spotty | hi. i've just installed devuan with sysvinit and there's some packages that provide systemd units that i want to enable. afaik i should be able to do `service service_name start` and that should work but it doesn't. i know where the unit files are on my system but `service` doesn't appear to look in that directory (/usr/lib/systemd/system). it also can't seem to find them if i symlink them to other | 14:42 |
---|---|---|
spotty | dirs specified in the man page (/{lib,etc}/systemd/system). could someone please point me in the right direction here? how can i get sysvinit to use these systemd units? | 14:42 |
djph | spotty: that's a systemd thing. If you want systemd, install Debian... | 15:10 |
djph | ... I mean, you could probably try installing systemd into Devuan and all, but it seems like a lot of extra effort | 15:11 |
spotty | djph: the package i've installed contains systemd units, and it's from devuan's repos | 15:12 |
spotty | i thought they could be used in the context of a non-systemd system somehow | 15:12 |
djph | weird | 15:12 |
djph | might be that it has both the systemd units, and init scripts (i.e. /etc/init.d/service start) | 15:13 |
djph | ? | 15:13 |
spotty | no, it only contains systemd units | 15:13 |
spotty | could it be that it's a package taken from debian and it hasn't had the sysvinit scripts added yet? | 15:14 |
spotty | it's not a very popular package so i can't imagine there would be great demand for them | 15:15 |
spotty | i just thought it might be possible to use the systemd units even without systemd, but maybe i'm wrong? | 15:16 |
fsmithred | spotty, the service command works on scripts in /etc/init.d | 15:32 |
spotty | fsmithred: yes, but according to the man page it also works on systemd units in /{lib,run,etc}/systemd/system | 15:34 |
fsmithred | debian has pulled sysvinit scripts from some packages, and some have been put in orphan-sysvinit-scripts | 15:34 |
fsmithred | yeah, I'm sure it does, but I guess you would need systemd for that. | 15:35 |
fsmithred | oh! Most of the .service files in my beowulf are symlinks to /dev/null | 15:37 |
spotty | i guess i just got confused, i assumed as there were systemd units in a devuan package that they could be used for something | 15:43 |
spotty | i guess they can be used to translate into sysvinit scripts | 15:43 |
spotty | thanks for input guys | 15:43 |
djph | kind of hard to use systemd on anything when systemd isn't actually "in" the OS. | 15:43 |
djph | :) | 15:43 |
spotty | i'm no expert at either systemd units or sysvinit scripts but it looks like .service files contain commands you can put in shell scripts | 15:46 |
spotty | and sysvinit scripts are basically just shell scripts right? | 15:46 |
spotty | figured there might be some magic to allow sysvinit to start systemd units | 15:46 |
fsmithred | some packages contain both init scripts and service files, and a few also contain runit runscripts. | 15:47 |
fsmithred | there is a script to convert service files to init scripts. | 15:47 |
fsmithred | I might be able to find it. | 15:47 |
spotty | fsmithred: i'd appreciate it if you could find it | 15:47 |
fsmithred | looking... | 15:48 |
fsmithred | sysvinit-utils: /lib/init/init-d-script | 15:49 |
fsmithred | I'm guessing you will have to do some tweaking on whatever it produces. | 15:49 |
fsmithred | another approach is to rescue the init script from an older version of the package. | 15:49 |
spotty | fsmithred: thanks! i'll look into this | 15:54 |
lunario | my devuan 4.0 setup always gives me the message on bootup "ima: Error communicating to TPM chip". I am running an x61 thinkpad without tpm, so i guess that is why it pops up. but is there some way to avoid getting this message? | 17:37 |
gnarface | probably... maybe a kernel flag or something, just a guess | 17:40 |
lunario | well, a quick sysctl -a | grep ima resulted in nothing | 18:05 |
gnarface | nah i meant like a kernel command-line parameter | 18:08 |
gnarface | or maybe it might require a rebuild | 18:08 |
gnarface | i don't know | 18:08 |
Guest17 | helo. we have a problem with network on another devuan 4 upgraded from 3.0 (2). services networking, wicd must be started? on this devuan 4 it works only with network-manager without networking and wicd. thank you. | 18:36 |
Guest17 | on this fresh devuan 4 we typped. | 18:40 |
ErRandir | lunario: It depends on CONFIG_IMA. In my Devuan kernel that is built-in. | 20:26 |
eyalroz | So... Daedalus... how's it comng along? | 20:42 |
eyalroz | Is it reasonably safe to upgrade to it from chimaera? | 20:43 |
brocashelm | yes | 20:44 |
brocashelm | >t. ceres user | 20:44 |
brocashelm | but it's usually not recommended unless you know what you're doing | 20:44 |
fsmithred | eyalroz, since debian still hasn't frozen bookworm, there are lots of softare updates. But it does seem to be working well. I'm only using it for limited things. (building isos and packages) | 20:45 |
eyalroz | Oh, looks like it won't be frozen before next year... | 20:46 |
fsmithred | 44 updates pending, last 'apt upgrade' was March 28. | 20:47 |
brocashelm | daedalus and ceres are pretty close | 20:49 |
fsmithred | I think there were over 80 when i did it on the 28th, and that was what accumulated in 9 days. | 20:49 |
brocashelm | but ceres isn't as badly affected by the freeze as testing versions | 20:49 |
brocashelm | there will be updates, albeit slow | 20:49 |
brocashelm | at least that's how it went when chimaera was frozen | 20:50 |
fsmithred | yeah, it slows down | 20:50 |
brocashelm | dmo and graphics-drivers repos still got updates and i had no problems | 20:52 |
bb|hcb | eyalroz: as stated above, knowing what you are doing helps, but in case you hit some half-baked transition it may become hard to fix or may need long time to 'self-heal'. normally the release cycle is ~2y, i would stay on stable after a release for about 1y+ then switch to testing before the next release (but that's my taste, don't take it as a recipe)... | 21:00 |
eyalroz | bb|hcb: I lived on debian testing for a few years, but not with my desktop | 21:02 |
brocashelm | IMO, either stay with stable and upgrade when the time comes or switch to unstable and just deal with some package bugs (usually minor IME) | 21:06 |
bb|hcb | isn't the desktop the place where you happen to need some more recent stuff more often than with servers? | 21:07 |
fsmithred | testing with unstable pinned low, but available for when you need the new version of that broken package right now | 21:07 |
bb|hcb | and for development there is ceres on a vm :) | 21:07 |
brocashelm | ceres runs fine on my desktops. it's not bleeding edge like arch, but is still very stable and it's on kernel 5.16.0-6 right now | 21:08 |
fsmithred | before devuan existed, I used to run debian testing when it went into freeze. | 21:08 |
brocashelm | yes, i recall that refracta switched to devuan from debian | 21:09 |
bb|hcb | fsmithred: in my experience that didn't work well, especially when the new thing from unstable pulled a ton of deps, which in turn broke other things :( | 21:09 |
fsmithred | you should have been running refracta :) | 21:09 |
brocashelm | rather than refracta being a "fork" of devuan | 21:09 |
brocashelm | ^^^ | 21:09 |
brocashelm | refracta avoids much of the metapackage breaks and has some other better defaults than mainline devuan IMO | 21:10 |
brocashelm | no avahi or pulseaudio installed on refracta isos | 21:11 |
brocashelm | it also uses abiword instead of libreoffice | 21:11 |
eyalroz | bb|hcb: Well, yes, but I also wanted some newer versions of some libraries etc. | 21:11 |
fsmithred | put /home on a separate partition and leave a partition for a daedalus or ceres root and a partition for a chimaera (stable) root, so that if your shiny new breaks, you can reboot and still get stuff done. | 21:12 |
bb|hcb | fwiw, there was no serious breakage on my headless ceres vms last 2y | 21:13 |
brocashelm | i've been using ceres for that long and can attest to that | 21:13 |
bb|hcb | it may be more, but my memory is short ;) | 21:14 |
brocashelm | ;) | 21:14 |
brocashelm | the whole "sid is evil" thing is a thing of the past. ceres is on all of my machines and they remain stable after tinkering with them | 21:15 |
brocashelm | so stable + backports or unstable and dealing with the consequences (if any) are the only two realistic paths for me | 21:15 |
brocashelm | security updates get to unstable sooner than testing | 21:16 |
brocashelm | iirc | 21:16 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!