15:04:11 <pwhalen> #startmeeting Fedora ARM and AArch64 Working Group Meeting
15:04:11 <zodbot> Meeting started Tue Apr 26 15:04:11 2022 UTC.
15:04:11 <zodbot> This meeting is logged and archived in a public location.
15:04:11 <zodbot> The chair is pwhalen. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
15:04:11 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:04:11 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_working_group_meeting'
15:04:18 * jlinton waves
15:04:38 <pwhalen> #chair pwhalen pbrobinson coremodule jlinton
15:04:38 <zodbot> Current chairs: coremodule jlinton pbrobinson pwhalen
15:04:40 * coremodule is here
15:04:45 <pwhalen> #topic roll call
15:04:56 <pwhalen> Sorry for the late start, got distracted
15:05:06 * nirik is in another meeting, but wants to bring up one 32bit arm bug in case anyone has ideas on it.
15:05:11 <jlinton> Your not the only one, had to run back to my desk.
15:05:40 <pwhalen> nirik: do you want to go ahead now?
15:06:54 <nirik> sure... https://bugzilla.redhat.com/show_bug.cgi?id=2077680 is a fun wacky bug... armv7 guest vms start with a time set to last year sometime.
15:09:41 <jlinton> Yah, not much to add besides its likely a qemu/edk2 mismatch that keeps the uefi time from being right.
15:11:18 <pwhalen> #info Bug 2077680 - 32bit arm vm on aarch64 has date wrong
15:11:23 <nirik> would a newer edk2-arm be worth trying?
15:11:43 <pwhalen> hrm, I havent hit it recently but have seen it before at some point
15:12:42 <nirik> I wouldn't care at all, but due to the date ssl certs are invalid. :(
15:13:40 <pwhalen> right, edk2-arm hasn't been updated since December
15:16:08 <pwhalen> its just happening on the container composes?
15:17:26 * pwhalen was just trying locally
15:18:32 <pwhalen> installation works ok for me on an f35 host
15:19:10 <pwhalen> oh, its an f34/35 install
15:19:45 <pwhalen> I tried an F35 installation yesterday and hit that resolv.conf bug on aarch64
15:20:34 <pwhalen> #topic 1) ==== Stable Releases ====
15:20:54 <pwhalen> Any new issues in F34/35?
15:22:43 <pwhalen> #info No issues reported.
15:23:00 <pwhalen> #topic 2) ==== Kernel Status ====
15:24:31 <pwhalen> #info kernel-5.18.0-0.rc4.33.fc37
15:24:31 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1955556
15:25:23 <pwhalen> I need to look at 5.18. Has anyone had a chance to do some testing?
15:26:20 <pwhalen> #info kernel-5.17.4-300.fc36
15:26:20 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1953622
15:26:44 <nirik> yeah, f34/f35 container images (failing now for like a month).
15:26:55 <pwhalen> #info Please test and report any issues to #fedora-arm or the mailing list.
15:27:44 <pwhalen> thanks nirik, I'll try to reproduce to see if I can add anything. When I've attempted installations in F35 I hit another bug
15:28:00 <pwhalen> This might explain why armhfp is no longer in the manifest
15:28:46 <pwhalen> #topic 3) ==== Bootloader Status ====
15:29:07 <pwhalen> I think we're ok here, no issues reported.
15:29:15 <jlinton> I've been running a couple machines (rpi/seattle) on with 5.18 series for a week or two.
15:29:38 <pwhalen> Excellent, at least someone is testing it :)
15:29:39 <pwhalen> thanks
15:29:41 <nirik> yeah, that stupid resolv.conf thing? supposed to be fixed in a new systemd update, not sure if it's stable tho. You can install without updates repo to avoid it.
15:30:33 <pwhalen> nirik: thats the one. It has broken f35 iot for a long time
15:30:42 <pwhalen> I'll try that
15:30:58 <pwhalen> #info No issues reported.
15:31:02 <nirik> it worked at GA, a systemd update broke it. ;( We have hit it installing infra machines too.
15:31:36 <pwhalen> Right. Frustrating.
15:31:52 <pwhalen> #topic 4) ==== Fedora 36 status ====
15:32:00 <pwhalen> How does F36 looks?
15:32:18 <jlinton> Well, I just discovered that the f36 virt-manager can't create a 32-bit guest properly
15:32:25 <pwhalen> #info https://fedoraproject.org/wiki/Test_Results:Fedora_36_RC_1.1_Installation?rd=Test_Results:Current_Installation_Test
15:32:45 <jlinton> "unsupported configuration: tpm model tpm-tis is only available for x86 and aarch64 guests"
15:32:46 <pwhalen> jlinton: yes, I discovered that yesterday as well. tpm-tis stuff?
15:32:53 <jlinton> yah
15:33:14 <pwhalen> I'll file a bug today if you dont beat me to it
15:34:10 <pwhalen> #link https://fedoraproject.org/wiki/Test_Results:Fedora_36_RC_1.1_Summary
15:34:41 <pwhalen> coremodule: how is Workstation looking on aarch64?
15:35:05 <jlinton> I guess I can go check off the ec2 tests, F36 has been pretty bulletproof on the gravaton3 I was running it on.
15:35:27 <pwhalen> #info Fedora 36 RC 1.1 did not include a 32-bit ARM Server image.
15:35:48 <coremodule> so far so good, no issues of note. have primarily done testing on pine64 so far
15:36:00 <pwhalen> The server image failed, looked to be an infra issue. So the arm swan song will be minus Server if 1.1 is GA
15:36:26 <pwhalen> Cool, does ethernet work for you on the pine64?
15:36:59 <pwhalen> Server is not release blocking for 32-bit arm
15:37:03 <jlinton> Although, i'm not sure some of those test cases are written properly for ec2, as you don't get an initial setup screens because the machine deploys with a working user/key configured by the deploy image.
15:37:19 <coremodule> yes, no issues with ethernet here
15:37:51 <pwhalen> coremodule: awesome. This confirms there is something wrong on my network.
15:38:15 <coremodule> it seems like local-network issues are the bane of our existence, hah
15:38:38 <pwhalen> jlinton: https://fedoraproject.org/wiki/Test_Results:Fedora_36_RC_1.1_Cloud
15:38:40 <pwhalen> those?
15:39:07 <pwhalen> Those are specific to cloud instances
15:39:25 <jlinton> I was looking at https://fedoraproject.org/wiki/Test_Results:Fedora_36_RC_1.1_Summary#aarch64 which has a ec2 column
15:39:53 <pwhalen> coremodule: it appears so :) , with your confirmation I will look into it here... very curious as to what it is
15:42:13 <pwhalen> Anything else for F36?
15:42:28 <jlinton> pwhalen: the test results link you have there isn't right either, the aarch64 public AMI's aren't listed only the direct download links
15:43:48 <pwhalen> jlinton: click on expand
15:44:10 <jlinton> ah yah, PEBCAK
15:44:45 <pwhalen> they have been missing recently though, so perhaps just another day you were looking
15:45:25 <pwhalen> #topic 5) ==== Open Floor ====
15:45:29 <pwhalen> Anything else for today?
15:47:44 <pwhalen> #endmeeting