[00:07] PR snapd#9073 closed: release: 2.45.3.1 [05:57] morning [07:23] PR snapd#9075 closed: daemon/api: use pointers to time.Time for debug seeding aspect [07:28] hello [07:32] gah [07:33] pedronis: hey [07:47] mborzecki: hi, can we chat after the desktop meeting? [07:47] pedronis: yes [07:48] pedronis: that unit test failure is only when building debs? [07:48] pedronis: i don't see it locally [07:48] PR snapd#9076 opened: interfaces: make the unmarshal test match more the comment [07:49] mborzecki: I opened this ^ [07:49] to tweaks the test because it wasn't actually testing what it said, but also add more debugging [08:00] pedronis: interesting, spread tests don't complain either, do you have more of a log where the failure occurred? [08:01] mborzecki: there's only false, [08:01] that's why I did we need more debugging [08:01] see my PR [08:01] *I think [09:19] PR snapd#9076 closed: interfaces: make the unmarshal test match more the comment [10:34] PR snapd#9077 opened: boot: add current recovery systems to modeenv [10:45] pedronis: ^^ [11:19] PR snapd#9078 opened: [RFC] boot: fancy marshaller for modeenv values [11:24] mborzecki: thx, finishing lunch break and then I will look [12:08] mborzecki: I reviewed #9074 , I dismissed Ian's review though becuase it has changed in some significative ways [12:08] PR #9074: bootloader: extend managed assets bootloader interface to compose a candidate command line [12:22] damn gofmt [12:28] pedronis: thanks for the review, i'll poke ijohnson or cmatsuoka when they are online [12:28] pedronis: #9077 should hopefully be easy [12:28] PR #9077: boot: add current recovery systems to modeenv [12:29] yes, I will look at it next [12:49] cmatsuoka: hey, do you have the script for booting in secure mode? something does not work in my setup and i'm not sure why [12:49] mborzecki: sure, just a sec [12:50] it's clearly doing something because the vm is significantly slower and getting stuck on low entropy occasionally [12:50] mborzecki: https://paste.ubuntu.com/p/mRHs9rx5QM/ [12:54] mborzecki: don't forget to re-sign shim with the snakeoil key [12:54] mborzecki: this is Ian's script to do that: https://paste.ubuntu.com/p/4CczzWSHzR/ [12:54] cmatsuoka: thanks! [12:59] cmatsuoka: which ovmf do you use? [12:59] (i mean version) [12:59] mborzecki: the one from the focal archive [12:59] cmatsuoka: hmm so it should be the same, idk i'm getting dropped to efi shell now [13:00] mborzecki: I'll check my steps here after the SU [13:35] PR snapcraft#3237 opened: spread: use host pip === ijohnson|lunch is now known as ijohnson [13:42] seems I forgot to leave lunch mode yesterday haha [13:47] mborzecki: dropping to efi shell happened to me if I didn't re-sign shim IIRC [13:48] cmatsuoka: ok, that would make sense, another question then, where do i get that snakeoil cert/key? or do you have a scipt to generate a custom one [13:48] mborzecki: it's in the ovmf package [13:49] ijohnson: can you take a look at https://github.com/snapcore/snapd/pull/9077 ? [13:49] PR #9077: boot: add current recovery systems to modeenv [13:49] mborzecki: sure [13:50] mborzecki: you can also get that snakeoil cert/key from the gadget snap repo in snapcore/pc-amd64-gadget [13:51] ijohnson: cmatsuoka: thanks, i see it in the ovmf package [13:51] heh funny how that package ships usr/share/OVMF and usr/share/ovmf [13:52] ofc both directories have differnt content [13:53] and a key with a password [13:53] it's "snakeoil" [13:54] yup, found that in docs [13:57] hmm still getting dropped to efi shell [13:58] btw. there's a problem with lxd tests, they reach kill timeouts, no matter the host (ubuntu, fedora etc) [14:01] yay, works with secure boot now [14:02] \o/ [14:03] cmatsuoka: i didn't have unit=0, unit=1 for the CODE/VARS entries, although code was listed first [14:11] hmmm Retrieving image: rootfs: 1% (99.00kB/s) [14:14] mborzecki: what's the thing about gofmt 1.9 in the Modeenv struct initialization? [14:14] I mean, what does it complain about? [14:15] cmatsuoka: gofmt changed the maximum line length around 1.9/10 iirc, so 1.n+1 formatting is different from earlier versions [14:15] ah ok [14:15] cmatsuoka: or not the max line length, but the maximum empty whitespace run, something like that [14:17] wow, gcp is also getting same downlaod speeds of lxd images: Retrieving image: rootfs: 3% (95.97kB/s) [14:17] wasn't there like a different remote rpeo we could use to get the images? [14:19] mborzecki: I had this kind of download rate with snaps some time ago, and it possibly triggered that write loop issue [14:20] mborzecki: yes but it has different images [14:21] mborzecki: ubuntu/foo [14:21] mborzecki: vs foo IIRC [14:21] mborzecki: those images don't have snapd and have other tweaaks [14:21] *tweaks [14:21] zyga-x240: hm well, the ubuntu repo is super slow atm, tests are failing [14:21] try lxc remote list [14:22] there are images/ubuntu-abc [14:22] and there's ubuntu/foo [14:22] and ubuntu-daily/foo [14:22] IIRC images is very fast [14:23] zyga-x240: it is, but our tests don't use that ;) [14:23] they could but they would need to be adjusted to cope with the different environment [14:24] cmatsuoka, about the reboots [14:24] cmatsuoka, could be possible we are rebooting the instace because of a race? [14:24] cmatsuoka, is it possible to track that? [14:25] wonder what's causing the load on cloud-images.ubuntu.com, boothole updates? [14:25] cachio: in our side? I don't think we're doing it [14:26] mborzecki: IIRC it was never fast [14:26] cachio: otherwise it would log something, there it looks like the VM is simply "power-cycled" [14:26] it was just some what okay [14:32] cmatsuoka, yes, makes sense [14:32] trying a new configuration now, I had 1 run without any reboot [14:32] but perhaps it was just lucky [14:32] I was [15:17] jdstrand, so it looks like we wont need any specific pcscd interface at all ... just an approved content interface (i.e. https://forum.snapcraft.io/t/auto-connections-for-pcsc-daemon/19170 ) should be enough ... [15:25] errands, bbl [15:31] jdstrand, also https://forum.snapcraft.io/t/module-blacklisting-interface/19171 ... [15:31] as requested 🙂 [15:32] (and i know you are busy with grub fixing ... so feel free to ignore as needed 😉 ) [15:49] * cachio lunch [15:56] cachio: I think 9027 is probably ready to be merged when it's green [15:56] I responded to your comment there [16:10] ijohnson: meh, hardly anything is green now [16:10] haha yeah probably [16:11] mborzecki: your fancy modeenv marshalling/unmarshalling looks good to me [16:11] mborzecki: I am going to do another pass on 9074 just so I can understand, but feel free to merge without my +1 [16:11] ijohnson: thanks, i think it'll be easier to put more complex things there [16:11] yeah [17:11] PR snapd#9074 closed: bootloader: extend managed assets bootloader interface to compose a candidate command line [17:21] PR snapd#9079 opened: gadget/install: retrieve command lines from bootloader [17:37] * cachio afk [18:24] PR core20#79 opened: Add secureboot-db package, try #2 [18:36] PR snapd#9077 closed: boot: add current recovery systems to modeenv [19:01] PR snapcraft#3237 closed: spread: use host pip [19:01] cmatsuoka, hey [19:01] I see this error: https://paste.ubuntu.com/p/pFMF5NJdDn/ [19:02] related to tpm [19:02] is it something expected? [19:03] let me see... [19:05] cachio: this is strange indeed [19:06] PR snapcraft#3236 closed: snap: use python3-apt stage-package [19:11] cachio: was it a normal installation run, or did you do something unusual in this test? [19:11] (like not clearing the tpm or something like that) [19:11] I am manually running [19:13] cmatsuoka, this is the full log https://paste.ubuntu.com/p/MVQCNz4RF3/ [19:16] cachio: the two error messages are different, so it failed twice? [19:18] cachio: the second one suggests that the tpm was not cleared before installing [19:19] sorry, the second was my fault [19:19] the first one is not something i've seen before [19:19] is it reproducible? [19:21] cmatsuoka, https://paste.ubuntu.com/p/zg3gKVQN5C/ [19:21] I am trying again [19:21] cachio: the last paste is a tail -f command :) [19:22] few lines before it is the error [19:22] this is the full log for that error [19:23] cachio: I mean, there's only a tail command in the paste [19:23] google-nested:ubuntu-20.04-64 .../tests/nested/core20/tpm# tail -f /tmp/work-dir/serial-log.txt [19:23] cmatsuoka, perhaps hte error was caused because a reboot [19:23] https://paste.ubuntu.com/p/zg3gKVQN5C/ [19:23] I see the full log there [19:23] line 2982 [19:23] cmatsuoka, [19:24] cachio: ah ok I got it, sorry [19:24] there's a lot of empty lines there [19:25] let me see where the reboots happened... [19:27] cachio: yes, maybe the crashes placed the tpm in some inconsistent state? [19:27] but that shouldn't happen, it's strange [19:28] cmatsuoka, how should I clean tpm [19:28] forgot that [19:29] you can delete the "permall" file in /var/snap/swtpm-mvo/current [19:30] thanks [19:30] tpm2-00.permall, but I think the number may change if you have more instances === Aavar_ is now known as Aavar [20:36] PR snapcraft#3235 closed: tests: fix assert ordering for error format tests [21:01] cachio: did you reproduce that error? [21:03] no [21:04] cmatsuoka, I was running the whole afternoon and didnt see that error again [21:04] hm interesting, ok, let me know if that ever happens again [21:05] in this case it seems that it was caused by the random reboots, which is also strange but... [21:47] PR snapd#9080 opened: osutil/disks: use xerrors to indicate a fs label wasn't found [22:22] PR snapd#9081 opened: secboot,cmd/snap-bootstrap: cross-check partitions before unlocking, mounting [22:47] PR snapd#9082 opened: interfaces/system-key: in WriteSystemKey during tests, don't call ParserFeatures