1201565021 Q * dowdle Remote host closed the connection 1201565343 Q * jescheng Remote host closed the connection 1201565361 J * jescheng ~jescheng@proxy-sjc-1.cisco.com 1201566029 Q * derjohn_mobil Ping timeout: 480 seconds 1201567995 Q * ViRUS Quit: Leaving 1201568919 Q * Infinito Quit: Quitte 1201570771 J * baldy baldy@weltkugel.baldy.biz 1201570802 M * baldy hi there 1201571477 Q * Punkie Remote host closed the connection 1201580278 Q * balbir__ Remote host closed the connection 1201581389 J * doener_ ~doener@i577B9781.versanet.de 1201581488 Q * doener Ping timeout: 480 seconds 1201583308 Q * grobie Ping timeout: 480 seconds 1201583474 P * etteyafed 1201586404 Q * jescheng Remote host closed the connection 1201586414 J * jescheng ~jescheng@proxy-sjc-1.cisco.com 1201586876 Q * hparker Quit: g'nite 1201587496 J * virtuoso_ ~s0t0na@ppp91-122-94-12.pppoe.avangard-dsl.ru 1201587664 Q * nkukard Quit: Leaving 1201587903 Q * virtuoso Ping timeout: 480 seconds 1201588147 J * derjohn_mobil ~aj@e180214150.adsl.alicedsl.de 1201588235 J * balbir ~balbir@59.145.136.1 1201588701 J * sharkjaw ~gab@shell.ormset.no 1201589930 J * Slydder ~chuck@194.59.17.53 1201589941 J * _bjh_ ~bjh@84.112.154.154 1201589956 M * Slydder morning all 1201590144 M * Linus hi Slydder 1201590208 Q * Slydder Remote host closed the connection 1201590428 J * DLange ~dlange@p57A30B10.dip0.t-ipconnect.de 1201590448 J * Slydder ~chuck@194.59.17.53 1201590459 Q * Slydder Remote host closed the connection 1201590484 J * Slydder ~chuck@194.59.17.53 1201590543 Q * Slydder Remote host closed the connection 1201590560 J * Slydder ~chuck@194.59.17.53 1201590872 Q * Linus Quit: i'll be back 1201592262 Q * DLange Quit: Bye, bye. Hasta luego. 1201592298 J * JonB ~NoSuchUse@130.227.63.19 1201592315 Q * JonB 1201593000 M * pmjdebruijn hmm 1201593042 M * pmjdebruijn I noticed it's not possible to do LABEL= based mounting from within a vserver 1201593051 M * pmjdebruijn blkid returns nothing as well 1201593066 M * pmjdebruijn do I need to give the vserver a certain capability? 1201593331 J * esa` bip@ip-87-238-2-45.static.adsl.cheapnet.it 1201593353 Q * esa Ping timeout: 480 seconds 1201593753 Q * balbir Ping timeout: 480 seconds 1201594816 J * chigital ~chigital@p5B0C54F2.dip.t-dialin.net 1201595017 J * dna ~dna@107-247-dsl.kielnet.net 1201595935 Q * chigital Ping timeout: 480 seconds 1201596124 J * larsivi ~larsivi@85.221.53.194 1201596509 J * jsambrook ~jsambrook@aelfric.plus.com 1201597211 J * gdistasi ~gdistasi@143.225.229.142 1201597222 Q * gdistasi Remote host closed the connection 1201599255 J * yaboo ~jsirucka@220-245-131-131.static.tpgi.com.au 1201599355 M * yaboo when I try to ssh into the vserver component it takes me to the vserver host 1201599360 M * yaboo any reason why 1201599420 M * daniel_hozac 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 1201599864 J * chigital ~chigital@91.90.144.102 1201600198 M * yaboo thanks daniel_hozac 1201600451 Q * arachnist Read error: Connection reset by peer 1201600649 J * arachnist arachnist@088156187175.who.vectranet.pl 1201600744 Q * derjohn_mobil Ping timeout: 480 seconds 1201600979 M * meebey if my host runs amd64, can vserver guests be x86/ia32 then? 1201601043 M * meebey I hit the memory barrier issue (3.3GB), so I would like to run a 64bit host OS, but 32bit bit guests, as none of the guest needs > 3.3GB 1201601079 M * daniel_hozac works fine. 1201601094 M * daniel_hozac (as long as you enable the ia32 compat) 1201601222 Q * ||Cobra|| Remote host closed the connection 1201601267 M * meebey daniel_hozac: I need to enable that in vserver? 1201601299 J * pusling_ pusling@77.75.162.71 1201601305 M * daniel_hozac no, in the kernel. 1201601336 M * Slydder trying to force a rotation of /var/log/messages using logrotate -f /etc/logrotate.conf does not yield a cleared /var/log/messages file. any ideas? 1201601353 Q * pusling Ping timeout: 480 seconds 1201601357 M * meebey daniel_hozac: ah ok, all vservers then will be able to get 3.3GB? and not just all vservers together? 1201601424 M * daniel_hozac every process, yes. 1201601445 J * tuxbubling ~omauras@62.80.120.214 1201601448 M * tuxbubling hello there 1201601576 M * meebey daniel_hozac: ok thanks 1201601590 M * tuxbubling what about the lattest change in kernel 2.6.24??? about the process name space things, does it mean that vserver patchset is being merged in the tree, or this will be totally incompatible with vserver 1201601602 M * daniel_hozac it'll be optional. 1201601628 M * tuxbubling so you'll continue providing a patchset ?? 1201601680 M * daniel_hozac oh yes, mainline is faaar from being where we are... 1201601690 M * tuxbubling ok 1201601718 M * tuxbubling i'm asking cause.... as there's no patchset for 2.6.23, and that new stuff in 2.6.24.... 1201601751 M * daniel_hozac the parts they merged in 2.6.23 was too broken for us to bother. 1201601752 M * tuxbubling was a bit affraid to not having any more kernel update to use vservers :) 1201601762 M * tuxbubling ok daniel_hozac 1201601765 M * daniel_hozac we've got prepatches for 2.6.24 already, but they need more work. 1201601765 M * tuxbubling nice to hear 1201601818 M * tuxbubling you're doing a really good work 1201601826 M * tuxbubling that's working flawlessly 1201602224 J * giovanni ~giovanni@143.225.229.142 1201602243 M * giovanni hi there 1201602245 J * ktwilight ~ktwilight@248.74-66-87.adsl-dyn.isp.belgacom.be 1201602252 M * daniel_hozac hello 1201602407 M * tuxbubling hi giovanni 1201602515 Q * ktwilight_ Ping timeout: 480 seconds 1201602855 Q * jsambrook Quit: Leaving. 1201602950 J * jsambrook ~jsambrook@aelfric.plus.com 1201604102 N * pusling_ pusling 1201604501 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl 1201605228 Q * chigital Ping timeout: 480 seconds 1201605287 J * Punkie ~Punkie@goc.coolhousing.net 1201605859 J * chigital ~chigital@213.131.236.199 1201605918 Q * chigital 1201606217 M * matti Morning folks :) 1201606980 J * mattzerah ~matt@121.50.222.102 1201607268 J * JonB ~NoSuchUse@130.227.63.19 1201607589 Q * Aiken Remote host closed the connection 1201607868 N * mattzerah mattzerah`afk 1201608455 N * Bertl_zZ Bertl 1201608461 M * Bertl morning folks! 1201608559 M * daniel_hozac morning Bertl! 1201608595 M * Bertl I updated to 2.6.24 yesterday (internal tree) do we have any pending patches from 2.2.0.6 for that? 1201608629 M * Bertl I remembered the AFSrpc issues, but that should be solved there, no? 1201608747 M * JonB when i do a vserver enter it complains about LANGUAGE and LC_ALL, but not when i SSH into a guest? 1201608758 M * Bertl environment 1201608773 M * Bertl actually environment and locale settings 1201608781 M * JonB how do i fix it? 1201608790 M * Bertl try: export LC_ALL=C LANG=C 1201608819 M * daniel_hozac AFS and AXRPC should be fixed in the 2.6.24 tree. 1201608823 M * JonB ok 1201608869 M * jsambrook so Linux's lack of RO bind mount I note :P 1201608874 M * jsambrook so is Linux's lack of RO bind mount I note :P 1201608897 M * daniel_hozac that hasn't been a problem in Linux-VServer kernels since 2.0.2. 1201608910 M * Bertl jsambrook: hehe, well, it only has been about 3 years, since the first patches were submitted (for mainline) 1201608919 A * jsambrook sighs 1201608930 M * Bertl will definitely be in 2.6.10 :) 1201608934 M * jsambrook well done - for the patches and the patience 1201608942 M * Bertl *2.10 I mean 1201608982 M * Bertl we have given up submitting them, some guy from IBM is doing that now on a regular basis 1201608994 M * giovanni hi Bertl 1201609630 M * pmjdebruijn jsambrook: I have that working, you do need a recent util-vserver 1201609648 M * pmjdebruijn is it possible to see device labels from within a guest/vserver? 1201609823 M * Bertl daniel_hozac: I removed you from the CC list on Mark's mails about routing (ML), hope you don't mind :) 1201609840 M * daniel_hozac i'm on the ML ;) 1201609849 M * Bertl I assumed so :) 1201609861 M * Bertl pmjdebruijn: device labels? 1201609863 Q * ruskie Read error: Connection reset by peer 1201610025 J * ruskie ruskie@ruskie.user.oftc.net 1201611873 M * pmjdebruijn Bertl: ie, blkid 1201611879 M * pmjdebruijn Bertl: of LABEL= in the fstab 1201611909 M * Bertl you need direct device access to retrieve that, AFAIK 1201611924 M * Bertl and that is something you want to avoid inside a guest for security reasons 1201611949 M * Bertl but if you are referring to the fstab in the guest config, that might work, daniel_hozac? 1201611985 M * daniel_hozac assuming you use "auto" as the filesystem type, or some other not-internally-processed filesystem. 1201612022 M * pmjdebruijn Bertl: no in the guest itself... 1201612038 M * Bertl what do you need the label there for? 1201612057 M * pmjdebruijn we want to mount our backup volume on the fly (not at boot) 1201612060 M * pmjdebruijn ie vserver startup 1201612070 M * daniel_hozac why does the guest need to do the mount? 1201612083 M * daniel_hozac i.e. why not do it on the host? 1201612096 M * Bertl that would have been my next question. tx :) 1201612102 M * pmjdebruijn because then it won't be visible in the guest 1201612136 M * Bertl first, why do you need it visible on the guest if you are doing backup, second, why not _make_ it visible to the guest? 1201612136 M * daniel_hozac that's what vnamespace is for.. 1201612139 M * daniel_hozac (and shared subtrees) 1201612217 M * Bertl pmjdebruijn: don't get me wrong, you can accomplish what you want too, but you have to pay a high price on lowered security 1201612230 M * pmjdebruijn well, how do I accomplish that? 1201612238 M * pmjdebruijn Bertl: it depends, I'm still evaluating options 1201612262 M * Bertl if you copy all device nodes for block devices into your guest, then label tools will be able to retrive it 1201612269 M * pmjdebruijn we plan on revising our backup solution within a couple of weeks 1201612271 M * daniel_hozac mount --make-shared /vservers//mnt; mount it on /vservers//mnt 1201612276 M * Bertl of course, guest root will also be able to rewrite your host disks 1201612313 M * pmjdebruijn Bertl: that the point of course 1201612318 M * daniel_hozac (of course, the --make-shared will need to happen before you start the guest) 1201612324 M * pmjdebruijn our host itself is initrd based 1201612340 M * pmjdebruijn daniel_hozac: that sortof defeat the purpose 1201612342 J * Infinito argos@200-103-111-131.gnace701.dsl.brasiltelecom.net.br 1201612350 M * daniel_hozac how so? 1201612352 M * pmjdebruijn each guest has it's own backup volume in our current setup 1201612359 M * daniel_hozac --make-shared is a one time thing. 1201612370 M * Bertl pmjdebruijn: why not mount them into the guest then? 1201612371 M * pmjdebruijn we don't want our backup volume permanently mounted 1201612375 M * daniel_hozac then, whenever you mount something in that area, the guests will have access. 1201612380 M * pmjdebruijn Bertl: that's an option 1201612388 M * pmjdebruijn however currently our backup script run inside the guest 1201612394 M * pmjdebruijn though that could change 1201612401 M * daniel_hozac i.e. you don't do it when there's something mounted, it's _before_ that. 1201612428 M * pmjdebruijn btw, so I could theoretically bind mount /dev right? 1201612434 M * pmjdebruijn though that does sound nasty 1201612445 M * Bertl trust me, it _is_ nasty :) 1201612457 M * pmjdebruijn we should probably mount the backup outside the guest, and run our backups scripts accordingly 1201612460 M * daniel_hozac cat /dev/zero > /dev/kmeme 1201612466 M * pmjdebruijn haha :) 1201612496 M * daniel_hozac seriously though, you want shared subtrees. 1201612521 M * Bertl well, it has one disadvantage over mounting it _into_ the guest though 1201612547 M * pmjdebruijn daniel_hozac: why? 1201612568 M * daniel_hozac because they were created for this type of thing? 1201612581 M * daniel_hozac Bertl: hmm? 1201612614 M * Bertl as he has different backup volumes for each guest, all of them would see the backup volume of the current guest, and concurrent backups would be impossible 1201612644 M * daniel_hozac well, there could be multiple shared subtrees in different places. 1201612646 M * Bertl mounting the proper backup volume into the guest would solve that easily 1201612735 M * pmjdebruijn our backups are _mostly_ serial, to keep our SAN from getting loaded to hard 1201612815 P * jsambrook Leaving. 1201613020 M * Bertl well, important thing is, you now have a bunch of options to choose from ... make your choise wisely :) 1201613032 M * Bertl okay, off for now .. wbbl 1201613036 N * Bertl Bertl_oO 1201613572 Q * Infinito Quit: Quitte 1201614418 J * balbir ~balbir@122.167.108.162 1201614438 J * lilalinux ~plasma@80.69.41.3 1201614653 J * ftx__ ~ftx@dslb-084-060-221-030.pools.arcor-ip.net 1201615222 Q * balbir Remote host closed the connection 1201615573 J * Infinito argos@200-103-111-131.gnace701.dsl.brasiltelecom.net.br 1201615607 Q * sharkjaw Quit: Leaving 1201615673 M * pmjdebruijn Thanks! 1201615822 J * ema ~ema@rtfm.galliera.it 1201615883 M * JonB does a vserver guest that mounts it's own partitions need ext3 fs utilities (ext3 filesystem is on that partition) 1201616384 J * yarihm ~yarihm@vpn-global-dhcp1-230.ethz.ch 1201616397 M * harry no 1201616420 J * Julius ~julius@p57B24478.dip.t-dialin.net 1201616995 M * pmjdebruijn JonB: lots of programs depends on e2fsprogs, so it's not much use to remove them 1201617005 Q * Julius Ping timeout: 480 seconds 1201617592 M * JonB and e2fslibs? 1201617671 J * Julius ~julius@p57B24478.dip.t-dialin.net 1201617686 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1201617742 M * marcfiu hello... I am trying to run qemu inside a vserver and am getting a Bus error very early in when running it. 1201617755 M * marcfiu Is there anything special required in terms of vserver setup for qemu? 1201617771 Q * larsivi Quit: Konversation terminated! 1201617791 M * marcfiu I'm using fc6 and qemu-0.8.2-4.fc6 1201617829 M * marcfiu the guest is running the latest release of fc6 1201617849 M * marcfiu and the host is running 2.6.20-1.2949.fc6.vs2.2.0.1.0.el5.2007.08.15smp, which I think was snarfed from daniel_hozac. 1201617914 M * JonB did you check the wiki page with problematik programs? 1201617925 M * marcfiu JonB: nope 1201617930 M * marcfiu let me check... brb 1201618072 M * marcfiu nothing there wrt qemu 1201618083 M * marcfiu I'm looking here: http://linux-vserver.org/Problematic_Programs 1201618132 Q * Infinito Quit: Quitte 1201618219 M * JonB it doesnt seem to be a problem 1201618257 M * marcfiu when I run it on a regular fc6 box then the invocation of the same version of qemu and identical image file just works. 1201618262 M * marcfiu But it does not work inside a vserver. 1201618266 M * marcfiu So something is up. 1201618272 M * marcfiu will dig further 1201618421 M * JonB enough memory? 1201618519 J * balbir ~balbir@122.167.108.162 1201618568 M * marcfiu no limits are set 1201618573 M * marcfiu and the box has 4GB of ram 1201618586 M * marcfiu host is 64bit 1201618589 M * marcfiu guest is 32bit 1201618611 M * JonB that might be why 1201618695 M * marcfiu a'ight 1201618703 M * marcfiu i'll switch to using a 64bit guest 1201618708 M * marcfiu and see if I can get further. 1201618728 M * marcfiu however, I do want to let folks on our systems use qemu from a 32bit guest on top of a 64bit host. 1201618821 M * marcfiu indeed qemu gets further in this instance. 1201618838 M * marcfiu but I am using qemu-0.9.0-3.fc7 in Fedora 7 1201618845 M * JonB why do you want them to be able to use a 32bit guest? 1201618855 M * JonB isnt it enough that the qemu is 32 bit? 1201618872 M * marcfiu our users are now purely 32bit guests. 1201618892 M * marcfiu switching them to 64bit may take more work than we really want to put in 1201618899 M * marcfiu put in right now 1201618913 M * JonB okay 1201618922 M * marcfiu if it was just me, I'd do the switch :) 1201618923 M * JonB why do you want to allow them to run qemu? 1201618957 M * marcfiu well, I really want to let them run qemu w/ kvm on our VT enabled boxes, but we have a number of non-VT enabled systems. 1201618976 M * JonB VT? 1201618985 M * marcfiu hardware support for virtualization 1201619015 M * marcfiu our project is www.planet-lab.org 1201619021 M * JonB okay 1201619026 M * marcfiu we have 800+ machines in 400+ sites in 40+ countries. 1201619039 M * marcfiu our users are distributed systems and networking researchers. 1201619057 M * marcfiu We use vserver to efficiently share system resources in a scalable manner. 1201619062 M * JonB okay 1201619076 M * JonB i suppose you vhasify the qemu 1201619090 M * marcfiu we currently don't use vhashify at all. 1201619118 M * marcfiu in fact, we are still in the process of upgrading from an ancient 2.6.14 fc4 based system. 1201619148 M * JonB okay 1201619213 M * marcfiu Are working on moving towards 2.6.22-14+ vs 2.3.0+ 1201619219 M * marcfiu probably using f7 or f8 1201619239 M * marcfiu ok... I'll try a fc6 x86_64 solution. 1201619299 M * pmjdebruijn fc6 will only be supported for just a few months right? 1201619307 M * marcfiu I know... 1201619326 M * marcfiu but we now build all of our software for fc4, fc6, f7, f8 for both 32 and 64 bit hosts. 1201619358 M * marcfiu I'm only going to see in fc6 whether the issue I was running into with qemu was a vserver 32bit & 64bit host problem. 1201619559 M * marcfiu ok... running qemu in a fc6 64bit guest on a 64bit host with the image in question just works. 1201619576 M * marcfiu So it seems that there is a 32bit guest + 64bit host incompatibility. 1201619597 M * marcfiu Now checking whether this problem has gone away with a more recent version of qemu (f7 and f8). 1201619757 A * ard6 guesses you can run the 64bit qemu in a 32 bit environment on a 64 bit kernel 1201619768 M * marcfiu problem exists in all variants 1201619777 M * marcfiu ard6: hmmm 1201619779 M * marcfiu interesting. 1201619823 M * ard6 qemu doesn't have 64 bit kernel->32 bit user space support, so just run the 64 bit version in the 32 bit user space ;-) 1201619851 M * marcfiu the image I'm pointing qemu at is all 32 bit 1201619855 A * ard6 actually just installed a 64 bit vserver environment just to get around it 1201619875 M * ard6 the image is not important... it's qemu itself and the kernel that's important 1201619895 M * marcfiu ard6: yeah... as I explained to JonB, I cannot switch to a 64bit environment for various reasons. 1201619904 M * marcfiu ard6: ok... will give it shot 1201619905 A * ard6 understands... 1201619922 M * tuxbubling marcfiu: i didn't get the reasons? 1201619947 M * marcfiu ard6: rawk! 1201619956 M * ard6 eh, do you need a reason not to switch? :-) 32 bits works :-) 1201619983 M * marcfiu qemu-system-x86_64 from inside the 32bit guest on the 64bit host "just works!". :) 1201619989 M * ard6 legacy 32 bit proprietary applications f.i. 1201619991 M * tuxbubling and if you use fc flavors, you'll always have 32b environment as FC is installing _all_ the i386 stuff .... 1201619993 M * marcfiu tuxbubling: www.planet-lab.org 1201619997 M * marcfiu we have 800+ machines 1201620000 M * marcfiu some are 64 bit 1201620006 M * tuxbubling btw that takes more space.... 1201620025 M * ard6 12 59 474.7M 11.7M 5h18m07 8h46m55 104d19h04 etchdev32 1201620025 M * ard6 13 61 710.4M 15.2M 8h00m41 3h27m25 105d01h34 etchdev64 1201620036 A * ard6 supports 32 bit and 64 bit environments 1201620039 M * marcfiu our machines usually have 250-500GB disks. 1201620060 M * marcfiu ok... ard6, you saved me a bunch of time. 1201620082 M * ard6 although native 64 bit all the way is just about 11% faster for a pure regex perl script 1201620104 M * marcfiu ard6: Are you in the UK (London)? If so, I might be there briefly in early April and will be glad to buy you a beer! :) 1201620112 M * ard6 heheh :-) 1201620142 M * ard6 You'd better buy Bertl_oO an alcohol free beer ;-) 1201620149 M * tuxbubling lol 1201620165 A * ard6 has seen too much drama with alcohol 1201620171 M * ard6 don't drink and type! 1201620191 M * ard6 rm -fr . /* 1201620198 M * ard6 someone really did that 1201620211 M * ard6 and it took him a week, and me a weekend to fix his server 1201620213 M * marcfiu ard6: indeed, we've been working together at times. 1201620229 M * marcfiu I hooked him up with OLPC, which is why the XO laptops run a vserver kernel. :) 1201620245 M * ard6 c00l 1201620258 M * marcfiu And we've done other work together, such as publish this paper on Vserver: http://www.cs.princeton.edu/nsg/papers/vserver_eurosys_07/ 1201620308 A * ard6 loves it when his tongue breaks over names 1201620323 M * tuxbubling (16:23:19) ard6: someone really did that << lol 1201620340 M * marcfiu as well as this OLS paper: http://www.kernel.org/doc/ols/2007/ols2007v2-pages-151-160.pdf 1201620380 M * ard6 but isn't princeton somewhere across an ocean? 1201620388 A * ard6 sits in amsterdam 1201620413 M * marcfiu indeed, we are in New Jersey, USA. 1201620437 M * marcfiu ard6: hmmm... not making it to Amsterdam any time soon. 1201620459 J * faheem ~faheem@cpe-071-077-007-143.nc.res.rr.com 1201620467 M * marcfiu But I'm traveling to South Africa at the end of Feb. Might fly using KLM via Amsterdam. 1201620476 M * marcfiu Though I doubt my stop over will be long enough. 1201620484 M * faheem Hi. Does anyone know of a 32 bit bigmem vserver enabled kernel in Debian? 1201620489 Q * _bjh_ Quit: leaving 1201620531 M * pmjdebruijn bigmem = PEA? 1201620546 M * pmjdebruijn the debian are crap anyways 1201620550 M * pmjdebruijn debian kernels* 1201620572 A * ard6 always compiles kernels by hand 1201620577 M * JonB me too 1201620586 M * faheem pmjdebruijn: Well, I need something that can handle 4 G memory. I think the bigmem can handle up to 64 G. 1201620590 A * ard6 doesn't trust anyone elses view on good .config files 1201620597 M * pmjdebruijn you don't need bigmem for 4GB 1201620607 M * pmjdebruijn any kernel should do 4GB 1201620610 M * faheem pmjdebruijn: In 32 bit I do. 1201620618 M * pmjdebruijn no 1201620621 M * pmjdebruijn you on't 1201620642 M * pmjdebruijn you get 840MB lowmem, en 3.3GB highmem = 4GB total 1201620649 M * faheem pmjdebruijn: Well, my direct experience contradicts your assertion. :-) 1201620668 M * pmjdebruijn it's possible to limit a kernel to 1GB, but distro rarely do that 1201620678 M * pmjdebruijn faheem: anyway, check your dmesg... 1201620679 M * faheem pmjdebruijn: I see 2.8 or so. 1201620696 M * faheem pmjdebruijn: I have 4 Gig. 1201620744 M * faheem I mean I see 2.8 with the 2.6.18 Debian vserver enabled kernel. 1201620744 M * pmjdebruijn faheem: please paste dmesg | grep LOWMEM 1201620748 M * pmjdebruijn and dmesg | grep HIGHMEM 1201620753 M * faheem pmjdebruijn: Ok. 1201620755 M * faheem Sec. 1201620788 M * faheem dmesg | grep LOWMEM : 896MB LOWMEM available. 1201620809 M * faheem dmesg | grep HIGHMEM : 3200MB HIGHMEM available. 1201620812 M * pmjdebruijn there 1201620818 M * pmjdebruijn 3200+896 makes? 1201620829 M * pmjdebruijn 4096MB=4GB 1201620831 M * faheem cat /proc/meminfo: MemTotal: 2845920 kB 1201620861 M * faheem pmjdebruijn: So, what am I doing wrong, then? 1201620956 M * pmjdebruijn I've never seen a machine with a MemTotal -eq RAM 1201621006 M * faheem pmjdebruijn: I don't follow you. 1201621189 J * dowdle ~dowdle@scott.coe.montana.edu 1201621233 M * pmjdebruijn faheem: it has something to do with a kernel/userland memory split if I remember correctly 1201621245 M * pmjdebruijn anyway, it should be proper behaviour 1201621257 M * pmjdebruijn but like I said, Debian kernel are real crap 1201621703 M * faheem I found http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=438086 1201621722 M * faheem which suggests that 4 G should work. Maybe there is a bios switch I need to turn on or something. 1201621949 M * pmjdebruijn no 1201621960 M * pmjdebruijn hardware with 4G or more RAM are not fully utilized. 1201622002 M * pmjdebruijn anyway, the best way to resolve this would be go to amd64 1201622240 M * faheem pmjdebruijn: Reasonable suggestion. However, some things like flash and java don't work in 64 bit. 1201622267 M * faheem I could run just a 64 bit kernel, with 32 userland, but the frigging nvidia drivers don't like that. 1201622282 M * faheem What I wouldn't give for a decent video card with free drivers. 1201622302 M * tuxbubling nvidia 64b drivers are working correctly 1201622314 M * tuxbubling and flash and java on 64b are ok too 1201622349 M * tuxbubling even if you have to use i386 flash plugin 1201622358 M * pmjdebruijn tuxbubling: why would you want to use vserver on a desktop machine? 1201622366 M * pmjdebruijn of course it's possible 1201622372 M * pmjdebruijn for flash there's gnash 1201622387 M * tuxbubling i was just answering faheem ;) 1201622392 Q * Slydder Quit: Leaving. 1201622395 M * pmjdebruijn tuxbubling: yeah sorry, I noticed (too late) 1201622404 M * tuxbubling pmjdebruijn: and actually i'm hosting my personnal server on my desktop ;) 1201622418 M * pmjdebruijn faheem: buy a decent cheap card like Radeon 9250, that will work out of the box with 64bit, without installing drivers, with 3D acel 1201622476 M * pmjdebruijn at least it used to work for me... now I have Intel Extreme 1201622492 M * tuxbubling i never had problems with 64b and nvidia drivers 1201622656 M * faheem tuxbubling: Not 64 bit per se. 64 bit kernel, 32 bit userland. 1201622682 M * faheem 64 works fine with nvidia, 32 bit works fine, mixture of 64 bit kernel and 32 bit userland, not. 1201622692 M * tuxbubling which userland? 1201622718 M * faheem tuxbubling: Sorry, don't understand the question. 1201622738 M * faheem I mean 64 bit kernel, on 32 bit system. 1201622785 M * tuxbubling an x86_64 kernel on an i686 system? 1201622833 M * faheem tuxbubling: Yes. 1201622860 M * tuxbubling i'm pretty sure that'd cause a lot of problems 1201622881 M * faheem Does flash and java really work on a 64 bit system? Pointers, links? 1201622882 M * tuxbubling just install a full x86_64 system and add your 32b apps that you need for compatibility 1201622933 M * tuxbubling java = use icedtea 1201622933 M * tuxbubling flash = use nspluginwrapper to make it work in your 64b firefox 1201622939 M * faheem tuxbubling: Debian doesn't like that; mixed system. 1201622943 Q * esa` Ping timeout: 480 seconds 1201622969 M * tuxbubling you still can run firefox 32b binaries with 32b plugins without problems 1201622995 M * tuxbubling faheem: then maybe think of changing distro.... x86_64 + 32b is not at all a problem 1201623033 M * tuxbubling btw maybe that's not the topic here :) 1201623085 M * faheem tuxbubling: Yes, sorry about the OT. I bought this here because I thought someone might know if there was a bigmem version of the Debian vserver kernel. 1201623092 M * faheem Going away now... 1201623110 M * faheem tuxbubling, pmjdebruijn: thanks for advice and suggestions. 1201623122 M * tuxbubling good luck with the stuff ;) 1201623279 Q * dowdle Quit: Konversation terminated! 1201623348 J * dowdle ~dowdle@scott.coe.montana.edu 1201623604 A * ard6 guesses faheem should switch to lenny or sid :-) 1201623616 M * ard6 it seems to just work on sid 1201623643 A * ard6 doesn't have any 64 bit systems in use which also need to run flash 1201623652 A * ard6 doesn't have any systems in use which also need to run flash 1201623685 Q * arachnist Ping timeout: 480 seconds 1201623968 J * esa bip@ip-87-238-2-45.static.adsl.cheapnet.it 1201624065 A * tuxbubling have :) 1201624080 A * tuxbubling has one ^^ 1201624193 M * faheem ard6: What just works? vserver + 4 gig? 1201624252 J * arachnist arachnist@088156187175.who.vectranet.pl 1201624265 M * ard6 faheem : 64 bits with 32 bits flash 1201624303 M * tuxbubling :) 1201624723 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa 1201625281 Q * tuxbubling Quit: Leaving. 1201626775 Q * JonB Ping timeout: 480 seconds 1201626897 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1201629309 Q * Julius Quit: Verlassend 1201629547 Q * pmenier Quit: Konversation terminated! 1201629606 Q * jescheng Remote host closed the connection 1201629625 J * jescheng ~jescheng@proxy-sjc-1.cisco.com 1201630151 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com 1201630261 J * JonB ~NoSuchUse@kg1-68.kollegiegaarden.dk 1201630685 J * Piet ~piet@tor.noreply.org 1201630742 J * hparker ~hparker@linux.homershut.net 1201631331 J * Flinx ~chatzilla@86.57.216.71 1201631346 M * Flinx Bertl_oO: Hi! 1201631378 M * Flinx Bertl_oO: I noticed there is no /proc/bus folder in guest os. 1201631447 M * Bertl_oO vprocunhide? 1201631451 M * Flinx Bertl_oO: This is because X server can't see device. 1201631473 P * marcfiu 1201631480 M * Bertl_oO well, I suggested that a few days ago, IIRC :) 1201631485 M * Flinx Bertl_oO: i just added all files from /proc and /dev to vprocunhide-files 1201631497 M * Flinx Bertl_oO: just for experiment 1201631507 M * Bertl_oO /dev doesn't make sense, and proc should be minimal 1201631509 J * Piet_ ~piet@tor.noreply.org 1201631572 M * Flinx Bertl_oO: I used find command ( find /proc > vprocunhide-files; find /dev >> vprocunhide-files ) 1201631585 M * Flinx Bertl_oO: What to unhide then? 1201631619 M * Bertl_oO as the name suggest, vprocunhide does only affect proc, not dev :) 1201631624 Q * giovanni Remote host closed the connection 1201631629 J * DLange ~dlange@p57A30B10.dip0.t-ipconnect.de 1201631733 M * Flinx Bertl_oO: In guest os when trying to run lspci i have error message first "pcilib: Cannot open /proc/bus/pci" and after this showing me devices. 1201631871 M * Flinx Bertl_oO: there is not /proc/bus in guest os 1201631888 M * Flinx Bertl_oO: This means something? 1201631913 Q * Piet Ping timeout: 483 seconds 1201632299 Q * Punkie Quit: Odcházím 1201632462 Q * yarihm Quit: Leaving 1201632601 Q * Flinx Quit: ChatZilla 0.9.80 [Firefox 2.0.0.11/2007112718] 1201633668 J * yarihm ~yarihm@vpn-global-dhcp1-75.ethz.ch 1201634050 M * arekm does anyone of you using 2.6.22 kernel? 1201634063 Q * ftx__ Ping timeout: 480 seconds 1201634072 M * arekm if yes then could you run this test program? http://pld.pastebin.com/f78cf53b1 1201634183 J * derjohn_mobil ~aj@dialin-145-254-195-034.pools.arcor-ip.net 1201634633 Q * derjohn_mobil Read error: Connection reset by peer 1201634762 M * Bertl_oO cpus = 0, ret = 0 1201634762 M * Bertl_oO ret = 128 1201634977 J * ftx ~ftx@dslb-084-060-156-137.pools.arcor-ip.net 1201635010 M * arekm on .22? 1201635023 M * Bertl_oO well, .22.9 actually 1201635046 M * arekm hm, could you try to run it several times? 1201635053 Q * ema Quit: leaving 1201635069 M * arekm I have .15 here and it hangs badly in D state 1201635117 M * Bertl_oO nah, worked fine here 1201635361 J * giovanni ~giovanni@89-97-35-68.ip15.fastwebnet.it 1201635365 J * giovanni_ ~giovanni@89-97-35-68.ip15.fastwebnet.it 1201635652 M * arekm Bertl_oO: what architecture? up/smp? 1201635673 M * Bertl_oO smp, x86_64 1201635702 M * arekm same here 1201635717 M * arekm well, I guess I'll have to do vanilla testing then ;/ 1201636039 J * larsivi ~larsivi@144.84-48-50.nextgentel.com 1201636353 J * Aiken ~james@ppp59-167-109-167.lns3.bne1.internode.on.net 1201637543 Q * yarihm Read error: Connection reset by peer 1201637548 J * yarihm ~yarihm@vpn-global-dhcp3-026.ethz.ch 1201637810 Q * yarihm Read error: Connection reset by peer 1201639904 J * Infinito argos@200-103-111-131.gnace701.dsl.brasiltelecom.net.br 1201640077 Q * FireEgl Read error: Connection reset by peer 1201640248 Q * ftx Ping timeout: 480 seconds 1201640307 Q * giovanni Remote host closed the connection 1201640307 Q * giovanni_ Remote host closed the connection 1201640308 J * Punkie ~punkie@home.pekelny.net 1201640379 J * ntrs ~ntrs@vs079.rosehosting.com 1201640408 J * lilalinux_ ~plasma@dslb-084-058-211-156.pools.arcor-ip.net 1201640457 M * ntrs When using mark/default to automatically start guests after a host reboot, will the starting stop if one of the guests stops booting due to an SSL certificate passphrase input wait, or will the other guests start regardless? 1201640488 Q * Infinito Ping timeout: 480 seconds 1201640514 J * jsambrook ~jsambrook@aelfric.plus.com 1201640520 P * jsambrook 1201640722 J * jsambrook ~jsambrook@aelfric.plus.com 1201640729 P * jsambrook 1201640783 Q * lilalinux Ping timeout: 480 seconds 1201641093 J * Linus ~Nuhks@bl7-135-169.dsl.telepac.pt 1201641243 J * ntrs_ ~ntrs@vs079.rosehosting.com 1201641266 Q * ntrs Read error: Connection reset by peer 1201641273 Q * DLange Quit: Bye, bye. Hasta luego. 1201642578 Q * lilalinux_ Remote host closed the connection 1201643585 J * ftx ~ftx@dslb-084-060-221-030.pools.arcor-ip.net 1201643807 M * ntrs_ anyone here? 1201644171 J * virtuoso ~s0t0na@ppp91-122-24-205.pppoe.avangard-dsl.ru 1201644236 Q * dna Quit: Verlassend 1201644383 Q * derjohn Ping timeout: 480 seconds 1201644405 J * derjohn ~derjohn@80.69.41.3 1201644433 Q * virtuoso_ Ping timeout: 480 seconds 1201645563 J * Infinito argos@200-103-111-131.gnace701.dsl.brasiltelecom.net.br 1201645608 M * ntrs_ How do I remove disk limits? 1201645710 M * JonB what does the great flowerpage say? 1201645733 M * JonB http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1201646582 J * derjohn_mobil ~aj@e180214150.adsl.alicedsl.de 1201646656 Q * JonB Quit: This computer has gone to sleep 1201646662 Q * ftx Remote host closed the connection 1201646712 Q * fatgoose Quit: fatgoose 1201646734 Q * Infinito Quit: Quitte 1201647341 J * yarihm ~yarihm@84-75-103-252.dclient.hispeed.ch 1201647573 J * _brent_ ~brent@mail.jivecommunications.com 1201647892 Q * Linus Remote host closed the connection 1201648364 Q * larsivi Quit: Konversation terminated! 1201648553 M * _brent_ i've got an application that's trying to set the TOS bit on its packets 1201648564 M * _brent_ i tried: echo CAP_NET_ADMIN > /etc/vservers/name/bcapabilities 1201648580 Q * meebey Remote host closed the connection 1201648587 M * _brent_ but the app is still whining that it can't set the TOS 1201648601 J * meebey meebey@booster.qnetp.net 1201648612 M * _brent_ am i going about this right and is there a better way that doesn't give full net admin control to a vserver instance? 1201648963 M * Bertl_oO what's TOS? 1201648971 M * _brent_ type of service 1201648978 M * _brent_ network type of service 1201648994 M * Bertl_oO ah, that kind of TOS 1201649006 M * _brent_ and i noticed that i was echoing the CAP into bcapabilities 1201649012 M * _brent_ which obviously is the wrong place 1201649027 M * _brent_ but putting it in ncapabilities doesn't work--the vps won't start 1201649033 M * Bertl_oO hmm, why? 1201649041 M * _brent_ says CAP_NET_ADMIN isn't valid 1201649043 M * Bertl_oO I mean, bcapabilities _is_ the right place 1201649052 M * _brent_ i tried NET_ADMIN, too 1201649075 M * _brent_ with CAP_NET_ADMIN in bcapabilities, my app still can't set the tos on its packets 1201649098 M * Bertl_oO so maybe CAP_NET_ADMIN isn't the capability controlling this? 1201649109 M * _brent_ http://64.233.167.104/search?q=cache:bqDqbiel98AJ:linux-vserver.org/Capabilities_and_Flags+%22type+of+service%22+linux-vserver&hl=en&ct=clnk&cd=1&gl=us&client=firefox-a 1201649115 M * _brent_ i found it there 1201649124 M * _brent_ i don't know that it's right or wrong 1201649155 J * Linus ~Nuhks@bl7-135-169.dsl.telepac.pt 1201649163 M * Bertl_oO hmm, yeah, that's from the mainline kernel, so that should be correct 1201649173 Q * mire Ping timeout: 480 seconds 1201649182 M * Bertl_oO try to strace -fF that application to see what goes wrong 1201649205 J * Infinito argos@200-103-111-131.gnace701.dsl.brasiltelecom.net.br 1201649234 M * Bertl_oO and what about changing that application to not setting TOS? does that fix the issue? 1201649260 M * _brent_ it doesn't whine about not being able to set tos, but that's not really a solution 1201649294 M * Bertl_oO so, it actually works, but you are missing the TOS stuff, yeah? 1201649308 M * _brent_ the application (asterisk) works, but can't set tos 1201649328 M * Bertl_oO does it set TOS depending on the connection or unconditionally? 1201649335 M * _brent_ unconditionally 1201649350 M * Bertl_oO so why not set it on the host with a mangle rule? 1201649391 M * _brent_ i'd prefer to have the host be as vanilla as possible 1201649420 M * Bertl_oO networking stuff belongs on the host, so a single iptables rule there would be appropriate, IMHO 1201649470 M * ntrs_ When using mark/default to automatically start guests after a host reboot, will the starting stop if one of the guests stops booting due to an SSL certificate passphrase input wait, or will the other guests start regardless? 1201649471 M * _brent_ so i could just tell the host to mangle the tos on all UDP traffic on the port range x:y 1201649475 M * Bertl_oO especially compared to the danger of adding CAP_NET_ADMIN :) 1201649482 M * _brent_ good point 1201649519 M * Bertl_oO ntrs_: depends, IIRC, guests not depending on eachother can be started in paralle 1201649538 M * Bertl_oO ntrs_: but ask daniel_hozac for the startup details 1201649552 M * ntrs_ what do you mean by "depending on each other"? 1201649586 M * Bertl_oO the guest startup is controlled by a 'make' file 1201649640 M * ntrs_ I see, but if I set all of them with the same mark/default file I guess that will not make them depend on each other, right? 1201649669 M * ntrs_ daniel_hozac, when using mark/default to automatically start guests after a host reboot, will the starting stop if one of the guests stops booting due to an SSL certificate passphrase input wait, or will the other guests start regardless? 1201649786 M * Bertl_oO without any dependancies set, and parallel startup in place, I would say no 1201649862 M * ntrs_ Bertl_oO, you can't answer an "or" question with a "no". Was that a no to the part before the or or after it? 1201649875 M * ntrs_ :) 1201649918 M * Bertl_oO well, actually I would have to answer that one with a 'yes' no? 1201649927 M * ntrs_ :) Lol 1201649939 M * Bertl_oO because of A or not A = true :) 1201649952 M * Bertl_oO but I was actually answering the one before that 1201650007 M * ntrs_ I see now. Thanks a lot. I like your humor a lot. 1201650031 M * ntrs_ Very subtle and scientific, and to the point. 1201650068 M * Bertl_oO thanks, I take that as compliment *G* 1201650080 M * ntrs_ Yes, it was meant as one. 1201650350 Q * Piet_ Quit: Piet_ 1201651066 P * _brent_