/srv/irclogs.ubuntu.com/2022/10/08/#snappy.txt

mupPR snapcraft#3945 closed: requirements: craft-store 2.3.0 for new credentials <Created by sergiusens> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/3945>01:47
dob1do the packages on snap get added only on release time or can be added before a new release is done?09:05
mupPR snapd#12246 opened: tests: add encrypted fakeinstaller spread test <Run nested> <Created by mvo5> <https://github.com/snapcore/snapd/pull/12246>10:32
mupBug #1992240 opened: Can't initialize on Raspberry Pi Compute Module 4 <Snappy:New> <https://launchpad.net/bugs/1992240>12:27
mupBug #1992240 changed: Can't initialize on Raspberry Pi Compute Module 4 <Snappy:New> <https://launchpad.net/bugs/1992240>12:51
mupBug #1992240 opened: Can't initialize on Raspberry Pi Compute Module 4 <Snappy:New> <https://launchpad.net/bugs/1992240>12:54
vanczAre seeds documented anywhere?12:56
vanczI'm trying to figure out how to safely remove things from /var/lib/snapd/seed12:56
ogravancz, probably in the code in livecd-rootfs 13:07
vanczI don't know my way around this infrastructure at all, what am I looking for?13:09
vanczI thought seeds are a snap thing? or are they only important when an image is first created?13:09
vanczdo they just assume youll leave a gig of all your old seeds around forever?13:09
ogra??13:13
ograon a desktop install snap keeps exactly 2 revisions of each snap 13:13
ograthough the seed directory is for initial seeding i think ... and gets copied 1:1 from the install media 13:15
vanczok thats roughly what I guessed so far13:16
vanczI did see mentions of deleting old revisions13:16
vanczdoes that also apply to the seed directory?13:16
ogralets take a step back ... what exactly do you want to achieve and why ? whats your goal13:17
vanczmeanwhile, im poking around the source13:17
vanczmy goal is to remove gnome and firefox from the seed directory to decrease space usage so I can fit the system in RAM13:17
ograwell, then you'd need to build an image from scratch and modify livecd-rootfs first i think 13:18
vanczwithout reinstalling13:18
ogra??13:18
vanczthis is an installed system13:18
ograand you want to fit that into ram ? how 13:18
vanczits not _that_ big (yet)13:19
ogranote that not even the live images "fit the raw system in ram" 13:19
vanczthe usage is around 12 gigs right now, I can definitely shave a few more off13:19
ograit ias all compressed squashfses ...13:19
vanczthis should really have been a server install but whoever installed the system *shrug*13:20
vanczok, and that implies...? <ogra> it ias all compressed squashfses ...13:21
ograthat the rootfs is massively smaller than what you currently try to fit into ram13:21
vanczI think I'm not understanding you13:22
ograa higly compressed squashfs can shrink the whole thing to 1/3 of its original size ... depending on the chosen copression13:22
ograyou'd simply never put a raw ext4 rootfs into ram 13:22
vanczok, so my plan is to do weird hacks and I'm not sure I can squeeze in a squashfs layer13:22
vanczthis isnt a long term thing but an intermediate step13:22
ograwhy exactl do you want to put it into rm in the first place ?13:23
ogra*ram13:23
* ogra needs a new laptop kbd 😞13:23
vanczgonna try to live migrate to zfs :D13:23
vancz(testing if its even feasible on a local vm first anywaY)13:24
ograwell, do that from an initrd shell 13:24
vanczI dont want to reboot13:24
ogralol13:24
ograyou will have to reboot at some point ... likely quite a few times13:24
vanczIf I end up having to reboot im not sure ill be able to get it to come up again without someone having to intervene. i plan to test the disk with the running system on a read only root and run a full boot in a vm and reboot once that succeeds13:25
ograi'd not waste that much time, a re-install of -server with zfs support would already be done since you asked first above 13:25
vanczsomething like taht13:25
vanczyeah this is a stupid weekend project, i dont have access to the adminstrative interface of the cm13:25
vancz* vm13:25
ograit is also likely to fail (r at least to result in an unexpectsd and unstable setup)13:26
vanczhence local testing13:26
vanczbut anyway13:26
ograanyway ... sudo apt purge snapd ... 13:26
vanczhaha13:27
ograand afterwards you can remove the seed dir 13:27
vanczwell Id love to but im running a service on a snap13:27
ogralater just re-install snapd and the firefox and lxd snaps 13:27
vanczgitea was easy enough to install13:27
vanczor can i export it or something and just add it back once ive done the reinstall of snapd?13:27
ograsnap help save 13:27
vanczcool thanks, ill take a look at that13:27
ograand snap help restore indeed 🙂13:28
ograsnaps have builtin backups 13:28
vanczyeah there is some nice stuff here and there13:29
ograyup 🙂13:29
vanczi wish i could have nixos with all the work in ubuntu and debian13:29
vanczis it save to try to find a way to remove the stuff in /usr/lib/x86_64-linux-gnu/dri ? if im not using graphics I expect I dont need that?13:29
vancz(not a snap issue ofc)13:30
vanczwait these are almost all the same size..wonder if its a bunch of hardlinks to something13:30
ograremoving files underneath the packagin system is always very risky in deb based systems ... bu you can use "apt-file search ..." to find the package names for files you delete temporary and use "apt install --reinstall ... " to reinstal the package later 13:31
ogra(or simply remove the packages instead of the files)13:31
vanczok thanks13:32
ograjust removing files obviously pulls out the carpet underneath the package db (which keeps track of every packaged file)13:32
vanczyeah i meant removing the packages I just dont know which they correspond to13:32
vanczbut the search will help with that13:32
ograyep13:32
ograapt-file is your friend13:32
vanczthe other half  of the question was "do i need these drivers" 13:33
ogranote though that some dri bits might be used for console output in the new kms based world13:33
vanczunless that breaks boot, i expect ssh should still work though13:33
vanczthey have different inode numbers but they are literally the same files, I wonder if I'm missing something13:35
vanczso this directory is reported as using a gig, but is it actually? 13:36
ogra$ du -hcs /usr/lib/x86_64-linux-gnu/dri13:37
ogra90M/usr/lib/x86_64-linux-gnu/dri13:37
ogranot here 13:37
vanczyoure right, i wasnt reading carefully, its actually 384mb here13:38
ograwow, still a lot more than it should be13:39
vanczhttps://bpa.st/7LCQ13:39
ograon my other (desktop) machine it is 61MB13:39
ograyou seem to have a lt more drivers there than i got on my default desktop installs13:41
vanczyeah I have no idea13:41
vanczim waiting for apt then ill search them all to make sure they all come from libgl1-mesa-dri13:41
vanczthat was the package for one of them13:41
vanczwhere do yours come from?13:42
ogradunno, havent checked 13:42
vanczok about five are from libgl1-amber-dri13:43
vanczmight be able to get rid of these https://bpa.st/EVPQ13:45
vanczalso i guess this isnt the kernel component either so it shouldnt result in graphics problems?13:45
ografor file in  /usr/lib/x86_64-linux-gnu/dri/*; do apt-file search $file; done | sed 's/: .*$//'|sort|uniq13:46
ograhttps://termbin.com/4cpp13:46
vanczyeah thats what I did13:46
ograthese are the ones installed here 13:46
lolekhi, I thought this is the best place to ask. Why snap is using keyword "refresh" instead of "update" to actually update it's packages? Everything else is using... update, upgrade or something similar. it's kind of counter intuitive. What was the reasoning behind this decision? 13:46
ogralolek, because canonicals CTO felt it was the better term 😛13:47
vanczogra: weird, looks like you even have a superser of what i do13:47
vancz*superset13:47
lolekso the logic was like "hey, let's make it the most confusing tool people has to use"?13:47
lolek*will need13:48
ogralolek, well, i think the logic was "lets make it distinct from the other tools to indicate it is different" 13:48
ograstart a thread on forum.snapcraft.io to discuss it though ... 13:48
ogralolek, also i think the idea was that "snap refresh" should hint towards "snap revert" since these two go hand in hand together 13:50
lolekwell for me snap itself and they way it works is totally frustrating... and using different word for update doesn't make things any better13:50
ograright, as i said, feel free to open a forum topic ... that way more people will actually see it ... 13:52
vanczogra: do you know what snaps like core20 and core18 are?14:05
vanczwell, I seem to have purged the dri directory14:05
vanczhm, guess I cant get rid of llvm, postgres seems to depend on it14:07
ogracore snaps are the bases the other snaps run on top 14:08
ograi.e. your execution environment to run snaps at all14:08
vanczok so its not ubuntu itself, good :pű14:11
vancz* :P14:11
vanczlets see if the whole thing doesnt break after this autoremove (lots of graphical stuff)14:11
vanczwill purging snapd not delete the snapshots dir?14:12
ograyou should indeed capture the arballs from it, yeah14:15
ogra*tarballs14:15
vanczok, doing that now14:21
vanczogra: ok, I've backed up and saved the backup of the one snap service im running, I just apt remove --purge snapd?14:24
vanczdisk usage down from 11g to 8 :p14:28
vanczogra: ok how do I deal with this?:14:32
vancz# snap install gitea --revision 609714:32
vanczerror: cannot install "gitea": Access by specifying a revision is not allowed for this Snap.14:32
vanczI want to install the revision the backup is for14:32
vanczogra: coincidentally if you give apt-file search a directory it will to it for all the files in it14:39
vanczogra: I have an old kernel version installed but I dont see its package, any idea why?: apt-file search /usr/lib/modules/5.15.0-47-generic/kernel/drivers/thunderbolt/thunderbolt.ko doesnt output anything15:07
mupPR snapd#12223 closed: many: add support for setting up encryption from installer <Run nested> <Created by alfonsosanchezbeato> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/12223>20:19

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