16:00:13 #startmeeting ELN (2022-07-29) 16:00:13 Meeting started Fri Jul 29 16:00:13 2022 UTC. 16:00:13 This meeting is logged and archived in a public location. 16:00:13 The chair is sgallagh. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:00:13 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:13 The meeting name has been set to 'eln_(2022-07-29)' 16:00:13 #meetingname eln 16:00:13 The meeting name has been set to 'eln' 16:00:13 #topic init process 16:00:13 .hi 16:00:14 sgallagh: sgallagh 'Stephen Gallagher' 16:00:33 Howdy 16:00:40 #meetingname eln 16:00:40 The meeting name has been set to 'eln' 16:00:42 sgallagh: Error: Can't start another meeting, one is in progress. 16:00:44 #topic init process 16:00:47 .hi 16:00:48 sgallagh: sgallagh 'Stephen Gallagher' 16:01:21 Ah there we go 16:01:32 Yep ... we're now officially a meeting. 16:01:53 Ahoy, Troy 16:02:10 I not only remembered this week, but I'm on time. :) 16:03:10 pbrobinson: Do you happen to be around to answer some questions about IMA signing? That's one of our agenda topics today. (Sorry for the lack of notice) 16:03:58 Davide Cavalca, Conan Kudo Are you around today? 16:04:35 we're both at SCALE this week 16:04:40 Ahh 16:04:41 I'm around right now though 16:05:05 I have two topics for discussion, both hopefully quick: 16:05:44 1) Do we want to enable IMA signing for ELN packages? 16:05:44 2) We've got kernel issues preventing the creation of aarch64 and ppc64le container images, which is blocking Anaconda efforts. 16:05:55 #topic IMA Signing 16:06:08 #link https://fedoraproject.org/wiki/Changes/Signed_RPM_Contents 16:06:51 This is currently enabled for all non-ELN Fedora packages. 16:07:26 I'll admit, I don't really know what value IMA signing provides to Fedora, so I'm not sure if we want it for ELN either. 16:08:12 is IMA signing something that will eventually end up enabled in RHEL 10 / CentOS 10? 16:08:14 Yes 16:08:27 then it seems reasonable to do it in ELN 16:08:45 But it's essentially a verification tool, confirming that files from RPM have not been tampered with 16:09:15 So naturally things will be signed with a different key in Fedora, CentOS Stream and RHEL 16:09:17 In my mind it's like rpm -V ... but not needing to go through rpm. 16:09:58 I'm uncertain if the value outweighs the resources needed to do the signing. 16:10:06 hello sorry late 16:10:20 And I'm not comfortable with "Eh, why not?" as a justification :-) 16:10:26 In Fedora, when do the packages get these IMA signatures? 16:10:58 I assume when the rest of the RPM is also signed, but I haven't looked. 16:11:17 my general feeling is that we want to be as close as possible to what the next RHEL/CentOS will be, so that folks can use ELN to get a preview of what's to come and get ahead of any surprises 16:11:22 I think the part of sigul to do this is being worked on currently 16:11:47 in this case, even if the key will be different, the existance of the feature and it being enabled makes it possible for folks that want to leverage it to start evaluating things and playing with it 16:11:51 sigul is the tool which does signatures of diferent types 16:12:02 smooge: nirik asked me the other day if we want it enabled, which I took to mean that the means already exist 16:12:21 Davide Cavalca: That's fair enough. 16:12:24 yes, it's just some config we need to set now... 16:12:33 well they were actively doing work on part of the system that day and I think it was a 'do I need to drop this config in' 16:13:22 I am with davide on this. It is going to take 'work' in various parts of enterprises to use this 16:13:38 .hello ngompa 16:13:39 * Eighth_Doctor waves 16:13:39 Eighth_Doctor: ngompa 'Neal Gompa' 16:13:46 * sgallagh waves back 16:13:56 .hi 16:13:56 IMA is already enabled in EL9 16:13:56 salimma: salimma 'Michel Alexandre Salim' 16:13:57 if most of the enterprises only get to use it when EL10 comes out, I doubt they will use/deploy IMA until EL11/12 16:14:02 so we should have it in ELN 16:14:28 smooge: As Conan Kudo noted, RHEL 9 also shipped with it 16:14:29 * Eighth_Doctor remembers the pain around IMA enablement in EL9 16:15:14 well colour me surprised.. I watched Fedora more closely than EL for once 16:15:33 thank you 16:15:41 * sgallagh throws a bucket of paint at smooge 16:16:02 I didn't know paint came in the color "suprise" 16:16:06 OK, I'm not hearing anyone say "we shouldn't do this". 16:16:20 tdawson, you clearly didn't have the painters we had 16:16:20 Anyone want to dissent, or shall we go ahead with signing? 16:16:37 I guess I haven't said ... I say go ahead with the signing. 16:16:38 tdawson: You might be surprised! 16:16:43 *laughs* 16:17:19 #agreed In order to remain as close to RHEL as possible, we will enable IMA signing for Fedora ELN 16:17:28 #topic aarch64 and ppc64le container image generation 16:17:28 #link https://koji.fedoraproject.org/koji/taskinfo?taskID=90211307 16:17:37 do note that this will just sign things as theyr are built moving forward, not everything... 16:17:42 * nirik can enable it. :) 16:18:20 nirik: That's fine, worst case is that we pick it up at the next mass-rebuild 16:18:43 #action nirik will enable the IMA signing for ELN 16:18:44 Thank you, nirik 16:18:56 would it help to have a releng ticket for that nirik ? 16:19:13 I've worked out almost all of the various compose issues finally, but the last is container images for aarch64 and ppc64le 16:19:45 meh, I was just gonna do it right now... 16:20:33 I'll be honest; I really have no idea where to go next in terms of fixing these two 16:22:03 (This is me not-so-subtly asking someone to step up and help) 16:23:17 ... 16:23:35 (notes everyone taking one step back) 16:24:18 Yeah, that's what I was afraid of 16:24:38 OK, I'll continue to knock my head against it for a while longer 16:24:40 So ... what is the actual error happening. 16:24:58 tdawson: That's the problem: I can't find out 16:24:59 how urgent is this? I'll have more time in a couple more weeks 16:25:11 and have a fast local aarch64 machine I can use to debug 16:25:14 The image build process is hitting the 300 minute timeout and bailing. 16:25:20 wow, that's... slow 16:25:23 Err, sorry. 300s timeout without disk access 16:25:50 But the only view we have of what's happening there is a screenshot of the VM which looks... normal 16:25:50 https://kojipkgs.fedoraproject.org//work/tasks/960/90190960/oz-ppc64le.log 16:26:46 good ol oz. ;( 16:27:04 Yeah, it's going to take a wizard to solve this one... 16:27:11 sgallagh: if you can start one I might be able to get in on the libvirt console and see more... 16:27:43 I'm not sure how to start just the image creation 16:27:55 And the full compose takes about 2 hours 16:28:08 it looks like it finished the install... something in post install failed. 16:28:28 After the install, it reboots to see if it runs 16:28:33 IIUC 16:28:41 The screenshot looks like it just hung mid-boot 16:28:58 #link https://kojipkgs.fedoraproject.org//work/tasks/1318/90211318/screenshot.ppm 16:29:36 Wait... this looks different this time 16:30:08 I think I may have been only looking at the aarch64 screenshot 16:30:53 This may be actionable... 16:31:05 OK, no need to hold the meeting open to debug this. 16:31:11 #topic Open Floor 16:31:27 humm... wait a sec... I think I might have seen this before. 16:33:28 or at least something very like it. can discuss in #fedora-releng I guess. 16:33:43 OK, thanks 16:33:44 Any topics for Open Floor? 16:33:53 Nothing from me 16:34:48 OK, I'll hold the meeting open two more minutes for anyone to chime in. 16:37:31 close it 16:37:45 its full of lava 16:37:58 aaaaaaaa <> 16:40:16 #endmeeting 16:45:03 and not being chair or maybe on the internet won't allow me to end the meeting 16:48:11 It's ok ... sgallagh's sense of two minutes is longer than other peoples. 16:48:25 #endmeeting