1114128001 M * Martin Serial console support appears to be buggered. 1114128004 M * Martin Fnar. 1114128007 M * Martin Take two. 1114128025 M * Bertl probably, I guess we do not really want to add too much to 2.4 1114128044 M * Martin fair 1114128172 A * albeiro > bed. have a good rest of night/day/whatnot ! 1114128223 M * Martin night 1114128237 M * romke nite albeiro :P 1114128249 M * albeiro romke: ! 1114128255 A * albeiro seen a ghost ;p 1114128257 M * romke albeiro: ! 1114128303 M * albeiro Bertl: i'm telling you, romke is a kind of ghost entity. anyway, off. 1114128311 M * romke albeiro: seen you there some time ago, but waited till you pay attention :P 1114128333 M * romke Bertl: don't listen to him, i'm just overworked :(... 1114128341 M * romke albeiro: :P 1114128345 M * albeiro i am here more or less about a year ;] 1114128350 M * albeiro *off* ;p 1114128354 M * Bertl night! 1114128418 A * romke is thinkin of preparing vserver-build subscript for preparing minimal PLD distro :> 1114128451 M * Bertl sounds very preparing to me ;) 1114128643 M * romke or i can go to sleep, but no, sleep is for wussies 1114128683 M * romke (hope albeiro won't read it :>) 1114128715 M * Bertl hehe 1114128790 M * romke Bertl: where change default shell on vserver enter ? 1114128799 A * romke dont like bash :P 1114128907 M * Bertl # shell 1114128907 M * Bertl Contains the pathname of the shell which will be used by the "vserver ... enter" command. 1114128916 M * Martin Hmmm... have really buggered this up. No kernel now boots :-\ 1114128938 M * Bertl romke: # /etc/vservers//shell (from the Flower page ;) 1114128965 M * Bertl Martin: what bootloader? 1114128979 M * Bertl and did you boot any 2.6 kernel yet? 1114128996 M * Bertl (because 2.6 kernels need other init tools ;) 1114129008 M * Martin Bertl: SILO 1114129009 M * Bertl s/init/module/ 1114129031 M * Martin I haven't used 2.6 on this machine before but I've got the tools needed 1114129054 M * Bertl proper initrd? 1114129056 M * Martin I'm not sure why my old kernel won't boot but the 2.6 one is an issue with the serial driver 1114129063 M * Martin initrd's fine 1114129076 M * Martin just rebooting with rescue disk 1114129094 M * Martin shouldn't take me too long... well... when it's POST'd 1114129707 M * Martin Removing the 2.6 kernel and hopefully returning things to normallity 1114130033 M * Bertl hmm, how would the 2.6 kernel interfere with the 2.4 one? 1114130041 M * Martin Don't know 1114130053 M * Martin Think my SILO config might be a bit odd 1114130382 M * Bertl doesn't like 2.6 kernels ;) 1114130404 M * Martin just persuaded it to play with 2.4.27 again 1114130451 M * Martin Think I'm going to try a pre-canned 2.6 first and then if that works, steal that config and build a 2.6+vserver 1114130537 M * Bertl well, what config did you take for the 2.6? 1114130549 M * Bertl (if you did take the 2.4 config, it's very likely to break) 1114130561 M * Martin And if the pre-canned don't work I should file a bug report anyway 1114130578 M * Martin Bertl: started with the config that came with the debina package 1114130620 M * Martin Then again that wasn't too good on 2.4 (serial drivers again IIRC) and so I ended up starting with a 'known good' for one of the binary kernels 1114130629 M * Martin They /should/ be the same as I understand it 1114130861 M * Martin There are many times when I have given thanks for the extensive checks the OpenBoot performs on reset and no doubt it has saved me much hasssel but now and again, taking more than a few minutes to reboot is annoying 1114130895 M * Bertl ah, btw, how do you install linux on a sparc64? 1114130911 M * Bertl is it able to boot from a debian CD? 1114130917 M * Martin Yep 1114130929 M * Martin Works just like x86 1114130946 M * Bertl what do you need to specify to the boot thingy? 1114130962 M * Bertl IIRC I once tried boot cdrom 1114130971 M * Martin `boot cdrom' normally 1114130985 M * Bertl hmm, do you need a 'special' cd for that? 1114131003 M * Martin Nope, usual burnt ones work. 1114131031 M * Bertl does it work with the cds this debian distro downloader generates too? 1114131042 M * Martin One some of the sparc (AKA sun4m) machines you need 512 bit block size CD-drives and CDs 1114131065 M * Martin I've downloaded stuff from cdimages.debian.org and it works 1114131072 M * Bertl 'jigdo' is the one I mean 1114131074 M * Martin OK, we have 2.6! 1114131082 M * Martin Pile of errors but 2.6 1114131096 M * Martin never tried with jigdo but don't see why not 1114131104 M * Martin XOPT $HOSTOPT $DOMAINOPT --secure $_SAVE_S_CONTEXT $__PKGSTATEDIR/$1.ctx $_CAPCHROOT $CHROOTOPT . $STARTCMD 1114131122 M * Martin Bollock. Problem not fixed. 1114131133 M * Bertl okay, then let's investigate it ... 1114131143 A * Martin slaps self 1114131145 M * Martin Ahem 1114131160 M * Martin When I've got a kernel with vserver patched in that is... 1114131170 M * Bertl ah, yeah ;) 1114131203 M * Martin Sorry for keeping you up - I fugure it's late where you are 1114131271 M * Bertl np, I'm not off to bed before 4am (usually) 1114131289 M * Bertl it's the time where I can get some things done ... 1114131302 M * Martin Likewise 1114131310 M * Martin I'm trying to sleep earlier 1114131318 M * Martin esp as I have to be giving a lecture at 10 1114131325 M * Martin but somehow it just doesn't happen 1114131892 A * romke is having problem with routing 1114132022 M * romke internet host -- 172.17.0.1 -- local net \- guests vservers (172.17.1.2 .3 etc...) 1114132047 M * Martin what appears to be the problem 1114132109 M * Bertl romke: vserver guests are not routed 1114132133 M * romke Bertl: yes, is there any way to fix this? 1114132134 M * Bertl so if you want them to reach the internet via some gateway you have to SNAT them to the public ip 1114132151 M * Bertl (they are not routed because they _are_ the host) 1114132241 M * Martin OK, kicked off compile of 2.6.8 with vserver patch. 1114132247 M * Martin Time for food I think 1114132256 M * Bertl 2.6.8? 1114132259 M * Martin Anyone else want some? :-) 1114132272 M * Martin 2.6.8 as provided by Debian 1114132281 M * Bertl hmm, okay ... so be it ... 1114132290 M * Martin Do you want something else? 2.6.11? 1114132394 M * Bertl well, I'd prefer the 2.6.11.7-vs2.0-pre1 1114132415 M * Bertl what version is the debian one? 1114132490 M * Martin Beyond 2.6.8 I don't know 1114132516 M * Martin If I can get this working I can quite easily try 2.6.11.7-vs2.0-pre1 later 1114132661 M * Bertl hmm, just because 2.6.8 before 2.6.8-15 IIRC is just broken (vserver wise) 1114132744 M * Martin Hmmm... not good - I'll see if I can find out 1114132848 M * Martin 2.6.8-15 1114132878 M * Bertl okay, so I presume it's the 1.9.5.x I ported ... (which should be fine) 1114132907 M * Martin opted for the Debian one so I get security nofication, all of their stabalising patches (v. useful on sparc) and cos it was there :-) 1114132942 M * Martin 1.9.5.3 1114132945 M * Bertl btw, do you know how to get yaboot/ybin to boot "compressed" kernels? or how to compress them ;) 1114132969 M * Martin not sure I've met yaboot 1114132988 M * Bertl hmm, okay, so a better choice for ppc (PowerBook G4) 1114133037 M * Martin doesn't ppc have a port of LILO? 1114133054 M * Martin the ppc machine I have are thin client and thus net boot 1114133099 M * Bertl hmm, well that explains why you didn't meet yaboot yet ;) 1114133106 M * Martin :-D 1114133122 M * Martin If it's Mac shouldn't it have OpenBoot - can you use that directly? 1114133144 M * Bertl well, no, that boots first (OF) and then I ahve to boot from a boot partition ;) 1114133222 M * Martin If you convert your kernel to a.out you /should/ be able to net boot and perhaps staright from disk 1114133228 M * Martin Well, works on sparc :-) 1114133279 M * Martin Having a dif through the patches supplied by Debian we seem to have patch-2.6.10-5-debian-vs1.9.4 and patch-2.6.11-rc3-vs1.9.4 as the opnly 2.6 patches 1114133300 M * Bertl hmm, not good ;) 1114133313 M * Martin Ah. 1114133325 M * romke Bertl: ha, it works now, i've changed MASQUERADE to SNAT :> (http://www.paul.sladen.org/vserver/archives/200205/0110.html) 1114133353 M * Bertl romke: yup, that's what I tried to tell you :/ 1114133372 M * Bertl Martin: http://vserver.13thfloor.at/Stuff/Debian/patch-2.6.8-15-vs1.9.5.x-4.diff 1114133382 M * Bertl this is the patch I did for the debian folks ... 1114133463 M * Martin Ola Lundqvist is in charge of the package and she's very prompt and helpful, mail her and I'm sure she'll add it in. 1114133474 M * romke Bertl: I've understanded you, but I do not know very well SNAT/DNAT/MASQUERADE issues, so google+SNAT+vserver helped me "a little" 1114133491 M * Martin It might be worth doing it via bugs@ ... YMMV 1114133529 M * Martin romke: In this case I can't think why MASQUERADE won't work and SNAT will but if it works... 1114133555 M * Bertl Martin: she? do you know something we don't? 1114133574 M * romke Martin: exactly as in mail from archives, SNAT works, MASQUERADE won't :| 1114133589 M * romke Martin: fortunatelly I've fixed ip so n/p 1114133601 M * Bertl Martin, romke: that's quite simple ... 1114133621 M * Bertl MASQUERADE is something which happens when you _forward_ packets 1114133635 M * Martin Bertl: ooops, my apologies 1114133643 M * Bertl i.e. you have to reserve a port to map the original port to ... 1114133645 M * Martin patch-2.6.8-15-vs1.9.5.x-4.diff.gz is in the unstable version 1114133676 M * Bertl now there is no _forwarding_ when the packet is generated on the host ... 1114133710 M * Bertl and therefore, masquerading will not happen .. but of course, you can SNAT a connection originating from the host 1114133713 M * Martin Hmmm... may be in testing aready 1114133734 M * Martin Ah, yes, in that case I see the differenc, I had got to topology wrong 1114133746 M * romke Bertl: heh, it seems simple when you tell it :P 1114133762 M * Bertl thanks ;) 1114133764 M * romke Bertl: what about SNAT from other hosts? 1114133785 M * Bertl will work, mut masquerade is the better choice here 1114133791 M * Bertl s/mut/but/ 1114133811 M * Bertl because of the different port option 1114133824 M * Martin Think we're nearly at the end of the compile... :-) 1114133894 M * Doener Bertl: got a minute for an apache question? 1114133898 M * romke Bertl: so, best choice is to divide local network computers and vserver guests into two separate nets, one to masquerade and one to snat - yes ? 1114133921 M * Bertl Doener: sure ... 1114133929 M * Bertl romke: yes, that's a good choice ... 1114133973 M * Bertl horray! 24 archs binutils support ;) 1114133994 M * Doener http://www.13thfloor.at/~doener/htaccess -- since i added the RewriteCond+RewriteRule, the second RewriteRule of the last block fails, while the third one still matches... any idea? 1114134093 M * Martin installing... 1114134131 M * Bertl Doener: that's for the wiki? 1114134202 M * Doener that's for a general php framework i'm working on... no idea if it's worth the effort at all, but who cares ;) 1114134225 M * Doener http://www.13thfloor.at/~doener/blah -- that's what I get... 1114134226 M * Bertl I'd say you add the ?_application=bla almost unconditionally 1114134244 M * Bertl which might easily result in a broken url ... 1114134274 M * Bertl let me check the rewrite log ;) 1114134288 M * romke Doener: \/ instead of / ? 1114134327 M * Doener how should that happen? that stuff ends up in $_GET (in PHP ;), it's not part of the filename 1114134346 M * Doener Bertl: that stuff is not on 13thfloor ;) 1114134354 M * Martin rebooting... 1114134369 M * Martin Place your bets now... :-D 1114134404 M * Bertl Doener: ah, okay ;) then please check the rewrite logs ;) 1114134588 M * Martin It works! 1114134592 M * Martin Well it boots 1114134597 M * Doener ah, didn't even know such a thing exists :) 1114134612 M * Martin Whole bunch of errors from /etc/modules - but that's my problem 1114134643 M * Bertl congrats! 1114134654 M * Martin "Bad address" for every entry in /proc and Bus error on vserver 1114134656 M * Martin Thanks 1114134670 M * Martin Also thanks for being patient 1114134730 M * Bertl the /proc actually might be the result of the 32/64 bit 1114134743 M * Bertl let me check that ... 1114134745 M * Martin http://pastebin.com/274415 1114134757 M * Martin for result of testme.sh 1114134813 M * Martin http://pastebin.com/274416 1114134824 M * Martin with results of vps and vtop :-( 1114134851 M * Bertl I guess you have absolutely no way to compile 64bit userspace? 1114134859 M * Doener hmm... that looks strange... http://www.13thfloor.at/~doener/foo 1114134905 Q * ndim Ping timeout: 480 seconds 1114134930 M * Martin Bertl: gcc -m64 :-D 1114134949 M * Bertl does that work with dietlibc? 1114134961 M * Bertl probably will, I guess ... 1114134985 M * Martin TBH, no idea. 1114134994 M * Bertl try to add it as compiler option ;) 1114135021 M * Bertl should work for the base tools 1114135085 M * Bertl Martin: the kernel is 64bit now, right? 1114135099 M * Martin All sparc64 kernels are 64 bit 1114135101 M * Bertl would it be possible to compile that 32bit? 1114135123 M * Martin No, sorry 1114135138 M * Bertl okay, was just a question ;) 1114135178 M * Martin Sorry. 1114135188 M * Bertl okay, give me a few minutes to get some test stuff ready ... 1114135212 J * ndim hun@helena.bawue.de 1114135226 M * Martin I'll start compiling with -m64... 1114135233 M * Bertl make that ... 1114135301 M * Bertl Martin: do you know sparc64 assembler? 1114135325 M * Martin Hmmm... util-vserver-0.30.206 doesn't build without vlan's. 1114135345 M * Martin sparc64 assembler - a bit rusty but yea. 1114135354 M * Bertl yeah, vlan-utils or so is required 1114135376 M * Bertl okay, the 'ta' is the trap used for the syscall 1114135379 M * Martin coming up... 1114135393 M * Bertl ta 0x6d actually 1114135423 M * Bertl glibc does a bcs,a,pt %%xcc, 1f after it returns ... 1114135435 M * Bertl (probably to fix up for the error path) 1114135453 M * Bertl then a 1114135459 M * Bertl sub %%g0, %%o0, %%o0; 1114135470 M * Bertl which as far as I can tell negates o0 1114135492 M * Bertl (the 1 label is right after that) 1114135535 M * Bertl so do I understand that correctly that the carry is set when the syscall succeeded, and the error code is 'positive' 1114135590 M * romke nite all, my battery goes off :( 1114135596 M * Martin night romke 1114135606 M * Bertl night romke! 1114135659 M * Martin I'm going to have ot look at the particular bit of code, can you give me a pointer? 1114135667 Q * ciphernaut Quit: 1114135706 M * Bertl hmm, no, but I can upload the extracted code ;) 1114135769 M * Bertl http://vserver.13thfloor.at/Stuff/syscall_sparc64.E 1114136011 M * Martin I /think/ so. 1114136038 M * Bertl okay, we'll se in a minute ;) 1114136059 M * Bertl s/se/know/ 1114136156 M * Martin Sorry, I'm being slow - where is CFLAGS in a ./configure'd Makefile? 1114136232 M * Martin Hmmm... now RTFM'ing 1114136236 M * Martin Obvious really 1114136239 A * Martin slap 1114136639 M * Martin On further exploration it appears that it was built against patch-2.6.8-15-vs1.9.5.x-4.diff.gz 1114136730 M * Martin Debian's online list of files per package was indeed out of date 1114137041 M * Bertl funny ... 1114137049 M * Bertl okay, I guess I have something to test with ;) 1114137124 M * Bertl Martin: http://vserver.13thfloor.at/Experimental/SYSCALL/syscall.h 1114137131 M * Martin Getting errors while trying to build 64 bit binaries, I'm going to build the whole thing first and then try again in 64 bit mode 1114137152 M * Bertl use this file with 1114137153 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/vdlimit-0.03.tar.bz2 1114137166 M * Bertl (i.e. replace the included syscall.h) 1114137183 M * Bertl then build the tool, and execute ./vdlimit -V 1114137187 M * Doener ok, i've managed to have the rewrite stuff recurse forever, had to wait for the oom killer to get rid of apache, the box was too slow for me to do it myself :) 1114137195 M * Doener guess i should go to bed ;) 1114137200 M * Bertl lol 1114137207 M * Doener good night folks! 1114137214 M * Bertl apache 'should' have a recrsion limit ... 1114137218 M * Bertl good night! 1114137255 M * Martin night 1114137264 N * Doener Doener_zZz 1114137309 M * Martin syscall sounds believable 1114137348 M * Martin Hmmm... at the moment it's not building at all 1114137365 M * Bertl output? 1114137367 M * Martin util-vserver that is 1114137372 M * Bertl ah, okay ;) 1114137374 M * Martin ld error 1114137387 M * Martin I've got the dependancies wrong I guess 1114137390 M * Bertl try the patched vdlimit ... 1114137391 M * Martin over to vd 1114137460 M * Martin OK, downloaded and unpacked the tarball, now replace syscall you say.. 1114137520 M * Bertl the syscall.h yes 1114137562 M * Martin error 1114137577 M * Bertl do you see a warning? 1114137596 M * Martin http://pastebin.com/274426 1114137605 M * Bertl k, tx 1114137622 M * Bertl yp, expected ... 1114137629 M * Bertl that's because you're 32 bit 1114137638 M * Bertl add the -m64 1114137651 M * Martin doing it as you said it :-) 1114137697 M * Martin http://pastebin.com/274427 1114137732 M * Bertl hmm, okay obviously I assumed that sparc_v9 is the 64bit marker ... 1114137745 Q * mcp Read error: Connection reset by peer 1114137746 M * Bertl let's do the following: 1114137755 J * mcp ~hightower@217.171.201.37 1114137761 M * Bertl echo | cpp -dM >/tmp/32.bit 1114137769 M * Bertl echo | cpp -m64 -dM >/tmp/64.bit 1114137783 M * Bertl diff -NurpP /tmp/32.bit /tmp/64.bit 1114137842 M * Martin http://pastebin.com/274428 1114137856 M * Bertl tx 1114137888 M * Bertl okay, sec 1114137891 M * Martin #define __arch64__ 1 ? 1114137940 M * Bertl okay, reload the syscall.h and replace it again 1114137973 M * Martin cc -m64 --static -g -Wall -O2 -Ivserver -o vdlimit vdlimit.c 1114137973 M * Martin strip vdlimit 1114137998 M * Martin ./vdlimit -V 1114137998 M * Martin vc_get_version: Operation not permitted 1114138070 M * Martin SYS_267(0, 0, 0, 0xffffffff, 0x1) = -1 EPERM (Operation not permitted) 1114138078 M * Martin I guess is the problem 1114138125 M * Bertl that looks good ... 1114138135 M * Martin :-) 1114138137 M * Martin woot! 1114138223 M * Bertl okay, let's see what kernel you really have ;) 1114138242 M * Bertl cat /proc/virtual/info 1114138257 M * Martin VCIVersion: 0001:0025 1114138257 M * Martin VCISyscall: 267 1114138257 M * Martin VCIKernel: 83000016 1114138290 M * Bertl okay, fine ... 1114138326 M * Bertl unfortunately they did disable debugging :( 1114138346 M * Bertl do you have the kernel tree (build) still available? 1114138352 M * Martin Yep. 1114138366 M * Bertl okay, please enable VSERVER_DEBUG 1114138384 M * Martin A full kernel rebuild is going to take a while 1114138398 M * Bertl that's why I asked, it should not rebuild too much 1114138444 M * Bertl but we'll continue in the meantime ... 1114138530 M * Bertl hmm, hmm, do you use the debian defaults? 1114138559 M * Bertl probably, please check if your capabilities module is loaded ... 1114138620 M * Martin mostly use debian defaults 1114138636 M * Martin capabilities module? Thought it was always built in 1114138645 M * hillct hmm 1114138649 M * hillct interesting 1114138650 M * Bertl yeah, I thought so too, but then there was debian ;) 1114138665 M * hillct looks like I have a version mismatch. 1114138708 M * hillct I'm using kernel patch 2.6.11.7-vs1.9.5.x5 and utils 0.30.206 1114138718 M * hillct when I attempt a vserver start I get: 1114138727 M * Bertl this one forces in the capability extension 1114138736 M * hillct chbind: vc_set_ipv4root(): Function not implemented 1114138754 M * Bertl hillct: arch, and did you disable the network legacy? 1114138763 M * hillct I disabled the legacy network support 1114138771 M * hillct this is x86_64 1114138772 M * Bertl well, then why do you wonder ;) 1114138781 M * hillct heh 1114138785 M * Bertl read the config option help ;) 1114138801 M * hillct I had my suspicions 1114138848 M * Martin CONFIG_VSERVER_HISTORY as well ? 1114138861 M * Bertl would not hurt, but it's not necessary 1114138878 M * hillct aha 1114138880 M * hillct For now, even recent tools use this interface of the x 1114138880 M * hillct x legacy API, so unless you know what you are doing, x 1114138880 M * hillct x leave that option enabled. 1114138881 M * hillct cute 1114138892 M * Bertl Martin: did you check regarding capabilities? 1114139000 M * Martin cat /proc/self/status lists capability info 1114139057 M * Bertl unfortunately that's no indication of anything :/ 1114139099 M * Bertl Martin: make sure to select the linux capabilities for the next kernel build (as kernel built in) 1114139214 M * Martin Where is the enable option - I don't ever recall seeing one (hence thinking it was in built) 1114139268 M * Bertl security models or so ... 1114139288 M * Bertl make menuconfig? 1114139297 M * Martin already doing so 1114139305 M * Bertl there is a search option 1114139308 M * Bertl ;) 1114139316 M * Martin CONFIG_SECURITY_CAPABILITIES 1114139321 M * Bertl yep 1114139394 M * Martin Only listed if CONFIG_SECURITY / Enable different security models is checked 1114139404 M * Martin and then not selectable 1114139421 M * Martin from the help i figure it's built in be default 1114139562 M * Bertl hmm .. maybe they changed it ... 1114139608 M * Bertl okay, please update the syscall.h once again, and now try without the -m64 1114139791 M * Martin Same output as before :-) 1114139854 M * Bertl means it compiles and gives -EPERM, right? 1114139874 M * Martin Yep 1114139878 M * Bertl great! 1114139891 M * Bertl means I got the 32bit version right too ;) 1114139898 M * Martin Woot! 1114139919 Q * lilo Quit: bbiab 1114139977 M * Martin Just upgrading ot (what should be) kernel + debugging 1114140068 M * Bertl but am I right that this machine is a v9 arch? 1114140077 M * Martin yep 1114140104 M * Martin Ultra I and Ultra II processors are sparc v9 1114140105 M * Bertl interesting, because the gcc obviously doesn't know about that (without some -mv9 option or so) 1114140119 M * Martin Ultra III are sparc v9.b 1114140138 M * Bertl what caps does you /proc/self/status show? 1114140142 M * Martin all of which are sun4u architectures 1114140177 M * Martin CapInh: 0000000000000000 1114140177 M * Martin CapPrm: 00000000fffffeff 1114140177 M * Martin CapEff: 00000000fffffeff 1114140197 M * Bertl so you should have CAP_CONTEXT ... strange ... 1114140218 M * Martin s_context: 0 1114140218 M * Martin ctxflags: none 1114140218 M * Martin initpid: none 1114140218 M * Martin ipv4root: 0 1114140218 M * Martin ipv4root_bcast: 0 1114140308 M * Bertl well, we should know with the debuging enabled ... 1114140334 M * Martin In POST 1114140390 M * Bertl btw, the G4 should be 64bit too, no? 1114140430 M * Bertl (powerpc, sorry for jumping topics ;) 1114140435 M * Martin No, G4 is 32 bit (excluding the 128 bit AltiVec unit), G5 is 64 bit 1114140449 M * Bertl ah, okay, that clarifies ... 1114140451 M * Martin No worries - I knew what you meant 1114140507 M * Martin All of the POWER series are designed to do both so the IBM ones all then to be 64 bit and (up until the G5) they tended to give Apple the 32 bit versions 1114140560 M * Martin OK kernel booted and it /should/ have debugging enabled. ANy way to tell for sure? 1114140586 M * Bertl echo 1 >/proc/sys/vserver/debug_switch 1114140610 M * Bertl then try the vdlimit -V again ... 1114140617 M * Bertl and check the klog output 1114140668 M * Martin -bash: /proc/sys/vserver/debug_switch: No such file or directory 1114140678 M * Martin I guess that means I got it wrong. 1114140680 M * Martin Hmmmm... 1114140684 M * Martin I'll go check 1114140701 M * Bertl hmm, check /proc/sys/vserver 1114140707 M * Bertl doesn't it exist at all? 1114140710 M * Martin Nope 1114140718 M * Bertl what about /proc/sys? 1114140724 M * Martin Yep. 1114140749 M * Martin Have /proc/virtual 1114140750 M * Bertl hmm, hmm, grep DEBUG .config ? 1114140790 M * Martin CONFIG_VSERVER_DEBUG=y 1114140813 M * Bertl strange ... 1114140887 M * Bertl vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o 1114140925 M * Martin Eh? 1114140946 M * Bertl means it will compile in the sysctl module when CONFIG_VSERVER_DEBUG is set 1114140970 M * Bertl which in turn makes the /proc/sys/vserver/* entries 1114140982 M * Bertl unless debian modified soemthing there ... 1114141000 M * Bertl what's in /proc/sys ? 1114141036 M * Martin abi debug dev fs kernel net proc vm 1114141093 J * lilo ~lilo@lilo.usercloak.oftc.net 1114141107 M * Bertl hmm, strange ... well, you could try with a kernel boot option 1114141194 M * Martin sure, which flag? 1114141201 M * Bertl sec, checking that ... 1114141308 M * Bertl hmm, no that's not aailable with this kernel :/ 1114141420 M * Martin So where does that leave us then? 1114141426 M * Bertl but that the /proc/sys/vserver is not there smells fishy 1114141485 M * Bertl well, I would suggest to compile a 2.6.11.7-vs2.0-pre1 with debug options enabled ... 1114141505 M * Martin OK... 1114141510 M * Bertl and see if a) the debug stuff works, and b) the tools work with that 1114141548 M * Bertl because as I said, except for the /proc and dlimit calls, everything should be fine even with mixed archs 32/64 1114141583 M * Bertl and the -EPERM is strange as a result to the syscall command interface version ... 1114141604 M * Bertl (unless you're missing CAP_CONTEXT ;) 1114141755 M * Martin In short - it should work but there are a few potential nasties in both kernel and userspace 1114141796 M * Bertl I designed the 'new' interface (used in 2.6) to be arch agnostic 1114141815 M * Bertl unfortunately I made an error, in not caring about the pointers ... 1114141834 M * Bertl and we use pointers to specify a file path 1114141857 M * Bertl (because I didn't want to put them in the syscall struct) 1114141869 M * Martin Fair 1114141876 M * Bertl so we are now considering two options 1114141890 M * Bertl (since we know that there is an issue ;) 1114141892 M * Martin (just setting up 2.6.11.7-vs2.0-pre1) 1114141930 M * Bertl 1) make a new revision (syscall interface is versioned), with an array, instead of the pointer 1114141955 M * Bertl 2) add the compatibility layer stuff to fix up the pointers ... 1114141986 M * Martin Given that 1 is a 'cleaner' way of doing things and v2 isn't out yet I'd opt for 1. 1114141992 M * Martin But I don't have to do the work... 1114142011 M * Bertl it's not the work, it's more the 'backwards' compatibility 1114142034 M * Bertl because if we look at x86_64, they might have 32 bit userspace and use older tools :( 1114142052 M * Bertl (otherwise 1 would be the logical choice ;) 1114142055 M * Martin Hmmm... yea, tricky. 1114142344 M * Martin 2.6.11.7-vs2.0-pre1 is compiling 1114142393 M * Martin I owuld really like to get this working - it would make my life my easier. 1114142423 M * Martin However I am very short on time and the server has to go to the server room today at 11 1114142439 M * Martin I do have other UltraSPARCs to work on though. 1114142450 M * Martin But it limits oding kernel compiles, etc. 1114142478 M * Martin What would seem like a good `roadmap' ? 1114142531 M * Bertl what's your local time now? 1114142541 M * Martin 5:00 AM 1114142548 M * Martin I have to lecture at 10 1114142579 M * Martin I didn't necessarily mean today - althought it would be cool if we could 1114142591 M * Bertl well, I'd say let's give the 2.6.11 a try, if that gives reasonable results ... 1114142601 M * Martin Sounds good. 1114142615 M * Martin Now I'm just wishing I'd fitted the extra 4 processors... 1114142617 M * Bertl (you can probably live without the proc and dlimit stuff) 1114142618 M * Martin :-) 1114142635 M * Martin that was my feeling as well 1114142645 M * Bertl and we can add the compatibility layer tomorrow or the day after ;) 1114142660 M * Martin should be OK, although... 1114142687 M * Martin I'm supposed to be traveling to Germany, leaving 4am Sunday 1114142698 M * Martin which will somewhat limit my time for the next two weeks 1114142762 M * Martin OK, more questions - what do you think the time frame of getting any changes needed propagated to stable releases of vserver-patch and util-vserver 1114142800 M * Martin For a whole bunch of reasons it would be a lot easier if I could just use the Debian provided packages 1114142809 M * mugwump ah, the curse of debian 1114142812 M * Martin (on to compiling modules) 1114142822 M * mugwump just set up your own package source, Martin 1114142840 M * mugwump (dpkg-scanpackages)++ 1114142843 M * Martin mugwump: I happen to like it and Ola is very responsive 1114142855 M * Bertl he is? 1114142864 M * mugwump which Ola is this? ;) 1114142864 M * Martin I could od my own packages but that becomes a pain to manage 1114142882 M * Martin Bertl: has been every time I submitted bugs - turn round in a couple of hours 1114142895 M * Martin Ola Lundqvist 1114142906 A * Bertl is now sure there are at least two of them ;) 1114143697 J * Hmmmm ~Hmmmm@202.144.29.194 1114143770 M * Hmmmm hello, anyone home? 1114143922 M * Bertl yep, I'm home ;) 1114144094 M * Bertl Martin: ppc syscall works too ;) just tested on my G4 ;) 1114144270 M * Hmmmm guys is there a good benchmark of vservers available somewhere on the web? 1114144313 M * Bertl what would you like to benchmark? the overhead is below 2% ... 1114144338 M * Hmmmm i want to benchmark the memory and CPU usage for Vservers 1114144347 M * Hmmmm i wana see how many vservers can sit on one box 1114144355 M * Martin Bertl: Cool vserver on PPC! 1114144382 M * Bertl Hmmmm: ah, well, that's something different ... and really depends on your vservers setup and the guests 1114144402 M * Hmmmm ok 1114144406 M * Bertl Hmmmm: lycos did put in a test run more than 100 guests on a dual PIII 1114144411 M * Hmmmm basically what i wana do is 1114144421 M * Hmmmm i wana have 5-6 vservers on each box 1114144434 M * Hmmmm and have some 20 odd users compiling code inside each vserver 1114144446 M * Hmmmm i wana know beforei try it if its gfeasable 1114144472 M * Hmmmm Bertl, any tips? 1114144473 M * Bertl it will work fine ... of course you will not have full compiling performance in each guest 1114144502 M * Hmmmm thats ok 1114144518 M * Hmmmm so would u know of any benchmark of a similar experiment? 1114144519 M * Bertl i.e. if you have a 2GHz machine ... then if just one guest compiles, it will feel like native 1114144537 M * Bertl actually you will lose less than 2% 1114144553 M * Hmmmm kewl 1114144557 M * Bertl now if there are 6 guests compiling at the same time, it really depends on your setup 1114144572 M * Martin For what it's worth I've never noticed the overheads 1114144575 M * Bertl basically you can expect the same as with 6 compile runs on the host 1114144607 M * Bertl i.e. the compile runs will add pressure to the file cache and depending on your memory it will slow down, or not ;) 1114144624 M * Hmmmm ah ic 1114144643 M * Hmmmm but thats normal right 1114144648 M * Bertl so basically you can think of it as a 'normal' linux host, just with separated compartments 1114144649 M * Hmmmm vserver or not 1114144658 M * Hmmmm kewl 1114144665 M * Hmmmm are u by chance a vserver develloper? 1114144668 M * Bertl that's the idea behind linux-vserver ... 1114144683 M * Bertl Hmmmm: well, yes I do some development ... 1114144705 M * Hmmmm vserver is a great tool 1114144720 M * Hmmmm is it written is C? 1114144738 M * Bertl it basically is a kernel patch, plus a bunch of sophisticated tools in C 1114144744 M * Hmmmm kewl 1114144752 M * Hmmmm i looked at chroot 1114144758 M * Hmmmm but vserver rocks 1114144769 M * Hmmmm how does is compare to bsd jails? 1114144787 M * Bertl I#d say it's very similar in 'how' it does it's work 1114144800 M * Bertl but the linux-vserver jails have much more 'features' 1114144814 M * Bertl e.g. resource limits and such stuff 1114144814 M * Hmmmm ya i noticed that 1114144832 M * Martin Also see zones in Solaris 1114144881 M * Hmmmm ok 1114144917 M * Hmmmm thanks a ton for ur advice 1114144929 M * hillct isn't it domains is solaris? 1114144930 M * Hmmmm i'll be back if i need more help 1114144977 M * Bertl k, cya! 1114145015 M * Bertl hillct: nobody really knows how it is called .. not even the solaris folks ... (slices, zones, domains ... choose thy name ;) 1114145076 M * hillct I recall hearing both Domains and Slices fro mtheir marketing guys 1114145309 Q * Hmmmm Quit: Leaving 1114145751 M * Martin Rebuilding kernel (again) after I made a mess of the last one 1114145780 M * Martin Hopefully should have 2.6.11.7-vs2.0-pre1 in a bit 1114145796 M * hillct Bertl what's the new mechanism for handling port passthrough that replaced the init scripts v_* ? 1114145807 M * hillct is it somewhere in the config tree? 1114145818 A * hillct roots around for documentation 1114145834 M * Bertl hmm ... v_ssh was replaced by configuring the host's sshd properly ... 1114145846 M * Bertl but maybe you are referring to the ngnet ? 1114145856 M * hillct I was 1114145867 M * hillct but my imediate problem is sshd 1114145895 M * hillct which is solved by listening to the specific host IP right? 1114145910 M * hillct or is there a more flexible solution? 1114145953 M * Bertl no, that's basically it ... if that is not feasable (for whatever reason) I always suggest to change the port number 1114146011 M * hillct bah 1114146023 M * hillct static assignment it is then... 1114146736 M * Bertl ah, fixed a minor bug for ppc ;) 1114146771 M * Bertl Martin: how is the status? 1114146816 M * Martin still compiling, 20 minutes in, onto modules now 1114146824 M * Martin Shouldn't be too much longer 1114146840 M * Bertl k, getting breakfast now ;) 1114146850 M * Martin That's a fine point. 1114146872 M * Martin It's getting light where I am and soon the shop will be open... 1114146887 M * Martin with luck I could toast our success :-) 1114146892 M * Martin Hmmm.... toast.... 1114147383 M * Martin Built! 1114147531 M * Bertl the toast or the ekrnel *G* ? 1114147546 M * Martin kernel 1114147552 M * Martin Shame it's not both 1114147576 M * Martin And despite the amount of heat this box makes I'm sure I'm rebooting the computer and not the toaster. 1114147583 M * Martin ... I think ... 1114147586 M * Martin :-) 1114147746 M * Martin Boots 1114147758 M * maharaja Bertl: after some straces and during sleep, i figured what the problem was 1114147758 M * Martin Still gives "Bad address" for all of /proc 1114147769 M * eyck wow, you are all still up? 1114147771 M * Martin No complaint about starting vservers 1114147806 M * Martin eyck: If that was aimed at me - then yes. Server has to go to server room at 11. 1114147807 M * maharaja Bertl: i saw that my vservers tried to open a network connection to 10.1.1.61 and i did not know where this came from 1114147850 M * Bertl Ma 1114147859 M * maharaja Bertl: 10.1.1.61 is my ldap server which i use for storing the users and i changed the ip which causes linux to hang on boot 1114147861 M * Bertl +rtin: the proc issue is expected 1114147902 M * Bertl maja: cool! 1114147904 M * Martin Have /proc/sys/vserver - with debugging switches in! 1114147924 M * Bertl could you try the vdlmit -V ? 1114147929 M * Martin Out of interest - what is the overhead of the debugging patch 1114147952 M * Bertl we did not measure it yet, but it should not be that high ... 1114147953 M * Martin vc_get_version: Operation not permitted 1114147976 M * Bertl so this is still there .. interesting ... 1114147976 M * Martin SYS_267(0, 0, 0, 0xffffffff, 0x1) = -1 EPERM (Operation not permitted) 1114147994 M * Martin However it didn't throw a bus error on chbind 1114147997 M * Bertl also with the 64bit version? 1114148006 M * Bertl (the EPERM) 1114148016 M * Martin I'll just check 1114148073 M * Martin Yep, same again. 1114148120 M * Martin vtop and vps both say vc_new_s_context(): Bad address 1114148156 M * Martin Hmmm... vserver still throws bus errors as well but by the looks of things they're further on than before 1114148166 M * Bertl okay, so at least we know that the debian kernel is not to blame for all of it ;) 1114148188 M * Bertl could you now enabled thedebug_switch 1114148205 M * Martin The weirdness - no, althought this one now does have the expected things in /proc/sys 1114148206 M * Bertl echo 1 >/proc/sys/vserver/debug_switch 1114148211 M * Martin Sure 1114148222 M * Bertl and then use the vdlimit -V 1114148347 M * Martin ./vdlimit -V 1114148347 M * Martin vxD: vc: VCMD_00_0[0], 0 1114148347 M * Martin vc_get_version: Operation not permitted 1114148353 M * Martin ./vdlimit64 -V 1114148353 M * Martin vxD: vc: VCMD_00_0[0], 0 1114148353 M * Martin vc_get_version: Operation not permitted 1114148376 M * Bertl okay, that's quite fine ... so I don#t udnerstand why you get that .... 1114148384 J * grecea ~grecea@h-195-22-237-74.mdl.net 1114148416 M * Bertl guess we have to postpone this (until my brains work again ;) 1114148441 M * Bertl you might look at the kernel/vserver/switch.c 1114148453 M * Martin That sounds fine to me 1114148475 M * Martin How do you think we should carry this forward? 1114148478 M * Bertl there is a single check (against CAP_CONTEXT) which should not fail ... 1114148490 M * Bertl it's the only one returning EPERM .. 1114148505 M * Bertl so I suspect the capability checks to fail somehow ... 1114148520 M * Martin do you want the contents of /proc/self/status? 1114148532 M * Bertl guess it's ffffffef 1114148538 M * Bertl (well, the caps) 1114148557 M * Martin Ummm... 1114148563 M * Martin I have a confession to make 1114148568 M * Bertl hmm? 1114148570 M * Martin I was getting mixed up 1114148580 M * Martin I ran vdlimit as user, not root 1114148585 M * Martin # ~martin/vdlimit-0.03/vdlimit -V 1114148585 M * Martin vxD: vc: VCMD_00_0[0], 0 1114148585 M * Martin version: 0001:0025 1114148587 M * Bertl LOL 1114148591 M * Martin # ~martin/vdlimit-0.03/vdlimit64 -V 1114148591 M * Martin vxD: vc: VCMD_00_0[0], 0 1114148591 M * Martin version: 0001:0025 1114148600 M * Martin I *officially* suck 1114148615 M * Bertl well, that explains the EPERM ;) 1114148629 M * Martin OK, two problems down then. 1114148634 M * Bertl okay, so we were looking for the wrong issues ;) 1114148644 M * Martin Sorry 1114148647 M * Bertl np 1114148652 A * Martin hangs head in shame 1114148655 M * Bertl shit happens now and then ;) 1114148673 M * Bertl after all we got a syscall implementation for ppc, sparc and sparc64 ;) 1114148689 M * Martin :-D 1114148748 M * Bertl okay, enough for me for today ... let's continue next time ... 1114148753 M * maharaja *off to study* 1114148773 M * Martin For me, that will have to be at least a week and a half. 1114148790 M * DaPhreak morning guys :) 1114148791 M * Bertl okay, so be it ... 1114148795 M * Bertl thanks for testing on non x86 archs ... 1114148801 M * Martin Thank you very much for all of your help 1114148808 M * Bertl you're welcome! 1114148812 M * Martin I'll see what others I can come up with. 1114148854 M * Martin See you (all) in a week and a bit then. 1114148854 M * Bertl the bus error issues look like 32/64 bit issues 1114148854 M * Bertl but I don#t see how the ipv4 interface should cause that 1114148858 M * Bertl anyway, night folks! cya 1114148864 M * DaPhreak night ;) 1114148868 M * Martin night 1114148868 N * Bertl Bertl_zZ 1114148874 Q * Martin Quit: using sirc version 2.211+KSIRC/1.1 1114148884 M * DaCa DaPhreak: make up your mind :) 1114148921 M * DaPhreak ah .. to early DaCa ;) and access (yeah i know) is still stabbing directly into my face ... 1114148969 P * erwan_taf Leaving 1114149956 M * hillct I imagine this too will turn out to be a stupid question but I'll ask anyway 1114149986 M * hillct in the standard fedora shutdown process there's a script that unmountsw all mounted fileststems 1114150033 M * hillct how does vserver incorporate the fstab in the config tree and a copy if it exists in /etc in the vserver itself? 1114150077 M * hillct the problem I'm trying to understand/solve is upon shutdown the vserver attempts to unmount /tmp and /proc 1114150103 M * hillct because they're listed in the fstab in the config tree 1114150133 M * hillct and unmounting these fileststems of course, fails 1114150307 J * stegbth ~stegbth@217.74.1.161 1114150317 M * stegbth guten morgen 1114150368 M * hillct hey 1114150434 M * stegbth i am playing with a already vserver (running on 2.4.30 with vserver 1.2.10 on debin woody) to migrate to an other machine (vserver 1.9.5rc4 on 2.6.11 debian sarge) 1114150459 M * stegbth i rsynced /var/lib/vserver/vsuse8 and when i start it looks good 1114150466 M * stegbth the ip-alias get created 1114150479 M * stegbth but when i try vserver vsuse8 enter i get 1114150501 M * stegbth chcontext: vc_new_s_context(): Invalid argument 1114150505 M * stegbth any ideas? 1114150531 M * stegbth i have two debian vserver running 4 test on the debian sarge without a problem 1114151268 M * stegbth i have to leave 4 now, will come back later 1114151277 Q * stegbth Quit: Verlassend 1114155100 M * kalou_learning hillct: In my vservers (2.4.29-vs1.2.10), the /proc filesystem gets mounted automatically at vserver start, no need for a line in /etc/fstab 1114155110 N * kalou_learning kalou 1114155193 Q * hillct Ping timeout: 480 seconds 1114156862 J * prae ~prae@84.14.106.134 1114157214 M * DaPhreak morning prae 1114158213 M * Hollow lo 1114158280 M * prae DaPhreak: yarglala :) 1114158754 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1114159043 M * DaPhreak ah, Hollow ;) 1114159049 M * Hollow hoi ;) 1114159093 M * DaPhreak skipped school ? 1114161179 M * Hollow parts .. 1114161180 M * Hollow :P 1114161547 M * eyck yeah, skool sukz 1114161703 Q * flock Remote host closed the connection 1114161755 M * Hollow yup 1114161770 M * Hollow so i thought better fix some bugs :P 1114171992 M * wurd can someone give me a hand today ? 1114171999 M * wurd still the same problem... 1114172022 M * wurd daniel_hozac maybe i should try your method, how do i know which packages to install? 1114172203 J * yarihm ~yarihm@vpn-global-dhcp2-21.ethz.ch 1114172261 Q * yarihm Quit: 1114172288 J * knoppix_ ~knoppix@dsl-213-023-140-036.arcor-ip.net 1114173180 J * monrad ~monrad@213083190130.sonofon.dk 1114174999 M * wurd anybody here ? 1114177415 M * BWare mostly there 1114177540 M * wurd would you know what this error means ? 1114177545 M * wurd -bash-3.00# ifconfig eth0 10.1.99.9 1114177545 M * wurd SIOCSIFADDR: Permission denied 1114177545 M * wurd SIOCSIFFLAGS: Permission denied 1114177562 M * wurd (I am inside my vserver while i execute this command) 1114177567 M * BWare that you're not allowed to configure the ethernet device 1114177596 M * BWare AFAIK you can only configure network devices from the host - default 1114177628 M * wurd you mean outside my vserver? 1114177638 M * BWare yep 1114177646 M * wurd uh... 1114177702 M * wurd well how am I supposed to communicate with my vserver then ? 1114177720 M * BWare I don't know your setup but mostly a vserver uses a named aliias as network device, eg: eth0:vserver1 1114177735 M * wurd thats what i did at first 1114177740 M * BWare through the aliased network device 1114177751 M * wurd but when lets say i try communicating with my vserver through ssh 1114177757 M * wurd i end up on my host's ssh server 1114177785 M * wurd the alias only gives eth0 a second name, right ? 1114177794 M * BWare that's probably because the host's ssh is listening on 0.0.0.0:22 ( check with netstat -nl) 1114177804 M * BWare yes, it does 1114177856 M * BWare to fix that; modify /etc/ssh/sshd_config and make ListenAddress IP.OF.YOUR.HOST 1114177868 M * BWare then restart sshd 1114177924 M * BWare to be on the safe side, modify /vservers/vservername/etc/ssh/sshd_config as well, but make that ListenAddress IP.OF.YOUR.VSERVER 1114177947 M * BWare and restart the vserver's sshd too 1114178119 M * wurd thanks 1114178191 M * BWare yw 1114178980 Q * monrad Quit: Leaving 1114179094 Q * knoppix_ Quit: Verlassend 1114181427 N * Bertl_zZ Bertl_oO 1114182222 J * kalou_ ~kalou@AToulon-201-1-30-46.w81-48.abo.wanadoo.fr 1114182341 Q * kalou Ping timeout: 480 seconds 1114182416 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.11 1114187079 Q * prae Quit: Client exiting 1114187420 J * hillct ~hillct@client200-5.dsl.intrex.net 1114188802 M * kevinp wurd: I have created a page that has all of the packages for a minimal install on FC3 at http://linux-vserver.org/FC3_min_install 1114188822 J * DuckMaster ~duckx@195.75.27.158 1114188921 M * kevinp wurd: put the contents into 01 and remove 02 and 03 1114188929 M * kevinp wurd: then run this command: 1114188936 M * kevinp vserver servername build -m yum --hostname domain.com --interface domain=eth0:192.168.0.136/24 --initstyle sysv --context 50 -- -d fc3 1114188955 M * kevinp And you will have a vserver that looks a lot more like what you are used to. 1114189037 M * kevinp Oh, and the install will take quite a while to download those packages 1114189235 M * hillct kevinp '-m yum' ? you can now use yum instead of apt to perform installations? 1114189338 M * kevinp if you have it setup right 1114189362 M * kevinp I was able to get it to work with 205 tools 1114189426 M * kevinp the 01 file I was referring to is in /usr/lib/util-vserver/distributions/fc3/pkgs (for FC3) 1114189457 M * hillct I'd like to do the same for FC2 1114189457 M * kevinp See http://linux-vserver.org/CentOS_HowTo to set it up for CentOS 1114189471 M * hillct should be similar I guess 1114189475 M * micah what is CentOS? 1114189483 M * kevinp then you should be able to do the same thing just in fc2/pkgs 1114189490 M * hillct RHEN with stuff ripped out 1114189494 M * micah is that the re-branded redhat? 1114189496 M * micah oh yeah 1114189509 M * kevinp And you would probably need to create a new list of pkgs that are for FC2 1114189511 N * Doener_zZz Doener 1114189518 M * Doener evening! 1114189523 M * kevinp CentOS == RHEL - support costs 1114189533 M * kevinp evening! 1114189608 M * kevinp To get the FC3 min install packages, I just installed FC3 on a test box, choose minimal (unchecking everything else first) and then did a yum update and then took the output of rpm -qa 1114189732 M * hillct you may be able to scale down that list by excluding dependencies 1114189840 M * kevinp true, you should also remove packages that you don't want installed 1114189866 M * kevinp My purpose was to just put out a list of the "minimal install" in FC3's install 1114189885 M * hillct ah 1114189888 M * kevinp as opposed to the very minimal install that the tools use 1114190245 J * DuckKing ~Duck@dyn-83-157-160-229.ppp.tiscali.fr 1114190338 M * hillct kevinp is there some introductory documentation of distribution management out there? the flower page talks about the config tree but i didn't see much/anything on how to achieve what yo outlined 1114190345 M * hillct er you 1114190385 M * hillct ie: what's the significance of the numbered files? I imagine each gets passed to yum in sequence 1114190432 M * hillct but in the case of /usr/lib64/util-vserver/distributions/fc2/pkgs it seems the files are in an odd order 1114190445 M * hillct 01 contains just glibc 1114190462 M * hillct 02 contains --reinstall 1114190462 M * hillct filesystem 1114190666 Q * duckx Ping timeout: 480 seconds 1114190899 M * kevinp Yeah, I don't know of any good documentation, I came up all of that through trial and error 1114190918 M * kevinp And all I can say is that "it worked for me" :) 1114190939 M * kevinp And some help from others here as well ... 1114190959 M * hillct folks here have been extremely helpful 1114190993 M * hillct I just dislike having to ask so many questions about things I should be able to do by reading documentation 1114191150 M * kevinp well, I documented what I learned, I guess that's all we can do 1114191216 M * hillct yup 1114191226 M * hillct thanks. It should be quite helpful 1114191288 M * hillct ah 1114192959 N * kevinp kevinp|gone 1114193012 P * kevinp|gone Leaving 1114193213 M * Doener hm, now that i finally believe that he's right regarding the minimality of the fc3 build, he leaves ;) (i just read the mail from enrico on the ml) 1114193348 M * hillct ? 1114193359 M * hillct I really should subscribe 1114193375 M * hillct but I really get enough mail as it is. Got a URL to the archive? 1114193411 M * Doener it's listed on http://linux-vserver.org ... don't remember it right now 1114193593 A * hillct browses 1114193595 M * hillct http://www.theregister.co.uk/2005/04/22/email_destroys_iq/ 1114193606 M * hillct this is one reason I don't need more email :) 1114193630 J * kjo nobody@p5484F71A.dip.t-dialin.net 1114193902 J * kjo_ nobody@p5484F71A.dip.t-dialin.net 1114193930 Q * kjo_ Quit: 1114194152 P * hillct 1114194619 J * nayco ~nayco@82.249.44.48 1114194631 M * nayco 'llo 1114194659 Q * nayco Quit: 1114195060 M * Snow-Man hah. 1114195085 M * Snow-Man According to Enrico using Debian in their vservers probably means using glibc is just fine. 1114195134 M * Snow-Man From a security standpoint anyway. 1114197244 M * daniel_hozac because Debian doesn't have any fancy tools :) 1114197284 M * eyck hmm, 1114197297 M * eyck but there is this new Girl Debian movements 1114197310 M * eyck they will probably bring something fancy to the project 1114197346 M * eyck OTOH, are the days of pornview numbered? 1114197596 M * Doener daniel_hozac: hmm... you say that rpm is fancy? (after all enrico says rpm is the [primary?] reason for using dietlibc...) ;) 1114197701 M * daniel_hozac Doener: vrpm and friends, yeah ;) 1114198495 J * nayco ~nayco@82.249.44.48 1114198521 M * Doener welcome nayco! 1114198596 M * nayco 'llo :) ! 1114198805 M * nayco Doener: Do you know where I can fnd a tutorial on programming design ? I mean, the steps before actually coding ? 1114198878 M * nayco I'm writing a program, and it's bigger than those I used to write: I need to take a pencil before coding, this time ;) 1114199136 M * Doener sorry, nothing i could help you with... i never read any 'complete' book/guide/tutorial on that stuff... i've just jumped into programming and from time to time, i read up on various topics... 1114199164 M * Doener but maybe some reference stuff would fit nicely into my browser's bookmarks... let's ask google :) 1114199199 M * nayco So, how do you usually go through the design step ? 1114199222 M * nayco Google isn't my freind this time ;) 1114199242 M * Doener heavily depends on the stuff i do, what language(s) i use and how many people are involved... 1114199306 M * nayco Well, let's imagine ( :P !!!) print accounting and reporting, PHP, 1 coder ;) 1114199333 M * Doener php4 or php5? 1114199348 M * Doener (IMHO oo-stuff doesn't make sense in php4) 1114199425 M * nayco Don't really need OO.... Anyway, I use PHP4 now, and I began this project using a few OO programming. But it's not a critical section 1114199503 M * Doener you can use OO with php4, but f.e. creating a singleton class, gave me more trouble than it's worth... it just misses sooo much ;) 1114199525 M * Doener anyways... you know uml? 1114199545 M * Doener (if you're going for oo...) 1114199788 M * nayco I'lm not really giong for OO... It's not mandatory. For UML, yeah, I learned a few about it at school, but it was too hard. Or maybe the teacher sucked ;). I think UML it a too big thing for my project(s): Seems it's not worth it. But maybe I'm wrong, and anyway I know it is the kind of method I need to learn to code projects larger than 100 lines ;) 1114199835 M * Doener 99% of the time you just need a small subset of uml 1114199912 M * nayco Good news ;) 1114199987 M * Doener as a start, i'd suggest to take a look at 'the whole'. what data are you gonna get, what data are you gonna give? (you can ignore most of the ui stuff for now) 1114200012 M * Doener think of the programs you'll need besides the one you're writing 1114200075 M * Doener then try to identify the different entities you're going to need. once you got them, try to figure out, what operations you'll need to do your work with them 1114200116 M * nayco My actual problem is that I have a description of wht the web site should contain, but that's all. I need to find a good method to design the program. How to cut the code in small parts, what to put together, what to ignore of decide later, relations, structure, etc... Then, when everything is written down, the only thing left is to code, and I know enough to do that. 1114200127 J * kevinp ~kevinp@ny.webpipe.net 1114200134 M * Doener kevinp: you won ;) 1114200143 M * Doener (regarding the fc3 minimal vserver) 1114200165 M * nayco I gonna verify if this chat is logged on my machine ;) 1114200187 M * Doener it's logged anyway ;) http://irc.13thfloor.at/LOG/ 1114200206 M * nayco I know, but I always forget the URL ;) 1114200211 M * nayco => bookmarks 1114200214 M * Doener heh :) 1114200269 M * kevinp Doener, oh yeah? 1114200289 M * Doener nayco: if you like to, i can give you the source code of a (pretty stupid) MySQL/PHP based email reminder thingy... 1114200370 M * nayco You mean, to learn PHP an d MySQL ? 1114200393 A * Doener just needs to copyright and gpl it ;) 1114200420 M * Doener no, just to see how i structured that code... like breaking it into units and such 1114200445 M * nayco Why not ? I'll be glad to learn from it. 1114200464 M * Doener it's probably not the best you can do (i already discovered some drawbacks), but it's not the worst either i guess 1114200500 M * nayco Oh, I just found the english word that I was missing to make Google help me: "Planning" !!! :) 1114200506 M * nayco Ok, send it to me... 1114200618 M * kevinp Doener, I get the ml in digest and I don't see the email on the archives from enrico 1114200629 A * kevinp just finished reading from when I was off 1114200675 M * kevinp did it go to you or the ml? 1114200823 M * Doener ml 1114200918 M * nayco brb 1114200938 M * Doener kevinp: hm... only paul's archives have it... http://archives.linux-vserver.org/200504/0230.html 1114201386 Q * albeiro Ping timeout: 480 seconds 1114201454 J * albeiro albeiro@albeiro.usercloak.oftc.net 1114202105 M * kevinp Doener, so it looks like I got it right except for replacing the already existent 01-03 files 1114202119 M * Doener yup yup 1114202133 M * kevinp I'll have to modify my doc appropriately 1114202170 M * Doener probably i'm just too used to debootstrap setting up a pretty complete minimal system ;) 1114202236 M * kevinp Bertl and I were talking awhile back about adding a page that contains a list of links to point to distro specific help/tips 1114202323 M * kevinp and then in your step by step doc for example instead of saying see google for this or that because distros vary, you could just add the one link to the page that has all the distro specific hints 1114202433 M * kevinp something like this: http://linux-vserver.org/distro_hints 1114202514 M * kevinp This is the part of your doc I would like to eventually replace with that link "This really depends on your setup, if you don't know how to do this at all, please ask google for some howto or take a look at your distribution's documentation." 1114202540 M * kevinp Sorry, Doener, I quess I never said that I was talking to you 1114202643 M * Doener feel free to do so.. it's a wiki, i got no problems with anyone changing 'my' stuff 1114202676 M * Doener (and it's probably a good idea to give those hints, i just don't know about that many distros, thus the advice to ask google) 1114202724 A * Doener .oO( maybe i shouldn't have used 'I' that much in the guide... ;) 1114202827 M * kevinp ok, I want to make a more complete page of hints first though 1114203175 Q * lilo Read error: Connection reset by peer 1114203181 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1114204032 J * lilo ~lilo@lilo.usercloak.oftc.net 1114204103 Q * nayco Quit: Bonne nuit ! 1114204133 Q * lilo_ Ping timeout: 480 seconds 1114205313 N * kevinp kevinp|gone 1114205435 Q * alexx xenon.oftc.net oxygen.oftc.net 1114205435 Q * bro xenon.oftc.net oxygen.oftc.net 1114205435 Q * locksy xenon.oftc.net oxygen.oftc.net 1114205435 Q * daniel_hozac xenon.oftc.net oxygen.oftc.net 1114205435 Q * Bertl_oO xenon.oftc.net oxygen.oftc.net 1114205435 Q * mikegrb xenon.oftc.net oxygen.oftc.net 1114205435 Q * ola_ xenon.oftc.net oxygen.oftc.net 1114205435 Q * jd86 xenon.oftc.net oxygen.oftc.net 1114205435 Q * SiD3WiNDR xenon.oftc.net oxygen.oftc.net 1114205435 Q * Doener xenon.oftc.net oxygen.oftc.net 1114205435 Q * Hollow xenon.oftc.net oxygen.oftc.net 1114205435 Q * mugwump xenon.oftc.net oxygen.oftc.net 1114205435 Q * TheSeer xenon.oftc.net oxygen.oftc.net 1114205435 Q * Beirdo xenon.oftc.net oxygen.oftc.net 1114205435 Q * berni xenon.oftc.net oxygen.oftc.net 1114205435 Q * aba xenon.oftc.net oxygen.oftc.net 1114205435 Q * DuckKing xenon.oftc.net oxygen.oftc.net 1114205435 Q * DuckMaster xenon.oftc.net oxygen.oftc.net 1114205435 Q * romke xenon.oftc.net oxygen.oftc.net 1114205435 Q * ntrs__ xenon.oftc.net oxygen.oftc.net 1114205435 Q * DaCa xenon.oftc.net oxygen.oftc.net 1114205435 Q * cereal xenon.oftc.net oxygen.oftc.net 1114205435 Q * rs xenon.oftc.net oxygen.oftc.net 1114205435 Q * sith xenon.oftc.net oxygen.oftc.net 1114205435 Q * gregster xenon.oftc.net oxygen.oftc.net 1114205435 Q * atsab xenon.oftc.net oxygen.oftc.net 1114205435 Q * logger xenon.oftc.net oxygen.oftc.net 1114205435 Q * Seraph xenon.oftc.net oxygen.oftc.net 1114205437 Q * kalou_ xenon.oftc.net oxygen.oftc.net 1114205437 Q * mcp xenon.oftc.net oxygen.oftc.net 1114205437 Q * ndim xenon.oftc.net oxygen.oftc.net 1114205437 Q * maharaja xenon.oftc.net oxygen.oftc.net 1114205437 Q * Loki|muh xenon.oftc.net oxygen.oftc.net 1114205437 Q * virtuoso xenon.oftc.net oxygen.oftc.net 1114205437 Q * SNy xenon.oftc.net oxygen.oftc.net 1114205437 Q * Medivh xenon.oftc.net oxygen.oftc.net 1114205437 Q * tchan xenon.oftc.net oxygen.oftc.net 1114205437 Q * DaPhreak xenon.oftc.net oxygen.oftc.net 1114205437 Q * sladen xenon.oftc.net oxygen.oftc.net 1114205437 Q * micah xenon.oftc.net oxygen.oftc.net 1114205437 Q * pusling xenon.oftc.net oxygen.oftc.net 1114205437 Q * nox xenon.oftc.net oxygen.oftc.net 1114205437 Q * stupidawy xenon.oftc.net oxygen.oftc.net 1114205437 Q * Zoiah xenon.oftc.net oxygen.oftc.net 1114205437 Q * cemil xenon.oftc.net oxygen.oftc.net 1114205437 Q * Snow-Man xenon.oftc.net oxygen.oftc.net 1114205532 J * DuckKing ~Duck@dyn-83-157-160-229.ppp.tiscali.fr 1114205532 J * DuckMaster ~duckx@195.75.27.158 1114205532 J * kalou_ ~kalou@AToulon-201-1-30-46.w81-48.abo.wanadoo.fr 1114205532 J * mcp ~hightower@217.171.201.37 1114205532 J * ndim hun@helena.bawue.de 1114205532 J * aba ~aba@sol.turmzimmer.net 1114205532 J * berni ~berni@svr01.mucip.net 1114205532 J * ola_ ~ola@c-adt-5.ataco.se 1114205532 J * jd86 ~jim@ip68-9-97-23.ri.ri.cox.net 1114205532 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1114205532 J * Beirdo ~gjhurlbu@beirdo.usercloak.oftc.net 1114205532 J * Doener doener@193.24.208.125 1114205532 J * Hollow ~Hollow@home.xnull.de 1114205532 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1114205532 J * TheSeer ~theseer@border.office.salesemotion.net 1114205532 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1114205532 J * cemil ~cemil@defiant.wavecon.de 1114205532 J * nox ~nox@noxlux.de 1114205532 J * Snow-Man ~sfrost@snowman.net 1114205532 J * Zoiah Zoiah@matryoshka.zoiah.net 1114205532 J * logger ~rs@vds.pas-mal.com 1114205532 J * Seraph kk@projects.verfaction.de 1114205532 J * atsab ~as@lotes.vtu.lt 1114205532 J * gregster ~gregor@greart.de 1114205532 J * sith sith@aaronp.com 1114205532 J * rs ~rs@194.98.28.10 1114205532 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1114205532 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1114205532 J * cereal ~cereal@217.20.124.153 1114205532 J * Medivh ck@paradise.by.the.dashboardlight.de 1114205532 J * sladen paul@starsky.19inch.net 1114205532 J * micah micah@micha.hampshire.edu 1114205532 J * SNy ~mfr@bmx-chemnitz.de 1114205532 J * pusling ~pusling@195.215.29.124 1114205532 J * virtuoso ~s0t0na@80.253.205.251 1114205532 J * DaCa ~danny@mail.limehouse.org 1114205532 J * Loki|muh loki@satanix.de 1114205532 J * ntrs__ ntrs@Dardeene-68.188.50.87.charter-stl.com 1114205532 J * romke romke@procyon.romke.net 1114205532 J * maharaja maharaja@ipax.at 1114205532 J * mikegrb ~michael@mikegrb.netop.oftc.net 1114205532 J * alexx ~alexx@82.225.136.176 1114205532 J * Bertl_oO ~herbert@janus.mc.tuwien.ac.at 1114205532 J * bro ~vanity@lanparty.lv 1114205532 J * locksy ~locksy@mrtg.sisgroup.com.au 1114205532 J * daniel_hozac ~daniel@h75n2fls33o829.telia.com 1114208906 Q * atsab Ping timeout: 480 seconds 1114209430 N * Bertl_oO Bertl 1114209445 M * Bertl evening folks! 1114209464 M * romke hi Bertl 1114209471 M * Doener hey Bertl! 1114209485 M * Bertl hey Doener! romke! 1114210138 Q * Vudumen Ping timeout: 480 seconds 1114210371 J * Vudumen vudumen@perverz.hu 1114210517 M * Bertl Doener: anything important I missed today? 1114210559 M * Doener no 1114210607 M * Bertl great! how is coding going? 1114210749 M * Doener Linux-VServer: port to .12-rc3 -> not started yet, code auditioning -> didn't feel like doing that today. Other stuff: extended my php framework... found new memory leaks in php4... ;) 1114210792 M * Doener (at least i guess those are memory leaks...) 1114210805 M * Bertl well, sounds fine ... 1114210837 M * Bertl finished (and tested!!) the syscall implementations for ppc/sparc and sparc64 yesterday ;) 1114210923 M * Doener not really... i'm like 'do a little here, do a little there' recently... nothing gets really usable... it's a little disappointing... but i don't know anything to do, that i feel capable of... (except porting/auditioning, but doing that all the time... well... ;) 1114210972 M * Bertl well, I'm now cleaning up the dlimit code (splitup and preparation for 32bit compatibility) and I think I have something for you if you're interested in some coding and testing? 1114210992 J * nayco ~nayco@82.249.44.48 1114211000 M * Bertl welcome nayco! 1114211006 M * nayco Hello !!! 1114211030 M * Doener wb nayco 1114211032 M * nayco enrico is still unjoinable here ? 1114211040 M * Doener got my mail? 1114211045 M * Doener Bertl: tell me :) 1114211046 M * nayco ....wait 1114211071 M * Bertl nayco: well, it seems so, I personally don#t know _what_ keeps him from showing up here ... but he will have his reasons ... 1114211102 M * Bertl Doener: there is this 'wait' syscall command, which should block a calling process until the 'requested' context exits ... 1114211130 M * Bertl it is incomplete, as I discovered the last time I looked at it, and it was obviously never tested ;) 1114211194 M * Bertl nevertheless I consider this to be important, especially now after the vshelper changes ... 1114211211 M * nayco Doener: Got it !!! Thx alot 1114211217 M * Doener you're welcome! 1114211245 M * nayco Bertl: Ok. I'll try to send him another mail... This flower page hurts my eyes :P ! 1114211261 M * Doener nayco: choose a different css style :) 1114211264 M * Bertl nayco: hmm, but you know that it has several stylesheets? 1114211313 M * nayco Oh, I forgot to say.... I read the blinking text, and sent him a (I hope ;-)) better stylesheet. Having no news, I was wondering... 1114211580 M * Bertl well, I guess he will add it ... 1114211620 M * Bertl Doener: does that sound like something which would be of interest to you? 1114211656 M * Doener Bertl: yep, seems like i could do that :) 1114212081 M * nayco http://nayco.free.fr/Depot/vservers/configuration.html ;-) 1114212200 M * Bertl hmm, is it intentional that the title floats over the contents? 1114212241 M * nayco yes... You thinks it degrades readability ? 1114212287 M * nayco -s 1114212289 M * Bertl well, never figured why somebody would like to have overlapping text ... but that's a matter of taste ;) 1114212350 M * nayco One thing I just remembered is that I haven't tested it on IE.... shame on me. 1114212356 M * nayco or maybe... ;) 1114212388 M * Doener IE doesn't support position:fixed ;) 1114212446 M * Doener IMHO, if you use position fixed, do it so that there's never any text that visibly scrolls under it... 1114212629 M * nayco Well, that was only cosmetic... But Enrico is free to remove it ;) 1114212927 Q * lilo Quit: bbiab 1114213103 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-pre2, ng9.4 -- 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 ;) 1114213143 M * Doener Bertl: the wakeup is supposed to happen when all tasks have exited? or when the context is really disposed? 1114213174 M * Bertl I'd say it should be _after_ the context got disposed 1114213177 M * DaCa Bertl: is there an easy way I can find out which delta's make a pre1 a pre2? 1114213194 M * Bertl DaCa: http://vserver.13thfloor.at/Experimental/delta-vs2.0-pre1-pre2.diff 1114213206 M * DaCa Bertl: great :) 1114213211 M * Doener ok, so the vx_state is no good there... 1114213242 M * Bertl no 1114213279 M * Bertl I guess the best would be to make a separate 'wake_up' function, which we can place wherever we like 1114213281 M * Doener hmm... logic strikes :) is that a "no, vx_state is good"? 1114213316 M * Bertl so it was a no, vx_state is too early ;) 1114213334 M * Bertl think of it like the normal wait() call 1114213350 M * Bertl you won't get that until the task is gone ... 1114213372 M * Doener btw, i got the mod_rewrite problem solved by changing the order of the rules... 1114213383 Q * kjo Ping timeout: 480 seconds 1114213681 Q * logger Ping timeout: 480 seconds 1114213817 J * lilo ~lilo@lilo.usercloak.oftc.net 1114214056 M * Bertl wb lilo!