06:00:11 <tojuvone> #startmeeting Fenix
06:00:12 <openstack> Meeting started Mon Mar 25 06:00:11 2019 UTC and is due to finish in 60 minutes.  The chair is tojuvone. Information about MeetBot at http://wiki.debian.org/MeetBot.
06:00:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
06:00:16 <openstack> The meeting name has been set to 'fenix'
06:00:46 <dangtrinhnt> o/
06:01:40 <tojuvone> Hi dangtrinhnt, anybody else around?
06:01:54 <dangtrinhnt> hi
06:02:22 <tojuvone> Well let's start
06:02:39 <tojuvone> #topic OpenStack upgrade status (ONS demo)
06:02:42 <dangtrinhnt> okie
06:03:39 <tojuvone> So got this now working and composed a video of it
06:03:48 <tojuvone> #link https://youtu.be/w5YPH5In0OQ
06:03:49 <dangtrinhnt> nice
06:04:13 <tojuvone> So the whole 1 controller + 4 compute upgrade takes an hour
06:04:34 <tojuvone> Migrations and DevsTack upgrade scripts takes time
06:04:35 <dangtrinhnt> Looks like CentOS :)
06:05:22 <tojuvone> Ubuntu Bionic on hosts and CentOS on HA VMs if I recall
06:05:43 <dangtrinhnt> okie
06:06:01 <tojuvone> Ubuntu 18.04.1 LTS
06:06:10 <tojuvone> That was foudn under /etc :D
06:06:42 <tojuvone> So the short video is now ~18min
06:07:13 <tojuvone> Also commented it has still some network downtime + surely API downtime during controller upgrade
06:07:59 <tojuvone> And yet there is no commenting on video as haven't went it through like it would be presented in ONS
06:08:15 <tojuvone> Meanwhile if any comment on the video if you look at it, let me know
06:08:41 <dangtrinhnt> okie
06:09:14 <tojuvone> Demo uses "5" different boxes
06:09:49 <tojuvone> 1 on top left is "Doctor" = application managers and admin
06:09:59 <tojuvone> top right is Fenix logging
06:10:11 <tojuvone> down left is Horizon
06:10:52 <tojuvone> and down right will apper 2 boxes running "surf" browser showing http server content runnign behind floating IPs
06:11:34 <tojuvone> meaning the active VM on HA instances. This demonstrates the switchover when needed on those VMs
06:11:47 <tojuvone> and as multitenant, there is 2 of these windows
06:12:41 <tojuvone> One tenant has 2x non_ha_app and 2x ha_app with active/stdby
06:13:20 <tojuvone> both tenants will be scaled down by one non_ha_app VM to gain empty compute host
06:13:40 <tojuvone> ok, that is some details. Any questions at this point?
06:13:51 <dangtrinhnt> No. Thanks.
06:14:03 <dangtrinhnt> LGTM.
06:14:19 <tojuvone> ok... maybe move on to next topic then
06:14:25 <dangtrinhnt> +1
06:14:33 <tojuvone> #topic status
06:15:06 <tojuvone> ok, so this should be what is done last 2 weeks and what will happen next
06:15:30 <tojuvone> I can start what I have been doing
06:15:47 <tojuvone> so surely the demo, that is obvious
06:16:09 <tojuvone> but then there was also those "ETSI  meetings" last week
06:16:35 <tojuvone> I will come to more details later, but to comment something already
06:17:30 <tojuvone> ETSI FEAT003 is still ongoing and details of the maintenance have not been agreed
06:18:03 <dangtrinhnt> okie, guess that there could be some changes on the interfaces? Should I expect some public documents about that?
06:18:20 <tojuvone> So this will still make some time and before that it is better not to proceed any related changes
06:18:28 <dangtrinhnt> +1
06:19:06 <tojuvone> I have to ask when and how much that can be showed to public then
06:19:22 <tojuvone> but would think when finalized
06:19:39 <tojuvone> just some brief about it anyhow:
06:20:15 <tojuvone> there will be constraints of how VNF can be upgraded (servers / groups)
06:20:32 <tojuvone> Fenix should be configured accordingly
06:20:53 <tojuvone> maybe some of thesse will land to affinity/anti-affinity groups too
06:21:07 <tojuvone> but that depends on how things will go
06:21:56 <tojuvone> We have some already in different way as we have the VNFM interaction
06:22:51 <tojuvone> and for PTG there should be coming a discussion over this
06:23:03 <dangtrinhnt> ok
06:23:49 <tojuvone> so that should be composed in a way that is does not use the official ETSI documents that cannot be shared..
06:24:42 <tojuvone> but still revealing needed details to work on changes to OpenSTack, like Nova (those server groups and live migration requirements)
06:25:03 <tojuvone> Any questions at this time?
06:25:57 <tojuvone> so coming 2 weeks I will have the ONS demo and I should find time to have some of those chagnes under review in Fenix
06:26:16 <dangtrinhnt> Is that mean we may have to change the current workflow including what hyungiskyang working on with tacker?
06:26:35 <tojuvone> csatari, will be driving the ETSI discussion
06:27:33 <tojuvone> dangtrinhnt, a good question
06:28:16 <tojuvone> There is still too options if our interaction cannot be used as it is
06:28:55 <tojuvone> have that and a one with ETSI definitions fulfilled
06:29:30 <tojuvone> It might also be complicated if there needs to be EM in VNF
06:30:02 <tojuvone> where VNFM is more or less a proxy for information coming from there
06:31:17 <tojuvone> Constraints for "maintenance" should be known by the VNF itself, nobody else
06:32:32 <tojuvone> otherwise if it doesn't have that, there could be only some default/configured way on what to do
06:33:18 <tojuvone> one is to think there is always extra capacity given to rolling maintenance
06:33:57 <tojuvone> so application can scale up the VMs involved, so service will not suffer when those VMs need to be removed
06:34:18 <tojuvone> as example
06:34:18 <dangtrinhnt> +1
06:35:33 <tojuvone> ok, dangtrinhnt, any other comment on this, or what is your status?
06:36:00 <dangtrinhnt> Ok, I think it's a good strategy. No update from me.
06:36:15 <tojuvone> ok, great :)
06:36:28 <tojuvone> so next topic then
06:36:42 <tojuvone> #topic Summit and PTG
06:37:16 <tojuvone> Well, we know the status of these, but in sake of minutes, I put it here
06:37:47 <dangtrinhnt> :)
06:38:26 <dangtrinhnt> got my flight ticket
06:39:09 <tojuvone> Fenix rolling OpenStack upgrade demo will be in ONS America summit in LFN booth 3rd - 5th of April
06:39:23 <tojuvone> dangtrinhnt,  great, we are so going now :)
06:39:23 <dangtrinhnt> nice
06:40:30 <tojuvone> Next will be Open Infrastructure summit 29th April to May
06:40:39 <dangtrinhnt> \
06:40:41 <dangtrinhnt> \m/
06:40:57 <tojuvone> Fenix forum session will be held there 30th May at 9AM
06:41:48 <tojuvone> Right after this will be PTG and Fenix session is scheduled on 4th May
06:42:18 <tojuvone> Otherwise there should be also session for ETSI FEAT03 during the PTG
06:42:54 <dangtrinhnt> wow
06:42:55 <tojuvone> And Fenix surely might be mentioned in some other discussion, but haven't looked those
06:43:51 <tojuvone> oh, dangtrinhnt, the forum allowed to put only one person somehow, but surely would like to do it together :)
06:44:04 <dangtrinhnt> ah, sure, no problem
06:44:19 <tojuvone> I mean the CFP process was so for that
06:45:10 <tojuvone> ok, that was what is coming as events, not to go to more in details?
06:45:44 <dangtrinhnt> nope. Thanks for the updates.
06:46:15 <tojuvone> ok then, we would heve the AoB (any other business) left
06:46:32 <tojuvone> #topic AoB
06:47:09 <tojuvone> Maybe something about Doctor
06:47:32 <tojuvone> There is currently a huge problem in getting testing done there
06:47:48 <tojuvone> So let's see if it is feasible to even continue there
06:47:53 <dangtrinhnt> what can we do to help?
06:48:27 <tojuvone> Meanwhile, I would then hurry to get Fenix testing working in OpenStack side
06:48:43 <tojuvone> Well, OPNFV is loosing people fast
06:48:54 <dangtrinhnt> I can see that
06:48:55 <tojuvone> And installers are not working
06:49:23 <tojuvone> so you cannot easily have environment to test. Or you should have a "huge" effort to set it
06:49:37 <dangtrinhnt> Could we do some roll-call for more contributors at the Summit / ONS
06:49:39 <dangtrinhnt> ?
06:49:40 <tojuvone> do the work that should be done by installer
06:50:13 <tojuvone> There is OPNFV meetup just befefore ONS
06:50:27 <tojuvone> So I will surely rise my concern there
06:51:02 <tojuvone> But stil las working currently to drive Fenix, it is double effort to keep Doctor just for that
06:51:33 <dangtrinhnt> what exactly testing environment that you need for Doctor? Could it be solve by providing some resources?
06:52:14 <tojuvone> Problem is more like the installers do not work
06:52:27 <tojuvone> so somebody should work on that
06:52:46 <tojuvone> but also, yes
06:53:26 <tojuvone> As far as Apex worked, I was able to test fault management and maintenance in Nokia POD
06:53:32 <tojuvone> that was kind of enough
06:53:49 <tojuvone> but fault management only for sample implementation, not congress
06:54:05 <dangtrinhnt> Do you have the bug report?
06:54:06 <tojuvone> that is know to be broken currently
06:55:04 <tojuvone> I do not have a bug report to Apex open currently
06:55:30 <tojuvone> DWG had one that is fixed now, but do not know if that fixed her virtual pod or not
06:55:43 <tojuvone> for me it was not enough with beremetal pod
06:56:36 <tojuvone> so yes, maybe should rise one with logs from the baremetal pod as fexes seems not to be enough for that
06:57:03 <tojuvone> Still haven't heard anybody will be continuing the Apex after this release
06:57:15 <dangtrinhnt> :o
06:57:31 <tojuvone> so, that is how deep things are
06:57:40 <dangtrinhnt> ok, got it.
06:58:06 <tojuvone> still, would be nice to have aworking release for Doctor
06:58:41 <tojuvone> no way to test the Fenix support and have it in release
06:59:09 <tojuvone> one minute left
06:59:31 <tojuvone> Any last words?
06:59:43 <dangtrinhnt> No. Thanks.
07:00:03 <tojuvone> ok, thanks for joining :)
07:00:19 <tojuvone> #stopmeeting
07:00:38 <dangtrinhnt> should be endmeeting I guess :)
07:00:47 <tojuvone> :D
07:00:53 <tojuvone> #endmeeting