1111536431 J * monrad ~monrad@0x535b06c0.ronxx3.adsl-dhcp.tele.dk 1111537210 J * keyser_soze ~keyser@host221.201-252-29.telecom.net.ar 1111539076 Q * DukeGangsta|aw Ping timeout: 480 seconds 1111542648 J * tad ~tad@h000f3d19edcc.ne.client2.attbi.com 1111543145 Q * tad Quit: using sirc version 2.211+KSIRC/1.3.11 1111543155 J * tad ~tad@h000f3d19edcc.ne.client2.attbi.com 1111543156 Q * monrad Quit: Leaving 1111543445 J * monrad ~monrad@213083190130.sonofon.dk 1111545346 Q * keyser_soze Quit: Abandonando 1111545833 J * DukeGangsta|aw ~uefgangst@p54869AE6.dip.t-dialin.net 1111550642 Q * tad Remote host closed the connection 1111551870 Q * Bertl_zZ Read error: Connection reset by peer 1111551875 J * Bertl_zZ ~herbert@janus.mc.tuwien.ac.at 1111552089 Q * kalou Ping timeout: 480 seconds 1111553581 Q * nox Ping timeout: 480 seconds 1111553610 J * nox ~nox@213.39.135.170 1111553913 M * ciphernaut how do i stop a vserver that has no label/name? just a context number 1111554269 J * nox_ ~nox@83.133.126.31 1111554317 Q * nox Killed (NickServ command used by nox_) 1111554334 N * nox_ nox 1111557495 N * DaPhreak|gone DaPhreak 1111559482 J * kalou ~kalou@80.13.75.17 1111560812 Q * stupidawy jupiter.oftc.net quasar.oftc.net 1111560857 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1111566018 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111566346 J * mhepp ~mhepp@r30s12p13.home.nbox.cz 1111566388 J * douglas ~douglas@douglas.user.oftc.net 1111566391 M * douglas anyone alive? 1111567140 Q * erwan_ho Remote host closed the connection 1111567433 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1111569307 N * BobR_oO BobR 1111569434 N * Bertl_zZ Bertl 1111569450 M * Bertl morning folks! 1111569455 M * Bertl douglas: yes, I am ;) 1111569472 M * BobR Guten morgen, Bertl! 1111569493 M * BobR [i.e. good morning, Bertl!] 1111569519 M * Bertl thanks, a good morning to you too, BobR! 1111569557 M * IceTi (Bertl): so i just need the patched kernel and the util-server tools? 1111569575 M * Bertl yes, that should be sufficient ... 1111569583 M * Bertl (at least that's all I use ;) 1111569624 M * IceTi ok i think i start with the wrong how to 1111569626 M * IceTi damn 1111569665 M * Bertl well, probably an older one, but the 'principles' are the same ... 1111569759 M * IceTi yes i see 1111569785 M * IceTi i ask because i write on my document in time ;-) 1111569915 M * Bertl did you ahve a look at the util-vserver pages yet? 1111570013 M * IceTi öhm 1111570029 Q * mhepp Quit: KVIrc 3.0.1.99 'Realia' 1111570035 M * IceTi site? 1111570037 M * meebey morning 1111570052 M * IceTi morning 1111570065 M * meebey Bertl: I still believe there is a bug in UDP 1111570073 M * meebey Bertl: should I explain it? 1111570133 M * Bertl yes, please, maybe there is ;) 1111570163 M * Bertl IceTi: http://linux-vserver.org/alpha+util-vserver 1111570169 M * IceTi :-) 1111570249 M * meebey Bertl: when a program inside the vserver uses a UDP socket with 0.0.0.0 and sends data to a IP it will always get invalid argument 1111570272 M * meebey Bertl: but when I use a virtual IP/device and do NAT and it binds to 0.0.0.0 it will work 1111570285 M * meebey Bertl: I found this when using snmpd 1111570315 M * meebey Bertl: its very strange since I allowed the native devices like eth0 for the vserver and it does not work 1111570345 M * meebey Bertl: when I create a eth0:foo with its own IP and do SNAT/DNAT for that internal/virtual IP, it works 1111570351 M * Bertl okay, slowly, first, what versions are we talking about 1111570378 M * meebey Linux galilei 2.4.29-vs1.2.10p #1 SMP Mon Feb 7 17:46:04 CET 2005 i686 unknown 1111570415 M * meebey Bertl: oh btw, using opeswan and snmpd together is a kernel killer :-P but that has nothing to do with vserver 1111570453 M * Bertl hehe 1111570498 M * meebey the kernel will oops and all network device related calls will block 1111570506 M * meebey :-P was not nice finding that out :) 1111570520 M * meebey its a openswan bug, it does some funny things on some probs 1111570525 M * meebey probes 1111570528 M * Bertl well, I guess that could be fixed ;) 1111570554 M * meebey :) anyhow thats not the problem I was going to report to you 1111570561 M * meebey its the 0.0.0.0 with UDP socket one 1111570601 M * Bertl okay, so we are talking about a 2.4/stable kernel here, how do you assing a network interface to a vserver there? 1111570609 M * Bertl (because that is not possible ;) 1111570785 M * meebey IPROOT=eth0 or more likely IPROOT=ALL 1111570796 M * meebey because the snmpd should gimme stats about all interfaces 1111570801 M * meebey thats the idea behind it :) 1111570988 M * Bertl yes, and here is the important difference ... 1111571003 M * Bertl because IPROOT=ALL means (and this _is_ possible) a list of all currently assigned ips ... 1111571042 M * Bertl so you do not specify any interface and the vserver will not be able to catch up new IPs later on with that config ... 1111571082 M * Bertl what you might want instead is to specify '0.0.0.0' as iproot value (or do not assign any IPs at all ... 1111571083 M * meebey hu? the devices/IPs all exist 1111571099 M * meebey that never worked for me 1111571108 M * meebey IPROOT=0.0.0.0 will not assign any IP to any device 1111571128 M * meebey ifconfig shows then all devices but no IPs at all for those 1111571140 M * meebey how come it works with a virtual IP and NATing? 1111571152 M * Bertl sec, checking your claims ... 1111571165 M * meebey k 1111571211 Q * DukeGangsta|aw Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1111571445 M * Bertl hmm, indeed, the code seems to block all interfaces if 0.0.0.0 is specified ... 1111571467 M * Bertl (I mean from listing, I guess they should work quite fine though) 1111571582 M * Bertl yep, that's a bug, but the 'binding' works as expected with 0.0.0.0, you should really try that ... 1111571620 M * Bertl Doener: you around by any chance? 1111571967 M * meebey Bertl: so you mean with 0.0.0.0 the programs should be able to work but ifconfig just gets wrong information? 1111571976 M * meebey Bertl: but snmpd will believe too that there is no IP? 1111571998 M * meebey its important because the snmpd does a look up for the IPs 1111572007 M * Bertl no, I guess it will bind to 0.0.0.0 which is quite fine ... 1111572030 M * meebey the IPROOT=0.0.0.0 problem I reported long though a long time ago to Doener I think 1111572041 M * meebey s/long// 1111572044 M * Bertl but yes, you are right, the interface hiding doesn't work as it should ... 1111572056 M * Bertl and it seems it got lost somehow :/ 1111572079 M * meebey ok, if it gets now fixed/handled I will be happy :) 1111572114 M * meebey this is just one real problem after using vserver for so long time, so its allright :) 1111572182 M * Bertl for sure it will (get fixed) ... 1111572653 M * meebey I have 3 vservers currently where I really don't want NAT, thats VPN (openswan), Samba and SNMP 1111572689 M * meebey samba and vpn are problematic protocols and snmp should get stats of the real interfaces 1111572740 M * Bertl did you try not to specify any address there, does that work with the tools or does it get replaced by something? 1111573421 M * meebey where? 1111573429 M * meebey leave IPROOT= iut? 1111573430 M * meebey out? 1111573470 N * BobR BobR_afk 1111573511 M * Bertl yep, try IPROOT="" same for all the other IP* fileds 1111573579 M * meebey according to the docs it should be the same as 0.0.0.0 1111573583 M * meebey I can try it though 1111573610 M * meebey ipv4root is now 0.0.0.0 1111573626 M * meebey seems to be the same, but like you said it should work anyhow 1111573629 M * meebey I will try that later 1111573636 M * Bertl k, thanks! 1111575240 N * BobR_afk BobR 1111575591 N * BobR BobR_afk 1111575867 J * keyser_soze ~keyser@host43.201-252-1.telecom.net.ar 1111576143 M * Bertl welcome keyser_soze! 1111576443 M * keyser_soze hi bert 1111576511 Q * keyser_soze Quit: Abandonando 1111577081 M * DaPhreak morning Bertl, ... ;) 1111577266 M * Bertl morning DaPhreak ! 1111578140 A * DaPhreak 's going home ;) cu later guys .. 1111578142 N * DaPhreak DaPhreak|gone 1111578350 M * Bertl cya 1111578952 M * erwan_taf cyo 1111580109 M * Doener morning 1111580109 Q * IceTi Read error: Connection reset by peer 1111580122 M * Bertl morning Doener! 1111580128 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1111580214 Q * virtuoso Read error: Connection reset by peer 1111581471 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net :: 1111581520 Q * kalou Ping timeout: 480 seconds 1111583673 J * virtuoso ~s0t0na@tranq.dorms.spbu.ru 1111583684 M * Bertl wb virtuoso! 1111583716 M * virtuoso Hey Bertl! 1111583743 M * virtuoso Just found out that you're also doing development for ipaq h1940. :) 1111583841 M * Bertl yes, but not too much recently ;) 1111583948 M * virtuoso I was recently offered to buy h1930 and was wondering if linux runs there. 1111583970 M * virtuoso But that's offtopic, right? 1111584193 M * Bertl cya at #h1940 then ... 1111584409 M * aba Bertl: the build problems with dietlibc on the different platforms are about to be fixed. Most of the problems are really dietlibcs. 1111584429 M * Bertl most? so which issues remain? 1111584462 M * aba Bertl: like, we can't build on m68k, as debian doesn't have a dietlibc-dev package or so. 1111584484 M * aba (and there are no patches till now, but just analysis) 1111584527 M * Bertl aba: hmm, no dietlibc dev package? IIRC dietlibc is 'just' a dev package? 1111584546 M * Bertl (because it contains no dynamic stuff/libs) 1111584569 M * aba The diet libc also supports dynamic linking, currently on the arm and 1111584569 M * aba i386 architectures only. This package provides the shared libraries. 1111584583 M * aba so, we have dietlibc, dietlibc-dev and dietlibc-doc packages. 1111584649 M * Bertl well, so be it ... libdietlibc and dietlibc-dev would be more intuitive IMHO, but as I said, so be it ... 1111584686 M * Bertl okay, and this isn't available on m68k, right? 1111584705 M * Bertl so then probably util-vserver should not be available there too, right? 1111584714 M * aba yes. 1111584732 M * Bertl guess that should not be an issue then? 1111584735 M * aba so, we're going in a very good direction. 1111584751 M * aba not really an issue, it just needs ftp-master action (but that can be done) 1111584767 M * aba also, the 2.6-kernel-patch is now there. 1111584792 M * Bertl which one? or a general package? 1111584832 M * aba 1.9.4 1111584977 M * Bertl how would updates to that work? 1111584978 M * Bertl (just curious ;) 1111584999 M * Bertl an ist this against debian kernel or vanilla? 1111585019 M * aba you upload a new version of the patch. 1111585034 M * aba against debian kernel. 1111585070 M * aba the package contains multiple patches, and during build/in preparation the right patch is selected and applied. 1111585121 M * aba (actually, there should be a primer about the way debian handles kernel packages - it can do such a lot of things that that would be really needed) 1111585150 M * Bertl okay, so this was tested by? (the kernel I mean) 1111585183 M * aba well, AFAI understood it was tested with the debian i386 2.6.8-kernel (which is the kernel that will be released with sarge) 1111585216 M * Bertl so it is a 2.6.8-vs1.9.4 kernel? 1111585235 M * aba yes. 1111585255 M * Bertl interesting ... 1111585289 M * aba we're definitly currently going into the right direction there ... 1111585308 M * Bertl yeah, agreed ... 1111585369 M * Bertl who is maintaining the kernel package? 1111585424 M * aba ola 1111585427 M * aba (as all) 1111585453 M * Bertl hmm, okay, so it's probably completely untested then ... 1111585482 M * Bertl maybe we could get debian folks (users) to test this stuff? 1111585539 M * aba well, I can blog about it :) 1111585541 M * Bertl I mean folks from the linux-vserver community of course ... 1111585578 M * aba I'll speak with ola about how we can test. 1111586005 M * Bertl @ debian folks, anybody able to provide a _single_ patch against 2.6.8 (vanilla) for 2.6.8-15 (or whatever debian kernel is current)? 1111586102 M * Bertl aba: btw, http://www.osdl.org/plm-cgi/plm?module=patch_info&patch_id=4327, so 1.9.5 will compile fine on all platforms except for a few warnings (which probably will be eliminated soon) 1111586115 M * aba ah, great. 1111586283 M * Bertl any arch specific issue with util-vserver (not dietlibc related) should be forwared to enrico asap ... 1111586315 M * aba Hm, you might want to look into the bug 297806 - bugs.debian.org/297806 1111586427 M * Bertl well, I had a look at that one (see my reply ;) 1111586525 M * Bertl but reading up now .. 1111586541 M * aba yes. But liw has done some research on different archs. 1111586677 M * Bertl in general it looks like the other packages (like dietlibc) are not really tested on all platforms either ... 1111586689 M * aba yes. 1111586695 M * Bertl (I'm pretty sure dietlibc has some kind of regress suite ;) 1111586702 M * aba we're doing quite some QA for dietlibc currently. 1111586750 M * Bertl excellent! 1111586805 M * Bertl sideonote (to whoever does the dietlibc testing): do not forget about the not so obvious env pointer issues we encountered on 64bit archs some time ago ... 1111590133 J * stegbth ~stegbth@stegbth.sim.tronicplanet.de 1111590141 M * stegbth hi 1111590315 M * Bertl welcome stegbth! 1111593068 J * ensc|w ~ensc@62.153.82.27 1111593088 J * clflush ~clflush@cm184.omega194.maxonline.com.sg 1111593116 M * Bertl good afternoon ensc|w! 1111593124 M * Bertl welcome clflush! 1111593149 M * clflush hi! 1111593313 P * clflush 1111593989 J * SVSwift ~SVSwift@pD95235D2.dip.t-dialin.net 1111593997 M * Bertl welcome SVSwift! 1111594002 M * SVSwift hello Bertl 1111594021 M * SVSwift I got two small questions: 1111594036 M * SVSwift Can I create a new vserver by copying another one? 1111594044 M * Bertl yes 1111594068 M * SVSwift the next: How do I set vserver's IP with alpha util? 1111594092 M * Bertl either on vserver creation with the --ip arguments 1111594099 M * SVSwift or? 1111594106 M * Bertl or after that in the config tree (see flower page for details) 1111594186 M * Bertl but there is a simple 'trick' how you can check various config things without crawling through the details ... 1111594210 M * Bertl the vserver tool (from util-vserver) supports the -m skeleton build method 1111594222 M * Bertl which creates a config and an empty vserver dir 1111594270 M * Bertl now you can use that with --ip ... and see what it writes to the config (just make sure to use a different name for the vserver) 1111594300 M * SVSwift ok, thank you ;) 1111594303 M * Bertl in the future, the tools (at least I guess so) will be able to change the config with similar commands 1111594313 M * Bertl you're welcome! 1111594453 M * Bertl SVSwift: so how do you like linux-vserver so far? 1111594462 M * stegbth hi all 1111594467 M * stegbth thanx for the personal greetinig 1111594483 M * Bertl heh, that's part of my service ;) 1111594517 M * nox *hug* Bertl 1111594534 M * Bertl hey nox! 1111594538 M * stegbth are there already a working version for 2.6.11? or only a rc, which doesn't correctly apply? 1111594568 M * nox Bertl: still listening *g* 1111594648 M * Bertl stegbth: http://vserver.13thfloor.at/Experimental/patch-2.6.11.5-vs1.9.5-rc4.diff 1111594664 M * nox Bertl: will vserver be represented on LT this year again ? 1111594683 M * Bertl LT being? 1111594689 M * nox linuxtag sorry 1111594719 M * SVSwift I think vserver is great but not that easy to install for a newbee 1111594722 M * Bertl well, I have to think about it, still didn't get the travelling expenses covered by the linuxtag folks :/ 1111594744 M * Bertl (although they confirmed that they would do so ;) 1111594750 M * nox oh thats uncool 1111594763 M * SVSwift one documentation in one place covering alle distributions would help a lot, I think 1111594777 M * stegbth Bertl: are this url anywhere noticed? cause i was unable to find it on the webpage? 1111594784 M * nox SVSwift Bertl made a wiki for you *g* 1111594785 M * Bertl nox: when will it be this yeara? 1111594794 M * SVSwift hrhr 1111594795 M * Bertl SVSwift: linux-vserver.org is that place 1111594813 M * SVSwift linux-vserver.org is greatly unorganized, I think 1111594821 M * SVSwift sorry for that 1111594826 M * daniel_hozac feel free to organize it ;) 1111594828 M * nox 22. bis 25. June Bertl 1111594853 M * stegbth berni: 404 Not Found 1111594882 M * SVSwift that's a lot of work 1111594882 M * SVSwift Let me use vserver for some time to get the knowledge what needs to go where 1111594893 M * SVSwift But basically: Why not 1111594911 M * nox afaik they are now thinking abaout fee .... seems to be lots og trouble inside 1111594938 M * SVSwift I'm now using some bind9.debs from a french page and can't find it again... 1111594968 M * Bertl why not use 'the default' packages? 1111594979 M * Bertl stegbth: unusual .. sure your resolving works? 1111595018 M * Bertl stegbth: hmm, you said 404, so resolving should be fine, did you copy the url correctly? try http://vserver.13thfloor.at/Experimental 1111595053 M * Bertl nox: but I will be at 'linuxwochen' 1111595079 M * nox so hope you have more time than lastyear *g* 1111595100 M * SVSwift uuh, copying a whole vserver takes some time 1111595101 M * nox aeh wochen ... 1111595121 M * Bertl http://linuxwochen.at/ (in Vienna) 1111595125 M * nox nice idea 1111595189 M * nox 6 weeks what an event 1111595219 M * Bertl yeah, well, you can consider every location a separate event ... 1111595278 M * stegbth Bertl: now it works, let see, what the compiler from sarge is saying ;) 1111595294 M * nox one event touring thru at 1111595315 M * Bertl stegbth: you are applying it to a vanilla kernel I hope, not debian one ... 1111595405 M * stegbth Bertl: yes 1111595422 M * stegbth i am applying to fresh vanilla kernel from ftp.de.kernel.org 1111595431 M * Bertl excellent! ;) 1111595471 M * stegbth need it to get activated? or is it always like on 2.4.x? 1111595504 M * Bertl there are some options, but it's pretty fine with the defaults ... 1111595518 M * stegbth where are the options hidden? 1111595536 M * Bertl in their own menu ... (or at places where they make sense) 1111595545 M * stegbth ;) 1111595609 M * Bertl the memory split option for example is in vicinity of the highmem option 1111595666 M * Bertl (but everything strictly vserver related is in the 'Linux VServer' menu) 1111596544 M * SVSwift how can I preserve owner-information when copying 1111596545 M * SVSwift ? 1111596581 M * Bertl cp -a, or use dump/restore or rsync ... 1111598066 N * BobR_afk BobR 1111598616 M * stegbth how do i get rid of proc filesystem error's when entering the vserver (debian woody) 1111598655 M * Bertl this is the result of a badly configured proc security ... 1111598699 M * Bertl easiest way to get it right is to run vprocunhide 1111598713 M * stegbth inside or outside vserver? 1111598718 M * Bertl on the host 1111598737 M * Bertl the tools (util-vserver) should add a runlevel script for doing that on host bootup 1111598766 M * stegbth doesn't exist 1111598784 M * stegbth dpkg -l util-vserver |grep proc 1111598790 M * stegbth doesnt bring anything 1111598831 M * Bertl which version of util vserver? btw: 1111598836 M * Bertl rpm -ql util-vserver | grep vproc 1111598836 M * Bertl /usr/lib/util-vserver/defaults/vprocunhide-files 1111598836 M * Bertl /usr/lib/util-vserver/vprocunhide 1111598854 M * stegbth ii util-vserver 0.30-14 1111598857 M * stegbth debian sarge 1111598890 M * stegbth what does vprocunhide? 1111598892 M * Bertl okay, you are using the stable tools, they do not have that 1111598930 M * Bertl http://linux-vserver.org/Proc-Security 1111598983 M * Bertl vprocunhide calles setattr (which is also part of the devel tools) with proper arguments, to 'unhide' certain entries in the procfs (for the vserver) 1111599006 M * Bertl you can get the same result by doing it yourself (with some script and the vproc tool) 1111599042 M * Bertl or if you do not care about the proc security, you can disable it completely either at compile time or after startup 1111599105 M * stegbth ah, ok thanx 1111599622 N * BobR BobR_oO 1111599931 M * stegbth lets hope util-vserver from sid comes to sarge 1111600014 M * Bertl well, it seems like debian folks are working on getting recent stuff into debian distros ;) 1111600173 J * matta-lt_ ~matta@69.93.28.254 1111600257 M * Bertl evening matta-lt_! 1111600702 Q * ciphernaut Ping timeout: 480 seconds 1111600890 M * matta-lt_ hi! 1111600891 M * stegbth i have to leave 1111600894 M * matta-lt_ what's new? 1111600896 M * stegbth bye 1111600904 M * stegbth thanx 4 all 1111600905 Q * stegbth Quit: Client exiting 1111600989 M * Bertl matta-lt_: well, not that much ... still busy with other stuff ... 1111601009 M * Bertl (but 1.9.5 will be released soon ;) 1111601342 M * matta-lt_ I notice the changelog on the site is no longer being developed 1111601347 M * matta-lt_ that is how I kept track 1111601471 M * Bertl hmm ... 1111601535 M * Bertl changelogs are up to date with releases ... 1111601564 M * Bertl (no entries for recent experimental releases though) 1111601628 M * daniel_hozac (the topic is also an rc behind ;)) 1111601686 M * Bertl well, if you notice, why not bring it up to date? 1111601742 M * daniel_hozac i just noticed it, and i feel weird settings the topic for others channels :) 1111601790 M * matta-lt_ yeah 1111601794 M * matta-lt_ it only does 1.9.4 1111601802 M * matta-lt_ not changes in 1.9.5-xxx 1111601807 M * matta-lt_ or 1.9.5.x 1111601819 M * Bertl yes, didn't get the feeling anyone cares ;) 1111601896 M * Bertl vserver.13thfloor.at/Experimental/FOR-1.9.5/ <-- 1.9.5 changes 1111602059 M * matta-lt_ oh bertl... you know people care :) 1111602068 M * matta-lt_ you know people watch vserver likea hawk (ie. me) 1111602077 M * matta-lt_ watching the development 1111602087 M * Bertl well, I now know that _you_ care about the changelogs ... 1111602097 Q * sebd Ping timeout: 480 seconds 1111602110 M * matta-lt_ what is this delta forks? 1111602110 M * Bertl so I'll try to keep them up to date in the future ;) 1111602127 M * matta-lt_ hrm... i'm trying to think of our idea 1111602154 M * matta-lt_ I know we had an idea that was much better than than trying to limit like 20+ different resources 1111602162 M * Bertl those deltas are the patches building teh 'delta' between 1.9.4 and 1.9.5 1111602169 M * matta-lt_ yeah 1111602173 M * matta-lt_ but what does 'forks' do ? 1111602196 M * Bertl well, let's take a look at it ;) 1111602210 M * Bertl +atomic_t total_forks;/* number of forks so far */ 1111602228 M * Bertl +"total_forks:\t%d\n" 1111602240 M * TheSeer Bertl: got a question for you too ;> 1111602249 M * matta-lt_ so it is accounting or a limit or what? 1111602253 M * Bertl looks like it's accounting the number of forks happened in a vserver so far ;) 1111602257 M * matta-lt_ ok 1111602277 M * Bertl TheSeer: hmm? 1111602288 M * TheSeer Bertl: i do have an "old" RH 7.3 based server... would it work to shut it down, move the hdd to a vserver enabled box, mount the hdd into /vservers/somename 1111602301 M * daniel_hozac what would you use that accounting for? 1111602302 M * matta-lt_ i'm too stupid these days bertl :( 1111602305 M * TheSeer Bertl: then create a matching config and run it as a vserver? *g* 1111602305 M * matta-lt_ i'm all in perl heaven 1111602319 M * TheSeer the vserver host would be a FC3 box... 1111602321 M * daniel_hozac matta-lt_: so that's a place where Perl doesn't exist? :) 1111602324 M * matta-lt_ Bertl: lycos is still using vserver? 1111602333 M * matta-lt_ daniel_hozac: funny :) 1111602361 M * Bertl TheSeer: guess so ... 1111602390 M * Bertl (a bunch of complaints because of services which can not be started, but it should work) 1111602407 M * matta-lt_ any work on ngnet / cowlinks? 1111602419 M * Bertl ngnet is working fine, except for the arp stuff ... 1111602425 M * matta-lt_ yeah? 1111602431 M * matta-lt_ does it still uses aliases? or proxy arp? 1111602442 M * Bertl no aliases, virtual interfaces 1111602448 M * matta-lt_ ok 1111602458 M * matta-lt_ but how do you limit what IP's a vserver can use 1111602461 M * Bertl as I said, arp is not done yet, you have to use static arp 1111602474 M * matta-lt_ oh, ok 1111602477 M * Bertl ips are limit via iptable rules 1111602482 M * matta-lt_ ok 1111602487 M * matta-lt_ based on a mark? 1111602493 M * Bertl special 'vnet' table ;) 1111602500 M * matta-lt_ ahhh 1111602504 M * matta-lt_ very unique 1111602523 M * Bertl yeah? is it? 1111602560 M * Bertl daniel_hozac: in the future, it might be used to adjust some things for a context ... 1111602826 M * Bertl matta-lt_: so what's up with your stuff? 1111602914 M * matta-lt_ eh 1111602915 M * matta-lt_ same old 1111602921 M * matta-lt_ i become more of a suit every day 1111602926 M * matta-lt_ and more stressed 1111602934 M * nox where moved ensc ? is there any download @ savannah? 1111602936 M * matta-lt_ got offered to be bought out, not worth it... 1111602955 M * matta-lt_ I did launch a service based on xen 1111602963 M * Bertl nox: try http://www.13thfloor.at/~ensc/ 1111602987 M * Bertl matta-lt_: hey cool, how does it work for you? 1111602989 M * matta-lt_ http://www.unixshell.com/index.php?page=technology 1111602999 M * matta-lt_ I wrote the teknic panel... 1111603014 M * matta-lt_ it's been stable for the most part, people like it... is it fast, i'll give them that 1111603028 M * matta-lt_ I do offer users a kernel with vserver 1.9.4 applied to it :) 1111603038 M * matta-lt_ so they can run vserver inside their VM's 1111603058 M * matta-lt_ I figure I can port Teknic to vserver in 2 weeks 1111603062 M * Bertl how many xen servers does a typical machine handle? 1111603079 M * matta-lt_ just waiting on ngnet and cowlinks (and yes, I know it will take a while) 1111603082 M * matta-lt_ hrm, it depends 1111603085 M * matta-lt_ look at the plans 1111603093 M * matta-lt_ xen is limited to 4GB right now 1111603097 M * matta-lt_ no PAE, no x86_64 1111603104 M * matta-lt_ so I can fit about 120 32MB VM's 1111603113 M * matta-lt_ or 62 64MB's 1111603115 M * matta-lt_ etc, etc 1111603122 M * matta-lt_ xen is my replacement for UML 1111603129 M * matta-lt_ I want vserver to be my replacement for virtuozzo :) 1111603142 M * Bertl i.c. so how does a server with 62 64MB machines perform? 1111603147 M * matta-lt_ not bad at all 1111603153 M * matta-lt_ the xen QoS is excellent 1111603162 M * matta-lt_ way better than expected 1111603167 M * matta-lt_ the disk I/O QoS also... 1111603174 M * matta-lt_ it cannot be oversold 1111603188 M * matta-lt_ since all memory in wired memory and I use LVM for partitions... 1111603208 M * matta-lt_ which helps to keep it fast 1111603218 M * matta-lt_ really xen is good, it's being backed by large corps 1111603227 M * matta-lt_ but xen/uml are different market than vserver/virtuozzo 1111603271 M * Bertl what are the main differences in your opinion? 1111603283 M * matta-lt_ between xen and vserver or vserver and virtuozzo? 1111603722 M * Bertl well if you ask me this way, both of course ;) 1111603750 M * matta-lt_ xen is for mainframe stuff, a true VM 1111603765 M * matta-lt_ I know their plan and they expect to let VM's boot their own kernels, etc 1111603772 M * matta-lt_ it will be like a true dedicated server... 1111603823 M * matta-lt_ vserver is lower level 1111603826 M * matta-lt_ or higher level 1111603847 M * daniel_hozac lol 1111603865 M * matta-lt_ the main advantage is 1111603881 M * matta-lt_ vserver/virtuozzo can be share resources and therefore provide higher consolidation 1111603890 M * matta-lt_ but if the main kernel crashes... go boom 1111603904 M * matta-lt_ with xen, each VM has a kernel... if they panic it, oh well.. they reboot their VM 1111603920 M * Bertl even with the host/main context? 1111603926 M * matta-lt_ this is why technology such as virtuozzo/vserver will attract hosting providers 1111603939 M * matta-lt_ with xen if the main VM crashes then it all goes down 1111603960 M * matta-lt_ the main domains providers networking, disk access (loop or LVM), and runs a daemon to control xend 1111603964 M * matta-lt_ so it doesn't do a whole lot 1111603966 M * Bertl hmm, so how is that different then? 1111603976 M * matta-lt_ xen is a hypervisor 1111603979 M * matta-lt_ it loads itself 1111603989 M * matta-lt_ and then boots a kernel with high privs 1111603994 M * matta-lt_ so the main VM itself is a 'domain' 1111604002 M * matta-lt_ context/vm in xen = domain 1111604008 M * matta-lt_ so main domain is dom0 1111604024 M * matta-lt_ the other domains talk directly to the hypervisor 1111604040 M * matta-lt_ it's more of a vmware type approach, but the guest kernel needs to be poirted to the 'xen' arch 1111604042 M * Bertl okay, so who does the disk stuff, for example? 1111604047 M * matta-lt_ the performance is no overhead 1111604050 M * matta-lt_ well, with me 1111604056 M * matta-lt_ I create a LVM partition on the host 1111604060 M * matta-lt_ one for root, one for swap 1111604071 M * matta-lt_ and assign them as sda1 and sda2 in the domain 1111604117 M * Bertl yeah, but the hardware driver for that disks has to be somewhere right? 1111604134 M * matta-lt_ it's why the guest needs to be ported to the xen arch 1111604136 M * Bertl it is not compiled into the guest kernels, they have a special driver 1111604143 M * matta-lt_ right 1111604160 M * matta-lt_ to the guest they see the xen block device and under that device are the sda1 and sda2 partitions 1111604178 M * Bertl yes, but where does the block device support come from 1111604192 M * matta-lt_ i can't say I know xen internals that well 1111604198 M * Bertl just let us make a simple example ... 1111604236 M * Bertl let's assume you have a harware raid controller from adaptec (or any other company) 1111604255 M * Bertl you have a special linux-kernel driver for that controller, right? 1111604259 M * matta-lt_ yep 1111604263 M * matta-lt_ that is part of dom0 1111604271 M * matta-lt_ dom0 has special privileges for direct hardware access 1111604296 M * matta-lt_ so you take a partition of that 1111604301 M * matta-lt_ and make a LVM volume group 1111604306 M * matta-lt_ then create logical volumes on that 1111604334 M * matta-lt_ and say volume name1-root is sda1 in domain 'name1' and name2-swap is swap for 'name1' 1111604347 M * matta-lt_ er, sda2 1111604348 M * Bertl precisely, so if the controller driver in dom0 (or the kernel in dom0) crashes, what will happen to the other doms? 1111604355 M * matta-lt_ they crash 1111604359 Q * daniel_hozac Quit: Oops... 1111604366 M * matta-lt_ oh, it's entirely likely that a host crash can happen 1111604375 M * matta-lt_ just much more unlikely than with a shared kernel 1111604386 M * matta-lt_ although virtuozzo is damn stable with a shared kernel 1111604386 M * Bertl ah, and why do you think so? 1111604391 M * matta-lt_ so both are goodf 1111604413 M * matta-lt_ what i'm saying is let's say someone tries to run wine w/ debugging in a VM 1111604422 M * matta-lt_ under Virtuozzo this used to c rash the host 1111604429 M * matta-lt_ they tracked down the problem and it is fine now 1111604444 M * matta-lt_ but under xen, if they run win w/ debugging i their VM it just crashes their kernel 1111604451 M * matta-lt_ since each VM loads it's own kernel 1111604455 M * matta-lt_ it's a complete waste of resources 1111604467 M * matta-lt_ shared kernel is the way to go for 'severe consolidation' 1111604481 M * matta-lt_ which appeals to a lot of people 1111604490 M * matta-lt_ also, with like xen 1111604501 M * matta-lt_ guests can run 2.4.28, 2.4.29, 2.4.10, 2.4.11, etc 1111604506 M * matta-lt_ all while the host is still 2.4.9 1111604512 M * matta-lt_ compile own kernel modules, etc 1111604517 M * matta-lt_ pro's and con's to each 1111604522 M * matta-lt_ it is why I label them 'different' 1111604531 M * matta-lt_ and not direct competitors 1111604547 M * matta-lt_ like I said, I consider uml/xen to compete and virtuozzo/vserver to compete 1111604613 M * matta-lt_ that's my reason for wanting the fully functional cowlinks implementation 1111604621 M * matta-lt_ just run the re-link script weekly 1111604634 M * matta-lt_ so it keeps re-saving the disk space (and memory due to less cache usage) 1111604658 M * matta-lt_ so on the same hardware and same workload, vserver can probably handle 20-30% more VM's than xen 1111604668 M * Bertl yeah, okay, I agree to that differences 1111604676 M * matta-lt_ since xen can't be oversold 1111604713 M * Bertl you are using 2.4 kernels for xen? 1111604715 M * matta-lt_ also xen supports most distros natively, since it's no different 1111604717 M * matta-lt_ I offer 2.4 1111604730 M * matta-lt_ but our benchmarks show just switching to 2.6 kernel is 40% faster 1111604734 M * matta-lt_ (using unixbench test) 1111604746 M * SiD3WiNDR hmm 1111604752 M * matta-lt_ if you look at the screen shot, I let users choose their kernel from a list that I provide 1111604753 A * SiD3WiNDR wants to switch a 1.2 to 1.9 1111604760 M * SiD3WiNDR are there any scripts to convert the config? 1111604779 M * Bertl you can use the old one for the time being ... 1111604786 M * matta-lt_ unixshell is for tinkerers, power users, etc 1111604797 M * matta-lt_ tektonic (main company) is for hostng VPS's 1111604806 M * Bertl SiD3WiNDR: conversion is trivial in theory so nobody bothered to write a script yet :/ 1111604808 M * matta-lt_ good support, control panels, all that jazz 1111604823 M * SiD3WiNDR Bertl: oh :) 1111604834 M * SiD3WiNDR Bertl: but the old one works? 1111604859 M * Bertl yes, that's part of the migration plan ... 1111604866 M * SiD3WiNDR cool :) 1111604910 M * Bertl old tools will work too (on the host) just not unleash new features .-.. 1111604933 M * Bertl the next step is probably to upgrade the tools, then switch to the new config ... 1111604947 M * SiD3WiNDR aha 1111604956 M * SiD3WiNDR but, I can just reboot with 2.6 kernel and it should work? very nice 1111604970 M * Bertl yes, 3 steps 1.2 _> 1.9 migration: 1111604994 M * Bertl 1) update module-init-tools and kernel 1111604997 M * SiD3WiNDR I saw 2.6 libata-dev finally has SMART support for SATA, so yay, gotta have that urgently - have 3 out of 4 failing 200G Maxtor drives :| So much for RAID5 ;) 1111605001 J * daniel_hozac ~daniel@h212n1fls33o829.telia.com 1111605006 M * Bertl 2) updated util-vserver tools 1111605018 M * Bertl 3) switch to new config 1111605022 M * SiD3WiNDR okay, that's nice :) 1111605023 M * SiD3WiNDR thanks 1111605029 M * Bertl you're welcome! 1111605039 Q * Vudumen Read error: Connection reset by peer 1111605227 M * Bertl matta-lt_: so cow links is a real issue for you? 1111605307 J * Vudumen vudumen@perverz.hu 1111605332 M * Bertl wb daniel_hozac! Vudumen! 1111605403 M * Vudumen hi Bertl :) 1111605700 M * daniel_hozac heh, thanks. 1111606491 Q * Vudumen Read error: Connection reset by peer 1111606698 M * Bertl matta-lt_: still there? 1111606718 M * matta-lt_ Bertl: cowlinks are vital for consolidation 1111606722 M * matta-lt_ it would beat out virtuozzo 1111606726 M * matta-lt_ they use an overlay fs 1111606738 M * Bertl ah, yes, indeed ... 1111606738 M * matta-lt_ so you know they HAVE to have double caching... 1111606763 M * Bertl okay, another xen question, how does networking work there from the admin point of view? 1111606764 M * matta-lt_ personally i'm more concerned with memory sharing 1111606771 M * matta-lt_ the binary size plus the inode cache usage 1111606778 M * matta-lt_ others might care more about saving disk space 1111606786 M * matta-lt_ Bertl: I use a proxy arp setup 1111606798 M * matta-lt_ each VM gets it's own psuedo interface 1111606806 M * Bertl with separate MAC? 1111606818 M * matta-lt_ so the start scripts add a route for the IP routed to their pseudo interface 1111606821 M * matta-lt_ yes, separate mac 1111606837 M * Bertl what MACs do you use for that? public range? 1111606841 M * matta-lt_ and the host server shows up as a "hop" 1111606844 M * matta-lt_ yeah 1111606849 M * matta-lt_ it's like aa:* 1111606851 M * matta-lt_ it public 1111606863 M * matta-lt_ examples: 1111606868 M * matta-lt_ AA:00:00:34:57:FE 1111606873 M * matta-lt_ AA:00:00:76:83:D6 1111606879 M * matta-lt_ AA:00:00:5B:35:5E 1111606886 M * Bertl i.c. and the host has to handle them ... 1111606891 M * matta-lt_ you get the picture now... it's randomly generated (or you can set it static) 1111606901 M * matta-lt_ it's proxy arp 1111606904 M * matta-lt_ i'd look into proxy arp 1111606906 M * matta-lt_ it rules 1111606915 M * matta-lt_ the only problem is forcing an update to upstream router/switch 1111606926 M * Bertl well, yes, but you cannot use those MACs on public lans ... 1111606932 M * matta-lt_ why not? 1111606935 M * matta-lt_ it works fine 1111606936 M * Bertl public lan, am I sleeping? 1111606966 M * Bertl no I meant, you cannot use those MACs on a network you do not controll 1111606990 M * matta-lt_ i'm not following, i just know it does work for xen 1111606995 M * Bertl (because the guy on the machine 2U higher might use the same ones) 1111607006 M * matta-lt_ due to proxy arp it never gets cached on other local servers 1111607020 M * matta-lt_ Bertl: for the random ones... I think it checks to see if it exists 1111607026 M * matta-lt_ arping or some similar technology 1111607038 M * matta-lt_ to make sure it's not duplicated on the lan 1111607041 M * matta-lt_ xen source for that one? 1111607047 M * Bertl hrm, well, yes that's a possibility, but no guarantee ... 1111607070 M * Bertl okay, anyway, so those MACs show up outside, right? 1111607075 M * matta-lt_ nope 1111607078 M * matta-lt_ for example 1111607081 M * matta-lt_ same lan... 1111607092 M * matta-lt_ 65.254.46.235 is a VM on another host 1111607101 M * Bertl k 1111607107 M * matta-lt_ --- 65.254.46.235 ping statistics --- 1111607107 M * matta-lt_ 3 packets transmitted, 3 received, 0% packet loss, time 2013ms 1111607107 M * matta-lt_ rtt min/avg/max/mdev = 0.454/2.666/7.017/3.076 ms, pipe 2 1111607107 M * matta-lt_ [root@vm3 cli]# arp -an | grep 65.254.46.235 1111607153 M * matta-lt_ oh, it only caches it if on the same subnet as the host 1111607163 M * matta-lt_ --- 65.254.46.1 ping statistics --- 1111607163 M * matta-lt_ 3 packets transmitted, 3 received, 0% packet loss, time 2019ms 1111607163 M * matta-lt_ rtt min/avg/max/mdev = 1.169/1.290/1.436/0.117 ms, pipe 2 1111607163 M * matta-lt_ [root@vm3 cli]# arp -an | grep 65.254.46.1 1111607183 M * matta-lt_ but.... 1111607184 M * matta-lt_ --- 72.9.242.1 ping statistics --- 1111607184 M * matta-lt_ 4 packets transmitted, 4 received, 0% packet loss, time 3022ms 1111607184 M * matta-lt_ rtt min/avg/max/mdev = 0.574/1.029/2.157/0.653 ms, pipe 2 1111607184 M * matta-lt_ [root@vm3 cli]# arp -an | grep 72.9.242.1 1111607185 M * matta-lt_ ? (72.9.242.199) at AA:00:00:34:57:FE [ether] on vif40.0 1111607186 M * matta-lt_ ? (72.9.242.196) at AA:00:00:57:DB:0E [ether] on vif47.0 1111607186 M * matta-lt_ ? (72.9.242.1) at 00:0D:BD:F6:5F:00 [ether] on eth0 1111607209 M * matta-lt_ it's all the same LAN, just different subnets 1111607227 M * Bertl hmm, okay, vif is? 1111607232 J * brc bruce@200165239109.user.veloxzone.com.br 1111607245 M * Bertl welcome brc! 1111607291 M * matta-lt_ the random interface it generates 1111607299 M * matta-lt_ what xen does is 1111607303 M * matta-lt_ internally it creates it 1111607318 M * matta-lt_ then it calls /etc/xen/scripts/vif-route with the args 1111607328 M * matta-lt_ ie. up, vif, ip, mac, etc 1111607340 M * matta-lt_ xen has already created the mac and vif id 1111607357 M * matta-lt_ vif i'm not sure what it is ... thought it was a tunnel, but not so sure 1111607377 M * matta-lt_ # ifconfig vif40.0 1111607377 M * matta-lt_ vif40.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF 1111607377 M * matta-lt_ inet addr:169.254.1.1 Bcast:169.254.1.1 Mask:255.255.255.255 1111607377 M * matta-lt_ UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 1111607377 M * matta-lt_ RX packets:5680 errors:0 dropped:0 overruns:0 frame:0 1111607378 M * matta-lt_ TX packets:9204 errors:0 dropped:56 overruns:0 carrier:0 1111607378 M * matta-lt_ collisions:0 txqueuelen:0 1111607380 M * matta-lt_ RX bytes:706196 (689.6 KiB) TX bytes:6928907 (6.6 MiB) 1111607380 M * matta-lt_ \ 1111607384 M * matta-lt_ that's on the host 1111607386 M * matta-lt_ it's eth0 inside the domain 1111607404 M * Bertl i.c. ... 1111607425 M * matta-lt_ so then you end up with something like... 1111607449 M * matta-lt_ # netstat -rn | grep 72.9.242.199 1111607451 M * matta-lt_ 72.9.242.199 0.0.0.0 255.255.255.255 UH 0 0 0 vif40.0 1111607457 M * matta-lt_ that is how IP access is controlled... 1111607479 M * matta-lt_ so then when the vm is destroyed it removes the vif and route 1111607483 M * matta-lt_ and starts over again 1111607894 M * Bertl any performace values regarding network? 1111607900 M * matti Bertl: ;] 1111607915 M * Bertl matti: `:] 1111608298 M * matta-lt_ Bertl: how could there be? 1111608325 M * Bertl well, I thought you maybe did some network throughput testing or such ... 1111608370 M * Bertl just curious how much overhead this setup adds ... 1111609747 M * matta-lt_ it's just routing 1111609758 M * matta-lt_ it's more more overhead than standard routing.... 1111609761 J * DukeGangsta|aw ~uefgangst@p54869AE6.dip.t-dialin.net 1111609763 M * matta-lt_ can saturate gigabit 1111609774 M * matta-lt_ now, if it is the bridging setup... then it limits you 1111609785 M * brc bertl! 1111609800 M * Bertl evening DukeGangsta|aw! 1111609814 M * DukeGangsta|aw hi 1111610580 J * Vudumen vudumen@perverz.hu 1111610638 M * Bertl wb Vudumen! 1111612437 J * prae ~prae@sherpadown.net 1111612527 M * Bertl evening prae! 1111612554 M * prae hi Bertl :) 1111613245 M * douglas hey bertly 1111613246 M * douglas err 1111613248 M * douglas -y 1111613251 M * douglas sorry 1111613270 M * douglas anyway, how would I enforce hardlimits on a vserver? whats the syntax I mean? 1111613305 M * Bertl depends on the limits ... 1111613308 M * Bertl rlimits? 1111613344 M * douglas rlimits I take it mean resource limits? 1111613371 M * douglas I'm talking in terms of cpu and memory usage. 1111613372 M * Bertl basically you 'just' set them ... when the vserver tries to get more resources than allowed, the request will be denied ... 1111613394 M * douglas ahh, what about soft limits, those are more perferred then hard limits of course right? 1111613408 M * Bertl cpu usage is not part of rlimits, you control that via the token bucket scheduler 1111613433 M * DukeGangsta|aw gn8 all 1111613435 Q * DukeGangsta|aw Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1111613451 M * Bertl douglas: in theory yes, in practice, the soft limits are not available yet ... 1111613497 M * douglas ok so whats the syntax in vserver1.conf for setting rlimits and where do I edit to control the token bucket scheduler, I'm sorry if my questions seem newbish, I dont quite understand everything :) 1111613721 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111613721 M * erwan_ho re 1111613764 M * Bertl douglas: vserver1.conf is old style (legacy) config and does not support rlimits 1111613819 M * douglas doh, ok I'm sure this is an obivous question, but what is the defacto standard now and is there a webpage that explains this? 1111614049 M * Bertl yep, sure, util-vserver (alpha tools) is defacto standard for 2.6 kernels 1111614049 Q * Vudumen Read error: Connection reset by peer 1111614069 M * Bertl http://linux-vserver.org/alpha+util-vserver 1111614086 M * Bertl see examples there, and the famous Flower Page for configuration details 1111614930 M * douglas bertl, vconfig is required, is there a vlan-utils package in debian? or is it only source? 1111614974 M * SiD3WiNDR I think I compiled it 1111614988 M * SiD3WiNDR yep, I didn't find a package 1111615266 J * Xearo ~steven@202.56.231.116 1111615282 M * Xearo i got a problem 1111615310 M * Xearo i'm tryin to get cyrus runnin on my fedora core 3 box 1111615362 M * Bertl evenening Xearo! 1111615367 M * Bertl douglas: vlan-utils? 1111615371 J * Vudumen vudumen@perverz.hu 1111615388 J * yarihm ~yarihm@80-218-4-232.dclient.hispeed.ch 1111615394 A * Bertl can't believe that debian has no package for that? 1111615417 M * Vudumen Bertl: for what? :) 1111615425 M * Zoiah Bertl: it's called vlan. 1111615430 M * Zoiah Bertl: and it's actually really neat. 1111615438 M * Vudumen (this machine sux. it's 5 years old and restarts 3-4 times a day :/) 1111615441 M * Bertl thought so ;) 1111615445 M * Zoiah auto eth0.11 1111615445 M * Zoiah iface eth0.11 inet static 1111615445 M * Zoiah address 10.1.1.1 1111615445 M * Zoiah netmask 255.255.255.0 1111615451 M * Zoiah And debian will take care of the rest. 1111615460 M * Bertl evening Zoiah, btw! 1111615467 M * Zoiah 'lo Bertl :) 1111615481 M * Bertl douglas: copy that? 1111615491 M * Xearo actually it's 3:36 am out here in India. Goodevenin to u Bertl 1111615529 M * Bertl ah, one of those famous half hour time zones, right? 1111615555 M * Bertl UTC +5:30 or so 1111615579 M * Xearo I have no Idea... 1111615584 M * Bertl Xearo: cyrus as in cyrus imap I guess ... 1111615598 M * Vudumen hmm i go to be. godd night 1111615602 M * Vudumen s/odd/ood/ 1111615604 M * Xearo cyrus Imap it is.... 1111615618 M * Bertl night Vudumen! 1111615626 M * Xearo I got a RPM for it... everythin was fin until it was time to login to it 1111615632 M * Bertl Xaand what issues do you encounter? 1111615646 M * Bertl *Xearo 1111615666 M * Xearo I dunno the default pass 1111615683 M * Bertl well, and that is in what way linux-vserver related? 1111615731 M * Xearo Well, I no way actually....Just came here cause there was no one at the fedora channel:) 1111615748 M * Bertl ahh, good idea, well, I use mandrake ;) 1111615754 M * Xearo *in 1111615793 M * Xearo ahh, that means u cant help me 1111615822 M * Bertl well, probably not ... did you read the man page/info for that beast? 1111615863 M * Xearo yes i did.... but all in vain 1111615910 M * douglas so the package is called vlan? 1111615944 M * douglas do I need to patch the kernel or enable an option? in debian is says you need a vlan patched kernel, does vserver patch it or is the discription on debian just wrong? 1111616018 M * Zoiah douglas: you need to enable 802.1q support. 1111616019 M * Bertl well, let me put it this way, you are better of with a non debian kernel for now ... (i.e. vanilla patched with latest linux-vserver patches) 1111616031 M * douglas bertl, of course. 1111616051 M * douglas so I do need to enable 802.1q? 1111616055 M * Bertl and there you can enable all the features you like to have .. vlans are not specifically used by linux-vserver 1111616067 M * Zoiah douglas: yes, you need to enable 802.1q support in the kernel. 1111616078 M * Bertl so if _you_ do not want to use them, you do not need to enable it 1111616086 M * douglas but vconfig is required to run devel releases 1111616122 M * Zoiah You need 802.1q support if you want to use vlans. 1111616133 M * Bertl util-vserver uses the vlan tools to configure vlan interfaces ... no vlan interfaces -> no use ;) 1111616142 M * douglas zoiah where is that in the kernel config? 1111616149 M * Zoiah dougles: network options... somewhere. :) 1111616156 M * douglas zoiah, exactly :) 1111616162 M * Zoiah douglas: I usually just edit the .config if I can't find something. :) 1111616180 M * douglas is tun/tap still required? 1111616191 M * Bertl tun/tap was never required 1111616198 M * douglas ok nevermind then :D 1111616279 M * Bertl ;) 1111616315 M * douglas oh for petes sake. they need to enable a search feature in menuconfig 1111616319 M * douglas I can't freakin fine vlan. 1111616322 M * douglas find 1111616341 A * Bertl travels back in time and implements it ... *tada* 1111616364 M * Bertl for Search 1111616374 M * douglas found it 1111616417 M * douglas recompiling 1111616418 M * douglas :D 1111616434 M * douglas time to go for a walk :) 1111616442 M * douglas or get in a quick game of unreal or quak3 1111616447 P * Xearo Leaving 1111616607 M * douglas wanna say thanks again bertl, been a big help. and you two zoiah, even tho you couldn't tell me where vlan was in the kernel! :) 1111616615 M * Zoiah ;) 1111616714 M * douglas I think I'm spoiled with faster technology nowadays 1111616749 M * douglas I member recompiling glibc about 6 years ago on my 486dx2, took 15 hours 1111616770 M * douglas went to bed, got up, took a shower, went to school, and came back and it was still compiling. 1111616803 M * Bertl douglas: you're welcome ... 1111616805 M * douglas maybe 8 years ago. I can't remember. 1111616844 M * douglas actually 10 years ago now that I've thought about it. 1111616847 M * Bertl kernel compiles should not take too long with 2.6, they are extremely short if the config is sane and you do a recompile for example ... 1111616865 M * douglas I always do a make clean. 1111616878 M * douglas its done 1111616975 M * douglas rebooting 1111617108 J * ciphernaut_zz ~a@61.88.18.130 1111617128 M * Bertl night ciphernaut_zz! ;) 1111617140 M * ciphernaut_zz night Bertl 1111617146 M * ciphernaut_zz whoa 1111617155 N * ciphernaut_zz ciphernaut 1111617553 M * Zoiah Bertl: ccache. :) 1111617575 M * Bertl does brake in too many cases ... 1111617642 M * Zoiah Never noticed it myself, but might happen, yes. 1111617765 Q * Vudumen Read error: Connection reset by peer 1111617980 M * jd86 hi bert 1111617993 M * Bertl evening jd86! 1111617995 M * jd86 sup 1111618004 M * Bertl 1.9.5 is almost done ... 1111618019 M * brc bertl 1111618027 M * brc do you have an idea about when it will be released ? 1111618051 M * brc i mostly wanted it because of the "bind thing" 1111618052 M * brc hehe 1111618068 M * Bertl hmm, yes ... what bind thing? 1111618086 M * brc being able to run named without paching it 1111618100 M * brc oops 1111618101 M * brc recompiling it 1111618102 M * Bertl ah, that bind thing ;) well should work with 1.9.4 too ;) 1111618109 M * brc hehe 1111618112 M * brc i am already with 1.9.4 1111618125 M * brc so i can use a normal bind package ? do i need to change anything ? 1111618187 M * Bertl configuration .. VXC_SET_RLIMIT 1111618221 M * brc Would that increase security risks ? 1111618294 M * Bertl not on the hosbutide the vserver, yes ... 1111618335 M * brc is it worth it ? :) 1111618355 M * Bertl excellent, I love it when the network connection does this ... 1111618387 M * Bertl no, not worth, bind is simply broken ... 1111618398 M * Bertl (well at least in the way it uses capabilities ;) 1111619453 M * brc well, setting the config flag it would be like a normal server, right ? 1111619455 M * brc the risky i mean 1111620589 M * Bertl yes, and no, basically you give CAP_SYS_RESOURCE (regarding rlimits) to everyone ... 1111620639 M * Bertl so root inside the vserver has less rights than a normal server, but user has slightly more rights than a normal serverver 1111620830 Q * erwan_ho Remote host closed the connection 1111621245 M * brc ah, ok ! gonna try it out 1111621280 M * douglas bertl, does 1.9.5-rc3 change that? or will people still need to use VXC_SET_RLIMIT ? 1111621336 M * douglas for the bind crap? 1111621344 M * Bertl this will be still needed with 1.9.5 1111621390 M * Bertl we might have different defaults in the future (probably with 1.9.6) 1111621467 Q * prae Quit: Pwet 1111621692 M * douglas ahh 1111621721 M * douglas so we would still need to recompile bind 1111621733 M * douglas unless we specify that VXC flag. 1111621745 M * Bertl with 1.9.5, yes ... 1111621806 M * douglas thanks 1111621816 M * Bertl np 1111621889 M * ciphernaut how can i kill a vserver by its context number? 1111621926 M * ciphernaut vserver-stat returns a vserver as running but without a name 1111622137 M * Bertl vkill