18:01:51 <SumitNaiksatam> #startmeeting networking_policy 18:01:52 <openstack> Meeting started Thu Mar 5 18:01:51 2015 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:01:55 <openstack> The meeting name has been set to 'networking_policy' 18:02:13 <SumitNaiksatam> #info agenda https://wiki.openstack.org/wiki/Meetings/GroupBasedPolicy#March_5th.2C_Feb_26th.2C_19th.2C_12th.2C_2015 18:02:38 <SumitNaiksatam> no announcements from me 18:02:48 <SumitNaiksatam> anyone like to share anything up front? 18:03:04 <SumitNaiksatam> #topic Bugs 18:03:27 <SumitNaiksatam> so as planned, we had the bug squashing day on Tuesday (March 2nd) 18:03:31 <SumitNaiksatam> #link https://wiki.openstack.org/wiki/GroupBasedPolicy/kilo-bug-squash#Pending_bugs_on_master_.28and_require_backport.29 18:03:55 <SumitNaiksatam> as you can see we fixed most of what we had planned to and managed to back port as well 18:04:08 <SumitNaiksatam> BIG thanks to everyone for participating 18:04:29 <SumitNaiksatam> there was an issue with the gate starting that evening 18:04:59 <SumitNaiksatam> and at one point mageshgv and I noticed that at least our stable branch patches we kicked out of the zuul queue 18:05:49 <SumitNaiksatam> so if you are noticing that your patch has not been voted on for a long time by Jenkins, always check status.openstack.org/zuul 18:05:53 <SumitNaiksatam> to see if its in the queue 18:06:13 <SumitNaiksatam> if not, you can retrigger it by putting “recheck” on the patch 18:06:42 <SumitNaiksatam> also one of the patches was in the merge queue and got bumped out 18:06:58 <SumitNaiksatam> so I had to do a +A again to get it back into the merge queue 18:07:05 <SumitNaiksatam> recheck was not enough 18:07:14 <rkukura> reverify? 18:07:45 <SumitNaiksatam> rkukura: i dont think reverify is supported anymore, at least i did not find it on the wiki page (but i might not be looking at the right place) 18:08:16 <SumitNaiksatam> rkukura: i think, reverify and recheck probably mean the same thing now, but i dont want to misinform folks here 18:08:46 <SumitNaiksatam> lastly, if your patch is in the zuul queue, it might take a long time for you to get back the jenkins vote, since the queue has been pretty deep 18:09:30 <SumitNaiksatam> in general, on the proccess, please feel free to check on #openstack-gbp any time you run into issues like these (or you can directly ask on #openstack-infra) 18:09:52 <SumitNaiksatam> the list of bugs that we fixed on the bug squash is prioritized subset of the pending buga 18:09:54 <SumitNaiksatam> *bugs 18:10:03 <SumitNaiksatam> we still need to take of the other pending bugs 18:10:34 <rkukura> I expect to have a patch for the last remaining bug squash bug, https://bugs.launchpad.net/group-based-policy/+bug/1416177, today 18:10:35 <openstack> Launchpad bug 1416177 in Group Based Policy " ptg wrongly created with null subnet, cannot be deleted either" [Medium,In progress] - Assigned to Robert Kukura (rkukura) 18:10:47 <SumitNaiksatam> rkukura: sweet! 18:11:16 <SumitNaiksatam> per rkukura’s suggestion last week, if we are able to get most of the prioritized bugs fixed and backported, we will do a new stable/juno release end of this weekend 18:11:42 <SumitNaiksatam> any comments/questions on the above? 18:12:02 <SumitNaiksatam> is jishnu here? 18:12:23 <rkukura> he’s logging in now 18:12:28 <SumitNaiksatam> rkukura: ah okay 18:13:07 <SumitNaiksatam> so jishnu has a uploaded a scenario testing library to pypi for running GBP tests 18:13:38 <SumitNaiksatam> this is an interim step while we enable these tests via the upstream gate 18:13:58 <SumitNaiksatam> i was hoping jishnu would have been able to summarize how to use this library 18:14:15 <SumitNaiksatam> ah right on cue 18:14:18 <jishnu> hi 18:14:32 <SumitNaiksatam> jishnu: hi, we were just discussing about your pypi lib 18:14:44 <SumitNaiksatam> jishnu: can you summarize for the team here how to use it? 18:14:47 <jishnu> Sure 18:15:01 <jishnu> The are the steps: 18:15:13 <jishnu> pip install noirogbptests 18:16:54 <SumitNaiksatam> jishnu: still there? 18:16:56 <jishnu> pkg will wil be installed at 18:16:57 <jishnu> /usr/lib/python2.7/site-packages/noirogbptests/ (for RHEL) 18:17:05 <SumitNaiksatam> ok 18:17:36 <jishnu> hi 18:17:49 <jishnu> Looks like my previous msgs did not appear 18:18:04 <jishnu> Let me repeat again .. sorry folks 18:18:12 <jishnu> pip install noirogbptests 18:18:21 <jishnu> pkg will be installed in: 18:18:31 <jishnu> /usr/lib/python2.7/site-packages/noirogbptests/ (for RHEL) 18:18:41 <jishnu> /usr/local/lib/python2.7/dist-packages/noirogbptests/ (for Ubuntu) 18:19:03 <jishnu> # Usage: User can run each test-script or entire test-suite(suite_run.py) in any of two ways: # 1. If the default location of the package is appended to the $PATH # then executable files can be run rom anywhere # 2. The executable can be run from the default location # # Test Report: Depending on the location from where the suite is run, a file "test_reports.txt" # get created in that 18:19:44 <SumitNaiksatam> jishnu: ok good, perhaps its a good idea to post these instructions on a wiki page, and point from the main GBP wiki page 18:20:23 <jishnu> yes .. i shall put it there 18:20:47 <SumitNaiksatam> jishnu: thanks, sorry to put you on the spot ;-) 18:20:53 <jishnu> folks: it takes 5 mins 10-12 secs to finish 52 functonal testcases 18:21:04 <SumitNaiksatam> jishnu: ok good to know 18:21:28 <SumitNaiksatam> these are end to end tempest scenario like tests 18:21:40 <SumitNaiksatam> jishnu: so the requirement is that devstack be installed before running these, right? 18:21:41 <ivar-lazzaro> are the testcase defined on the library itself? 18:22:00 <ivar-lazzaro> should we include them upstream instead? 18:22:01 <jishnu> yes.. devstack needs to be installed first 18:22:13 <banix> could these become part of the tree? 18:22:20 <banix> is that the plan eventually? 18:22:36 <SumitNaiksatam> banix: yes 18:22:38 <jishnu> As Sumit mentioned, I will put the instructions clearly in the Wiki.. 18:22:51 <SumitNaiksatam> ivar-lazzaro: yes, we should upstream this 18:22:58 <SumitNaiksatam> so banix ivar-lazzaro on that - 18:23:29 <SumitNaiksatam> i have started the process of triggering an upstream gate job to run GBP functional tests 18:23:35 <SumitNaiksatam> have posted two patches: 18:23:42 <SumitNaiksatam> #link https://review.openstack.org/#/c/161511/ 18:23:49 <SumitNaiksatam> that one is in infra 18:23:57 <SumitNaiksatam> and: #link https://review.openstack.org/#/c/161532/ 18:23:59 <SumitNaiksatam> in GBP 18:24:47 <banix> nice 18:24:59 <ivar-lazzaro> good! 18:25:40 <ivar-lazzaro> just wondering, do they give you the ability to define arbitrary topologies? Or is it all single node testing? 18:26:00 <yapeng> great, how to run the functional test, is there any instruction? 18:26:38 <jishnu> its is topology agnostic.. 18:28:51 <SumitNaiksatam> sorry i got disconnected 18:29:08 <SumitNaiksatam> any other bugs we need to discuss in this meeting? 18:29:41 <SumitNaiksatam> #topic Re-factor Group Based Policy with Neutron RESTful APIs 18:29:48 <SumitNaiksatam> Yi: yapeng: over to you 18:29:54 <SumitNaiksatam> how is this coming along? 18:30:16 <Yi> almost done.. 18:30:32 <Yi> the client and api are ready 18:30:33 <jishnu> As the datapath testcases are getting developed, the approach taken is: presence of two end-points agnostic of their location(so that you can run the same set of tests in two nodes or single node) 18:30:37 <SumitNaiksatam> Yi: wow! :-) 18:30:57 <SumitNaiksatam> jishnu: ok good to know 18:30:59 <Yi> we are working on the refactoring the RMD and UT cade 18:31:19 <SumitNaiksatam> Yi: you ran into some issues with the latest merged patches? 18:31:59 <Yi> it was done. -- As some new test cases were introduced in recent bug fixes, we are working on patching them 18:32:08 <SumitNaiksatam> Yi: ok 18:32:11 <SumitNaiksatam> so in terms of review, we are discussing the following chaing of patches: 18:32:16 <SumitNaiksatam> #link https://review.openstack.org/#/c/159725 18:32:23 <SumitNaiksatam> #link https://review.openstack.org/#/c/156776/ 18:32:32 <SumitNaiksatam> #link https://review.openstack.org/#/c/156856/ 18:33:10 <Yi> SumitNaiksatam: correct 18:33:14 <SumitNaiksatam> i did one pass of quick reviews over the former two patches 18:33:15 <yapeng> the first link is ready for review. 18:33:21 <Yi> the first two are ready 18:33:24 <SumitNaiksatam> yapeng: ah good to know 18:33:35 <Yi> the last one is almost... 18:33:50 <SumitNaiksatam> yapeng: yi, apart from the UTs, have you tested this in a devstack setup? 18:34:01 <Yi> yes, we did. 18:34:10 <SumitNaiksatam> Yi: sweet!! 18:34:22 <yapeng> yes, last week I did test against devstack 18:34:31 <SumitNaiksatam> yapeng: nice! 18:34:46 <Yi> I have not got a chance to test the external segment in devstack 18:34:54 <Yi> but otherwise, it seems fine 18:35:02 <SumitNaiksatam> Yi: okay, ivar-lazzaro and I can help you with that 18:35:21 <Yi> SumitNaiksatam: ivar-lazzaro: Thanks! 18:35:23 <SumitNaiksatam> would really appreciate if the rest of the team can help review these patches 18:35:39 <rkukura> SumitNaiksatam: will do 18:35:47 <SumitNaiksatam> Yi: and yapeng have been working pretty hard on this, and we owe them some good feedback at the earliest 18:35:51 <SumitNaiksatam> rkukura: great, thanks! 18:36:00 <Yi> Will do one more round test in devstack once we finish pacthing the RMD and UT. 18:36:04 <mageshgv> SumitNaiksatam: sure, will review these 18:36:10 <SumitNaiksatam> mageshgv: thanks 18:36:25 <SumitNaiksatam> yapeng: Yi: do you have any blockers at this point? 18:36:31 <Yi> rkukura: mageshgv: Thanks! 18:36:48 <Yi> SumitNaiksatam: So far so good 18:36:51 <yapeng> SumitNaiksatam: not at this moment 18:37:25 <SumitNaiksatam> Yi: yapeng: great to hear that, so lets focus our attention on reviewing and testing these 18:37:39 <SumitNaiksatam> any other questions for Yi and yapeng at this point? 18:37:39 <yapeng> SumitNaiksatam: ok 18:37:44 <Yi> great! 18:38:23 <SumitNaiksatam> yapeng: yi: thanks much for your work on this! 18:38:31 <SumitNaiksatam> #topic Floating IP support 18:39:11 <SumitNaiksatam> there were lots of good discussions on this since rkukura is in town, and some of us were meeting 18:40:07 <SumitNaiksatam> most of the discussions were sparked by rkukura’s comments on the latest patch set: 18:40:20 <SumitNaiksatam> #link https://review.openstack.org/#/c/157298/2/specs/kilo/gbp-floating-ip-support.rst 18:41:05 <SumitNaiksatam> at this point i think most of us are on the same page in terms of understanding where we want to be in terms of trying to capture the user intent 18:42:14 <SumitNaiksatam> and I think that the current spec puts us on a patch towards that, although might not entirely be the case (the explicilt use of the network service policy is a sticky point) 18:42:34 <SumitNaiksatam> that said, our current plan is to use this spec as a basis for the first iteration of implementation 18:42:53 <SumitNaiksatam> we will no doubt learn with that, and we can revise accordingly 18:43:01 <SumitNaiksatam> that said, over to you mageshgv 18:43:44 <mageshgv> SumitNaiksatam: True, Capturing the intent exactly is the important aspect here 18:44:59 <mageshgv> It might help in getting a clear picture when we can try out a patch for this functionlity 18:45:39 <Yi> SumitNaiksatam: mageshgv: regarding the intent, I do think in the case of FIP, it's to allow an endpoint to be accessed from external environment 18:46:13 <Yi> i.e., the external users could be the initiator of the session 18:46:37 <mageshgv> Yi: Yes, that is indeed the usecase for this feature 18:47:19 <Yi> while the intent, for the non-FIP NAT case, is to allow endpoint to access external ... 18:47:38 <Yi> the difference is "who is the initiator"... 18:48:39 <mageshgv> Yi: are you referring to nat_pool when you mean non-FIP NT case ? 18:48:52 <Yi> mageshgv: yes.. 18:49:02 <SumitNaiksatam> sorry, folks having problems with my client again 18:50:13 <mageshgv> Actually I too wanted to discuss this with ivar-lazzaro and SumitNaiksatam 18:51:04 <SumitNaiksatam> mageshgv: my apologies, i was typing not realizing that what i was typing was not geting relayed 18:51:10 <SumitNaiksatam> and i missed some of the context here too 18:51:21 <SumitNaiksatam> i will go back and check the logs once the meeting is done 18:51:35 <SumitNaiksatam> #chair rkukura mageshgv ivar-lazzaro 18:51:36 <openstack> Current chairs: SumitNaiksatam ivar-lazzaro mageshgv rkukura 18:51:47 <SumitNaiksatam> in case i drop off again 18:52:17 <Yi> mageshgv: do we want to discuss here? 18:52:18 <SumitNaiksatam> mageshgv: are we at a logical point in this discussion where we can take this offline? 18:52:18 <ivar-lazzaro> Yi, mageshgv: nat pool is intended for FIP cases 18:52:34 <SumitNaiksatam> i mean onto #openstack-gbp 18:52:36 <ivar-lazzaro> Yi, mageshgv: non-FIP cases are in the PAT domain imp 18:52:46 <ivar-lazzaro> s/imp/IMO 18:52:53 <mageshgv> SumitNaiksatam: ok, that will be fine 18:53:03 <SumitNaiksatam> i want to take a few minutes to discuss the GBP project in the open discussion 18:53:20 <SumitNaiksatam> mageshgv: thanks 18:53:38 <Yi> ivar-lazzaro: I would agree, but there seems some inconsistency about terms used in openstack regard nat/pat/FIP/snat/dnat 18:54:03 <SumitNaiksatam> mageshgv: to Yi’s point perhaps we can clarify in the spec 18:54:16 <ivar-lazzaro> Yi: well, PAT only mean one thing I think. And NAT pool could be renamed if we need to 18:54:49 <Yi> ivar-lazzaro: that's why I tried to avoid using nat-pool previously. Instead, using FIP/non-FIP. 18:54:55 <SumitNaiksatam> Yi: that said, we are not reinventing the floating IP (or NAT) definition, we intend to use it exactly as neutron uses it 18:55:18 <SumitNaiksatam> #topic Open Discussion 18:55:25 <Yi> SumitNaiksatam: I agree. but we need to make sure all of us on the same page.:-) 18:55:33 <SumitNaiksatam> Yi: absolutely 18:56:28 <SumitNaiksatam> i had one item i wanted to bring up - as a team it seems that we are in agreement that we want to take the step towards openstack “projectification” per the new “big tent” policy 18:56:38 <SumitNaiksatam> *governance policy 18:57:04 <SumitNaiksatam> everybody mostly in agreement on this? 18:57:10 <ivar-lazzaro> +1 18:57:18 <yapeng> +1 18:57:31 <mageshgv> +1 18:57:47 <SumitNaiksatam> ivar-lazzaro: yapeng mageshgv: great 18:57:50 <banix> +1 18:57:58 <SumitNaiksatam> banix: great 18:58:01 <Yi> +1 18:58:12 <LouisF> +1 18:58:17 <SumitNaiksatam> towards that end - here is the proposed mission statement for the project: 18:59:05 <SumitNaiksatam> “To provide a policy framework consisting of declarative, intent-based abstractions for cloud infrastructure coupled with an enforcement engine designed to enable scalable automation.” 18:59:33 <SumitNaiksatam> sound okay to everyone? 18:59:40 <banix> sounds reasonable 18:59:44 <SumitNaiksatam> banix: okay 18:59:46 <Yi> +1 18:59:56 <SumitNaiksatam> this is of course in no way set in stone 19:00:18 <SumitNaiksatam> projects keep evolving, so can be adjusted as we go along 19:00:31 <SumitNaiksatam> anyway, let me know if you have further thoughts on any of this 19:00:42 <SumitNaiksatam> and we have the #openstack-gbp channel to discuss as well 19:00:48 <SumitNaiksatam> we are at the hour 19:00:50 <SumitNaiksatam> thanks all! 19:01:02 <Yi> thanks, SumitNaiksatam! 19:01:07 <SumitNaiksatam> reminder - kilo-2 deadline for us is March 16th 19:01:25 <SumitNaiksatam> bye! 19:01:29 <Yi> bye 19:01:33 <banix> thaks. bye 19:01:35 <mageshgv> bye 19:01:38 <SumitNaiksatam> #endmeeting