04:30:06 <gongysh> #startmeeting tacker
04:30:07 <openstack> Meeting started Wed May 31 04:30:06 2017 UTC and is due to finish in 60 minutes.  The chair is gongysh. Information about MeetBot at http://wiki.debian.org/MeetBot.
04:30:08 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
04:30:10 <openstack> The meeting name has been set to 'tacker'
04:30:17 <tung_doan> o/
04:30:20 <tung_doan> hi all
04:30:28 <YanXing_an> o/
04:30:32 <claire1006> o/
04:30:45 <gongysh> #topic Roll Call
04:30:59 <gongysh> claire1006, hi
04:31:00 <YanXing_an> hello, claire1006, wellcome
04:31:01 <gongysh> welcome
04:31:06 <claire1006> hi,all
04:31:19 <bennco1> hello
04:31:32 <gongysh> bennco1, hi welcome
04:31:39 <bennco1> thank you
04:32:22 <gongysh> it seems we need some time for other guys to join.
04:32:38 <tung_doan> gongysh: agree
04:34:18 <sridhar_ram> hi
04:34:27 <tung_doan> hi sridhar_ram
04:34:34 <gongysh> sridhar_ram, great to have you on meeting.
04:34:52 <sridhar_ram> sorry for the delay, will get used to the new time :)
04:35:14 <gongysh> ok. today, we have two new guys. bennco1 and claire1006
04:35:22 <bennco1> :)
04:35:27 <claire1006> hi~
04:35:31 <gongysh> bennco1, claire1006 , could you please introduce yourself?
04:36:00 <bennco1> @claire1006 can go first
04:36:00 <bennco1> :)
04:36:19 <YanXing_an> claire1006 is a girl, from china mobile :)
04:36:50 <bennco1> Well hello My name is Conrad from Verizon Wireless
04:37:01 <gongysh> YanXing_an, great to hear china mobile give one more member to this project.
04:37:02 <bennco1> Conrad Bennett
04:37:10 <sridhar_ram> welcome claire1006 and bennco1  to Tacker!
04:37:23 <gongysh> bennco1,  welcome.
04:37:29 <tung_doan> welcome claire1006 and bennco1 to Tacker :)
04:37:38 <claire1006> my name's bao fangyan,i'm yanxin'an's colleague.
04:37:56 <gongysh> verizon is acquired by someone, right?
04:38:34 <bennco1> Verizon Wireless? No, not that I have heard
04:38:42 <gongysh> bennco1, ok
04:38:49 <gongysh> lets start
04:38:53 <gongysh> #topic bp
04:39:29 <gongysh> remember June 8th, the deadline of P2 milestone is the deadline of BP merge
04:39:46 <gongysh> if there is no much exception, we will abide by this rule.
04:40:11 <gongysh> #link https://review.openstack.org/#/q/project:openstack/tacker-specs
04:40:29 <gongysh> we have some specs pending for review or update.
04:40:47 <gongysh> first one is from tung_doan, https://review.openstack.org/434974
04:40:52 <gongysh> tung_doan, hi
04:41:05 <tung_doan> gongysh: hi
04:41:24 <tung_doan> gongysh: I am trying to wrap my bp this week
04:41:37 <gongysh> tell us the progress on this spec, how far it is from being merge?
04:42:15 <tung_doan> gongysh: I think it could be merge soon. We just need some modification in NFVO
04:42:39 <tung_doan> gongysh: related to some support for connection points
04:43:03 <tung_doan> at the same time. i also open the patch for this bp: https://review.openstack.org/#/c/469209/
04:43:04 <gongysh> tung_doan,  is it related to https://review.openstack.org/#/c/448109/ vnffg in NSD?
04:43:43 <tung_doan> gongysh: firstly, i will go with vnffg, then move to nsd
04:44:01 <gongysh> ok
04:44:09 <tung_doan> tung_doan: the scope of my bp only focuses on VNFFG
04:44:36 <tung_doan> gongysh: I think code will be lauched soon for this feature
04:44:51 <gongysh> tung_doan, lets get bp merged first.
04:45:00 <tung_doan> gongysh: got it
04:45:08 <gongysh> second bp
04:45:09 <gongysh> VNFFG support for Network Service Descriptor
04:45:15 <gongysh> trozet, hi
04:45:16 <sridhar_ram> tung_doan: does the scope include VNFFG fixup due to a VNF respawn ?
04:45:29 <tung_doan> sridhar_ram: right
04:45:40 <sridhar_ram> tung_doan: my understanding ^^ shd not need any TOSCA changes, can u confirm?
04:45:51 <tung_doan> sridhar_ram: exactly
04:46:22 <gongysh> tung_doan, I think so too.
04:46:35 <tung_doan> sridhar_ram: we need auto-heal VNFFG when some VNF go down
04:46:54 <gongysh> trozet,  dharmendra.kushwaha are both absent here.
04:47:11 <gongysh> so let skip the NSD bp
04:47:39 <sridhar_ram> tung_doan: so, the users need not do anything to get this feature .. that is unclear from the BP
04:48:52 <tung_doan> sridhar_ram: when some VNFs are respawned, their connection points will change. That break VNFFG.
04:49:03 <sridhar_ram> tung_doan: i'll add some comments to the BP .. basically L39 (gongysh already flagged this)
04:49:20 <sridhar_ram> tung_doan: understood, thanks
04:49:24 <tung_doan> sridhar_ram: sure. really need your comment. thanks
04:49:38 <gongysh> tung_doan, I am also wondering what we need to do under scaling.
04:50:23 <sridhar_ram> gongysh: i'm not sure if we support VNFFG + Auto-Scaling combo
04:50:28 <tung_doan> gongysh: ok. when vnf is overload, the new chain should be created.
04:50:50 <sridhar_ram> gongysh: it would involve add new neutron ports to port-pair groups dynamically
04:50:55 <sridhar_ram> *addming
04:51:18 <tung_doan> sridhar_ram: you got the point
04:52:06 <gongysh> sridhar_ram, our VNFFG is fixed on a certain CP path.
04:52:53 <tung_doan> sridhar_ram: gongysh: in Barcelona summit I heard some questions about auto-healing in VNFFG. It should be up to Tacker
04:53:03 <sridhar_ram> gongysh: that's what i was afraid of .. we first shd work towards enabling FFG + Scaling before taking that up ..
04:53:30 <YanXing_an> it's a big problem, i think we need fix it to support vnffg+scaling
04:53:33 <sridhar_ram> i suggest we split this BP into these two scenarios
04:53:40 <gongysh> sridhar_ram, yes.
04:53:42 <sridhar_ram> .. into two BPs
04:54:14 <gongysh> tung_doan, do you got it? I agree with sridhar_ram.
04:54:42 <tung_doan> gongysh: one for vnffg+scaling, another for what? could you elaborate?
04:54:55 <gongysh> another is: vnffg to deal with respawn policy action
04:55:18 <tung_doan> gongysh: sridhar_ram: got it. i know they are different
04:55:49 <sridhar_ram> tung_doan: let's split those into two diff BPs .. so that they are scoped appropriately
04:56:06 <tung_doan> sridhar_ram: got it. thanks
04:56:13 <gongysh> tung_doan, so you have lot to do. :)
04:56:29 <gongysh> next bp is Implement VNF monitor policies with Mistral workflow
04:56:36 <gongysh> https://review.openstack.org/#/c/467479/
04:56:37 <tung_doan> gongysh: this time I can deal with these :)
04:56:52 <gongysh> sridhar_ram, I need your review
04:56:55 <YanXing_an> tung_doan, i can help to write one if possible, after barbican finished.
04:57:25 <tung_doan> YanXing_an: vnffg+scaling may need your support
04:57:36 <tung_doan> YanXing_an: I can deal with vnffg+respawning
04:57:45 <sridhar_ram> gongysh: sure, will do
04:57:47 <tung_doan> YanXing_an: thanks.
04:58:03 <sridhar_ram> tung_doan: +1 for that sequence
04:58:29 <gongysh> next bp is https://review.openstack.org/#/c/453442/
04:58:34 <gongysh> block storage bp
04:58:42 <gongysh> from chinamobile
04:59:02 <gongysh> zhouzhihong is busy with company job.
04:59:10 <gongysh> so who can take this?
04:59:16 <YanXing_an> zhouzhihong is not here, i will help to remove the WIP fields.
04:59:48 <gongysh> YanXing_an, you need to update the spec according to review too.
05:00:16 <YanXing_an> gongysh, ok
05:00:29 <gongysh> 30 mins passed
05:01:01 <gongysh> #topic bugs
05:01:16 <gongysh> or codes.
05:01:41 <gongysh> https://review.openstack.org/#/c/465080/
05:01:48 <gongysh> barbican patch
05:01:56 <YanXing_an> tung_doan, seems vnffg+scaling is difficult than vnffg+respawning :)
05:02:11 <gongysh> it seems almost done.
05:02:25 <YanXing_an> gongysh, i will commit a PS EOD.
05:02:33 <tung_doan> YanXing_an: that's why you and me should co-work ;)
05:03:00 <tung_doan> * you and I
05:03:12 <gongysh> https://review.openstack.org/#/c/457843/
05:03:36 <claire1006> https://review.openstack.org/#/c/469304/
05:03:44 <gongysh> this also need review, so that I can go with vim reachability bp coding.
05:04:20 <claire1006> resources like vnf which got stuck in PENDING_DELETE status can't be deleted correctly.
05:04:29 <sridhar_ram> gongysh: is the patchset for tacker-conductor out?
05:05:05 <gongysh> sridhar_ram,  not, it is on tacker API to start workflow.
05:05:32 <gongysh> tacker-conductor will be next patch to implement VIM monitoring.
05:05:53 <gongysh> claire1006, I know it is a bug, could you please update doc to talk about the state transition?
05:05:59 <sridhar_ram> gongysh: soudns good
05:06:03 <claire1006> It's better to add a state transition,or just make the deletion process continue?plz give some suggestion.thx.
05:06:46 <claire1006> i didn't take the state transition way.
05:07:17 <gongysh> claire1006, I am talking to write a doc about the state transition, so that we can make decision if the change is reasonable.
05:08:12 <sridhar_ram> claire1006: is the heat stack ID available in the vnf db for those VNFs stuck in PENDING_DELETE ?
05:09:07 <claire1006> sridhar_ram,i will check it later
05:09:31 <sridhar_ram> i remember we discussed 'vnf-delete --force' option way back to handle situations like this,
05:10:03 <YanXing_an> sridhar_ram, i think the VNF is already deleted in heat in this bug.
05:10:13 <sridhar_ram> YanXing_an: I see ..
05:10:20 <gongysh> claire1006,  make a clear state transition in a devref doc and then make the change will help us.
05:10:38 <claire1006> It can't be found by _get_vnf_db() when it got stuck in PENDING_DELETE.
05:10:59 <claire1006> gongysh, ok
05:11:11 <gongysh> claire1006 thanks
05:11:35 <gongysh> #topic open discussion.
05:11:59 <tung_doan> sridhar_ram: this bp is a really good feature: https://blueprints.launchpad.net/tacker/+spec/tosca-csar-mgmt-driver
05:12:07 <YanXing_an> gongysh, this state transition need another patch?
05:12:10 <tung_doan> sridhar_ram: do you have a plan to do it?
05:12:19 <diga> o/
05:12:26 <tung_doan> sridhar_ram: if i finish my work load, i will look into this
05:12:32 <diga> is our meeting timing changed ?
05:12:38 <gongysh> YanXing_an, we can make it in the same patch. but another is ok too.
05:12:59 <gongysh> diga, it is one hour earlier than before.
05:13:04 <sridhar_ram> tung_doan: yes, that BP  was a long standing request
05:13:04 <diga> ohh
05:13:15 <YanXing_an> gongysh, ok
05:13:20 <diga> gongysh: that;s the reason I am missing it
05:13:26 <sridhar_ram> tung_doan: it makes even more relevant now that we support NSD, FFGD, etc
05:13:39 <tung_doan> sridhar_ram: seems heat-translator already support csar. it is really nice :)
05:14:01 <diga> gongysh: will talk you after the meeting, I think we are about to finish the meeting
05:14:16 <sridhar_ram> tung_doan: yeah, i remember that .. we can use that to "unpack" the CSAR but we need to pass it thru' tacker for parsing
05:14:22 <gongysh> diga, we have still time about 10 mins.
05:15:29 <diga> gongysh: okay
05:16:01 <diga> gongysh: I will start pushing the patches on per module weekwise
05:16:15 <tung_doan> sridhar_ram: agree. In the future, i think we can separate out a repo like tacker-catalog :) This will process TOSCA, YANG,...
05:16:20 <diga> gongysh: I didn't get time last month to work on this peace due to some office releases
05:16:33 <sridhar_ram> tung_doan: yep
05:17:01 <diga> gongysh: I will start with VIM api
05:17:01 <gongysh> diga, never mind.  let push tacker forward together.
05:17:13 <diga> gongysh: Yeah
05:17:36 <diga> :)
05:17:47 <sridhar_ram> diga: fyi, there is a general convergence in using uwsgi and riding on existing apache (using mod_wsgi)
05:17:56 <sridhar_ram> .. in other openstack projects
05:18:03 <diga> sridhar_ram: ohh..kk
05:18:25 <diga> sridhar_ram: do you have any link on that ?
05:18:31 <diga> sridhar_ram: I will go through it
05:18:59 <sridhar_ram> diga: don't have it handy, will pass on when i find it
05:19:17 <diga> sridhar_ram: sure! Thanks
05:19:45 <diga> sridhar_ram: is this a first meeting after summit or second ?
05:20:06 <diga> sridhar_ram: I dont know how many I have missed in between
05:20:28 <sridhar_ram> diga: please use this to catchup -> http://eavesdrop.openstack.org/meetings/tacker/2017/
05:20:34 <gongysh> #info, YanXing_an and I proposed a topic in openstack china day to talk about tacker vnf.
05:20:47 <sridhar_ram> gongysh: wonderful !
05:21:15 <diga> sridhar_ram: yep, thanks for sharing the link
05:21:34 <gongysh> if someone talks about tacker at any conference, please share it.
05:22:12 <gongysh> anything else to talk?
05:22:59 <gongysh> #topic core team adjustment
05:23:35 <gongysh> I will make some adjustment about core team at Pike 3 milestone
05:24:04 <gongysh> so please stay active to make project hit.
05:24:42 <diga> gongysh: +1
05:25:06 <gongysh> if there is no more to talk, lets end the meeting. thanks everyone.
05:25:16 <diga> bye
05:25:17 <gongysh> #endmeeting