freenode/#devuan/ Friday, 2019-07-12

enyci SEEM to have a config that works-to-a-fashion, sort of!00:00
guru:(00:00
enyc[ 1270.464947] [drm:radeon_uvd_cs_parse [radeon]] *ERROR* Invalid UVD handle!00:00
enyc[ 1270.465011] [drm:radeon_cs_ioctl [radeon]] *ERROR* Invalid command stream !00:00
guruDo you use yours to game as well?00:00
enycno00:00
guruAh00:00
enychrrm, i add virtualbox to the mix ;p00:00
guruI have no idea what packages affected my FPS but I'm loving it00:05
enycguru: what sort of gpu do you use?  what games you play?00:42
enycfsmithred: i did a  bind  (not rbind)  mount of / into /bind  ... /bind/dev ... theres only liwmimted items there,  console core fd full kmem loop* mem null port ptmx pts ram* random shm std* tty tty0 urandom zero00:46
enycfsmithred: none of the sd??  etc that booting was complainnig about00:46
slvrshould a non-active /dev (such as with bind mounts) have anything in it at all?00:55
fsmithredthanks for checking, enyc. I don't think that's the problem.00:57
fsmithredslvr, you can go either way with that00:58
fsmithredenyc, I don't get the error if I mount the hosts /run/udev to the chroot00:59
fsmithredso I'm thinking in your case maybe /run/udev doesn't exist when it's needed01:00
fsmithredsearching the error message led me to this: https://github.com/systemd/systemd/issues/1131401:20
agris If I'm using LXC2 on Devuan ASCII, is root inside a container the same as root on the hypervisor?04:24
agrisevery time I try to start a container i get xc-start: cgroups/cgfsng.c: create_path_for_hierarchy: 1306 Path "/sys/fs/cgroup/cpuset//lxc/voip" already existed.04:59
agrisI need some help04:59
agrisno containers successfully start unless i add lxc.aa_allow_incomplete = 1 to their config05:11
gnarfaceagris: i think they need their own roots06:11
gnarfacei don't know about lxc2 specifically but in general, containers tend to have their own roots06:11
agrisgnarface, what do you mean?06:11
agrisevery container has it's own rootfs in /var/lib/lxc/nameofcontainer/rootfs/06:12
gnarfacethat makes sense06:12
agrisalso, Where is the manual page for start-stop-daemon?07:29
agrisDoes Devuan have problems with OpenRC?07:34
agrisI'm trying to write a script and it executes $command in forground07:34
agrisalso it doesn't seem to have the same start-stop-daemon as gentoo07:35
agrisIs this the wrong channel for support?07:35
FunkyBobno, it can jus be a bit slow at times07:36
rwpagris, On your system 'man start-stop-daemon' should work.  Doesn't it?  You can look online here: https://manpages.debian.org/buster/dpkg/start-stop-daemon.8.en.html07:44
agrisno it doesn't07:44
agrisI've also just upgraded from ascii to beowulf07:44
agrisnow it does07:44
agrisbut only in beowulf07:44
agrisalso OpenRC isn't working correctly07:45
rwpOkay. Glad to hear it is working for you.07:45
agrisIt's starting the command variable in foreground07:45
agrisIsn't that the whole point of init? to not start things in foreground?07:45
agrisalso, the net virtual doesn't work properly07:46
agrisany services that claim a dependency on network connectivity fail, when network connectivity is clearly working, otherwise I wouldn't be sshed in07:46
agrisalso07:46
agrisOpenRC is not respecting start_stop_daemon_args= at all07:47
agrisand in start-stop-daemon, it's missing parameters like --progress07:48
agrisAt this point so many things are broken or having errata I don't even know where to begin07:48
agrisI'm so confused07:48
rwpI am not running openrc so have little experience to help there. Sorry.07:49
agrisalso, daemons are crashing and rc-status is reporting them as online07:49
agriswhen they are clearly missing in top and their pid files don't resolve to any proccesses07:49
agris# rc-service fnet stop07:52
agris * Caching service dependencies ...                                                                                                                 [ ok ]07:52
agris * Stopping the fnet Daemon ...07:52
agrisstart-stop-daemon: warning: failed to kill 916: No such process07:52
agris1 pids were not killed07:52
agrisNo process in pidfile '/var/run/fnet.pid' found running; none killed.                                                                               [ !! ]07:52
agris * ERROR: fnet failed to stop07:52
agrisSo you think after openrc literally telling me No process in pidfile '/var/run/fnet.pid' found running; none killed. it would mark the service as 'stopped' right?07:53
agriswrong07:53
agrisfnet is still running according to rc-service and rc-status07:53
agriswtf?07:53
gnarface agris you might be running into beowulf-specific bugs, i don't know.  but i do know two things about openrc: 1) it works fine for a lot of people in devuan as-is 2) it is not the same setup as gentoo's, which (i think) has stripped out sysv style script compatibility08:02
gnarfaceso you may be seeing unfamiliar behavior that is expected behavior but just different from gentoo's... that shit about daemons running in the foreground though, that sounds like user error to me.08:03
agrisanother weird thing is the daemons act differently than when executed by hand08:04
gnarfacethat is expected behavior for several of them that i know of, nothing to do with openrc08:04
gnarfacealot of daemons are known to default to stderr logging if exected by hand, but not as a daemon08:04
gnarfacei can't imagine that'd be different in gentoo08:05
gnarfacehowever some of these daemons may be different versions than the ones in gentoo, so you may be seeing version related differences unrelated to the distro or the init system08:05
agrisI see08:06
gnarfacelemme rephrase that; most of these daemons are probably different versions than the ones in gentoo.  some differences in behavior should be expected08:06
agrisHow do i run a command as a specific uid?08:06
agrisIn Devuan08:06
gnarfacethere are several ways to do that08:06
agriswith OpenRC08:06
gnarfacenone of them are devuan specific either08:06
gnarfaceafaik start-stop-daemon should work08:06
gnarfacebut only if run as root08:06
gnarfaceyou can try out runas, sudo, or su too08:07
agrisOpenRC is supposed to use start-stop-daemon in the backend08:07
gnarfacehmm. runas is wrong.  i'm thinking of something with a similar name, but i don't think "runas" is the name08:07
agrisBut what's the Devian OpenRC way to do that? because command_user="_somedaemonaccountd" doesn't appear to work08:07
agrisI have command="", command_args="", and command_user=""08:08
gnarfacehmmm, sorry unfortunately i'm not familiar with openrc either, but i know that some people here are08:08
gnarfaceyou just gotta be patient and wait till one shows up08:08
gnarfaceor check out the forum and mailing list08:08
agrissure08:08
gnarfacedev1galaxy.org08:09
gnarfacedid you look here?08:09
gnarfacea lot of these repeating questions do get put on the forum, with answers08:09
gnarfaceand you've gotta be about the 4th or 5th person at least who has come in here complaining about openrc being "broken" compared to gentoo.  so far they all have gotten it figured out without my help.08:09
gnarfaceone of these days i might have to install it myself to figure out what's going on, but i suspect that you've run into something fairly simple08:11
agrisok search function on the forum isn't working right either08:22
gnarfacereally?  hmmm08:22
agrisI search for posts containing command_user and it brings up posts with absolutely zero occurrences of that string08:23
gnarfacedid you search for openrc?08:23
agrishold it, there's a cooldown period for searches08:23
gnarfacehttps://dev1galaxy.org/search.php?search_id=147076303608:24
gnarfacethe openrc installation in Devuan is probably more like the one in Debian than Gentoo, but Debian's documentation on it seems pretty sparse as well...08:25
agrishttps://dev1galaxy.org/search.php?search_id=80537578408:25
agrisNo, I got OpenRC installed. That's not the issue08:25
gnarfacedid you try my search link there?08:26
gnarfacei just searched for "openrc" and came back with a few pages of results08:26
gnarfacethe first thing you need to figure out is if your issues are self-inflicted or part of a ongoing problem with beowulf08:27
gnarfacei don't suppose you have an ascii install to test against, do you?08:27
agrisIs there search doing something special with underscores?08:27
agrisgnarface, Is there a way to downgrade back down to ascii?08:27
agrisIf not I'll just reinstall08:27
gnarfaceagris: i don't know about the underscores.  i do know that, like Debian, downgrading is not supported.08:27
gnarfaceyou could try ascii in a VM or something instead though08:28
gnarfaceit might save you some time08:31
gnarfacethat forum software is some well known open source one i believe, so it probably handles underscores in a documented fashion08:31
gnarfaceit might just be ignoring them or converting them to whitespace like google search does08:32
agrisgnarface, one second pleae08:38
agrisI'm trying something08:39
agrisok08:44
agrisI"m almost certain this is a bug in beowulf08:44
agriscommand_user="user:group" does not work08:44
agrisIf you set both a group and a user, starting the service will fail with start-stop-daemon: user 'user:group' not found08:45
agrisif if you set the init script command_user="user"08:45
agrisopenrc will start the init script as root even though you specified a user08:46
agriseffectively opening up the system to attack08:46
agrislet me try specifying a different process supervisor. maybe start-stop-daemon is the buggy part in beowulf08:47
agrisI will test this again in ascii later08:48
agrisok08:49
agrisadding supervisor=supervise-daemon to the script makes it work properly08:49
gnarfaceagris: glad you got it figured out.  just from your description of the behavior change i can't tell for sure if that's a bug or just a surprise change in behavior.  maybe they moved the group name to a separate variable or something?  i'd check to see if they have updated documentation somewhere for the version in beowulf08:56
agrisgnarface, I'm going by the official OpenRC documentation here https://github.com/OpenRC/openrc/blob/master/service-script-guide.md08:58
agristhe syntax specified is command_user="user:group"08:59
gnarfacenothing in there about command_group ?08:59
agrisno, that would be redundant08:59
gnarfacewell it would be redundant unless they decided to remove the ability to specify a group along with a user in command_user09:00
gnarfacei'm not saying that's what happened; this is pure speculation on my part09:00
gnarfaceyou're probably right that it's a bug, but often the documentation is the last thing to be updated, too...09:00
agriswell they didn't from OpenRC, as when I changed the proccess supervisor from the default of start-stop-daemon to supervise-daemon the priv-dropping worked correctly09:01
agrisso somewhere the translation of user and group parameters to start-stop-daemon got borked09:02
gnarfaceplausible09:03
gnarfacehmmm09:04
gnarfaceno bug reports on openrc in debian09:04
gnarfaceonly one in bugs.devuan.org09:04
gnarfacewell, only one unresolved09:04
gnarfacei don't think it's this though09:05
agrispity09:31
gnarfaceagris: i recommend filing the bug report about it09:34
agrisok09:34
agrisabout filing bugs for INIT systems other than systemd09:38
agrisIs this something I should file a bug report with Debian or Devuan?09:39
agrisBecause I feel like I'd get a hostile or apathetic response filing this with Debian09:39
gnarfaceagris: file it at bugs.devuan.org10:34
gnarfaceagris: try the email submission method if the form fails10:34
gnarfacei've heard it advised to use the email submission method10:35
gnarfacemake sure that one other open bug on openrc isn't the same thing first10:35
FatPhilagris: is it possible to set -x on the rc scripts? It might be verbose, but you would be able to see which of your parameters were being picked up and how they were being parsed?11:06
agrisFatPhil, that's not necessary. OpenRC scripts have their own interpreter like bash, perl, or python11:07
agrisopenrc-run11:07
FatPhilOK, back in the day it was standard sh, and -x would have come to your aid.11:13
agrisFatPhil, yeah OpenRC is sysv compatible, but it has extra capabilities11:25
FatPhilhave you looked at the source?11:26
agrisno11:26
FatPhildoes openrc use the default start-stop-daemon, or come with its own?11:29
FatPhilI'm not sure that its use of command_user makes sense11:29
FatPhilI see the following from doing $ git grep -C4 command_user11:30
FatPhilsh/start-stop-daemon.sh:                ${command_user+--user} $command_user \11:30
agrisby default is uses start-stop-daemon for supervisory11:31
FatPhiland man start-stop-daemon says:11:31
FatPhil       -u, --user username|uid11:31
FatPhil              Check for processes owned by the user specified by username  or  uid.11:31
agrisyou can also override the start() function and write a sysv-like system by hand if you need11:31
FatPhil       -c, --chuid username|uid[:group|gid]11:31
FatPhil              Change to this username/uid before starting the process. You can also11:31
gnarfacestart-stop-daemon is part of dpkg itself here on ceres.   try dpkg -S start-stop-daemon11:32
agrisstart-stop-daemon works fine by itself, albeit missing features11:33
agrisIt's the translation that's missing11:33
FatPhildo you agree that it would make sense for sh/start-stop-daemon.sh to use --chuid rather than --user, or am I misinterpreting the script. I have no experience either with openrc or even running start-stop-daemon directly, I'm just a pattern-matcher.11:37
g0zzyJust been phoned up about a box hanging on boot at a message about a dns resolver. Any ideas?12:44
gnarfacesomething is probably wrong with the DNS resolver12:45
gnarfacethat's more helpful than it sounds12:46
gnarface127.0.0.1localhost.localdomainlocalhost12:46
gnarfacefor good measure, make sure you have this line in /etc/hosts though too ^12:46
g0zzyWhy would that prevent booting though? Why not carry on booting with no network?12:47
gnarfacewell it's not actually hanging, it just has a really long timeout when it can't figure out it's own name12:47
gnarfaceprobably12:47
gnarfacesomething they're running12:48
gnarfaceoften the MTA12:48
gnarface(exim4 by default)12:48
g0zzyYes, that is in use12:48
gnarfaceit should continue booting after 60 seconds12:48
FatPhilare you using resolvconf?12:49
gnarfacethen you can either fix the DNS or fix the /etc/hosts file so you don't need DNS12:49
g0zzyProbably. Can't quite remember12:49
gnarfacebut if it's the ISP's DNS servers that are down, this change is external to the box and still needs to be corrected12:49
g0zzyNetwork ok for other devices apparently12:50
gnarfacetry doing something that requires DNS12:50
gnarfacelike surf to a webpage not in your cache12:50
gnarfaceIP traffic would still work even without DNS12:50
g0zzyYes, OK. That's probably a good idea12:50
FatPhilor just ping12:50
FatPhilifdown/ifup will at least let you see the DHCP negiotiation take place. however, that should be in your logs anyway12:51
FatPhilbut I do recommend resolvconf to make management of resolv.conf easier.12:52
gnarfaceoh, that's a good point.  a DNS problem could be a red herring... it might be DHCP failing to assign DNS properly12:52
g0zzyRight. SHall look into it12:52
gnarfacebut in most home networks these are both coming from the plastic router box they gave you anyway12:52
g0zzyYes, i'll call them and ask them to restart that piece of tat before rebooting12:53
FatPhilI have a shitty box whose ETH PHY took a minute to work out which way was on, and always timed out with udhcpc, so I had to change dhcp client.12:55
FatPhil(back to the default, so you should be fine)12:55
* FatPhil whistles innocently as he removes a horribly wrong line defining his own IP address from /etc/hosts...12:57
* g0zzy started using dnscrypt-proxy on this machine yesterday. Found it was broken in Stretch12:58
g0zzyThanks folks13:26
ejrhi. when i log into devuan ascii, audio output is muted by default and can only be enabled with the laptop's 'volume up' button (i.e. alsamixer is not sufficient). is there a way to still automatically unmute the system without needing to open my laptop lid and reach for the volume-up button?13:28
ejri am using a thinkpad on an ultrabase that is in a sort of drawer, which makes it tedious to unmute devuan as i need to move the entire docking station in order to open the lid13:30
Humpelst1lzchenejr: sounds like an older thinkpad?13:33
ejryes, an x6013:33
Humpelst1lzchenejr: because older had hardware mute, check if you have a bios setting for it13:34
Humpelst1lzchenejr: or search in /proc/acpi/ibm13:36
ejri am using libreboot... but i will check /proc, thanks!13:38
Humpelst1lzchennice13:38
ejroh well, there is the file "volume" which has two lines: level:   12 / mute:   on; but if I try to manually edit it (replacing "on" with "off" i cannot save it because it has already been overwritten again by the system13:40
ejrwhen dealing with files like that that had just one line i would just use "echo 'mute: off' > volume" but this does not work when there are two lines obviously13:40
Humpelst1lzchenejr: https://www.kernel.org/doc/Documentation/laptops/thinkpad-acpi.txt13:41
Humpelst1lzchencheck the Volume control section13:42
fsmithredecho -e "first-line\nsecond-line" > file13:42
fsmithredoh, you might need a second \n at the end13:43
Humpelst1lzchen"echo unmute > /proc/acpi/ibm/volume" in this case13:43
ejri tried both of your suggestions, fsmithred and Humpelst1lzchen, but always get "Operation not permitted" although I'm root13:45
gnarfaceno pulseaudio right?13:47
gnarfacei don't know for sure that it wouldn't do this13:49
ejrno pulse, right13:51
Humpelst1lzchenejr: did you read the kernel.org link?13:55
fsmithredis alsactl store still a thing13:59
fsmithred?13:59
fsmithredguess so, I got a man page for it14:00
fsmithreddon't know if it'll help14:01
ejrHumpelst1lzchen: yes, it says i need to enable volume_control=1 as a module parameter, but i don't know whether that means I need to create a file called volume_control with a 1 in it, or add that line to the volume file. in either case i cannot do so as i am not permitted to14:01
fsmithredwhat are the permissions on the file?14:02
ejr0 -rw-r--r-- 1 root root 014:03
fsmithreddid you have to make any changes to get the mute or volume buttons to work after you installed?14:05
ejrno14:06
Humpelst1lzchenejr: thats a module parameter14:09
Humpelst1lzchenejr: aka create a file /etc/modprobe.d/thinkpad.conf with content "options thinkpad_acpi volume_control=1"14:11
Humpelst1lzchenand reboot14:12
ejrwill do, brb14:15
Guest80950thanks, that fixed it. when i now do echo "unmute">volume, i get sound as desired14:19
Humpelst1lzchen:)14:31
FatPhilJust dist-upgrading one of my jessie boxes to ascii - is there any reason why my fairly minimal install (has only barebones X, no "desktop environment"), should bloat by half a gig in the process?15:11
gnarface --no-install-recommends15:12
gnarfaceadd this^15:12
gnarfaceFatPhil:15:12
FatPhilAPT::Install-Suggests "0";15:14
FatPhilAPT::Install-Recommends "0";15:14
gnarfaceyikes15:14
gnarfacei don't know then15:14
gnarfacethe estimates could be wrong15:15
FatPhilThe first lines in my new machine deployment script set that15:15
gnarfaceor it might not count what gets removed after later running autoremove to prune unused stuff...15:15
FatPhilI'll check the logs for packages which aren't uprades but are new packages, and see if I can spot a pattern.15:16
fsmithredalso, clean the cache15:16
fsmithreddu -sh /var/cache/apt/archives15:16
fsmithredto see how significant it is15:16
fsmithredapt-get autoclean (for obsolete) or apt-get clean (to empty it)15:17
FatPhiloops, I've run out of space! need to do some cleanup, crtain;ly15:17
fsmithredI hate when that happens15:17
fsmithredbbl15:18
FatPhiloooh, it's in a messed-up state now, I can't just type dist-upgrade again15:23
FatPhilI'll just pop into aptitude, update, and upgrade everything apart from what I want to hold (browsers that get worse with every new version)15:23
FatPhilWhy the fuck does /etc/X11/xdm/Xservers.dpkg-new contain 40 blank lines at the start of it? did some debain dev fall asleep at his keyboard?15:25
FatPhil    N or O  : keep your currently-installed version15:26
FatPhilvery easy decision to make15:26
drawkulachck the diff15:26
drawkulahttps://tinyurl.com/ommmmmmmmm15:27
* drawkula waits for a bercow-ommmm-meme...15:27
FatPhilthe diff had 40+ lines of nothing but '+', that's what made me look at the actual file, as I thought the pager showing the diffs had got its knickers in a twist.15:29
FatPhilI'm pleased to say that on another file, the 3-way merge worked perfectly.15:30
FatPhilOK, what's the difference between pkg-config and pkgconfig, just a cosmetic rename?15:32
g0zzy>>well it's not actually hanging, it just has a really long timeout when it can't figure out it's own name15:33
g0zzyActually it turns out that it's been like that for hours. So it's definitely a hang. It won't boot15:33
FatPhilcan you turn on early logins, so you can get in as root and sniff around before everything's started?15:34
g0zzyFatPhil: Is that to me?15:36
g0zzygnarface: Any ideas short of installing a different OS? There's nothing i can think of doing if it's not even booting15:40
FatPhilg0zzy: yup, alternatively you could go in in single user mode and then bring it up by hand15:40
g0zzyI did go in there rw init=/bin/bash15:40
g0zzyBut i just checked resolv.conf and it looked normal15:41
FatPhildoes DELAYLOGIN=yes in /etc/default/rcS let you go in and examing what task is hanging?15:42
FatPhil=no, sorry, obviously15:42
g0zzyI can try that i suppose15:43
gnarfaceg0zzy: it doesn't boot after 60 seconds?15:44
g0zzyAbso not. Stuck like that forever15:44
gnarfaceeven 120s?15:46
gnarfacedid you add localhost to the hosts file?15:47
gnarfacetry adding your own ip and hostname too15:47
g0zzyThe machine had been on for at least two hours when i got to it just now with that on the screen15:47
gnarfaceis it network boot or something?15:47
g0zzyStandard boot15:48
g0zzy"Couldn't get address for 'resolver1.opendns.com': failure"15:48
gnarfaceis that even something it should be looking for?15:50
drawkulaip6 problem?15:50
g0zzyA wicd hook?15:50
g0zzy(problem)15:50
g0zzyOne of the scripts that i use (wip == what is my ip) contains this: dig +short myip.opendns.com @resolver1.opendns.com15:55
g0zzyBut i don't see why that would be getting called so early. And it's been fine up until now15:55
gnarfacelook at the LSB headers in the script g0zzy15:57
gnarfacesee if it depends on networking?15:57
g0zzyI can't remember how that would be getting called now. The idea is for that box to mail me its IP address15:58
gnarfacegrep -ni 'opendns' /etc/init.d/15:59
* g0zzy thinks of the kludge dig +short myip.opendns.com @208.67.222.222 ;)16:00
gnarfacegrep -ni 'opendns' /etc/16:00
gnarface?16:00
gnarfacegrep -rni 'opendns' /etc/16:01
g0zzyI'd probably need to grep for 'wip' actually16:01
gnarfaceoh16:02
gnarfaceor whatever16:02
gnarfaceanyway it should be easy to find16:02
g0zzyI think we might be closing in on it though16:02
g0zzySo i'd have put a hook in networking to send that mail. It's ascii standard desktop. Where do you think i put that hook? I can't remember ;)16:06
gnarfacewell the startup scripts for the system should all be in /etc/init.d but you could have put it in your home directory or probably a number of other places16:08
gnarfaceif you had put it in /etc/rc.local it probably wouldn't be running before networking though...16:09
FatPhilg0zzy: does your IP address change often? I have a machine which does, so I have an (ana)cron job that just lets me know its address (by hitting my webserver, but mail would work too).16:10
jonadab^ I do that.16:11
jonadabActually, in my case, it's infrequent, so I have the cron job just hit the web server once an hour.16:11
jonadabBut if you're fetching a very small file, or requesting one that doesn't exist, you could do it much more often without having much impact.16:11
FatPhilI double-duty the script - my webserver caches some scraped information, so I use my pinging script to just keep those caches fresh so I don't have to wait 5 seconds for it to perform a new scrape!16:20
FatPhilg0zzy: udhcpc theoretically llets you run a script after various DHCPevents, perhap y cout16:24
FatPhilfkn kbd btry16:24
FatPhilerhaps you could hook your script into that?16:25
FatPhillet's see if the new battery is better. yup, seems so.16:27
jonadabRight, in my case I just grep the error log for "doesnotexist".16:50
jonadabSince it happens once or twice a year, this is adequate for my needs.16:50
teamjal_k-dawgHey, guys, just stopping by in here to say thanks for the great OS. I've been on ASCII for maybe about 9 months, and it's just been great. I really appreciate that you guys were able to branch off and get rid of systemd. I'll continue to buy Devuan merch in the future to support the development. Thanks everyone!16:52
FatPhilseconded17:01
ryuwhola17:04
ryuwdiavoletti17:04
jonadabDevuan isn't just great:  it's _necessary_.  Otherwise we'd have all ended up doing some hybrid between staying on Wheezy forever, and manually compiling by hand everything that really needed to be upgraded, which eventually lands you in LFS territory.17:06
Tom-_well, there's the debian-init-diversity effort.  it's possible to install buster using packages from experimental and unstable.  i've written up instructions but i haven't posted them on the www yet.  sorry if this is offtopic17:11
g0zzyGot it folks. There was a symlink to my mailing script in if-up.d. DNS was failing and making it hang. I'm not needing DNS now as using exim4 to send the admin mail instead of sendEmail17:11
Tom-_(the debian-init-diversity effort seems to be helped in equal parts by Devuan and other distributions)17:12
jonadabThe way I see it, in the long term, things can only really go three ways.  1. Ideally, the majority of distros eventually realize they should not be using systemd, and Devuan or something likes it effectively displaces Debian as the distro lots of other distros are based on (or Debian itself reverts to non-systemd; but that becomes more unlikely with every passing version).  2. Users who actually care17:24
jonadababout configurability gradually migrate away from Linux distros altogether, landing on another OS (possibly BSD).  Or 3. It all becomes moot because of some catastrophy that knocks out civilization as we know it, probably by destroying all the power grids worldwide at once.17:24
xinomilo_do you think missing init.d scripts in packages/daemons should be reported to bugs.debian as well ?17:24
jonadabWill it do any good?17:25
jonadabAt this point, I just sort of figure Debian is poetteringed (which is a synonym for broken and useless).17:25
xinomilo_i mean with the current init-diversity effort.. does it make sense or?17:26
jonadabDunno.17:26
jonadabIf there's a real chance Debian will recant and bring back reasonable init systems as a viable option, then sure.17:27
jonadabBut if they're going to keep marking systemd as a dependency for tons of different completely unrelated packages...17:27
James1138There are already websites against systemd - http://without-systemd.org/wiki/index.php/Main_Page17:31
James1138https://appuals.com/remove-systemd-ubuntu-permanently/17:31
James1138The 2nd link also says one of the choices is go Devuan.17:32
golinuxJames1138: without-systemd.org has been around for nearly 5 years17:39
James1138Ahh.. my bad GoLinux17:40
golinuxThe second link is a little scary.17:40
golinuxJames1138: Well, it's new to you.  :)17:40
golinuxThose of us who've been around since fall 2014 have a different perspective.17:41
golinuxDealing with an influx of former Ubuntu users would be "interesting".17:43
jonadabSure.  The fundamental design problems with systemd are well documented, and its implementation issues (bugs and so on) are reasonably well known too.17:43
golinuxBut that is really getting OT for this channel.  #debianfork is the place for that.17:44
jonadabBut ultimately something like 98% of Linux users were using either one of two "core" distros (Debian or RedHat), or a distro that is directly or indirectly derived from them.  So ultimately the decision to adopt systemd was made by a very small number of people.17:44
jonadabDevuan provides an important alternative.17:45
Tom-_jonadab, were you asking me the question "will it do any good"?17:52
Tom-_is talking about Debian and trying to get them to not be all systemd all the time on topic here?17:52
Tom-_i'm writing a message to the debian-init-diversity mailing list but I haven't sent it yet for fear that it will waste their time i guess17:53
golinuxIt is off topic in this channel17:53
golinuxto vent on systemd17:53
golinuxPlease that that discussion to #debianfork17:54
b17wiseHie all :)17:57
James1138Going back to Devuan... question. I see things like "flashcache" and "Memcache" - is there a suggested memory and/or disk cache application that is SSD-friendly?  Or is the whole cache thing unneeded?18:01
James1138I already have "preload" installed.18:02
jonadabFlash memory has usually faster read times than a spinning-platter disk, so caching reads is somewhat _less_ necessary.  But not pointless:  DRAM is faster than Flash memory, and SRAM is faster yet.18:03
jonadabNot sure about relative write speeds.  I know flash write speeds are much slower than flash read speeds, but I don't know how they compare to spinning-platter HD speeds.18:04
jonadabPersonally I tend to be annoyed at write caching on principle, but I can't deny the performance advantages in some use cases.18:05
slvrit buffers and flushes. You can mount with sync to prevent.18:05
slvrnote that flash has faster SEEK time than spinning rust, not always more throughput.18:07
James1138My Thinkpad T500 is already using Write-cache. Sounds like using SSD plus preload and write-cache... installing any extra cache and/or buffer application maybe a waste of time.18:09
b17wiseWhat device manager Devuan uses? udev or mdev? Is that Devuan packaging system vastly different from Debian?18:09
jonadabThe packaging system is the same.18:10
jonadabThough certain packages are different.18:10
FatPhilb17wise: packaging's the same. eudev replaces systemd's udev.18:12
golinuxb17wise: Devuan has eudev18:16
b17wiseI see18:18
Evilhamand it has eth0 and the likes by default, eudev is pretty cool19:45
Walex#20:03

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!