freenode/#devuan/ Monday, 2019-08-12

nexgenhello, please let me know how get a backup copy of the current whole repository (all debs)03:10
gnarfacenexgen: did you try wget?03:17
nexgenthere are no ISOs like for Debian?03:19
nexgenhaving complete set of debs?03:19
golinuxnexgen: Do you want just the Devuan packages or also the Debian packages fetched by redirect?03:19
golinuxNo iso set ot the packages03:20
nexgenI need a backup if Devuan for some reason gets offline for a significant time03:20
nexgenbtw, do you have a plan how to deal with situation if Devuan is attacked by systemD corpo invaders like it happened to Debian?03:21
nexgenand wounder how actually it happened to Debian03:21
nexgenit seems it was very slow and unnoticeable earlier03:22
golinuxThat discussion should probably go on #debianfork03:22
nexgenbut finally I was forced to use that systemD shit03:22
golinuxThat is the systemd bashing channel03:23
nexgen#devuan or @debianfork is bashing systemD ?03:23
nexgenthanks, got it, already there03:25
nexgendo you plan to have reproducible builds sometimes?03:26
nexgenor it is not possible with apt/dpkg?03:26
rseveroHow can I list the scripts called at boot time in a plain Devuan server?15:59
r3bootls /etc/rc`runlevel`.d/S??16:02
onefangI use sysv-rc-conf for that sort of thing.16:02
r3bootor something along those lines. Find the current runlevel (see the runlevel command or /etc/inittab), check the S prefixed symlinks in the rc dir for that runlevel (/etc/rc<number>.d/)16:02
r3bootonefang: of the nicer things about SysV init, is that you can just look at how the symlinks are created, no need for external tools iig16:03
r3boot(also, dont forget to check /etc/rc.local, since that can also contain commands that'll be started @ boot)16:04
onefangsysv-rc-conf is a nice little command line tool for listing and changing which sysv init scripts start up when.  Much nicer than dealing with symlinks and stuff.16:05
r3bootmja, to each his/her/its own ofc. I'd rather stick to tried & tested UNIX methods :)16:05
r3bootsysv init is fairly trivial to figure out, just like setting symlinks, so personally, mja :)16:06
r3bootoh lol :) That's yet another clone of update-rc.d/chkconfig :D16:07
r3boot(why do people keep re-inventing wheels ... idk)16:08
r3boot &
r3boot(first is SGI + RH + CentOS, 2nd is debian-likes)16:08
retak(I know why people do re-invent wheels. But I keep it for me ;) )16:09
r3bootNiH is strong ;P16:09
onefangThe "get off my lawn" is strong here.  sysv-rc-conf isn't just reinventing that particular wheel.  :-P16:14
rseveroGreat! Just installed sysv-rc-conf. Thanks you all.16:17
onefangYou are welcome.16:17
r3bootMja, it all depends onefang :) I have been raised to work with a bunch of unix systems (solaris, hp/ux, aix, *bsd, linux, bsdi, sco, tru64), and one thing that works across all of them (well, atleast the SysV-based ones), is the inittab+ln trick I just described. But, to each his/her/its own ofc. I am not trying to recommend any particular approach16:19
r3bootonly differences are where the OS stores it's rc.d dirs; For hp/ux for instance, thats /sbin/rc.d :)16:20
TwistedFatehi, i just booted devuan live and i don't have sound20:49
TwistedFatei checked alsamixer and it's not muted20:49
debdogrun alsa-info.sh20:49
debdogif it's there20:49
furrywolfis the live built with alsa or pulse?  I've never used it.20:49
debdogif not:
furrywolfif it has pulse, nothing you do with alsa is going to work properly.20:50
debdogif it has pulse, fsmithred should be... what's the word...20:50
debdog      Installed - Yes (/usr/bin/pulseaudio)20:51
debdog      Running - Yes20:51
debdogcheck whether it has pasuspender installed, too20:52
TwistedFatedebdog, it has20:52
debdogok, gimme some time to read through the link...20:53
debdogopen alsamixer and see which device it selects. top left corner20:54
fsmithredI think the live has pulseaudio, pretty sure20:54
furrywolfif you have pulseaudio, alsamixer is useless.  you should use pulse's volume control.20:55
fsmithredit's basically the same as installing xfce desktop from tasksel in the installer20:55
debdogtry pasuspender -- speaker-test -c2 -twav -l5 -Dplughw:1,020:55
furrywolfyou mean 'try apt-get purge pulseaudio'?  :P20:55
debdoghehe, that too20:55
debdogalso: pasuspender -- speaker-test -c2 -twav -l5 -Dplughw:0,020:56
fsmithredthat might not work so well20:56
furrywolfthat fixes >95% of my linux audio problems.20:56
TwistedFatedebdog, works20:56
TwistedFateso it's pulse problem20:56
fsmithredyeah, but you don't install task-anything-desktop20:56
furrywolfit's always a pulse problem.20:56
debdogTwistedFate: which one, the 0,0 or the 1,0?20:56
TwistedFatethe first one20:56
debdogok, seems audio's defsault is set to HDMI, TwistedFate20:57
debdog0 [HDMI           ]: HDA-Intel - HDA ATI HDMI20:57
debdog1 [Generic        ]: HDA-Intel - HD-Audio Generic20:57
TwistedFateat least i know my sound card works20:58
debdogto have sound create (if possible on live) /etc/asound.conf with content:21:00
debdogdefaults.ctl.!card Generic21:00
debdogdefaults.pcm.!card Generic21:00
debdognot sure 'bout ther Generic part21:00
TwistedFatedebdog, done21:01
TwistedFatehow do i restart alsa? :321:01
debdogyou don't. each program reads that file on startup. so you have to restart the program you play sound with21:02
TwistedFatemeh, still doesn't work21:02
TwistedFatei'm shocked that pulseaudio is a default on devuan21:03
debdogwait, in yopur case that's prolly pulseaudio21:03
debdogfucking everything up21:03
debdogso you might have to restart that one21:03
debdogif that doesn't work and it PA does not honor this file, it's yet another reason to get rit of it21:04
furrywolfasound.conf is useless when using pulseaudio.21:04
debdoghehe, why is that not surprising21:05
furrywolfif you're going to use pulse, you need to deal with pulse's configuration.  something that's not using alsa's libraries will never even open alsa's config file.21:05
debdogfsmithred: task, replace PA by apulse :D21:05
fsmithredwill bring it up at the next meeting21:06
furrywolfas you said, each program reads that file on startup...  when it initializes alsa.  which it doesn't do if it's using pulse instead.21:06
debdogoh, thought you're the head maintainer of the live-cd21:06
fsmithredI just tried sound in the desktop-live iso, and it works fine...21:06
furrywolfyes, eliminating pulse would be a good thing for devuan, and something I've asked for many, many times.  :P21:06
fsmithreduntil I open the PA volume control, and then the sound stutters21:06
debdogfurrywolf: I hoped PA is reading it on startup as well and then use "Generic" as device21:07
furrywolfyes.  pulse causes stuttering.  always has.  and other less-obvious audio defects.  also always has.21:07
fsmithredI had no idea. We just inherit the deps from debian. Sorry.21:07
debdogI see21:08
TwistedFatehow to restart alsa?21:08
TwistedFatejust purged PA21:08
fsmithredit's supposed to be easy to change the default device in PA21:08
fsmithredoh, ok21:08
fsmithred /etc/init.d/asla-utils restart21:08
debdogalsa is a kernel module, no need to restart it21:08
furrywolfas debdog said, you need to restart each program that uses alsa.21:08
furrywolfalso, note that purging pulse often does not actually stop the pulse daemon from running.21:09
furrywolfand you need to manually kill it.21:09
fsmithredwow, it is easy to remove, and yes, it does keep running after you remove it21:11
furrywolfyeah, I'd label not-actually-killing-the-daemon-you-just-deleted-all-the-conf-files-for a bug, but it's been like that _forever_ and no one seems to care.21:12
debdogI'd thnink that's a good solution21:13
debdogrunning processes should generally not be touched by the packaging software, well, with some exceptions, of course21:14
debdoglike updates. but in that case it asks one whether it should restart the deamons or not21:14
furrywolfI'm pretty sure the pre-removal scripts for just about everything else stop the appropriate software first.21:15
debdoghmm, what could be a nice process to test that with?21:15
debdoghah, PA21:16
debdogwait, no21:16
fsmithreddebdog, I just tested with tilda - it stayed open after I removed the package.21:38
fsmithredI think only services shut themselves down before removal21:38
debdoghmm, ok, I thought that applies to deamons, too21:40
debdogthank god I am not a server adminitrator :D21:40
debdoglol +s21:40
TwistedFatebtw, does sound work in firefox when alsa is used?21:45
debdogTwistedFate: if not, try apulse21:46
TwistedFatewhen devuan is installed with netinst, does it use alsa by default?21:50
fsmithreddepends on what you select21:50
debdogI think that depends on...21:51
fsmithredif you select a desktop environment, you get pulseaudio21:51
fsmithredif you just select standard system utilities you get alsa only21:51
fsmithredand no graphical environment21:51
fsmithred(all the cool people know that's the way to go)21:52
fsmithredthen add what you want and minimize the metapackages21:52
TwistedFatei wanted to install gentoo, originally.. but i'm to tired for that21:53
TwistedFateso devuan will have to do :S21:53
fsmithredI could never figure out what options to use on the kernel build21:53
TwistedFateit took me a while to figure out how to configure & compile a kernel21:54
TwistedFatetime to experiment a bit21:59
TwistedFategonna go with kde this time22:00
TwistedFatedoes kde work nice in devuan?22:00
fsmithredkde works22:01
fsmithredif you like that sort of thing22:01
fsmithredand have a lot of RAM22:01
slvr_will kde work better with a radeon?22:01
fsmithredno idea22:01
slvr_It sucks with a geforce. :/22:01
fsmithredyou're using nouveau or nvidia driver?22:02
slvr_nouveau, yes.22:02
ErRandirFYI, I did not use pulseaudio on Devuan Jessie and the upgrade to Ascii did not add it. Sound works like before. So you should only have to purge it once22:02
TwistedFatePhoneDo I need a /boot partition if I'm creating an EFI system partition?22:11
fsmithredTwistedFatePhone, no22:12
fsmithredif you're encrypting the system, you might want a separate /boot, but it's not necessary22:12
fsmithredbut if you do the full-disk encryption (full except for efi part) then grub is v e r y  s l o w  about responding to the passphrase.22:13
fsmithredare you installing from the live?22:14
TwistedFatePhonefsmithred: heck no, netinst22:15
TwistedFatePhonefsmithred: what are 'standard system utilities'?22:18
TwistedFatePhonein software selection (expert installer)22:18
_abc_Hello. Why would mkfs.ext2 be extremely slow writing a new fs to a new usb stick (usb 2.0 but clocked at 15MBps)22:20
_abc_It takes ages. -c option is not used. Devuan ascii.22:20
_abc_Something is seriously wrong here22:21
fsmithredTwistedFatePhone, things like less, bz2 and some other stuff22:22
fsmithredstuff you would expect to be there on the command line22:22
fsmithrednot bloat22:22
_abc_So, how long should a mkfs.ext2 on a brand new flash 32GB stick take? It seems to take forever here. Under devuan ascii?22:25
specingWhy not mkfs.btrfs?22:25
_abc_mkfs.ext2 -vDL label1 /dev/sdb122:26
slvr__abc_: did you partition it?22:26
_abc_^ this is my command running now, I had to interrupt it 3 times so far, without -D, now with -D22:26
_abc_slvr_: it was partitioned with a single fat32, commercial new stick. I changed the partition flag to 83 using fdisk and then formatting it.22:27
_abc_I've formatted umpteen sticks on this system, first time I see such a mess22:27
slvr_I strongly prefer to remove all partitions and create a new table.22:27
_abc_"Writing inode tables" proceeds at about 1 table every 5 seconds for 232 tables.22:27
_abc_slvr_: not relevant. There is only one partition.22:27
_abc_So how long should it take?!22:28
_abc_Is it linear with disk size?22:28
_abc_I mean this is ridiculously slow. Something is seriously amiss?22:29
slvr_remove the partition. Make a new one. Start at sector 2048. type 8e.22:30
_abc_Why 8e?22:30
debdoganything suspicious on dmesg?22:30
slvr_because you need some LVM in your life.22:30
_abc_nothing in dmesg22:30
_abc_slvr_: do get a life.22:30
slvr_you can always strace the process22:30
slvr_not like I'm a storage admin or anything, but that's what I'd do.22:31
_abc_I'd rather hear from others how long it takes to format a silly partition?22:31
_abc_At least with -D it seems to not get stuck.22:31
fsmithred_abc_, not that long22:31
gnarfacewell wait... how long is long here?  because it could take *hours*22:32
gnarfaceit could take *days* with a badblocks pass added in22:32
_abc_fsmithred: Hi. I've never seen it go this slow in 25 ? years of linux (since 1995 or so)22:32
specingMaybe the stick is a dud22:32
fsmithredI've seen old sticks slow down22:32
_abc_The stick is brand new and I tested it by filling it to the end.22:32
gnarfaceext2/3/4 are known to format and resize the slowest, and cheap flash can have dramatic speed differences, even brand new stuff22:32
_abc_No slow down. Kingston DataTraveller G4 32GB22:33
gnarfaceyou can use "iotop -a" to see if it is actually writing anything22:33
_abc_good idea22:33
slvr_I bet the partition extends past the end of the disk.22:33
_abc_slvr_: no22:33
slvr_Wifes new stick was like that.22:33
specing_abc_:ext2 format takes less than a second for me22:33
_abc_I did not buy it in an Asian market.22:33
slvr_just by a few bytes, a mistake, not a scam.22:34
specing_abc_: modprobe zram num_devices=4 ; printf "32G" > /sys/block/zram0/disksize; mkfs.ext2 /dev/zram022:34
gnarfacespecing: you might be doing the lazy format thing though?22:34
_abc_specing: on what hw22:34
slvr_man you have a hell of a tall high horse. haha.22:34
gnarfaceext2 won't take "seconds" unless you're using a small SSD22:34
specinggentoo, core2duo, zram emulated storage device (in-memory)22:34
gnarfaceor are doing the lazy non-format format pass22:34
gnarfaceoh, or a ram disk :-p22:34
specingyes :)22:34
specing_abc_: repeat on zram to see if the problem is OS22:35
_abc_It is not the os for sure, the hw is a little bottleneck. USB2.0 only available now for this22:35
specingI'd bet on dud stick though22:35
_abc_I already said I filled it to the end, speed was nice.22:35
gnarface_abc_: i'm not sure it is a dud honestly.  i've seen this on regular working stuff that was just cheap.22:36
_abc_Well with -D it is slow and constant. Running iotop now, just installed22:36
_abc_Kingston is not particularly cheap or bad. I like it more than other lower cost brands.22:36
gnarface_abc_: other things that can slow it down are lack of DMA access (using PIO mode should be detectable with hdparm) or known issues performance with the usb block device drivers on ARM with the recent late end of the 4.x kernels...22:37
_abc_This is Intel22:37
_abc_but older 32 bit coreduo22:37
gnarfaceah, right.  nevermind the ARM thing probably then22:37
gnarfacealso try "hdparm -i [device]22:38
_abc_iotop says the process is io bound and mkfs writes22:38
gnarfaceer,  "hdparm -i [device]"22:38
_abc_mkfs leads the iotop table by a long margin.22:39
_abc_Disk Write in iotop is per second or total?22:39
gnarface_abc_: if you run "iotop -a" for a while (note "-a" for cumulative) you should see one of the processes slowly gaining a lead in total throughput.  basically as long as it is counting up, i'd let it keep going until it says some value ridiculously higher than the total space on the flash card, and i would not assume it is dead unless it froze before reaching that22:39
gnarface_abc_: yea, without "-a" it is just a per second speed rating.  that's not quite as helpful to you in this scenario.22:40
_abc_gnarface: it's not supposed to fill the card. But it is writing and increasing.22:40
_abc_Useful command, noted22:40
_abc_I'll let it run to the (bitter) end22:41
gnarface_abc_: well, here's the thing.  if you're formatting with ext2/3/4, it absolutely *does* fill the whole card.  but if you mean that you partitioned it and you're only formatting one partition, obviously do some math and just count to the size of that partition22:41
_abc_Well it's one partition and just about fills the disk apart from the mbr22:42
gnarfacewhat type of write speeds was iotop showing you without "-a" ?22:42
_abc_Without -a it's hopping in and out but speeds are like 7.5 Kbytes/sec ?!22:43
_abc_I assume it writes lots of small files.22:43
_abc_It is better to ext2fs format a binary image in ram then write it out with dd?22:43
_abc_I have no idea how this thing is coded but the speed (lack of) is ridiculous.22:44
_abc_Why would a ext2fs format fill the disk? It should write only inode and superblock tables, no?22:45
_abc_kernel is Linux devuan 4.9.0-8-686 #1 SMP Debian 4.9.110-3+deb9u6 (2018-10-08) -- latest?22:45
_abc_for ascii22:45
gnarface_abc_: ext* is primitive as fuck.  if you want fast formats use xfs or reiserfs.22:46
_abc_What percentage of the disk would ext2 housekeeping structures occupy? 5%22:46
_abc_Isn't reiser deprecated?22:46
gnarfaceonly by the debian installer22:46
_abc_Hm. Is Reiser still in jail?22:46
gnarfacedefault reserved space for ext* is 5% i think, but you probably only need to really reserve space on the / partition and whatever partition your system daemons log to22:47
gnarfaceheh, yes he is still in jail22:47
_abc_I know, I was asking about what the housekeeping consumes. How much space.22:47
gnarfaceprobably a lot less than 5% on anything of modern size22:47
gnarfacexfs is really good and won't get you any flak22:49
_abc_I've done a lot of 8-16GB ext2 formatting sticks lately never saw it so slow, but that was before upgrading the kernel. Is the new kernel (above) known to have some hiccups?22:49
gnarfaceif you're using commercial software like Steam however, using something other than ext4 has historically proven to be more trouble than it is worth22:49
gnarface(but note that is Valve's fault, not XFS)22:49
_abc_I'm not into steam. I'm into disk image manipulation, for openwrt and other purposes.\22:49
gnarfaceyou might really like xfs22:50
gnarfaceit is good with big files22:50
_abc_I might, or not, I need these things to be conservative so I can open them on ancient devices.22:50
_abc_xfs is in xfstools?22:51
_abc_Also does that not need a kernel module support?22:51
gnarfaceevery filesystem requires a kernel module and dedicated userspace tools.  ext2 is not an exception to that.22:52
gnarfaceyour stock debian kernel should have all these modules already built22:52
gnarfacebuilt and installed, maybe or maybe not automatically loaded22:53
_abc_cat /proc/filesystems |grep xfs -> nada22:53
gnarfacelsmod |grep xfs22:53
_abc_sudo modprobe xfs -> success22:54
gnarfacefind /lib/modules/`uname -r`/ -iname 'xfs*'22:54
_abc_Yeah overkill I guess, it is in now22:54
gnarfaceyea you'll need xfsprogs too22:54
_abc_What is 'nodev' in cat /proc/filesystems22:54
gnarfaceunused, i assume22:55
gnarfaceor unused by physical devices at least?22:55
_abc_strange because some are also not used and do not have that22:55
gnarfacenon-kernel mounts?  i dunno22:55
_abc_btrfs is good for what usually?22:55
gnarfaceit's supposed to be the thing that will save you from the shitty choice of ext* or xfs.  but multiple sources tell me it is not advised for production use yet.  you said you wanted old and reliable; i'd avoid btrfs22:56
_abc_I've used reiserfs on etch, it was smooth for me, others complained at the time.22:56
TwistedFatePhonei got disconnected22:56
TwistedFatePhonecan i upgrade to ceres from stable, or i gotta switch to testing first?22:57
gnarface_abc_: the biggest problem with reiserfs is that if your disk fsck process triggers on boot before DMA is engaged for the block devices, say on a raid or something, fsck.reiserfs can take forever and a half to complete in PIO mode22:57
_abc_Okay the mkfs finished.22:57
TwistedFatePhoneMinceR: fsmithred halp22:58
gnarface but yea, formatting and resizing are things inherently faster in xfs and reiserfs due to fundamentals of the design22:58
_abc_I should have timed it. -D was crucial, would not finish without it, time between screen updates seemed to go up exponentially22:58
_abc_I think it's the new kernel doing this.22:58
MinceRTwistedFatePhone: no idea, sorry22:58
_abc_I't s hunch for now, I'll reboot into the previous kernel and try some things tomorrow.22:58
TwistedFatePhonedo you know?22:59
gnarface_abc_: high cpu load during block writes could be a driver or kernel issue, that's not unheard of.  you might want to try some adjacent versions to compare speeds22:59
gnarfaceTwistedFatePhone: downgrades are unsupported by debian23:00
_abc_I also had an unrelated problem with a SD card today, vfat, I had to use ddrescue with raw device access on read to read it without errors. The 4.x series kernels seem to be less than perfect sometimes.23:00
TwistedFatePhonewhat?, i don't wanna downgrade, i wanna upgrade23:00
gnarfaceTwistedFatePhone: i'm pretty sure ceres still comes after stable23:01
TwistedFatePhoneceres is unstable, is it not?23:01
TwistedFatePhonebeowulf is new testing, right?23:03
gnarfacei don't think it is in sync with debian yet23:04
gnarfaceso it corresponds to devuan stable but it is not actually "new testing"23:04
gnarfacecorresponds to *debian* stable i mean23:04
gnarfacestill currently in testing for devuan23:04
TwistedFatePhoneshould i just use 'testing' and 'unstable' in sources?23:04
gnarfaceno, this is actually an example of why NOT to use those keywords23:05
TwistedFatePhoneand do i also keep using security and updates if i'm switching to ceres?23:05
gnarfaceno they only exist for stable23:05
TwistedFatePhonehow do i upgrade to ceres if devuan is in beowulf limbo?23:06
gnarfacedebian's official recommendation is to reinstall fresh23:06
gnarfacemerge config files by hand23:06
TwistedFatePhonethis is fresh netinst23:06
gnarfacewhy not just do a fresh ascii netinst then upgrade to beowulf then?23:07
TwistedFatePhonethis is ascii netinst23:07
TwistedFatePhoneso, apt-get dist-upgrade?23:08
gnarface"apt-get update && apt-get upgrade && apt-get dist-upgrade"23:08
gnarface(it is advised to do "upgrade" and "dist-upgrade" as separate steps though it rarely matters)23:08
TwistedFatePhonedone, swching to ceres in sources niw23:10
gnarfacei thought you said you were trying to upgrade ascii to beowulf23:10
gnarfacewhy are you referring to ceres at all?23:10
gnarfaceTwistedFatePhone: ^23:10
TwistedFatePhonei want to upgrade to ceres23:11
TwistedFatePhonei was just wondering if i can do it from stable23:11
gnarfaceupdate from ascii to beowulf first, then from beowulf to ceres23:11
TwistedFatePhoneor do i have to switch to testing first23:11
TwistedFatePhonegot it23:11
TwistedFatePhonealready done23:11
gnarfaceyou DO have to go through testing first, but your initial question was about downgrading not upgrading.  downgrading is not supported.23:11
TwistedFatePhoneso, do i change sources to 'beowulf'?23:12
TwistedFatePhoneor is apt-get dist-upgrade enough?23:12
gnarfaceyou need to change the sources and run "apt-get update" first, every time23:13
gnarfacenote that apt-get does not consider "update" and "upgrade" to be synonyms23:14
TwistedFatePhoneJesus, this thing is freaking fast23:15
TwistedFatePhoneuh oh, after upgrading, i can no longer use 'reboot' and 'halt' frm terminal..23:37
TwistedFatePhone'bash: reboot: command not found'23:37
slvr_try /sbin/reboot or /usr/sbin/reboot. Something broke with the paths for root recently.23:38
fsmithredtry /sbin/reboot23:40
fsmithredoh, sorry. was already said.23:40
TwistedFatePhonethat works23:40

Generated by 2.17.0 by Marius Gedminas - find it at!