16:01:07 #startmeeting tacker 16:01:09 Meeting started Thu Oct 8 16:01:07 2015 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:12 The meeting name has been set to 'tacker' 16:01:21 #topic Roll Call 16:01:29 o/ 16:01:30 who is here for tacker mtg ? 16:01:35 o/ 16:01:35 o/ 16:01:48 o/ 16:02:03 gud' morning co-horts! 16:02:29 lets start then.. 16:02:34 #topic Agenda 16:02:43 #link https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_Oct_8.2C_2015 16:02:50 #topic Announcements 16:03:35 A doodle poll is up for Mitaka summit developer meetup time slot.. 16:03:38 #link http://doodle.com/poll/yv7mr45z2qeyhtrq 16:04:20 Please pick a slot that you can make it .. hint: there is no main conference events on Friday :) 16:05:01 nothing else from my side .. 16:05:18 any general heads up from anyone ? 16:05:42 sridhar_ram: do you want to update on the gate jobs enabled? 16:06:10 sripriya: sure 16:06:14 Hi Everyone 16:06:34 sridhar_ram: probably also let us know which is voting and non voting...? 16:07:30 folks - we now have many gate tests enabled for tacker repo - the check pipeline (when you upload a patchset) now runs pep8, py27 (unit tests) and dsvm-functional 16:07:51 Rajkumar_: hi! 16:08:25 dsvm-functional - will bring up a full functional devstack and run 'tox -e functional' which creates vnfd templates and starts vnfs .. 16:08:26 Just curious - Our tacker is on Liberty release? 16:08:45 Rajkumar_: lets touch on that in a bit 16:08:57 Sure. Thanks Sridhar.. 16:09:36 whenever you upload a patchset make sure jenkins posts SUCCESS for all these three test runs (pep8, py27, dsvm-functional) 16:10:40 dsvm-functional - is currently marked as non-voting - meaning it won't block us from merging. Cores can decide to give exceptions to merge a patchset even if dsvm-functional reports FAILURE (case-by-case) 16:11:04 once we get confidence, we will switch even that to voting ... 16:11:29 any questions on this ? 16:12:19 this is actually a big deal, we were hoping to get these for a long time... thanks sripriya (for py27) and santosh (for functional) ! 16:12:23 sridhar_ram: thanks for the update! 16:12:47 #topic Liberty Release Status 16:13:37 Rajkumar_: we are waiting for health-mon patchset and few critical bugs to fix before we can pull a stable/liberty 16:14:16 the plan is do a release by end of next week.. at least pull the release branch 16:14:29 Ok Sridhar. 16:14:31 Thanks 16:14:32 we still haven't decided on posting the tarballs in pypi 16:15:00 I could not see anything about tacker in the liberty release note 16:15:09 that was the reason asked about it 16:15:25 Rajkumar_: no, given we are not in governance tent we won't be listed there anyway.. 16:15:55 Hello Sridhar and team 16:16:24 ok. Do we have any release notes for tacker? 16:16:34 I'm anticipating after pulling stable/liberty we will still make dot releases ... for stability and other cross-project needs (like OPNFV-SFC needs Tacker-SFC in liberty) 16:16:35 Hi Shrinath 16:16:48 shrinathsuresh: hi there 16:17:02 Rajkumar_: no, but that's a good point.. 16:17:29 anyone here know the process to write a release notes for openstack projects ? 16:18:01 sridhar_ram: can we link the tar files of the 3 projects on tacker launchpad page? 16:18:14 Sridhar - we're not sure. but we can help if assistance is provided. 16:18:52 sripriya: I need to find that out .. will take an AI 16:19:29 meanwhile I would nominate the doc sub-team - sripriya & tbh - to help out here :) 16:20:29 #action - sridhar_ram to follow up with infra how to post tacker release tarballs 16:20:49 Sridhar - Please share us the details in mail and our team will help on release notes. 16:21:13 Rajkumar_: sure 16:21:31 Thanks 16:22:04 Rajkumar_: we are still looking for functional tests on parameterization from your side.. it will be great to get that in before stable/liberty pull (next week) 16:22:19 sridhar_ram: ack, doc sub-team hopefully grows :-) 16:22:31 yes Sridhar. am working on it. will send few test cases tomorrow 16:22:48 Sure Sridhar. I'll update it shortly 16:22:53 sripriya: I'll help as much I can :) 16:23:07 shrinathsuresh: Rajkumar_: thanks! 16:23:27 anything else related Liberty Release ? 16:24:12 #topic API Deprecation 16:25:10 folks here will know we still have the old servicevm based APIs (and in turn CLI cmds) for "device" and "device-template" 16:25:43 they have now been superceded by new MANO contructs like "vnf" and "vnfd" 16:25:57 it is time to start the deprecation process... 16:26:44 had a discussion with folks from unitedstack (in our #tacker channel) and they are open to switch over to the new MANO apis 16:27:28 I propose we start deprecation process in liberty - by adding the warning headers in API and CLI in liberty - and remove these API in mitaka 16:28:03 anyone have concerns removing it or removing it starting Mitaka cycle ? 16:28:11 sridhar_ram: does that mean we deprecate the servicevm standards from the API layer only or should it be deprecated in servicevm references too in the entire project? 16:28:52 sripriya: first order of business to remove it from our external interfaces - API layer 16:29:16 the rest is more a clean up and tech debt cleanups - IMO 16:29:21 what do you think? 16:29:46 sridhar_ram: sounds good, that affects tacker and python-tackerclient projects 16:30:08 sripriya: agree, it need to removed from both these projects 16:30:22 I'll also send a note on this to the ML 16:30:27 sridhar_ram: we still have the db layer untouched then 16:31:34 sripriya: we can be opportunistic there .. if we can rename our db schema away from "device*" tables we can still do it 16:31:52 but it is not a must for deprecation to complete 16:32:28 sridhar_ram: should be doable, but again it would require db-migration after deprecation on master 16:33:28 sripriya: true, I'd rather do it sooner than later before more tacker adoption w/ real operators 16:33:43 sridhar_ram: agree 16:33:51 lets move to bugs (the topic of the week!) 16:33:56 #topic Bugs 16:34:26 as called out last week ... we need some serious bug squashing... 16:35:09 I've marked some of the critical ones using the tag "liberty-critical" .. 16:35:21 #link https://bugs.launchpad.net/tacker/+bugs?field.tag=liberty-critical 16:36:09 there are bunch of them related to health-monitoring and VNF states / cleanup on deletion 16:37:09 lets discuss few of them here... 16:37:11 https://bugs.launchpad.net/tacker/+bug/1503480 16:37:11 Launchpad bug 1503480 in tacker "Deletion of vnf having issues when doing in between Active-->Dead State" [Critical,New] 16:37:34 bobh: tbh: this seems to happen with monitoring code 16:38:06 after a successful respawn .. looks deletion is getting stuck in pending_delete 16:38:19 sridhar_ram: I tried to reproduce but could not 16:38:33 sridhar_ram: if the delete is happening while the respawn is in progress that may cause the problem 16:38:58 sridhar_ram: we might need to have a way to 'lock' the status during the respawn and not allow the delete until it completes 16:39:20 sridhar_ram: new state :-) 16:39:28 bobh: make sense, mark it device busy 16:39:48 bobh: nooooo - no new state allowed ! 16:39:51 sridhar_ram: :-) 16:40:09 sridhar_ram: I'll take a look at it 16:40:11 bobh: just ask the operator to come back later after a coffee! 16:40:33 bobh: cool, thanks! 16:41:10 https://bugs.launchpad.net/tacker/+bug/1501079 16:41:10 Launchpad bug 1501079 in tacker "tacker mgmt-driver support in vnf-create flow" [Medium,New] - Assigned to bharaththiruveedula (bharath-ves) 16:41:27 tbh: who have any further questions on this ? 16:41:35 s/who/you/ 16:42:05 sridhar_ram: Iwant to test with vnf descriptors 16:42:52 sridhar_ram: I tried to look at the problem in vnf create, in config_device method, it is unable to get config params 16:43:24 sridhar_ram: so, if we correct that, I think we can solve that issue 16:44:18 tbh: yeah, there might be two part problem here .. vnf-config correctly carried over from CLI / horizon to the backend 16:45:09 (2) the backend .. shd wait for ... boot_wait and try to invoke the configured mgmt driver 16:45:17 sridhar_ram: I tried with hrizon, I can see the config params reaching the backend 16:46:00 tbh: I don't think horizon has that support yet.. 16:46:52 Sridhar - I've few critical requirements. This is very much needed and learnt from our experience 1) Consolidated Physical & Virtual Network inventory (Standard reconciliation is needed if not consolidated) 2) Interoperability of Tacker Orchestrator across different versions of cloud (hybrid cloud also if possible) 3) Workflow management & dynamic policy management (Convergence workflow & Jboss Drools if possible) 4) 16:47:01 we 16:47:12 we'll discuss it in next topic 16:47:12 sridhar_ram: config params from yaml is passed to create_vnf 16:47:37 sridhar_ram: and config_device is waiting for 30 sec https://github.com/stackforge/tacker/blob/master/tacker/vm/plugin.py#L243 16:47:38 bobh: btw, I'm realizing the boot_wait ... monitoring delay .. needs to apply for other things as well like mgmt_driver 16:47:56 Rajkumar_: great. User requirements. 16:48:40 without those functionalities zero touch NFV orchestration is tedious 16:49:00 Rajkumar_: "different versions of cloud"? You mean supporting something other than OpenStack? 16:49:07 tbh: config and params are 2 different attributes supported from cli 16:49:15 sridhar_ram: agree - there may be a need to monitor different things and trigger config after one of them 16:49:23 tbh: horizon only supports param attribute support 16:49:37 initially within openstack (juno,kilo, liberty) and later like AWS..etc 16:49:47 tbh: sripriya has some insights into these code flows .. perhaps you can sync up to make some progress 16:50:08 Rajkumar_: s3wong: folks - hang in there.. lets have one mtg please 16:50:10 tbh: we can discuss it on tacker IRC after the meeting 16:50:11 sripriya: okay using --config ? 16:50:18 tbh: yes 16:50:23 sripriya: sridhar_ram sure 16:50:28 Sure Sridhar.. 16:50:38 any other high priority bug ? 16:50:39 10 minutes left 16:51:13 looks for folks to flag critical / high priority if you see them! 16:51:27 sridhar_ram: may be some global issues with delete 16:51:46 bobh: good point.. there is a nasty one there 16:51:49 * sridhar_ram looking up... 16:52:20 here it is ... 16:52:22 https://bugs.launchpad.net/tacker/+bug/1475798 16:52:22 Launchpad bug 1475798 in tacker "Deleting a VNF in PENDING_CREATE state does not delete the associated VM" [Critical,In progress] - Assigned to vishwanath jayaraman (vishwanathj) 16:52:34 vishwana_: are you actively working on this ? 16:52:36 sridhar_ram: definitely related 16:52:42 yes 16:53:03 thanks! 16:53:16 sridhar_ram: vishwana_ there is no validation done when executing delete, and different things need to happen based on the current state 16:53:59 bobh: yeah.. needs some locks here! 16:54:32 alright .. lets get the floor back to Rajkumar_ & s3wong! 16:54:37 #topic Open Discussion 16:54:58 Rajkumar_ put up some user requirements about 5 minutes ago 16:55:13 1) Consolidated Physical & Virtual Network inventory (Standard reconciliation is needed if not consolidated) 16:55:18 Rajkumar_: s3wong: agree, quite useful list there! 16:55:49 Inventory is big issue in our experience 16:56:20 Rajkumar_: there are good inputs.. lets put this in the Mitaka roadmap discussion and talk it out... 16:56:38 Rajkumar_: any one from Comcast going to attend Tacker meetups in Tokyo ? 16:56:52 Sridhar - Do you want us to schedule meeting for those requirement? 16:57:09 Sridhar - Let me check with my team and update you in mail if any of us coming there 16:57:18 Rajkumar_: I'm wondering if anyone from your team going to Mitaka summit .. 16:57:27 we plan to discuss many of these face 2 face 16:57:42 when is this happening and venue? 16:57:57 Oct. 27th, Tokyo 16:58:10 #link https://www.openstack.org/summit/tokyo-2015/ 16:58:19 * sridhar_ram 2 mins left 16:58:23 Sure. We'll update it shortly 16:58:37 Our informal design session will be on Friday? 16:58:38 if meeting is needed, I can send out invite for those req? 16:58:59 Rajkumar_: sure... I think it is fair to say we need more detail on the requirements 16:59:00 s3wong: doodle poll is out (discussed earlier in the mtg)... 16:59:31 Rajkumar_: can you stick in these requirements to Mitaka etherpad.. 16:59:51 #link https://etherpad.openstack.org/p/mitaka-tacker-design-summit 16:59:57 Sure 17:00:00 I'm doing it 17:00:07 times up... 17:00:13 tty all next week! 17:00:17 bye 17:00:22 #endmeeting