IRC Logs for #crux-devel Tuesday, 2012-11-20

prologicjue, jaeger ok ... points taken00:54
prologicthis is specifically for supporting venet0 in openvz guests00:54
prologicso maybe I should create a crux-custom openvz set of templates?00:54
*** pedja has quit IRC01:31
jueRomster: mine has two but configured differently, external nic via dhcp, internal has a static ip02:36
jueI think the requirements wrt network setup differ much and the CRUX way to leave this up to the user is the most simple and most flexible way ;)02:42
Romsteradding another file for eth1 teth2 in /etc/rc.d/ is not a bad idea.02:43
Romsterprologic, you could just leave /etc/rc.d/net alone and have a netvz or something script.02:44
Romsteras part of the scripts to setup the template02:44
prologicjue, I agree02:45
prologicI think I'll just have custom scripts for openvz02:45
prologicmight package some of this stuff up as optional ports02:45
prologicwhen running crux in an openvz environment02:45
prologic(like I'm doing)02:45
Romsterand others can use them too.02:49
Romsteri should of ordered my SSD last week ugh i hate waiting.02:50
jueRomster: I don't say that it is a bad idea, but I'm running 4 different CRUX hosts here and don't need it on any of it02:50
Romsteryour right on the let user configure that sort of stuff.02:50
jueRomster: please don't offer such things ;)04:14
prologicwhat was he offering ?:)04:16
jueto build a CRUX with NLS stuff04:17
prologicoh hmm yeah04:19
prologiclet's not04:19
prologicthe user should add that him/her self04:19
prologiclike we've always done04:19
prologicit's not that hard to do right?04:19
frinnstwe could install /etc/rc.d/eth0 and /etc/rc.d/dhcpcd with dhcpcd for example04:20
frinnstthats how i've done it at home04:20
prologicI'm happy maintaining and shipping custom openvz templates04:20
prologicthat are 99.999% CRUX standard :)04:20
Romstermmmkay, just trying to help him04:21
juesure I know, but if we, the official maintainer, follow every user request CRUX wouldn't be the same as it is until now04:23
frinnstthen it would be fedora :)04:23
prologicit wouldn't04:23
prologicit would become Arch :)04:23
prologicor worse04:23
prologicit should and needs to remain clean and simple yet powerful and flexbile (attributes of simplicity)04:24
juePer sometimes said, that his most important job is to say "no" :)04:24
jueremark: not to everything, but at least in the first step04:25
Romsterwont ever be arch without systemd :P04:28
prologicbut that is a true statement04:28
Romsteri know my younger years of crux i was like this and that, but i see jues point. effort.04:28
prologicit's not so much effort alone04:29
prologicit's also added complexity04:29
prologiccomplex systems are hard to maintain and even harder to understand and use04:29
Romstersimple systems are more effort to get to where your going too and updates to end users but maybe not so bad as a binary distro to the end user. i dunno. our biggest selling point is simple to write ports for.04:30
*** deus_ex has joined #crux-devel04:44
prologicit's a simple architecture05:29
prologicbut con't confuse simple with "lacking in features"05:30
prologicYou need to watch:
prologicSimple made Easy05:35
Romsterhmm bookmarked06:00
Romsteri don't even know half these uses we are getting the past few weeks.07:26
horrorStruckjue: is it just me or /etc/rc.d/unbound restart fails for you too? i changed killall -q /usr/sbin/unbound to kill `pidof unbound` and it restarts fine now.07:51
Romsteri've came across issues with killall if yo do not stop it before you update then start.07:55
Romsterif you update then try to restart it'll fail.07:55
horrorStruckIIRC i had the case without update but i would need to recheck to be sure07:56
Romsterjaeger, when you got time can you look at talloc-32it is fine in a chroot but when i compile it on my system it seems to skip the python check and then fails on the mv lib ... lib32. seems everything goes into lib32 just fine on my system bar the python bindings. haven't ahd much luck figuring out a fix.08:01
Romsteri'm heading to bed, it fails on mine and another persons setup (vee_ iirc) but works in the clean chroot. g'night08:26
jaegerpaste some logs when you come back08:27
juehorrorStruck: works for me, but as Romster said, it might fail after an update of the binary08:36
horrorStruckjue: ok thanks.08:46
*** c0x` has quit IRC13:04
*** c0x has joined #crux-devel13:13
jaegerstarted working on my python irc bot again, fun project17:33
jaegeronce it's robust enough I'll replace clb with it since supybot is abandoned17:34
jaegerjue: gvim is missing a buildtime dep on xorg-libxt18:04
jaegerI tried to install it on a system without xorg already installed and it didn't build the GUI part because X11/Intrinsic.h wasn't available18:05
Romstermy SSD just arrived O_O20:06
jaegerthat was quick20:06
Romsteryeah really this scorptech company don't muck around20:06
Romsterit even has a mounting bracket too that'll be handy20:07
Romsteri wasn't too fused with that though as half my disks live on the desk itself.20:07
Romsteruntil i move those to a enclosure later.20:07
Romsteranything special i need to do with ext4 when i format this?20:08
Romsterfdisk aligns to 1MiB now so that's probably ok.20:08
Romsteror is a journal a bad idea on a ssd20:09
Romsterext4 has TRIM20:10
jaegeralign and mount with discard20:11
jaegerjournal is fine20:11
jaegermake sure all your partitions align, not just the firstr20:12
Romsterso really make sure all partions align to 1MiB and mkfs.ext420:12
jaegerswap also supports discard these days20:14
jaegerafter you've installed, set your scheduler for the SSD block device to "noop"20:14
Romsteroh godo call i'm using BFQ on everything else20:14
Romsteri hadn't thought about swap on SSD that would be a good idea too.20:15
Romsterhmm i can set priority to use ssd first and if need be use my rest of my other 2 LV swaps as a lower priority.20:16
jaegerI'd suggest just getting rid of the ones that aren't on the SSD20:17
Romsternot sure if but i would be best to initramfs and root in LV with the swap and ccache in there own LV on the SSD20:17
jaegerif you're gonna do it that way make sure TRIM doesn't get eaten by lvm somehow20:17
Romsteri really don't know what my root size wil come too, probably not as big as i have it now.20:17
Romsteri need to check up on that.20:18
Romsteri want to do this right the first time that way i wont have to do ugly moves redo copy back.20:19
Romsteri was just gonna do a parttition and root on entire SSD but it's very tempting for swap and ccache use too.20:20
Romsterbut then i have to watch that i don't run out of room20:20
Romsteri'll go the lvm way if it supports TRIM20:20
Romsterelse fixed partitions i guess i can always use a partition tool if i get into trouble with space.20:21
jaegerI avoid lvm these days, leads to annoyances when trying to rescue20:21
Romsterok i'm heading back to work now,  i'll check this out after work.20:21
jaegergreat if nothing breaks, though :)20:22
Romsteryeah i've got backup copies just in case.20:22
Romsterand i know the metadata format well to be able to recover stuff if need be, helped a few users out in #lvm hence i got chanop out of helping. and a few small patches.20:22
Romsteryou done a initramfs script? if so could you post the stuff you've done, would save me some effort. i'll look when i get back after work. later.20:23
jaegerit has no lvm support yet, for what that's worth20:32
jaegerstill too young20:32
*** _mavrick61 has quit IRC21:48
*** _mavrick61 has joined #crux-devel21:50

Generated by 2.11.0 by Marius Gedminas - find it at!