1308009620 M * Bertl looks good at first glance
1308009627 M * wmp and work
1308009636 M * Bertl btw, you can avoid the -- --arch amd64 if your host is amd64
1308009648 M * wmp O, thx ;)
1308009663 M * Bertl it is usually required on other distros, as they call it x86_64 instead
1308009750 M * wmp Bertl: it is possible to mount on guest folder from host?
1308009769 M * wmp without nfs and samba and other ;)
1308009786 M * Bertl yes, you can (bind) mount any host path into the guest
1308009793 M * wmp nice ;)
1308009799 M * Bertl just add an entry to the guest config's fstab
1308009806 M * Bertl (not the guest's fstab :)
1308009840 M * wmp O, www-debian work ;)
1308009873 M * Bertl http://linux-vserver.org/Upstart_issues
1308009885 M * Bertl you want to go through that for ubuntu with upstart
1308009913 M * wmp http://static.thedudes.org/initpost - dont work
1308009978 M * Bertl yeah, life is hard, use one of the other links :)
1308010011 M * wmp omg, i'm blind ;)
1308010045 M * wmp and i must reinstall www ?
1308010095 M * Bertl yes, create a new guest, you can use the same hostname though
1308010127 Q * derjohn_mob Ping timeout: 480 seconds
1308010207 M * wmp hmmm
1308010214 M * wmp i installed on www-debian sshd server
1308010222 M * wmp and i want to connect:
1308010230 M * wmp ssh root@188.165.18.196
1308010241 M * wmp and I connecting to host
1308010245 M * wmp no to guest
1308010252 M * Bertl known problem, listed in the FAQ, you need to restrict the host's sshd to host only IPs
1308010276 M * Bertl http://linux-vserver.org/Frequently_Asked_Questions#When_I_try_to_ssh_to_the_guest.2C_I_log_into_the_host.2C_even_if_I_installed_sshd_on_the_guest._What.27s_wrong_here.3F
1308010287 M * wmp ;)
1308010360 M * wmp ssh: connect to host 188.165.18.196 port 22: Connection refused
1308010378 M * Bertl restart the guest, sshd died inside the guest :)
1308010411 M * wmp command reboot? dont work
1308010442 M * Bertl well, reboot works, but that is more complicated
1308010446 M * wmp ok vserver www-debian restart
1308010450 M * Bertl yep
1308010467 M * wmp work! :)
1308010471 M * wmp I love Vserver ;)
1308010490 M * Bertl reboot (without -f) fails, because it tries to contact init
1308010502 M * Bertl and by default, a Linux-VServer guest does not run an init
1308010517 M * Bertl (it's simpler and uses less resources)
1308010535 M * wmp so in quest i cant make, eq: service mysql restart ?
1308010539 M * Bertl with ubuntu, you configure the guest to run upstart as init, so there the reboot will work
1308010553 M * wmp nice ;
1308010559 M * Bertl service mysql restart will work, that doesn't require init
1308010624 M * wmp Bertl: how to make smal partiton on VM?
1308010634 M * wmp acctualy i have size form partition on host
1308010665 M * Bertl you have several options, depending on what you use it for
1308010677 M * Bertl - separate partition/lvm for each guest
1308010679 M * wmp give me how to or manual ;)
1308010687 M * Bertl - shared space for several guests
1308010704 M * Bertl depending on the filesystem you can set disk limits for each guest
1308010719 M * wmp maybe separate
1308010746 M * Bertl then just create an lvm for your guest (or an partition) and mount that as rootfs
1308010757 M * Bertl (again via the fstab in the guest config)
1308010765 M * Bertl util-vserver will then mount it for you
1308010766 M * wmp hmmm, isn;t simply method?
1308010790 M * Bertl what's simpler than adding a single line to the fstab file?
1308010798 M * wmp ok ;)
1308010816 M * wmp i think about methon on kvm/vmware
1308010831 M * Bertl Linux-VServer is quite different from that
1308010842 M * Bertl think more chroot on steroids
1308010885 M * wmp ok ;)
1308011327 M * wmp Bertl: /usr/lib/util-vserver/vserver-build.debootstrap: line 170: /usr/lib/util-vserver/distributions/lucid/initpost: Permission denied
1308011346 M * wmp -rw-r--r-- 1 root root
1308011347 M * Bertl wrong permissions I guess :)
1308011359 M * wmp what permissions is good?
1308011385 M * Bertl -rwxr-xr-x 1 root root 4225 2011-04-19 02:36 initpost*
1308011393 M * nicholi 755 methinks
1308011433 M * wmp hmmm: http://wklej.org/id/546252
1308011511 M * wmp Bertl: ^ ?
1308011536 M * Bertl is there a guest called www?
1308011549 M * wmp yes
1308011563 M * Bertl so what's in /etc/vservers/www ?
1308011579 M * wmp ls: cannot access /etc/vservers/www: No such file or directory
1308011594 M * wmp but i doing: vserver www build ...
1308011598 M * Bertl looks like there is no guest called 'www' (anymore)
1308011600 M * nicholi doesn't vserver delete the files on an unsuccessful build ?
1308011602 M * wmp and i have this error with permission
1308011610 M * Bertl yep
1308011625 M * wmp so i can install new www?
1308011627 M * wmp :)
1308011629 M * nicholi only places i check is /etc/vservers and /var/vservers 
1308012235 M * nicholi Bertl: how exactly do you approach using vhashify when you update the root files being linked to on all the guests?
1308012308 M * Bertl with vhashify you do not 'update' the 'root' files
1308012324 M * Bertl you simply update the guests and rehash them when needed
1308012549 M * wmp all work great ;)
1308012557 M * wmp thx Bertl
1308012561 M * wmp and thal nicholi
1308012565 M * wmp thx*
1308012567 M * wmp bye
1308012572 M * Bertl you're welcome!
1308012576 M * Bertl have fun!
1308012581 M * wmp ;)
1308012595 Q * wmp Remote host closed the connection
1308013639 M * nicholi Bertl: but if you update the guests, won't those files never rehash with the 'root' files?
1308013665 M * Bertl there are no 'root' files in a typical setup
1308013676 M * Bertl there are guests, and there are unified files
1308013695 M * Bertl if the guests stay reasonably similar, the number of unified files is high
1308013763 M * nicholi and there is no real issues with unifying executable binaries?
1308013775 M * Bertl not that we know of
1308013859 M * nicholi since unifying is a hardlink, does that mean memory would be shared between C libraries and such when running between the guests?
1308013879 M * Bertl read only mappings will be shared, yes
1308014452 M * nicholi sort of get it now, what are the vapt/vyum/vrpm wrappers for though exactly ?
1308014479 M * Bertl executing apt/yum/rpm on guests
1308014515 M * nicholi ahh ok, nothing to do with unify/hashify then
1308014521 M * Bertl nope
1308014574 M * nicholi only issue then still with hashify is multiple links to the same file in a guest then eh ?
1308014611 M * Bertl yup, but only if the guest assumes them to be writeable
1308014621 M * Bertl which is not very common
1308014695 M * nicholi something i was looking to do was hashify specifially was java runtimes... you wouldn't know offhand of any nightmare stories of that yet would you ?
1308014829 M * Bertl I know many horror stories about java, but none of them are Linux-VServer specific :)
1308014851 M * nicholi lol ok :)
1308016912 M * Bertl off to bed now ... have a good one everyone!
1308016917 N * Bertl Bertl_zZ
1308023967 Q * ryker Quit: Leaving.
1308025322 J * sannes ~ace@cm-84.209.81.224.getinternet.no
1308028783 Q * hijacker Read error: Connection reset by peer
1308028784 J * hijacker_ ~hijacker@213.91.163.5
1308033622 J * derjohn_mob aj@88.128.65.44
1308034125 Q * derjohn_mob Ping timeout: 480 seconds
1308034638 J * ghislain ~AQUEOS@adsl2.aqueos.com
1308035969 Q * ghislain Read error: Connection reset by peer
1308036759 J * derjohn_mob ~aj@213.238.45.2
1308036778 J * ghislain ~AQUEOS@adsl2.aqueos.com
1308038496 J * kir ~kir@swsoft-msk-nat.sw.ru
1308038675 J * petzsch ~markus@dslb-092-078-234-171.pools.arcor-ip.net
1308039122 J * thierryp ~thierry@zanzibar.inria.fr
1308041164 P * kir Leaving.
1308042436 N * Bertl_zZ Bertl
1308042441 M * Bertl morning folks!
1308042452 M * daniel_hozac morning Bertl
1308042458 M * josef_ Morning Bertl, a good day as usual? :)
1308042481 M * daniel_hozac Bertl: http://people.linux-vserver.org/~dhozac/p/k/patch-3.0-rc3-vs2.3.1-pre1.1.diff seems to compile now at least... still missing quite a few rejects though
1308042654 M * Mr_Smoke If I may interject something here, isn't stabilization more likely to happen with the 2.3.0 series ?
1308042660 M * Mr_Smoke (vserver stabilization that is)
1308042752 M * daniel_hozac than what?
1308042757 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com
1308043273 M * Mr_Smoke Than 2.3.1 with 3.0 kernel ?
1308043324 M * Mr_Smoke All I'm saying is the current stable is still 2.6.22-2.2.0.7, and I'm wondering how much sense it makes to spend time on 2.3.1 
1308043335 M * Mr_Smoke No hard feeling, just an honest question :)
1308043599 M * BenG I've got a wierd problem here
1308043627 M * BenG my guest don't seem to want to use swap
1308043694 M * BenG a set of cyrus processes I have complain of having no memory but there's 3Gb of swap left they aren't using
1308043710 M * BenG anyone any idea what this is about?
1308043827 M * Mr_Smoke I'm tempted to say swap ain't memory
1308043844 M * BenG tell me more :)
1308043861 M * BenG I'm complete up for accepting it's my own lack of knowledge causing my confusion
1308043879 M * BenG no swap shows as used on either host or guest
1308043986 M * Mr_Smoke how much free memory is there in the guest ?
1308044013 M * Mr_Smoke Also, what's the error message and what are the circumstances?
1308044029 M * BenG there's a lot of imapd processes running
1308044048 M * BenG when it kicks off, there's about 10mb free or so in the guest
1308044074 M * BenG according to free running there
1308044078 M * BenG according to free running there
1308044231 M * BenG http://pastebin.com/VUDUR9hF
1308044236 M * BenG that's the cyrus error
1308044265 M * BenG it repeats as each run of lmtp is attempted - and of course fails
1308044314 M * Bertl cyrus imapd is known to hang around forever consuming resources
1308044331 M * Bertl windoof clients are known to open imapd connections every few seconds
1308044356 M * Bertl the result usually is imapd spawning many many threads consuming too much memory
1308044407 M * BenG yep
1308044422 M * BenG okay, it seems I can look at some db4 config options too
1308044434 M * BenG I should try those out and then get back to you
1308044455 M * BenG perhaps my server is not using swap simply because it doesn't need to?
1308044969 M * BenG hmmm, looks like this is a bug in the Berkeley DB package I'm using
1308046301 Q * thierryp Remote host closed the connection
1308047410 M * BenG okay, should be fine now, berkeley DB issue
1308047430 M * BenG was one of those "just decided to ask for help when I figured out the problem" things
1308047461 M * BenG "vs2.3.0.37-donotuse"
1308047464 M * BenG lol
1308047480 M * BenG not seen that tag before
1308047499 M * BenG usually "technology preview" or "alpha"
1308047534 M * daniel_hozac it's before that.
1308047611 M * daniel_hozac Mr_Smoke: it doesn't make sense to "stabilize" on an old unmaintained kernel.
1308047956 Q * petzsch Quit: Leaving.
1308050420 J * sannes1 ~ace@62.101.223.178
1308050471 M * Bertl off for now ... bbl
1308050476 N * Bertl Bertl_oO
1308050531 Q * sannes1 Quit: Leaving.
1308051764 J * thierryp ~thierry@zanzibar.inria.fr
1308053377 Q * thierryp Remote host closed the connection
1308053581 Q * LuckyLuke Quit: testing 2.6.39 on HPPA SMP 64bit
1308053619 J * cuba33ci_ ~cuba33ci@111-240-174-224.dynamic.hinet.net
1308053678 Q * cuba33ci Write error: connection closed
1308053687 J * thierryp ~thierry@zanzibar.inria.fr
1308053687 N * cuba33ci_ cuba33ci
1308053699 Q * thierryp Remote host closed the connection
1308054315 J * LuckyLuke ~luca@host65-83-static.228-95-b.business.telecomitalia.it
1308054615 J * sannes1 ~ace@62.101.223.178
1308055447 M * sannes1 daniel_hozac: To make the cgroup mount work in util-vserver in gentoo I had to do the following: http://www.sannes.org/misc/fix_gentoo_util-vserver_initscript.patch
1308055500 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net
1308055536 M * daniel_hozac there's more that needs to be done.
1308055673 A * LuckyLuke is running vserver on gentoo but hadn't ever looked in cgroup mount
1308056109 M * sannes1 daniel_hozac: What kind of stuff?
1308056142 M * daniel_hozac CGROUP_MNT_PER_SS for instance.
1308056167 M * sannes1 daniel_hozac: I have some problems when the mount is specified in /etc/vservers/.defaults/cgroup/mnt .. vserver xyz start does not create the xyz directory
1308056221 M * daniel_hozac do you have a kernel where it should?
1308056227 M * daniel_hozac is it configured to?
1308056235 M * daniel_hozac do you have a --debug run?
1308056257 M * sannes1 daniel_hozac: I mount it at /sys/fs/cgroup since that seems to be where it is rumored one should mount it .. :P
1308056320 M * sannes1 I can certainly make one, just vserver --debug vs1 start > debug.txt then?
1308056328 M * daniel_hozac 2>
1308056390 M * sannes1 ah :) stderr
1308056581 M * sannes1 got 2980 installed by the way.. and http://www.sannes.org/misc/debug_cgroup.txt
1308056761 M * daniel_hozac do you have /sys/fs/cgroup/tasks?
1308056811 M * sannes1 yes
1308056846 M * daniel_hozac is /sys in your namespace-cleanup-skip?
1308056880 M * sannes1 namespace-cleaup-skip?
1308056890 M * sannes1 :P probably not
1308056909 M * sannes1 ah, so it might be unmounted already at that point? :P
1308056919 M * daniel_hozac yep...
1308056948 M * sannes1 ah, that explain why that test fails, .. how do I skip it? and does it remain in the vserver then?
1308056955 M * sannes1 (in the vserver guest)
1308056975 M * daniel_hozac add it to namespace-cleanup-skip
1308057005 M * daniel_hozac along with /sys/fs/cgroup
1308057089 M * sannes1 That worked beautifully :)
1308057527 J * thierryp ~thierry@zanzibar.inria.fr
1308058293 J * petzsch ~markus@dslb-092-078-234-171.pools.arcor-ip.net
1308058631 M * sannes1 daniel_hozac: my patch missed a ; :P
1308058916 Q * thierryp Remote host closed the connection
1308058991 Q * BenG Quit: I Leave
1308059056 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com
1308059083 Q * BenG 
1308059145 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com
1308059195 Q * BenG 
1308059892 J * FireEgl FireEgl@2001:470:e056:1:646e:1e65:758:5c71
1308060758 M * pmjdebruijn Bertl_oO: kernel compiles and boots, haven't been able to test more than that yet
1308061436 N * Bertl_oO Bertl
1308061523 M * LuckyLuke uhm, something somehow must have changed the way routing is handled, now my vserver does not route correctly any more on the second interface. I'll try looking into source policy routing as in http://oldwiki.linux-vserver.org/Recipes
1308061598 M * LuckyLuke packets try to go out with the wrong source address
1308061606 M * daniel_hozac what kernel?
1308061609 M * Bertl LuckyLuke: kernel/patch version_
1308061616 M * Bertl s/_/?
1308061823 J * thierryp ~thierry@zanzibar.inria.fr
1308061956 Q * thierryp Remote host closed the connection
1308062379 J * dowdle ~dowdle@scott.coe.montana.edu
1308062600 Q * petzsch Quit: Leaving.
1308063083 M * LuckyLuke I'm not able to bisect it, I'm now running on 2.6.38.4-vs2.3.0.37-rc15. The host has ETH0 with one subnet and default gw to internet, and TUN0 with another subnet and that's an openvpn to a third, remote, lan. The vserver has an ip on eth0 (with access to the internet via default gw) and a DUMMY0 with an ip con the openvpn interface
1308063152 M * LuckyLuke from the vserver if I ping an host on the 'remote vpn side' the packets go through the vpn but with the LAN source address
1308063183 M * LuckyLuke hope my explanation is understandable
1308063891 M * LuckyLuke daniel_hozac, Bertl: I tried to explain better here - please have a look: http://pastebin.com/nbWXvqYG
1308063891 M * sannes1 Could you try with 2.6.38.8-vs2.3.0.37-rc17 ? I think there was a fix for some source address selection
1308063913 M * LuckyLuke sannes1: yeah I'm downloading it right now :)
1308063927 M * LuckyLuke I was busy trying 2.6.39 stuff this morning (without vserver patches)
1308064154 M * LuckyLuke uhm the pastebin I made is not entirely correct about IPs in the root server, but I think it is understandable anyway
1308064652 M * Bertl yep, recent ekrnel fixes it
1308064656 M * Bertl *kernel even
1308065247 M * sannes1 Bertl: How long has that not worked?
1308065276 M * Bertl not for long
1308065650 J * petzsch ~markus@dslb-092-078-234-171.pools.arcor-ip.net
1308067152 Q * FireEgl Quit: Leaving...
1308067408 J * bonbons ~bonbons@2001:960:7ab:0:31fb:7170:7a85:c8a6
1308067726 M * daniel_hozac started failing in 2.6.38.
1308067811 J * FireEgl ~FireEgl@173-16-9-3.client.mchsi.com
1308067962 Q * derjohn_mob Ping timeout: 480 seconds
1308068482 J * thierryp ~thierry@home.parmentelat.net
1308068628 Q * thierryp Remote host closed the connection
1308070350 M * Mr_Smoke daniel_hozac: you mean 2.6.38/39 are old and unmaintained ? 
1308070902 M * daniel_hozac 2.6.38 is, yes.
1308071538 M * Bertl note 2.6.38 vs 2.6.38.8
1308071586 M * daniel_hozac well, the 2.6.38 series is retired already
1308071592 M * daniel_hozac upstream anyway.
1308071594 J * manana ~mayday090@nat049-252-205-109.tvoe.tv
1308071665 M * Bertl it's still listed as stable and I'm somewhat confident it will become longterm
1308072105 M * daniel_hozac i thought i read somewhere that it was not going to be maintained any further.
1308072883 M * nicholi i've never compiled the kernel before...is there a recommended gcc version to use? or rather one you should definitely not use?
1308072924 M * Bertl don't use experimental gcc versions, don't use broken gcc/toolchains 
1308072939 M * Bertl that's about it I guess :)
1308072968 M * nicholi lol :3
1308073404 M * Bertl time for a nap now .. bbl
1308073417 N * Bertl Bertl_zZ
1308076709 Q * tokkee Read error: Connection reset by peer
1308076858 J * Piet_ ~Piet__@28IAABZJD.tor-irc.dnsbl.oftc.net
1308077272 Q * Piet Ping timeout: 480 seconds
1308077509 M * _Shiva__ Bertl_zZ: http://thread.gmane.org/gmane.linux.kernel.stable/12220
1308078193 M * trippeh_ nicholi: Generally whatever the distro provides by default is OK.
1308078626 M * trippeh_ Oh and the -longterm criteria is "somebody steps up to maintain it after gregkh drops it", not "this is a good release so we should let it live" :P
1308078665 Q * petzsch Quit: Leaving.
1308078812 M * sannes trippeh_: I usually pit a couple of versions against eachother in a battle royal and then see which comes out alive ..
1308078959 Q * cehteh Ping timeout: 480 seconds
1308078978 M * trippeh_ And then you step up to longterm maintain the one that wins? :)
1308079063 M * sannes trippeh_: hell no, I know for certain that it killed the other one, so I shoot it in the head, good riddance :P
1308079455 Q * transacid Remote host closed the connection
1308079650 J * transacid ~transacid@84.19.167.254
1308080539 M * trippeh_ awwh
1308081458 J * ghislain1 ~AQUEOS@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr
1308081727 Q * ghislain Ping timeout: 480 seconds
1308081972 J * derjohn_mob ~aj@d045167.adsl.hansenet.de
1308082034 M * nicholi trippeh_: yeah cept i'm noticing the one provided for ubuntu is slightly older, and the first version with cgroup implementation, wasn't sure if i should try using a newer patched one
1308082075 M * trippeh_ Oh I was thinking about the toolchain/gcc
1308082081 M * nicholi haven't really had any issues yet...which is usually a good sign to not upgrade i suppose :3
1308082096 M * nicholi trippeh_: oh ok....i was talking in terms of kernels
1308082101 M * nicholi i hear ya otherwise on toolchain
1308082290 N * Bertl_zZ Bertl
1308082293 M * Bertl back now ...
1308083212 Q * sannes Remote host closed the connection
1308084471 M * nicholi does the vunify-excludes list need to be moved to /etc/vservers/.defaults/whatnot ? or if nothing is specified it gets used from util-server location
1308084689 M * nicholi and some of the documentation seems to contrast on what the link in /etc/vservers/.defaults/apps/vunify/ is supposed to be called which points to the dir of unified files
1308084699 M * nicholi flower page says id, wiki says root...and my default install actually has 00
1308084846 M * daniel_hozac it's just an id
1308084849 M * daniel_hozac it can be anything
1308084868 M * Bertl and the flower page says nothing about  /etc/vservers/.defaults/apps/vunify
1308084896 Q * bonbons Quit: Leaving
1308084977 M * nicholi it does for me...
1308085026 M * nicholi http://www.nongnu.org/util-vserver/doc/conf/configuration.html <-- this is the flower page ?
1308085049 M * Bertl yup
1308085108 M * Bertl so what does it say about /etc/vservers/.defaults/apps/vunify for you?
1308085168 M * nicholi talks about exclude/pkgmgmt-force/ignore and the /hash subdir
1308085218 M * Bertl yep, but nothing about the directory itself
1308085222 M * nicholi sorry i meant vunify/hash before...if thats what you are referring to
1308085250 M * nicholi the link in that dir
1308085252 M * nicholi not the dir itself
1308085296 M * Bertl k
1308085313 M * Bertl and what is the problem with the wiki?
1308085334 M * nicholi the wiki example shows the link being called root
1308085347 M * nicholi but i guess it doesn't matter, just finds first link in dir ?
1308085361 M * daniel_hozac it finds whichever matches the filesystem your guest lives on.
1308085371 M * Bertl there should be one link per filesystem
1308085394 M * Bertl i.e. you can have a setup with several filesystems per guest
1308085410 M * nicholi oh, and the link is the name of the filesystem on guests to unify ?
1308085412 M * Bertl (and want all of them to be unified as far as possible)
1308085416 M * nicholi ok
1308085456 M * Bertl id -> description says: Points to a directory within the filesystems which are used for the vservers. There must be not more than one of such a directory per filesystem.
1308085489 M * Bertl which probably should read: There must not be more htan one ...
1308085526 M * Bertl (but no native speaker, so it might just be fine the way it is)
1308085594 M * nicholi no it makes more sense now, the dir on guest to unify...i was thinking in terms of host
1308085613 M * nicholi but how would you have 2 directories with same name on a filesystem ?
1308085666 M * nicholi it seems sort of redundant to mention that...because you can't make two dirs with same name
1308085669 M * nicholi or am i missing something ?
1308085760 M * nicholi maybe i'm making it more complex...i was asking before what the name of the link should be, not where it points to
1308085768 J * cehteh ~ct@pipapo.org
1308085817 M * nicholi you can't have multiple unified dirs on a single filesystem, is the warning? each unified dir must be on a different filesystem?
1308085969 M * daniel_hozac one link per filesystem
1308085994 M * nicholi per guest filesystem ?
1308086005 M * daniel_hozac per filesystem that contains guests, yes.
1308086017 M * nicholi ok, and the name of the link, is the dir on guest which you wish to unify ?
1308086021 M * daniel_hozac i.e. if you have guests in /vservers and /vservers2
1308086035 M * daniel_hozac the name of the link can be anything
1308086055 M * daniel_hozac foo, bar, 00, 01, monkey, elephant
1308086059 M * nicholi ok
1308086071 M * nicholi but /vservers and /vservers 2 sound like host paths to me
1308086074 M * nicholi not guest filesystems
1308086075 J * petzsch ~markus@dslb-092-078-234-171.pools.arcor-ip.net
1308086089 M * daniel_hozac of course.
1308086271 M * nicholi it only works if all the guests are contained within the same directory ... that's all?
1308086287 M * daniel_hozac on the same filesystem
1308086394 M * nicholi if /vservers2 was actually another filesystem then it would be ok ?
1308086406 M * daniel_hozac what?
1308086423 M * daniel_hozac what would be okay?
1308086460 M * nicholi unify/hashify ?
1308086491 M * nicholi how would the links work if you had 2 separate fileystems of guests ?
1308086502 M * daniel_hozac you'd have a link per filesystem
1308086508 M * Bertl you cannot unify across filesystems, but you can unify the files on each filesystem
1308086517 M * nicholi ahhh
1308086792 Q * ryker Quit: Leaving.
1308086805 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net
1308087527 J * wmp ~wmp@62.244.136.94
1308087533 M * wmp hello
1308087545 M * wmp i want to make CGroup
1308087557 M * wmp but when i run: mount -t cgroup -o memory /dev/cgroup
1308087577 M * wmp i have help from mount ;)
1308087597 J * cuba33ci_ ~cuba33ci@111-240-169-200.dynamic.hinet.net
1308087740 M * daniel_hozac why don't you let the util-vserver initscript do it?
1308087764 M * wmp cna i use util-vserver on ubuntu?
1308087776 M * daniel_hozac of course?
1308087782 M * wmp ok
1308087785 M * daniel_hozac just build it yourself.
1308087823 M * wmp ok, work ;)
1308087941 M * wmp daniel_hozac: and, how to limited memory?
1308087950 Q * cuba33ci Ping timeout: 480 seconds
1308087962 N * cuba33ci_ cuba33ci
1308088053 M * wmp daniel_hozac: in file: /etc/vservers/www/cgroup/memory.memsw.limit_in_bytes i shoud write limit?
1308088229 M * nicholi that sounds right to me
1308088235 M * nicholi just read all that yesterday
1308088439 M * wmp ...
1308088441 M * wmp  /usr/lib/util-vserver/vserver.functions: line 1590: /dev/cgroup/www/memory.memsw.4G: Permission denied
1308088461 M * Bertl is the swap controller enabled in the kernel?
1308088474 M * daniel_hozac the file name is limit_in_bytes.
1308088481 M * daniel_hozac the contents would be the limit.
1308088488 M * wmp Bertl: yes
1308088489 M * wmp daniel_hozac: thx
1308088495 M * wmp Bertl: i dont limited swap
1308088538 M * wmp echo "4G" > cgroup/memory.memsw.limit_in_bytes
1308088540 M * wmp yes?
1308088561 M * wmp  /usr/lib/util-vserver/vserver.functions: line 1590: /dev/cgroup/www/memory.memsw.limit_in_bytes: Permission denied
1308088732 M * wmp Bertlm daniel_hozac: ?
1308088834 M * Bertl sure you enabled the swap controller?
1308088846 Q * petzsch Quit: Leaving.
1308088851 M * wmp emmm, how to?
1308088864 Q * manana Remote host closed the connection
1308088869 M * Bertl what does /proc/mounts say about /dev/cgroup
1308088893 M * wmp vserver on /dev/cgroup type cgroup (rw,relatime,blkio,freezer,devices,memory,cpuacct,cpuset)
1308088909 M * Bertl so at least there the swap controller is not lsited
1308088911 M * Bertl *listed
1308088923 M * wmp ok
1308088924 M * Bertl i.e. I presume it isn't even enabled in the kernel config
1308088934 M * wmp so i use memory.limit_in_bytes 
1308089052 M * wmp hmmm
1308089059 M * wmp serwer is runnging
1308089065 M * wmp but have all memory
1308089068 M * wmp no 4GB
1308089330 M * wmp Bertl: ?
1308090223 M * nicholi you using a kernel that supports cgroups ?
1308090242 M * nicholi and new enough util-vserver ?
1308090373 M * wmp nicholi: hmmm, how to check this? 
1308090489 M * nicholi uname -a and vserver --version
1308090494 M * nicholi and following alot of this page
1308090496 M * nicholi http://linux-vserver.org/util-vserver:Cgroups
1308090687 M * wmp i kave kenrel with vserver patch
1308090714 M * wmp but i don't know it have support for cgroups
1308090724 M * wmp and i have new vserver
1308090898 M * wmp from my kenrel config: CONFIG_RESOURCE_COUNTERS=y CONFIG_CGROUP_MEM_RES_CTLR=y # CONFIG_CGROUP_MEM_RES_CTLR_SWAP is not set
1308090918 M * nicholi i read somewhere that cgroups came in with vs2.3.0.36.29
1308090937 M * wmp vserver 0.30.216-pre2981 
1308090997 Q * ryker Quit: Leaving.
1308091145 M * nicholi and you've already used more than 4gb in the guest ?
1308091165 M * nicholi probably have to restart guest after making the setting
1308091186 M * wmp i restarting
1308091594 M * nicholi were there any extra utils to observe what was being vhashed/unified in the guests? or just observe yourself how many links to an inode there are for files?
1308091822 M * wmp nicholi: ?
1308091886 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net
1308091936 M * nicholi question for Bertl/the rest
1308091946 M * wmp ohh ;)
1308092062 M * Bertl nicholi: the hash directory will contain a link to each unified file
1308092073 M * Bertl wmp: cgroups setup is not really Linux-VServer specific
1308092098 M * wmp Bertl: so what i dhoud doing to have only 4GB in quest?
1308092110 M * nicholi yeah i see, i mean any tools to analyze whats being unified/link...other than typical linux tools?
1308092135 M * nicholi my .hash is just getting alot bigger than i expected...since the guests are practically identical
1308092168 M * Bertl which is fine, no actual space wasted there
1308092171 M * nicholi i think i read somewhere it doesn't bother unifying "small files", correct ?
1308092181 M * nicholi oh right i guess
1308092188 M * nicholi hardlinks are hardlinks :3
1308092533 J * ghislain ~AQUEOS@adsl2.aqueos.com
1308092790 P * wmp Konversation terminated!
1308092800 Q * ghislain1 Ping timeout: 480 seconds
1308092897 J * WMP ~oftc@auburn.sored.pl
1308093226 Q * ghislain Quit: Leaving.
1308093252 J * chrissbx ~chrissbx@69-196-152-229.dsl.teksavvy.com
1308093298 M * WMP Bertl: so, how to limited memory?
1308093335 M * chrissbx Hi. How do you let the host offer files to guests securely (it's only a "give" operation, I would throw away changes done by the guests). Those are GIT repositories, actually small enough that I could just copy them.
1308093357 M * chrissbx But even copying needs atomic moves then to make it safe, and unlink would have to be done from within the guest, messy.
1308093370 M * chrissbx But maybe someone has already written the scripts for that?
1308093397 M * chrissbx I remember there's a kernel patch for readonly bind mounts, but sounds risky to me and I don't want to do so much patching.
1308093464 M * chrissbx And I don't want to use NFS nor samba. Anything else I missed?
1308093523 M * chrissbx Hm, I could use cow hardlinks to save space for the copy above.
1308093542 M * chrissbx So, where's the tool for that?
1308093656 M * Bertl not sure what you are looking for, but read-only bind mounts have been included in mainline for some time now
1308093708 M * Bertl WMP: what does your cgroup file contain?
1308093721 M * chrissbx hm ok, checking those, thanks Bertl.
1308093732 M * WMP Bertl: 
1308093734 M * WMP cat /etc/vservers/www/cgroup/memory.limit_in_bytes 
1308093735 M * WMP 4G
1308093758 M * Bertl does 4G look like a sane 'byte' value?
1308093772 M * WMP "values are stored in bytes. When writing to those files you can use suffixes: K,M,G."
1308093778 M * WMP from: http://linux-vserver.org/util-vserver:Cgroups#using_cgroup_to_enforce_memory_limits
1308093804 M * Bertl what does the guest dir (in cgroup contain)?
1308093862 M * WMP Bertl: in /etc/vservers/www/cgroup/ is only memory.limit_in_bytes
1308093888 M * Bertl and in /dev/cgroup/www ?
1308093902 M * WMP http://pastebin.com/UjUAKxUA
1308094053 N * ensc Guest4624
1308094063 J * ensc ~irc-ensc@p5DF2F5F5.dip.t-dialin.net
1308094070 M * Bertl and what does memory.limit_in_bytes and memory.stat contain?
1308094114 M * WMP memory.limit_in_bytes 
1308094116 M * WMP 4294967296
1308094122 M * WMP http://pastebin.com/SQ4B7vgV
1308094154 M * Bertl looks good, no?
1308094160 M * WMP yes
1308094196 M * Bertl so, what's the problem then?
1308094222 M * WMP http://wklej.org/id/546795
1308094337 M * nicholi the display in a guest is a different settings isn't it ?
1308094352 M * nicholi doesn't mean it's not limited
1308094352 M * WMP yes
1308094368 M * WMP i have this same output on host
1308094391 M * WMP http://wklej.org/id/546796
1308094398 M * nicholi When you wish the guests to see only their limited memory pool, be sure to include VIRT_MEM in your cflags config file. 
1308094445 M * WMP where is cflags?
1308094460 M * nicholi vserver-stat will show you how much memory the vserver is actually using
1308094465 M * nicholi thats what i use at least
1308094472 Q * Guest4624 Ping timeout: 480 seconds
1308094495 M * nicholi WMP: i believe they are referring to cflags when you compile the kernel
1308094501 M * WMP ok ;)
1308094505 M * nicholi otherwise not really sure, thats from the wiki i quoted (without quotes, lol)
1308094535 M * WMP cat /root/kernels/2.6.38.8-vserver/linux-2.6.38.8-vs2.3.0.37-rc17/.config | grep VIRT_MEM
1308094539 M * WMP return nothing
1308094548 M * nicholi yeah mine either
1308094642 M * nicholi oh nm...thats referring to context flags
1308094647 M * WMP yes ;)
1308094650 M * WMP http://linux-vserver.org/util-vserver:Capabilities_and_Flags
1308094650 M * nicholi http://linux-vserver.org/util-vserver:Capabilities_and_Flags
1308094653 M * nicholi lol
1308094654 M * WMP :)
1308094661 M * WMP first!
1308094663 M * WMP :)
1308094748 M * WMP nice ;)
1308094772 M * WMP Bertl: can i set this flag to default?
1308094789 M * WMP ;)
1308094792 M * WMP  /etc/vservers/.defaults 
1308095078 M * Bertl no default AFAIK
1308095179 M * WMP ?
1308095280 M * Bertl as far as I know, there are no default entries for cflags
1308095359 M * WMP i make file flags with  VIRT_MEM in /etc/vservers/.defaults 
1308095496 M * Bertl and, does it work?
1308095569 M * WMP i wait for new installation
1308095948 Q * dowdle Remote host closed the connection