[01:23] I got it working. Thanks. [01:24] PR snapd#10151 opened: wrappers/services.go: refactor helper lambda function to separate function [02:24] PR snapd#10152 opened: quota: new package for managing resource groups [02:40] PR snapd#10153 opened: [RFC] wrappers, quota: implement quota groups slice generation [05:51] morning [07:01] morning [07:06] pstolowski: hey [07:30] hmm hmm dropped console=ttyS0 from cmdline.full and spent a while wondering why my qemu serial is so silent [07:35] PR snapd#10154 opened: cmd/snap, boot: add snapd_full_cmdline_args to dumped boot vars [07:39] mborzecki uh oh [07:40] mborzecki approved [08:21] thx [09:11] PR snapd#10151 closed: wrappers/services.go: refactor helper lambda function to separate function [09:16] PR snapd#10154 closed: cmd/snap, boot: add snapd_full_cmdline_args to dumped boot vars [09:22] zyga, omg, krakow is attacked by the french ?!?! https://www.todayonline.com/world/is-it-bird-is-it-iguana-no-its-croissant [09:26] PR snapd#10155 opened: [RFC] o/snapstate: handle refresh hints [09:32] ogra I'm wearing a krakow t-shirt today ;-) [09:32] does it have a croissant on it ? [09:32] ogra only if we have some left upstairs [09:33] (I'm afraid not) [09:35] and runing a spread test with cmdline.full & secure boot [09:46] hm when running nested tests we compress the images using xz -k0 [09:46] wonder if using bzip/gz/zstd would speed this up a bit [09:47] also when we repack the kernel snap, we use xz again, we could use zstd there right? [09:48] duh, s/zstd/lzo/ for the snap [09:51] huh, how is that possible? ` Mount snap "snapd" (11588) (snap is unusable due to missing files; contact developer)` [09:51] mborzecki: gzip is probably fastest [09:51] mborzecki: uh, I never saw this message [09:51] mvo: that's from Adio's log [09:53] mborzecki: there should be more details from snap.ValidateContainer for this error [09:54] mborzecki: I have a meeting in 5 unfortunately, what PR fails like this? [09:55] mvo: that's not from a PR [09:56] Adio: can you `snap changes` and paste the output to some pastebin? [09:58] mborzecki: aha, sorry, missed that because I joined late [10:02] Adio: can you collec the output of `snap tasks 6` ? [10:03] pedronis: hi, i've proposed #10155 for early feedback re the direction taken [10:03] Bug #10155: System reboots after halt [10:03] PR #10155: [RFC] o/snapstate: handle refresh hints [10:04] Adio: hm interesting, can you collect the journal? `sudo journalctl -u snapd --since 15:36` [10:11] thanks [10:11] Adio: does /snap/snapd exist in your system? if so what does /snap/snapd/current point to (`ls -l /snap/snapd` will show that) [10:16] pstolowski: I'll try to look at it after lunch, I have few older PRs still in my review queue [10:17] pstolowski: thanks [10:17] hm no clue what's happening, it's as if the snapd snap file is corrupted? but that would probably give a different error [10:17] pedronis: sure. as i said naming TBD, tests are missing and existing tests will fail atm [10:29] Adio: can you try running `snap install --beta snapd` ? [10:29] Adio: what system is it anwyay? what does `snap version` show? [10:55] Adio: hm probably something changed in unsquashfs, what's the old and the new version? [11:23] Adio: thanks, does the new version even work at all? can you try `unsquashfs -l /var/lib/snapd/snaps/snapd_.snap`? [11:51] PR snapd#10156 opened: boot, bootloader, bootloader/assets: support for full command line override from gadget [11:56] mborzecki, hi [11:56] cachio_: hey [11:56] any idea if this is a new error? https://github.com/snapcore/snapd/pull/10135#issuecomment-819955829 [11:56] PR #10135: tests/lib/nested: proper status return for tpm/secure boot checks [11:57] it is failing because of memory problems [11:57] this is the only test failing [12:00] cachio_: not really, that's the minimal test in https://github.com/snapcore/snapd/pull/10135 which means we were not executing that test with secure boot enabled, and maybe it jsut doesn't work with 384MB of ram [12:00] PR #10135: tests/lib/nested: proper status return for tpm/secure boot checks [12:00] mvo: ^^ that's probably interesting to note too [12:01] mborzecki: in a meeting will have a look [12:01] let me add that to the standup notes so that we don't forget [12:03] mborzecki, but we were not using secure boot during the last 20 days [12:04] after we merged core and core20 suites [12:04] before we were using it [12:05] not sure if any change we did lastly could affect that test [12:05] cachio_: afaiu the change in 10135 and your last commit there, make it used by default now on ubuntu-20.04 [12:06] mborzecki, you are right [12:06] manual tests were not using sec boot [12:07] mborzecki, should that test run with sec boot? [12:07] or it is ok to run without? [12:11] mborzecki, I can add environment: [12:11] NESTED_ENABLE_TPM: false [12:11] NESTED_ENABLE_SECURE_BOOT: false [12:11] to the test [12:11] cachio_: that's a good question and the answer depends on what we promise ;), let's discuss it during the standup [12:11] mborzecki, sure [12:55] pstolowski: thanks for https://github.com/snapcore/snapd/pull/10155 , left a bunch of comments and some worries, let me know [12:55] PR #10155: [RFC] o/snapstate: handle refresh hints [14:12] mborzecki, I have tested minimal smoke with 512MB and works [14:12] cachio_: with secure boot enabled right? [14:12] yes [14:13] cachio_: can you also try to run it with 384MB like before, but with disabled tpm/secure boot? [14:13] sure [14:13] eventually maybe we should keep both variants of the test [14:13] mvo: ^^ [14:14] mborzecki: +1 (but in a meeting right now) [14:19] pedronis: sorry I think I forgot to look at the frankfurt notes to see that it was decided to _not_ allow having different quota groups for services in a single snap, that does simplify things I suppose, but to be clear that's the implementation we want for now ? [14:19] to have a single group for all services in a snap ? [14:20] yes [14:20] that does go a bit against our recommendation that we give to customers to bundle services into a single snap, since now all the services in a single snap would be subject to the same resource quota [14:21] mborzecki, you mean to run with 384 and not secboot and with 512 and secboot right? [14:21] cachio_: yeah [14:21] mborzecki, I can try that [14:24] ijohnson: I was not even aware of this reccommendation [14:24] is that just because then it's easier to make them talk to each other? [14:25] pedronis yes, at least that's what was the recommendation when I was on the field team [14:25] The other reason is due to refreshes etc and keeping different services in sync [14:27] either way, we can't start doing services as first step, it gets pretty convoluted quickly state wise [14:29] ok, I thought it was fairly straight forward to just put a quota on an appinfo like I showed, but as mentioned I have not finished plumbing everything together [14:29] ijohnson: but we don't really keep app infos around like that [14:29] that's why as I said there's need some thinking about what we actually stick into the state [14:30] oh I see, I didn't realize we don't serialize app-infos like that [14:30] ijohnson: SnapState doesn't contain a snap.Info nor snap.AppInfos [14:30] we read them always from disk combined with SideInfo data from SnapState [14:31] mmm right [14:31] so not even clear that putting the group in snap.Info is useful or confusing [14:32] Okay, I'll have a look today [14:32] at the wrappers level it could come via SnapServicesOptions [14:32] but not sure [14:33] there is more than one way to do this, but ultimately we need to keep what we did/need to do in state [14:33] Right [14:46] pedronis: ijohnson: i've updated #10148 [14:46] Bug #10148: xfree86: new changes from Debian require merging [14:46] PR #10148: overlord/devicestate, overlord/snapstate: add task for updating kernel command lines from gadget [14:53] ack [15:29] mborzecki: I re-reviewed it, I'm looking at the full one next. OTOH that one probably can use as many eyes as possible [15:29] I mean the task one [15:30] pedronis: understood, thanks for the review [15:57] PR snapd#10157 opened: o/snapstate: remove unused DeviceCtx argument of ensureInstallPreconditions [15:59] * cachio_ lunch [16:20] mborzecki around? [16:20] I reviewed https://github.com/snapcore/snapd/pull/10156 thanks [16:20] PR #10156: boot, bootloader, bootloader/assets: support for full command line override from gadget [16:32] PR snapd#10158 opened: packaging: build without dwarf debugging data === popey8 is now known as popey [18:17] PR snapd#10135 closed: tests/lib/nested: proper status return for tpm/secure boot checks [18:28] * cachio_ afk [19:10] PR snapcraft#3503 opened: deb: do not filter python3 packages on core20 [19:37] PR snapd#10159 opened: secboot: use secboot.{ActivateVolumeWithKeyData,SnapModelChecker} [23:15] PR snapcraft#3495 closed: conda v2 plugin: support for more architectures