1458708544 Q * FireEgl Ping timeout: 480 seconds
1458709106 J * FireEgl FireEgl@2604:2d80:8410:8067:d0c7:d8f7:e302:cc0a
1458715841 N * Bertl_zZ Bertl
1458715846 M * Bertl morning folks!
1458720233 J * Ghislain ~aqueos@adsl1.aqueos.com
1458720360 Q * derjohn_mob Ping timeout: 480 seconds
1458720890 J * nikolayK ~nikolay.k@199.91.137.248
1458722547 Q * nikolayK Read error: Connection reset by peer
1458722563 J * nikolayK ~nikolay.k@199.91.137.248
1458722853 J * PowerKe_ ~tom@d54C69995.access.telenet.be
1458722948 Q * PowerKe Read error: Connection reset by peer
1458723954 Q * PowerKe_ Ping timeout: 480 seconds
1458724974 M * Ghislain hello
1458724975 M * Ghislain http://www.phoronix.com/scan.php?page=news_item&px=CGroup-Namespaces-Linux-4.6
1458725022 M * Ghislain woooot ! systemd should be usable in 4.6 with vserver with some premount of device and mountpoint with that
1458725113 M * Ghislain daniel , get ready for this ;p
1458726063 J * derjohn_mob ~aj@fw.gkh-setu.de
1458729826 M * Bertl looking forward to it :)
1458732585 J * bonbons ~bonbons@2001:a18:223:8801:e0:3233:877a:bf9e
1458732637 J * Gremble ~Gremble@cpc87151-aztw31-2-0-cust755.18-1.cable.virginm.net
1458735211 Q * Aiken Remote host closed the connection
1458737838 Q * Gremble Quit: I Leave
1458738622 Q * bonbons Quit: Leaving
1458742159 M * Ghislain there seems a general move to systemd that want to make linux like windows ( ONE program that does everything instead of the unix one thing but weel done)
1458742174 M * Ghislain then there was a move to dockers and containers para virt etc..
1458742197 M * Ghislain then the two colide and explode has systemd want to be a container and broke everything
1458742215 M * Ghislain then they start to find ways around this while keeping making linux like windows
1458742232 M * Ghislain BUT in the paravirt/docker space
1458742252 M * Ghislain at last we will have revenge, at last we will reveal ourself to the kernel !
1458743393 M * Bertl :)
1458743419 M * Bertl I recently updated gnome on my laptop, and it seems the windows trend continues
1458743472 M * Bertl the gnome-terminal has now a huge title bar and even bigger tab bar with a gravestone and a big ugly menu button on it :)
1458743687 M * Bertl anyway ... off for now ... bbl
1458743691 N * Bertl Bertl_oO
1458745508 M * Ghislain i am looking forward to unbuntu 16.04 with kde neon, seems promising
1458745627 M * Guy- kde packaging in Debian seems to have stalled
1458746061 M * Ghislain well most desktop users moved to ubuntu i think
1458746079 M * nikolayK Gentoo with OpenRC and xfce4 for DE - simply amazing! No systemd involved ;-)
1458746091 M * Ghislain the issue is that servers starts too as packages are lagging too much for a lot of uses
1458746130 M * Ghislain i tried gentoo looooong time ago, stopped when emerge was recompiling openoffice  for 4h each update, i think they have  binary package now
1458746181 M * nikolayK indeed ;-)
1458746197 M * nikolayK openoffice-bin, etc
1458746232 M * nikolayK it is all about the freedom of choice ...
1458746258 M * Ghislain antegros is a rolling distro too, hard to find the spot between rolling when everyday is a bet, and stable that has still apache 1.3 and php4 (joking..not so much but...)
1458747532 J * PowerKe ~tom@d54C69995.access.telenet.be
1458748992 M * Guy- I've been using Debian sid almost universally since about 1999
1458749002 M * Guy- no, 1997
1458749046 M * Guy- this way I can avoid upgrading lots of machines in bursts, when new stable releases happen
1458752085 M * gnarface Ghislain: they're up to apache 2.2 and php5 now ;)
1458752342 M * gnarface what's gonna happen then anyway? will linux-vservers start to require systemd like everything else?
1458752381 M * gnarface or remain systemd-free? or actually successfully make it optional despite the fact nobody else has figured out how so far?
1458752406 A * gnarface votes for remaining systemd-free
1458752459 M * gnarface or will it just be up to the distro at that point, leaving vservers basically systemd-agnostic?
1458752621 Q * derjohn_mob Remote host closed the connection
1458752885 Q * nikolayK Quit: Leaving
1458754471 J * Gremble ~Gremble@cpc87151-aztw31-2-0-cust755.18-1.cable.virginm.net
1458754501 M * Guy- I've managed to keep my servers and vservers systemd-free so far and intend to keep doing so as long as feasible
1458754514 M * Guy- (I use runit instead)
1458754935 M * CcxCZ Heh, I don't think I've seen anyone even propose tying down vserver to something like that.
1458755026 M * CcxCZ But cgroup namespaces would certainly be nifty to have inside vservers. Cgroups are actually pretty easy to use and way more than systemd can do so. The implementation in openrc was like one afternoon of work.
1458755052 M * gnarface hmm. interesting
1458755110 M * CcxCZ My runit scripts assign processes into cgroups too, whenever available.
1458755140 M * CcxCZ It's one added line of shell in most cases.
1458755183 M * Guy- CcxCZ: can you post an example runit script?
1458755285 M * Guy- I'd also be interested in knowing how you benefit from putting your services into cgroups, specifically (I can think of a few things)
1458755369 M * CcxCZ Well, my zsh-based runit *framework* is up here: http://wpr.cz/ccx/bzr/zsv/ but basically you can just echo 0 >/sys/fs/cgroup/cgroupname/tasks in ./run before exec
1458755402 M * Guy- oh, it's an entire framework :)
1458755407 M * Guy- thanks, I'll look at it
1458755412 M * CcxCZ http://wpr.cz/ccx/bzr/zsv/head/zsh-functions/zsv_set_cgroup.html is the stuff related to cgroups, not sure if it would benefit from an update wrt. unified hierarchy
1458755412 M * Guy- I also use zsh anyway
1458755444 M * Guy- so how do you benefit from the cgroups?
1458755495 M * CcxCZ No particular benefit as of now, main use was supposed to be more granular monitoring, but haven't got that far with it yet.
1458755521 M * Guy- too bad
1458755554 M * CcxCZ Due to not finding anything I'd find usable for monitoring I'm building my own thing. Ditto for configuration management.
1458755605 M * CcxCZ So yeah, lot of work before I get to the nice extras. :]
1458755624 M * CcxCZ https://xkcd.com/1319/
1458755662 M * Guy- yeah, been there :)
1458755702 P * undefined 
1458755731 M * CcxCZ But if you do zsh for system management, I have nice thing for you. Gimme a sec…
1458756018 M * CcxCZ Oh, apparently I don't have any docs for this one besides what's in the source code. But the main executable is pretty well commented. http://wpr.cz/ccx/bzr/confz/ It's babushka/marelle-style "framework" (function runner really, that is able to do per-action variable scoping)
1458756099 M * CcxCZ I can do stuff like this with it: http://wpr.cz/ccx/paste/2016-03-23/0/ http://wpr.cz/ccx/paste/2016-03-23/1/
1458756243 M * CcxCZ Those fs_* requires come from another project of mine: http://wpr.cz/ccx/bzr/fileset/ … I should really work on those READMEs and setup some overview page or something. :]
1458756293 M * Guy- I don't have time to look at this right now
1458756299 M * Guy- but I'll check it out later, thanks
1458756301 M * CcxCZ I try really hard to build generic reusable tools whenever I find functionality missing, hence it's all spread out around half a dozen repos.
1458756559 M * Guy- fwiw, my most complete zsh project so far is this: https://github.com/akorn/zfsbackup
1458756666 M * CcxCZ Ah. Contrast with http://wpr.cz/ccx/bzr/snaprep/ :P
1458756711 M * Guy- hah, quite similar :)
1458756740 M * CcxCZ I do hardlinks instead of snapshots, but there's no reason that couldn't be added as a feature.
1458756758 M * Guy- snapshots are expensive with lvm
1458756774 M * CcxCZ I think I have a rudimentary btrfs support in there even.
1458756774 M * Guy- (due to cow)
1458756786 M * Guy- I only do zfs
1458756794 M * CcxCZ Well, so it is on COW filesystems. :]
1458756801 Q * Gremble Quit: I Leave
1458756819 M * Guy- zfs isn't cow :)
1458756828 M * Guy- which is why snapshots are cheap
1458756833 M * Guy- anyway, gotta run
1458756834 M * Guy- bbl
1458757015 M * CcxCZ Everywhere I've looked I've seen it described as such. (Just checked wikipedia articles on zfs and on cow)
1458757022 M * CcxCZ bye, ttyl
1458758632 J * Aiken ~Aiken@d63f.h.jbmb.net
1458761809 J * derjohn_mob ~aj@145.253.118.26
1458762247 J * undefined ~undefined@00011a48.user.oftc.net
1458763082 M * Guy- CcxCZ: well, it's not cow in the sense lvm2 is cow
1458763114 M * Guy- CcxCZ: lvm2, when you issue a write, first reads the original block and copies it to the snapshot, then overwrites the original
1458763150 M * Guy- CcxCZ: zfs on the other hand never overwrites data in place, so that having a snapshot just means not marking the previous data block as free
1458763157 M * Guy- there is no "copy"
1458763532 M * CcxCZ Yeah, otoh with lvm you only need to do COW as long as you are creating backup. TBH the benchmarks I've been seeing were rather unconvincing wrt. zfs performance, being only neglibly faster for RDBMS workload compared to lvm cow/journalling fs and essentially half of the speed than without cow.
1458763585 M * CcxCZ Perhaps it depends on how often the blocks you are rewriting are in cache anyway, then it doesn't matter.
1458764087 M * Guy- CcxCZ: in my own experience, since lvm introduces a sync write on each snapshot copy (the copied block has to be committed to disk before the original can be overwritten), performance of a filesystem drops by an order of magnitude if you make a snapshot
1458764109 M * Guy- (if you don't have sufficient bbwc, and as long as you're overwriting "old" blocks)
1458764164 M * Guy- I'll grant you zfsonlinux is not blazing fast, but it's fast enough; and you can help its performance by adding SSDs
1458764191 M * Guy- I think the other benefits outweigh the performance hit
1458764437 Q * derjohn_mob Ping timeout: 480 seconds
1458764595 M * CcxCZ So far the performance hit of lvm snapshot wasn't problem for me, so I had no need to try anything else (though I'm aware of zfsonlinux, as well as BSD support for it). I'm currently using btrfs for replaceable data storage too, haven't had issues since linux 3.0
1458765157 M * Guy- CcxCZ: ah, I remember seeing this zsv thing before
1458765503 M * CcxCZ It's not as interesting by itself (except perhaps for the usage of preconditions instead of dependencies). But the plan is to integrate it with a tool that is able to tell when running processes are out of date compared to their configuration and reload/restart what's needed. That's partly implemented using confz.
1458768623 J * derjohn_mob ~aj@fw.gkh-setu.de
1458771276 Q * Ghislain Quit: Leaving.
1458773092 J * Ghislain ~aqueos@adsl1.aqueos.com
1458773176 Q * Ghislain 
1458775665 M * Bertl_oO off to bed now ... have a good one everyone!
1458775667 N * Bertl_oO Bertl_zZ