| ServiceRobot | greetings, earthers! | 00:33 |
|---|---|---|
| ServiceRobot | does anyone here know if it's possible to use openrc-init in beowulf? I've looked on the forum, but the suggestions for fixes don't seem to work | 00:35 |
| golinux | ServiceRobot: Parazyd left this link for you after you'd left: https://wiki.gentoo.org/wiki/OpenRC#openrc-init | 00:41 |
| ServiceRobot | that isn't related to debian's openrc implementation. yes, it skips over inittab, thus the gettys don't start. but I added the service manually and it still gets stuck at the sysinit runlevel | 01:44 |
| ServiceRobot | it appears openrc on debian uses patches to work with old sysvinit scripts. I was hoping to remove clutter like sysvinit-core and just use openrc-init, but it's proving difficult | 01:45 |
| ServiceRobot | I wish there was actual support for it | 01:46 |
| drawkula | start a wiki page about all these openrc stuff | 01:50 |
| gnarface | ServiceRobot: my advice is don't | 02:05 |
| gnarface | just don't try to cripple it before you actually get it working | 02:05 |
| gnarface | it depends on sysvinit stuff for compatibility purposes | 02:06 |
| ServiceRobot | it doesn't depend directly on it though. it is possible to use openrc-init. otherwise it might as well not exist | 02:10 |
| gnarface | agris: can you help this guy out by any chance? ^ | 02:14 |
| agris | ServiceRobot, what does /etc/init.d/rcS contain? | 02:16 |
| golinux | ServiceRobot: I've heard that in Devuan you still need the the sysvinit scripts. | 02:19 |
| ServiceRobot | yes, you still need the initscripts package, but you don't need sysvinit-core I believe | 02:20 |
| ServiceRobot | openrc-init should be able to handle it, as long as you manually install an agetty service | 02:20 |
| golinux | I can't anser to that | 02:20 |
| ServiceRobot | well I'd like to know if it actually can work, otherwise I'll have to rely on sysvinit-core with openrc on top | 02:21 |
| golinux | Too bad that KatolaZ isn't around these days | 02:22 |
| * golinux sheds a tear . . . | 02:22 | |
| ServiceRobot | who? | 02:39 |
| golinux | You don't have a search engine? | 02:42 |
| ServiceRobot | I mean, there's more than one KatolaZ on google... | 02:43 |
| agris | ServiceRobot, installing OpenRC in Devuan currently does not provide openrc-supervisory | 02:46 |
| ServiceRobot | because all the scripts are for sysvinit? I guess using openrc at all is a waste of time then? | 02:46 |
| agris | as you'll see when a script isn't using #!/sbin/openrc-run or is overriding the start() function it's basically in sysV compatibility mode with start-stop-daemon supervisory | 02:47 |
| agris | but start-stop-daemon supervisory without integration into OpenRC. OpenRC does in fact have start-stop-daemon as one of the supervisory options | 02:47 |
| ServiceRobot | yeah, and from the wiki I'm on my own according to openrc. but it's still weird that it gets stuck | 02:47 |
| agris | unless you want to maintain replacement init scripts yourself as far as i'm aware it's mostly useful for your own daemons, and having a consistant rc-status | 02:48 |
| agris | ServiceRobot, yeah, I asked you what your /etc/init.d/rcS contained | 02:48 |
| agris | you still have not answered | 02:48 |
| ServiceRobot | ah, sorry. let me look | 02:48 |
| ServiceRobot | I'm currently testing in a vm | 02:48 |
| gnarface | fucking triple deja-vu! | 02:52 |
| gnarface | crazy | 02:52 |
| gnarface | i need you all to please stop trying to force me to evolve the ability to see into the future | 02:52 |
| ServiceRobot | uh what? also here it is. it's a normal bash script with exec /sbin/openrc sysinit at the bottom | 02:53 |
| ServiceRobot | it says it's called from inittab, but openrc-init skips inittab | 02:54 |
| ServiceRobot | so.... | 02:54 |
| gnarface | paste.debian.net has no ads | 02:55 |
| drawkula | wiki! | 02:55 |
| gnarface | seconded on the wiki | 02:55 |
| ServiceRobot | you want me to paste the script there? | 02:57 |
| ServiceRobot | http://paste.debian.net/plain/1094297 | 02:59 |
| ServiceRobot | well here it is | 02:59 |
| gnarface | ServiceRobot: agris went through all of this last night. you should really listen to what he has to say | 03:01 |
| ServiceRobot | I did though. I just posted a link to the rc script I have | 03:02 |
| ServiceRobot | *rcS | 03:02 |
| agris | ServiceRobot, remove the last bit 'sysinit' | 03:02 |
| agris | then reboot | 03:02 |
| ServiceRobot | okay | 03:03 |
| agris | wait | 03:03 |
| agris | before you do that | 03:03 |
| agris | run rc-status | 03:03 |
| ServiceRobot | how can I if the system gets stuck and never reaches a tty? | 03:03 |
| agris | it says runlevel is sysinit yes, but does it say those services are started or stopped? | 03:03 |
| agris | anyways first try changing sysinit in /etc/init.d/rcS to nothing, and then if that doesn't work try changing sysinit to default | 03:04 |
| agris | I've had mixed 50/50 depending on the system what works | 03:04 |
| agris | I still don't know why yet | 03:04 |
| ServiceRobot | that's just odd. let me try it though | 03:05 |
| ServiceRobot | I can chroot in and adjust it | 03:05 |
| ServiceRobot | nah, it still gets stuck. darn | 03:06 |
| gnarface | wasn't there a second change you had to make too, agris ? | 03:09 |
| agris | at this point I'm not even sure anymore | 03:11 |
| agris | some of my containers with with sysinit | 03:11 |
| agris | someone of them only work with blank | 03:12 |
| agris | others only work with default | 03:12 |
| agris | I do have a custom inittab in my deploy script | 03:12 |
| ServiceRobot | that's so odd... if I'm not writing my own services, perhaps I shouldn't bother using openrc. | 03:12 |
| gnarface | agris: oh that's right, didn't you say something about agetty? | 03:13 |
| agris | http://dpaste.com/0SCPYAY | 03:13 |
| gnarface | agetty or getty? | 03:13 |
| agris | I'm not using gettys at all | 03:13 |
| agris | so perhaps this isn't really effecting me | 03:13 |
| gnarface | you just removed them ll? | 03:13 |
| gnarface | all? | 03:13 |
| gnarface | that might matter if you removed them | 03:14 |
| agris | I'm either logging in via ssh or spawning a shell inside the container | 03:14 |
| agris | but no, I didn't remove them | 03:14 |
| gnarface | hmm | 03:14 |
| agris | I can still get to a getty login prompt if I start my containers in foreground mode | 03:14 |
| gnarface | it could just be down to which startup scripts run, too | 03:14 |
| gnarface | couln't it? | 03:14 |
| gnarface | *couldn't | 03:14 |
| agris | It looks like Devuan might do sysinit runlevel in SysV, and then do the rest of the init proccess in openrc | 03:15 |
| agris | It's really weird | 03:15 |
| agris | I don't know anyone yet who can fully explain it | 03:15 |
| agris | i've hammered at it long enough to where it works again, but I still don't fully understand why | 03:15 |
| gnarface | i think just building the gentoo version and installing it is an option as well, albeit unsupported | 03:16 |
| ServiceRobot | if you have sysinit-core installed with openrc as a replacement for sysv-rc, sysinit runlevel is in sysv, and then it starts openrc from there | 03:16 |
| ServiceRobot | honestly, I may have less trouble if I just do that and say screw it to openrc-init | 03:16 |
| ServiceRobot | but then what are the advantages to using openrc+sysvinit? | 03:17 |
| gnarface | you get to use either type of script? | 03:17 |
| gnarface | best of both worlds? | 03:17 |
| gnarface | i am sure they had a good reason | 03:18 |
| gnarface | but i'm also now having a vague recollection of someone finding out you could just make a simple edit to bypass all the sysvinit stuff and just run openrc the way gentoo does it | 03:18 |
| gnarface | like without changing out the package | 03:19 |
| ServiceRobot | yeah, by not installing sysvinit-core, symlink openrc-init to /sbin/init, and create an agetty service. though that hasn't worked for me... hmmm | 03:19 |
| gnarface | does it fail a different way, ServiceRobot? | 03:19 |
| ServiceRobot | it gets stuck at "Cleaning up temporary files..." aka it gets to that part and never continues | 03:20 |
| gnarface | see what that script is trying to do | 03:21 |
| gnarface | it might have nothing to do with the first issue | 03:21 |
| ServiceRobot | I did. it doesn't seem like it would cause the system to freeze | 03:24 |
| ServiceRobot | and I looked at each line pretty carefully | 03:24 |
| ServiceRobot | but just to be clear, if I'm not creating my own openrc-run scripts, it's not worth having openrc installed at all? | 03:24 |
| gnarface | o dpm | 03:25 |
| gnarface | * i don't | 03:25 |
| gnarface | even know | 03:25 |
| gnarface | it might be faster | 03:25 |
| gnarface | it might have other benefits | 03:25 |
| gnarface | lower attack surface perhaps | 03:25 |
| gnarface | easier to understand, if you ever do want to make your own scripts | 03:26 |
| gnarface | some compatibility with gentoo documentation ... | 03:26 |
| gnarface | ServiceRobot: find the exact command it is hanging on, make sure it is something that works inside a container. it is not a guarantee that it would be. | 03:27 |
| gnarface | container/vm | 03:27 |
| gnarface | whatever you call it | 03:27 |
| gnarface | it could be an entirely different bug do you understand? | 03:28 |
| agris | ServiceRobot, If you want to go about replacing the sysinit runlevel, and most daemon scripts with OpenRC-run scripts, I would like to join you in that endeaver | 03:28 |
| ServiceRobot | yes, I'm aware. every setup is different | 03:29 |
| agris | I want an Alpine but based in Devuan | 03:29 |
| ServiceRobot | I tried alpine once. wasn't a big fan | 03:29 |
| agris | neither was I | 03:29 |
| ServiceRobot | oh I would love to help out with that | 03:29 |
| agris | that's why I run Devuan OpenRC container systems | 03:29 |
| ServiceRobot | but making it not conflict with what's already in /etc/init.d would be a pain | 03:30 |
| agris | You think there's a way to do that without stepping on sysv user's toes? | 03:30 |
| ServiceRobot | I wish devuan was more modular. it would set itself apart from debian more. as of right now, I don't know what the difference is besides saner defaults | 03:30 |
| agris | I would like to replace sysv start-stop-daemon scripts with OpenRC-Supervisory | 03:31 |
| agris | or s6 | 03:31 |
| ServiceRobot | supervise-daemon. I tried s6 once. couldn't wrap my head around it | 03:31 |
| ServiceRobot | I find runit better than s6, but runit has its own problems | 03:31 |
| agris | but currently openrc-supervisory works well enough | 03:31 |
| ServiceRobot | I'm constantly fighting with myself wether to use openrc or runit | 03:31 |
| ServiceRobot | I've used both, but they're both different | 03:31 |
| gnarface | what don't you like about sysvinit? | 03:32 |
| agris | OpenRC provides s6 as one of the optional daemon supervisors, along side it's internal one and start-stop-daemon | 03:32 |
| gnarface | oh, right, that was the other advantage about openrc | 03:33 |
| agris | gnarface, I like sysvinit, but OpenRC provides a lot of the functionality that systemd provides, without going outside the scope if init and without being a giant monolithic peice of software | 03:33 |
| gnarface | process monitoring | 03:33 |
| ServiceRobot | I'm worried about its maintainability. I'm not sure how long it will stay around. probably a while because compatibility, but openrc just feels cleaner | 03:33 |
| gnarface | ServiceRobot: it was process monitoring, that's what that guy kept telling me about openrc, or daemon management or something like that | 03:33 |
| ServiceRobot | I rather use openrc's supervision though | 03:33 |
| gnarface | ServiceRobot: yea, the primary value of openrc may be after it is booted | 03:34 |
| ServiceRobot | yeah, the rc-status command, rc-update, etc are very handy. easier than manually symlinking, which runit requires you to do | 03:34 |
| gnarface | so it is really a question of what you like | 03:34 |
| ServiceRobot | which is hard to support, because everyone's tastes are different | 03:34 |
| gnarface | but if you're having trouble with sysvinit, it might be easy for you to get help with that in here too | 03:34 |
| ServiceRobot | I'd be glad if both runit and openrc got the support they deserve, since they both are very usable | 03:35 |
| agris | openrc also provides easy ways to define cgroups per daemon, daemon uptime | 03:35 |
| agris | https://0x0.st/zOXa.png | 03:35 |
| ServiceRobot | yeah, uptime can be handy information | 03:35 |
| agris | also /etc/conf.d is very nice | 03:35 |
| agris | that pic I sent is an example of openrc's internal daemon supervisor being put to use for a custom daemon in Devuan ASCII | 03:36 |
| ServiceRobot | yeah, runit doesn't have a conf.d style directory | 03:36 |
| agris | you also can do periodic health checks | 03:36 |
| ServiceRobot | it's incredibly minimalistic, for better or worse | 03:36 |
| ServiceRobot | didn't know about health checks | 03:37 |
| agris | keeps track of which daemons are part of the runlevel configuration and which daemons are running manually, but not configured as part of a runlevel | 03:37 |
| agris | well, It comes from Gentoo | 03:37 |
| agris | so you expect a level of competency' | 03:38 |
| ServiceRobot | but so far, devuan has the same problems using a different init other than sysvinit as debian does. what sets apart devuan, really? just saner defaults? | 03:38 |
| agris | I'm typing from a Gentoo system right now | 03:38 |
| agris | ServiceRobot, systemd still runs in Devuan even if you install a different init | 03:38 |
| agris | you can see it there in the process table | 03:38 |
| ServiceRobot | I thought it was just libsystemd0 library? | 03:38 |
| agris | no | 03:39 |
| ServiceRobot | in that case, why use devuan over debian? | 03:39 |
| agris | because it force you to run systemd | 03:39 |
| agris | that attack surface isn't there | 03:39 |
| ServiceRobot | I managed to use debootstrap to not have systemd installed though? | 03:39 |
| ServiceRobot | just as a test in a vm | 03:39 |
| onefang | There's no systemd even installed on any of my Devuan boxen. | 03:39 |
| ServiceRobot | I really wish there was a comparison chart of what is actually different | 03:40 |
| ServiceRobot | that would help clear up some misconceptions | 03:40 |
| agris | you can look at the diffs | 03:40 |
| ServiceRobot | from where? | 03:40 |
| agris | also take for example all the stupid political bs debian is dealing with now | 03:41 |
| ServiceRobot | political bs? systemd related? | 03:41 |
| agris | you know Devuan is largly made of ex-debian maintainers | 03:41 |
| agris | well | 03:41 |
| agris | systemd somtimes related | 03:41 |
| gnarface | uh | 03:41 |
| gnarface | agris: false, there is no systemd running | 03:41 |
| agris | but take for example, removing a packages and creating a big PR nightmare that had the string 'boob' in the name | 03:42 |
| ServiceRobot | yeah, but I'm worried about development. buster was released not too long ago, and I can't find much information on when devuan will release beowulf | 03:42 |
| agris | or how they released buster with ecryptfs broken, because they had to work around systemd's shortcomings | 03:42 |
| ServiceRobot | I'm not saying this to diss on devuan. I'm trying to get a better idea on where both distros stand | 03:42 |
| agris | when it's ready | 03:43 |
| ServiceRobot | they removed a package over naming? | 03:43 |
| gnarface | ServiceRobot: note that if agris has systemd running in devuan, it's because he installed it from a debian repo accidentally | 03:43 |
| agris | there's a lot to un-fuck from debian | 03:43 |
| gnarface | ServiceRobot: there is no systemd in devuan | 03:43 |
| ServiceRobot | yeah, and I can't use another STABLE distro that doesn't have systemd because a lot of game server software I want to use wouldn't work | 03:44 |
| agris | gnarface, I don't have systemd running anywhere, I was talking about debian still have systemd helpers in the proc table even if you replace it with another init | 03:44 |
| ServiceRobot | yes, I know devuan doesn't come with systemd, but the argument is you can remove systemd from debian, or install it without systemd | 03:44 |
| gnarface | ServiceRobot: there was an old udev version with systemd in the name (obnoxiously) but that's not systemd | 03:44 |
| ServiceRobot | yeah, I notice that sysvinit uses systemd-udev | 03:44 |
| ServiceRobot | very odd | 03:44 |
| gnarface | agris: sorry i must have misread you | 03:45 |
| gnarface | or you typed Devuan accidentally | 03:45 |
| agris | ServiceRobot, did you actually try removing systemd from debian? | 03:45 |
| ServiceRobot | I installed it from scratch without systemd | 03:46 |
| ServiceRobot | with openrc-init, etc | 03:46 |
| agris | I had the same standpoint as you until I actually tried removing systemd from a debian install | 03:46 |
| agris | then I opened htop | 03:46 |
| agris | installed a RDBMS | 03:46 |
| agris | and boom, there systemd was, back again | 03:46 |
| agris | even though 'service' still worked, the attack surface was still there | 03:46 |
| ServiceRobot | couldn't you modify the /etc/apt files to ensure systemd won't install back? | 03:46 |
| agris | ServiceRobot, no, not if you like databases or desktop enviroments | 03:47 |
| ServiceRobot | I plan to use it for servers | 03:47 |
| ServiceRobot | nothing else | 03:47 |
| agris | One of the things Devuan does is rebuild packages that link in systemd libs, to no longer do that | 03:47 |
| ServiceRobot | but I thought devuan gave up on avoiding libsystemd0? | 03:48 |
| ServiceRobot | this argument is used against devuan all the time on other forums when I do research | 03:48 |
| gnarface | that's my understanding as well ServiceRobot | 03:49 |
| gnarface | but they do rebuild some packages that depend on systemd | 03:49 |
| ServiceRobot | people don't like the fragmentation devuan "caused". I understand their sentiments. I'm curious to hear from you guys | 03:49 |
| agris | ServiceRobot, Devuan ascii replaces libsystemd0 with a NOOP. Beowulf it will be completely removed | 03:49 |
| ServiceRobot | noop? | 03:49 |
| agris | no-operation | 03:50 |
| MinceR | it has given up on that, unfortunately. but in beowulf, libsystemd will be replaced with libelogind, afaik | 03:50 |
| MinceR | which is similar, but with a less hostile upstream | 03:50 |
| ServiceRobot | ah, goold ol' elogind | 03:50 |
| ServiceRobot | *good | 03:50 |
| gnarface | on the daemon, they do rebuild what packges they can that depend on systemd. they ban the rest: https://pkgmaster.devuan.org/bannedpackages.txt | 03:50 |
| ServiceRobot | elogind has worked great for me on Artix Linux. consolekit2 has been dead for 2 years | 03:50 |
| agris | A lot of the work that goes into Devuan comes from Gentoo | 03:50 |
| agris | eudev | 03:50 |
| agris | for example | 03:51 |
| ServiceRobot | but isn't banning packages a bit excessive? what if people need them? are their replacements? | 03:51 |
| ServiceRobot | oh i see. the list isn't that long | 03:51 |
| agris | if people need systemd-crond and can't use cronie, crond, or any other cron implementation they can go use debian | 03:52 |
| ServiceRobot | ah, I want to use beowulf right now. but debian buster also has elogind? | 03:52 |
| ServiceRobot | I currently am testing things in a vm | 03:52 |
| gnarface | ServiceRobot: it's mostly a representation of the limitations of the work power available, but stuff is generally prioritized less if systemd is very entrenched in its functionality | 03:52 |
| agris | like GNOME3 | 03:53 |
| ServiceRobot | I wouldn't get anywhere near GNOME | 03:53 |
| agris | I liked gnome2 | 03:53 |
| ServiceRobot | I only started using linux in 2016 so I wouldn't know what gnome2 was like | 03:53 |
| agris | but i'm glad the "Debian is not GNOME" movement is also Devuan | 03:54 |
| agris | It was like a fancier version of XFCE | 03:54 |
| agris | look up screenshots of Debian Sarge | 03:54 |
| ServiceRobot | but that's just it though. if you're installing from scratch a minimal system with no desktop environment, what does devuan provide for me that debian can't? | 03:54 |
| agris | http://pre09.deviantart.net/64b6/th/pre/i/2006/004/3/a/debian_3_1_sarge_gnome_desktop_by_accodeum.jpg | 03:55 |
| gnarface | ServiceRobot: it provides a system without systemd | 03:56 |
| gnarface | ServiceRobot: also some various arm device support :) | 03:56 |
| ServiceRobot | right, but I've gotten debian to not provide systemd either? if I use debootstrap that is | 03:56 |
| gnarface | it's very possible that you can debootstrap a system that is identical in debian and devuan. that is sort of the point actually. what's not to get? | 03:57 |
| ServiceRobot | if I can debootstrap debian to be like devuan, why use devuan? | 04:01 |
| gnarface | devuan won't sneak the systemd daemon in if you upgrade to a graphical environment for example | 04:01 |
| ServiceRobot | well I don't plan to use a graphical environment? | 04:01 |
| gnarface | there are scenarios where it literally won't matter, you're right. i'm saying you're right about that. | 04:01 |
| ServiceRobot | also won't it ask first? or does it really sneak it in? | 04:01 |
| gnarface | most the packages in devuan actually are the literal debian packages, unchanged | 04:01 |
| ServiceRobot | yeah, which calls into question once again. I guess it depends on use case | 04:02 |
| gnarface | i think you can just look on the devuan repo in a web browser or something to see which packages had to be rebuilt. | 04:02 |
| ServiceRobot | my use case is servers without systemd. no graphical environment. no xorg | 04:02 |
| gnarface | a basic repo doesn't even mirror all the debian packages, the unchanged ones get served by redirect | 04:03 |
| agris | One by one more packages pull in systemd | 04:03 |
| agris | when you do that, while that may work today your actively working against the goals of upstream development | 04:03 |
| agris | upstream will continue to get further away from your goals | 04:04 |
| ServiceRobot | even without a graphical environment? | 04:04 |
| gnarface | hmmm, dbus might do it to you i think | 04:04 |
| ServiceRobot | that just seems odd. they still provide sysvinit | 04:04 |
| agris | sure, you particular combination of packages might not pull in systemd today, but there's no guarantee an update might change that | 04:04 |
| ServiceRobot | even if I have sysvinit installed already? | 04:05 |
| agris | yeah, and openssh-server now pulls in systemd for some stupid reason | 04:05 |
| ServiceRobot | wait what? | 04:05 |
| ServiceRobot | I'm going to need openssh-server | 04:05 |
| agris | crap | 04:05 |
| ServiceRobot | it has a hard dependency on systemd? | 04:05 |
| gnarface | ServiceRobot: i think it will warn you which packages are going to get removed but you have to actually read it | 04:05 |
| agris | sorry, I meant to say openssh-server now pulls in dbus | 04:05 |
| ServiceRobot | ah | 04:05 |
| agris | my mistake | 04:05 |
| ServiceRobot | yeah you scared me | 04:05 |
| gnarface | it says it in all caps too | 04:06 |
| gnarface | but it is sandwiched between all the additions so , you know, pay attention | 04:06 |
| ServiceRobot | well of course. that goes without saying when configuring anything | 04:06 |
| gnarface | you could pin it out too | 04:07 |
| gnarface | in the configuration | 04:07 |
| ServiceRobot | I'm just worried about development rot if I use devuan. if devuan were to focus on better init support for openrc and runit, I'd have a reason to use it | 04:07 |
| gnarface | but then if you forget you did that, you'll be trying to figure out later why you can't upgrade or install something | 04:07 |
| ServiceRobot | does it not tell me why? | 04:08 |
| gnarface | i'm not sure | 04:08 |
| gnarface | aptitude might, apt-get might not | 04:08 |
| gnarface | i forget for sure, but i think it has been a problem for me in the past, yes | 04:08 |
| gnarface | or the error might have just been confusing or misleading | 04:08 |
| ServiceRobot | does devuan have any long term plans for diverging from debian, or is development too hard pressed? If debian causes more issues with systemd and doesn't listen to people who don't want it installed, devuan would benefit from actual choice | 04:11 |
| ServiceRobot | development time is an issue, but there's got to be a way to make it simpler | 04:11 |
| ServiceRobot | it shouldn't even be this hard, ffs | 04:11 |
| gnarface | the plan is NOT to fork it any further, actually, if at all possible | 04:12 |
| gnarface | but what actually ends up getting done does depend a lot on what Debian does first, and the work power available | 04:12 |
| gnarface | but the project is gaining volunteers and users all the time | 04:13 |
| ServiceRobot | has there been a large influx? | 04:13 |
| gnarface | i wouldn't say huge but steady | 04:13 |
| ServiceRobot | well that's a good sign at least | 04:14 |
| gnarface | it's not big news, but demand is certainly not decreasing | 04:14 |
| agris | ServiceRobot, Devuan wants to merge back with Debian if and when Debian stops making stupid decisions | 04:14 |
| ServiceRobot | demand would increase if openrc and runit weren't broken as hell on devuan. I have a lot of free time this summer. I'm literally willing to devote time organizing how sysvinit, openrc, and runit are packaged | 04:15 |
| gnarface | i think systemd itself will continue to slowly whittle away at its own userbase in this fashion for years, perhaps decades, while people slowly drift one by one over to devuan | 04:15 |
| ServiceRobot | I'm willing to change packaging if it means better support, backwards compatibility be damned | 04:15 |
| ServiceRobot | not like debian has backwards compatibility working anyway | 04:15 |
| gnarface | ServiceRobot: there seems to be some problems with using it in a container or vm that people keep tripping on | 04:16 |
| gnarface | ServiceRobot: you're like the 6th or so person to go through all the same phases of grief in the same order | 04:16 |
| agris | ServiceRobot, gnarface I'm already working on Devuan OpenRC containers | 04:16 |
| agris | We could work on it together | 04:16 |
| agris | I could show you what I've already got | 04:17 |
| ServiceRobot | sure | 04:17 |
| ServiceRobot | what have you done so far? | 04:17 |
| agris | OpenRC isn't going to get any better on Devuan until someone works on it | 04:18 |
| agris | ServiceRobot, I've gotten a full templat and config working to auto-create containers with apparmor | 04:18 |
| ServiceRobot | well then let's get started already. tired of waiting for devuan to provide a better experience | 04:18 |
| ServiceRobot | I don't mean that to be rude. lol | 04:18 |
| agris | I was going to wait ti'll I figured out why that weird thing was happening with /etc/init.d/rcS but I guess I could put up a git repo now | 04:19 |
| ServiceRobot | never worked with apparmor though | 04:19 |
| agris | apparmor support is imcomplete in ascii for containers | 04:19 |
| ServiceRobot | oh, so it's like lxc? | 04:20 |
| agris | I hear Beowulf is releasing with apparmor on by default though | 04:20 |
| agris | no, apparmor works in conjunction with lxc to further harden | 04:20 |
| agris | and isolate | 04:20 |
| ServiceRobot | ah | 04:20 |
| agris | apparmor provides mandatory access control | 04:20 |
| ServiceRobot | well I'll figure that out as I go. that's how I've always used linux so far | 04:21 |
| ServiceRobot | just a question. have devuan in the past tried at all to cooperate with debian, and if so, was it met with a lukewarm response? | 04:28 |
| ServiceRobot | the biggest argument against devuan from other users is that they could "just work together with debian" | 04:28 |
| gnarface | it seems like some of the debian maintainers started to warm up to devuan lately | 04:29 |
| gnarface | there is collaboration happening | 04:29 |
| gnarface | just not really equally on everything yet | 04:29 |
| ServiceRobot | that's good to know | 04:29 |
| ServiceRobot | so this means in the long term devuan won't be needed, or is it here as a contingency plan if things go south? | 04:30 |
| gnarface | i don't know that any of that has been finalized | 04:31 |
| gnarface | but i suspect there is miles of discussion on it in the mailing list archive | 04:31 |
| gnarface | i think devuan will remain as long as they can afford to keep the servers on | 04:31 |
| gnarface | whether it re-merges back with debian or not at some distant date in the future, is probably not known | 04:32 |
| ServiceRobot | the problem I see happening, and this is from someone who has briefly looked into this (I was following the devuan mailinglist for a while), is that debian in its goal to remain backwards compatible, won't be willing to make big changes for other init systems | 04:32 |
| ServiceRobot | I actually reported a bug with runit's stage files, and the fixes they made were less than optimal. but it's for backwards compatibility | 04:32 |
| ServiceRobot | as far as I know, if we truly want support for other inits, backwards compatibility will hinder that goal | 04:33 |
| gnarface | i'm gonna go afk for a while, but you guys have fun patching openrc | 04:34 |
| gnarface | something to make it lxc compatible without breaking anything else would have a strong chance of getting accepted upstream even | 04:35 |
| gnarface | there is precedent for that, for a patch getting submitted to devuan and devuan successfully getting it accepted in debian | 04:35 |
| ServiceRobot | one thing that can be worked on is init-system-helpers. helpers for openrc and runit for things like package upgrades, etc, would help | 04:38 |
| Arcans | Hi! | 16:56 |
| furrywolf | heyas! | 16:57 |
| Arcans | I have been using ASCII for a while now, but not having some software in their latest version is getting annoying, so I was wondering if switching to Ceres is realisitic without too much trouble. | 17:04 |
| fsmithred | upgrading to beowulf might be realistic, depending on what's installed. Upgrading a minnimal install works pretty well, upgrading a desktop can be tricky (a little or a lot) | 17:07 |
| furrywolf | unless you absolutely need the version ceres, I'd suggest beowulf instead. unstable tends to break randomly. | 17:07 |
| furrywolf | bbl | 17:07 |
| fsmithred | beowulf with judicious use of ceres packages can work well, because fixes go into ceres first. | 17:08 |
| fsmithred | but you need to be paying attention to what's happening in the repos | 17:08 |
| Arcans | Yeah, I have updated Debian stable to Sid in the past, I remember tricky it can become. | 17:10 |
| Arcans | Ok. | 17:11 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
