[03:26] amurray: if you're happy with https://github.com/snapcore/snapd/pull/8943, could you mark your review as approved? [03:26] PR #8943: wrappers: generate D-Bus service activation files [04:22] PR snapd#9647 opened: vendor: upgrade to godbus v5.0.3 [06:52] morning [07:21] good morning [07:28] zyga: hey [08:08] morning [08:09] hey pstolowski [08:24] mvo: pstolowski: hey [08:27] o/ [08:28] good morning mborzecki ! [08:28] mborzecki: anything I should review urgently? [08:28] mvo: not really, nothing urgent to review [09:03] PR snapd#9645 closed: bootloader/grub: replace old reference to Managed...Blr... with Trusted...Blr [09:05] mborzecki: I removed "blocked" from 7614 but it does need a security review, yes? [09:06] zyga: 9384 has some conflicts now, should I merge master and de-conflict? [09:08] mvo: it was reviewed by amurray, or you mean we need another pass before landing? [09:09] mborzecki: oh, I did not see his +1 [09:10] mborzecki: the discussion is long, if it did got a security review I will remove the label. I guess it means we can merge this now, yes? [09:11] mvo: yes [09:13] cool [09:18] PR snapd#7614 closed: cmd/snap-confine: implement snap-device-helper internally [09:18] mvo: wooooooooot [09:55] is the store flaky again? tests are failing randomly [10:06] mborzecki: it's possible [10:18] PR snapd#9648 opened: spread: UC20 no longer needs 2GB of mem [10:38] PR snapd#9649 opened: seed: make a shared seed system label validation helper [11:03] PR snapd#9648 closed: spread: UC20 no longer needs 2GB of mem [11:22] mborzecki: one remark re #9649 [11:22] PR #9649: seed: make a shared seed system label validation helper [12:39] PR snapd#9650 opened: tests: skip boot state test on arm devices [12:55] looks like f33 switched to systemd-resolved completely [12:59] PR snapd#9643 closed: bootloader: use ForGadget when installing boot config [13:05] mborzecki: I commented on https://github.com/snapcore/snapd/pull/9649 [13:05] PR #9649: seed: make a shared seed system label validation helper [13:19] pedronis: thanks, you mean `^[a-zA-Z0-9](?:-?[a-zA-Z0-9])*$` then? [13:23] hm `^[a-zA-Z0-9]+(?:-[a-zA-Z0-9]+)*$` a model name maybe by model-like-names separated with dashes [13:26] mborzecki: look again, sorry, formatting issues [13:27] mborzecki: also I remembered something, this might need a tweak, a different tweak :/ [13:41] mvo: I reviewed #9555 [13:41] PR #9555: asserts: implement "storage-safety" in uc20 model assertion [13:42] mborzecki: can you join the standup a bit earlier to talk about labels? [13:42] pedronis: sure [13:42] morning folks [13:44] ijohnson: hey [13:44] hey zyga_ === zyga_ is now known as zyga-x240 [13:59] PR snapd#9651 opened: tests: fix basic20 test on arm devices [14:48] PR core20#95 opened: Rpi kernel names [14:48] waveform: https://github.com/snapcore/core20/pull/95/files this seems to work, and do the right thing on rpi2, 3, 4 did a subset of models. [14:48] PR core20#95: Rpi kernel names [14:49] waveform: will build that into a core20 image and will attempt deployment via testflinger, if that is good, will merge that. [14:52] hmmm lxd-services-smoke failed in https://github.com/snapcore/snapd/pull/9556 too [14:52] PR #9556: tests: testing new fedora 33 image [14:55] mvo, hi [14:55] mvo, I see this https://paste.ubuntu.com/p/RkNGtbMFYc/ [14:55] in pi4 [14:57] mborzecki, ~ [14:57] no more info in dmesg and journal logs [14:59] cachio: oh, interessting. I have not seen this one yet [14:59] mvo, not sure where else I could check [15:00] cachio: not sure either, have a meeting now unfortunately [15:00] mvo, ok, np, I'll continue the research [15:00] tx [16:11] * cachio lunch [16:13] aaah i see the bug in services [16:19] doServiceControl keeps state locked when calling wrappers [16:38] * zyga EODed his day job [16:38] well, a while ago, now fixed wifi [16:38] pedronis did you have the time to look at export manager patches? [16:39] zyga: not today [16:39] that's okay, if you can make a review pass by EOW I can spend some time on it this weekend [16:39] I think it's close, the "cork" patch, and the rest should be easier to work with [16:54] pstolowski: I updated 8929 (install-mode: disable), I think it needs some more tests but a quick check if it looks sane (it's short) woudl be great [16:54] pstolowski: no rush, does not have to be today of course [16:55] mvo: ok [16:55] PR snapd#9625 closed: snap: add new "fde-setup" hooktype [17:04] pstolowski: I reviewed #9580 [17:04] PR #9580: store: download timeout [17:04] pedronis: ty [17:10] PR snapd#9652 opened: o/servicestate: unlock state before calling wrappers in doServiceControl [17:18] pstolowski: ^ I left something there too [17:18] pedronis: good point, thanks === ijohnson is now known as ijohnson|lunch === ijohnson|lunch is now known as ijohnson [20:30] cachio: when you get a chance could you take a look at https://github.com/snapcore/snapd/pull/9646 too ? [20:30] PR #9646: tests/many: enable some uc20 tests, delete old unneeded tests or TODOs [20:30] I just reviewed #9651 for you, I think you can borrow some of the checks I wrote there [20:30] PR #9651: tests: fix basic20 test on arm devices [20:44] ijohnson, sure [20:44] thanks [20:49] ijohnson, left a comment [20:49] let see if the results [20:50] tests are apssing [20:50] passing [20:51] there is a weir error building arch [20:52] it is not related to this change [20:54] cachio thanks I responded [20:55] Yeah I'm looking into the other errors there they are all unrelated afaict [20:56] today I was working with the test core/fsck-on-book [20:57] ijohnson, for uc20 on arm64 [20:57] It is trying to umount umount /run/mnt/ubuntu-seed/systems/*/snaps/snapd_*.snap but it doesnt exist [20:57] I update that to umount /run/mnt/ubuntu-seed/snaps/snapd_*.snap instead [20:57] on arm [20:58] but still the appearing the text "Volume was not properly unmounted" after reboot [20:58] ij do you have any idea about how to make he umount on arm? [20:58] ijohnson, [21:04] cachio: hmm that's interesting [21:05] cachio: on such a system what is `snap list` and `cat /var/lib/snapd/modeenv` ? [21:16] ijohnson, let me reserve one [21:16] because I dont have a pi4 here [21:17] cachio: ok, I imagine that what is probably happening is that /run/mnt/snapd is only mounted on the very first boot, and on subsequent boots it is not mounted at all [21:17] the logic should be smart enough to avoid that but it might not be [21:19] ijohnson, the test has this comment [21:19] # FAT uses a specific byte to effectively indicate that the file system is [21:19] # dirty. The precise details as to how this byte is used by each system vary, [21:19] # but Linux sets it on a non-read-only mount, and clears it on unmount. We [21:19] # can set it manually, verify it when the image is mounted and observe fsck [21:19] # clearing it. Note that larger block devices use FAT32 and the offset [21:19] # differs. FAT12 and FAT16 uses 37 while FAT32 uses 65. [21:19] so it does the umount and reboot [21:19] but next time the dirty byte seems to be set again [21:20] but that works fine on amd64 [21:20] well if the umount fails then the dirty byte will be set [21:20] the umount does not fail [21:20] I fixed that [21:20] the umount is done correctly [21:21] so [21:21] it is a bug or I am not umounting correctly [21:21] not sure which one [21:22] hmm [21:22] let me look at the test again [21:23] cachio: ah I see the problem [21:23] cachio: the test does this: [21:23] dd if=one of=/dev/sda2 seek=65 bs=1 count=1 conv=notrunc [21:24] /dev/sda2 is wrong there on arm devices, the actual partition that it should be using is ubuntu-seed, which on a pi using an SD card is probably /dev/mmcblk0p1 [21:24] cachio: probably you should instead just do the easy thing and instead change that line to something like [21:24] dd if=one of=/dev/disk/by-label/ubuntu-seed seek=65 bs=1 count=1 conv=notrunc [21:25] cachio: same for the line below it too [21:25] dd if=/dev/sda2 of=dirty skip=65 bs=1 count=1 conv=notrunc [21:25] that should become [21:25] dd if=/dev/disk/by-label/ubuntu-seed of=dirty skip=65 bs=1 count=1 conv=notrunc [21:26] nice, I'll try that [21:26] ijohnson, thanks [21:29] ijohnson, is this the same for all the arm devices right? [21:29] pi2 pi3 on uc16 and uc18? [21:29] cachio: if you use the /dev/disk/by-label/ubuntu-seed, that location will be correct for all uc20 devices, amd64 or arm64 or armhf, etc. [21:29] cachio: for uc16 and uc18 it would be a different label [21:29] cachio: for uc16 and uc18 you would need to use something like this: [21:30] dd if=/dev/disk/by-label/system-boot ... [21:30] because the vfat partition on uc16 / uc18 is with the label system-boot [21:31] ijohnson, nice, thanks