1178152565 M * Bertl okay, uploaded the first rc for 2.6.21 1178152580 M * Bertl will need some testing, especially in the following areas: 1178152591 M * Bertl - hard cpu limits with dynaticks 1178152605 M * Bertl - sys_sysinfo() virtualization 1178152626 M * Bertl I'm off to bed now ... have fun! 1178152632 N * Bertl Bertl_zZ 1178152707 Q * bzed Quit: Leaving 1178152946 J * neuralis ~krstic@solarsail.hcs.harvard.edu 1178156071 Q * cehteh Ping timeout: 480 seconds 1178156412 Q * eDog Remote host closed the connection 1178156949 J * eDog ~edog@office.aichyna.com 1178158841 Q * besonen_ Read error: Connection reset by peer 1178158846 J * besonen ~besonen@dsl-db.pacinfo.com 1178158894 J * tudenbart ~willi@xdsl-213-196-251-42.netcologne.de 1178159322 Q * dothebart Ping timeout: 480 seconds 1178159757 Q * ensc Ping timeout: 480 seconds 1178159844 J * cehteh ~ct@pipapo.org 1178160831 Q * besonen Ping timeout: 480 seconds 1178161676 J * ktwilight ~ktwilight@216.109-66-87.adsl-dyn.isp.belgacom.be 1178162088 Q * ktwilight_ Ping timeout: 480 seconds 1178162394 J * besonen ~besonen@dsl-db.pacinfo.com 1178163199 J * Johnnie ~jdlewis@c-67-163-247-109.hsd1.pa.comcast.net 1178163847 J * ensc ~irc-ensc@p54B4FE0A.dip.t-dialin.net 1178164477 J * mugslave ~samv@leibniz.catalyst.net.nz 1178164479 Q * mugslave 1178169049 J * sharkjaw ~gab@216-158-6.0503.adsl.tele2.no 1178170117 Q * DavidS Quit: Leaving. 1178171363 Q * tamitall Ping timeout: 480 seconds 1178172116 J * dna ~naucki@96-204-dsl.kielnet.net 1178172847 Q * eDog Remote host closed the connection 1178173044 Q * cdrx Ping timeout: 480 seconds 1178174382 N * DoberMann[ZZZzzz] DoberMann[PullA] 1178176565 J * meandtheshell ~markus@85-124-174-45.dynamic.xdsl-line.inode.at 1178177728 Q * bavi 1178178539 J * cdrx ~legoater@blueice4n2.uk.ibm.com 1178179179 Q * DoberMann[PullA] Ping timeout: 480 seconds 1178179729 J * suser ~tux@bzq-82-81-19-143.red.bezeqint.net 1178179734 M * suser hi all ! 1178179751 M * suser How this vserver is different from openvz? 1178179770 M * suser any article/reviews on that topic? 1178179836 J * DoberMann[PullA] ~james@AToulouse-156-1-131-216.w90-30.abo.wanadoo.fr 1178179931 M * harry suser: don't know any articles or reviews regarding that 1178180148 J * bavi foobear@193.30.161.200 1178180153 M * bavi <3 Bertl_zZ 1178180291 J * DavidS david@chello062178045213.16.11.tuwien.teleweb.at 1178181266 Q * sharkjaw Quit: Leaving 1178182111 Q * glut synthon.oftc.net oxygen.oftc.net 1178182111 Q * bXi synthon.oftc.net oxygen.oftc.net 1178182111 Q * eyck synthon.oftc.net oxygen.oftc.net 1178182111 Q * djbclark synthon.oftc.net oxygen.oftc.net 1178182111 Q * svenk synthon.oftc.net oxygen.oftc.net 1178182111 Q * neuralis synthon.oftc.net oxygen.oftc.net 1178182111 Q * nebuchadnezzar synthon.oftc.net oxygen.oftc.net 1178182111 Q * Bertl_zZ synthon.oftc.net oxygen.oftc.net 1178182111 Q * harry synthon.oftc.net oxygen.oftc.net 1178182140 M * suser harry>: u still here? 1178182167 M * daniel_hozac no, he netsplitted. 1178182174 M * suser ok 1178182183 J * neuralis ~krstic@solarsail.hcs.harvard.edu 1178182183 J * svenk ~sven@pulsar.digital.udk-berlin.de 1178182183 J * Bertl_zZ herbert@IRC.13thfloor.at 1178182183 J * eyck ~eyck@nat.nowanet.pl 1178182183 J * bXi bluepunk@irssi.co.uk 1178182183 J * djbclark dclark@opensysadmin.com 1178182183 J * nebuchadnezzar ~nebu@zion.asgardr.info 1178182183 J * harry ~harry@d54C2508C.access.telenet.be 1178182183 J * glut glut@no.suid.pl 1178182191 M * suser anyway: how is this different from openvz generally? 1178182264 M * daniel_hozac OpenVZ does virtualization in some places where we do isolation, which increases the OpenVZ overhead. 1178182296 M * suser daniel_hozac>: how is the memory model isolated ? 1178182318 M * suser that is: if some program attacks kernel memory space somehow, can it leak to other VPSes? 1178182362 M * daniel_hozac well, that's true for all OS-level virtualization technologies. 1178182420 M * suser or how is the networking isolation works ? that is: iptables (userland) use stnadard kernel system calls but how it's all isolated? 1178182436 M * suser Do you replcae netfilter with vserver netfilter to foll iptables? 1178182441 M * daniel_hozac no. 1178182443 M * suser "fool" i meant 1178182453 M * daniel_hozac iptables doesn't work inside a guest. 1178182471 M * suser ahhh.. I think in OpenVZ it works 1178182480 M * daniel_hozac yes, because they virtualize the network stack. 1178182495 M * daniel_hozac while we do IP address isolation. 1178182510 M * suser daniel_hozac>: and how about routing tables? isolated? 1178182521 M * daniel_hozac they too are shared. 1178182538 M * suser that is not very good IMHO 1178182545 M * suser from secury point of view 1178182713 M * daniel_hozac oh? why's that? 1178182713 M * suser different servers shouldn't see one other's routing table 1178182713 M * daniel_hozac would they be different? 1178182713 M * suser yes, I prefer to have different routing tables for each VPS 1178182713 M * daniel_hozac so you'd use policy routing and you'd only want the guest to have access to its own routing table? 1178182772 M * suser I'd prefer that each VPS has it's own routing table, like a real server 1178182838 M * daniel_hozac that means virtualizing the network stack, which should generate roughly a 100% increase in overhead... 1178182860 M * suser yep 1178182879 M * suser however it would be nice to have. At least optionally. 1178182890 M * daniel_hozac and once mainline has it, so will we. 1178182912 M * suser mainline has only very basic virtualization support 1178182927 M * daniel_hozac for now. 1178182933 M * suser just KVM and VMI 1178182946 M * daniel_hozac hmm 1178182957 M * daniel_hozac that's not relevant for OS-level virtualization at all. 1178182957 M * suser nothing like network stack virtualization is even planned AFAIK 1178182964 M * suser true 1178182966 M * daniel_hozac uh, it is. 1178182976 M * daniel_hozac there are even multiple patchsets implementing it. 1178182987 M * suser so I don't know of any mainline-kernel features that can help OS-level virtualization 1178183028 M * daniel_hozac the uts an IPC namespaces introduced in 2.6.19 do basically the same thing we did. 1178183033 M * daniel_hozac s/an/and/ 1178183038 M * suser btw: If I want OS-level virtualization for Windows - what are my options ? 1178183054 M * daniel_hozac i have absolutely no idea. 1178183065 M * daniel_hozac this is _Linux_-VServer ;) 1178183074 M * suser I know :) 1178183080 M * baldy suser: maybe vmware 1178183093 M * suser baldy>: he he - I know VMware 1178183105 M * suser baldy>: and it's NOT OS-level ! 1178183116 M * baldy isnt it? 1178183119 M * suser no 1178183135 M * baldy oke 1178183139 M * baldy ignore my idea ;) 1178183150 M * daniel_hozac only one i'm aware of is Virtuozzo. 1178183168 M * suser daniel_hozac>: me too. 1178183284 M * suser daniel_hozac>: thanks for your time, but for my needs network-stack virtualization is preffered, so I will look at openvz 1178183288 M * suser thanks again. 1178183384 M * daniel_hozac bye bye 1178183815 M * harry suser: i do per vps routing 1178183830 M * harry other vps'es don't know anything about any other vps 1178183838 M * harry (in linux-vserver) 1178183870 M * harry it was a bitch to setup initially, but now i just fixed it in 2 scripts... a .default pre-start and a post-stop 1178183966 M * suser harry>: ok, that's nice, but what about iptables ? can you do iptables in each VPS ? 1178183985 M * harry yes 1178183990 M * harry well... no 1178183999 M * harry but that enhances my security ;) 1178184010 M * harry i do have a "per vps" firewall 1178184018 M * suser which one? 1178184018 M * harry admin'ed on the host 1178184021 Q * phreak`` Ping timeout: 480 seconds 1178184023 M * suser ahh ok 1178184024 M * harry iptables :) 1178184037 M * harry that means: if someone should root my vps 1178184044 Q * Hollow Ping timeout: 480 seconds 1178184045 M * harry they still can't disable firewall or do crazy routing stuff 1178184056 M * harry to get their rootkits in etc.. 1178184062 M * suser I understand, you use host as a router + as a firewall for your VPSes 1178184075 M * harry yup 1178184106 M * harry at first i thought: this sucks... (like you said ;)) 1178184133 M * harry but... right now... i think differently 1178184137 M * suser well, I don't like this. I prefer to be able to do firewall in each VPS 1178184149 M * harry that's your choice :) 1178184218 M * suser harry>: that's because my target is different. Not servers at all. but network-simulation 1178184315 M * harry aha 1178184320 M * harry then you can't use vservers ;) 1178184328 M * harry i use it for different applications 1178184332 M * harry that need different servers 1178184347 M * harry but are... well... mostly crap that doesn't do anything (1 server) 1178184358 M * harry or very heavy servers which have to be easily moved 1178184413 M * suser until now I work with full-virtualizers, but I'm considering a move to partial virtualizers: Xen/openvz/vservers 1178184606 J * chand ~chand@212.99.51.254 1178184764 M * harry network testing stuff is generally not such a good idea with partial virtualizer stuff imho 1178184774 M * waldi why? 1178184788 M * harry because performance/behaviour isn't allways 100% correctly virtualized 1178184808 M * harry may respond differently on partially virtualized stuff 1178184819 M * waldi why do you expect that performance is always part of the test? 1178184822 M * harry performance/response times/... 1178184833 M * harry i don't 1178184834 M * waldi you can't predict them in ip anyway 1178184857 M * waldi and behaviour. a linux kernel responds like a linux kernel 1178184875 M * waldi this does not differ between paravirt and real 1178184886 M * suser I'm not testing performance. Just the functionality. 1178184892 M * suser correctness 1178184898 M * harry what makes you think that the patches in the virtualisation layer of openvz change NOTHING compared to a standaar linux kernel? 1178184932 M * harry no oddities/weirdnesses in the implementation of virtualisation of the ip stack 1178184940 M * harry network stack... sorry 1178184942 M * waldi err, openvz \in paravirt, not openvz == paravirt 1178184955 M * waldi xen also \in paravirt 1178184997 M * harry well... you can use what you want... if you want to test/check correctness of that kind of stuff... you should be very careful that your test situation is "realistic" 1178186742 J * bianchi ~nano@174.Red-88-16-29.dynamicIP.rima-tde.net 1178186768 M * bianchi hi 1178186932 Q * cdrx Read error: Connection reset by peer 1178187464 M * bianchi hello, can anybody help me? 1178187474 J * phreak`` ~phreak``@deimos.barfoo.org 1178187509 Q * phreak`` 1178187524 M * bianchi help 1178187544 M * bianchi sorry, this last "help" wasn't write for this chat. :-S 1178187548 J * phreak`` ~phreak``@deimos.barfoo.org 1178187602 N * bianchi bianchi_away 1178188252 J * Piet hiddenserv@tor.noreply.org 1178189301 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1178189337 Q * suser 1178189530 N * Bertl_zZ Bertl 1178189544 M * Bertl morning folks! 1178189709 M * Bertl waldi: openvz is paravirtualization now? 1178189722 M * Bertl bianchi_away: what help do you need? 1178189738 M * waldi Bertl: dunno 1178191360 M * bavi Bertl: You are my sweet sweet love 1178191365 M * bavi I think i am inlove 1178191388 M * Bertl bavi: again? :) 1178192516 Q * shedi Quit: Leaving 1178193452 J * cruser ~chatzilla@72.242.194.162 1178193804 M * Bertl welcome cruser! 1178193828 M * cruser Hi 1178193844 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1178193860 M * Bertl wb renihs! 1178193867 M * renihs anyone wants to take a look at this 1178193869 M * renihs http://rafb.net/p/TXTDTE70.html 1178193872 M * renihs hey Bertl 1178193886 M * renihs thats a 2.6.20-vs2.3.0.11-gentoo #2 SMP 1178193893 M * renihs seems to be an old pty io race bug 1178193916 M * renihs i am sitting in #gentoo-vps in freenode usually 1178193924 M * renihs completely forgot about this channel here :) 1178193931 M * Bertl try with vs2.2.0 and let me know if it is still there 1178193940 M * Bertl (2.3.x needs to be updated) 1178193968 M * renihs hmm 1178193981 M * renihs mkay 1178194001 M * renihs the problem didnt exist on linux-2.6.18-vserver-2.1.1-rc48/ 1178194026 M * renihs i will switch to vs2.2.0 1178194039 M * renihs problem is the machine was running for +18 days before that incident occured 1178194044 M * renihs hard to reproduce 1178194064 M * renihs also the last files where +200m and who -l listed more than 12.000 1178194093 M * renihs also there was alot load without any process being affiliated to it 1178194219 M * cruser Bertl: I've been reading the princeton white paper comparing xen and vserver. Now I am wondering if I should use a UP kernel instead of the SMP kernel I have installed on this dual core box. 1178194243 M * daniel_hozac why would you?! 1178194249 M * renihs can they be compared? 1178194259 M * daniel_hozac certainly you want to use both cores, no? 1178194279 M * cruser It seemed to say that the UP kernels got better performance 1178194342 M * Bertl cruser: for a single core, yes :) 1178194353 M * daniel_hozac well, SMP adds some overhead, but since you get 100% as much CPU... 1178194367 M * daniel_hozac +english +sense 1178194379 M * cruser Thank you. Some of it is over my head so I wanted to make sure on this channel. 1178194462 M * cruser How do I find the XID of a vserver? 1178194487 M * Bertl by choosing a number between 2 and 49151? 1178194514 M * cruser I wanted to use "vdu" on an existing vserver. 1178194560 M * renihs also, since you all are so active 1178194563 M * daniel_hozac cat /etc/vservers//context 1178194637 M * Bertl renihs: load can arise from kernel threads too (since recently) 1178194656 M * renihs Bertl, i noticed 1178194657 M * renihs :p 1178194696 M * renihs also, is there any news on using the normal baselayout within a vserver guest on gentoo? 1178194709 M * Bertl IIRC, that works since some while 1178194716 M * daniel_hozac baselayout-2 should work fine with util-vserver trunk. 1178194735 M * cruser daniel_hozac: Okay, there is no "context" file I wil lput one with a number between 2 and 49151. Do I need to do any processing after that? 1178194766 M * Bertl cruser: you should stop the guest first, and you want to move files to that context before restarting 1178194771 M * daniel_hozac cruser: if you're using dynamic contexts, vdu hopefully won't do anything... 1178194805 M * daniel_hozac note that vdu only counts tagged files, so it requires a filesystem mounted with tagxid, as well as, well, tagged files. 1178194841 M * cruser I have build my guests with default settings so I assume I am using dynamic contexts. 1178194873 M * Bertl cruser: dynamic context have been deprecated since several years now 1178194873 M * daniel_hozac in which case you're hopefully not using xid tagging, as that would break as soon as you restart a guest... 1178194912 M * Bertl cruser: are you using debian? 1178194921 M * cruser Oh boy. Fedora 1178194956 M * Bertl interesting ... 1178194973 M * cruser Shall I pass the command I used to build a guest and later copy it? 1178195029 M * daniel_hozac cruser: hey, you had problems booting my kernels, right? 1178195036 M * cruser yes. 1178195053 M * daniel_hozac could you try building a vanilla kernel with highmem and a 2/2 memory split? 1178195085 M * cruser Building a kernel is beyond me :( 1178195122 M * cruser You tell me what do to and I'll do it. 1178195227 T * daniel_hozac http://linux-vserver.org/ | latest stable 2.2.0, 2.0.3-rc2, devel 2.3.0.12, stable+grsec 2.0.2.1, 2.2.0 | util-vserver-0.30.213 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the Wiki, and we'll forget about the minute ;) 1178195260 M * daniel_hozac micah, ensc, Hollow, phreak``, everyone else: util-vserver 0.30.213 1178195388 M * daniel_hozac cruser: get a 2.6.21.1 kernel, copy the config file to .config, and run make 1178195493 M * cruser daniel_hozac: Okay. 1178195532 M * daniel_hozac make sure it's the config from a non-working kernel. 1178195561 M * Bertl daniel_hozac: hmm? 1178195812 M * Bertl daniel_hozac: why a non-working kernel? 1178195846 M * daniel_hozac because i built one without the memory split. 1178195863 M * daniel_hozac (which worked) 1178195877 M * Bertl well, I think I don't get it :) 1178195896 M * cruser daniel_hozac: I have a 2.6.20-1.2307.fc5smp kernel and the 2.6.20-1.2312.fc5.vs2.2.0.1smp 1178195904 M * Bertl daniel_hozac: i.e. I would not advise anybody to use a config from a _non_ working kernel :) 1178195925 M * daniel_hozac Bertl: it's to see if HIGHMEM + memory split is broken for mainline too. 1178195983 M * daniel_hozac cruser: the first doesn't work, right? 1178196028 M * phreak`` daniel_hozac: 'hokay 1178196030 M * Bertl ah, now I get it, tx :) 1178196050 M * cruser daneil_hozac: That is what I am trying to remember. If there is a 2312 kernel I would say it was that one. 1178196070 M * phreak`` daniel_hozac: any huge changes compared to _rc6 ? 1178196088 M * daniel_hozac phreak``: i think Hollow fixed quite a few Gentoo-related things. 1178196095 J * FireEgl FireEgl@2001:5c0:84dc:0:48bc:dffa:e138:be22 1178196112 M * phreak`` daniel_hozac: 'k, I'll let hollow do the work then ;) 1178196116 M * daniel_hozac cruser: you're not running either of the kernels right now? 1178196152 M * cruser daniel_hozac: I am running the 2312 vserver kernel. 1178196168 M * daniel_hozac cruser: so 2307 would be broken. 1178196175 M * cruser The one you got working for me. 1178196180 M * daniel_hozac right, 1178196201 M * daniel_hozac (i'll be building a new one RSN, with the recent security updates etc) 1178196276 M * daniel_hozac Bertl: do you think it could have anything to do with http://cvs.hozac.com/viewvc/rpms/kernel/fedora-5/linux-2.6-split-wo-embedded.patch?revision=1.1 ? 1178196335 M * Bertl could be .. 1178196506 M * cruser daniel_hozac: /boot shows a 2307 kernel but looking at /usr/src/kernels there is not any with that name. 1178196534 M * daniel_hozac cruser: that's expected. 1178196546 M * cruser where so I find the config file? 1178196558 M * daniel_hozac /boot/config- 1178196610 M * cruser okay. Let me find a 2.6.21.1 kernel to download. 1178196731 M * daniel_hozac cruser: note that FC5 is going to be EOL in a couple of months. 1178196849 M * cruser daneil_hozac: meaning maintance will stop on it? 1178196856 M * daniel_hozac yep. 1178196921 M * cruser If upgrading means doing a complete reinstall...I do not know 1178196951 M * daniel_hozac well, anaconda has an upgrade mode too. 1178196972 M * daniel_hozac and it's usually possible to yum update to a new release with some minor fiddling. 1178197079 M * cruser Thanks. 1178197182 Q * Aiken Quit: Leaving 1178197196 M * cruser daniel_hozac: I am worried about possibly having built my guests using the deprecated "dynamic" way. 1178197222 M * cruser I am just building this development box and am most of the way through. 1178197251 M * daniel_hozac just set it now then. 1178197265 M * daniel_hozac stop the guest, echo > /etc/vservers//context, start it again. 1178197343 M * cruser daniel_hozac: whew...that's easy. A different number for each guest...right? 1178197357 M * daniel_hozac yep. 1178197401 J * tamitall ~assmaster@gw.nettam.com 1178197474 M * Bertl welcome tamitall! 1178197730 M * cruser daniel_hozac: I have downloaded linux-2.6.21.1.tar.gz my box is dual core. Is this alright? 1178197760 M * daniel_hozac yerp. 1178197876 M * cruser daniel_hozac: I have untarred it into /usr/src to a make a ln -s to /usr/src/linux? 1178197960 M * daniel_hozac that's not important. 1178198194 M * cruser daneil_hozac: I have compiled programs but not the kernel. How do I compile using the config from the other kernel? 1178198245 M * daniel_hozac hmm, maybe it would be better to use a 2.6.20.11 kernel... 1178198256 M * daniel_hozac 2.6.21 is bound to have lots of new options. 1178198277 M * daniel_hozac sorry i didn't think of that before. 1178198292 M * cruser Okay. Which one do you want me to download? 1178198371 M * daniel_hozac linux-2.6.20.11.tar.bz2 should be fine. 1178198429 M * cruser okay 1178198591 Q * Piet Quit: Piet 1178198759 Q * infowolfe Quit: Leaving 1178199117 M * cruser daniel_hozac: Okay. I have /usr/src/linux-2.6.20.11 how do I compile using the config file? 1178199144 M * daniel_hozac cp config- .config 1178199147 M * daniel_hozac make oldconfig 1178199151 M * daniel_hozac make 1178199163 M * daniel_hozac wait, not make. 1178199172 M * daniel_hozac you'll have to change a few settings first to get the split. 1178199320 M * daniel_hozac open .config, and change "# CONFIG_EMBEDDED is not set" to "CONFIG_EMBEDDED=y", and rerun make oldconfig 1178199357 J * Hollow ~hollow@styx.xnull.de 1178199367 M * cruser daniel_hozac: okay. /usr/src/linux-2.6.20.11 # cp /boot/config-2.6.20-1.2307.fc5smp .config 1178199377 M * daniel_hozac right. 1178199390 M * Hollow Bertl: ping 1178199403 M * Bertl hehe 1178199442 M * Bertl Hollow: what's up? 1178199446 M * Hollow Bertl: the new box is already running, i would like to migrate during the weekend... just need your help for dns 1178199464 M * Bertl okay, just give me a short notice (and the new ip) 1178199484 M * Bertl will reduce the TTL now ... 1178199485 M * Hollow ok, what about the mailing list? should we move it? 1178199489 M * Hollow ok, great 1178199506 M * Bertl well, we have to get the current subscription list from martin (somehow) 1178199546 M * Hollow ok, do you have his mail address? 1178199550 M * cruser daniel_hozac: Okay CONFIG_EMBEDDED=y . But I do not know where "oldconfig" is. Or is that new .config? 1178199577 M * daniel_hozac cruser: it's a make target. you run make oldconfig to have it look over your configuration file. 1178199606 M * cruser got it. 1178199627 M * Bertl Hollow: martin at list-petersen dot dk 1178199643 M * Hollow ok, thanks, will contact him 1178199667 A * phreak`` pokes Hollow 1178199671 M * daniel_hozac good luck... ;) 1178199698 A * Hollow pokes back 1178199709 M * phreak`` Hollow: wanna do the bump ? 1178199724 M * Hollow bump for? 1178199731 M * phreak`` (1) I don't feel like doing anything else today and (2) you did some of the changes in trunk 1178199734 M * phreak`` util-vserver :P 1178199742 M * Hollow ah.. 0.30.213! yay! 1178199744 M * Hollow :) 1178199748 M * phreak`` Hollow: exactly 1178199748 M * Hollow will do it 1178199754 M * phreak`` Hollow: thanks a lot :P 1178199916 J * ema ~ema@rtfm.galliera.it 1178199956 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1178200945 J * dna_ ~naucki@96-204-dsl.kielnet.net 1178201298 M * Bertl Hollow: hmm, we are already on 4H, so I guess that should be fine :) 1178201319 M * Hollow ok :) 1178201349 Q * dna Ping timeout: 480 seconds 1178201969 M * svenk I am having a problem with /proc/self I cannot stat it even though ls shows it 1178201995 M * svenk stat says there is no such file or directory 1178202086 M * Bertl sounds interesting ... kernel/patch version? 1178202115 M * svenk latest stable: 2.6.19.7-grsec2.1.10-vs2.2.0-rc21 1178202189 M * svenk this does not happen in all vservers 1178202334 M * cruser daniel_hozac: "make" is finished. What do I do next. Without over writing my working kernel :) 1178202712 M * Bertl svenk: please try with final vs2.2.0 and without grsec 1178202798 M * Bertl svenk: if the issue can be reproduced there, we will investigate 1178202905 M * svenk Bertl: ok will build a kernel without grsec 1178203106 J * dothebart ~willi@xdsl-87-78-53-48.netcologne.de 1178203283 M * daniel_hozac cruser: try to boot arch/i386/boot/bzImage 1178203479 Q * tudenbart Read error: Connection reset by peer 1178203844 M * svenk Bertl: without grsec everything is fine 1178204184 J * fluor- ~fluor@tanneries.squat.net 1178204200 M * fluor- hi there 1178204223 M * fluor- I just did something very stupid: rm -rf /etc/vservers/ on a Debian box 1178204235 M * fluor- now I'm reconstructing the directory structure 1178204245 M * fluor- and everything's pretty obvious, apart from the 'context' file 1178204250 M * Bertl svenk: okay :) 1178204265 M * fluor- can I put any arbitrary number in there, or is there a specific scheme to be aware of? 1178204276 M * Bertl fluor-: if you use xid tagging, the files will show the xid 1178204288 M * cruser daniel_hozac: bzImage would be the same as /boot/initrd-2.6.20-1.2307.fc5smp.img . I am wanting to update my grub/menu.lst. 1178204300 M * Bertl fluor-: if not, you can choose any number between 2 and 49151 (which is unique) 1178204352 M * fluor- Bertl: how can I know if I'm using xid tagging? 1178204395 M * Bertl hmm, if you mount your filesystem with tagxid or tagctx, that is a good indication 1178204416 M * fluor- Bertl: no such thing in my fstab 1178204442 M * fluor- I'm running a pretty standard setup, apart from the fact that vservers are sitting on a LUKS encrypted partition 1178204474 M * Bertl k, then unique number between 2 and 49151 should be fine 1178204479 M * fluor- Bertl: thanks! 1178204503 M * fluor- Bertl: one last thing - one of the vservers I deleted the config of is still running - can I get its context? 1178204532 M * Bertl vserver-stat should show it 1178204553 M * fluor- Bertl: is that the first field? 1178204579 M * fluor- because I have 49154 - which is outside the possible values you gave me 1178204594 M * Bertl means you are outdated and using dynamic contexts 1178204607 M * Bertl (they are deprecated for a few years now) 1178204615 M * fluor- oh my 1178204703 M * fluor- it's weird though 1178204714 M * fluor- it's a recent Debian Etch install 1178204719 M * fluor- and I'm using the config defaults 1178204733 M * sid3windr nooooooo 1178204736 M * sid3windr don't say that 1178204743 M * sid3windr now Bertl has a chance to bash debian again :P 1178204762 M * Bertl nah, we _know_ that debian is outdated ... 1178204772 M * sid3windr we do? :p 1178205014 J * BobR odie@IRC.13thfloor.at 1178205273 M * fluor- how can I switch to non-dynamic contexts? seems like I have the proper set of tools - might be lacking the kernel feature though 1178205278 N * BobR BobR_oO 1178205296 M * sid3windr echo 31337 > /etc/vservers/bomba/context 1178205328 M * fluor- is it as simple as choosing a number? 1178205494 M * Bertl fluor-: kernel supports it since the beginning (static contexts) and so do the tools :) 1178205521 M * Bertl fluor-: yes, it is as simple as choosing an unique number :) 1178205536 M * fluor- I did, and my vserver now refuses to start 1178205555 M * Bertl what error message? 1178205626 M * fluor- nothing shows 1178205632 M * fluor- is vserser supposed to tell in syslog? 1178205656 N * Bertl Bertl_oO 1178205666 M * sid3windr there is no error when you start it? 1178205676 M * Bertl_oO dinnertime ... back shortly 1178205694 M * fluor- just "Failed to start vserver 'dns'" 1178205709 M * fluor- removed the 'context' file, and now it starts ok 1178205733 M * sid3windr hmz 1178205748 M * fluor- is there any good reason to abandon dynamic contexts, btw? 1178205759 M * fluor- I mean, is it important to do so? 1178205806 J * stefani ~stefani@flute.radonc.washington.edu 1178206284 M * svenk Bertl_oO: without the grsec chroot restrictions it works fine too 1178206982 M * harry svenk: did you read the readme in my homedir? 1178206997 M * harry regarding the chroot restrictions? 1178207168 M * daniel_hozac cruser: no, bzImage is the same as vmlinuz 1178207208 M * daniel_hozac cruser: you'd generate the initrd with mkinitrd /boot/initrd-.img after running make modules_install. 1178207268 M * svenk harry: no i didn't 1178207274 M * harry do that then ;) 1178207282 M * harry it would have saved you the trouble you had :) 1178207341 M * cruser daniel_hozac: I would like to follow my grub configuration. So I would copy bzImage to /boot/vmlinuz-2.6.20.11 . 1178207366 M * daniel_hozac cruser: yep. 1178207380 M * svenk harry: thank you 1178207411 M * cruser daniel_hozac: Then where/when do I run "make modules_install"? 1178207474 M * daniel_hozac cruser: in the kernel source tree. 1178207540 M * cruser daneil_hozac: sorry, that will not touch any of my other kernel bits (paraniod) 1178207611 M * daniel_hozac no. 1178207723 M * fluor- ok, so it seems I can't assign a static context to a vserver that was installed with dynamic context 1178207729 M * fluor- they don't start 1178207884 M * cruser daniel_hozac: Looking to "mkinitrd" but I see no /boot/initrd-.img 1178207905 M * cruser I have run make modules_install 1178207952 M * daniel_hozac did you run mkinitrd? 1178207961 M * daniel_hozac fluor-: sure you can. 1178208060 M * cruser no. Where/parameters do I run it? 1178208081 J * gypsymauro ~Io@84.18.151.77 1178208084 M * gypsymauro hi 1178208104 M * gypsymauro I'm trying to use screen on a vserver guest but I get this error Cannot open your terminal '/dev/pts/0' - please check. 1178208108 M * gypsymauro any hint? 1178208151 N * bianchi_away bianchi 1178208185 M * daniel_hozac cruser: mkinitrd /boot/initrd-.img 1178208204 M * daniel_hozac gypsymauro: did you use vserver ... enter with old utils to get into the guest? 1178208239 M * gypsymauro daniel_hozac: uh? 1178208256 M * Bertl_oO gypsymauro: you are not using screen inside the guest, no? 1178208267 M * gypsymauro I'm trying to use it inside it 1178208283 M * Bertl_oO gypsymauro: you are using it on the host _after_ you entered a guest, yes? 1178208292 M * gypsymauro yes 1178208298 M * gypsymauro (sorry for my bad english) 1178208303 M * cruser daniel_hozac: I see no /boot/initrd-.img file which I assume mkinitrd wants. 1178208308 M * gypsymauro vserver guest enter 1178208312 M * gypsymauro screen :) 1178208313 M * Bertl_oO gypsymauro: so your options are: 1178208323 M * Bertl_oO - screen, vserver guest enter 1178208327 M * daniel_hozac cruser: no, that's what creates it. thus the name, make initrd... 1178208329 M * Bertl_oO - update the tools 1178208336 M * Bertl_oO - ssh into the guest, screen 1178208420 M * gypsymauro why are u sayng to udpate te tool? 1178208445 M * daniel_hozac because you're using really old ones. 1178208446 M * Bertl_oO gypsymauro: because older tools do not detach the terminal, which is what you are experiencing 1178208476 M * gypsymauro I'm using sarge util-vserver 0.30.204 1178208488 M * Bertl_oO we are at 0.30.213 now :) 1178208517 M * daniel_hozac sarge is ancient... 1178208557 M * cruser daniel_hozac: Okay. mkinitrd /boot/initrd-2.6.20.11.img 2.6.20.11 let me get that into grub/menu.lst 1178208557 M * gypsymauro I know I know :/ but is on a production machine, so.. what if I update tools? 1178208569 M * gypsymauro vservers still works? 1178208669 J * bonbons ~bonbons@ppp-111-238.adsl.restena.lu 1178208670 M * Bertl_oO I suppose so :) 1178208698 M * gypsymauro :) 1178209518 J * bzed ~bzed@dslb-084-059-104-121.pools.arcor-ip.net 1178209648 J * webcruiser ~webcruise@201.90.175.132 1178209749 Q * gypsymauro Quit: Lost terminal 1178209770 M * cruser daniel_hozac: Okay. ~ # uname -r 1178209772 M * cruser 2.6.20.11 1178209793 M * daniel_hozac cruser: and you set the different memory split? 1178209819 M * cruser I made the change you told me to in the config file 1178209823 M * daniel_hozac what does grep VMSPLIT /usr/src/linux-2.6.20.11/.config say? 1178209910 M * cruser nothing 1178209931 M * daniel_hozac hmm 1178209935 M * daniel_hozac and grep EMBEDDED? 1178209998 M * cruser # CONFIG_EMBEDDED is not set. :( I was sure I set that. 1178210019 P * webcruiser 1178210031 M * cruser yes I have it in the .config file. 1178210049 M * cruser No! I don't 1178210068 M * cruser Do I recompile? 1178210268 M * daniel_hozac yeah. 1178210285 M * daniel_hozac and you have to redo make oldconfig after changing it in .config. 1178210291 Q * FireEgl Ping timeout: 480 seconds 1178210449 M * cruser daniel_hozac: I have changed .config . Now I run "make oldconfig"...? 1178210488 M * bianchi hi 1178210558 M * bianchi I'm trying to setup a vserver with quota support but when I do quotaon (inside the vserver) returns with quotaon: using //quota.group on /dev/hdv1 [/]: OperaciĆ³n no permitida 1178210578 Q * fluor- Quit: . 1178210614 M * bianchi "Operacion no permitida" (spanish) means "Operation not permitted" 1178210619 M * cruser daniel_hozac: I ran "make oldconfig" and it has taken out the EMBEDDED I set in .config 1178211374 M * cruser daniel_hozac: after running "make oldconfig" I opened .config and put in the CONFIG_EMBEDDED=y and then ran "make". Which gave me this: .config:54:warning: trying to reassign symbol EMBEDDED 1178211395 M * cruser daniel_hozoc: ...going to lunch 1178211411 Q * daniel_hozac Ping timeout: 480 seconds 1178213067 N * Bertl_oO Bertl 1178213071 M * Bertl back now ... 1178213114 M * bonbons derjohn: ping 1178213127 M * derjohn bonbons, re ;) 1178213234 M * bonbons have you checked which patch causes the missing per-cpu issue to araise? 1178213344 M * bonbons don't know what exactly uses per-cpu data in all modules (I'm running with similar config, but ipv6 compiled in and a few patches less) 1178213997 Q * chand Ping timeout: 480 seconds 1178214521 M * Bertl nap attack ... back later ... 1178214526 N * Bertl Bertl_zZ 1178214682 Q * bianchi Quit: using sirc version 2.211+KSIRC/1.3.12 1178215269 M * cruser daniel_hozoc: back from lunch Are you there? 1178215283 M * nebuchadnezzar hello everybody 1178216163 P * stefani I'm Parting (the water) 1178216566 Q * ema Quit: leaving 1178216855 N * DoberMann[PullA] DoberMann 1178217920 J * duckx ~Duck@tox.dyndns.org 1178218378 J * spoi ~asdf@p54AD094A.dip0.t-ipconnect.de 1178218382 P * spoi 1178218387 J * spoi ~asdf@p54AD094A.dip0.t-ipconnect.de 1178220012 Q * virtuoso Read error: Connection reset by peer 1178220441 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de 1178221528 J * daniel_hozac ~daniel@c-2f1472d5.08-230-73746f22.cust.bredbandsbolaget.se 1178223999 J * ema ~ema@rtfm.galliera.it 1178224290 J * virtuoso ~s0t0na@80.253.205.251 1178224744 Q * bonbons Quit: Leaving 1178224801 M * cruser daniel_hozac: ping 1178225401 J * sasasasa VistaScr@88.243.102.11 1178225656 Q * sasasasa 1178225734 J * Hurga nobody@p508a8adb.dip0.t-ipconnect.de 1178225755 M * Hurga Hi. 1178226181 Q * spoi 1178226557 J * Aiken ~james@ppp222-137.lns2.bne1.internode.on.net 1178227328 M * daniel_hozac cruser: pong 1178227378 M * cruser daniel_hozac: make oldconfig over wrote the setting 1178227404 M * daniel_hozac really? 1178227417 M * daniel_hozac and you changed # CONFIG_EMBEDDED is not set to CONFIG_EMBEDDED=y? 1178227432 M * cruser yes. 1178227443 M * daniel_hozac that's weird. 1178227462 M * cruser I changed it after make oldconfig and "make" choked on it. 1178228220 M * cruser daniel_hozac: do want to keep trying? 1178228392 M * daniel_hozac cruser: maybe tomorrow, when i've tried doing it myself ;) 1178228411 M * cruser okay...going home. 1178228420 P * cruser 1178229048 Q * lilalinux Remote host closed the connection 1178230228 Q * meandtheshell Quit: Leaving. 1178230334 Q * dna_ Quit: Verlassend 1178231027 N * Bertl_zZ Bertl 1178231790 Q * ema Quit: leaving 1178232146 J * Cichetto ~morfheus2@217.220.105.226 1178232155 P * Cichetto 1178232836 Q * dreamind Quit: dreamind 1178233427 N * DoberMann DoberMann[ZZZzzz] 1178233781 J * DoberMann_ ~james@AToulouse-156-1-134-78.w90-30.abo.wanadoo.fr 1178233909 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds 1178234464 M * derjohn daniel_hozac, did you use ipv6 as module or compiled in hard ? 1178234497 M * derjohn reading bonbons i might make a difference concerning the percpu stuff 1178234708 J * toidinamai__ ~frank@i59F75AF3.versanet.de 1178235034 M * tanjix Bertl: are you there? 1178235036 M * Bertl derjohn: it definitely makes a difference, and the issue happened with ipv6 as module, IIRC 1178235041 M * Bertl tanjix: yep 1178235056 M * tanjix Bertl: I reinsalled that "vps024" with amd64 now 1178235068 M * tanjix and we recheck some things to see if it brought some performance? 1178235105 M * Bertl okay, do you have some guests running on it? 1178235113 M * tanjix you remember, that was the machine with exemely high loads 1178235120 M * tanjix extremely 1178235129 M * tanjix yes, all guests are runnin 1178235130 M * Bertl yep, did you split up the guests to the two disks? 1178235133 Q * toidinamai_ Ping timeout: 480 seconds 1178235146 M * tanjix no, not yet 1178235162 M * tanjix i would do that if amd64 itself will not speed up the performance 1178235229 M * tanjix the amd64 has same kernel and tools etc as before 1178235279 M * Bertl hmm, I hope it is a 64bit kernel now :) 1178235306 M * tanjix i hope? can i verify this? 1178235313 M * Bertl and the tools should be 64bit too, although 32bit versions would work ... 1178235332 M * Bertl tanjix: okay, let's upload the output of: 1178235356 M * Bertl - 'vserver-info - SYSINFO' 1178235363 M * Bertl - 'cat /proc/meminfo' 1178235372 M * Bertl - 'vmstat and iostat' 1178235423 M * Bertl yeah, that should do for now 1178235484 M * tanjix http://paste.linux-vserver.org/1619 1178235548 M * Bertl hmm 24G swap, nice :) 1178235564 M * tanjix hehe 1178235573 M * tanjix that was on the old machine too 1178235582 M * tanjix i thought the loads were because of too less swao 1178235584 M * tanjix swap 1178235622 M * Bertl basically, the rule for Linux-VServer system si, when you need swap, get a new hardware/more memory 1178235629 M * Bertl *systems is 1178235655 M * Bertl i.e., you don't want a system to go into swapping/trashing :) 1178235674 M * derjohn Bertl, yes. Then i'll try a "compile v6 statically" setup. O thought alloc_percpu would be called in either way: module or compiled in. thats why i asked. 1178235697 M * Bertl yes, it is, but there is a difference depending on the kernel config 1178235710 M * Bertl i.e. if modules are enabled, the percpu space is pre-reserved 1178235722 M * Bertl while in a monolithic kernel, it is hard allocated 1178235753 M * Bertl tanjix: and how is the load now? 1178235788 M * tanjix do you really want to know? :) 1178235797 M * tanjix 01:43:00 up 1:21, 3 users, load average: 67,96, 68,98, 39,60 1178235808 M * tanjix but maybe as the guests are statting right now? 1178235814 M * tanjix starting 1178235830 M * Bertl are they? 1178235835 M * tanjix yes 1178235842 M * tanjix finished now 1178235846 M * tanjix 61 guests 1178235850 M * Bertl well, for that it's not too bad :) 1178235889 M * tanjix load is @ 39 now 1178235899 M * tanjix still decreasing 1178235917 M * tanjix now raised from 33 to 41 1178235958 M * tanjix and again decreasing.. 1178236009 M * tanjix vtop shows the blow-ups when the confixx_counterscript inside the guests are runnign (cronjob) 1178236050 M * tanjix although those raises load is decreasing, currently @ 21 1178236139 M * tanjix dmesg shows: 1178236140 M * tanjix vxW: !!! limit: ffff810113af2078[NPROC,6] = 255 on exit. 1178236140 M * tanjix ntpd[10764:#49210]: segfault at 00000000000001b0 rip 0000000008062403 rsp 00000000ff87df50 error 4 1178236146 M * tanjix is that from any meaning? 1178236164 M * Bertl it means that ntpd inside a guest segfaulted 1178236173 M * tanjix and the line above? 1178236205 M * Bertl is unusual, and you should move away from dynamic context :) 1178236257 M * tanjix this is planned for one of the next things once we solved the current load problem :) 1178236258 M * Hurga BTW, any ideas why segfaults get logged on AMD64 but not on P4? 1178236306 M * Bertl Hurga: I guess because the actual page fault logic is per arch, and x86_64 has a log entry there? 1178236332 M * derjohn Bertl, thx! And: Good night ! 1178236389 M * Hurga Bertl: maybe... just surprised that the 32 and 64 bit vendor kernels are differently behaving. I'll check, thx. 1178236625 Q * Hurga Remote host closed the connection