17:02:50 <sridhar_ram> #startmeeting tacker
17:02:50 <openstack> Meeting started Tue Nov 17 17:02:50 2015 UTC and is due to finish in 60 minutes.  The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:02:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:02:55 <openstack> The meeting name has been set to 'tacker'
17:02:59 <sridhar_ram> #topic Roll Call
17:03:05 <natarajk> hi
17:03:09 <sridhar_ram> who is here for tacker ?
17:03:13 <vishwanathj> o/
17:03:14 <sridhar_ram> natarajk: hi
17:03:31 <bobh> o/
17:03:41 <natarajk> o/
17:04:44 <sridhar_ram> lets get started...
17:04:54 <sridhar_ram> #topic Announcement
17:05:21 <sridhar_ram> Agenda #link https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_Nov_17.2C_2015
17:05:27 <tbh> o/
17:05:37 <sridhar_ram> rather lite today.. any open items welcome!
17:05:42 <sridhar_ram> tbh: hi there
17:06:04 <tbh> sridhar_ram, Hi
17:06:10 <brucet> hello
17:06:25 <santoshkumark> Hello all
17:06:25 <brucet> I did not see any new Blueprints this past week
17:06:26 <sridhar_ram> Sripriya is offically a new core member!
17:06:46 <sridhar_ram> brucet: hi there, hang on a bit.. we will get to it
17:06:52 <brucet> OK
17:07:09 <sridhar_ram> I believe Sripriya is traveling and not here today
17:07:32 <sridhar_ram> #topic Mitaka Blueprints and RFEs
17:08:18 <sridhar_ram> bobh: any quick update on the tosca-parser activities ?
17:08:56 <bobh> I created the tosca-parser BP and Sahdev has replied that they will review it soon.  I believe I need to create a corresponding heat-translator BP and one in tacker-specs too
17:09:28 <bobh> I've been working on the tosca-parser changes in the background, hopefully something will be working soon
17:10:00 <sridhar_ram> bobh: sounds good. provide gerrit links when available..
17:10:07 <bobh> will do
17:10:14 <s3wong> sorry, a bit late
17:10:34 <sridhar_ram> bobh: have you thought about how the tacker part of the BP will look like ?
17:11:44 <bobh> sridhar_ram: There are several pieces - (1) incorporating the new tosca-parser library, (2) changes to the existing tosca parsing code to handle the "new" and the "old" formats, and (3) incorporating the new heat-translator capabilities
17:12:08 <bobh> (2) could be an issue - there are significant differences between what we support today and the official TOSCA NFV Profile
17:12:36 <sridhar_ram> bobh: exactly .. that's what I'm curious about.. the level of feature parity between the two
17:13:23 <sridhar_ram> bobh: anyway, it is fair for you to write up your thoughts in the BP and have everyone here review & comment
17:13:26 <bobh> sridhar_ram: I need to do a full comparison but I think the differences may be significant
17:13:32 <brucet> What about existing heat resources? Anything needed there??
17:13:55 <bobh> brucet: Do you mean new resources?
17:14:05 <brucet> Yes. Heat resources
17:14:13 <sridhar_ram> bobh: the two stand out items for us to carry over is mgmt_driver and mon_driver to the new format
17:14:24 <brucet> Tosca translator >> Heat template, right??
17:15:16 <bobh> brucet: right - I think that the existing heat-translator supports (many) more Heat resources than our home-grown translator does today
17:15:42 <brucet> Oh. This is not built on existing Tosca >> Heat translator??
17:15:42 <bobh> brucet: we may find additional things we need in the heat-translator bp/spec after the tosca-parser work is done
17:16:16 <bobh> brucet: Existing tacker code generates heat from tosca "manually"
17:16:40 <sridhar_ram> brucet: at high level, what is currently in heat-translator is tosca-simple-profile --> heat .. for Mitaka we want to get to tosca-nfv-profile --> heat
17:16:44 <bobh> brucet: so by moving to the official tosca-parser/heat-translator codebase we will gain a lot of capabilities
17:16:55 <sridhar_ram> brucet: that is the scope of this effort, bobh correct if I'm wrong
17:17:00 <brucet> OK
17:17:01 <bobh> sridhar_ram: correct
17:17:06 <brucet> Got it, thx
17:17:18 <sridhar_ram> brucet: sure
17:18:00 <sridhar_ram> lets discuss more on tosca-parser once some BPs write up show up
17:18:18 <sridhar_ram> bobh: thanks for the update.. looks things chugging nicely here!
17:18:43 <sridhar_ram> tbh: I know you picked up Auto Flavor / Auto Network creation.
17:19:02 <sridhar_ram> tbh: can you provide your thoughts & updates ?
17:19:35 <tbh> sridhar_ram, I have seen the new rfe reported by you for flavor
17:19:50 <sridhar_ram> https://bugs.launchpad.net/tacker/+bug/1516193
17:19:50 <openstack> Launchpad bug 1516193 in tacker "Automatic Flavor Creation based on VNFD Template" [Medium,New] - Assigned to bharaththiruveedula (bharath-ves)
17:20:23 <sridhar_ram> tbh: I still need to write on for Auto Network creation :(
17:20:28 <tbh> sridhar_ram, but I think we will continue with automatic flavor creation once we get the done with tosca parser intergration?
17:20:29 <sridhar_ram> s/on/one/
17:21:15 <sridhar_ram> tbh: no, I would prefer for us not to put tosca-parser dependency for the things in flight
17:21:15 <tbh> sridhar_ram, I will write that
17:21:51 <tbh> sridhar_ram,  you mean to use existing tosca parse/heat translator for that?
17:22:43 <sridhar_ram> tbh: bobh: this is the part I need some clarification
17:23:04 <sridhar_ram> tbh: bobh: how do you envision tosca-nfv-parser work will help in this auto-flavor effort ?
17:23:35 <bobh> sridhar_ram: We need to determine if the existing heat-translator code can generate a heat template that auto creates a flavor, network, etc
17:24:05 <bobh> sridhar_ram: if it does/can, we will get that functionality for "free" when we move to using tosca-parser/heat-generator
17:24:27 <tbh> bobh, no, it wont create. It just maps vnf details -> flavor name
17:24:46 <bobh> tbh: ok, so the next question is where should that capability go?
17:24:50 <tbh> bobh, sridhar_ram  but that too heat-translator is not using nova client to get flavor details
17:25:17 <bobh> tbh: sridhar_ram I think we could argue that this is an enhancement to heat-translator?
17:25:39 <bobh> sridhar_ram: Since we want to get out of the heat generation business right?
17:26:04 <tbh> bobh, sridhar_ram  this is related BP https://blueprints.launchpad.net/heat-translator/+spec/match-custom-flavor
17:27:04 <tbh> bobh, yes, I think it must go in heat translator
17:27:40 <bobh> tbh: Agree - and if there are admin issues around creating custom flavors that will need to be addressed too - maybe Tacker is "admin"-enough that it's not a problem
17:27:41 <sridhar_ram> bobh: true, but flavor has lot more things riding on it .. it is not just cpu, mem. there are aspect of efficient VNF placement that goes into Nova flavor
17:28:16 <bobh> sridhar_ram: agree
17:28:44 <sridhar_ram> looks we need more data here.
17:29:21 <sridhar_ram> vishwanathj: you also need to get your efficient placement requirements into the mix and have it supported by what bobh and tbh are building
17:30:02 <vishwanathj> sridhar_ram, agree
17:30:51 <sridhar_ram> but this criss-cross dependency scares me a bit.. lets see if we can break these into smaller work items among us
17:31:19 <sridhar_ram> if tacker needs to be in the business of flavor create - say using vnfd-create
17:31:50 <sridhar_ram> that will give us some control to handle different performance level metadata
17:32:07 <bobh> sridhar_ram: that would be one solution - initial analysis during vnfd-create, but it doesn't prevent the available flavors from changing between vnfd-create and vnf-create
17:32:46 <sridhar_ram> bobh: true, in fact I'd suggest we SHOULD NOT rely on existing flavors in the nova flavor-list
17:33:12 <sridhar_ram> bobh: flavor shd always be created using custom uuid and live thru' the lifetime of the vnfd
17:33:35 <sridhar_ram> there is a section in tosca called host_properties: .
17:34:01 <bobh> sridhar_ram: so it might be that we want to put a layer between the tosca-parser output and the heat-translator input where we do flavor analysis and inject some custom flavor creation into the tree
17:34:32 <bobh> sridhar_ram: need to ensure that heat-generator will support creating custom flavors
17:34:36 <sridhar_ram> bobh: spot on.. that's a clear value that tacker can bring in
17:34:37 <tbh> sridhar_ram, bobh yeah, but I don't think new flavor for every vnfd, but analyse flavor details of tacker created flavors
17:36:03 <tbh> bobh, heat generator you mean tacker heat generator, or heat code?
17:36:10 <sridhar_ram> tbh: in trying to analyze existing flavors, we are mostly trying to be db efficient..
17:36:10 <bobh> tbh: Flavors are "cheap" - they don't use any resources by themselves, so I think it is reasonable to tie them to the VDU/VNFD
17:36:21 <sridhar_ram> tbh: I don't think that's an issue
17:36:34 <sridhar_ram> bobh: +1
17:36:50 <bobh> tbh: sorry - heat-translator - my brain keeps saying heat generator
17:36:51 <vishwanathj> heat-generator is expected to work only for the OpenStack VIM, right? if we support other VIMs, a different tosco to that particular VIM generator would be needed
17:37:01 <bobh> vishwanathj: correct
17:37:49 <sridhar_ram> vishwanathj: that's a good point.. current code intermingles tacker functions in heat-parser.
17:37:50 <tbh> agree
17:38:05 <sridhar_ram> hope we can clean that a bit and get a nice layer of separation
17:38:15 <sridhar_ram> bobh: no pressure ;-)
17:38:36 <bobh> sridhar_ram: yeah right
17:39:15 <sridhar_ram> tbh: once you've some data .. you can use the RFE bug to write up how you are planning to implement
17:39:32 <sridhar_ram> tbh: we can get to impl after that. fair ?
17:39:39 <tbh> sridhar_ram, sure
17:40:47 <sridhar_ram> tbh: quickly on network creation, the note I got from tosca folks is .. if network param indicates a "name" we should assume the network is already available (like the way it is today)
17:41:20 <sridhar_ram> if the network: is described as ip_addr/netmask/vlan .. that's the cue to do auto-network creation
17:42:12 <sridhar_ram> anything else to discuss auto-flavor / net area?
17:42:34 <tbh> sridhar_ram, I think even in second case we can specify name
17:42:39 <tbh> is that correct?
17:43:11 <sridhar_ram> tbh: sure.. it become more a well known name
17:43:35 <sridhar_ram> lets move on to the next topics
17:43:44 <sridhar_ram> #topic Bugs / Docs / Infra
17:44:14 <sridhar_ram> anyone (include tacker users) want to bring up any significant bugs to our attention ?
17:44:52 <santoshkumark> sridhar_ram i have a bug on monitoring, looking for bugid..
17:45:10 <sridhar_ram> santoshkumark: sure...
17:45:28 <sridhar_ram> one thing from my side that is bothering me is https://bugs.launchpad.net/tacker/+bug/1505468
17:45:28 <openstack> Launchpad bug 1505468 in tacker "Error status after creating a vnf - TypeError: <type 'NoneType'> can't be decoded" [High,New]
17:46:10 <sridhar_ram> this  shows up periodically with "Port in conflict vdu1_mgmt_port still in use" or something like that
17:46:39 <sridhar_ram> if anyone hits this issue please provide some data in the above bug
17:47:44 <sridhar_ram> santoshkumark: got the bug id ?
17:47:47 <santoshkumark> https://bugs.launchpad.net/tacker/+bug/1513972
17:47:47 <openstack> Launchpad bug 1513972 in tacker "vnf monitoring fails after multiple vnf create/delete operations" [Undecided,New]
17:48:37 <santoshkumark> this bug is seen when vnf creation goes to ERROR state,
17:48:43 <s3wong> sorry, need to drop off
17:49:19 <santoshkumark> Why will vnf create might be error: one reason is if number of vm limits are exceeded while vnf create..
17:50:04 <sridhar_ram> santoshkumark: in those situation i would expect the system to gracefully fail...
17:50:12 <santoshkumark> sridhar_ram: this bug is not a stopper..but can cause monitoring to fail..
17:50:44 <sridhar_ram> santoshkumark: looks a respawn failed due to resource limits exhaustion ?
17:51:15 <tbh> sridhar_ram, do we need to monitor, if the VNF goes to ERROR state?
17:51:16 <sridhar_ram> anyway, lets see if anyone have bandwidth to take it up..
17:51:26 <sridhar_ram> tbh: no
17:51:38 * sridhar_ram 10mins left
17:51:58 <santoshkumark> sridhar_ram: not sure if respawn..
17:52:16 <tbh> any comments on this bug patch https://bugs.launchpad.net/tacker/+bug/1503477 will be appreciated
17:52:16 <openstack> Launchpad bug 1503477 in tacker "Provide respawn limit for respawn failure policy" [Medium,In progress] - Assigned to bharaththiruveedula (bharath-ves)
17:52:30 <tbh> it's been so long
17:53:07 <bobh> tbh: I've looked at it a couple of times but never had time to do a thorough review
17:53:23 <bobh> tbh: I'll try to spend some time on it this week
17:53:36 <tbh> bobh, sure thanks
17:54:17 <sridhar_ram> tbh: we can take some of monitoring related enhancements after we make some progress in Mitaka priorities.
17:54:36 <tbh> sridhar_ram, okay
17:54:38 <sridhar_ram> I need another help from the team ..
17:54:59 <sridhar_ram> anyone here know how to waddle through openstack-infra project-config ?
17:55:23 <sridhar_ram> I could use some help to land tacker repos to pypi
17:56:09 <sridhar_ram> reach out if someone has some expertise here..
17:56:24 <sridhar_ram> moving on...
17:56:34 <sridhar_ram> #topic Open Discussion
17:57:00 <sridhar_ram> quick poll.. how many of here would like to install Tacker in a non-devstack environment ?
17:57:19 <bobh> o/
17:57:22 <tbh> o/
17:57:50 <natarajk1> o/
17:58:06 <sridhar_ram> okay.. there is multiple asks coming in the irc channel regardign this
17:58:36 <sridhar_ram> we should take it up as a work item then
17:58:50 <sridhar_ram> anything else ?
17:58:57 <bobh> sridhar_ram: Also seems like we need a way t validate that heat/nova are working
17:59:15 <bobh> Lots of questions come in because VNF 'failed to deploy'
17:59:24 <sridhar_ram> bobh: you mean a stack validate ?
17:59:32 * sridhar_ram almost out of time
17:59:33 <bobh> more like an openstack-validate
17:59:42 <bobh> is heat working?  is nova working?
17:59:53 <bobh> if the answer to either is no then tacker isn't going to work
17:59:57 <sridhar_ram> i think we should write a RFE and work through it ..
18:00:07 <sridhar_ram> bobh: quite valid..
18:00:15 <sridhar_ram> times up folks..
18:00:21 <sridhar_ram> talk to you next week
18:00:24 <vishwanathj> bye all
18:00:25 <sridhar_ram> bye all
18:00:29 <bobh> bye
18:00:35 <sridhar_ram> #endmeeting