01:00:12 #startmeeting ovn4nfv 01:00:12 Meeting started Tue Nov 28 01:00:12 2017 UTC. The chair is trinaths. Information about MeetBot at http://wiki.debian.org/MeetBot. 01:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 01:00:12 The meeting name has been set to 'ovn4nfv' 01:00:17 #chair trinaths 01:00:17 Current chairs: trinaths 01:00:30 Good Morning All 01:00:49 good morning 01:00:57 #info Trinath Somanchi - NXP 01:01:26 #info Aakash KT 01:02:22 #topic Openstack-OVN charm – design 01:02:32 hi ak_dev 01:03:21 hello trinaths 01:04:11 this is my first meeting, so forgive me if I do not follow the rules (if any) 01:04:15 ak_dev: as per our ML discussion, 01:04:27 you have any input further ? 01:04:53 I do, was looking up a bit on this topic 01:05:04 but unfortunately, do not have any code as of yet 01:05:29 agree. 01:06:19 ak_dev: since I dont have any team with me internally I was checking with JOID JIRA tickets 01:06:47 #link http://lists.openstack.org/pipermail/openstack-dev/2017-August/121587.html 01:07:29 as per the discussion in the above ML, 01:07:49 charm-neutron-api-ovn <--> charm-ovn <--> charm-neutron-ovn 01:08:02 might be the implementation 01:08:19 yeah according to my further readings, that seems to be accurate 01:08:36 similar to charm-neutron-api-odl <--> charm-odl-controller <--> charm-openvswitch-odl. 01:09:24 with the above pattern, and your expertise 01:09:44 https://www.irccloud.com/pastebin/fOXebgYK/ 01:09:46 can you please post a outline charms for all these 3 blocks for review ? 01:10:12 ak_dev: yes agreee. 01:10:59 ak_dev: if you see, for ODL too the installation differs 01:11:11 trinaths: sorry I could not post the design yet, I will do them soon, been a lot of work lately (still) 01:11:20 trinaths: yes 01:11:56 trinaths: I believe openstack has some template for charms, which we can use, and some other layers to help with installation 01:11:57 ak_dev: one doubt 01:12:16 ak_dev: where do we need to post this? opnfv or openstack-charms project ? 01:12:55 trinaths: even I am not so sure about this 01:13:12 ak_dev: from where k8s-ovn was taken ? 01:13:13 trinaths: you are talking about the review code right? 01:13:27 ak_dev: yes, the charms actually 01:13:38 ak_dev: I see two things here, 01:13:46 ak_dev: 1. joid charms repo 01:13:54 2. openstack-charms repo 01:14:02 oh so for k8s-ovn, the charms were developed in my own repo, on git 01:14:19 then we pushed the charms to Juju 01:14:28 ak_dev: great. 01:14:34 and then the logic to deploy was done in JOID 01:14:36 ak_dev: how did you test that charms 01:14:51 trinaths: initially, I tested on CENGN pod 01:15:04 later on intel pod 01:15:26 and most of the time, for quick tests, on google cloud (my account) 01:15:28 i was unable to get those PODs. since many at holidays 01:16:00 trinaths: the our new charms, I can test them on my local PC, built just a few days back 01:16:01 ak_dev: ok. what is the environment at those PODs ? 01:16:52 trinaths: if I remember right, something like one jumphost, 6 compute nodes 01:17:26 ak_dev: do we need such env for developing and testing charms ? 01:17:37 trinaths: I have been able to deploy "openstack-on-lxd" on my PC using juju, we can do the same for our project 01:17:57 trinaths: I think for final testing, we will require, but for development, my PC, for now, should suffice 01:18:04 oh sounds great 01:18:52 can you share your fndings on how to bring your dev env on my lab 01:19:28 trinaths: I have not look at that, but that will be a problem since my PC is on the college network 01:19:51 to get access, a VPN needs to be created, and that requires a college ID 01:20:13 ak_dev: ok. confused. not your lab setup, but the procedure to bring the same 01:20:32 trinaths: oh sorry I misunderstood 01:21:18 ak_dev: since I too want to have my lab for testing the charms 01:21:40 mine is 16 GB ram, amd ryzen 1500 X, and 128 GB SSD 01:21:50 which is close to what is suggested here : https://docs.openstack.org/charm-guide/latest/openstack-on-lxd.html 01:21:51 ak_dev: same here 01:22:09 #link https://docs.openstack.org/charm-guide/latest/openstack-on-lxd.html 01:22:25 oh great! I think then you can follow the steps mentioned there to set this up 01:22:42 ak_dev: great 01:23:01 ak_dev: how do we go with 3 charms required for ovn 01:23:17 any inputs from narinder 01:23:36 trinaths: I have not spoken to him recently 01:23:41 ok. 01:24:19 ak_dev: how do we design the new charms? can you share your ideas 01:25:39 the thing I have been able to understand is that the charm-ovn and charm0ovn-api will be standalone charms 01:26:06 ak_dev: ok. standalone ? 01:26:22 and charm-sorry mistake there 01:26:53 ok 01:27:03 charm-neutron-api-ovn and charm-ovn are standalone 01:27:12 meaning 01:27:21 they will be deployed directly on machines 01:27:24 ok 01:27:32 how about the middle one ? 01:27:38 and charm-neutron-ovn will be subordinate 01:27:48 to nova deployments 01:28:02 that means that it will be present along side all nova deployments 01:29:32 okay. 01:29:59 with these charms, do we need to code for joid integration 01:30:01 ? 01:31:03 trinaths: that we will have to check with Narinder 01:31:14 to add a JOID scenario 01:32:42 ok. for test the above written charms, do openstack-on-lxd suffucient ? 01:33:08 I believe we need a juju env to integrate and test the charms before hand 01:33:53 trinaths: openstack-on-lxd will have to be installed via juju, so we will have the environment 01:34:02 great 01:34:03 +1 01:34:33 i think it should be sufficient, but I have never personally tried doing it, so can't guarantee 01:34:40 but most likely it will be alright 01:34:44 cool :) 01:34:51 :) 01:35:11 when is your travel planned. ? 01:35:32 3rd morning 01:35:38 to 13th night 01:35:56 dec 15 is the deadline for nosdno-noha scenarios 01:36:45 I think we have some time for openstack-ovn-noha scenario. but we need to start it before hand rather than having it start in tight schedule. 01:37:40 agree +1 01:37:57 I am not sure how my schedule is, when I am there though 01:38:14 and here, it is highly unpredictable as of now too 01:38:23 but I will do my best to keep things going 01:39:52 ak_dev: thank you for supporting the project. 01:40:09 how to get CENGEN POD access? is there a procedure ? 01:41:25 trinaths: for CENGN pod, I had contacted Dave Urschatz 01:41:27 dave.urschatz@cengn.ca 01:41:42 along side being in touch with Narinder 01:41:46 ok. let me also put a mail for him 01:42:01 Narinder is not active responding the mails. 01:42:12 okay yeah, that should be great 01:42:18 might be on thanksgiving week leave 01:42:34 Dave will send out the further instructions 01:42:37 oh okay 01:42:42 ok 01:44:15 #topic JOID and APEX PODs 01:44:34 I just dropped a mail to dave and jack 01:44:37 for POD access 01:45:00 ak_dev: were there any new interns registering for LF ? 01:45:46 trinaths: not that i know of, recently 01:46:18 ak_dev: ok. 01:46:33 do you know any one working for apex installer ? 01:46:43 trinaths: do you plan to push this as an intern project? 01:47:16 intern project ? 01:47:17 trinaths: no idea about apex / who is working on it 01:47:33 ok 01:48:18 something like here 01:48:23 #link https://wiki.opnfv.org/display/DEV/Intern-projects-page 01:49:04 We work on ARM SoC. so my work mostly be on testing the scenarios on ARM. 01:49:18 ak_dev: no intern project. 01:49:33 I need a helping hand to move with the project. 01:49:38 trinaths: oh okay 01:50:46 trinaths: a request, if there is nothing else to talk about, could we end the meeting? I have a presentation in around 3 hours 01:51:41 ak_dev: ok. 01:52:01 thanks you ak_dev for attending the meeting and share your thoughts. 01:52:09 #endmeeting