1156723714 Q * nebuchadnezzar Remote host closed the connection 1156724181 Q * Johnnie Quit: G'bye! 1156724193 J * Johnnie ~jdlewis@static-acs-24-154-32-33.zoominternet.net 1156724193 Q * Johnnie 1156724204 J * Johnnie ~jdlewis@static-acs-24-154-32-33.zoominternet.net 1156725153 J * nebuchadnezzar ~nebu@zion.asgardr.info 1156725249 N * matled_ matled 1156725686 J * glua ~glua@4-mail.net 1156725691 M * glua hi 1156725740 M * glua i've got some problems with quota inside a vserver..can anybody help me? 1156725786 M * daniel_hozac what's the problem? 1156726016 M * glua quotaon: using //aquota.user on /dev/hdv1 [/]: Function not implemented 1156726063 M * glua i've used this tutorial: http://linux-vserver.org/Standard+non-shared+quota 1156726198 M * daniel_hozac do you have the quota modules loaded? 1156726210 M * glua i compiled quota into the kernel 1156726227 M * daniel_hozac so grep QUOTA .config shows what? 1156726265 M * glua CONFIG_XFS_QUOTA=y 1156726265 M * glua CONFIG_QUOTA=y 1156726265 M * glua CONFIG_QUOTACTL=y 1156726305 M * daniel_hozac ah, are you using xfs? 1156726312 M * glua jep 1156726329 M * daniel_hozac what version are you using? 1156726342 M * daniel_hozac also, what does grep QFMT .config show? 1156726367 M * glua gentoo stable linux-2.6.15-vserver-2.0.1-r5 1156726388 M * glua # CONFIG_QFMT_V1 is not set 1156726388 M * glua CONFIG_QFMT_V2=y 1156726417 M * glua i just set V2=y..haven't recompiled it yet 1156726442 M * glua do i need v2? 1156726449 M * daniel_hozac well, yes. 1156726561 M * glua thx, i'll try it 1156726582 M * glua btw - do you speak german? 1156726676 M * daniel_hozac no, not really. 1156726715 M * glua well, i thought because uf your surname 1156726733 M * daniel_hozac http://irc.13thfloor.at/LOG/2006-08/LOG_2006-08-25.txt may be of interest to you, nayco set up quota on XFS (and found some bugs). 1156726796 M * daniel_hozac i didn't think kk was common in german. 1156726930 M * glua kk? 1156727099 M * glua gentoo doesn't serve any newer packages than 2.0.1..maybe i've to upgrade it myself 1156727110 M * glua but thanks for the hint 1156727115 M * daniel_hozac IIRC all the newer ebuilds were moved to an overlay. 1156727135 M * daniel_hozac but the bugs discovered shouldn't hit stable at all. 1156727201 M * glua i'll try to catch some sleep now 1156727209 M * daniel_hozac http://www.mail-archive.com/vserver@list.linux-vserver.org/msg09086.html 1156727212 M * glua thanks for your help 1156727221 M * daniel_hozac good night! 1156727261 M * glua thx..good night 1156727277 Q * glua Remote host closed the connection 1156727536 N * Bertl_oO Bertl 1156727542 M * Bertl back now ... 1156728302 M * daniel_hozac wb! 1156728340 M * Bertl tx 1156728403 M * daniel_hozac did you see the bug waldi reported? 1156729591 J * mire ~mire@212-167-222-85.COOL.ADSL.VLine.Verat.NET 1156729808 M * Bertl the remount issue? 1156729884 M * daniel_hozac yeah. 1156729922 M * Bertl we discussed that some time ago, and IIRC, even discussed a solution (waldi had a patch) 1156729954 M * Bertl any new issues or patches there? 1156730033 M * daniel_hozac hmm, are we talking about the same thing? you're allowed to remount a ,nodev filesystem as ,dev if you have secure_remount. 1156730078 M * daniel_hozac http://lophos.multibuild.org/linux/vserver-mount.patch 1156730089 M * Bertl yes, thats it .. problem is that the flags are not 'propagated' on remount 1156730170 M * Bertl IIRC, I suggested to move it down into the separate mounts, which seems to be done in this patch 1156730573 M * daniel_hozac did you also see the rbind problem? 1156730615 M * Bertl no? 1156730633 M * daniel_hozac 22:39 < waldi> Bertl_oO: recursive bind does only apply new flags to the top mount 1156730675 M * Bertl ah, okay, yes, that is generic problem of rbind 1156730760 M * daniel_hozac ok, so not something we want to fix, then? 1156731159 M * Bertl well, I'm not 100% sure we _can_ fix it that easily without adding bloat 1156731185 M * Bertl maybe the best approach (for now) would be to add nodev unconditionally for secure_* 1156731208 M * Bertl (to all mounts done inside a context) 1156731397 M * Bertl in the future I want to allow dev inside a guest without restrictions 1156731520 M * daniel_hozac yeah. 1156734743 Q * anonc Ping timeout: 480 seconds 1156737091 J * anonc ~anonc@staffnet.internode.com.au 1156737284 J * abi` ~abi@enz.schiach.de 1156737446 Q * abi Read error: Connection reset by peer 1156737697 Q * Zaki Ping timeout: 480 seconds 1156738362 M * Bertl okay, off to bed now .. have a good one! 1156738369 N * Bertl Bertl_zZ 1156740853 Q * derjohn2 Ping timeout: 480 seconds 1156741265 J * derjohn2 ~aj@dslb-084-058-213-136.pools.arcor-ip.net 1156743208 J * yarihm ~yarihm@whitehead2.nine.ch 1156744004 J * dna ~naucki@159-251-dsl.kielnet.net 1156746741 Q * Hollow Remote host closed the connection 1156746778 J * Hollow ~hollow@2001:a60:f026::1 1156746838 Q * Hollow Remote host closed the connection 1156746874 J * Hollow ~hollow@2001:a60:f026::1 1156747316 Q * olilo Remote host closed the connection 1156747346 J * olilo hiddenserv@tor.noreply.org 1156749822 J * meandtheshell ~markus@85-125-230-140.dynamic.xdsl-line.inode.at 1156750635 M * nayco_work Hello, all !!! 1156751411 J * phedny ~mark@volcano.p-bierman.nl 1156751735 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl 1156752954 M * tanjix hi together 1156753656 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156753656 Q * ntrs Read error: Connection reset by peer 1156754314 J * cdrx ~legoater@242.32.96-84.rev.gaoland.net 1156754341 M * cdrx hi 1156754816 Q * BenBen Ping timeout: 480 seconds 1156754831 Q * cemil Ping timeout: 480 seconds 1156755274 M * Hollow Bertl_zZ, daniel_hozac: http://home.xnull.de/work/vserver/press/20060828-announce-202-wiki.pdf 1156755428 J * BenBen ~benny@defiant.wavecon.de 1156755441 J * cemil ~cemil@defiant.wavecon.de 1156755653 J * coocoon ~coocoon@p54A05E63.dip.t-dialin.net 1156755990 J * pagano ~pagano@131.154.5.20 1156756460 J * eyck eyck@ghost.anime.pl 1156756495 M * tanjix when i try to enter a vserver with "vserver NAME enter" I get: chcontext: vc_new_s_context(): Invalid argument 1156756498 M * tanjix any ideas? 1156756837 M * daniel_hozac tanjix: what versions are you using? 1156757036 M * tanjix vs-master:~# vserver-info 1156757038 M * tanjix vs-master:~# vserver-info 1156757038 M * tanjix Versions: 1156757040 M * tanjix Versions: 1156757040 M * tanjix Kernel: 2.6.12.4-vs2.0 1156757042 M * tanjix Kernel: 2.6.12.4-vs2.0 1156757042 M * tanjix VS-API: 0x000100ff 1156757044 M * tanjix VS-API: 0x000100ff 1156757044 M * tanjix util-vserver: 0.30.204; Dec 20 2005, 16:58:50 1156757046 M * tanjix util-vserver: 0.30.204; Dec 20 2005, 16:58:50 1156757073 M * daniel_hozac Hollow: haven't we always supported ppc? 1156757116 M * Hollow daniel_hozac: well, i wondered too, but 2.0.1.2 introduced the powerpc syscall 1156757123 M * Hollow was sys_ni_syscall before 1156757141 M * Hollow at least interdiff says that 1156757149 M * daniel_hozac well, didn't it just get reworked in mainline? 1156757155 M * daniel_hozac ppc/ppc64 -> powerpc? 1156757165 M * Hollow ah, yeah.. probably 1156757195 Q * olilo Remote host closed the connection 1156757217 M * Hollow daniel_hozac: fixed ;) 1156757262 J * olilo hiddenserv@tor.noreply.org 1156757338 M * daniel_hozac was strace really broken in 2.0.1? didn't we introduce that bug with the reaper changes? 1156757353 M * Hollow i copied that from the changelog.. dunno 1156757372 M * tanjix daniel_hozac: unsure if mirc sent the version info - did you get them ? 1156757379 M * Hollow because interdiff sucks if kernelversion++ 1156757382 M * daniel_hozac tanjix: yep, twice for each line... 1156757386 M * daniel_hozac Hollow: http://daniel.hozac.com/vserver/2.0.2.deltas 1156757403 M * tanjix daniel_hozac: and you have any hint? ;) 1156757403 M * Hollow ah, nice 1156757453 M * Hollow daniel_hozac: this list is complete? 1156757456 Q * virtuoso Quit: leaving 1156757474 M * tanjix daniel_hozac: Bertls testme.sh gives secceeded for all tests 1156757479 J * virtuoso ~s0t0na@shisha.spb.ru 1156757482 M * daniel_hozac Hollow: AFAIK. there may be a delta or two missing, and i didn't think anything changed before 2.0.2-rc1. 1156757499 M * Hollow ok, thanks, will update the announcement 1156757523 M * Hollow what about BME? iirc bertl said it got into mainline? 1156757547 M * daniel_hozac i was wondering about that too the other day. 1156757564 Q * shedi Quit: Leaving 1156757603 M * daniel_hozac tanjix: it's likely the fact that you enabled the legacy version id that screws everything up. 1156757638 M * tanjix daniel_hozac: correct, it is legacy mode as i dont know yet the new config style 1156757654 M * daniel_hozac so you're using ancient utils as well? like 0.30? 1156757675 M * daniel_hozac though i guess 0.30.204 is ancient enough :) 1156757702 M * tanjix daniel_hozac: was this for me ? 1156757710 M * daniel_hozac yes. 1156757737 M * tanjix so i shall upograde or what should i do? 1156757754 M * daniel_hozac Hollow: i couldn't find anything like it in 2.6.18-rc4 or -mm3 though. 1156757778 M * daniel_hozac tanjix: any particular reason you're using such old versions? 1156757813 M * tanjix no - i set up a vserver host based on a howto on the linux-vserver website - the util-vserver comes from apt 1156757820 M * daniel_hozac i mean, 2.0 isn't even the most recent stable release. 1156757823 M * Hollow daniel_hozac: ok, i'll mention it in the announcement then.. 1156757853 M * daniel_hozac tanjix: yeah, i figured. onl 1156757879 M * daniel_hozac +y Debian supplies ancient util-vservers :) 1156757893 M * tanjix ;) 1156757911 M * tanjix so when i use self-compiled tools my problem will go away wuth newer tols ? 1156757911 M * daniel_hozac i don't think it would fix your problem, but you might want to add backports.org and get a more recent util-vserver. 1156757958 M * daniel_hozac i'm not sure. i'd say this is a kernel issue, but i don't really know how legacy works. 1156757981 M * tanjix and when using non-legacy ? 1156757982 M * daniel_hozac Hollow: did you see Bertl_zZ's paste with the changelog? 1156758008 M * tanjix because i want to change from legacy to non legacy but did not find help on how to do so? maybe you can assist me ? 1156758027 M * daniel_hozac Hollow: http://paste.linux-vserver.org/283 1156758041 M * daniel_hozac tanjix: just the config, or what? 1156758054 M * Hollow daniel_hozac: ah, even nicer 1156758068 M * tanjix yes, that i can use the new functions from vserver - just like cpu and ram throttling for each vserver 1156758115 M * daniel_hozac http://linux-vserver.org/ConfigConverters 1156758162 M * daniel_hozac tanjix: you would still be using the legacy APIs until you recompile your kernel (which you should do anyway, that kernel is _really_ old) and disable the legacy version id. 1156758182 M * daniel_hozac what howto were you following? 1156758239 M * tanjix daniel_hozac: http://linux-vserver.org/Step-by-Step+Guide+2.6 1156758261 M * daniel_hozac that uses 2.6.14.3-vs2.01. 1156758291 M * tanjix hm then it was another howto ;) 1156758299 M * tanjix i will re-setup the machine based on that howto ;) 1156758374 M * daniel_hozac Hollow: ource control management or a 1156758374 M * daniel_hozac public anonymous FTP archive. 1156758390 M * daniel_hozac wtf, stupid evince, anyway, shouldn't that or be and? 1156758493 M * Hollow yep, fixed 1156758563 M * daniel_hozac great work! 1156758651 M * Hollow what is "ext attribute sync" in bertls log? 1156758754 M * daniel_hozac hmm, i don't know. 1156758878 M * Hollow ok.. pdf has been updated 1156759025 M * daniel_hozac i guess we should probably mention the ccapabilities vulnerability that was fixed? 1156759071 M * daniel_hozac CVE-2006-2110 1156759128 M * Hollow ah, yep... the ccaps entry in bertls log didn't seem that important ;) 1156759144 M * daniel_hozac hehe. 1156759187 M * Hollow ok, updated 1156759231 M * tanjix daniel_hozac: the howto uses 2.6.14.3 - can i use the latest kernel 2.6.17.11 too ? 1156759251 M * Hollow hm.. the only question left is: how can i put all footnotes on the last page :o 1156759253 M * daniel_hozac tanjix: sure, but you'll have to use a different patch as well. 1156759262 M * tanjix is there a patch for that kernel ? 1156759277 M * daniel_hozac yes. 1156759299 N * Belu_zZz Belu 1156759364 M * daniel_hozac Hollow: i think it looks great now. 1156759378 M * Hollow ok, thanks for the logs! 1156759381 M * daniel_hozac tanjix: http://vserver.13thfloor.at/Experimental/ 1156759511 M * tanjix daniel_hozac: which one should i use ? 1156759608 M * daniel_hozac tanjix: patch-2.6.17.11-vs2.0.2-rc31.diff would seem like a wise choice, don't you think? 1156759653 M * tanjix sounds great ;) 1156759810 Q * cdrx Ping timeout: 480 seconds 1156759829 M * Hollow daniel_hozac: the pdf should look better now, with the changelog on one page 1156759899 M * daniel_hozac yeah, that does look better. 1156760390 M * tanjix daniel_hozac: what options should i select in the kernel for the new config style ? 1156760426 M * tanjix regarding legacy - you said something to enable/disable in kernel ? 1156760433 M * daniel_hozac as long as you don't select the legacy version id, it should be good. 1156760479 M * Hollow apropos... IMO - before 2.0.2 - the DISABLE_LEGACYNET should be renamed to ENABLE_LEGACYNET just like the other legacy options 1156760485 M * daniel_hozac you'll probably want the legacy support though, as it has some implications for how vshelper works. (and util-vserver's vshelper doesn't work the way it's supposed to yet) 1156760577 M * daniel_hozac does LEGACYNET still need to be its own option? 1156760614 M * Hollow well, consolidation of both options would be ok too 1156760632 M * Hollow but it is quite confusing like it is atm 1156760640 M * daniel_hozac yeah, i agree. 1156760665 M * matti Hi Hollow, daniel_hozac, * :) 1156760676 M * daniel_hozac hey matti! 1156760683 M * Hollow huhu matti ;) 1156760700 J * shake_spear ~chris@chello080108214151.4.13.univie.teleweb.at 1156760751 M * matti :> 1156760870 M * daniel_hozac Hollow: http://daniel.hozac.com/vserver/delta-config-clean01.diff 1156760896 M * daniel_hozac -depends on EXPERIMENTAL. 1156760941 M * waldi daniel_hozac: i can disable that if i use 0.30.210? 1156760947 M * Hollow yep looks good.. is VSERVER_NGNNET actually used? 1156760962 Q * shake_spear 1156760969 M * daniel_hozac waldi: that's the idea. 1156760998 M * daniel_hozac Hollow: nope, at least not in stable. 1156760999 A * Hollow doesn't even remember running legacy API ;) 1156761035 M * Hollow daniel_hozac: could be dropped then, no? 1156761063 M * daniel_hozac i guess. 1156761081 M * Hollow well, since it is not visible to the user it's doesn't matter really.. 1156761103 M * daniel_hozac yeah. 1156761244 M * Hollow haha... ikea now has pax support 1156761259 M * Hollow ikea.de/PAX 1156761260 M * Hollow ;) 1156761279 M * daniel_hozac they've had pax for quite some time :) 1156761280 M * Wonka lol 1156761294 M * Hollow hehe 1156761379 J * shedi ~siggi@dsl-149-109-85.hive.is 1156761717 N * abi` abi 1156761916 Q * kaner Remote host closed the connection 1156762019 J * kaner kaner@strace.org 1156763131 M * Hollow daniel_hozac: anny suggestions to whom the announcement should be sent? i currently have kerneltrap, newsforge, slashdot, linux-mag and some local german news pages.. 1156763142 M * waldi daniel_hozac: do you know, why vserver checks for the secure_mount capability instead of priviledged context? the later check should be more clear and don't break of some other code in the caps check is broken 1156763178 M * daniel_hozac Hollow: lwn? 1156763195 M * Hollow ah yep 1156763205 M * daniel_hozac waldi: hmm? privileged context? 1156763228 M * waldi hmm, this may change the behaviour ... 1156763242 M * daniel_hozac isn't a privileged context one which has the secure_mount ccapability? 1156763251 M * waldi no, it is the context 0 1156763266 M * daniel_hozac context 0 doesn't have ccapabilities at all. 1156763273 M * daniel_hozac so the check doesn't apply. 1156763357 M * Hollow what about lkml? 1156763382 M * Hollow not sure if announcements should be sent to lkml 1156763464 M * daniel_hozac me neither... i guess it would be ok though. 1156763496 M * daniel_hozac http://www.gatago.com/linux/kernel/14660378.html heh :) 1156763514 M * Hollow hehe 1156763722 M * waldi daniel_hozac: hmm, a quick fix may be: vx_ccaps(VXC_SECURE_MOUNT) ||vx_ccaps(VXC_SECURE_REMOUNT) 1156763755 M * daniel_hozac waldi: did you see Bertl_zZ's response last night? 1156763804 M * waldi yep, but I won't propose this patch for debian stable update 1156763833 M * waldi and checking for secure_remount at the same place should be a valid fix 1156763855 M * daniel_hozac yeah. 1156763963 M * cehteh i am up to try to use bridge devices instead eth's for my vserver, i guess that will cause no problem or does anyone have other experiences? 1156765122 Q * olilo Remote host closed the connection 1156765149 J * olilo hiddenserv@tor.noreply.org 1156766004 M * derjohn daniel_hozac, waldi : ideas about rc32 ? 1156766045 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1156766170 M * daniel_hozac i'd rather see 2.0.2 ;) 1156766353 M * daniel_hozac Hollow: http://daniel.hozac.com/vserver/util-vserver-0.30.210-vshelper.patch what do you think? (ugly as hell for now, just a hack to test the theory behind it) 1156766463 M * Hollow uhf 1156766482 M * Hollow ugly syncing... ;) 1156766500 M * daniel_hozac very. 1156766520 M * Hollow back then my ugly patch looked like this: http://home.xnull.de/misc/vshelper-workaround.patch 1156766543 M * Hollow but probably even hackier :D 1156766560 M * daniel_hozac that won't execute the stop scripts for init-less guests though. 1156766575 M * Hollow ah yes.. init-less crap.. 1156766597 M * Hollow but it seems i'm forced to reintroduce it to vcd ;) 1156766647 M * daniel_hozac hehehe :) what convinced you? 1156766650 M * Hollow bertl 1156766652 M * Hollow :P 1156766667 M * daniel_hozac heh. 1156766865 M * Hollow he said no init-less support is K.O. for him 1156766866 M * Hollow ;) 1156766923 M * daniel_hozac ah. 1156767760 M * nayco_work http://wiki.linux-vserver.org/Welcome_to_Linux-VServer.org Nice job, guys ! 1156767782 M * nayco_work test 1156767801 M * nayco_work *damn irrsi scroll* 1156767818 M * daniel_hozac hehe. 1156767832 M * Wonka nayco_work: failed 1156768236 M * Hollow nayco_work: thanks 1156768304 M * Hollow ah, finally got my coffee machine back from repair :) 1156768341 M * Hollow vitally important 1156768342 M * Hollow ;) 1156768346 M * tanjix daniel_hozac: any idea about the following issue: 1156768356 M * daniel_hozac Hollow: lol. 1156768364 M * nayco_work Hollow: I read your annoncement... Do you need an announcement in a french highly-visited linux site ;-) ? 1156768367 M * tanjix daniel_hozac: http://paste.linux-vserver.org/310 1156768397 M * Hollow nayco_work: sure, i'm actually gathering sites 1156768422 M * daniel_hozac tanjix: that's a known issue with testme, try http://www.13thfloor.at/~doener/vserver/tools/testme-legnet.sh instead. 1156768444 M * sid3windr omg that mediawiki looks great 1156768452 M * Hollow tztz.. doeners fix is around for ages now 1156768452 M * tanjix daniel_hozac: that works - thanks ;) 1156768473 M * daniel_hozac indeed, over 4 months old :) 1156768484 M * Hollow sid3windr: yeah, because it does not look like mediawiki :D 1156768489 A * daniel_hozac delayed-pokes Bertl_zZ 1156768494 M * derjohn Hollow: announcement? 1156768508 M * Hollow derjohn: http://home.xnull.de/work/vserver/press/20060828-announce-202-wiki.pdf 1156768514 M * Hollow beware this is not official yet 1156768520 M * Hollow please don't spread the link 1156768553 M * derjohn Hollow, as I didnt transfer all FAQ yet. I build a template first and now do s'n'r on the old wiki src. 1156768574 M * Hollow yep, i already saw it, nice job so far 1156768599 M * derjohn Hollow, nice paper ;) !! 1156768604 M * Hollow :) 1156768613 M * derjohn Hollow, can I get a "peoples" account, too? 1156768618 M * Hollow sure 1156768622 M * Hollow ssh pubkey? 1156768630 M * nayco_work Hollow: Ok, when do you need it ? What should be included in the announcement as a summary (Your message is too long to translate, so, summary is needeed !) 1156768649 M * derjohn Hollow, and: I would like to offer a Mirror. As soon as you decide to make it mirror-able, let me know 1156768651 M * nayco_work "14:35 < Hollow> please don't spread the link" Nope ;-) 1156768653 M * Hollow nayco_work: it would be ok to just translate the abstract (the first two paragraphs without headline) 1156768684 M * Hollow and then link to the announcement or so 1156768709 M * sid3windr I can mirror as well if needed 1156768713 M * Hollow derjohn: you're talking about the ftp archives? 1156768733 M * derjohn Hollow, in an ideal world I would mirror the whole thing. 1156768749 M * Hollow well, i don't think mirroring the wiki makes sense 1156768750 M * derjohn but I dont have a concept to sync to mysql .... 1156768758 M * nayco_work Hollow: Ok for the abstract. I'll include the rest of the article in english. Oh, what is the license for the article, I mean, can I freely reproduce it entirely (One never knows... ;-)) ? 1156768759 M * sid3windr mysql has replication :) 1156768767 M * Hollow mysql replication is horrible 1156768768 M * derjohn well, sense != practical issues 1156768773 M * derjohn Hollow, ack. 1156768793 M * Hollow nayco_work: sure, public-domain, as-is, GNU FDL, whatever, i don't care :D 1156768794 M * sid3windr is it? I'm using it for a bunch of things (with 4.1) and haven't had any sync problems yet 1156768804 M * nayco_work Hollow: Oh, and which is the date for the official announcement ? I mean, the one before nothing must be said ? 1156768808 M * derjohn no replication, that causes more problems than it solves - at least in the of a wiki mirror 1156768810 M * Hollow sid3windr: try to change the collation of one server -> boom 1156768825 M * daniel_hozac nayco_work: preferably not before 2.0.2 is actually released ;) 1156768825 M * derjohn sid3windr, hm, what I one of the mirrors goes down? 1156768829 M * nayco_work Hollow: ok for the license. 1156768837 M * sid3windr derjohn: ...? :) 1156768866 M * derjohn sid3windr, you donr have the bin logs on that machine and have to resync (?) somwehow 1156768867 M * nayco_work daniel_hozac: And when will it happen ? 1156768874 M * Hollow i'm curious how long the announcement has to wait... *looks at bertl* 1156768875 M * Hollow ;) 1156768876 M * daniel_hozac Hollow: the patch seems to work fine, so i'll start cleaning it up. 1156768881 M * sid3windr that works fine for me ... :) 1156768896 M * sid3windr anyway, mirroring a wiki via replication is indeed not the way to go :) 1156768899 M * derjohn sid3windr, as long as the main wiki does not get slow ... 1156768907 M * Hollow nayco_work: i _guess_ that it will be announced in 1-3 days 1156768913 M * sid3windr but if you need mirror space & bandwidth just gimme a yell :) 1156768929 M * Hollow sid3windr: sure, TIA 1156768966 M * sid3windr it will be hosted inside a vserver ;> 1156768969 M * Hollow derjohn: well, the box for the new webservices is idle most of the time anyway 1156768975 M * Hollow sid3windr: indeed 1156768977 M * Hollow :D 1156768981 M * Hollow s/will/is/ 1156769001 M * sid3windr uhu, if I'd mirror it as well, hehe :) 1156769012 M * Hollow ah ;) 1156769034 M * Hollow daniel_hozac: sounds good, tbh i don't know much about all that syncing stuff in util-vserver 1156769043 M * harry making new grsec + vserver patch!!!!!! :) 1156769056 M * Hollow harry: want an announcement? *gg* 1156769059 M * nayco_work Hollow: Ok, I ask this because there are delays on linuxfr.org for validating/moderating articles ;-) 1156769091 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156769095 M * daniel_hozac that's probably true for all the places the announcement will be sent to (except lkml). 1156769103 M * Hollow nayco_work: yeah, i guess there will be no difference for other news sites 1156769115 M * nayco_work good point 1156769211 J * id23 ~id@p50811AD0.dip0.t-ipconnect.de 1156769232 M * id23 greetings #vserver 1156769248 M * Hollow hi id23 1156769314 M * harry Hollow: nono... just informing... :) 1156769325 M * harry but a statue would be fine :) 1156769329 M * Hollow harry: just joking ;) 1156769335 M * id23 ? insider ;) ? 1156769340 M * id23 k 1156769346 M * harry Hollow: i still want a statue :p 1156769363 A * Hollow goes ordering a statue for harry 1156769365 M * Hollow ;) 1156769368 M * harry :) 1156769382 M * Hollow will place it my vserver altar then.. 1156769388 M * id23 hehe 1156769393 M * Hollow :) 1156769453 M * harry http://paste.linux-vserver.org/311 1156769455 M * harry grmvl 1156769458 M * harry grmbl 1156769473 M * Hollow aweful 1156769487 M * harry yeah 1156769496 A * harry interdiffs rc28-rc31 1156769524 M * harry blerk... won't work 1156769680 J * coocoon ~coocoon@p54A054C2.dip.t-dialin.net 1156769686 M * derjohn harry will have much for some hours :) 1156769688 Q * cehteh Remote host closed the connection 1156769704 M * harry derjohn: naah... /me will fix it quite fast 1156769712 M * derjohn harry, what kind of statue? 1156769719 M * harry logic thinking + experience :) 1156769737 M * derjohn harry, i'll chip in for a large gina wild, if you do grsec+devel ;) 1156769748 J * cehteh ~ct@cehteh.homeunix.org 1156769749 M * harry derjohn: 5m at least, i must look devine and be placed @ the vatican... with a pentagram around my neck 1156769779 A * harry wants to see ratzi's face :) 1156769829 M * derjohn i guess they wont accept a statue of gina wild with a pentagram at the vatican :) 1156769853 M * matti Hmm... 1156769860 M * matti New patches from harry? 1156769866 M * matti Where? When? Who?! 1156769867 M * matti ;] 1156769882 A * matti hugs harry ;-) 1156769889 M * derjohn harry, i'll switch to you grsec kernel as soon as stable get the capability masking feature. /me hopes for 2.0.3. 1156769914 M * matti derjohn: Capability masking/ 1156769926 M * matti s/\//\?/ 1156769957 M * id23 gina - lol - you enjoy too much digital fulfilment 1156769967 M * id23 ;) 1156769986 M * id23 but i am pretty sure razzi would be very happy ... 1156769994 M * derjohn id23 she used to live about 10km away from here. I only know this starlet from the press of course ;) 1156769994 M * matti Who is razzi? 1156770006 M * derjohn matti, the current pope 1156770016 M * matti Oh. 1156770019 M * matti And gina? 1156770022 M * id23 hehe 1156770035 M * id23 a ex porn star ? 1156770043 M * matti Well. 1156770048 M * matti I am only simple geek ;p 1156770054 M * derjohn matti, that makes it possible to run bind9 und pureftp within a guest, without patching the stuff (no need ro recompile with --nocapability) 1156770056 M * matti I do not know any ex port stars ;p 1156770058 M * daniel_hozac harry: why didn't you use the interdiffs? :) 1156770080 J * lylix ~eric@dynamic-acs-24-154-53-234.zoominternet.net 1156770087 M * matti derjohn: I see. Sounds nice. 1156770237 M * harry daniel_hozac: i would... if i could 1156770255 M * harry config:/usr/local/config/kernel# interdiff -p1 patch-2.6.17.11-vs2.0.2-rc31.diff patch-2.6.17.7-vs2.0.2-rc28.diff 1156770258 M * harry 1 out of 1 hunk FAILED -- saving rejects to file /tmp/interdiff-1.TEiW9z.rej 1156770261 M * harry interdiff: Error applying patch1 to reconstructed file 1156770274 M * daniel_hozac ok, so they're not strictly interdiffs. 1156770387 M * matti harry: This may not work - too much was changes in kernel itself :< 1156770406 M * daniel_hozac mostly just the Makefile and timer.c. 1156770457 M * daniel_hozac harry: want the interdiffs? or are you just using the individual deltas? 1156770513 Q * sladen Ping timeout: 480 seconds 1156770761 J * sladen paul@starsky.19inch.net 1156770805 M * harry daniel_hozac: don't know yet 1156770817 M * harry i'll see what i'm gonna do 1156770829 M * derjohn sladen, hello! Did you make an progress in the ubuntu/VServer "project"? I admit, that I didnt write the spec yet, but made a edgy kernel with VS 1156770839 J * glua ~glua@4-mail.net 1156770857 M * glua hi 1156770864 M * id23 hi glua 1156770909 M * glua does anybody know whether vserver supports ipv6 in the near future? 1156770931 M * derjohn glua, bonbons knows, he made the current beta patch for that ;) 1156770985 M * glua is the patch included in the dev. sources? 1156771001 M * derjohn glua, you in the the experimental srces? no, it not in yet 1156771008 M * derjohn *it's 1156771026 M * harry daniel_hozac: can you give me the url to the interdiffs? 1156771035 M * derjohn glua, ppl reported it to work, I didnt test myself yet. 1156771117 M * glua i've seen, there is a link to a ipv6-patch, do i have to patch util-vserver also? 1156771142 M * derjohn glua, hm, i think so, but do know know. daniel_hozac ? 1156771411 M * glua i like vserver, because it doesn't depend on a commercial product like openvz 1156771421 M * harry daniel_hozac: !!!!!!!! 1156771428 M * harry didlyduuuuuuuuude? 1156771436 M * harry where areth thou! ;) 1156771442 M * harry or something to that effedt 1156771443 M * harry effect 1156771511 M * glua but openvz supports ipv6 1156771630 M * harry there is some IPv6 support i think 1156771667 M * harry http://daniel.hozac.com/vserver/delta-2.6.17.4-vs2.1.1-rc26-ipv6.patch <== don't know ifyou want to use that... but this means there is some upport :) 1156771810 M * glua i'll try it 1156771819 M * harry ask daniel_hozac first 1156771850 M * Hollow doener: around? 1156771939 M * glua k.. 1156772040 M * glua i think there are only some small details that should be fixed in vserver. then it would be perfect for productional use 1156772047 M * Hollow daniel_hozac: do you have time to move your vserver stuff to your new devspace until the announcement? 1156772066 M * Hollow glua: we do our best ;) 1156772090 M * Hollow but many people use it in production (including me) 1156772097 M * glua just 2 things..xfs quota support and ipv6 support :) 1156772107 M * Hollow xfs quota will be in 2.1.1 1156772123 M * Hollow ipv6 has to wait until 2.3 1156772172 N * Bertl_zZ Bertl 1156772176 M * Bertl morning folks! 1156772179 M * Hollow morning Bertl 1156772183 M * harry hey Bertl ! 1156772184 M * glua ipv6 is a very important feature for me..i hope, the patches will work 1156772187 M * id23 hi Bertl 1156772187 M * glua hi berb 1156772191 M * glua bertl 1156772196 M * harry Bertl: do you have interdiffs for rc28-29-30-31? 1156772201 M * Bertl glua: indeed, patches for xfs quota work fine, nayco_work tested them 1156772234 M * Bertl glua: ipv6 on the host works too (out of the box) and bonbons patches give you guest support too 1156772242 M * Hollow Bertl: http://home.xnull.de/work/vserver/press/20060828-announce-202-wiki.pdf 1156772258 M * Bertl harry: hmm, the deltas do not suffice? 1156772274 M * harry Bertl: sure , if you tell me which ones :) 1156772286 M * Bertl harry: thing is, the kernel version changed inbetween, which is the major change in 29-30 1156772310 M * Hollow yep, interdiff is broken if kernelversion++ 1156772316 M * Hollow i noticed that too this morning ;) 1156772333 M * harry this kinda sucks 1156772340 M * Bertl well, interdiff breaks if you look at it the wrong way :) 1156772357 M * harry thats why i have to interdiff -p1 1156772362 M * Bertl harry: wouldn't it be better to keep your changes in a patch? 1156772363 M * harry but ... 1156772374 M * harry i have that.. 1156772382 M * Hollow let's praise the day when our git trees are setup ;) 1156772386 M * Bertl so you can _simply apply that ontop of rc31, e.g. 1156772387 M * harry linux-2.6.17.11-g-vsrc28 1156772400 M * harry so i have a 2.6.17.11 with grsec (latest) 1156772406 M * harry patched now, with rc28 vserver 1156772426 M * harry created from a diff from rc28 kernel 1156772441 M * harry but now, i need to upgrade to rc31 1156772455 M * harry so kernel: ok, grsec: ok... vserver @ rc28 level 1156772460 M * harry now up to 31 :) 1156772570 M * harry i can even do them manually 1156772584 M * harry it would be nice to have interdiffs tough... 1156772602 M * harry then i can allways just use those to update my patchset 1156772606 M * Bertl check the changelog and which deltas are in devel 1156772661 M * Bertl then simply apply them, but be careful, as the 2.6.17.8-> 11 change _did_ break vserver (without adjustments) 1156772701 M * harry yeah, i know :) 1156772740 M * harry http://daniel.hozac.com/vserver/2.0.2.deltas <== correct? 1156772778 M * Bertl that for stable, but derjohn is also doing a changelog 1156772815 M * derjohn Bertl, as far as the community feeds /me with informatiion: Yes :) 1156772830 M * derjohn harry, http://linux-vserver.org/ChangeLogStableExperimental 1156772834 M * harry Bertl: i don't do patches for 2.1.2-rc stuff 1156772839 M * Bertl and if you look at http://vserver.13thfloor.at/Experimental/ 1156772848 M * derjohn 2.1.2 ??? 1156772865 M * Bertl you also see all (potential) changes in chronological order 1156772877 A * glua is back 1156772900 M * Hollow Bertl: did you reach enrico? 1156772918 M * Bertl nope, didn't answer yet, will try again today 1156772933 M * Hollow ok 1156772947 M * Hollow do we release today? 1156772960 A * mnemoc raise the ears 1156772963 M * harry bloody hell 1156772987 M * harry i need daniel_hozac 's interdiff 1156772992 M * harry this won't work 1156772996 M * Hollow thing is: i'm in berlin for a week starting tomorrow morning and will only have limited net access 1156772997 M * daniel_hozac why? the deltas don't work? 1156772999 M * Bertl not sure, have to double check the implications of the secure_mount stuff and something else I forgot :/ 1156773044 M * harry daniel_hozac: if i know which delta's 1156773055 M * harry but your deltas differ from derjohn's deltas 1156773058 M * Bertl Hollow: when exactly will you be back? 1156773059 M * harry so which ones are the right ones 1156773076 M * harry and how will i EVER be able to use just rcXX-rcXY interdiffs? 1156773078 M * Hollow Bertl: september 2nd at around 22pm 1156773087 M * derjohn harry, mine? then they differ from /dev/null ;) 1156773115 M * daniel_hozac harry: my list of deltas should be correct, if not, let me know. 1156773123 M * Bertl okay, so I'd suggest we simple postpone it till then, should give me enough time to get 2.1 almost ready too, and maybe one or the other fix for stable 1156773159 M * derjohn Bertl, s/almost// ;) 1156773177 M * Hollow Bertl: ok, fine with me too, there are still some things (beside patches) to do as well 1156773184 M * Bertl then we do a week of releases :) including tools? 1156773187 M * Hollow e.g. your logo enhancements? 1156773189 M * daniel_hozac Bertl: did you see http://daniel.hozac.com/vserver/delta-config-clean01.diff ? 1156773208 M * derjohn tools? will they move to the new secret svn, too? 1156773217 M * Hollow Bertl: hm.. not sure about vcd release.. we'll see 1156773227 M * Bertl daniel_hozac: yep, looks fine 1156773238 M * daniel_hozac Bertl: i usually /msg him and he eventually gets around to answering, at least he has so far. 1156773251 M * Hollow derjohn: probably 1156773262 M * harry + bool "Enable Legacy Networking Kernel API" 1156773264 M * harry + default y 1156773268 M * harry + depends on EXPERIMENTAL && !VSERVER_LEGACYNET 1156773269 M * harry + default n 1156773308 M * harry i think i better wait some more for a stable 2.0.2 release it seems ;s 1156773369 M * daniel_hozac why? there will almost always be a new release just around the corner :) 1156773378 M * daniel_hozac (by release i mean new patch) 1156773411 M * daniel_hozac Hollow: what stuff? daniel.hozac.com/vserver/? 1156773423 M * Hollow daniel_hozac: yep 1156773460 M * daniel_hozac sure, i'll move it. i guess i should move my other patches there too (findutils, portmap, bind, etc.) 1156773471 M * Hollow ok, great.. 1156773491 M * Hollow Bertl: i have setup people.linux-vserver.org for dev webspace... can you add the dns entry? 1156773503 M * Hollow and do you have time to migrate your data as well? 1156773506 M * daniel_hozac so, to where do i move it? helios.dev.croup.de:~/public_html? 1156773512 M * Hollow daniel_hozac: exactly 1156773519 M * daniel_hozac ok, will do. 1156773523 M * Hollow thanks 1156773536 M * Bertl Hollow: I can add that, what do you mean with 'your' data? 1156773547 M * Hollow vserver.13thfloor.at/Experimental 1156773558 M * daniel_hozac not /Stuff? :) 1156773560 M * Hollow and probably /stuff 1156773562 M * Hollow heh 1156773585 M * Bertl hmm, not at the moment, it's my direct link between my development system and the web guest 1156773585 M * daniel_hozac oh, and can we get a non-legacynet compatible testme released as well? 1156773618 M * Bertl Hollow: will se what I can do with that, but definitely needs some special adjustments to make certain scripts work for me 1156773624 M * Hollow Bertl: ok, if you need some special setup for your devspace just tell me 1156773633 M * Hollow heh 1156773636 M * matti Eh. 1156773694 M * matti Hi Bertl. 1156773695 M * Bertl but we have an svn and a git repostiroy now, right? 1156773711 M * Bertl are they working and backed up? 1156773730 M * daniel_hozac isn't git distributed? 1156773733 M * Bertl because if so, I could start moving stuff in there 1156773739 M * Hollow Bertl: yep, regarding git i just cloned a linux-2.6 tree, so i'm not sure the setup meets our requirements 1156773753 M * Hollow Bertl: best would be to send me your ssh pubkey first :) 1156773761 M * Bertl yep, right :) 1156773783 M * Bertl will do so in a few minutes ... 1156773790 M * Hollow sure, np.. 1156773819 M * Hollow Bertl, daniel_hozac: http://home.xnull.de/work/vserver/press/20060828-announce-202-wiki.TODO anything else? 1156773846 M * daniel_hozac harry: if you still want the interdiffs, let me know and i'll upload them. 1156773955 M * harry daniel_hozac: i'm at rc29 now 1156773959 M * daniel_hozac looks fine to me. 1156773959 M * harry (i hope :s) 1156774008 M * daniel_hozac http://daniel.hozac.com/vserver/delta-2.0.2-rc29-rc30.diff http://daniel.hozac.com/vserver/delta-2.0.2-rc30-rc31.diff 1156774011 A * harry did an interdiff for rc28-rc29 1156774015 M * harry ahaaaa 1156774039 M * harry i'll do the 29-30 by hand... 1156774051 M * daniel_hozac http://daniel.hozac.com/vserver/delta-2.0.2-rc28-rc29.diff is my -rc28-rc29 diff. 1156774052 A * matti cross fingers for harry :) 1156774104 M * harry identical to mine, daniel_hozac :) 1156774111 M * matti :-) 1156774138 J * lilalinux ~plasma@h1-gw.of.net-lab.net 1156774182 A * matti hands harry coffee. 1156774200 A * matti hands daniel_hozac coffee and cookie :) 1156774221 M * daniel_hozac mmm, cookies! :) 1156774263 M * Bertl careful, might be one from google :) 1156774271 M * matti Hehehe. 1156774276 M * matti Bertl: :-))) 1156774284 M * Hollow or one with thc in it 1156774360 M * matti No. 1156774366 M * Hollow :p 1156774372 M * matti This is the good old cookie. 1156774373 M * matti :) 1156774425 M * harry done! 1156774554 M * matti harry: Ready?! The -rc31? 1156774554 M * matti ;] 1156774582 M * harry -rw-r--r-- 1 root staff 1422438 Aug 28 16:16 patch-2.6.17.11-vs2.0.2-rc31-grsec2.1.9.diff 1156774584 Q * yarihm Quit: Leaving 1156774585 M * harry wiiiiiiii 1156774598 A * matti hands harry a cookie ;) 1156774608 M * matti harry: If you want a THC one, ask Hollow ;p 1156774618 M * Hollow i don't share my drugs 1156774619 M * Hollow :P 1156774624 M * matti Hehehe. 1156774625 M * harry hehe 1156774631 A * harry now 0xc0ffee 1156774631 M * matti Hollow++ 1156774633 M * Hollow too expensive ;) 1156774635 M * harry then compiling the kernel 1156774646 M * harry it's online now... so if you want to test it... feel free 1156774650 M * harry if not: mkay;) 1156774658 A * matti goin' to test it. 1156774659 M * harry let me know if there are any probs 1156774668 M * harry but now... /me gone for a min :) 1156774951 M * sladen derjohn: haven't done anything with it except talk to BenC (Ubuntu kernel dude) in Paris 1156774999 M * derjohn sladen, yes, I mailed him aldreay and got answers :) 1156775018 M * sladen derjohn: what answers did you get? 1156775037 M * derjohn i'll forward the mail to you 1156775039 M * sladen derjohn: I think he tends to like git trees, so we may need to ask Bertl to set one up 1156775137 M * Bertl sladen: I'm already on it :) 1156775166 M * Bertl i.e. after the 2.1 release, (including 2.02 and 2.1.1) we will switch to git 1156775183 M * derjohn sladen, cool. BTW: do you run a ubuntu vserver host ? 1156775358 M * daniel_hozac Hollow: could you extend the name field in the directory listings? http://helios.dev.croup.de/~dhozac/patches/util-vserver/ for instance is pretty non-descriptive. ;) 1156775420 M * matti harry: You probably should remove the "dontdiff" from your patch :) 1156775467 M * matti harry: It looks kinda funny, hehe :-) 1156775567 J * Andrico ~Andrico@lncswibas01-pool5-a97.lncswi.tds.net 1156775608 M * matti k, WiKi updated. 1156775702 M * Andrico rad. 1156775718 M * matti We need a slightly longer/bigger topics on OFTC :) 1156775737 M * Andrico ok. 1156775741 T * matti http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc31, stable+grsec 2.0.2-rc31 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1156775757 M * matti Andrico: Hello stranger! 1156775759 M * matti :) 1156775808 M * sladen derjohn: nope, just Debians ones; but the difference really isn't much. When I next install a box it'll probably be Ubuntu 1156775925 M * sladen derjohn: (I tend to normally have Ubuntu CDs on me) :) 1156775949 M * derjohn sladen, I'll looking for someone, who I could warmwalk to make unionfs BME aware. I was only able to null it out ;) 1156775962 M * Andrico Hello, matti. 1156775988 M * matti :> 1156776006 M * sladen derjohn: I'm lost, can you give me some more context, where are you using unionfs? 1156776013 M * sladen derjohn: and what's 'BME' ? 1156776192 M * derjohn sladen, edgy's kernel has unionfs patched in. I wanted to provide a clean to that applies to edgy's kernel src. BME is bind-mount-extension, which will appear in mainline soon anyway. 1156776338 M * kir anybody have vserver guest images for powerpc? 1156776526 M * harry ? 1156776571 M * Bertl kir: yes, I have some on my powerbook? 1156776669 M * kir Bertl, could you put those to some accessible place? the best one would be FC4, FC5 or smth like that... 1156776676 M * sladen derjohn: is that an extensions to the namespace stuff? 1156776685 M * sladen derjohn: unionfs is used for the LiveCD 1156776701 M * Bertl kir: only debian and mandrake here, and it will take some time (no fast internet access here) 1156776714 M * Bertl kir: what do you plan? 1156776723 M * kir Bertl, oh...in that case just don't worry 1156776727 M * derjohn sladen, yes, you could also use unionfs for something like base guest setup to spare space (like unification) 1156776748 M * derjohn sladen, BME is located in the VFS 1156776761 M * kir Bertl, looks like I have ported openvz to ppc (oh, well...let's say it boots:) and wanted to run running something :) 1156776781 M * kir Bertl, no troubles, I will port my vzpkg tools tomorrow and will create some images... 1156776789 M * Bertl well, would be simple to get util-vserver and use that to network install guests? 1156776825 M * derjohn kir, cant you 'debootstrap ' ? 1156776832 M * kir vzpkgtools is pretty much the same as util-vserver 1156776838 M * Hollow daniel_hozac: i could, but you can do it yourself too, with .htaccess 1156776848 M * derjohn you wouldnt need external tools, nor? 1156776856 M * kir derjohn, the host is suse :) and I know that gentoo is the only non-debian that has "debootstrap" in repository 1156776875 M * kir derjohn, well I can probably find debootstrap for suse (or compile from sources:)) 1156776883 M * Bertl kir: debootstrap works on all distros in util-vserver 1156776900 M * derjohn kir, sry: I dint have any PPC running here (at least not anymore ....) 1156776942 M * kir derjohn, so you had it before? what happened? 1156776980 M * Hollow gentoo has yum and rpm too, btw 1156777003 M * kir Hollow, yep......although rpm is terribly old and conflicts with beecrypt from stable 1156777010 M * Hollow i know.. *bah* 1156777029 M * Hollow guess we have to resolve that conflict sooner or later 1156777047 M * kir Hollow, the problem with rpm is 4.2, 4.3 and 4.4 has an incompatible database format 1156777082 M * kir Hollow, so if you want to host VPSs that use rpm-4.3 inside and want to install packages to that VPS from the host system, you have to have rpm-4.3 on the host system 1156777082 M * Hollow i have absolutely NFC about rpm *g* 1156777094 M * Hollow i see.. 1156777096 M * Hollow the rpm hell 1156777097 M * Hollow ;) 1156777103 M * kir Hollow, so you have to end up with zoo of three rpm versions... 1156777108 P * Andrico 1156777156 M * kir in fact you can use earlier version and then later version of rpm -- but after later version touches the DB you can not go back :) 1156777240 M * derjohn kir, well, the PPC 603e got a little sluggish ... ah, and we had a G4 as 'desktop', but even browsing got slow, no flash player AFAIR ... and some guy brought the machine from us. 1156777240 A * Hollow hugs portage 1156777362 M * kir Hollow, portage db is also non-compatible across versions as I understand...so I am not sure can you use host portage from Gentoo 2006 to emerge something for guest running gentoo, say, 2004-something 1156777429 M * Hollow kir: don't confuse profiles with portage versions... but yep.. the cache format has changed, although this does not influence installed packages, it just made package searches faster 1156777465 M * Hollow OTOH, i never tried portage as external package manager 1156777472 M * Hollow i use it only from inside 1156777820 J * hallyn ~xa@adsl-75-2-92-167.dsl.chcgil.sbcglobal.net 1156777886 M * Bertl wb hallyn! 1156777893 A * Belu is away (i´ll be back later...) 1156777895 N * Belu Belu_zZz 1156777965 Q * ||Cobra|| Remote host closed the connection 1156778427 J * Zaki ~Zaki@212.118.126.43 1156778821 J * stefani ~stefani@tsipoor.banerian.org 1156779445 J * yarihm ~yarihm@84-75-121-93.dclient.hispeed.ch 1156779536 M * Bertl Hollow: hmm, mutt messed up your name alias, did bounce you the email again ... 1156779655 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156779786 J * coocoon ~coocoon@p54A054C2.dip.t-dialin.net 1156779826 Q * FireEgl Read error: Connection reset by peer 1156780077 M * matti harry: k, pasting. 1156780078 M * matti http://romke.net/paste/d99T 1156780093 M * matti Anyone have some hints about this nasty warnings? 1156780111 M * matti harry also uses smp, but he didn't noticed any of those. 1156780377 J * FireEgl FireEgl@Sebastian.Atlantica.US 1156780848 M * Hollow Bertl: hehe.. what about username@linux-vserver.org as aliases to our real mailbox? 1156780884 M * Hollow Bertl: username? bertl? 1156780907 M * Bertl yeah, good idea, but it actually was a bug in mutt not escaping the umlaut 1156780962 Q * hallyn Quit: leaving 1156781014 M * Hollow Bertl: ok, should work.. 85.10.237.61 1156781077 M * waldi daniel_hozac: CVE-2006-4243 1156781093 M * Bertl Hollow: yep, works 1156781104 J * Viper0482 ~Viper0482@p54976744.dip.t-dialin.net 1156781137 M * Hollow Bertl: /var/git 1156781175 M * daniel_hozac waldi: ok. 1156781226 M * daniel_hozac Hollow: another CVE to add to the announcement ;) 1156781232 M * Hollow :) 1156781269 M * waldi which announcement? 1156781305 M * Hollow Bertl: actually, if you change the MX record for linux-vserver.org it should already work, user@linux-vserver.org goes to ~/.Maildir/ if no ~/.forward configured 1156781320 M * Hollow waldi: http://home.xnull.de/work/vserver/press/20060828-announce-202-wiki.pdf 1156781392 M * waldi ah 1156781419 M * Bertl okay, will do so soon 1156781438 M * waldi Bertl: how do we want to fix that? 1156781553 M * Bertl waldi: for now, I think we will go for unconditional nodev on every mount/remount inside 1156781573 M * Bertl waldi: long term solution is to allow dev inside guests 1156781606 M * waldi change the check to vx_ccaps(VXC_SECURE_MOUNT) ||vx_ccaps(VXC_SECURE_REMOUNT) ? 1156781683 M * Bertl more like, if (!vx_check(0,VX_ADMIN)) ... |= nodev 1156781695 M * waldi ah 1156781745 M * waldi hmm, does root in the vserver have CAP_SYS_MOUNT? 1156781783 M * Bertl by default no 1156781793 M * Bertl i.e. we could use that as condition too 1156781805 M * waldi looks weird than 1156781813 M * waldi this check is okay 1156781981 J * essobi_ ~kstone@216.26.137.75 1156781991 M * essobi_ Afternoon. 1156781994 M * waldi for the debian kernels i'll adopt that to the dev-bindmount, i need that currently on any machine which hosts buildds 1156782067 M * phreak`` harry: use diff -Nup :) its much more readable :P 1156782079 M * harry http://ludit.kuleuven.be/software/vserver/ipt_stealth.patch 1156782081 M * harry there :) 1156782095 M * phreak`` harry: better ;) 1156782146 M * essobi_ Is there a way to see which guests are consuming the most CPU usage on the host server? 1156782200 M * daniel_hozac not with some fancy utility. 1156782230 M * essobi_ daniel_hozac Umm. Okay.. HOW do I see the CPU utilization of each guest then? 1156782234 M * daniel_hozac /proc/virtual/*/sched should provide you with some hints though. 1156782241 M * essobi_ roger that.. 1156782253 M * daniel_hozac (or alternatively /proc/virtual/*/cvirt's load average) 1156782254 M * harry tnx to Bertl for helping me out :) 1156782263 M * essobi_ I'm green to the whole Vserver project but all be in far froma unix newbie.. 1156782271 M * essobi_ s/in/it 1156782282 M * harry lets see what grsec people think about the patch... 1156782325 M * daniel_hozac Bertl: i have to say, i like waldi's fix better. 1156782361 M * essobi_ Mmm... I see a /proc/virtual/49153/... where'd the 49153 come from? Is that the process ID of my guest1? 1156782375 M * essobi_ Mmm. Newp.. not a PID.. 1156782388 M * daniel_hozac it's the xid. 1156782396 M * daniel_hozac and you are using dynamic ones. 1156782408 M * daniel_hozac it would be better if you set a static one. 1156782424 M * essobi_ O... 1156782433 M * essobi_ Mmm. 1156782469 M * essobi_ .. /etc/vservers? 1156782540 M * daniel_hozac /etc/vservers/.../context, to be exact. 1156782547 M * daniel_hozac but you should stop the guest before setting it. 1156782556 M * daniel_hozac (or it won't be able to shut it down) 1156782578 M * essobi_ roger that.. 1156782675 M * essobi_ find /etc/vservers -name "context" is turning up empty.. 1156782689 M * daniel_hozac right, you'll have to create it. 1156782694 M * essobi_ Ahh. 1156782705 M * daniel_hozac the lack of it is what's making it use dynamic contexts. 1156782734 M * essobi_ so... echo "1" > /etc/vservers/vs1/context would assign XID 1 to vs1? 1156782750 M * daniel_hozac theoretically, but xid 1 is special (it's the spectator context). 1156782760 M * daniel_hozac 2-49151 are valid static contexts though. 1156782761 M * essobi_ oh.. 1156782784 M * essobi_ Danke. 1156783098 A * Hollow is stunned 1156783115 M * Hollow http://www.heise.de/newsticker/meldung/77382 1156783126 M * waldi why? 1156783142 M * Hollow do i need to translate? ;) 1156783157 M * daniel_hozac yes please :) 1156783229 M * Hollow basically, there was a artwork project in our parliament, and a politician made a photograph, posted it in her blog, and got admonished for not paying license fees 1156783252 M * Hollow the interesting thing here is: tax payers payed for it 1156783294 M * Hollow and is exhibited in a public building 1156783296 M * Hollow wtf.. 1156783354 A * Hollow feels like he'd moved to america 1156783359 M * daniel_hozac lol 1156783513 M * cehteh heh ... ever tried to google for "great flower page" ? :) 1156783531 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156783535 J * adm-ybanafa ~admin@198.69.12.92 1156783542 M * daniel_hozac haha. 1156783548 M * Hollow cehteh: actually you won't find the flower page on the first page of results if you search for "util-vserver" ;) 1156783564 M * daniel_hozac hmm, second hit for me. 1156783569 M * Hollow o.o 1156783582 M * cehteh first 2 or 3 hits for me 1156783598 M * Hollow indeed... well, it's been a while since i searched for it with "util-vserver" :P 1156783608 M * daniel_hozac heh. 1156783678 M * cehteh (while i would prefer if the theme was build on some kinds of mushrooms ;)) 1156783781 M * Hollow hehe 1156783794 P * adm-ybanafa 1156784354 M * essobi_ So vtop is just like top but it shows ps top for of all the servers at once? 1156784561 P * lylix 1156784894 M * Bertl essobi_: yep, it is actually just a wrapper around top 1156786195 J * coocoon ~coocoon@p54A069AA.dip.t-dialin.net 1156786606 M * essobi_ Neat.. 1156786619 M * essobi_ Anyone used OpenVPS? 1156786892 M * daniel_hozac i never quite understood what it was... 1156787000 M * essobi_ Looks like a web interface built to manage your vserver guests.. 1156787032 M * daniel_hozac OpenVCP also does that. 1156787088 Q * shedi Quit: Leaving 1156787424 J * |gerrit| ~kvirc@dslb-084-060-234-250.pools.arcor-ip.net 1156787488 J * shedi ~siggi@dsl-149-109-85.hive.is 1156787974 Q * lilalinux Remote host closed the connection 1156788436 J * bonbons ~bonbons@83.222.36.236 1156788519 Q * olilo Quit: brb 1156788531 M * nayco_work For those of you who read french : http://nayco.free.fr/wiki/doku.php?id=linux-vserver-quota <= XFS disk quotas inside vservers ;-) 1156788552 M * daniel_hozac i.e. without regular quota? 1156788895 J * olilo hiddenserv@tor.noreply.org 1156789468 Q * Viper0482 Remote host closed the connection 1156789694 M * nayco_work daniel_hozac: What do you mean ? Maybe I use the wrong words to describe the purpose of this doc, so it is important to know ;-) 1156789744 M * nayco_work daniel_hozac: this doc is about user disk quotas with XFS (and certainly valid for others fs), using the quota-tools inside the guest (setquota, repquota, ...) 1156789939 M * daniel_hozac ok, so not the XFS specific quota then. 1156790016 M * nayco_work Well, In fact, I only tried with XFS. I never use anything else ;-) But I should try with ext2/3 or others : Maybe this works too... 1156790031 J * ybanafa ~admin@198.69.12.7 1156790048 M * nayco_work Ok, "galagohome". 1156790112 P * ybanafa 1156790471 M * ntrs_ Hi all. 1156790495 M * ntrs_ Bertl, is there some known problem with devel and server performance and especially mysql issues/corruptions? 1156790534 M * ntrs_ I upgraded a few servers to latest devel and customers from every single one is complaining about mysql problems and database corruptions 1156790611 M * ntrs_ Hello? 1156790625 M * ntrs_ I guess nobody is here atm. 1156790889 Q * FireEgl Ping timeout: 480 seconds 1156790956 Q * shedi Quit: Leaving 1156791507 Q * michal_ Ping timeout: 480 seconds 1156791637 M * daniel_hozac ntrs_: so what version were you using previously? what's causing the corruption? 1156791690 M * Bertl ntrs_: not that we would know of? 1156791720 M * ntrs_ Bertl, the only thing that was changed was the kernel upgrade 1156791734 M * Bertl well, maybe the kernel changed something? 1156791760 M * Bertl but I doubt that it would cause corruption in userspace 1156791795 M * phreak`` harry: seems Brad liked your patch :P 1156791841 M * ntrs_ Bertl, the guests are generally quite slower as well 1156791848 J * michal_ ~michal@www.rsbac.org 1156791951 M * Bertl ntrs_: then something is wrong, performance in testing is much higher with devel than stable 1156791977 M * Bertl but I have to leave now, so I'd suggest double checking with stable, just to make sure 1156791984 N * Bertl Bertl_oO 1156792997 J * FireEgl FireEgl@Sebastian.Atlantica.US 1156793069 J * shedi ~siggi@inferno.lhi.is 1156793363 Q * glua Quit: using sirc version 2.211+KSIRC/1.3.12 1156793402 M * ntrs_ daniel_hozac, it seems that mysql instances are just being forked all the time and are never closed and then when killed are corrupting the database. 1156793458 M * daniel_hozac sounds more like a MySQL problem than anything else, IMHO. 1156793479 M * ex check swap, maybe mysql instances takes too much ram, and kernel kills it 1156793497 M * ex there should be some OOM killer notices in dmesg, and some info in mysql.err 1156793651 Q * |gerrit| Remote host closed the connection 1156793682 M * ex if OOM kills mysql there should be such record: 16:19:31 h1:Out of memory:err Killed process 6592 (mysqld) 1156793704 M * ex mysql restarts (mysqld_safe forks new mysql instance), but tables are crashed 1156794063 M * ntrs_ ex, no there is no such thing. There are no overages in /proc/.../limits 1156794096 M * daniel_hozac OOM wouldn't necessarily generate a hit on the memory limit. 1156794128 M * daniel_hozac for instance, if the entire host is OOM. 1156794223 M * ex daniel_hozac, I've MySQL in root, and 5 vservers which uses it (apache, etc). when vservers was taking too much ram (1G + 1G swap) OOM killed mysql 1156794244 M * ex maybe there are some OOM tuning.. dont know :) 1156794717 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156794724 Q * ntrs_ Read error: Connection reset by peer 1156794897 M * derjohn is there any particular reasin, why a rlimit "AS" is called "VM" in /proc/virtual//limit ? 1156794902 M * derjohn *reason 1156794978 M * daniel_hozac probably not :) 1156795153 M * derjohn if we would call it VM, when most people would understand it, without any expainations. adding /etc/vservers//rlimit/vm as some kind of alias to "AS" ... this would be only a part of the utils only? 1156795297 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156795297 Q * ntrs__ Read error: Connection reset by peer 1156795331 M * daniel_hozac i suppose. 1156795695 M * derjohn cehteh, more funny is even the if you google for a "great flower", the GFP is the second hit ;) 1156795805 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156795805 Q * ntrs Read error: Connection reset by peer 1156796337 J * restill ~restill@c-24-11-195-139.hsd1.mi.comcast.net 1156796533 Q * bonbons Quit: Leaving 1156797154 Q * restill Quit: Leaving 1156797415 M * ntrs_ there must be some problem with devel. more and more customers are reporting mysql problems after the upgrade to 2.1.1-rc29 1156797431 M * daniel_hozac why did you upgrade to an old version? :) 1156797476 M * ntrs_ It was not old when I upgraded. Did anything important change from .rc29 to .rc31? 1156797497 M * daniel_hozac http://linux-vserver.org/ChangeLogDevelExperimental 1156797519 M * daniel_hozac nothing relevant. 1156797530 M * daniel_hozac you still haven't told us what version you used earlier, i.e. what didn't show the problem. 1156797534 M * ntrs_ I think there is nothing relevant too. 1156797559 M * ntrs_ previously it was 2.6.16.16 with 2.0.2 rc20 1156797748 M * daniel_hozac so you upgraded to a major new version of the kernel as well as switched tree. 1156797761 M * ntrs_ yes 1156797783 M * daniel_hozac have you tested 2.0.2-rc31 yet? 1156797805 M * ntrs_ another interesting thing is that the loadavg on a guest is 1.00 1.00 1.00 even though the cpu with top is 99% idle 1156797811 M * ntrs_ No, have not tried that yet 1156797839 M * ntrs_ Would it be too complicated to put the named patch from devel in stable? 1156797853 M * daniel_hozac hmm? 1156797859 M * daniel_hozac ah, you mean the capability masking? 1156797860 Q * meandtheshell Quit: bye bye ... 1156797998 M * ntrs_ yes 1156798031 M * ntrs_ from top: 1156798033 M * ntrs_ top - 17:46:59 up 2 days, 5:13, 3 users, load average: 1.00, 1.00, 1.00 1156798034 M * ntrs_ Tasks: 153 total, 1 running, 152 sleeping, 0 stopped, 0 zombie 1156798034 M * ntrs_ Cpu(s): 0.0% us, 0.5% sy, 0.0% ni, 99.5% id, 0.0% wa, 0.0% hi, 0.0% si 1156798034 M * ntrs_ Mem: 1048576k total, 121608k used, 926968k free, 0k buffers 1156798034 M * ntrs_ Swap: 0k total, 0k used, 0k free, 0k cached 1156798060 M * daniel_hozac what about it? 1156798185 Q * cdrx Ping timeout: 480 seconds 1156798222 M * ntrs_ lod average 1.00 with 99.5% idle 1156798229 M * ntrs_ isn't it weird to say the least? 1156798288 M * ntrs_ Is it also normal for the swap to be absent from the guest? 1156798307 J * nayco ~nayco@lns-bzn-26-82-254-108-216.adsl.proxad.net 1156798342 M * daniel_hozac do you have swap on the host? 1156798345 M * ntrs_ yes 1156798350 M * daniel_hozac virt_mem? 1156798358 M * ntrs_ on the host 1156798359 M * ntrs_ # free 1156798359 M * ntrs_ total used free shared buffers cached 1156798359 M * ntrs_ Mem: 902736 868504 34232 0 50776 506076 1156798359 M * ntrs_ -/+ buffers/cache: 311652 591084 1156798360 M * ntrs_ Swap: 3068404 576 3067828 1156798378 M * ntrs_ yes virt_mem is enabled 1156798385 M * daniel_hozac that'd be why then. 1156798423 M * daniel_hozac load average of 1 with 99.5% could just mean that you have some process stuck in D state. 1156798423 M * ntrs_ hmm, did that change from stable into devel? I can see the swap just fine inside guests on stable. 1156798521 M * daniel_hozac what AS/RSS limits do you have? that's the most deciding factor. 1156798555 M * ntrs_ no limitonly RSS set to 262144 1156798565 M * daniel_hozac no AS limit at all? 1156798621 M * daniel_hozac i don't know if that is handled at all... 1156798644 M * daniel_hozac and does it really make sense to limit the guest to 1 GiB of RAM when your host has less? 1156798699 M * daniel_hozac hmm, looks like swap should shine through to the guest if there's no AS limit. 1156798870 M * derjohn ntrs_, where is your 'top' from. guest? 1156798882 M * ntrs_ yes 1156798899 M * derjohn and you enables virt loadavg ? 1156798903 M * derjohn *enabled 1156798920 M * derjohn what does the top on the host say? 1156799039 M * ntrs_ sched_prio 1156799040 M * ntrs_ virt_mem 1156799040 M * ntrs_ virt_uptime 1156799040 M * ntrs_ virt_cpu 1156799040 M * ntrs_ ^19 1156799040 M * ntrs_ hide_mount 1156799042 M * ntrs_ hide_netif 1156799048 M * ntrs_ that is all that is set for the guest 1156799060 M * derjohn what load does the host show? 1156799115 M * derjohn BTW: ^19 is a really silly line for a config.... yes, it it right, but who the hell makes such "config options2 ß 1156799147 J * mkhl mkhl@200-148-41-184.dsl.telesp.net.br 1156799239 M * ntrs_ the tools were unaware of that option when the guest was initially created and it works just fine. That was the only way to add that option at the time. And it was done by recommendation from Bertl 1156799253 Q * dna Quit: Verlassend 1156799440 M * daniel_hozac yep. 1156799481 M * ntrs_ Ok, so the question still remains why the swap is no longer visible inside a guest 1156799531 M * ntrs_ daniel_hozac, can the capability masking be done easily in stable? 1156799551 M * daniel_hozac shouldn't be _too_ hard. 1156799857 A * waldi decides to not add swap to this machine ... 1156799871 M * daniel_hozac the 32 GiB RAM one? 1156799918 M * waldi yes 1156799944 M * waldi it have only 140 GiB or disk space 1156799966 M * daniel_hozac damn. 1156800223 M * ntrs_ Is there any way I could prevent a single guest from using any swap? 1156800246 M * waldi bad idea 1156800254 M * daniel_hozac and no. 1156800527 M * daniel_hozac ntrs_: ok, i can reproduce the no-swap issue. 1156800536 M * ntrs_ Ok, great 1156800539 M * ntrs_ another thing... 1156800564 M * ntrs_ I have a D process inside of a guest. Is a hardware reboot the only fix? 1156800595 M * daniel_hozac in quite a few cases, yes. 1156800620 M * daniel_hozac it depends on why it's in D state. 1156800941 N * Bertl_oO Bertl 1156800948 M * Bertl okay, back ... 1156800956 M * Bertl so what did you figure till now? 1156800966 A * Bertl is reading up ... 1156801001 Q * mkhl 1156801014 M * daniel_hozac vx_vsi_swapinfo just returns all zeroes. 1156801048 M * daniel_hozac oh, i was looking at the wrong tree :| 1156801059 M * Bertl that's normal (and intended) for devel with virt_mem 1156801077 M * Bertl when the hard and the soft limit have the same values 1156801086 M * Bertl (RSS limits that is) 1156801089 M * daniel_hozac yeah, i realize that now. 1156801101 M * daniel_hozac i was looking at the stable tree for some reason. 1156801135 M * Bertl derjohn, daniel_hozac: regarding VM vs AS, I'd prefer to use AS, because it _is not_ VM what we account and limit there :) 1156801161 M * Bertl we might (at some point) have a special VM limit in addition to the existing ones) 1156801188 M * Bertl ntrs_: did you check for a kernel panic or stack dump in the logs? 1156801192 M * daniel_hozac yeah, this is what i thought. 1156801237 M * Bertl ntrs_: a process in 'D' state often means that the 'device' did vanish or some lock is still held by the kernel (which almost always makes a kernel unresponsive) 1156801254 M * ntrs_ Bertl, are you saying that it is now normal for swap not to be shown inside a guest? 1156801256 M * Bertl (or at least strangely behaving) 1156801273 M * Bertl ntrs_: with virt_mem, you get a virtualized view, which shows: 1156801278 M * derjohn Bertl, then the value in /proc shoudl be called AS. Calling it VM causes confusion, nor? 1156801293 M * daniel_hozac derjohn: agreed. 1156801306 M * Bertl RSS soft limit as memory max, and RSS hadr limit - soft limit as swap 1156801317 M * Bertl derjohn: correct 1156801353 M * ntrs_ Bertl, does that mean that this server will not use any swap? 1156801363 M * derjohn Bertl, BTW (always forgot to ask) if i set sched_hard and sched_prio .. which one is taken? 1156801404 M * daniel_hozac ntrs_: no. 1156801410 M * Bertl derjohn: both 1156801484 M * ntrs_ so what is the point of the soft and hard rss? 1156801486 M * ntrs_ http://paste.linux-vserver.org/315 1156801487 M * derjohn Bertl, foo? *kopfkratz* could you shortly point out, what happens if the guest gets out of tokens? 1156801491 M * ntrs_ kernel panic from the host with devel with "D" process 1156801525 M * derjohn Bertl, I mean: the scheduler can "stop" the guest or it can take it down to loooooow prio. not both. 1156801533 M * ntrs_ http://paste.linux-vserver.org/316 1156801542 M * daniel_hozac ntrs_: after the soft limit is reached, the guest's pages likelyhood of being swapped out will increase. 1156801562 M * ntrs_ http://paste.linux-vserver.org/317 1156801570 M * ntrs_ Ok, got it 1156801576 N * ntrs_ ntrs 1156801588 M * Bertl derjohn: priorities do not depend on putting contexts on hold, they are 'adjusted' according to the fill level of the bucket 1156801599 M * daniel_hozac Bertl: http://helios.dev.croup.de/~dhozac/patches/util-vserver/experimental/util-vserver-0.30.210-initenv.patch (wow, the URLs got really long from all that directorization, heh) 1156801624 M * Bertl actually too long for my taste 1156801675 M * daniel_hozac yeah, me too... 1156801693 M * Bertl ah, great, adds the env stuff, right? 1156801700 M * daniel_hozac i'll rework the directory structure... 1156801701 M * daniel_hozac yeah. 1156801708 M * daniel_hozac does it look sane? 1156801743 M * Bertl not sure, no idea atm what it does? 1156801747 J * PowerKe ~icuser@d5153A0DA.access.telenet.be 1156801780 M * Bertl adds a config file, called environment, which gets sourced, yes? 1156801797 M * daniel_hozac apps/init/environment, one variable per line. 1156801801 M * Bertl welcome PowerKe 1156801815 M * Bertl daniel_hozac: okay, but we also need a default for that, see 'man init' 1156801816 M * PowerKe Hey 1156801823 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156801828 Q * cehteh Ping timeout: 480 seconds 1156801829 M * daniel_hozac Bertl: yeah, i agree. 1156801834 M * Bertl daniel_hozac: no problem if that default is placed into some file though 1156801850 Q * ntrs Read error: Connection reset by peer 1156801868 M * derjohn Bertl, i think i dont get it.... I thought if the context is out of tokens, there are two possibe ways to proceed, until new tokens are filled in: wait 'hard' or run 'low prio'. I think my understanding how it works is wrong? both's scheds share the same config ..../schedule , nor? 1156801890 M * Bertl ntrs_: do you have a build tree of that kernel? to do some addr2line stuff on that? 1156801934 M * ntrs_ yes I do 1156801946 M * Bertl derjohn: yeah, you got it wrong .. the 'shared' part of both scheduling decisions (and the one and only criterion, actually) is the token bucket 1156801979 M * Bertl ntrs_: okay, let's start with addr2line -e vmlinux c018c260 1156802005 M * derjohn ... and of course the token buckets fill state. 1156802021 M * Bertl derjohn: now, the priority based 'scheduling element' cals a vavavoom value from the bucket fill level 1156802027 M * Bertl *calcs 1156802047 M * Bertl and uses that to slightly adjust the task priority (up/down 50%) 1156802069 M * derjohn task == all processes of that guest? 1156802087 M * Bertl while the hard scheduling 'element' only looks at the range between min fill and nothing 1156802123 M * Bertl and acts as a hysteresis loop, putting tasks on hold when the bucket is empty, rescheduling them when min fill is reached 1156802145 M * derjohn so, I i set both scheds, it can happen that the processes are set lowprio and then they are set on hold (in low prio state) ? 1156802159 M * tanjix daniel_hozac: there ? 1156802164 M * daniel_hozac sure. 1156802169 M * Bertl derjohn: 'tasks' _might_ be all processes of a context, but in general, some tasks do not run always, so they are neither scheduled nor putted on hold 1156802181 M * Bertl derjohn: yes 1156802187 M * derjohn *ouch* 1156802192 M * tanjix daniel_hozac: you gave me the converter for the config files today - i tried that but it will not work? 1156802204 M * daniel_hozac tanjix: how come? 1156802215 M * Bertl hey tanjix! LTNS! :) 1156802216 M * daniel_hozac (it should be noted i've never used it, but i assume whoever wrote it used it) 1156802226 M * tanjix hi Bertl! 1156802228 M * derjohn now i can understand one of customer complaints ;) 1156802239 M * Bertl daniel_hozac: might be problem between keyboard and chair ... 1156802261 M * tanjix daniel_hozac: what do you mean? 1156802266 M * Bertl derjohn: which was? 1156802279 M * daniel_hozac tanjix: why won't it work? 1156802294 M * derjohn "even my shell doenst really react more then one chars per two seconds" 1156802331 M * derjohn and: I gave that guest 16/32 CPU ... 1156802348 M * tanjix daniel_hozac: if i understood it right, these scripts convert from config style to directory style (in order to not using legacy mode) - but i cannot start the vservers. the tools say they were built without legacy supprt (ok, thats another problem) but why dont they start with new style ? 1156802393 M * daniel_hozac tanjix: have you removed the old configs? do you have directories in /etc/vservers (assuming that's your destination directory)? how did you compile your kernel? 1156802475 M * tanjix daniel_hozac: no, i did not remove the config file - i looked to /etc/vservers/ now - there is still the config file and no other things (directory called by the vservers name) 1156802501 M * tanjix daniel_hozac: and what do you want to know regarding the kernel ? 1156802521 M * derjohn Bertl, no. it was on sched_hard only as I see just now. but under constant pressure of a large sendmail setup, that might "pause" all incl. bash 1156802571 M * ntrs_ derjohn, what is a good sane sched config you normally use? 1156802571 M * tanjix daniel_hozac: i must correct: in /etc/vservers there is a directory called S_HOSTNAME="ubuntu" 1156802582 M * Bertl derjohn: 'normal' interactivity 'guesstimations' of the kernel still apply to the guests 1156802582 M * tanjix that matches the 1st line of the config file 1156802605 Q * yarihm Quit: Leaving 1156802606 M * derjohn ntrs_, depends on the CPU and number of guest and the phase of the moon. 1156802612 Q * mire Quit: Leaving 1156802626 M * Bertl ntrs_: do we have any results yet? 1156802664 M * derjohn ntrs_, I aim towards 10 guest per single CPU COTS machine, with 1/4 to 1/2 OF CPU power 1156802678 M * Bertl ntrs_: hum, let me ask you a simple question, regarding the 315'er pastebin upload 1156802682 M * ntrs_ Bertl, sorry for the delay 1156802684 M * ntrs_ # addr2line -e vmlinux c018c260 1156802684 M * ntrs_ ??:0 1156802696 M * Bertl ah, your addr2line is broken too, so we need gdb 1156802723 M * ntrs_ derjohn, what are the actual numbers in the config file? 1156802736 M * ntrs_ Bertl, ok gdb it is. what do you need from gdb? 1156802745 M * Bertl 'gdb vmlinux 1156802760 M * Bertl then l *0xc018c260 1156802791 M * derjohn ntrs_, that a big company secret! http://paste.linux-vserver.org/318 . Now it's no secret anymore. 1156802844 M * derjohn ntrs_, the script i posted is my first try to automate the setting of my guests ... just a proof of concept. 1156802858 M * Bertl derjohn: what is HZ set to? 1156802862 M * derjohn 250 1156802886 M * derjohn (in former times 1000 ... as long as you poined out thats shit for servers) 1156802886 M * Bertl so the context will get on hold for 400/250 seconds everytime :) 1156802900 M * Bertl 'roughly 400 that is' 1156802910 M * derjohn 400 ? how so ? 1156802918 M * derjohn ah 1156802918 Q * ntrs_ Read error: Connection reset by peer 1156802927 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156802929 M * derjohn *klong* *coin fallen* 1156802932 M * tanjix daniel_hozac: or might it be i have uses the converter wrong? 1156802936 M * Bertl 200*32/14 = 457 1/7 th 1156802959 M * derjohn a large pause ... 1156802966 M * Bertl means 1.828 seconds pause 1156802978 M * Bertl sounds like what you observed, no? 1156802979 M * daniel_hozac tanjix: you have a _directory_ named S_HOSTNAME="ubuntu"? 1156802983 M * derjohn how do i get the HZ from a running kernel 0 1156802984 M * derjohn ? 1156802987 M * tanjix daniel_hozac: correct 1156802997 M * tanjix daniel_hozac: that matches the 1st line of the config file 1156802998 M * daniel_hozac tanjix: probably you used the converter wrong then. 1156803004 M * daniel_hozac tanjix: or the script is broken. 1156803018 M * Bertl derjohn: easiest way from the config in /proc/config.gz :) 1156803042 M * Bertl one reason we will switch to msecs for those values at some point 1156803050 M * matti Hi Bertl :) 1156803057 M * tanjix daniel_hozac: i removed everything now and will try to re-use the converter 1156803058 M * Bertl hey matti! everything fine? 1156803062 M * matti Which vmware is more like xen? esx? 1156803085 M * derjohn whoooh: 2.6.15-amd64-vs2.1.0.5.1, 1000 Hz in that case 1156803099 M * tanjix daniel_hozac: when is use toDir.sh i get some "errors" 1156803106 M * Bertl matti: sounds about right, but I don't use vmware so I don't know the details 1156803111 M * derjohn I think i'll upgrade herbert now ;) 1156803124 M * matti Bertl: Me too, that's why I asking. 1156803148 M * Bertl derjohn: so in this case only 0.457 seconds 1156803158 M * daniel_hozac i gotta get up early tomorrow so i'm gonna head to bed, good night all! 1156803169 M * Bertl daniel_hozac: have a good one! cya! 1156803179 M * tanjix daniel_hozac: good night and thanks so far 1156803203 M * derjohn daniel_hozac, bye ~ ~ 1156803222 M * ntrs Bertl, can you recommend some good numbers for the sched file for sched_hard? 1156803232 M * tanjix Bertl: maybe we can continue? ;) I was wondering on how to convert the old config files for vservers to the new style? 1156803250 M * ntrs tanjix, we did all of them manually. 1156803265 M * ntrs We did one directory structure and then just copied them and changed the values. 1156803305 M * derjohn Bertl, so lowering the minfill would shorten the pause, nor? 1156803319 M * tanjix ntrs: and is there a list or s.th. on how this structure must be? just creating a directory under /etc/vservers called by the vservers name ? 1156803330 M * derjohn Bertl, isnt 0.5 sec even pretty much, too if ones types? 1156803345 M * Bertl yes, I'd say so ... 1156803345 M * ntrs tanjix, yes and then the flower page 1156803364 M * Bertl derjohn: definitely not a very interactive setup ... 1156803381 M * derjohn Bertl, i switch to a 250 HZ kernel and try out with that -always loaded- guest 1156803575 M * tanjix ntrs: maybe it's late here - could you please give a hint, what else i need to get a vserver to work? i created the directory now called like the vserver 1156803646 M * ntrs tanjix, all I can tell you is look at the documentation at the flower page http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1156803667 M * tanjix ntrs: i am on that page but i dont really understand :( 1156803675 M * ntrs Bertl, can you recommend some good numbers for the sched file for sched_hard? 1156803695 M * ntrs Bertl, preferrably low latency but not able to step on each other's toes? 1156803750 M * Bertl depends on the number of guests, and the applications running 1156803822 M * derjohn Bertl, another thing: as a guest can cause load on the hosts with kernels therads, would it not make sense to add something of the host's load to teh virtualized guest load? or even add all the hosts load to the guests one? or hostload/#guests ? 1156803843 M * ntrs Bertl, about 30 guests out of which about 5-8 are heavy users the rest pretty much doing nothing. All 30 running apache, mysql, pop3, imap, php, and such. 1156803942 M * Bertl okay, how much cpu do you want to give a guest (max)? 1156803960 M * Bertl 100%? 1156803964 M * Bertl 80%? 1156804050 M * derjohn ntrs, ping ;) [/me waits for the results, too] 1156804069 M * derjohn 100% ? what's that for a "limit" ? 1156804079 M * ntrs Bertl, 50% 1156804087 M * Bertl let's assume 100% with a 50% 'guarantee for all guests (together) when the host is doing something, okay? 1156804104 M * ntrs actually 30% so that all 6-7 heavy users can chug along. 1156804109 M * ntrs Ok 1156804130 M * ntrs derjohn, I don't think I'll have the results. wrong tree 1156804132 M * Bertl good, so we want to have 50% total in case all 30 guests are running 1156804152 M * ntrs Bertl, no, i'd use up to 100% when all 30 are running. 1156804165 M * ntrs There is nothing else but those 30 guests 1156804172 M * derjohn ntrs, the host! 1156804176 M * Bertl okay, but we want 50% if the host is running 1156804186 M * ntrs Bertl, BTW I have 100 Hz in the kernel 1156804189 M * tanjix fakerunlevel: open("/var/run/utmp"): No such file or directory 1156804190 M * derjohn ntrs, the cron foo, xfs joural bla etc. 1156804199 M * ntrs derjohn, what about the host? 1156804200 M * tanjix i get that now when trying to start the vserver - but that file exists ? 1156804220 M * Bertl so we basically give a 'guarantee' for 50% total for the guests, assuming that 6 of them are running at once 1156804261 M * matti Eh. 1156804263 M * matti Well. 1156804274 M * Bertl means 50%/6 = 8.33% for each guest 1156804303 M * Bertl and we want the guests to 'share' the rest equally amongh them in case we are otherwise idle 1156804318 M * Bertl so we give 1/30 to each (idle time) 1156804332 M * Bertl now the min and max values 1156804361 M * Bertl as you said interactivity, we don't want longer gaps than 1/5th of a second 1156804383 M * Bertl with 100Hz, that gives 20 tokens min 1156804453 M * Bertl OTOH, we do not want one context to do bursts for longer than a 2 or 3 seconds 1156804482 M * Bertl which 'roughly' (I spare the details here) gives 200-250 tokens bucket max 1156804526 M * Bertl now the question is, how do we adjust the 1/12th and 1/30th 1156804555 M * tanjix no ideas @ all ? :) 1156804568 M * Bertl as we said, we want short delays, and we want to give contexts a chance to figure that 1156804586 M * Bertl tanjix: maybe the file/dir is missing? 1156804597 M * tanjix Bertl: No, that exists 1156804610 M * Bertl tanjix: inside the guest? 1156804628 M * tanjix Bertl: should that be *inside* i thought outside? ;) 1156804657 M * Bertl so we choose to make that 'chunks' of 10 tokens for each 1156804660 M * tanjix Bertl: anyway - the file exists in the guest too 1156804665 M * Bertl which gives us the final values: 1156804694 M * Hollow Bertl: you seem the best candidate to write the scheduler section for my manual :D 1156804705 M * Hollow (or improve it at least) 1156804731 M * Bertl min = 20, max = 200, R/I= 10/120, R'/I'= 1/30 1156804800 Q * nayco Quit: Bonne nuit ! 1156804913 M * Bertl waldi: still around? 1156805074 M * derjohn Bertl, the magic question is to set a "max pause time" and a "max brust time" to get the numbers. wasnt there a util around to calc values ( Hollow ? ) 1156805093 Q * ntrs Read error: Connection reset by peer 1156805096 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156805110 M * Hollow http://dev.croup.de/proj/gentoo-vps/browser/util-vserver/tools/vschedcalc 1156805126 M * Hollow but it does not support idle time settings yet 1156805298 M * ntrs Bertl, ok, so the numbers should look like this: 1156805302 M * ntrs 10,1 1156805305 M * ntrs 120,30 1156805318 M * ntrs 20 1156805320 M * ntrs 200 1156805356 M * Bertl if that is rate,rate' interval,interval' and min max, yes:) 1156805384 M * Bertl has daniel_hozac added support for idle time to the config yet? 1156805404 M * Hollow not that i know of 1156805405 M * Bertl s/yet/already/ 1156805429 M * Bertl ah, just 'looked' to me like a config file :) 1156805456 M * ntrs Bertl, I made it look like it. :) sorry. 1156805463 M * Bertl np 1156805463 M * Hollow Bertl: i'll be off in half an hour or so, until saturday as i said.. if you need anything (especially infra related) bug me now :) 1156805505 M * Bertl haven't got to testing anything, main question on my side is what is missing in 2.1.1 from your POV? 1156805508 M * derjohn Hollow, off ? holiday? 1156805526 Q * PowerKe charon.oftc.net cation.oftc.net 1156805526 Q * Zaki charon.oftc.net cation.oftc.net 1156805526 Q * virtuoso charon.oftc.net cation.oftc.net 1156805530 M * Bertl derjohn: vacation I guess :) 1156805547 J * PowerKe ~icuser@d5153A0DA.access.telenet.be 1156805547 J * Zaki ~Zaki@212.118.126.43 1156805547 J * virtuoso ~s0t0na@shisha.spb.ru 1156805551 M * Hollow Bertl: retrieve information in cvirt (nr_*, load*, uptime) using vx_info or sth like that.. the rlimit accounting mask.. 1156805558 M * derjohn Hollow, what about the "announcement" ? delayed to next week ? 1156805563 Q * ntrs Read error: Connection reset by peer 1156805563 M * Hollow derjohn, Bertl: actually, i'm looking for a flat ;) 1156805570 M * Hollow derjohn: yep 1156805571 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156805574 M * derjohn Hollow, where ? 1156805577 M * Hollow Berlin 1156805596 M * Bertl Hollow: okay, plz put that into a mail and send it to me ... so I can look it up easily :) 1156805598 M * derjohn Hollow, got the university place ? 1156805608 M * Hollow yep.. Freie Universität Berlin 1156805618 M * derjohn Hollow, congrats !!! 1156805624 M * Hollow thanks :) 1156805627 M * Bertl yeah, from me too! 1156805651 M * Hollow i guess it will be a lot of fun, i'm moving with a friend.. 1156805660 M * derjohn Hollow, thats means 10 semesters full time linux-vserer coding ;) 1156805674 M * Hollow that's the spirit :D 1156805722 M * derjohn Hollow, the announcements next week? Bertl: changes for the release or stable / devel till then ? 1156805740 M * Hollow derjohn: yep, we decided to delay it until sa/so/mo 1156805766 M * derjohn Hollow, thats gives me time for restruct the FAQ ;) 1156805782 M * Hollow Bertl: sent you a bunch of attachments 1156805788 M * Hollow derjohn: :) 1156806046 M * PowerKe hmm, bringing the host network interface that is also used for the guests down and back up seems to break networking inside the vserver guest 1156806066 M * PowerKe is there any simple way to restore networking inside the guest without restarting it? 1156806106 M * PowerKe (not that it's an actual problem to restart it in this case, just wondering) 1156806367 P * stefani I'm Parting (the water) 1156806831 M * Bertl PowerKe: sure, you can add back the ips with 'ip addr add' 1156806861 M * Bertl PowerKe: but there is no 'restore ips for devices' feature in the tools (yet) 1156806934 M * Hollow ok folks.. i'm off now 1156806942 M * Hollow cu on saturday 1156806975 M * PowerKe Bertl: Thanks, that does the trick 1156807135 M * Bertl np 1156807213 M * PowerKe I'm trying to get vhashify to work with Gentoo guests 1156807248 M * PowerKe It's a php script that changes the mtimes of the installed package files and updates the portage database 1156807302 M * derjohn Hollow, still there ? 1156807330 M * derjohn Hollow, ah. i see, youre over and out. k. 1156807391 J * cehteh ~ct@cehteh.homeunix.org 1156807658 M * cehteh mhm .. i have 2 rtl8139 cards in my server but they swap their interface names on some boots, how can i fix that by assigning pci-id's to them 1156807765 M * Bertl better identify them via mac 1156807992 M * cehteh why is that better? 1156808011 M * cehteh i seen some dectection my mac mentioned but dont know how it works 1156808493 M * Bertl most init scripts can do that out of the box 1156808530 M * derjohn cehteh, in fedora it well made, in debian you have to look for /usr/share/doc/ifupdown/examples/.. for a mac mapper script 1156808563 M * cehteh ok but whats the preference for a mac in this case? 1156808568 M * derjohn (you can even to hotplug stuff with those mapper scripts, or network detection) 1156808603 M * cehteh binding to a pci-id seems more reasonable for me in my case 1156808613 M * derjohn cehteh, you always know that you got the right card, no matter in which slot ypu plug it. no matter which time of day and phase og the moon is. 1156808633 M * cehteh (one nic is soldered on the mainboard and i want to use it for the dsl line) 1156808663 M * cehteh and if one would switch the cards then he has to connect the wires diffrently 1156808738 M * Bertl I 1156808756 M * Bertl I'm pretty sure you can do that with the hotplug stuff and udev 1156808781 M * Bertl but you have to figure that yourself (the pci-id variant) 1156808809 M * cehteh well i can do it by mac too .. but is not well documented 1156808813 M * cehteh lemme try 1156808820 M * derjohn cehteh, the script i mentioned is bash and awks from 'ip addr'. you can replace that by "lspci -vv ...." 1156808857 M * derjohn cehteh, the docs are a mess , i fiddled about 20 mind around, but only because i am silly ;) 1156808875 M * cehteh derjohn: it tried get-mac-address.sh eth0 and it just hangs :P 1156808916 M * Bertl even ancient mandrake supports having MACADDR=... in the ifcfg-* files 1156808917 M * derjohn put it to the mapper part in the interfaces. it realyl works :) 1156808946 M * Bertl I doubt your distro doesn't have something similar 1156808969 M * derjohn Bertl, yup. fedora does that too. I wonder why debian needs an external scripts for that stuff, but ... "ueber opensource darf man nicht weinen, sondern patchen" 1156808990 J * mire_ ~mire@212-167-222-85.COOL.ADSL.VLine.Verat.NET 1156809021 M * cehteh derjohn: do you have a example for such a mapping entry and can you paste it to me? 1156809023 M * derjohn Bertl, IMVHO Debian does not yet in that "easy" manner. we even got the vlan on a bond very recently. (ifup hooks) 1156809036 M * derjohn cehteh, the notebook's at home .... 1156809042 M * cehteh heh ok 1156809046 A * cehteh reads the manpage 1156809062 M * derjohn cehteh, but in some mins, I just leave to go home 1156809079 M * cehteh i am prolly not finished until then :P 1156809080 M * derjohn cehteh, did lile I did and will hate the example they let out ;) 1156809088 M * cehteh hehe 1156809097 M * derjohn *left 1156809226 M * cehteh Bertl: do you know space/time overhead of using bridge devices? i am now setup my vserver to use aliases on bridges actually need one for vpn and just tried a 2nd one (anonymous bridge with no devices connected) for a dmz network between vservers 1156809267 M * cehteh looks like it works but i didnt tested it fully yet .. i tihnk its a nice alternative for a virtual networking abstraction 1156809446 M * Bertl well, I don't see why you actually want bridges for that, but it will not be different from mainline values 1156809483 M * Bertl i.e. I doubt that your guests actually use the bridge :)