14:00:20 <pwhalen> #startmeeting Fedora IoT Working Group Meeting 14:00:20 <zodbot> Meeting started Wed Apr 1 14:00:20 2020 UTC. 14:00:20 <zodbot> This meeting is logged and archived in a public location. 14:00:20 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:20 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:00:20 <zodbot> The meeting name has been set to 'fedora_iot_working_group_meeting' 14:00:20 <pwhalen> #chair pwhalen pbrobinson bcotton tdawson puiterwijk 14:00:20 <pwhalen> #topic roll call 14:00:20 <zodbot> Current chairs: bcotton pbrobinson puiterwijk pwhalen tdawson 14:00:39 <bcotton> .hello2 14:00:39 <zodbot> bcotton: bcotton 'Ben Cotton' <bcotton@redhat.com> 14:00:58 <tdawson> I'm here ... and ready for the right meeting this time. :) 14:01:10 <puiterwijk> Hi 14:01:11 <pwhalen> heh, morning tdawson 14:01:30 <pwhalen> howdy puiterwijk 14:01:57 <jsmith> Morning! 14:02:53 <pwhalen> Good morning jsmith! 14:03:08 * pwhalen is just giving it a couple minutes for others to join 14:04:25 * pbrobinson is here 14:04:27 <pwhalen> ok, lets get started 14:04:29 <pwhalen> #topic 1) ==== Working Group process and admin ==== 14:05:19 <pwhalen> So the vote for the new name is over, have we heard back from legal? 14:05:40 <pbrobinson> not that I'm aware of 14:05:50 <bcotton> we haven't even sent it yet :-) 14:05:56 <pwhalen> ah, well :) 14:06:01 <bcotton> but if IoT is the winner, i'm not sure we need to 14:06:09 <pbrobinson> right 14:06:16 <pwhalen> right, are we going with IoT then? 14:06:19 <bcotton> unless we want to say "to heck with your votes, we don't like this" 14:06:22 <pwhalen> heh 14:06:40 <pbrobinson> well we can't really got with Frontier as it's a US telco 14:07:00 <bcotton> mattdm and i had a conversation about this yesterday. i like the idea of not having a separate brand name to dilute the fedora brand (as I worry silverblue, e.g., does), he is wrong ;-) 14:07:20 <jsmith> "It's complicated..." 14:07:34 <pbrobinson> right 14:07:34 <bcotton> yeah, i probably should have removed frontier initially because telcos are pretty into the iot space 14:07:49 <jsmith> (I say that, having been in mattdm's shoes, and having had many many conversations about trademark law) 14:08:30 <pbrobinson> it also is a little "wild west frontier" which is probably a little like "bleeding edge" (pun intended) depending on how you look at it 14:09:20 <bcotton> so yeah, unless someone says "no, we are not going to use Fedora IoT is the name. pick something better", i think we can just call this done 14:10:12 <pbrobinson> well we could look at going to Fedora Edge, but it's a bit swings vs round-a-bouts 14:12:13 <bcotton> i'm very content to go with 'Fedora IoT' and call it done :-) 14:12:31 <pwhalen> +1 14:13:27 <pwhalen> Anything else admin? 14:13:56 <tdawson> I'm good with either, (IoT or Edge) ... but if I had the choice, I'd pick Edge because it's easier to type. 14:14:09 <bcotton> nothing here 14:14:18 <pwhalen> heh :) 14:14:24 <pbrobinson> nothing from me 14:14:34 <pwhalen> #topic 2) ==== Fedora 31 status ==== 14:15:07 <pwhalen> I dont think anything has changed here, any new issues reported? 14:15:17 <pbrobinson> nothing new from me 14:15:50 <pwhalen> #info No issues reported. 14:16:07 <pwhalen> #topic 3) ==== Fedora 32 status ==== 14:16:38 <pwhalen> #info Latest compose available for testing 14:16:39 <pwhalen> #link https://kojipkgs.fedoraproject.org/compose/iot/Fedora-IoT-32-20200401.0/compose/ 14:17:14 <pwhalen> #info Testing results - https://fedoraproject.org/wiki/User:Pwhalen/QA/Fedora-IoT-32-20200401.0 14:17:56 <pwhalen> started testing this morning to check the tpm2-tss update, working as expected. 14:18:21 * lorbus just joined 14:18:39 <lorbus> sorry I'm late. 14:18:46 <pwhalen> Ignition is also working better on the disk images, I ran a script and a few have shown up, install on the Fitlet2 showed up as well.. was able to get the ssh key copied over via zezere 14:19:43 <tdawson> Does that mean I can try with just a dd install and it should/might show up on the site? 14:19:59 <tdawson> Or that you ran a script after it was up? 14:20:00 <pwhalen> ignition failed on the RPi3 - https://paste.centos.org/view/bf4008d1 14:20:31 <pbrobinson> tdawson: that's the hope 14:20:46 <puiterwijk> pwhalen: yeah, I know that the actual ignition run for now fails. But at least it should put the key in place 14:21:35 <pbrobinson> puiterwijk: what's the issue there if it's a quick summary? 14:21:55 <tdawson> puiterwijk: That doesn't make sense. How can it put the key in place if it can't run and/or connect to the server. 14:21:58 <puiterwijk> pbrobinson: it doesn't like that the system is already fully up and stuff are mounted. 14:22:05 <puiterwijk> tdawson: it runs and connects to the server 14:22:06 <pwhalen> puiterwijk, ok, the rpi3 didnt show up when I looked in zezere, perhaps I didnt wait long enough 14:22:22 <pbrobinson> puiterwijk: ah, OK 14:22:27 <puiterwijk> It's just that there are like 6 different phases of Ignition, and I'm just running all, while some of them won't work in a fully booted system 14:22:54 <pbrobinson> puiterwijk: should be the same use case as a cloud image right? 14:22:55 <tdawson> I waited 10 or 15 minutes, and it never showed up. 14:23:00 <puiterwijk> I just didn't analyze which of the stages/phases performed which tasks, so I just run them all in sequence. The one putting SSH key there is working though 14:23:06 <puiterwijk> tdawson: after enabling and starting the timer? 14:23:13 <pbrobinson> tdawson: using today's compose? 14:23:23 <puiterwijk> pbrobinson: cloud image runs ignition in initramfs 14:23:26 <tdawson> Not todays compose 14:23:46 <pbrobinson> tdawson: it's likely only fixed in today's compose 14:23:46 <lorbus> should I set up a meeting with the CoreOS team so we can do a deep dive into Ignition? 14:24:09 <puiterwijk> lorbus: eh, I know how it works (had to read the code because there's not really documentation that's useful). But I just didn't bother 14:24:10 <tdawson> Ah, ok. Writting it on a card now ... 14:24:35 <puiterwijk> lorbus: it's just a case of me not wanting to spend too much time on that part of the stuff if it works enough for the key insertion for now. 14:24:47 <lorbus> I feel like there communication between IoT and CoreOS could be improved...also wrt to other tools like Zincati and Airlock which could be very useful for IoT's purposes 14:24:59 <pwhalen> tdawson, great. Let us know how it goes 14:25:07 <lorbus> ack. just let me know, happy to help :) 14:25:25 <puiterwijk> Anyway, pwhalen, if the SSH key doesn't get installed or the device does not show up, just ping me out-of-meeting with the full logs of the last service run 14:25:39 <puiterwijk> Because right now I only got the last 3 or so lines 14:26:16 <pwhalen> puiterwijk, ok, will do. Will also try the Pine64 and see how it goes. 14:26:54 <pwhalen> lorbus, did you have a chance to work out a quick test for Greenboot? 14:27:08 <pbrobinson> lorbus: maybe, but I don't necessarily agree 14:28:45 <pbrobinson> lorbus: because Airlock is for cluster co-ordination, it's a focus on k8s, I don't see how that would help here for device edge 14:29:19 <lorbus> No the focus for Airlock and Zincati is not for K8s at all 14:29:30 * coremodule is here, got the wrong time again... 14:29:33 <lorbus> it's for managing a fleet of devices 14:29:35 <puiterwijk> Anyway. Let's do this out of meeting? 14:29:45 <lorbus> but we exclude both in OKD for example 14:29:48 <lorbus> because we manage updates differently there 14:30:16 <pbrobinson> lorbus: airlock according to the docs seems to be a cluster coordinator so that all the cluster doesn't reboot at once 14:30:24 <pbrobinson> puiterwijk: right 14:31:08 <lorbus> yeh, it's a cluster of multiple devices. But those devices wouldnt have to run K8s 14:31:18 <pbrobinson> anyway, we need testing with tpm2 and the new tpm2-tss package 14:31:44 <pwhalen> pbrobinson, I tested todays compose on the fitlet2, added karma 14:31:59 <lorbus> it would help with rolling out a change to those devices with a policy, i.e. not all at once 14:32:13 <pbrobinson> lorbus: please move the discussion out of the meeting 14:32:24 <pbrobinson> or at least to general business 14:32:41 <pwhalen> our testing template for iot lives here - https://fedoraproject.org/wiki/User:Pwhalen/QA/IoT_Test_Template 14:32:54 <lorbus> sorry matrix is lagging here. that's all I have for this :) 14:33:31 <pwhalen> I need to finish the zezere-ignition test case, do something for http boot. 14:33:39 <pwhalen> Is there anything else we should be testing? 14:33:59 <pwhalen> we also need something for Greenboot. 14:35:02 <pbrobinson> pwhalen: I think lorbus was working on some tests for greenboot 14:35:06 <lorbus> greenboot needs some fixes still for the auto-reboot behavior. I've written the code and I'll test them tonight or tommorrow when I get to it. 14:35:27 <pbrobinson> lorbus: what's the ETA on a new release there? 14:35:44 <pbrobinson> we need to get things stable for next Monday which is freeze 14:36:22 <lorbus> I have some prio work from my mgtm, but I can squeeze it in before Friday 14:36:34 <pwhalen> lorbus, we can help with testing as well 14:36:39 <lorbus> we'll probably have a new build tomorrow if it works out 14:37:22 <lorbus> https://github.com/LorbusChris/greenboot/tree/devel contains the fixes, and I have a build on COPR already 14:37:27 <pwhalen> lorbus, sounds good, thanks! Ping us to test when you're ready 14:37:30 <lorbus> https://copr.fedorainfracloud.org/coprs/lorbus/greenboot/ 14:38:08 <pwhalen> if folks could also test things and add results to the wiki that would be greatly appreciated 14:39:55 <pwhalen> #info IoT Test Day scheduled next week - April 8th. 14:40:58 <pwhalen> In addition to the tests we ran the last couple of times, I'll have the zezere-ignition test ready 14:41:13 <pwhalen> also Clevis 14:41:31 <pwhalen> Anything else we want to test before final? 14:42:39 <pwhalen> Anything else for F32? 14:43:04 <pwhalen> #topic 4) ==== Open Floor ==== 14:43:30 <pwhalen> Anything for open floor? 14:43:41 <bcotton> me! 14:44:33 <bcotton> #info pbrobinson will be the guest of honor at the next Fedora Council video meeting. Wednesday 8 April at 1400 UTC 14:47:03 <pwhalen> looks like thats it. 14:47:05 <pbrobinson> right on time for our test day! 14:47:24 <pwhalen> heh. 14:47:30 <pwhalen> #endmeeting