LouisF: I think we 17:47:49 got to address Kyles comment on test sufficiency 17:47:57 vikram_: +1 17:48:27 we need to add api, full-stack tests 17:48:39 LouisF: + functional 17:49:05 LouisF: Existing patches doesn't have these.. 17:49:25 LouisF: Getting all the changes done by 10th Oct might be a risk 17:49:34 vikram_: right, I can look at api tests 17:49:35 LouisF: As we discussed in the last meeting 17:50:42 Here's the review to add python3-jobs #link https://review.openstack.org/227445 17:51:50 pcarver: thanks 17:52:18 LouisF: We also got to write the dedvstack changes 17:52:34 LouisF: I think mohan_ is doing that 17:53:03 mohan_: vikram_ thanks 17:53:08 vikram : ys 17:54:06 lets discuss testing in more detail next week 17:55:07 LouisF: ok 17:55:40 I think we made good progress this week - thanks everyone 17:56:43 bye 17:56:48 bye 17:56:52 bye 17:56:58 bye 17:57:01 #endmeeting