17:30:29 #startmeeting networking_lib 17:30:30 Meeting started Wed Jan 20 17:30:29 2016 UTC and is due to finish in 60 minutes. The chair is dougwig. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:30:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:30:33 The meeting name has been set to 'networking_lib' 17:30:35 anyone here this morning? 17:30:46 sorry for overage dougwig 17:30:59 sam i already apologied :) 17:31:00 hi 17:31:08 absolutely no worries, sdake SamYaple . :) 17:33:16 pc_m: looks like it's you and me. 17:33:35 agenda: https://wiki.openstack.org/wiki/Network/Lib/Meetings 17:33:37 #link https://wiki.openstack.org/wiki/Network/Lib/Meetings 17:33:38 maybe a different time would be appropriate for the IRC? 17:34:08 not a bad thought. i think we could go later than a typical meeting, given who is usually involved. 17:34:31 * HenryG is here, had the time wrong 17:34:46 #topic Current Work Items 17:35:03 just to highlight stuff that needs eyes... 17:35:07 RPC routines and a 'legacy' section 17:35:07 https://review.openstack.org/#/c/267807/ 17:35:26 note that the above is about more than rpc. it introduces a section in the library for things we need to move, but aren't super fond of the interface. 17:35:45 Moving base db gunk 17:35:45 https://review.openstack.org/#/c/267214/ 17:36:05 I will post a new patch set on that today 17:36:06 and this one is henry's, setting up how to do the base model stuff. he says it's getting opinionated, but that's probably a good thing. 17:36:52 I tried to get rid of common_db_mixin, but that did not go well 17:36:59 kill it with fire. 17:37:34 unfortunately it is used in aweful ways :( 17:37:54 can we get rid of it, or are we stuck with mixins? 17:38:46 not sure yet 17:39:09 anything you want to chat about, or should we wait for the next patchset? 17:39:41 We can discuss in next patchset 17:39:45 ok 17:40:11 #topic Open Discussion 17:40:26 pc_m: let's talk about neutron tests first? 17:40:47 ok 17:41:00 My latest change fixed the API test. 17:41:11 for the hook, i noticed that you might also be able to use LIBS_FROM_GIT instead of doing the git clone 17:41:32 Did not fix the LB test (but I didn't check why it failed). 17:41:38 and for the non hook tests, i think we have several *very similar* jobs in the experimental queue that are actually controlled by hooks now, so we could cover that base there. 17:42:16 There are 4 tests, IIRC. 17:42:18 pc_m: lb test will likely have to be done here: https://review.openstack.org/#/c/242220/ 17:42:41 which ones? 17:43:02 * pc_m looking up 17:43:27 unless by lb you meant linuxbridge. it's hookless, so it'll just have to fail. 17:43:49 gate-tempest-dsvm-neutron-full 17:43:59 gate-grenade-dsvm-neutron 17:44:21 gate-tempest-dsvm-neutron-dvr 17:44:34 gate-tempest-dsvm-neutron-lib-linuxbridge 17:44:52 The LB one is gate-neutron-lbaasv2-dsvm-minimal 17:45:08 it has a gate hook, but failed. May be an unrelated failure. 17:45:27 , so -dvr and -linuxbridge are the same as -full, but with dvr and LB. i think we can ignore those two. -plus in the experimental queue is -full with a few extra things turned on, so it can provide coverage there. 17:45:27 I recall, before I went on PTO, some failure with that test. 17:45:44 since these are just pre-release sanity checks. 17:46:18 So do we want to make them non-voting (the two)? 17:47:37 no, because it's only in that one review where they fail. hmm, so this is about how to test that changes to neutron work against lib changes, before those changes are released. maybe we can flip this around and make a neutron-lib change that is dependent on the neutron fixes, once this merges: https://review.openstack.org/#/c/268240/ 17:49:46 if api and unit are passing, that's actually a pretty good sanity screen, too. 17:50:07 dougwig: not quite sure I'm following. 17:50:19 i think maybe we just update the process for making changes to note which tests you should expect to pass/fail ? 17:51:12 * pc_m sorry - I'm still in PTO mode 17:52:03 How does 268240 help? 17:52:17 pc_m: we're talking about item 4 in 'submission procedures' on this page https://wiki.openstack.org/wiki/Neutron/Lib . We don't need infra changes for that, just enough info from existing tests to know if neutron related lib changes are correct, before the lib changes ship (because it's easier to do them at the same time.) for that, we don't need a perfect 17:52:17 pass rate. right? 17:52:29 pc_m: no worries, let's chat offline. 17:53:21 dougwig: OK. Would like to clearly understand the procedure for migrating. 17:53:31 ok 17:53:39 anything else to chat about today? 17:53:42 Unfortuantely, I don't right now. 17:53:45 Do you want to move neutron-lib to using constraints based jobs? 17:54:02 I did this for VPN. Can do it for neutron-lib, if desired. 17:54:05 yes, but there is a change to how that's done as of yesterday's TC meeting. 17:54:22 no more -constraints jobs. just constrain the normal jobs. 17:54:49 ok. Any pointers to outcome of that meeting/changes?\ 17:55:05 * pc_m wondering how it affects the VPN changes done/ 17:55:13 meeting minutes are likely the only thing so far. they'll have a link to the resolution that was passed. 17:56:55 will look at them. 17:57:15 ok, anything else, or we get a half hour back ? 17:57:26 nope 17:57:29 nope 17:57:34 #endmeeting