1285720701 J * imcsk8 ~ichavero@201.156.204.25 1285724862 Q * imcsk8 Quit: This computer has gone to sleep 1285732781 Q * C14r_ Remote host closed the connection 1285732783 J * C14r ~C14r@mail.cipworx.de 1285737494 J * derjohn_mob ~aj@d046038.adsl.hansenet.de 1285741584 Q * derjohn_mob Ping timeout: 480 seconds 1285741687 J * petzsch ~markus@p4FF45EBF.dip.t-dialin.net 1285742165 M * Bertl off to bed now ... have a good one everyone! 1285742171 N * Bertl Bertl_zZ 1285742900 J * derjohn_mob ~aj@213.238.45.2 1285743622 J * ntrs_ ~ntrs@77.29.115.18 1285744330 J * ghislain ~AQUEOS@adsl2.aqueos.com 1285745339 M * _Shiva_ fLoo: make that a "still no .7 patch" ;-) 2.6.35.7 is out ;-) 1285745366 M * _Shiva_ (but yes - .5-vs applies cleanly to .6 and .7) 1285748067 J * barismetin ~barismeti@zanzibar.inria.fr 1285750840 M * Alteisen hi all 1285750879 M * Alteisen which CONFIG_CGROUPS options do i need for linux-vserver? 1285750969 M * Alteisen i am looking for CONFIG_CGROUP_MEM_RES_CTLR, but i can not find this option im my 2.6.35.5 config 1285751114 J * imcsk8 ~ichavero@201.144.130.18 1285751226 Q * imcsk8 1285753734 M * Alteisen ah, got it 1285760706 Q * petzsch Quit: Leaving. 1285761206 M * fLoo is pptp usable inside a guest ? 1285761873 Q * ntrs_ Ping timeout: 480 seconds 1285762111 J * petzsch ~markus@p4FF45EBF.dip.t-dialin.net 1285762809 M * Alteisen daniel_hozac, _are_: i have switched back to the original debian squeeze package, and reconfigured my kernels .config; CONFIG_VSERVER_LEGACY_MEM=y and enabling some CONFIG_CGROUPS options helped me 1285762895 M * _are_ I never worked with cgroups myself, but might be I am forced to in near future 1285763026 M * _are_ io bandwidth is more of an issue here 1285763716 J * manana ~mayday090@84.17.25.149 1285764487 M * Alteisen not here. my server doesnt know anything about performance or bandwidth 1285764549 M * Alteisen http://pcengines.ch/ - i run such a system for linux-vserver... 1285764561 J * Fbi-bot-83 ~Love@188.247.88.35 1285765444 Q * manana Remote host closed the connection 1285765920 N * Bertl_zZ Bertl 1285765925 M * Bertl morning folks! 1285766210 M * fback hello Bertl :) 1285766247 J * BenG ~bengreen@cpc6-aztw22-2-0-cust100.aztw.cable.virginmedia.com 1285766293 Q * BenG 1285766478 P * is_null 1285766842 M * ncopa hi 1285766857 M * ncopa are there any plans for 2.6.32.23 kernel? 1285768687 M * Bertl any reason for picking that banch? 1285768804 M * Bertl *branch 1285768883 Q * Fbi-bot-83 Ping timeout: 480 seconds 1285769245 M * fback daniel_hozac: did you have time to take a look at my issue? 1285769259 Q * petzsch Quit: Leaving. 1285769396 M * Bertl fback: what was the issue again? 1285769484 M * fback Bertl: dpkg-buildpackage made empty .debs 1285769520 M * fback structure under debian/tmp is perfectly well 1285769743 M * Bertl interesting ... 1285769902 M * fback Bertl: http://people.fback.net/fb/build.log I can't find anything wrong with it 1285770038 M * Bertl and all .deb are empty? 1285770066 M * fback yup 1285770090 M * Bertl could it be that you hit some disk limit on ../ ? 1285770131 M * fback http://paste.linux-vserver.org/18011 1285770157 M * Bertl and that's 'empty'? 1285770224 M * Bertl (didn't expect empty debian packages to be 90k large, sounds more like word documents to me ... :) 1285770224 M * fback http://paste.linux-vserver.org/18012 1285770363 M * Bertl to be honest, sounds more like a debian specific issue to me 1285770418 M * Bertl I don't know how dpkg-deb gets the list of files to package, maybe there is a debug option for that somewhere? 1285770442 M * fback and http://paste.linux-vserver.org/18013 1285770445 M * Bertl (on rpm based systems, you usually get a list of unpackaged files and a warning) 1285770459 M * Alteisen Debian... :-/ 1285770475 Q * arekm Ping timeout: 480 seconds 1285770515 M * Alteisen i am wondering how many people are switching away from debian to Arch Linux or other "simpler" distributions... 1285770522 J * ntrs ~ntrs@77.29.85.164 1285770535 M * fback Alteisen: not me ;) 1285770537 M * Bertl fback: I guess asking a debian person (best one intimate with package building) might help here 1285770555 J * petzsch ~markus@p4FF45EBF.dip.t-dialin.net 1285770567 M * Bertl at least I do not see any error or warning related to this on the log ... 1285770928 M * daniel_hozac fback: there is no indication that it wasn't run in the log. 1285770951 M * fback Bertl: that was also my plan, but all debian devs I know currently are too busy ;) 1285770971 M * Bertl micah: ping? got a minute for fback? 1285770999 M * daniel_hozac fback: does running contrib/make-deb-manifest util-vserver debian contrib/manifest.dat files work? 1285771218 M * fback daniel_hozac: http://paste.linux-vserver.org/18014 so seems it does 1285771282 M * daniel_hozac was that before or after you ran it manually? 1285771342 M * fback daniel_hozac: it's hard to tell now :) 1285771362 M * fback daniel_hozac: I'll test it a bit later, now have to pick kids from school 1285771732 J * dowdle ~dowdle@scott.coe.montana.edu 1285772180 M * mnemoc daniel_hozac: hi, did you have time to try to reproduce there my degradation with the mounts? 1285772199 M * daniel_hozac no, not yet. but i think i know what's causing it... 1285772239 M * mnemoc i thought i could work around it with umotun -l and perfect backward order, but it's still killing my server :< 1285772244 M * mnemoc daniel_hozac: that's good :) 1285772286 M * daniel_hozac umount -l is recursirve 1285772342 M * mnemoc .oO 1285772382 M * daniel_hozac as for why that isn't helping... it may be that it is still part of the other namespace. 1285772545 M * daniel_hozac you can check with vspace --enter --index 0 --fs --mount -- cat /proc/mounts 1285772560 M * daniel_hozac and/or vspace --enter --index 1 --fs --mount -- cat /proc/mounts 1285772737 M * mnemoc gosh 1285772742 M * mnemoc # vspace --enter builder-urr --index 0 --fs --mount -- cat /proc/mounts | wc -l 1285772742 M * mnemoc 65692 1285772755 M * mnemoc # vspace --enter builder-urr --index 1 --fs --mount -- cat /proc/mounts | wc -l 1285772755 M * mnemoc 10 1285772770 M * daniel_hozac yeah 1285772780 M * daniel_hozac that's basically what i anticipated... 1285772798 M * daniel_hozac i guess mount propagation isn't really ready for prime time yet :-) 1285772805 M * mnemoc :) 1285772811 M * mnemoc seems so :) 1285772904 M * mnemoc anything I can do about it? 1285772920 M * daniel_hozac well 1285772931 M * daniel_hozac the hack way to get around it is to comment that in /usr/lib*/util-vserver/vserver.functions 1285772957 M * daniel_hozac in mountVserver, the -o shared,rec 1285773022 M * mnemoc ! $_VSERVER_INFO - FEATURE PIVOT_ROOT || \ 1285773022 M * mnemoc $_SECURE_MOUNT -n -o shared,rec "$real_vdir" "$real_vdir" 1285773027 M * mnemoc ^--- those out? 1285773055 M * daniel_hozac right 1285773074 M * mnemoc done, i'll restart the vserver now 1285773176 M * daniel_hozac that should fix it. 1285773181 M * daniel_hozac or, well, work around it. 1285773196 M * mnemoc thanks a lot :) 1285773221 M * mnemoc who's the culprit? vanilla kernel, coreutils or the vserver patch? 1285773262 M * daniel_hozac this is a vanilla feature. 1285773268 M * daniel_hozac been around since 2.6.15. 1285773293 M * mnemoc so I should report the problem to lkml, right? 1285773294 M * daniel_hozac i'll need to dig a little deeper to see why the umount isn't propagating like it should. 1285773304 M * mnemoc ah, cool 1285773311 J * ntrs_ ~ntrs@77.29.85.164 1285773318 M * daniel_hozac could be that pivot_root resets that or something. 1285773345 Q * ntrs Read error: Connection reset by peer 1285773429 M * daniel_hozac it's kind of strange that mounts do propagate, but the umount doesn't... 1285773541 J * ntrs__ ~ntrs@nimitz.rosehosting.com 1285773972 Q * ntrs_ Ping timeout: 480 seconds 1285774120 M * mnemoc something went wrong... it seems i lost /dev... vlogin: openpty(): No such file or directory 1285774162 M * mnemoc (my guest's /dev is a tmpfs so i can keep /vservers/* nodev) 1285774439 J * arekm arekm@carme.pld-linux.org 1285774451 J * ntrs_ ~ntrs@77.29.85.164 1285774523 M * mnemoc daniel_hozac: is it normal to have all that duplication in 'index 0' proc/mounts? http://nopaste.opensde.net/?N2RlMD 1285774534 M * mnemoc the guest is fresh 1285774760 M * daniel_hozac yeah 1285774762 M * daniel_hozac that's normal. 1285774803 J * ntrs ~ntrs@77.28.168.188 1285774804 Q * ntrs_ Read error: Connection reset by peer 1285774902 Q * ntrs__ Ping timeout: 480 seconds 1285775288 J * BenG ~bengreen@cpc6-aztw22-2-0-cust100.aztw.cable.virginmedia.com 1285775377 M * mnemoc daniel_hozac: i had to add the shared,rec thing again :( 1285775407 M * daniel_hozac why? 1285775439 M * daniel_hozac that should only be required for vnamespace to work. 1285775439 M * mnemoc i was not able to ssh or vserver enter 1285775462 M * daniel_hozac how do you setup your /dev? 1285775480 Q * barismetin Remote host closed the connection 1285775504 M * mnemoc I get a /dev tmpfs, then copy nodes and mount devpts from init 1285775519 M * daniel_hozac at what stage? 1285775554 M * daniel_hozac pre-start? 1285775556 M * daniel_hozac prepre-start? 1285775571 M * mnemoc 1m 1285775590 M * mnemoc pre-start 1285775597 M * mnemoc should i move to prepre? 1285775638 M * daniel_hozac pre-start should be good... 1285775646 M * daniel_hozac but you can try switching it, i suppose. 1285775654 M * daniel_hozac prepre is before fstab is mounted though, JFYI 1285775687 M * mnemoc uhm :( 1285775717 M * mnemoc i'm using if grep -q '[ ]/dev[ ]' "$__CONFDIR/$2/fstab"; then to see if guest needs me to populate /dev 1285775719 M * daniel_hozac do you get any of the other mounts? 1285775727 M * mnemoc how can i know? 1285775731 M * daniel_hozac in your guest. 1285775735 M * daniel_hozac cat /proc/mounts 1285775741 M * mnemoc i can't enter the guest 1285775742 M * daniel_hozac from e.g. vserver ... exec 1285775749 M * mnemoc ok, i'll try exec 1285775874 Q * BenG Quit: I Leave 1285775897 M * mnemoc daniel_hozac: everything except dev/pts is there... 1285775930 M * daniel_hozac odd. 1285775962 M * mnemoc actually i can't find where i mount that one :p 1285775965 M * mnemoc 1m 1285776149 M * fback daniel_hozac: seems manifest didn't work 1285776257 M * mnemoc daniel_hozac: dev/pts goes in post-start.... no idea why 1285776280 M * mnemoc daniel_hozac: i'll merge both 1285776288 M * daniel_hozac you have to do that in pre-start 1285776353 M * mnemoc $_SECURE_MOUNT or mount? 1285776407 M * daniel_hozac doesn't really matter. 1285776412 M * fback with manual-generated .install files .debs build properly 1285776412 M * mnemoc ok 1285776416 M * mnemoc ok 1285776418 M * mnemoc err 1285776494 M * daniel_hozac fback didn't work how? 1285776511 M * mnemoc daniel_hozac: \o/ thank you! 1285776539 M * daniel_hozac mnemoc: does it also fix your issue? :) 1285776552 M * mnemoc don't know yet :p 1285776557 M * mnemoc tomorrow i can tell you 1285776570 M * mnemoc needs to get some smoke out of this machine first 1285776581 J * ncopa_ ~ncopa@3.203.202.84.customer.cdi.no 1285776589 M * mnemoc takes some hours to get degraded 1285776603 M * daniel_hozac okay 1285776605 M * daniel_hozac keep us posted. 1285776622 M * mnemoc i will 1285776656 M * fback daniel_hozac: *.install files were not created 1285776757 M * fback daniel_hozac: but because I didn't remove them by hand before another build, .debs are ok now 1285776811 M * daniel_hozac so most likely the override for install never ran the first time... 1285776880 M * daniel_hozac what distribution are you on? 1285776906 M * daniel_hozac and how did you build it, exactly? 1285776911 M * fback lenny 1285776946 M * fback with dpkg-buildpackage -us -uc command 1285777036 Q * ncopa Ping timeout: 480 seconds 1285778510 Q * bzed Remote host closed the connection 1285778517 J * bzed ~bzed@devel.recluse.de 1285778531 J * emcepe ~mcp@wolk-project.de 1285778824 Q * mcp Ping timeout: 480 seconds 1285778824 N * emcepe mcp 1285779475 J * ntrs_ ~ntrs@77.29.85.164 1285779475 Q * ntrs Read error: Connection reset by peer 1285779706 Q * derjohn_mob Ping timeout: 480 seconds 1285779857 Q * balbir_ Ping timeout: 480 seconds 1285780066 Q * ncopa_ Ping timeout: 480 seconds 1285780077 J * ncopa_ ~ncopa@ti211310a081-0128.bb.online.no 1285780318 J * Sav ~Fear@95.142.142.126 1285780323 P * Sav 1285781703 M * fLoo re 1285781720 M * fLoo back to my question, is it possible to run a vpn (pptp) in a guest ? 1285781737 M * fLoo like i have 10 guests with 10 vpn-connections at the same time ? 1285781805 M * Bertl depends on the setup 1285781821 M * fLoo tell me more :) 1285781832 M * fLoo depends on what exactly 1285781841 J * AndrewLee ~andrew@210.240.39.7 1285781841 M * Bertl you can (without network namespaces) assign a tun device to a guest, and use that to inject packets into the host networking (ip layer) 1285781871 M * Bertl this is sufficient for e.g. openvpn to work properly, so it should do for ppp as well 1285781887 M * fLoo ok 1285781891 M * fLoo thats the way i thought of 1285781924 M * fLoo but this way does not allow me to run 10 connections at a time 1285781928 M * Bertl the problem is, when you try to go below the IP layer, the IP isolation will not work and access will be denied 1285781942 M * fLoo yep 1285781959 M * Bertl pptp uses TCP port 1723 by default 1285782001 M * Bertl so, as long as you provide enough IPs to connect to, that should be fine 1285782019 M * fLoo well 1285782038 M * fLoo i got one public ip and internal ips 1285782049 M * fLoo and i want to route the internal ips over vpn into the world 1285782061 M * fLoo so they are reachable by their own (more or less) dedicated ip 1285782066 M * fLoo thats what my plan is 1285782099 M * fLoo so i thought of running 10 vpn instances (for each guest one) to assign the vpn-ip to them 1285782103 M * fLoo and route the traffic over vpn 1285782114 M * Bertl sounds a little contradictional 1285782142 M * Bertl first, how do you plan to provide dedicated IPs with a single public one? 1285782169 M * Bertl (or does that just refer to prive IPs seen on the vpn end?) 1285782174 M * Bertl *private 1285782197 M * fLoo nope i got 1 single ip from my isp (at home) 1285782200 M * Bertl second, what's the point in using PPTP for that? 1285782216 M * fLoo i would like to test this with a free vpn provider first ;) 1285782218 Q * ntrs_ Read error: No route to host 1285782226 M * fLoo and they only offer pptp 1285782247 J * ntrs_ ~ntrs@77.28.168.188 1285782258 M * fLoo they provide me with tunnel-ips 1285782265 A * Bertl googles vpn provider ... 1285782270 M * fLoo proxpn ;) 1285782402 M * Bertl okay, obviously I'm missing something here 1285782461 M * Bertl looking at proxpn, I see a bunch of claims which are just wrong IMHO 1285782493 M * fLoo ya 1285782499 M * fLoo some facts or false on their website 1285782504 M * fLoo *are 1285782525 M * Bertl but basically they provide a (somewhat) dynamic public IP for you as gateway to the world, yes? 1285782620 M * Bertl the idea of adding security by encrypting only a part of the connection (from you to the public gateway) is amusing :) 1285782652 M * fLoo yep 1285782658 M * fLoo :D 1285782663 M * Bertl (note that this service seems to require special software to work) 1285782668 M * fLoo but i just wanted to test the scenario i had in my mind 1285782674 M * fLoo only for windows users 1285782674 M * fLoo ;) 1285782706 M * Bertl okay, so, taking one step back, what is the purpose of the 'final' setup? 1285782726 M * Bertl (IMHO no point in testing setup A, when you actually want/need setup B) 1285782751 M * Bertl you have a bunch of guests, with private IPs, yes? 1285782780 M * Bertl and you have a host with 'only' one public IP to connect to the internet 1285782801 M * Bertl am I correct so far? 1285783639 M * fLoo mmh 1285783643 M * fLoo after reading a bit 1285783657 M * fLoo yes 1285783657 M * fLoo :) 1285783680 M * fLoo i got 1 public ip - 10 internal (192.168.X.X) 1285783686 M * fLoo correct so far 1285783720 M * Bertl okay, so now the idea is to get a different 'provider' which allows you to get 10 public IPs, which in turn will be tunneled to your host (to access the private IPs of those guests), right? 1285784182 J * ntrs ~ntrs@77.28.164.137 1285784294 M * fLoo yes 1285784297 M * fLoo you're insane :D 1285784303 M * fLoo thats exactly what i wanna do 1285784390 M * Bertl okay, so besides the fact that all traffic will go through the vpn provider (potential security issue), why not just 'request' the 10 IPs from the hoster? 1285784572 Q * ntrs_ Ping timeout: 480 seconds 1285784607 M * fLoo well 1285784612 M * fLoo the node is at my home 1285784616 M * fLoo and i only get 1 ip :D 1285784676 M * Bertl okay ... probably not the best place for hosting 10+ IPs :) 1285784709 M * Bertl anyway, assuming that you get a good priced service for the tunneling/vpn, I would put all the vpn stuff on the host 1285784732 M * fLoo ok 1285784732 M * Bertl mapping the vpn addresses (via iptables) to the private guest IPs 1285784775 M * fLoo well this is not about good or bad, i just want to see if this and if - how it might be possible 1285784775 M * Bertl no need to put any vpn stuff on a guest, and more important, no need to have several vpn connections either (if your vpn provider manages this) 1285784785 M * fLoo ok 1285784864 M * fLoo but i dont know any vpn provider who offers more then 1 ip / customer 1285785754 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1285786404 Q * petzsch Quit: Leaving. 1285786564 J * imcsk8 ~ichavero@201.174.32.227 1285786956 M * fLoo mmh 1285786956 M * fLoo hun 1285786958 M * fLoo u r off ? :P 1285786964 M * fLoo or just had a nap-attack ? :P 1285786980 M * Bertl np :) 1285787013 M * Bertl so, connecting to several of them might as well be possible 1285787054 M * Bertl but I doubt that this will become a stable solution ... 1285787085 M * Bertl did you consider using a single (maybe your) public IP and different dns entries? 1285787092 M * fback wouldn't it be better to find cheap (virtual) hosting, where you can buy more public IPs? 1285787099 M * fback s/better/easier/ 1285787112 M * Bertl yeah, probably way better and easier 1285787135 M * Bertl but there might be a reason for hosting them at home 1285787177 M * fback lack of serious infrastructure? ;-) 1285787332 M * Bertl well, let's see if I can come up with a plausible explanation ... hmm, fLoo is agoraphobic and paranoid and thus needs to keep the host @ home for security reasons :) 1285787533 A * ard has enough with 1 ip 1285787542 M * ard for the usual webserving that is 1285787551 M * fback fLoo: but seriously, did you try to write to you provider and ask him for more public IPs? 1285787555 M * ard I've got 2 at home 1285787577 M * fback I have /64 at home :> 1285787579 M * ard 64 at my colorack 1285787608 M * ard ow, and I've got 2 /64's and one /60 at home and a /64 at my colo rack :-) 1285787617 M * ard and at my work I got a /32 in ipv6 space :-) 1285787633 M * Bertl I presume it is a xDSL line for personal use, where strictly speaking it isn't even allowed to host a single server 1285787635 M * fLoo ok re 1285787636 M * fLoo :) 1285787662 M * fLoo lol 1285787675 M * fLoo i am NOT agoraphobic, nor i am paranoid 1285787676 A * ard had that discussion with an cable isp once 1285787682 M * fLoo i just want to get some more experience 1285787688 M * fLoo why are we all here ? 1285787692 M * fLoo isnt it all about learning ? :P 1285787696 M * fLoo yessir, it is 1285787708 M * ard I was not allowed to run an X-server, but they would only act if I use too much traffic 1285787715 M * fLoo Bertl : i have cable over here and its allowed to serve content from home 1285787732 M * ard Heh, it's changed here too 1285787738 M * fLoo furthermore i produce around 1.5 TB traffic each month 1285787741 M * fLoo no problems so far 1285787760 M * fLoo and its just 32 mbit .. 1285787769 M * fLoo in 2 months i am able to upgrade to 100 mbit here :) 1285787787 M * ard up/down / symmetric? 1285787791 M * fLoo 32/2 1285787792 M * fLoo 100/10 1285787820 M * fLoo down/up obviously ;) 1285787823 M * ard ah :-) 1285787832 Q * imcsk8 Remote host closed the connection 1285787834 A * ard was already typing the obvious :-) 1285787838 M * fLoo hehe+ 1285787868 M * ard the only 100 I can get is a fiber, which would cost me 230Euro/month ex. VAT, and ex. internet service 1285787886 M * fLoo ouch 1285787890 M * ard the internet will be another 100 euro/month at least for a 10 Mb/s 95 percentile 1285787893 M * fLoo i'll pay 39,90 1285787894 M * fLoo euro 1285787899 M * fLoo including phone flat 1285787906 M * ard So I have 2 adsl lines :-) 1285787912 M * fLoo its fine too ;) 1285787918 M * ard that can barely hold 10 down and 1 up 1285787923 M * fLoo 2 adsl lines load-balanced ? 1285787941 M * ard no, 1 dedicated for voip (since it is stable) 1285787955 M * ard and 1 for all the other things, like mail and stuff 1285787966 M * ard I only host websites at my colo-space 1285787968 M * ard and a nameserver :-) 1285787973 M * fLoo :) 1285787975 M * fLoo so do i 1285787992 M * fLoo well i use a rootserver i've been given for development 1285788003 M * fLoo but i am free to host my sites on it 1285788007 A * ard has 1 24U rack 1285788011 M * fLoo yay! 1285788012 M * ard half rack actually 1285788020 A * fLoo hands over some cookies 1285788021 M * ard but only 4A of power 1285788022 M * fLoo :) 1285788026 M * fLoo got some spare space left ? 1285788036 M * ard space enough, not enough A 1285788053 M * ard but at my work things are better 1285788061 M * ard 34Gb/s interwebs :-) 1285788085 M * fLoo ouch 1285788185 M * ard actually we only have a 10Gb/s ring :-( 1285788201 M * ard so we can get at most 20Gb/s to our datacenter :-( 1285788216 M * mnemoc shame on you 1285788219 M * ard ow wait, no, we are multiplexing over dark fibre right now 1285788226 M * ard so we have 30Gb/s now :-) 1285788238 M * fLoo well 1285788242 M * fLoo ard where is your dc located ? 1285788286 M * ard heh, 15 meters downwards, and about 100meters east from here :-) 1285788301 M * fLoo i mean geolocated 1285788302 M * ard that's my works dc 1285788305 M * ard amsterdam 1285788309 M * fLoo cool 1285788321 M * fLoo may u perform a speedtest for me pls (off one of my vservers located in germany) ? 1285788323 M * fLoo its on gbit 1285788336 M * ard allright, but I am behind a shaped firewall 1285788340 M * ard doing 200MB/s max 1285788343 M * fLoo http://188.64.45.134/100mb.bin 1285788345 M * fLoo its fine 1285788362 M * ard 2010-09-29 21:25:59 (17.3 MB/s) - `/dev/null' saved [104857600/104857600] 1285788364 M * fLoo i am testing this box and promoting their services a bit 1285788369 M * fLoo mmh alright 1285788371 M * fLoo sounds good 1285788380 M * ard that's the shaping talking 1285788402 M * fLoo -..- 1285788411 M * ard 2010-09-29 21:26:48 (64.7 MB/s) - `/dev/null' saved [104857600/104857600] 1285788416 M * ard from unshaped :-) 1285788418 M * fLoo :D 1285788458 M * fLoo lemme get a gb file on that box 1285788460 M * fLoo for better check 1285788463 J * imcsk8 ~ichavero@201.174.32.227 1285788494 J * petzsch ~markus@p4FF45EBF.dip.t-dialin.net 1285788519 M * ard you know you can use dd to make a sparse file? 1285788538 M * ard seek=1gb count=1 bs=1 1285788542 M * ard something like that 1285788551 M * fLoo already done :D 1285788565 M * ard Length: 638124032 (609M) [application/octet-stream] 1285788569 M * ard 2010-09-29 21:29:27 (63.5 MB/s) - `/dev/null' saved [638124032/638124032] 1285788573 M * fLoo http://188.64.45.134/1000mb.bin 1285788576 M * fLoo now 1285788579 M * fLoo takes some time lol 1285788583 M * fLoo dont be mean -.- 1285788601 M * ard it's slower now 1285788606 M * fLoo well 1285788610 M * ard 2010-09-29 21:30:02 (61.7 MB/s) - `/dev/null' saved [1048576000/1048576000] 1285788614 M * fLoo kk 1285788620 M * ard there is some kernel network overhead btw 1285788637 M * ard this is: 2.6.31.6 1285788638 M * fLoo best result i got was 97 mb/s to .nl 1285788641 M * ard and that's not the fastest 1285788646 M * fLoo fact 1285788669 M * mnemoc 137KB/s from my home in spain 1285788676 M * ard need at least 2.6.34 to get at least a 10% increase 1285788683 M * fLoo ya 1285788689 M * fLoo but is it that much ? 1285788690 M * fLoo 10 % ? 1285788699 M * ard o wait: these are bnx2 cards. They are slow by design 1285788705 M * ard maybe more 1285788733 M * ard nevermind :-) 1285788736 M * ard brain:~# cat /proc/cpuinfo 1285788738 M * ard model name : Intel(R) Pentium(R) III CPU family 1133MHz 1285788743 M * ard WOW it's so fast :-) 1285788758 M * ard But I've done a test from 2.4.30 -> 2.6.35 1285788766 M * ard the speed increase was 1000% 1285788794 M * ard because we were being DDOS'd until the switch to 2.6.35, then we could handle te traffic :-) 1285788795 M * fLoo lol 1285788803 M * fLoo :D 1285788833 M * fLoo processor : 7 1285788833 M * fLoo model name : Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz 1285788835 M * fLoo ftw ;) 1285788842 M * fLoo (my .dev box) 1285788860 M * ard anyway: really got to go.. (BTW from 77kp/s -> server unreacable to 660kp/s server being idle...) 1285788875 M * fLoo model name : Intel(R) Celeron(R) CPU 2.40GHz 1285788877 M * fLoo (home) 1285788911 M * ard (660k SYN's /s that is) 1285789037 M * fLoo mmh 1285789038 M * fLoo :D 1285789046 M * fLoo i'm bored 1285789053 M * fLoo gonna rape the lkml :D 1285789463 J * derjohn_mob ~aj@d065143.adsl.hansenet.de 1285789647 M * _are_ Alteisen: we use this very Board (Alix 2d3) for redundant firewall/vpn routers based on Voyage linux (Debian enhancement for systems on flash memory), OpenVPN and heartbeat 1285789673 M * _are_ I would not have come to the idea to run VServers on them 1285789874 J * ntrs_ ~ntrs@77.28.7.8 1285789875 Q * ntrs Read error: Connection reset by peer 1285790004 M * Bertl _are_: why not? 1285790143 M * fLoo i would run vserver on my girlfriend i that would be possible 1285790147 M * fLoo :) 1285790152 M * fLoo *if 1285790209 M * Bertl girl/boyfriends usually utilize wetware which doesn't support Linux (yet) 1285790312 J * petzsch1 ~markus@p4FF46685.dip.t-dialin.net 1285790318 M * Alteisen _are_: i was looking for the smallest, low power but complete system i could get for debian, and that board was it - quite cheap for all those features it has 1285790480 M * Alteisen _are_: and running a server in a virtualized environment is a question of principle as well 1285790551 M * Alteisen actually, when i started with debian on that little board, i have stolen many settings from voyage (which is debian), but i wanted an original debian system 1285790551 M * _are_ Bertl: for my part: lack of RAM mostly, we use if for firewalling and OpenVPN, both things that require adding quite a few bcapabilities. So in the end there is no enefit in stuffing it into a VServer 1285790569 Q * petzsch Ping timeout: 480 seconds 1285790605 M * _are_ I still look for something similar sized with similar power requirements with GBit ethernet 1285790677 A * Alteisen too - and with at least 2 sata ports 1285790721 M * _are_ the typical VServer system I set up is more like 2*Quadcore CPU, 24-72GB RAM in company with DRBD and Heartbeat for failover servers. They run anything: file, mail, web, db, backupsytems 1285790768 M * _are_ and if VServer is not 'virtualization' enough i throw in a KVM 1285790849 M * Alteisen we have a linux-vserver host with 3 vserver guests running as guest itself on a ESX Host 1285790952 M * _are_ well, we tried that, 1 setup with ESX + FC-SAN, 1 with XEN and iSCSI SAN. I think it is a bad choice in the long run. you have a VServer that is just native performance with disk IO and stuff it into ESX/XEN that is just a nightmare with disk IO. 1285791018 M * _are_ the ESX setup was up and running for 3 years, 2 weeks ago we removed all linux hosts there and stuffed them onto a seperate DRBD/Heartbeat/VServer cluster, way, way better performance 1285791086 M * _are_ the Xen cluster will follow next month for all linux servers that have disk io 1285791114 M * Alteisen performance doesn't matter in that case 1285791144 M * Alteisen i just like the way i can rsync a vserver within seconds to another location 1285791190 M * Bertl how's that related to ESX? 1285791198 M * _are_ well, here it does, IMAP server, 100 users, 350GB mail, samba servers with up to 3TB disk requirements, postgres/mysql servers and similar. 1285791210 M * Alteisen and setting up the host is a scripted setup ( preseed.cfg with postinst.sh ) 1285791237 Q * ntrs_ Read error: Connection reset by peer 1285791247 J * ntrs ~ntrs@77.28.164.137 1285791257 M * Alteisen all i do is booting via pxe, selecting "myserver" menu entry, wait for 20 minutes and then rsync back my /vservers and /etc/vservers files 1285791379 Q * fback Ping timeout: 480 seconds 1285791380 M * _are_ Bertl: just in the way that VServer basically runs perfect on a ESX/XEN/KVM guest. 1285791475 M * _are_ I just personally consider it a waste, it runs way better on plain hardware and if you really need somthing with a seperate kernel you can run KVM guests next to it 1285791752 M * _are_ for the files: yes, I run VServers in VM-whatever guests myself and I find the exitence of a debian package 'linux-image-2.6-xen-vserver' facilitates the discussion very much when you run into one who already rusn a policy of 'our virtualization solution is VMware/XEN' as soon as they see the V in VServer 1285791982 J * fback fback@red.fback.net 1285792499 M * harry Bertl: will you update the 2.6.32 patch to 2.6.32.23? 1285792694 M * fLoo and 2.6.35.5 to .35.7 :) 1285792765 Q * petzsch1 Quit: Leaving. 1285792957 M * arekm there were no big changes 1285793004 M * harry no , but there are "errors" when applying patches to 2.6.32.23 1285793151 Q * ntrs Ping timeout: 480 seconds 1285793153 M * arekm easily fixable I guess 1285793490 M * Bertl harry, fLoo: very likely 1285793517 J * ntrs ~ntrs@77.28.7.8 1285793986 Q * bonbons Quit: Leaving 1285794007 M * fLoo thanks Bertl :) 1285794036 J * petzsch ~markus@p4FF46685.dip.t-dialin.net 1285794073 J * ntrs_ ~ntrs@77.28.7.8 1285794120 Q * ntrs Read error: Connection reset by peer 1285794175 Q * ghislain Quit: Leaving. 1285794598 Q * ntrs_ Ping timeout: 480 seconds 1285794939 J * ntrs ~ntrs@77.28.7.8 1285795392 J * ntrs_ ~ntrs@nimitz.rosehosting.com 1285795652 J * manana ~mayday090@84.17.25.149 1285795835 Q * ntrs Ping timeout: 480 seconds 1285796052 J * ntrs ~ntrs@nimitz.rosehosting.com 1285796052 Q * ntrs_ Read error: Connection reset by peer 1285796355 J * ntrs_ ~ntrs@77.28.7.8 1285796795 Q * ntrs Ping timeout: 480 seconds 1285796840 Q * ntrs_ Ping timeout: 480 seconds 1285796856 J * ghislain ~AQUEOS@adsl2.aqueos.com 1285797341 Q * tokkee Ping timeout: 480 seconds 1285797662 M * Bertl nap attack .. bbl 1285797666 N * Bertl Bertl_zZ 1285797677 Q * ghislain Quit: Leaving. 1285799429 Q * petzsch Quit: Leaving. 1285801361 J * tokkee tokkee@osprey.tokkee.org 1285802347 M * Chlorek where can I find full cgroups documentation? 1285802924 M * fLoo well 1285802926 M * fLoo google :D 1285802950 M * fLoo what u need may be found here -> http://linux-vserver.org/util-vserver:Cgroups 1285802963 M * fLoo the rest can be found here -> http://www.mjmwired.net/kernel/Documentation/cgroups.txt 1285803883 Q * imcsk8 Quit: Leaving