Tuesday, 2015-12-01

*** sridhar_ram has quit IRC00:02
*** sridhar_ram has joined #tacker00:03
*** sridhar_ram has quit IRC00:31
*** sridhar_ram has joined #tacker00:32
*** sridhar_ram has quit IRC01:03
*** sridhar_ram has joined #tacker01:10
*** u_kozat has quit IRC01:15
*** karimb has quit IRC01:20
*** openstackgerrit has quit IRC01:22
*** openstackgerrit has joined #tacker01:22
*** s3wong has quit IRC02:03
*** sripriya has quit IRC02:17
*** bobh has joined #tacker02:17
*** sridhar_ram has quit IRC02:30
*** yfujioka has quit IRC03:19
*** prashantD has joined #tacker03:52
*** prashantD has quit IRC04:32
*** sridhar_ram has joined #tacker05:00
*** sridhar_ram has quit IRC06:09
*** sripriya has joined #tacker06:16
*** sripriya_ has joined #tacker06:19
*** sripriya has quit IRC06:22
*** lhcheng has joined #tacker06:30
*** lhcheng_ has joined #tacker06:32
*** lhcheng has quit IRC06:35
*** gperal has joined #tacker07:08
*** gperal has quit IRC07:13
*** gperal has joined #tacker07:26
*** jhan1 has joined #tacker07:41
jhan1hello, I'm new to tacker. I just want to confirm whether tacker support NFVO currently? Could you help on it?07:42
sripriya_jhan1: hello there. welcome to tacker! tacker currently does not support the NFVO functionality yet. it supports VNFM feature with monitoring and management of VNFs. These features are currently in pipeline: https://etherpad.openstack.org/p/tacker-mitaka-priorities07:46
jhan1Got it. Thank you very much, sripriya_.07:50
sripriya_jhan1: sure feel free to catch us on IRC, and if you would like to learn more on Tacker, we have weekly IRC meetings. in fact we have one happening today at 1700 UTC on Tuesday. Feel free to join and introduce yourself07:53
jhan1Thanks for inviting. I will join you later.07:58
*** sripriya_ has quit IRC08:02
*** mdnadeem has joined #tacker08:13
*** zeih has joined #tacker08:13
*** karimb has joined #tacker08:48
*** karimb_ has joined #tacker09:05
*** karimb has quit IRC09:07
*** mbound_ has joined #tacker09:09
*** mbound_ has quit IRC09:27
*** mbound_ has joined #tacker09:35
*** lhcheng has joined #tacker09:55
*** lhcheng_ has quit IRC09:59
*** changzhi has joined #tacker10:16
*** mbound_ has quit IRC10:58
*** mbound_ has joined #tacker11:03
*** zeih has quit IRC11:07
*** changzhi has quit IRC11:48
*** gperal has quit IRC12:21
*** lhcheng has quit IRC12:33
*** lhcheng has joined #tacker12:34
*** mbound_ has quit IRC12:35
*** zeih has joined #tacker12:50
*** mbound_ has joined #tacker13:14
*** gongysh has joined #tacker13:39
*** mbound_ has quit IRC14:08
*** gongysh has quit IRC14:12
*** zeih has quit IRC14:12
*** amotoki_ has joined #tacker14:24
*** zeih has joined #tacker14:27
*** mbound_ has joined #tacker14:44
*** amotoki_ is now known as amotoki15:01
*** mbound_ has quit IRC15:06
*** mbound_ has joined #tacker15:09
*** zeih has quit IRC15:35
*** mah has joined #tacker15:49
*** mah has left #tacker15:49
*** mbound_ has quit IRC15:55
*** mbound_ has joined #tacker16:03
*** sripriya has joined #tacker16:26
*** vishwanathj has joined #tacker16:47
*** bobh has joined #tacker16:51
*** sridhar_ram has joined #tacker16:53
*** amotoki has quit IRC16:55
*** mbound_ has quit IRC16:55
*** u_kozat has joined #tacker16:57
*** mbound_ has joined #tacker16:59
*** santoshk has joined #tacker16:59
sridhar_ramGood morning Tackers!17:00
vishwanathjgood morning17:00
sridhar_ramhead over to #openstack-meeting-4 for our weekly17:00
u_kozatgood morning!17:00
*** brucet has joined #tacker17:01
*** brucet has quit IRC17:03
*** brucet has joined #tacker17:03
*** masterbound has joined #tacker17:04
*** mbound_ has quit IRC17:08
*** brucet has quit IRC17:11
*** brucet has joined #tacker17:12
trozetsridhar_ram: i just joined the meeting room, sorry im late17:12
sridhar_ramtrozet: no worries, we haven't started sfc yet17:13
*** lhcheng_ has joined #tacker17:17
*** s3wong has joined #tacker17:18
*** lhcheng has quit IRC17:20
*** brucet has quit IRC17:23
*** brucet has joined #tacker17:24
*** ksantoshk has joined #tacker17:26
*** santoshk has quit IRC17:29
*** lhcheng_ has quit IRC17:29
*** lhcheng has joined #tacker17:29
*** prashantD has joined #tacker17:32
*** karimb_ has quit IRC17:37
*** masterbound has quit IRC17:37
*** ksantoshk has quit IRC17:43
*** lhcheng_ has joined #tacker17:46
*** lhcheng has quit IRC17:49
sripriyabrucet: sridhar_ram: hello18:01
brucethello18:03
sripriyabrucet: we had multi-vim bip on the agenda today. because of time crunch for SFC we could not cover that up. i understand you had few comments on the heat orchestration part, since sridhar_ram is also here, i thought we could discuss it out here18:05
*** brucet has quit IRC18:07
sripriyabrucet: VIM Capacity / Resource Reservation / Allocation scenarios are important use cases we want to take it up for follow on iterations. we are starting out something simple to answer the immediate requirement of telcos who want to try out Tacker as a PoC in their exisitng OpenStack environments18:07
s3wongsripriya: your question is so harsh that bruset decided to quit :-)18:08
sridhar_rams3wong: lol18:08
s3wong* brunet (no longer on IRC :-)  )18:08
s3wongwow, never able to type his IRC nic correctly...18:08
sridhar_rambruce is hitting some irc client issue18:08
sripriyas3wong: lol, was that a question i asked? i thought i was concerned about his feedback ;)18:09
*** bobh has quit IRC18:09
s3wongsripriya: according to his IRC client, his Mac went to sleep18:10
*** bobh has joined #tacker18:10
s3wongsripriya: so most likely he slam his Macbook lid close after seeing your comment :-)18:10
sripriyas3wong: yup, i can just wait for his Mac to be awake then :p18:10
*** brucet has joined #tacker18:10
brucetOK18:10
s3wongback :-)18:11
sripriyas3wong: then it should have been the opposite and woken up from the sleep18:11
sripriyas3wong: :-)18:11
sripriyabrucet: hi again18:11
sripriyabrucet: not sure if you received my previous ping18:12
brucetI did18:12
*** lhcheng_ is now known as lhcheng18:12
brucetMy main concern is that I think we could have significant problems if we develop functionality in Tacker that operlaps with either existing or planned functionality in Hest18:13
brucetHeat18:13
brucetThe main advantage of the Heat engine as I understand it is that it is able to back out of any deployment process cleanly in the condition of an error18:15
brucetSo I think the more things Tacker can hand off to the Heat engine, the better.18:16
sripriyabrucet: regarding 1st point, are you talking about the multi-region feature in heat?18:16
brucetThat's one of them18:16
brucetI think SFC will be another18:16
brucetNeutron SFC will be exposed as Heat resources18:17
brucetTacker should take advantage of the SFC resources that exposed in Heat18:17
sridhar_rambrucet: Heat is a general purpose openstack orchestrator.. we are indeed planning to use Heat as much as possible, including for SFC when it is available18:17
brucetSo why not drive SFC definition and implementation into Heat?18:18
brucetIn general, I was thinking the Tacker orchestration architecture would be:18:19
brucetTosca NFV Template >> Heat Template >> Heat engine18:19
brucetThe goal would be to have a minimal (or ideally no) state machine used to drive the Heat engine18:20
sridhar_rambrucet: IMO, that's the classic misconception that Heat is sufficient for NFV orchestration needs18:20
brucetMy point was not about whether the current implementation of Heat is sufficient or not18:21
sridhar_rambrucet: I see Heat as a resource orchestrator (to create components like compute, network, etc) and not a workflow engine18:21
brucetI would  argue that the current Heat engine is likely sufficient for NFV.18:22
sridhar_rambrucet: again, specifically for SFC, if & when neutron-sfc finishes its work and when those constructs are available in Heat .. tacker will make use of it18:22
*** vishwana_ has joined #tacker18:22
brucetThe new functionality required for Heat will consist of new APIs exposed as Heat resources.18:23
sripriyabrucet: for multi-region feature, we will surely support in Tacker by using the existing functionality in heat , however as i understand that would still not work for telco environments who are already running multiple OpenStack deployments where the concept of multi region is not in picture18:23
sridhar_rambrucet: if Heat is sufficient for your NFV orchestration needs then that's perfectly fine!18:23
*** brucet has quit IRC18:23
*** vishwanathj has quit IRC18:24
* sridhar_ram waiting for bruce to join back18:24
*** bobh has quit IRC18:25
*** brucet has joined #tacker18:27
sridhar_rambrucet: you are back!18:27
brucetDid you get my last comments?18:27
brucetI am concerned that Tacker will end up building another execution engine for Orchestration on top of the Heat Engine18:27
brucetIf this is really required, than we may want to look at the Mistal project which is implementing a workflow engine that works with Heat.18:28
sridhar_rambrucet: some folks who have started this journey before (before Tacker started) are indeed still managing with Heat.18:28
sridhar_rambrucet:  But Tacker's stand is slightly different. ..18:28
sridhar_rambrucet: Our goal is not in any particular underlying technology components like Heat .. instead focus on NFV Orchestration as envisioned by ETSI MANO and implement with whatever that is available *TODAY*18:28
brucetIsn't Tacker an Openstack project??18:29
sridhar_rambrucet: like I said if some users things Heat and Mistral are sufficient .. they can very well build a solution off it18:29
sridhar_rambrucet: it very well is.. but IMO it differs slightly from other project (IMO)18:30
brucetI thought the idea behind Tacker would be to drive requirements into other OpenStack projects to support NFV.18:30
brucetAnd then develop functionality that does not overlap with other OpenStack priojects18:30
sridhar_rambrucet: we shouldn't be too cross about overlap.. some amount of overlap is healthy18:30
sridhar_rambrucet: I might be wrong, but the whole idea with big-tent is to allow many different projects .. sometimes every competing projects to evolve18:31
brucetBut overlap with core components??18:32
sridhar_rambrucet: again, we are not planning to re-implement anything that Heat already has.. we will just use it18:32
brucetAnd what about planned functionality?18:33
sripriyabrucet: we want to use as many features of heat if it is supporting certain use cases for NFV for ex: multi-region feature for new OpenStack deployments in telco environments18:33
brucetHow about Heat support for SFC?18:33
sridhar_rambrucet: again, for SFC we will use Heat SFC resource when it is available..18:33
sridhar_rambrucet: not sure if you attened the neutron-sfc design summit session...18:34
brucetOK. Then I suggest that we develop in a way that is going to be compatible with SFC functionality in Heat when it is available18:34
sridhar_rambrucet: let me say this, if neutron-sfc is available *today* .. Tacker will be using it..18:34
sridhar_rambrucet: we are closely working with neutron-sfc team on this..18:34
sridhar_rambrucet: that is the plan w.r.t SFC...18:36
brucetTo ensure we don't develop something incompatible, we should know how SFC is going to be exposed as Heat Resources. In fact, I think we should be helping to drive that definition.18:36
sridhar_rambrucet: tacker and neutron-sfc is maintaining a slide-deck on our iteration plans to reach there..18:36
brucetOK18:36
sridhar_rambrucet: in fact we are quite excited about each other!18:36
sridhar_rambrucet: many neutron-sfc members expressed interest to contribute to Tacker18:37
brucetNot sure if I saw any mention of Heat SFC support in Tacker SFC spec18:37
*** vishwana_ has quit IRC18:37
sridhar_rambrucet: neutron-sfc is a low-level SFC api..18:37
brucetSounds like we basically agree.18:37
brucetYes18:37
*** bobh has joined #tacker18:38
sridhar_rambrucet: integrating with heat resources, where available, is always the plan..18:38
sridhar_rambrucet: .. but at the same time, I'm very clear, we are not going to wait - say 12 months, for a downstream project to project something and then Heat resources to show up and then for Tacker to make use of it.18:39
brucetThe way I understand things now, it looks like all the higher level NGV functionality above what is available in Heat will be expressed in a Tosca NFVtempate18:39
sridhar_rambrucet: that will slow the features we can deliver to our user community18:39
brucetI understand. You don't want to wait.18:39
sripriyabrucet: just to add for the mult-vim support itself, we are starting out something very simple that can support existing openstack deployments without modifying too many components, in the follow on iterations, we will be supporting the heat multi region and multi-cloud features as they evolve18:39
brucetHowever, you can make sure that both projects are moving in the same direction.18:40
brucetSo whoever is implementing the Heat SFC implementation should review the Tacker SFC spec and vica versa.18:40
sridhar_rambrucet: side note, I've seen enough things go around in circles in opernstack for multiple cycles (read, over years) and nothings gets done18:40
*** vishwanathj has joined #tacker18:40
sridhar_rambrucet: yes, s3wong is core in both Tacker and neutron-sfc..18:41
brucetIsn't Heat multi region available now??18:41
sridhar_rambrucet: we have cross pollinated enough folks cross these two projects now18:41
brucetOK18:42
sridhar_ramsripriya: do you know ? I thought is still in early BP stage...18:42
sripriyabrucet: does it work for exisitng OpenStack deployments where there are multiple OpenStack instances running their own identity and other core services?18:42
sridhar_rambrucet: sripriya: I got to split for another mtg that is coming up..18:42
sripriyasridhar_ram: no multi-region feature is supported in heat, multi-cloud feature is still in early BP stage18:42
brucetMulti region is for single keystone instance I think18:42
brucetRight. Multi region is supported now.18:43
brucetSo why not start by having tacker use that???18:43
sripriyabrucet: yes! that is what i wanted to bring out where it does not support exisitng openstack instances if they are already running their own identity services18:43
brucetOK. Didn't see this in spec18:44
sripriyabrucet: that is the plan for the follow on phase of multi-vim18:44
sridhar_rambrucet: sripriya: my suggestion is .. we should consider few other keystone-region deployments if it make sense in the current Tacker multi-vim BP and if the effort estimate is manageable within one BP18:44
brucetSo first iteration will be focused on using Heat multi region support?18:44
sripriyabrucet: not yet, as the spec is as of today18:45
brucetBut you will add this?18:45
sridhar_rambrucet: first iteration will focus on individual heat endpoints..18:45
sridhar_rambrucet: but I agree we should put multi-region in scope and see if it make sense & can be absorbed18:46
sripriyasridhar_ram: brucet: if we think multi-region is also something we should plan in for the 1st iteration itself, then we should think about the effort estimate involved18:46
brucetI think the simplest thing would be to focus on just multi-region for first iteration18:46
sridhar_ramsripriya: yeah, I believe it is a very good idea to consider ...18:47
brucetI have seen other projects that have implemented orchestration using multiple Heat endpoints18:47
brucetThere was a talk on this in Vancouver18:48
sridhar_rambrucet: we need data to make that assumption... our initial users mostly had individual openstack installations18:48
sridhar_ramsridhar_ram: to be specific, this request came from Comcast18:48
brucetWell.. If that's important then you may want to look at the way this was implemented in the project I am referring to. I will try to find it and send a link.18:49
sridhar_rambrucet: sure... please send18:49
sripriyabrucet: when you mean only multi-region for 1st iteration, what happens to those cases where multiple heat end poinst exist and need imeediate support of this feature18:49
*** mbound_ has joined #tacker18:50
brucetNo, I actually meant using Heat multi-region support as is.18:50
sripriyabrucet: OK18:50
brucetI think this type of project could actually be subsumed by the Tosca NFV >> Heat Translator project18:50
sridhar_ramsripriya: beyond what it is in the BP, lets see if one of the VIM endpoint is a multi-region and have taker discover the regions and make it available for VNF placement18:51
brucetIf Tosca NFV object includes the region as an attribute, then the translator could just translate that to a region for Heat template18:51
brucetHmmmmm....18:52
brucetI thought OpenStack multi Region === multi VIM18:52
*** prashantD_ has joined #tacker18:52
brucetThe main difference is that there is one Keystone instance.18:52
sripriyasridhar_ram: does TOSCA call out vim parameter in placement policy in their document?18:52
sridhar_ramsripriya: not yet, this is something we need to implement what is in ETSI MANO and contribute back to TOSCA-NFV18:54
brucetAgree. Tacker can drive requirements in both directions.18:54
sridhar_ramthe discussion is too interesting to leave.. but I got to go..18:54
*** prashantD has quit IRC18:54
brucetOpenStack + NFV18:54
sripriyabrucet: agree on the last point, that is the point/gap we are addressing to have "backward/existing compatibility" multi-VIM for multiple keystone instances in near term!18:54
sridhar_ramwill catchup in the irc logs...18:54
sripriyasridhar_ram: thanks for your time!18:54
brucetOK18:54
sridhar_rambrucet: in short, I think openstack is missing an NFV api endpoint...18:55
sridhar_rambrucet: we are asking the operators to go to nova, heat, mistral, etc.. them all of them funnel thru Tacker!18:55
*** sridhar_ram has quit IRC18:56
*** u_kozat has quit IRC18:56
sripriyabrucet: i guess we are on the same page now, 1. to support existing deployments running multiple Keystone instances 2. to support new deployments running single Keystone instances18:57
brucetI was thinking that the main glue between Openstack and NFV would be Tosca NFV Template + Tosca >> Heat Translator18:57
sripriyafor 2--> we use the multi-region feature from Heat as is18:58
brucet2 == second iteration?18:58
brucetAh18:58
brucetOK18:58
sripriyabrucet: we can plan in for the 1st iteration itself if the effort seems reasonable18:59
brucetLet me look for the project that supported multiple OpenStack instances (no multi region)18:59
brucetI understand now.18:59
sripriyabrucet: though i have not thought through that in depth yet18:59
sripriyabrucet: that will be great!18:59
brucetOK. How do I get your direct email?19:00
sripriyabrucet: we can take a look at it19:00
brucetOr should everything be done through the review process?19:00
sripriyabrucet: that would be better since everybody can access and take a look at it19:00
brucetOK19:01
brucetLtr then19:01
sripriyabrucet: please post it on the BIP gerrit link and i will take a look19:01
sripriyabrucet: thank you for your time and feedback!19:02
*** brucet has quit IRC19:03
*** brucet has joined #tacker19:06
*** brucet has quit IRC19:11
*** prashantD_ has quit IRC19:11
*** prashantD has joined #tacker19:14
*** mbound_ has quit IRC19:35
*** sridhar_ram has joined #tacker19:36
*** vishwanathj has quit IRC19:42
*** vishwanathj has joined #tacker19:42
*** karimb has joined #tacker19:46
*** santoshk has joined #tacker19:49
*** karimb has quit IRC19:51
*** u_kozat has joined #tacker19:56
*** u_kozat has quit IRC20:03
*** sripriya has quit IRC20:21
*** sripriya has joined #tacker20:48
*** ksantoshk has joined #tacker21:02
*** sripriya_ has joined #tacker21:03
*** santoshk has quit IRC21:05
*** sripriya has quit IRC21:06
*** u_kozat has joined #tacker22:00
*** u_kozat has quit IRC22:05
prashantDhi sridhar_ram : ping22:07
sridhar_ramprashantD: pong22:27
sridhar_ramprashantD: sorry, was away in a mtg22:27
sridhar_ramprashantD: what's up ?22:27
openstackgerritMerged openstack/tacker: Add Tacker monitor test with new template  https://review.openstack.org/24786222:41
ksantoshks3wong thankyou for merging my patchset..22:47
s3wongksantoshk: your IRC nic keeps on changing :-)22:48
ksantoshk:-)22:48
sripriya_bobh: ping22:54
bobhsripriya_: hello22:54
sripriya_bobh: there are few patchsets in pipeline that need rebasing, just wanted to follow up22:55
bobhthey are on my list of things to do - just seems to be getting longer instead of shorter.22:55
bobhhoping to have some time this week to catch up22:56
sripriya_bobh: understand, sure thank you.22:56
*** bobh has quit IRC22:57
*** u_kozat has joined #tacker23:03
sridhar_ramtrozet:ping23:15
sridhar_ramtrozet: ping23:15
*** lhcheng has quit IRC23:48
*** lhcheng has joined #tacker23:48

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!