mupPR snapd#9510 opened: tests: new system-state tool <Created by sergiocazzolato> <https://github.com/snapcore/snapd/pull/9510>04:10
zygagood morning05:08
zygaamurray hey05:08
zygaI saw your PR, it looked good, I will review it now05:08
zygaamurray did you hit any bumps on the way? anything that I could have mentioned to make it easier?05:08
amurrayhey zyga  - thanks - I notice the test I added fails on arch - is there an easy way for me to try and debug this locally?05:08
zygaamurray sadly no, we don't have qemu images of arch05:09
amurrayzyga: no you were really helpful - thanks again05:09
zygaperhaps mborzecki - who is off for a family wedding today - could help05:09
zygayou can debug that locally if you have a GCE credentials05:09
zygathen you can run spread commands locally easily05:09
zygaI will look at the arch failure now05:09
amurrayI tried to create one locally but I don't have much experience with arch and have spent an hour or so trying to create an image without much success so can't justify much more on that approach...05:10
zygaamurray btw, can you rebase on master, I think mvo merged some more fixes so we could see a better (more complete) run this time05:10
zygayeah, it's hopeless05:10
zygathere are some helpers05:10
zygalike the systemd mkosi thing05:10
zygabut I have not tried using it05:10
zygabut it'd be my best bet if I were to try05:10
zygathe problem is that everyone on the core team has google cloud credentials so there's no incentive to improve local testing05:11
zygaamurray https://wiki.archlinux.org/index.php/Mkosi05:11
amurrayzyga: oh nice - I'll give it a go :)05:13
zygaamurray it can create various images05:13
zygasome more easily than others05:13
zygabut systemd is using it for testing itself05:13
zygaso it must work somehow: )05:13
amurrayalso I just rebased it on master05:13
zygatests are not in a happy place today05:14
zygaE: Failed to fetch https://packages.microsoft.com/ubuntu/16.04/prod/dists/xenial/main/binary-amd64/Packages  Writing more data than expected (934716 > 934130)05:14
zygaoh well, I'll just review what I can05:14
zygaand thank you so much for the spread test05:14
zygathat is really really valuable05:15
amurraythat spread test is failing on arch but I fixed it on core-16/18 (and was able to generate a local core-18 qemu image for testing so that was handy)05:16
zygaamurray btw, why did you use daemon: simple for the apps?05:17
amurrayso they would run automatically on install of the snaps05:17
zygait's usually easier to just have apps you run directly05:17
zygaand observe the failures05:17
amurrayoh ok that makes more sense.. I'll change it then05:17
zygain addition, daemon: simple apps are restarted by systemd on failure, no?05:18
zygaanyway, I'll leave a comment05:18
amurrayperhaps... another good point then for why I should change that :) thanks05:18
zygaamurray I know why the test failed on arch05:25
zygaI'll add a comment about that05:25
amurrayzyga: thanks - I couldn't see how to get extra details for it from the github actions output.. is it hidden somewhere?05:26
zygayou should normally see the full log05:27
zygabut again, may be limited to committers, I don't know05:27
zygaamurray please have a look at https://github.com/snapcore/snapd/pull/9509#pullrequestreview-51010272605:30
mupPR #9509: multipass/docker-support: Mount /etc/apparmor* from the base snap <Created by alexmurray> <https://github.com/snapcore/snapd/pull/9509>05:30
zygaI'll grab some food and return in 3005:30
zygaamurray arch didn't work because it doesn't have /snap05:31
amurraythanks zyga05:33
zygaeh, forgot to clean the kitchen last night05:53
zygaanyway, just coffee05:53
zygaamurray the test I meant in that comment about base refreshing is more tricky05:53
zygaamurray you must keep an app running (typically via systemd run for best containment) and then refresh the base (core)05:53
zygaso that core moves from x1 to x205:53
zygathen the mount profile should be different (and we should measure that)05:54
zygaand ideally, we'd measure that snap-update-ns did not fail, though that is harder05:54
zygarepackaging core to add a canary file would be easiest but more time consuming at runtime05:54
zygabut this kind of test will regularly happen for people running multipass, in practice, as bases refresh05:55
zygaso I'd like to see what we can do05:55
amurrayyep... ok I'll address the other comments and then will come back to that later - can you please put these hints etc in a comment on the PR so I don't forget :)05:55
zygaamurray sounds good05:55
zygaamurray yeah, I may try to push that test myself05:55
zygaso please don't worry about it05:55
zygaI'm just pointing out why I think it is relevant05:55
zygaif that part is broken it's like a delayed fuse, it will go off in production05:55
amurrayzyga: yeah that sounds bad - I am more than happy if you want to try and tackle that as a test ๐Ÿ˜€05:56
* zyga reboots macos for updates06:00
zyga-x240but I'm also here06:00
zyga-x240mvo: good morning06:04
zyga-x240how are you today?06:04
mvogood morning zyga-x24006:05
zyga-x240mvo: tests are not in a happy place, but not because of anything we did06:05
mvozyga-x240: good! less crazy, what can I do for you?06:05
zyga-x240apt cannot fetch the index06:05
mvozyga-x240: anything I can review?06:05
zyga-x240mvo: a 2nd review on https://github.com/snapcore/snapd/pull/9406 would be great06:05
mupPR #9406: many: allow ignoring running apps for specific request <Needs Samuele review> <Created by zyga> <https://github.com/snapcore/snapd/pull/9406>06:05
zyga-x240it's +1 by samuele and is fairly short06:06
zyga-x240I've restarted some tests, the Azure ubuntu mirror has some issues06:06
mvozyga-x240: cool, will do06:06
zyga-x240so the unit tests task fails early06:06
zyga-x240mvo: amurray has mostly fixed the apparmor 3 transition issue, we've just discussed some test improvements06:07
zyga-x240so not sure if there's a .1 coming but that would be my candidate06:08
mvozyga-x240: ok06:11
* amurray relocates.. back in 20mins06:12
zygathanks for the review!06:28
zyga-x240... value *errors.errorString = &errors.errorString{s:"cannot sign assertion: bad GPG produced signature: it does not verify: openpgp: invalid signature: RSA verification failure"} ("cannot sign assertion: bad GPG produced signature: it does not verify: openpgp: invalid signature: RSA verification failure")06:34
zyga-x240but at least the apt error is gone06:34
zygaI resolved conflicts on https://github.com/snapcore/snapd/pull/944606:42
mupPR #9446: overlord,usersession: initial notifications of pending refreshes <Created by zyga> <https://github.com/snapcore/snapd/pull/9446>06:42
zygaI think that could land soon but needs a 2nd review too :)06:43
zygaoh,sorry, 1st review actually06:43
zygamvo https://github.com/snapcore/snapd/pull/9507 can probably force-land06:57
mupPR #9507: overlord/snapshotstate/backend: specify tar format for snapshots <Simple ๐Ÿ˜ƒ> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/9507>06:57
zygamvo https://github.com/snapcore/snapd/pull/9502#pullrequestreview-510201180 is an easy win07:00
mupPR #9502: tests: more output for sbuild test <Created by sergiocazzolato> <https://github.com/snapcore/snapd/pull/9502>07:00
zygapstolowski good morning07:10
zygapstolowski I reviewed https://github.com/snapcore/snapd/pull/9501#pullrequestreview-51020166607:10
mupPR #9501: [RFC] wrappers: do not error out on read-only /etc/dbus-1/session.d filesystem on core18 <Bug> <Created by stolowski> <https://github.com/snapcore/snapd/pull/9501>07:10
zygaamurray https://github.com/snapcore/snapd/pull/9509#pullrequestreview-51020816907:13
mupPR #9509: multipass/docker-support: Mount /etc/apparmor* from the base snap <Created by alexmurray> <https://github.com/snapcore/snapd/pull/9509>07:13
* zyga is on a review spree today07:13
amurraythanks (I keep forgetting to run spread-shellcheck...)07:16
amurrayzyga: ok with any luck that should take care of your review comments - thanks again for taking the time to help me with this - let me know if you want me to squash these all down to just a few commits or not and if there is anything else need to do07:23
amurrayok I am going eod - thanks again zyga, I may be back later tonight but otherwise will catch up on scrollback next week07:27
zygaamurray not squashing is fine, I think that's okay07:54
zygaamurray perhaps mvo will decided to squash07:54
zygaamurray see you next week, thanks!07:54
zygaI'll push a test for the base refresh test into your branch later07:54
zygaamurray: have a lovely weekend!07:54
mvozyga what PR?07:54
zygamvo: https://github.com/snapcore/snapd/pull/950907:55
mupPR #9509: multipass/docker-support: Mount /etc/apparmor* from the base snap <Created by alexmurray> <https://github.com/snapcore/snapd/pull/9509>07:55
mupPR snapd#9511 opened: o/snapstate: re-order remove tasks for individual snap revisions to remove current last <Needs Samuele review> <Created by stolowski> <https://github.com/snapcore/snapd/pull/9511>08:01
zygapstolowski do you think we can re-order assertion vs snap download to fetch assertions before pulling, potentially huge snap?08:02
pstolowskizyga: theoretically yes, but i don't know the details of the code and don't know how easy or difficult that would be08:03
zygaI was hoping it's swapping two tasks around as those seem to have distinct descriptions in a snap change08:03
pstolowskiyeah, bth i don't remember if it's that simple or not08:04
pstolowskizyga: do you have a moment for HO? i'd like to talk about undo & interfaces08:18
mvozyga: looking08:21
zygasure, I need a moment08:28
zygapstolowski standup?08:31
mvozyga, amurray are you okay with me squash merging 9509 ?08:36
mvoso that I can cherry pick into 2.47?08:36
zygamvo but it's not fully ready yet, I will work on that later today08:36
mupPR snapd#9406 closed: many: allow ignoring running apps for specific request <Needs Samuele review> <Created by zyga> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9406>08:36
mvozyga do you know what the expectation is, should this go into groovy today so that there is a chance for it to be in the image?08:36
zygaI wanted to add a critical test08:37
zygaotherwise it's a risk of regression later08:37
zygaas in, it could be wrong now and we don't know08:37
mvozyga ok, sure, let chat in a bit (I think you are in a HO right now :) ?08:37
zyganot that it is okay now but breaks later08:37
zygawaiting for pawel08:37
zygaso hop in08:37
zygaand I'll explain08:37
mvozyga let's chat when you are done, I don't want to interrupt you guys :)08:38
zygahe's afk for a moment08:38
zygaso hop in :)08:38
zygapstolowski hop in :)08:42
amurraymvo: yep that's fine with me (squashing) - I just realised I used 'not' incorrectly in the task.yaml so I just pushed one more commit to fix that08:54
=== mup_ is now known as mup
mvoamurray: thanks for this! zyga works on one additoinal spread test and then this should be good to go (sorry for the delay, was in a meeting)09:15
amurraymvo zyga: I have one more commit in-the-works to try and fix the failures due to aa-status not existing on other distros... (I am changing it to use the one from the core snap)09:18
zygaas in /etc has no apparmor.d*09:18
zygathis may be a more complex problem09:18
zygaas you know, we share host /etc09:19
zygaso we may not be able to easily create that directory09:19
zygawe could do a writable mimic on /etc to do that09:19
zygaI think we should09:19
zyga(I mean automatically009:19
zygabut I'll wait for you to finish09:19
amurrayzyga: no I don't think that is it - it is just that in task.yaml I try and run aa-status to see if the profile got loaded but it doesn't exist - so instead I am going to run /snap/core/current/sbin/aa-status to look for it09:22
zygaamurray will that work if you don't have libapparmor?09:22
zygae.g. centos 709:22
zygaperhaps on systems without apparmor we should just test that we don't fall over entirely09:22
zygabut not measure apparmor (which is not present anyway)09:23
zygamvo https://github.com/snapcore/snapd/pull/9498#pullrequestreview-51020739609:24
mupPR #9498: client,daemon,snap: auto-import does not error on managed devices <Needs Samuele review> <Run nested> <Created by mvo5> <https://github.com/snapcore/snapd/pull/9498>09:24
amurrayhmm perhaps, I am not sure.. I've gotta run now so feel free to run with this however you feel is good - I'll try check back again later09:24
mvozyga \o/09:34
pedronismvo: zyga: https://github.com/snapcore/snapd/pull/9498#discussion_r50610993709:38
mupPR #9498: client,daemon,snap: auto-import does not error on managed devices <Needs Samuele review> <Run nested> <Created by mvo5> <https://github.com/snapcore/snapd/pull/9498>09:38
* zyga afk09:52
amurraymvo: sorry one more commit incoming on PR #9509 to try and make this use of aa-status in the regression test a bit more robust so hold off on squashing/merging for a bit - then with any luck the test should be more reliable on non-ubuntu platforms10:07
mupPR #9509: multipass/docker-support: Mount /etc/apparmor* from the base snap <Squash-merge> <โš  Critical> <Created by alexmurray> <https://github.com/snapcore/snapd/pull/9509>10:07
pedronispstolowski: I reviewed #9511,  also what is blocking #9391 ?10:08
mupPR #9511: o/snapstate: re-order remove tasks for individual snap revisions to remove current last <Needs Samuele review> <Created by stolowski> <https://github.com/snapcore/snapd/pull/9511>10:08
mupPR #9391: o/assertstate: introduce ValidationTrackingKey/ValidationSetTracking and basic methods <Needs Samuele review> <validation-sets :white_check_mark:> <Created by stolowski> <https://github.com/snapcore/snapd/pull/9391>10:08
pstolowskipedronis: thanks10:09
pstolowskipedronis: merged 9391, sorry i missed it10:10
mupPR snapd#9391 closed: o/assertstate: introduce ValidationTrackingKey/ValidationSetTracking and basic methods <Needs Samuele review> <validation-sets :white_check_mark:> <Created by stolowski> <Merged by stolowski> <https://github.com/snapcore/snapd/pull/9391>10:12
pedronispstolowski: thx10:15
pedronismvo: should we archive the 2.46 lane and create 2.48 ?10:15
pedronis(in the trello board)10:15
mvopedronis: yes10:17
pstolowskizyga: do you have a moment for https://github.com/snapcore/snapd/pull/9511 (simple, needs 2nd review)?10:17
mupPR #9511: o/snapstate: re-order remove tasks for individual snap revisions to remove current last <Needs Samuele review> <Created by stolowski> <https://github.com/snapcore/snapd/pull/9511>10:17
zygaI had it open but was digging into other tasks10:17
zygapstolowski LGTM10:20
zygaI was confused for a second because I though the intent was to reverse the sequence10:21
zygabut I think this is fine10:21
pedronisbecause of blocked revision, current is not always the top10:22
zygammm, I see10:23
pedronismvo: done10:25
pedronismvo: the bug in https://trello.com/c/OnPCcZSM/342-https-bugslaunchpadnet-ubuntu-cosmic-source-snapd-bug-1825298 seems wrong ?10:28
mvopedronis: looking10:29
pedronisor I'm confused10:29
pedronisI'm missing the relation with zfs there10:29
pedronisfor sure is not mentioned in the bug itself10:30
mvopedronis: the comment is wrong, the bug was real at some point, not sure if it still is10:30
mvoyeah, the comment is bogus10:30
pedronisI would archive that card, unless we plan to do something, the bug still exist as bug10:31
pedronisand create a different card about testing with zfs10:31
pedronismvo: ^10:31
zygamvo FYI: there's an optional include syntax in apparmor10:32
zygabut not sure if available all the way back10:32
mvopedronis: sorry for the delay - yes, card about zfs would be good10:41
zyga-x240https://github.com/snapcore/snapd/pull/9204 should pass on this iteration10:52
mupPR #9204: sandbox: track applications unconditionally <Created by zyga> <https://github.com/snapcore/snapd/pull/9204>10:52
zyga-x240I will need to prepare some more things there10:52
zyga-x240mvo: I will pick up Alex's PR now10:53
zyga-x240pedronis: I updated https://github.com/snapcore/snapd/pull/9422 - if you have some time to give it another look it would be great to get going on export manager10:53
mupPR #9422: overlord: add link participant for linkage transitions <Needs Samuele review> <Created by zyga> <https://github.com/snapcore/snapd/pull/9422>10:53
pedroniszyga-x240: yes, I will look at it after lunch, hopefully before the standup11:01
zyga-x240the remaining part of export manager is large but this is crucial to make progress11:02
rogpeppezyga: FYI that raspberry pi went down again the day before yesterday, and still required two power cycles to bring it up again11:02
zyga-x240rogpeppe: yes, that was our doing11:03
zyga-x240rogpeppe: thank you for rebooting it11:03
rogpeppezyga-x240: ah, np11:03
zyga-x240rogpeppe: we have more ideas but we are exploring them locally11:03
zyga-x240rogpeppe: we should have more information later on11:03
rogpeppezyga-x240: sorry, i had to reboot my computer and forgot to restart my IRC client if you were trying to get in touch11:03
zyga-x240rogpeppe: that's okay, you are really the most helpful bug reporter we ever had11:03
rogpeppezyga-x240: i've very glad to be of help.11:04
zyga-x240rogpeppe: some fixes for the bugs we discovered while analyzing your device are already fixed11:05
zyga-x240rogpeppe: some more are in progress11:05
zyga-x240rogpeppe: I sent you a message the other day11:05
rogpeppezyga-x240: in which messaging medium?11:05
zyga-x240rogpeppe: I'm not sure if you are willing to do this, but buying a low-cost USB-TTL adapter would allow you to collect boot logs that will show what happens when the device fails to reboot for update11:06
zyga-x240rogpeppe: here on IRC11:06
zyga-x240it was an amazon UK link11:06
zyga-x240I can find that again11:06
zyga-x240it's a USB-to 3.3V serial port adapter11:06
zyga-x240just a cable you plug into a laptop11:06
zyga-x240open screen/minicom/putty and collect the serial logs on "snap refresh core18"11:07
zyga-x240we'd know exactly what fails11:07
zyga-x240if you could do that that would be amazing11:07
zyga-x240but if not we are re-creating that environment here, I should have some answers today11:07
rogpeppezyga-x240: if you could provide detailed instructions for doing that on a windows machine, that would be very helpful - then i can walk my dad through doing it, maybe this weekend11:08
zyga-x240yes, that's totally fine (I mentioned putty for a reason)11:08
zyga-x240rogpeppe: it requires a hardware attachment, you'd have to purchase that and your dad would have to have it11:09
rogpeppezyga-x240: if you sent the link by PM on IRC, i don't think i got it. i can't find it in my logs anyway.11:11
zyga-x240let me find it again11:11
zyga-x240rogpeppe: this looks sensible: https://www.amazon.co.uk/DSD-TECH-adapter-FT232RL-Compatible/dp/B07BBPX8B8/ref=sr_1_11_sspa?dchild=1&keywords=USB+TTL&qid=1602846716&sr=8-11-spons&psc=1&spLa=ZW5jcnlwdGVkUXVhbGlmaWVyPUEyUU41UjZVS0hGUTQ4JmVuY3J5cHRlZElkPUEwMTQ5NjIzMk1BOUJPSVUyT05VQyZlbmNyeXB0ZWRBZElkPUEwMzQxMzE2MTBRNTlWTTlEN0FMNiZ3aWRnZXROYW1lPXNwX210ZiZhY3Rpb249Y2xpY2tSZWRpcmVjdCZkb05vdExvZ0NsaWNrPXRydWU=11:12
zyga-x240it should work out of the box on windows11:12
zyga-x240we'd only need that, a laptop and putty you can download easily11:13
rogpeppehow do you connect that to the pi?11:13
zyga-x240rogpeppe: there are wires you connect to the pins on the GPIO header11:13
zyga-x240with the ethernet port on the bottom left11:13
zyga-x240you take the row of pins on the right11:14
zyga-x240there are two columns of pins11:14
zyga-x240the right column is the one we need11:14
zyga-x240skip the first two pins from the top, they supply power11:14
zyga-x240the next pin is ground, connect it to the adapters GND pin with the wire11:15
zyga-x240(the adapter has a jumper that can select 5V or 3.3V mode, we need the 3.3V mode for pi)11:15
zyga-x240the next two pins are TX/RX11:15
zyga-x240just plug the wires to the corresponding pins on the USB adapter11:15
rogpeppezyga-x240: if you could find a diagram, that would be very helpful11:15
zyga-x240yeah, it's very well documented11:15
rogpeppezyga-x240: i think i should probably order it here and try it out myself before attempting to get someone to do it remotely with no visuals :)11:16
zyga-x240that's a good idea11:16
rogpeppeok, it should arrive here tomorrow.11:17
zyga-x240that's quick, we can sync tomorrow, just send me a message11:18
zyga-x240the good way to check if it works is to reboot a pi11:18
zyga-x240there's all kinds of log messages shown11:18
zyga-x240on windows you need putty 115200, 8N1, no hardware flow control11:18
zyga-x240on linux screen /dev/ttyUSB0 115200 is equivalent11:18
mupPR snapd#9036 closed: snapshots: import of a snapshot set <Needs Samuele review> <Skip spread> <Created by slimjim777> <Closed by mvo5> <https://github.com/snapcore/snapd/pull/9036>11:22
mupPR snapd#9506 closed: boot: skip some unit tests when running as root <Simple ๐Ÿ˜ƒ> <Skip spread> <Created by bboozzoo> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9506>11:22
zyga-x240amurray: can you rewrite your history to use canonical email address?11:31
zyga-x240mvo: ^ some annoyance to overcome11:31
mupPR snapd#9511 closed: o/snapstate: re-order remove tasks for individual snap revisions to remove current last <Needs Samuele review> <Created by stolowski> <Merged by stolowski> <https://github.com/snapcore/snapd/pull/9511>12:02
mvozyga-x240: we just squash merge12:19
mvozyga-x240: I mean, that should be fine, no?12:19
pedronismvo: zyga-x240: I fear #9422 doesn't do anything atm so should be fine, but it shows some problems that will need some snapstate fixes that might be a bit risky before 2.4812:36
mupPR #9422: overlord: add link participant for linkage transitions <Needs Samuele review> <Created by zyga> <https://github.com/snapcore/snapd/pull/9422>12:36
mupPR snapd#9502 closed: tests: more output for sbuild test <Created by sergiocazzolato> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9502>12:37
mupPR snapd#9469 closed: snapshots: import of a snapshot set <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9469>12:42
zyga-x240pedronis: ack12:45
zyga-x240pedronis: I'll read your review12:45
zyga-x240mvo: hmmm, I guess that solves the problem, yes12:46
marcustomlinsonhey, does Arch Linux keep snaps somewhere other than /snap? looks like /var/lib/snapd/snap/?12:51
zygamarcustomlinson correct12:53
marcustomlinsonI see thanks12:53
marcustomlinsonzyga: is the home directory structure the same though? i.e. ~/snap/blah12:55
zygamarcustomlinson yes, it is the same everywhere12:56
zygamy pleasure :-)12:56
mupPR snapd#9501 closed: wrappers: do not error out on read-only /etc/dbus-1/session.d filesystem on core18 <Bug> <Created by stolowski> <Merged by stolowski> <https://github.com/snapcore/snapd/pull/9501>13:32
ograzyga-x240, shoudnt that be "DISK 1/12" ?13:34
zyga-x240mvo: how much data did you have to process?13:35
mvozyga-x240: how much data?13:37
zygathat localization thing you mentioned13:41
mvozyga: the apt-ddtp stuff? it's not small, maybe a gig of distro release to churn through but a lot of redundant things of course13:43
mvozyga why do you ask :) ?13:43
zygato compute the number of floppy disks needed for ogra ;)13:44
zygamvo I won't make it with that PR today, it requires a few more hours of careful coding and testing13:47
zygamvo is it okay if we have it by Monday?13:47
mvozyga yes13:49
zygaI just don't want to rush it13:49
mvozyga I would prefer sooner but if it does not happen it does not happen13:49
mvozyga I will let sil2100 know13:49
zygait should be a connected plug for most sensibility but that also changes how the service starts13:49
zygabecause we IIRC first start then auto-connect (if no then  that's less of a problem)13:49
zyga-x240mvo: I've added my analysis to https://github.com/snapcore/snapd/pull/9509#pullrequestreview-51049594513:57
mupPR #9509: multipass/docker-support: Mount /etc/apparmor* from the base snap <Squash-merge> <โš  Critical> <Created by alexmurray> <https://github.com/snapcore/snapd/pull/9509>13:57
zyga-x240I'll fix it today, just not in the next 30 minutes13:57
pedroniszyga-x240: not sure it's useful but we run setup-profiles twice for core and snapd though13:59
pedronisalso it seems I will to review this13:59
zygathanks, I'll add a comment when I'm done13:59
zygaI'll make one last coffee of the day, vent the office and take a 30 minute break14:54
zygathen back to finish this14:54
zygaback to coding15:52
* cachio lunch16:10
mbeierlIs it possible to allow a snap to read a file that I have mounted on NFS in an arbitrary locations?  I have an NFS mount under ~/, but of course get permission denied when using a snap to read a file from there16:22
zygambeierl under /home?16:24
zygambeierl it should work, unless we mis-detected nfs16:24
zygambeierl if we detect nfs mounts we enable a special mode that compensates for that16:24
zygambeierl but regular restrictions apply, e.g. snaps need to connect the home interface to access files in home16:24
mbeierllet me give a quick test with jq (as it's simple) to make sure.  This is with my osmclient snap and I may have missed something there16:24
zygambeierl if you have a moment, could you tell me two things:16:25
zygambeierl: dmesg | grep DENIED16:25
zygambeierl and how you've set up NFS to be mounted (systemd unit, fstab, something else)16:25
mbeierlA few denied: https://pastebin.ubuntu.com/p/pDq9djmsdy/16:27
zygaah, hgfs (vmware)16:28
zygabut is that the thing being denied? or is there an actual NFS there as well?16:28
mbeierlAnd I was wrong.  This is not NFS, this is VMware hgfs (host/guest virtual filesystem)16:28
zygahgfs is mounted in /mnt/hgfs16:29
zygawe don't have any official support for that16:29
zygabut we can think of adding some16:29
mbeierlodd.  jq works, but my snap does not, so I need to double check on that16:29
zygacould you please report a quick bug for that on launchpad.net/snapd/16:29
zygait won't be tomorrow but we can sort this out, I suspect16:29
zygaas a variant of home interface mode16:29
mbeierlbut jq works, so I think it is the way I set up and published the osmclient snap, so I will do a comparison of the two first16:30
zygaespecially that it is quite common16:30
mbeierlI just jumped the gun, as I had a problem with jq like 6 months back and it did not work back then on NFS.  Things have changed, but I did not keep up, so let me make sure I really have a bug here and not pebkac16:31
zygaI fixed a few more variants of nfs support16:31
zygaI'd love if you could confirm there is nothing more16:31
zygaor tell me what I've missed16:31
mbeierlzyga, ok I am stumped.  jq has no connections?  How is this magic done then?16:33
mupPR snapd#9512 opened: o/snapstate: move setting updated SnapState after error paths <Created by pedronis> <https://github.com/snapcore/snapd/pull/9512>16:38
mbeierlI can confirm: my osmclient snap works with NFS, just not HGFS.  The jq snap works with reading files from both16:39
mupPR snapd#9504 closed: many: verify that unit tests work with nosecboot tag and without secboot package <Run nested> <Test Robustness> <Created by bboozzoo> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9504>16:48
mupPR snapd#9449 closed: interfaces: PTP hardware clock interface <Needs security review> <Created by slimjim777> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9449>16:53
mupPR snapd#9498 closed: client,daemon,snap: auto-import does not error on managed devices <Needs Samuele review> <Run nested> <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9498>16:53
mupPR snapd#9503 closed: tests: use tests.backup tool <Run nested> <Created by sergiocazzolato> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9503>16:53
mupPR snapd#9420 closed: tests/nested/manual/minimal-smoke: use 384MB of RAM for nested UC20 <Run nested> <UC20> <Created by bboozzoo> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9420>17:19
zygambeierl re17:38
zygasorry, I had to go away17:38
zygambeierl what is the jq command you invoked?17:39
zygambeierl please report the bug about hgfs, I'll try to do something about that, I'm also using vmware often for things like that17:39
* cachio afk17:53
mupPR snapd#9507 closed: overlord/snapshotstate/backend: specify tar format for snapshots <Simple ๐Ÿ˜ƒ> <Created by bboozzoo> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/9507>18:19
mupPR snapd#9513 opened: snapshotstate: detect duplicated snapshot imports <Needs Samuele review> <Created by mvo5> <https://github.com/snapcore/snapd/pull/9513>18:24
mbeierlzyga, simply jq /mnt/hgfs/home/test.json and it is able to read the file.  snap connections jq shows no connections19:02
zygambeierl can you share your "snap version" please?19:10
mbeierlzyga, https://pastebin.ubuntu.com/p/3Mk9zyHzVF/19:10
zygathat is really interesting!19:11
mbeierljq                   1.5+dfsg-1              6      latest/stable    canonicalโœ“      -19:11
mbeierlhow so?19:11
zygacan you pastebin /var/lib/snapd/apparmor/profiles/snap.jq.jq19:11
zygabecause without connections, what should be granting access to hgfs?19:11
zygathis looks correct19:12
zygawas it really just19:12
zygajq /mnt/hgfs/home/test.json19:12
zygajq < /mnt/hgfs/home/test.json19:12
zygaand perhaps most importantly19:12
zygacommand -v jq19:12
zygais it /snap/bin/jq?19:13
zygaor something else?19:13
mbeierlcommand -v jq => /snap/bin/jq19:14
mbeierlI know... it's got me confused19:15
zygaand jq < test.json?19:15
mbeierlno message, no output19:15
zygadmesg | grep DENIED19:15
zygaperhaps it was rejected?19:16
mbeierlwait - I added content to test.json so jq would have something to do, and it worked as expected, no denied19:16
zygawell, technically without an interface that bridges /mnt, it just doesn't exist from that snap point of view19:16
mbeierlzyga, I am silly.  jq syntax error.  I forgot the ".", so: jq . test.json gives me permission denied finally19:20
zygaokay, no alarm then19:20
zygambeierl does jq have any interfaces?19:21
* zyga checks himself19:21
zygasomeone would have to add an interface to it19:22
mbeierlside note: jq works on the NFS share, which again is odd with no connections19:22
zygato read stuff from removable media and other places19:22
zygambeierl where is that share?19:22
zygabut yes, it is odd19:22
mbeierlI have nfs mount /home/mark/git.  So it's under home19:22

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!