1144195251 M * brc hehehhe it is already 21:00 here 1144195264 M * Bertl well, it's 2am here :) 1144195274 M * brc if you mail it to me i will have the test enviroment done as soon as i get the mail 1144195282 M * brc a vmware enviroment is fine ? 1144195295 M * Bertl if it is for you, then yes 1144195305 M * brc for me it is ok 1144195326 M * brc i have to wake up at 7:00 AM 1144195327 M * brc :( 1144195351 M * Bertl well, start with setting up the test environment and preparing a few tests tomorrow 1144195376 M * Bertl all tests should be doable on 'normal kernels with quota too 1144195382 M * Bertl just to get a reference 1144195404 M * Bertl I will see, maybe we have something tomorrow to test 1144195437 M * Bertl otherwise it will probably have to wait till after the weekend, as I'm travelling to germany over the weekend 1144195456 M * brc No problem. I will get two servers. one patched and one ok 1144195467 M * brc so i can test both and compare results 1144195476 M * brc Aren't you on germany ?? 1144195487 M * Bertl nope, I'm in Austria :) 1144195498 M * brc i though you were from germany 1144195513 M * Bertl you thought wrong ... 1144195583 M * brc i am leaving right now. if you want to tell me something just /msg, i am always idling here but i often don't read the chaennl. i read the erceived messages 1144195599 M * Bertl okay, good 1144195599 M * brc I will tell you as soon as i have the test enviroment done 1144195615 M * Bertl looking forward to it ... 1144195796 M * brc cya tomorrow! 1144195800 M * brc gnight 1144196515 Q * FireEgl Quit: Bye... 1144197107 Q * michal` Ping timeout: 480 seconds 1144197450 J * michal` ~michal@www.rsbac.org 1144197625 M * Bertl wb michal`! 1144197758 Q * doener Quit: Good night! 1144199047 M * ray6 Bertl: I still don't know whether I'll be traveling in germany this weekend :) 1144199204 M * Bertl well, we'll see on the weekend then :) 1144199238 M * ray6 Bertl: but I got mail from cngw about not really slots being available, did you get an answer from Jule? shuld I forward the mail? 1144199260 M * Bertl no and yes please 1144199419 M * ray6 Bertl: the schedule online available doesn't look as full as the mail implies though. And it seems as one could always do something after 21:00. And there's a "compiler VM ans script basteln" on day4 1144199461 M * ray6 (mail is forwarded) 1144199488 J * gerrit gerrit@dhcp64-134-126-220.sjcc.sjc.wayport.net 1144199495 M * Bertl welcome gerrit! 1144199525 M * gerrit Hi Bertl! 1144199538 M * ray6 Bertl: on day 3 there's a FreeBSD jail introduction, possibly one could attach after that %-) 1144199573 M * Bertl ray6: did you make some suggestions yet? 1144199603 M * ray6 Bertl: where/for what? 1144199611 M * ray6 Bertl: to Teemu? 1144199672 M * Bertl well, in general :) 1144199683 M * ray6 Bertl: just the mail I forwarded you 1144199710 M * ray6 Bertl: isn't that general enough? :) 1144199991 M * Bertl just asking ... 1144200159 M * micah Bertl: in case anyone is looking for them, I'm working on backports to sarge of libeecrypt, dietlibc and util-vserver from the unstable versions 1144200192 M * Bertl ah, good! 1144200194 M * micah Bertl: beecrypt and dietlibc are finished, and util-vserver will be soon as well, so once they are done, people running sarge/stable can install those 1144200219 M * micah (from backports.org) 1144200229 M * Bertl excellent work! 1144200243 Q * gerrit Ping timeout: 480 seconds 1144200246 M * micah I expect them to be available in the next two days, hopefully less 1144200249 M * Bertl I assume the barrier issues have been resolved 1144200267 M * Bertl and did we get any feedback/testing/idea regarding sparc64? 1144200274 M * micah yes, those were fixed earlier this morning 1144200322 M * micah nebuchadnezzar was the sparc64 person, right? 1144200335 M * Bertl yup, I think so 1144200355 M * micah that person put this up: http://linux-vserver.org/sparc64 1144200359 M * micah and that was the last I saw of them 1144200424 M * Bertl nebuchadnezzar: ping? 1144200519 M * micah that sparc64 page has latest info of last year 1144200551 M * Bertl well,y eah, he was just referring to it ... 1144200583 M * Bertl I know that it worked quite fine a few months ago, amybe a year 1144200598 M * Bertl so it should not be a problem on sparc64 1144200618 M * Bertl what made me wonder was the wait4() issues we saw 1144200638 M * Bertl which would point into a general kernel issues 1144200640 M * Bertl -s 1144200893 J * gerrit gerrit@dhcp64-134-126-250.sjcc.sjc.wayport.net 1144201319 N * ourseven_ ourseven 1144201342 M * Bertl wb ourseven! 1144201434 Q * gerrit Quit: Client exiting 1144203852 Q * cryo helium.oftc.net jupiter.oftc.net 1144203852 Q * mids helium.oftc.net jupiter.oftc.net 1144203852 Q * click helium.oftc.net jupiter.oftc.net 1144203852 Q * sladen helium.oftc.net jupiter.oftc.net 1144203852 Q * bubulak helium.oftc.net jupiter.oftc.net 1144203852 Q * nebuchadnezzar helium.oftc.net jupiter.oftc.net 1144203852 Q * Vudumen helium.oftc.net jupiter.oftc.net 1144203852 Q * alexx helium.oftc.net jupiter.oftc.net 1144203852 Q * ddlp helium.oftc.net jupiter.oftc.net 1144203852 Q * dsoul helium.oftc.net jupiter.oftc.net 1144203852 Q * mugwump helium.oftc.net jupiter.oftc.net 1144203852 Q * Cru helium.oftc.net jupiter.oftc.net 1144203852 Q * nokoya helium.oftc.net jupiter.oftc.net 1144203852 Q * Greek0 helium.oftc.net jupiter.oftc.net 1144203852 Q * michal` helium.oftc.net jupiter.oftc.net 1144203852 Q * wenchien helium.oftc.net jupiter.oftc.net 1144203852 Q * pollux helium.oftc.net jupiter.oftc.net 1144203852 Q * ourseven helium.oftc.net jupiter.oftc.net 1144203852 Q * mkhl helium.oftc.net jupiter.oftc.net 1144203852 Q * mire helium.oftc.net jupiter.oftc.net 1144203852 Q * insomniac helium.oftc.net jupiter.oftc.net 1144203852 Q * phreak`` helium.oftc.net jupiter.oftc.net 1144203852 Q * Skram helium.oftc.net jupiter.oftc.net 1144203852 Q * alamar helium.oftc.net jupiter.oftc.net 1144203852 Q * matti helium.oftc.net jupiter.oftc.net 1144203852 Q * Wonka helium.oftc.net jupiter.oftc.net 1144203852 Q * phedny helium.oftc.net jupiter.oftc.net 1144203852 Q * _Gonzo_ helium.oftc.net jupiter.oftc.net 1144203852 Q * pusling helium.oftc.net jupiter.oftc.net 1144203852 Q * mef helium.oftc.net jupiter.oftc.net 1144203852 Q * nox helium.oftc.net jupiter.oftc.net 1144203852 Q * SiD3WiNDR helium.oftc.net jupiter.oftc.net 1144203852 Q * rmoriz helium.oftc.net jupiter.oftc.net 1144203852 Q * sannes helium.oftc.net jupiter.oftc.net 1144203852 Q * Adrinael helium.oftc.net jupiter.oftc.net 1144203852 Q * cohan helium.oftc.net jupiter.oftc.net 1144203852 Q * wibble helium.oftc.net jupiter.oftc.net 1144203852 Q * Wenix helium.oftc.net jupiter.oftc.net 1144203852 Q * meebey helium.oftc.net jupiter.oftc.net 1144203852 Q * samuel_ helium.oftc.net jupiter.oftc.net 1144203852 Q * kilian helium.oftc.net jupiter.oftc.net 1144203914 J * michal` ~michal@www.rsbac.org 1144203914 J * wenchien ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1144203914 J * cryo ~say@psoft.user.matrix.farlep.net 1144203914 J * pollux foobar@image4.cpe.fr 1144203914 J * ourseven ~ourseven@tor-irc.dnsbl.oftc.net 1144203914 J * mkhl ~mkhl@200-148-40-113.dsl.telesp.net.br 1144203914 J * mids mids@mids.student.utwente.nl 1144203914 J * mire ~mire@48-166-222-85.COOL.ADSL.VLine.Verat.NET 1144203914 J * nebuchadnezzar ~nebu@zion.asgardr.info 1144203914 J * click click@ti511110a080-4050.bb.online.no 1144203914 J * insomniac ~insomniac@slackware.it 1144203914 J * phreak`` ~phreak``@134.68.220.30 1144203914 J * sladen paul@starsky.19inch.net 1144203914 J * Skram ~mark@admins.sentiensystems.net 1144203914 J * phedny ~mark@volcano.p-bierman.nl 1144203914 J * alamar ~alamar@sol.diddens.de 1144203914 J * matti matti@linux.gentoo.pl 1144203914 J * Wonka produziert@chaos.in-kiel.de 1144203914 J * bubulak ~bubulak@cicka.wnet.sk 1144203915 J * Vudumen ~vudumen@perverz.hu 1144203915 J * alexx ~alexx@proxy.ikse.net 1144203915 J * ddlp ~ddlp@sarayi.kariva.org 1144203915 J * dsoul darksoul@vice.ii.uj.edu.pl 1144203915 J * _Gonzo_ gonzo@langweiligneutral.deswahnsinns.de 1144203915 J * mugwump ~samv@watts.utsl.gen.nz 1144203915 J * Cru ~mindwarp@turbodiesel.e.de.wahlich.com 1144203915 J * nokoya young@hi-230-82.tm.net.org.my 1144203915 J * mef ~mef@targe.CS.Princeton.EDU 1144203915 J * Greek0 ~greek0@85.255.145.201 1144203915 J * nox ~nox@nox.user.oftc.net 1144203915 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1144203915 J * rmoriz ~roland@moriz.name 1144203915 J * sannes ~ace@simula-084.simula.no 1144203915 J * Adrinael adrinael@hoasb-ff09dd00-79.dhcp.inet.fi 1144203915 J * cohan ~cohan@koniczek.de 1144203915 J * wibble wibble@vortex.ukshells.co.uk 1144203915 J * Wenix ~wenix@81.7.189.11 1144203915 J * meebey meebey@booster.qnetp.net 1144203915 J * samuel_ ~samuel@levinux.UQAR.UQUEBEC.CA 1144203915 J * kilian kk@projects.verfaction.de 1144203915 J * pusling pusling@195.215.29.124 1144203925 J * kilian_ kk@projects.verfaction.de 1144204062 Q * mef Ping timeout: 480 seconds 1144204077 Q * kilian Ping timeout: 480 seconds 1144204077 N * kilian_ kilian 1144204146 J * mef ~mef@targe.CS.Princeton.EDU 1144204314 M * Bertl wb mef! 1144205101 J * djforradalmar ~djforrada@86.35.124.246 1144205104 M * djforradalmar hello guys 1144205111 M * Bertl welcome djforradalmar! 1144205139 M * djforradalmar can samba working inside in vserver host? 1144205145 M * djforradalmar in vserver guest... 1144205157 M * djforradalmar i have some problem with netbios resolve... 1144205174 M * djforradalmar (samba working on browsing only with ip address) 1144205186 M * Bertl well, yes it can work, to some extend 1144205210 M * Bertl linux-vserver is ip based, samba is 40% ip based 1144205226 M * Bertl i.e. you can do all ip related stuff out of the box 1144205250 M * djforradalmar so, samba is can't work property in vserver? 1144205267 M * Bertl it can, but not in a secure and isolated way 1144205270 M * daniel_hozac i'm using samba in several vservers, works fine. 1144205294 M * daniel_hozac i'm using DNS for names though, rather than NetBIOS. 1144205300 M * Bertl daniel_hozac: but not with ipx based stuff, right? 1144205326 M * daniel_hozac does anything use IPX these days? 1144205333 J * FireEgl Atlantica@Atlantica.DollarDNS.Net 1144205340 M * daniel_hozac except semi-old games on Windows. 1144205367 M * djforradalmar this is on only-xp network 1144205377 M * Bertl daniel_hozac: I guess most windows environments still use that, no? (maybe I'm wrong, have no windows :) 1144205456 M * daniel_hozac IIRC XP doesn't even have IPX installed by default. 1144205486 M * Bertl djforradalmar: well, as long as it is ip based, it will work 1144205595 M * djforradalmar ok thx :) 1144205646 Q * djforradalmar Read error: Connection reset by peer 1144207262 M * Bertl okay, the schedmod-feat02 should be usable by now ... 1144207295 M * Bertl anyway, I'm off to bed now ... so have a good one everyone! 1144207303 N * Bertl Bertl_zZ 1144207373 Q * softi42 Ping timeout: 480 seconds 1144208006 J * softi42 ~softi@p549D4FDF.dip.t-dialin.net 1144213135 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1144213544 J * Tj- Clean@61.247.246.75 1144213571 M * Tj- helo 1144213576 M * Tj- ny1 there 1144213636 P * Tj- 1144214319 M * Skram eee, just got a debian dedi, may but vpses on it 1144214321 M * Skram i am not allowed to change the os :( 1144215022 M * Hollow Skram: why? 1144215035 M * Skram its from a company i also do work with 1144215042 M * Skram it is their server, they dont like gentoo, bleh 1144215049 M * Skram i agree, pfft. 1144215062 M * Hollow well, you can use vserver with debian too ;) 1144215069 M * Skram yeah 1144215079 M * Skram i just need to make sure we have a tech there for a reboot 1144215088 M * Skram lemme find the stats on this box 1144215660 Q * Hollow Quit: Konversation terminated! 1144215802 J * Hollow ~hollow@home.xnull.de 1144215995 M * Skram okay 1144216003 Q * Hollow Remote host closed the connection 1144216030 M * nebuchadnezzar pong 1144216067 M * nebuchadnezzar ok, time to work 1144216130 J * Hollow ~hollow@home.xnull.de 1144216263 M * eyck eeeh 1144216304 M * eyck IPX is not IP based ;) 1144218171 J * kir_home ~kir@64.119.159.226 1144219789 J * ||Cobra|| ~cob@146.50.22.204 1144219875 J * hue ~hue@218.20.51.109 1144221347 J * Smutje_ ~Smutje@xdsl-87-78-99-161.netcologne.de 1144221478 Q * Smutje Ping timeout: 480 seconds 1144221478 N * Smutje_ Smutje 1144221928 Q * shedi Read error: Connection reset by peer 1144222963 J * shedi ~siggi@inferno.lhi.is 1144223338 Q * shedi Read error: Connection reset by peer 1144224188 J * shedi ~siggi@inferno.lhi.is 1144224642 Q * michal` Remote host closed the connection 1144225676 Q * shedi Quit: Leaving 1144225696 J * pagano ~pagano@lappagano.cnaf.infn.it 1144230063 Q * SNy Read error: Connection reset by peer 1144230810 J * lilalinux ~plasma@dslb-084-058-202-043.pools.arcor-ip.net 1144233899 J * cehteh foobar@cehteh.homeunix.org 1144234015 J * dimebag Bogner@192.116.240.34 1144234031 M * dimebag Hey Is Breti here? 1144234068 M * dimebag I have a question about memory corruption 1144234506 J * shedi ~siggi@tolvudeild-206.lhi.is 1144235893 M * SiD3WiNDR for you 1144235896 M * SiD3WiNDR woops, wrong win. 1144236944 M * pollux seems there is another problem with the debian package for util-vserver: 1144236945 M * pollux Setting up util-vserver (0.30.210-3) ... 1144236946 M * pollux lsattr: Inappropriate ioctl for device While reading flags on /var/lib/vservers/ 1144237175 J * SNy 99f3e61d7f@bmx-chemnitz.de 1144237290 M * dimebag loool i need Bertl_zZ Plz 1144237800 M * dimebag Bertl_zZ When you get up lord pong me loool 1144238756 Q * cryo Ping timeout: 480 seconds 1144238969 J * cryo ~say@psoft.user.matrix.farlep.net 1144239296 J * Loki|muh loki@satanix.de 1144239315 Q * Loki|muh_ Remote host closed the connection 1144239351 J * id23 ~id@p54A01C38.dip0.t-ipconnect.de 1144239359 M * id23 hi #vserver 1144239400 M * daniel_hozac hello 1144239732 Q * sannes Ping timeout: 480 seconds 1144240179 J * oliwel ~oliwel@ldvpc07.ldv.e-technik.tu-muenchen.de 1144240275 Q * oliwel Quit: 1144240711 M * id23 cya laters 1144240713 Q * id23 Quit: Leaving 1144242211 Q * cryo Ping timeout: 480 seconds 1144242486 Q * kir_home Quit: Ухожу я от вас 1144242641 J * cryo ~say@psoft.user.matrix.farlep.net 1144242743 J * doener ~doener@i5387FDA2.versanet.de 1144244465 J * kir_home ~kir@solution4.masslink.com 1144245377 Q * mids Quit: Lost terminal 1144245563 J * KraL ~aSi@81.215.138.118 1144245575 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144245711 Q * matta Read error: Connection reset by peer 1144246141 Q * KraL Quit: y akao 1144246149 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144246393 N * Bertl_zZ Bertl 1144246399 M * Bertl morning folks! 1144246406 M * Bertl dimebag: pong! 1144246605 J * sannes ~ace@simula-084.simula.no 1144246697 Q * cryo Ping timeout: 480 seconds 1144246824 Q * kir_home Quit: Ухожу я от вас 1144246881 Q * pagano Quit: Leaving 1144247055 J * cryo ~say@psoft.user.matrix.farlep.net 1144247135 M * softi42 morning Bertl 1144247319 M * Bertl morning softi42! 1144247397 J * Viper0482 ~Viper0482@p54976F77.dip.t-dialin.net 1144247742 M * brc Bertl: the test enviroment will be ready this afternoon 1144247782 M * Bertl excellent, I have a test patch available soon, will probably make your system explode :) 1144247821 M * Bertl brc: you should figure a way to capture and paste kernel traces somewhere, btw 1144247897 M * brc hehe cool 1144247912 M * brc how ? in dmesg ? 1144248292 M * Hollow hey Bertl! i need some help documenting flags, didn't find those: http://phpfi.com/111043 1144248433 M * harry Bertl!!!!!!!!! 1144248443 M * harry (yes, i'm here, yes, with a question :)) 1144248468 M * harry how do i backup files (in the host) that are mounted in a vserver 1144248495 M * harry gandalf:/vservers/ftpserv/home# ls -l 1144248496 M * harry gandalf:/vservers/ftpserv/home# ls -total 0 1144248496 M * harry gandalf:/vservers/ftpserv/home# vserver ftpserv enter 1144248496 M * harry ftpserv:/# ls /home/ 1144248496 M * harry achg borisdh christam dc_docs ekonomik ftp giessen harry jp lpq med-atl pelagant plotter sgielen stevenj testtest u0041345 ue005687 1144248499 M * harry angihypo carmel ckpdf demol euromarg geofoto gpsbasis inpac kulpabx manscape mez peter potgieter spm11 tboxuser tom u0041496 vasomod 1144248502 M * harry bioman chicken cvut dipu fswftp georeac grotenhout janelsen labeco marijke p5240000 pfeest sadl spoedts tech u0039351 ue005393 1144248511 M * harry sry for the spam :S 1144248591 M * Hollow harry: cp, rsync, dd, tar .. you have plenty of options :) 1144248602 M * harry Hollow: we use tsm for backup 1144248622 M * Hollow but you need to enter the namespace 1144248622 M * harry so basicly, i need to make these files available in the host context 1144248632 M * Bertl harry: what is the problem with backing them up? 1144248650 M * harry Bertl: as you can see, gandalf is the host 1144248654 M * harry which has a backup client 1144248657 M * Bertl harry: you could use vcontext to enter the guest's namespace 1144248660 M * harry i can backup filesystems there 1144248682 M * harry but not the filesystems INSIDE vservers it seems 1144248697 M * Bertl s/vcontext/vnamespace( 1144248699 M * harry Bertl: and then? 1144248706 M * harry how would that solve my problem? 1144248708 M * Bertl *argh* not fully awake yet :) 1144248723 M * Hollow harry: it would give you all the mounts of your guest, so you can access them 1144248736 M * harry hmm... how? 1144248739 M * Bertl harry: well, all filesystems the guest sees will be seen in vnamespace -e 1144248789 M * Bertl try vnamespace -e -- cat /proc/self/mounts 1144248839 M * harry how do i know the xid's? these are not ctx'es 1144248926 M * Hollow harry: if you don't use dynamic contexts, the id is in /etc/vservers//context 1144248948 M * Hollow (or you can simply use vserver-state) 1144248948 M * harry aha 1144248949 M * Bertl harry: well, if there is no context running, there is no guest, and therefore no guest filesystem 1144248961 M * harry 42 13 57.7M 16.2M 1m46s68 4m07s90 7d00h29 ftpserv 1144248962 M * Hollow s/state/stat/ 1144248966 M * Bertl harry: if the guest _is_ running, then vserver-stat will show the xid 1144248983 M * harry so 42 ??? 1144248999 M * harry it is 42 :) 1144249040 M * harry so ok... i can see them now... how does that help me backup my stuff? 1144249053 M * Bertl Hollow: but you had a look at http://linux-vserver.org/Caps+and+Flags 1144249084 M * Bertl harry: guess you can backup what you see, no? 1144249127 M * harry yes 1144249163 M * harry but i have to run a command to get to the information 1144249181 M * harry tsm is just a client in which you say e.g.: backup /vservers 1144249187 M * harry and it backs up everything there 1144249207 M * Hollow Bertl: no, anyway: what are binary mounts? and are these two INFO_* legacy or not? they still appear int the source, and you paper says legacy, but this page says the meaning is different.. 1144249263 M * Bertl binary mounts are mounts which take binary data as last argument 1144249275 M * Bertl this includes _most_ of the network mounts 1144249307 M * Bertl the flags _are_ legacy flags, but some of them are reused nowadays 1144249323 M * Bertl like the INIT and HIDE 1144249340 M * Hollow yeah, i grepped the kernel source and only these two seem to be left 1144249346 M * Bertl probably the PRIVATE and LOCK will get reused in the future 1144249353 M * Hollow mhm 1144249418 M * Bertl okay, off for dinner now .. back later ... 1144249425 N * Bertl Bertl_oO 1144249495 M * harry so, there is no way to just make 1144249509 M * harry the fs visible inside the host? 1144249942 M * jake- na toll. ich wrde das aber eher ber einen vpn tunnel machen. da kannst das ganze subnet tunneln 1144249958 M * jake- ups sorry. tabbed to the wrong window 1144250099 N * Bertl_oO Bertl 1144250102 M * Bertl back now 1144250688 M * harry hmm.... so no sollution? 1144250793 M * Bertl harry: sure, just mount it _on the host_ 1144251004 M * harry is that still possible now? 1144251035 M * Bertl sure, usually you do it the other way round 1144251046 M * Bertl i.e. mount it on the host and --bind mount it into the guest 1144251455 N * weasel w 1144251510 N * w weasel 1144251621 M * nebuchadnezzar hi 1144251682 M * Bertl hey nebuchadnezzar! 1144251733 M * nebuchadnezzar I see your ping :-) 1144251759 M * Bertl good :) 1144251762 M * bragon lo nebuchadnezzar & Bertl 1144251798 M * Bertl nebuchadnezzar: so I assume the sparc issues didn't vanish, yes? 1144251824 M * nebuchadnezzar no 1144251838 M * nebuchadnezzar is there is some update to utils/kernel patch ? 1144251862 M * Bertl but you only get them with the utils, right, not with other apps 1144251866 M * nebuchadnezzar I don't understand why the wait4 maybe a kernel issue 1144251874 M * nebuchadnezzar Bertl: right 1144251902 M * nebuchadnezzar maybe a little test case to give a try 1144251914 M * nebuchadnezzar ho I'll try something 1144251927 M * nebuchadnezzar I use the optimized libc6-sparcv9b 1144252029 Q * mountie Remote host closed the connection 1144252044 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1144252120 M * wenchien micah: ping 1144252196 M * nebuchadnezzar Bertl: the libc6 version is not the issue 1144252196 M * micah wenchien: yes? 1144252242 M * wenchien micah: have you tried util-vserver 0.30.210-3 ? 1144252279 M * micah wenchien: I've even tried 0.30.210-4 1144252295 M * micah wenchien: why do you ask? 1144252332 M * wenchien micah: I can't install 0.30.210-3.... I'd like to know if you have the same problem.. 1144252355 M * micah wenchien: you should try installing -4 1144252382 M * eyck Bertl: am I not being clear, there is some problem with my language skills or sth? Or are you actively trying not to understand that cutting one's off his remote machine for no reason but 'well, the script work like this, that they cut you off' is a bad thing? 1144252385 M * Bertl nebuchadnezzar: could you try to build the tools from scratch with some ./configure options I give you? 1144252398 M * nebuchadnezzar ok ok 1144252405 M * eyck I don't understand how this is even a matter for discussion, 1144252411 M * eyck I must be missing something... 1144252415 M * eyck isn't it obvious? 1144252430 M * micah wenchien: you can get it from here, if your local mirror has not updated: http://incoming.debian.org/ 1144252444 M * Bertl eyck: 'ifconfig eth0 down' will do the same, but I do not know any distro which will disallow that or ask you twice? 1144252479 M * Bertl micah: please confirm that debian doesn't do that :) 1144252483 M * eyck Bertl: this is not similiar at all 1144252490 M * wenchien micah: downloading... 1144252503 M * Bertl eyck: so how is that different? 1144252533 M * micah Bertl: you are free to do 'ifconfig eth0 down' in debian if you want :) 1144252534 M * eyck well, ... how is it similar? 1144252546 M * Bertl eyck: I would understand if you would argue that a guest which tries to create an IP which already exists, refuses to start and write an error message 1144252569 M * eyck hmm 1144252572 M * Bertl eyck: that's something I would support, and you can file a bug report anytime 1144252599 M * eyck how can I file a bugreport if you dismiss it beforehand 1144252603 M * Bertl eyck: but there is absolutely no reason to _disable_ functionality 'jsut' because some folks get it wrong 1144252626 M * Bertl eyck: but maybe I'm still not understanding what you are bitching about 1144252639 M * wenchien micah: ya! -4 installed. 1144252644 M * micah wenchien: :) 1144252656 M * eyck oh my god... someone replaced Bertl 1144252993 M * Bertl eyck: so I haven't seen your reply yet with the 'sane eyck semantics' 1144253112 M * eyck suppose your guest is going down, and utils come to the part with 'remove the ip' .. it checks if some other guest is not using that IP.. (or maybe host created the ip) and then, IF user requested deletes the IP 1144253135 M * Bertl how would you 'know' that the host created it? 1144253153 M * eyck i see no reason why would you think that 'bind to IP' means 'delete the IP when you're done' 1144253164 M * Bertl that is _wrong_ 1144253165 M * eyck well, either utils created the ip, or the host 1144253174 M * Bertl bind to ip doesnt mean that you delete it 1144253187 M * Bertl but _create_ and ip means that it is cleaned up afterwards 1144253188 M * eyck if I have not created the IP, I'm not going to touch it 1144253196 Q * FireEgl helium.oftc.net oxygen.oftc.net 1144253196 Q * matta helium.oftc.net oxygen.oftc.net 1144253196 Q * alexx helium.oftc.net oxygen.oftc.net 1144253196 Q * Vudumen helium.oftc.net oxygen.oftc.net 1144253196 Q * click helium.oftc.net oxygen.oftc.net 1144253196 Q * ddlp helium.oftc.net oxygen.oftc.net 1144253196 Q * bubulak helium.oftc.net oxygen.oftc.net 1144253196 Q * dsoul helium.oftc.net oxygen.oftc.net 1144253196 Q * sladen helium.oftc.net oxygen.oftc.net 1144253196 Q * nebuchadnezzar helium.oftc.net oxygen.oftc.net 1144253196 Q * nokoya helium.oftc.net oxygen.oftc.net 1144253196 Q * Cru helium.oftc.net oxygen.oftc.net 1144253196 Q * mugwump helium.oftc.net oxygen.oftc.net 1144253196 Q * Greek0 helium.oftc.net oxygen.oftc.net 1144253196 Q * Viper0482 helium.oftc.net oxygen.oftc.net 1144253196 Q * doener helium.oftc.net oxygen.oftc.net 1144253196 Q * ||Cobra|| helium.oftc.net oxygen.oftc.net 1144253196 Q * kilian helium.oftc.net oxygen.oftc.net 1144253196 Q * samuel_ helium.oftc.net oxygen.oftc.net 1144253196 Q * meebey helium.oftc.net oxygen.oftc.net 1144253196 Q * Wenix helium.oftc.net oxygen.oftc.net 1144253196 Q * wibble helium.oftc.net oxygen.oftc.net 1144253196 Q * cohan helium.oftc.net oxygen.oftc.net 1144253196 Q * nox helium.oftc.net oxygen.oftc.net 1144253196 Q * _Gonzo_ helium.oftc.net oxygen.oftc.net 1144253196 Q * Wonka helium.oftc.net oxygen.oftc.net 1144253196 Q * Skram helium.oftc.net oxygen.oftc.net 1144253199 Q * phreak`` helium.oftc.net oxygen.oftc.net 1144253199 Q * insomniac helium.oftc.net oxygen.oftc.net 1144253199 Q * mire helium.oftc.net oxygen.oftc.net 1144253199 Q * alamar helium.oftc.net oxygen.oftc.net 1144253199 Q * pollux helium.oftc.net oxygen.oftc.net 1144253199 Q * matti helium.oftc.net oxygen.oftc.net 1144253199 Q * mkhl helium.oftc.net oxygen.oftc.net 1144253199 Q * ourseven helium.oftc.net oxygen.oftc.net 1144253199 Q * SiD3WiNDR helium.oftc.net oxygen.oftc.net 1144253199 Q * phedny helium.oftc.net oxygen.oftc.net 1144253199 Q * rmoriz helium.oftc.net oxygen.oftc.net 1144253199 Q * wenchien helium.oftc.net oxygen.oftc.net 1144253199 Q * Adrinael helium.oftc.net oxygen.oftc.net 1144253199 Q * jkl helium.oftc.net oxygen.oftc.net 1144253199 Q * teukka helium.oftc.net oxygen.oftc.net 1144253199 Q * dhansen helium.oftc.net oxygen.oftc.net 1144253199 M * Bertl precisely 1144253199 M * eyck what? 1144253213 M * Bertl that is precisely the behaviour now 1144253222 M * eyck suppose I've got machine with 1 IP, a bind two vservers to that IP... 1144253228 M * Bertl when the tools (util-vserver) creates the ip 1144253240 M * Bertl then the ip will be removed on shutdown. period. 1144253245 M * eyck no. 1144253250 J * doener_ ~doener@i5387FAF4.versanet.de 1144253250 J * Viper0482 ~Viper0482@p54976F77.dip.t-dialin.net 1144253250 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144253250 J * doener ~doener@i5387FDA2.versanet.de 1144253250 J * ||Cobra|| ~cob@146.50.22.204 1144253250 J * FireEgl Atlantica@Atlantica.DollarDNS.Net 1144253250 J * kilian kk@projects.verfaction.de 1144253250 J * wenchien ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1144253250 J * pollux foobar@image4.cpe.fr 1144253250 J * ourseven ~ourseven@tor-irc.dnsbl.oftc.net 1144253250 J * mkhl ~mkhl@200-148-40-113.dsl.telesp.net.br 1144253250 J * mire ~mire@48-166-222-85.COOL.ADSL.VLine.Verat.NET 1144253250 J * nebuchadnezzar ~nebu@zion.asgardr.info 1144253250 J * click click@ti511110a080-4050.bb.online.no 1144253250 J * insomniac ~insomniac@slackware.it 1144253250 J * phreak`` ~phreak``@134.68.220.30 1144253250 J * sladen paul@starsky.19inch.net 1144253250 J * Skram ~mark@admins.sentiensystems.net 1144253252 J * phedny ~mark@volcano.p-bierman.nl 1144253252 J * alamar ~alamar@sol.diddens.de 1144253252 J * matti matti@linux.gentoo.pl 1144253252 J * Wonka produziert@chaos.in-kiel.de 1144253252 J * bubulak ~bubulak@cicka.wnet.sk 1144253252 J * Vudumen ~vudumen@perverz.hu 1144253252 J * alexx ~alexx@proxy.ikse.net 1144253252 J * ddlp ~ddlp@sarayi.kariva.org 1144253252 J * dsoul darksoul@vice.ii.uj.edu.pl 1144253252 J * _Gonzo_ gonzo@langweiligneutral.deswahnsinns.de 1144253252 J * mugwump ~samv@watts.utsl.gen.nz 1144253252 J * Cru ~mindwarp@turbodiesel.e.de.wahlich.com 1144253252 J * nokoya young@hi-230-82.tm.net.org.my 1144253252 J * Greek0 ~greek0@85.255.145.201 1144253252 J * nox ~nox@nox.user.oftc.net 1144253252 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1144253252 J * rmoriz ~roland@moriz.name 1144253252 J * Adrinael adrinael@hoasb-ff09dd00-79.dhcp.inet.fi 1144253252 J * cohan ~cohan@koniczek.de 1144253252 J * wibble wibble@vortex.ukshells.co.uk 1144253252 J * Wenix ~wenix@81.7.189.11 1144253252 J * meebey meebey@booster.qnetp.net 1144253252 J * samuel_ ~samuel@levinux.UQAR.UQUEBEC.CA 1144253252 T * xenon.oftc.net 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}-rc15 | 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 ;) 1144253259 M * eyck please, please 1144253271 M * eyck tell my why 'bind to ip' means 'delete the IP when you're done' ? 1144253274 M * Bertl nebuchadnezzar: please try first with: --disable-alternative-syscalls 1144253288 M * Bertl eyck: as I said, it doesn't and that is not how it works 1144253303 M * eyck good. 1144253307 M * eyck only that it works like this. 1144253312 M * Bertl nope 1144253317 M * eyck UNLESS you put 'nodev' or some such 1144253338 M * Bertl if you do not put nodev there, but have a dev entry instead it _means_ create the IP for me 1144253354 M * eyck nope. 1144253357 M * Bertl yes 1144253362 M * eyck it just means that I have not put 'nodev' there 1144253370 M * eyck why would I put random stuff in configuration 1144253382 M * eyck still, 1144253388 M * Bertl why would you adhere to any configuration? 1144253399 M * eyck what? 1144253414 M * eyck listen, those semantics are insane. 1144253432 M * eyck or maybe I'm insane. 1144253453 M * Bertl well, what you basically say is: I did write a config for IPs to be created by the tools, but I didn't intend to, so can we change the config semantics? 1144253474 J * jkl eric@c-67-172-156-116.hsd1.co.comcast.net 1144253477 M * eyck are you trying to trick me? 1144253484 M * Bertl and my only question to that is: what do you consider 'sane' semantics? 1144253495 Q * FireEgl arion.oftc.net jupiter.oftc.net 1144253495 Q * alexx arion.oftc.net jupiter.oftc.net 1144253495 Q * Vudumen arion.oftc.net jupiter.oftc.net 1144253495 Q * click arion.oftc.net jupiter.oftc.net 1144253495 Q * ddlp arion.oftc.net jupiter.oftc.net 1144253495 Q * bubulak arion.oftc.net jupiter.oftc.net 1144253495 Q * matta arion.oftc.net jupiter.oftc.net 1144253495 Q * dsoul arion.oftc.net jupiter.oftc.net 1144253495 Q * sladen arion.oftc.net jupiter.oftc.net 1144253495 Q * nebuchadnezzar arion.oftc.net jupiter.oftc.net 1144253495 Q * nokoya arion.oftc.net jupiter.oftc.net 1144253495 Q * Cru arion.oftc.net jupiter.oftc.net 1144253495 Q * mugwump arion.oftc.net jupiter.oftc.net 1144253495 Q * Greek0 arion.oftc.net jupiter.oftc.net 1144253495 Q * samuel_ arion.oftc.net jupiter.oftc.net 1144253495 Q * meebey arion.oftc.net jupiter.oftc.net 1144253495 Q * Wenix arion.oftc.net jupiter.oftc.net 1144253495 Q * wibble arion.oftc.net jupiter.oftc.net 1144253495 Q * cohan arion.oftc.net jupiter.oftc.net 1144253497 Q * nox arion.oftc.net jupiter.oftc.net 1144253497 Q * _Gonzo_ arion.oftc.net jupiter.oftc.net 1144253497 Q * Wonka arion.oftc.net jupiter.oftc.net 1144253497 Q * Skram arion.oftc.net jupiter.oftc.net 1144253497 Q * phreak`` arion.oftc.net jupiter.oftc.net 1144253497 Q * insomniac arion.oftc.net jupiter.oftc.net 1144253497 Q * mire arion.oftc.net jupiter.oftc.net 1144253497 Q * doener arion.oftc.net jupiter.oftc.net 1144253497 Q * Viper0482 arion.oftc.net jupiter.oftc.net 1144253497 Q * doener_ arion.oftc.net jupiter.oftc.net 1144253497 Q * alamar arion.oftc.net jupiter.oftc.net 1144253497 Q * pollux arion.oftc.net jupiter.oftc.net 1144253497 Q * kilian arion.oftc.net jupiter.oftc.net 1144253497 Q * matti arion.oftc.net jupiter.oftc.net 1144253497 Q * mkhl arion.oftc.net jupiter.oftc.net 1144253497 Q * ourseven arion.oftc.net jupiter.oftc.net 1144253497 Q * SiD3WiNDR arion.oftc.net jupiter.oftc.net 1144253497 Q * ||Cobra|| arion.oftc.net jupiter.oftc.net 1144253497 Q * phedny arion.oftc.net jupiter.oftc.net 1144253497 Q * rmoriz arion.oftc.net jupiter.oftc.net 1144253497 Q * wenchien arion.oftc.net jupiter.oftc.net 1144253497 Q * Adrinael arion.oftc.net jupiter.oftc.net 1144253497 M * Bertl welcome jkl! 1144253510 M * Bertl seems like splitting day on oftc again :) 1144253533 Q * jkl helium.oftc.net oxygen.oftc.net 1144253569 M * eyck Bertl: I didn't write the config to create those IPs, I wrote a basic config with no frills, like you people told me to 1144253583 M * eyck when I tried to check new config style, 1144253622 M * eyck and then you tell me, that I did wrong, because I haven't put 'Please don't kill me system' line in configuration. 1144253628 M * eyck oh well, my fault, 1144253648 M * eyck everyone knows that unless you ask nicely, utills will kill you and your family... 1144253659 M * eyck ...well, I didn't knew that, until now. 1144253678 M * eyck what I'm asking for, is, for dangerous stuff to be explicitly requested 1144253700 M * eyck so, if I want the utils to delete the ips, there should be an option 'deleteips' 1144253731 M * Bertl eyck: there is no _explicit_ delete ips 1144253760 M * eyck so, what I did wrong? 1144253774 M * Bertl look, it's like memory ownership (without garbage collector), you allocate it, you have to free it 1144253798 M * eyck ok, but it was someone else's memory 1144253824 M * Bertl first, you probably started to create the config from scratch, which, after you didn't read the Flower age had different semantics than you intended 1144253828 M * eyck with garbage collector you've got usage counter 1144253829 J * bonbons ~bonbons@83.222.39.180 1144253838 M * Bertl wb bonbons! 1144253849 J * jkl eric@c-67-172-156-116.hsd1.co.comcast.net 1144253850 M * bonbons Hi Bertl! 1144253857 M * eyck and you know that someone is using that memory, 1144253864 M * Bertl eyck: then you did ignore the warning which gave a good hint that something is wrong (the NETLINk stuff) 1144253883 M * eyck what warning? 1144253894 M * Bertl something like address already exists 1144253904 M * eyck ha! 1144253912 M * Bertl this is the point where I would suggest to make it an error 1144253916 M * eyck exactly, the tools ignored the error 1144253936 M * eyck which meant - you're not the owner of that IP, the host is, 1144253940 M * Bertl well, it's not an error by default, but I would make it one 1144253959 M * Bertl the ip could also be left over from a previous incomplete startup 1144253996 M * Bertl but as I already said twice, I'd support you in a quest to make that a critical error 1144254027 M * eyck oh well, better then nothing, 1144254045 M * eyck but still, deleting the IP is pretty aggressive stance 1144254048 M * daniel_hozac eyck: you realize that if you do not put dev in the configuration, the sanity check will tell you that you should put nodev in it, right? 1144254069 M * eyck daniel_hozac: nope, you guys told me to put 'dev' there... 1144254084 M * Bertl ah, now we are there, it's our fault :) 1144254092 M * eyck well, it is. 1144254104 M * eyck my fault is in trusting too easily 1144254105 M * Bertl well, eyck, so be it, it's my fault, and only mine. period. 1144254133 J * mugwump ~samv@watts.utsl.gen.nz 1144254133 J * Cru ~mindwarp@turbodiesel.e.de.wahlich.com 1144254133 J * nokoya young@hi-230-82.tm.net.org.my 1144254133 J * Greek0 ~greek0@85.255.145.201 1144254133 M * eyck Bertl: why would you say that, if you claim that there was nothing wrong ? 1144254145 J * doener_ ~doener@i5387FAF4.versanet.de 1144254145 J * Viper0482 ~Viper0482@p54976F77.dip.t-dialin.net 1144254145 J * ||Cobra|| ~cob@146.50.22.204 1144254145 J * kilian kk@projects.verfaction.de 1144254145 J * wenchien ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1144254145 J * pollux foobar@image4.cpe.fr 1144254145 J * ourseven ~ourseven@tor-irc.dnsbl.oftc.net 1144254145 J * mkhl ~mkhl@200-148-40-113.dsl.telesp.net.br 1144254145 J * mire ~mire@48-166-222-85.COOL.ADSL.VLine.Verat.NET 1144254145 J * insomniac ~insomniac@slackware.it 1144254145 J * phreak`` ~phreak``@134.68.220.30 1144254145 J * Skram ~mark@admins.sentiensystems.net 1144254145 J * phedny ~mark@volcano.p-bierman.nl 1144254145 J * alamar ~alamar@sol.diddens.de 1144254145 J * matti matti@linux.gentoo.pl 1144254145 J * Wonka produziert@chaos.in-kiel.de 1144254145 J * _Gonzo_ gonzo@langweiligneutral.deswahnsinns.de 1144254145 J * nox ~nox@nox.user.oftc.net 1144254145 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1144254145 J * rmoriz ~roland@moriz.name 1144254145 J * samuel_ ~samuel@levinux.UQAR.UQUEBEC.CA 1144254145 J * meebey meebey@booster.qnetp.net 1144254145 J * Wenix ~wenix@81.7.189.11 1144254145 J * wibble wibble@vortex.ukshells.co.uk 1144254145 J * cohan ~cohan@koniczek.de 1144254145 J * Adrinael adrinael@hoasb-ff09dd00-79.dhcp.inet.fi 1144254161 M * Bertl eyck: to make you happy and to end a useless discussion. 1144254171 J * FireEgl Atlantica@Atlantica.DollarDNS.Net 1144254172 M * eyck it's not a matter of fault or guilt, it's a matter of fixing the utils. 1144254185 M * Bertl eyck: we do not fix what is not broken 1144254192 J * dhansen ~dave@sprucegoose.sr71.net 1144254192 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144254192 J * nebuchadnezzar ~nebu@zion.asgardr.info 1144254192 J * click click@ti511110a080-4050.bb.online.no 1144254192 J * sladen paul@starsky.19inch.net 1144254192 J * bubulak ~bubulak@cicka.wnet.sk 1144254192 J * dsoul darksoul@vice.ii.uj.edu.pl 1144254192 J * ddlp ~ddlp@sarayi.kariva.org 1144254192 J * alexx ~alexx@proxy.ikse.net 1144254192 J * Vudumen ~vudumen@perverz.hu 1144254199 M * eyck that's why I'm trying to have a discussion 1144254220 M * Bertl well, we are having one, but we are missing your _good arguments_ 1144254230 M * eyck and I'm missing yours... 1144254247 M * Bertl my argumentation is simple: 1144254268 M * eyck principle of least surprise.. 1144254271 M * Bertl IF you specify 'dev' then you say 'please create that IP for me on the device dev' 1144254283 M * eyck ok 1144254308 M * Bertl and further, ips created from the tools on startup, will get removed on shutdown 1144254324 M * eyck why would it try to delete the ip, if - as you said yourself, I haven't requested it's deletion... 1144254337 M * eyck why would they get removed on shutdown? 1144254354 M * Bertl because the tools are the 'owner' of those IPs 1144254366 M * Bertl like the allocated memory which should be freed (cleanup) 1144254393 M * Bertl otherwise all kind of issues would happen with dangling IPs 1144254395 M * eyck ok, and that's my problem... the tools are not the 'owner' of those IPs 1144254411 M * Bertl maybe, but you told them they are :) 1144254419 M * eyck well 1144254426 M * eyck I told them to 'bind' to those ips.. 1144254428 M * Bertl otherwise you would not have put the 'dev' there 1144254431 M * doener_ it's not broken, it's all about semantics... 'ip' is "bind the vserver to this ip address", 'dev' is "create and destroy the ip address for me, and do that on this device", 'nodev' is "the address in 'ip' is already there, I take care of it" 1144254433 Q * Vudumen arion.oftc.net neutron.oftc.net 1144254433 Q * alexx arion.oftc.net neutron.oftc.net 1144254433 Q * bubulak arion.oftc.net neutron.oftc.net 1144254433 Q * click arion.oftc.net neutron.oftc.net 1144254433 Q * matta arion.oftc.net neutron.oftc.net 1144254433 Q * dhansen arion.oftc.net neutron.oftc.net 1144254433 Q * ddlp arion.oftc.net neutron.oftc.net 1144254433 Q * dsoul arion.oftc.net neutron.oftc.net 1144254433 Q * sladen arion.oftc.net neutron.oftc.net 1144254433 Q * nebuchadnezzar arion.oftc.net neutron.oftc.net 1144254469 M * Bertl eyck: and more important, all this is described in detail on the flower page 1144254481 M * doener_ the bind part is the 'ip' file _only_, and the tools additionally need either 'dev' or 'nodev', but that's no longer the plain binding 1144254483 J * dhansen ~dave@sprucegoose.sr71.net 1144254483 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144254483 J * nebuchadnezzar ~nebu@zion.asgardr.info 1144254483 J * click click@ti511110a080-4050.bb.online.no 1144254483 J * sladen paul@starsky.19inch.net 1144254483 J * bubulak ~bubulak@cicka.wnet.sk 1144254483 J * dsoul darksoul@vice.ii.uj.edu.pl 1144254483 J * ddlp ~ddlp@sarayi.kariva.org 1144254483 J * alexx ~alexx@proxy.ikse.net 1144254483 J * Vudumen ~vudumen@perverz.hu 1144254538 M * eyck ok, ok, what I'm saying, is this is wrong, 1144254549 M * eyck you can't go doing bad things to people, 1144254555 M * Bertl personally I had eliminated the 'nodev' and made the 'without dev' default 1144254580 M * eyck and describe that you're doing it on 'flower page', and that's it. 1144254675 M * Bertl eyck: nobody does bad things to people, you are the only person who did bad things to yourself (by choosing an unfortunate config) 1144254676 M * eyck let's take it around - 'nodev' is the default ? 1144254690 M * daniel_hozac neither is the default. 1144254696 M * Bertl eyck: no, there is no default, you get a warning 1144254697 M * daniel_hozac one or the other must be set. 1144254779 M * eyck as far as I can see, the problem is that the tools perform their actions blindly 1144254808 M * Bertl well, maybe we should add some code from espdiff :) 1144254817 M * eyck what's espdiff? 1144254838 M * Bertl the diff tool we use for merging complicated patches 1144254871 M * Bertl (part of patchutils) 1144254890 M * eyck Please concentrate... 1144254890 M * eyck No brainwave activity detected. Aborting. 1144254893 M * eyck this? 1144254925 M * Bertl works here ... 1144254936 M * eyck maybe you've got some barinwave activity 1144254937 M * daniel_hozac lol 1144255046 M * Bertl eyck: so IMHO if you are interested in improving things, file that feature request 'make the vserver start fail when the ip already exists, in case the creation/deletion (dev) is specified' 1144255088 M * Bertl eyck: but if you just want to complain, then there is no real point in it 1144255142 M * nebuchadnezzar Bertl: ok, I compile 0.30.210-4 with --disable-alternative-syscalls and it do not work 1144255147 Q * shedi Ping timeout: 480 seconds 1144255156 M * nebuchadnezzar wait(): No child processes 1144255160 M * nebuchadnezzar snif 1144255236 M * eyck no point in filing feature requests/sending patches, when the authors don't grasp that hurting people is bad, 1144255367 Q * Viper0482 Ping timeout: 480 seconds 1144255399 M * Bertl eyck: okay, that's your decision ... 1144255476 M * eyck well, I won't be affected by this, I already know that unless you put 'nodev' utils will try to hurt you, only new users will get bitten, 1144255482 M * eyck and if you don't think that's a problem, 1144255490 M * eyck well... 1144255534 M * Bertl new users will use the tools to create guests, which do put the nodev there 1144255569 M * eyck oh, so we're only alienating converts, 1144255572 M * eyck that's good too 1144255597 M * eyck it's their fault, they should've waited before starting using vservers 1144255662 M * Bertl no, but they should read the docu, before they create random configs :) 1144255759 M * daniel_hozac Bertl: does sparc64's clone have the arguments in flags, stack order? 1144255785 Q * Vanzetti Quit: Leaving 1144255793 M * Bertl let me check the syscall args from shiny 1144255837 J * shedi ~siggi@tolvudeild-199.lhi.is 1144255853 M * Bertl all arguments in registers 1144255906 M * Bertl the only difference is the syscall command (between sparc32/64 1144255956 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1144255956 M * daniel_hozac Bertl: didn't waldi say something about s390 having sys_clone's arguments in the opposite order? 1144255974 Q * shedi Quit: 1144255976 M * Bertl hmm, okay, let me check the kernel code 1144256027 M * daniel_hozac although i guess sys_clone should return an error in that case, as SIGCHLD|CLONE_VFORK|VLONE_NEWNS shouldn't be a valid address for the child stack. 1144256107 Q * ||Cobra|| Remote host closed the connection 1144256109 J * Viper0482 ~Viper0482@p54975DE0.dip.t-dialin.net 1144256261 M * waldi daniel_hozac: yes, it returns an error 1144256365 M * Bertl basically we already ruled out the syscall implementation and dietlibc, so what is left? 1144256392 M * Bertl could somebody craft a clone/wait test tool while I translocate? 1144256394 M * nebuchadnezzar Bertl: my buil was with dietlibc 1144256411 M * Bertl nebuchadnezzar: yes, but the one yesterday was without, no? 1144256432 M * nebuchadnezzar right, but it may impact --disable-alternative-syscalls ? 1144256470 M * Bertl well, you can try without dietlibc and alternative-syscalls jsut to make sure 1144256477 M * nebuchadnezzar ok 1144256508 M * Bertl interesting detail, there is no sys_clone() for sparc as it seems 1144256560 M * daniel_hozac hmm, arch/sparc64/kernel/entry.S? 1144256564 M * daniel_hozac sys_clone: 1144256577 M * Bertl ah, yes, cscope misses .S files ... 1144256710 M * Bertl mov 0, %o3 1144256714 M * Bertl interesting ... 1144256754 M * Bertl yes, the arguments are definitely reversed on sparc 1144256772 M * Bertl sparc_do_fork(unsigned long clone_flags, unsigned long stack_start, struct pt_regs *regs, unsigned long stack_size) 1144256792 M * Bertl basically sys_clone calls that with stack_size = 0 1144256863 M * Bertl arg0,1: flags,usp 1144256879 M * Bertl (for sparc32, the case we have here) 1144256890 M * Hollow ARGH! 1144256896 M * Hollow it's snowing again! 1144256898 M * Hollow wtf 1144256975 M * Bertl okay, translocating now, back later ... 1144256986 N * Bertl Bertl_oO 1144257650 Q * cohan Remote host closed the connection 1144257687 J * trash trash@databerlin.org 1144257691 M * nebuchadnezzar Bertl_oO: I got an error src/vlogin.o: In function `do_vlogin':src/vlogin.c:254: undefined reference to `openpty' 1144257693 M * nebuchadnezzar strange 1144257717 M * daniel_hozac nebuchadnezzar: very. is that with glibc? 1144257724 M * nebuchadnezzar yes 1144257729 M * nebuchadnezzar and --disable-alternative-syscalls 1144257778 M * daniel_hozac Hollow: does openpty require a library in glibc? 1144257784 M * daniel_hozac nebuchadnezzar: http://daniel.hozac.com/vserver/wait-test.c 1144257794 M * nebuchadnezzar ok 1144257797 M * daniel_hozac nebuchadnezzar: gcc -o wt wait-test.c; ./wt 1144257802 M * daniel_hozac what output do you get? 1144257832 M * nebuchadnezzar need to be root ? 1144257834 M * nebuchadnezzar Parent: clone(): Operation not permitted 1144257834 M * nebuchadnezzar Parent: wait(): No child processes 1144257838 M * daniel_hozac yeah. 1144257854 M * nebuchadnezzar Parent: wait(): No child processes 1144257854 M * nebuchadnezzar Parent: pid 30524 exited with status 65280 1144257903 M * Hollow daniel_hozac: i don't think so, but if you link with diet you have to use -lutil IIRC 1144257918 M * daniel_hozac Hollow: hmm, seems to be the other way around. 1144257937 M * daniel_hozac glibc requires -lutil, at least. 1144257941 M * Hollow indeed. 1144257980 M * daniel_hozac nebuchadnezzar: how did you manage to get both those lines? they are in different clauses of the same if... 1144258029 M * nebuchadnezzar well, under root I got: 1144258029 M * nebuchadnezzar Parent: wait(): No child processes 1144258029 M * nebuchadnezzar Parent: pid 30561 exited with status 65280 1144258076 M * daniel_hozac nebuchadnezzar: could you paste the file somewhere? 1144258084 M * nebuchadnezzar which one ? 1144258088 M * daniel_hozac wait-test.c 1144258105 M * nebuchadnezzar I make a wget http://daniel.hozac.com/vserver/wait-test.c 1144258106 M * daniel_hozac because if you have the same file as i have, that's some serious gcc bug. 1144258113 M * nebuchadnezzar I made 1144258121 M * nebuchadnezzar gcc 4.03 1144258123 M * daniel_hozac right, i just want to rule out that i updated it and you got an old copy. 1144258144 M * daniel_hozac or better, what does line 36 say in your file? 1144258144 M * daniel_hozac else? 1144258190 M * nebuchadnezzar http://www.asgardr.info/~nebu/vserver/wait-test.c 1144258267 M * daniel_hozac yep, you've got the same file. the fact that both those lines get printed shows your gcc really isn't working as it should... 1144258284 M * nebuchadnezzar ok ok 1144258291 M * nebuchadnezzar I'll upgrade to 4.1 1144258400 M * daniel_hozac actually, wait a minute. 1144258406 M * daniel_hozac let me add another test. 1144258430 M * daniel_hozac http://daniel.hozac.com/vserver/wait-test.c 1144258440 M * daniel_hozac how many dots do you get? 1144258441 J * cohan ~cohan@koniczek.de 1144258473 M * nebuchadnezzar 2 1144258477 M * daniel_hozac ok. 1144258486 M * nebuchadnezzar I use gcc 4.1 1144258504 M * daniel_hozac reget it now, and paste the output. 1144258512 M * daniel_hozac it's not returning the expected values, it seems... 1144258544 M * nebuchadnezzar pid = 30797 1144258544 M * nebuchadnezzar Parent: wait(): No child processes 1144258544 M * nebuchadnezzar pid = 30798 1144258544 M * nebuchadnezzar Parent: pid 30798 exited with status 65280 1144258603 M * daniel_hozac hmm, it seems sys_clone returns the parent's pid in the child thread. 1144258626 M * daniel_hozac reget it, and paste. 1144258649 M * nebuchadnezzar Child: pid 30811 1144258649 M * nebuchadnezzar pid = 30811 1144258649 M * nebuchadnezzar Parent: pid 30811 exited with status 0 1144258658 M * daniel_hozac ok, thanks a lot. 1144258768 M * nebuchadnezzar since when we get the parent pid in the child ? 1144258776 M * nebuchadnezzar ho, it maybe specific to fork 1144258776 M * nebuchadnezzar ? 1144258854 M * daniel_hozac typically, clone() and fork() return 0 in the child. 1144258874 M * nebuchadnezzar well, POSIX require this, IIRC ? 1144258898 M * daniel_hozac i'm not sure about clone(), IIRC it's Linux specific. 1144258912 M * daniel_hozac but for fork(), yeah. 1144258922 M * nebuchadnezzar ok 1144258926 M * Hollow clone is evil. 1144258928 M * Hollow :) 1144258958 M * daniel_hozac only way to get a new namespace ;) 1144258991 M * Hollow yeah, fortunately you don't have to mess with the stack for namespaces ;) 1144259185 A * waldi adds optional namespace cleanup to util-vserver 1144259225 M * Hollow ah, the cleanup is another one of these eval syscalls :) 1144259230 M * Hollow *evil 1144259256 M * waldi no, i use the userspace version 1144259334 M * daniel_hozac won't that make vnamespace -e ... mount --bind /path/on/host /vservers/... impossible? 1144259355 M * Hollow daniel_hozac: do you know what these are for? 1144259356 M * Hollow #define VXC_CAP_MASK 0x00000000 /*!< ?? */ 1144259357 M * Hollow #define VLIMIT_SEMARY 20 /*!< ?? */ 1144259357 M * Hollow #define NXA_MOD_BCAST (1<<8) /*!< ?? */ 1144259357 M * Hollow #define NXA_TYPE_ANY (~0) /*!< ?? */ 1144259376 M * daniel_hozac NXA_MOD_BCAST is to set the broadcast address of the network context. 1144259387 M * daniel_hozac NXA_TYPE_ANY is to remove all addresses. 1144259403 M * waldi daniel_hozac: yes, it removes any mount except / and anything under the vhost root 1144259413 M * daniel_hozac i've never seen the other two. 1144259414 M * Hollow and where do i set the broadcast? in ip[4] or mask[4]? 1144259421 M * daniel_hozac ip[0] 1144259428 M * daniel_hozac (IIRC) 1144259428 M * Hollow ok.. 1144259452 M * Hollow semary has something to do with ipc semaphores, but i don't know exactly what it is 1144259460 M * daniel_hozac waldi: so then you'll have to restart a vserver to mount things in it. is that really the best idea? 1144259490 M * waldi daniel_hozac: no, you can always use the device 1144259498 M * daniel_hozac waldi: but that would require the device on the inside. 1144259506 M * waldi daniel_hozac: hu? 1144259511 M * daniel_hozac waldi: as the /dev on the host would be inaccessible, no? 1144259511 M * waldi daniel_hozac: you have the original /dev 1144259554 M * daniel_hozac waldi: if the host's / is not in the guest's namespace, i.e. mount --bind's won't work, why should device mounts work unless the guest has the device? 1144259573 M * waldi daniel_hozac: EPARSE 1144259632 M * daniel_hozac waldi: it's simple. if you've cleaned up the namespace, the host's / is no longer accessible, how do you expect device mounts to work? 1144259666 M * waldi daniel_hozac: you can't umount / 1144259690 M * daniel_hozac even in the namespace? 1144259697 M * waldi yes 1144259706 M * waldi even in an initramfs 1144259728 M * daniel_hozac ok then, nevermind me :) 1144259734 M * waldi you can only move a filesystem over / but the original will never go away 1144259755 M * waldi okay, maybe i should do an exception for /dev 1144259757 M * daniel_hozac Hollow: i guess that's just the array index for the semary accounting (whatever that is...). 1144259773 M * waldi so that a complete /dev on a udev system is available 1144259787 M * Hollow yeah, whatever it is, guess i have to wait for Bertl :) 1144259840 M * Hollow anyway... the api doc for libvserver is now up to date... http://home.xnull.de/work/vserver/libvserver/doc/ 1144259944 M * waldi daniel_hozac: something like http://paste.debian.net/5710 1144260200 M * doener_ so... not having followed development closely lately, are the legacy interfaces (in 2.0.2) still needed with 0.30.210, and are there any patches I should apply to util-vserver? 1144260270 J * shedi ~siggi@inferno.lhi.is 1144260279 M * Hollow doener_: all patches that i know of can be found at http://dev.croup.de/proj/gentoo-vps/browser/util-vserver/patches/0.30.210-r7 (they include daniel_hozac's patches, mine and gentoos) 1144260299 M * daniel_hozac doener_: no. 0.30.210 removed the need for legacy. 1144260311 M * daniel_hozac doener_: assuming you don't want to use vshelper to reboot guests ;) 1144260351 M * Hollow daniel_hozac: ah-hm... i.e. !CONFIG_LEGACY = vshelper won't reboot properly? 1144260359 M * daniel_hozac Hollow: right. 1144260365 M * Hollow d'oh. 1144260381 M * doener_ daniel_hozac: both legacy APIs? (ie the "Disable Legacy Networking..." option) 1144260383 M * Viper0482 hi 1144260386 M * daniel_hozac well, i'm not sure if that's true for 2.0.2-rc15. 1144260399 M * daniel_hozac doener_: legacy net is not required. 1144260404 M * Hollow hm.. what's the status on the new vx_wait in u-v? 1144260408 M * doener_ ok 1144260457 M * nebuchadnezzar hmm, interessting, in my kernel with vs2.1.1-rc15, it says that I must prefer enable legacy net 1144260477 M * daniel_hozac Hollow: i'm not sure how the helper patches in 2.0.2-rc15 changed the behaviour. 1144260488 M * daniel_hozac nebuchadnezzar: hmm? 1144260513 M * Hollow mhm.. 1144260572 M * daniel_hozac Hollow: it's possible it works with -rc15. and as far as the new helper is concerned, Enrico said he'd do it, and i haven't heard anything yet. 1144260586 M * Hollow ok, will probably test it tomorrow.. 1144260625 M * nebuchadnezzar daniel_hozac: well, Disable Legacy Networking Kernel API: CONFIG_VSERVER_NGET: Do not disable it unless you exactly know what you are doing 1144260625 M * daniel_hozac legacy seems to still be required for proper vshelper functionality. 1144260645 M * daniel_hozac nebuchadnezzar: that's a different option. 1144260650 M * nebuchadnezzar ha ok 1144260694 M * daniel_hozac nebuchadnezzar: hmm, apparently not. 1144260696 M * nebuchadnezzar doener_: legacy net is not required. <-- I juste try to understand 1144260725 M * daniel_hozac nebuchadnezzar: i thought those were separate options, and that configuration help was written back when util-vserver required it. 1144260731 M * daniel_hozac (i.e. before 0.30.210) 1144260757 M * nebuchadnezzar ok, so I can set Yes to this disable 1144260782 M * daniel_hozac that's the idea. 1144260846 M * nebuchadnezzar ok 1144261080 M * Viper0482 is it possible to run dhcpd inside a vserver? 1144261099 M * bonbons I'm running a kernel 2.1.1rc15 with disabled legacy (v-u 1.0.4 branch), works fine 1144261105 M * nebuchadnezzar Viper0482: there's a howto for this 1144261109 M * nebuchadnezzar IIRC 1144261167 M * nebuchadnezzar google give me http://lena.franken.de/linux/debian_and_vserver/dhcpd.html 1144261203 M * daniel_hozac bonbons: right, but util-vserver's vshelper won't. 1144261241 M * Viper0482 yes, but it is with the "old" vserver 1144261250 M * nebuchadnezzar http://linux-vserver.org/Linux-Vserver+FAQ 1144261260 M * nebuchadnezzar G 2 1144261261 M * bonbons don't rememer on which box I tried vs_helper (just a script, not u-v's) 1144261269 M * brc bertl around? Someone please tell bertl that the test enviroment for user quota is done. 1144261275 M * bonbons reboot this box to get 2.1.1rc15 running on it 1144261339 Q * bonbons Quit: Leaving 1144261344 M * Viper0482 thank you, will try that 1144261406 A * nebuchadnezzar love pringles hot&spicy 1144261412 M * nebuchadnezzar oops, wrong buffer 1144261530 J * bonbons ~bonbons@83.222.39.180 1144261959 M * nebuchadnezzar daniel_hozac: if I understand correcly, you need to fix all the clone calls in util-vserver for sparc ? 1144261978 M * nebuchadnezzar in fact all the test to see if we are in child or parent 1144261987 M * daniel_hozac nebuchadnezzar: well, that, or make the kernel return 0 in the child as is expected. 1144262023 M * nebuchadnezzar it can be intrusive with other tool wich call clone and expect a not 0 return value for children ? 1144262071 M * daniel_hozac yeah. 1144262081 M * nebuchadnezzar the clone manpage speak about the caller, but do not mention the chile 1144262083 M * nebuchadnezzar child 1144262089 M * daniel_hozac yeah, i noticed that too. 1144262178 M * nebuchadnezzar erf 1144262222 M * nebuchadnezzar so, what's the expected behavior ?... 1144262236 M * daniel_hozac that the clone returns 0 in the child. 1144262243 M * daniel_hozac as it does on every other architecture :) 1144262249 M * nebuchadnezzar ha ok 1144262253 M * nebuchadnezzar it's sparc specific 1144262267 M * nebuchadnezzar I can ask on sparclinux 1144262278 M * nebuchadnezzar David S. Miler is very active as I see 1144262286 M * nebuchadnezzar :-) 1144262461 M * doener_ daniel_hozac: i just wonder... is there a return code in the child at all? 1144262479 M * daniel_hozac doener_: yeah, it returns the parent's pid. 1144262504 M * waldi okay, works 1144262579 M * Hollow daniel_hozac: where can i find the latest vlogin patch? 1144262582 M * doener_ daniel_hozac: ah damn, libc vs. kernel clone... I always mess that up 1144262588 M * nebuchadnezzar daniel_hozac: I compile your code on my Athlon, and see this 1144262595 M * nebuchadnezzar pid = 0 1144262595 M * nebuchadnezzar Parent: wait(): No child processes 1144262595 M * nebuchadnezzar pid = 5740 1144262595 M * nebuchadnezzar Parent: pid 5740 exited with status 65280 1144262612 M * daniel_hozac nebuchadnezzar: yeah, i customized it to track down your issue. 1144262618 M * nebuchadnezzar ok 1144262653 M * daniel_hozac Hollow: http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch 1144262662 M * Hollow thx 1144262675 M * daniel_hozac Hollow: or in CVS, http://cvs.hozac.com/viewcvs/rpms/util-vserver/fedora-5/?root=rpms 1144262691 M * daniel_hozac umm, i mean http://cvs.hozac.com/viewcvs/util-vserver/fedora-5/?root=rpms 1144262812 M * doener_ daniel_hozac: hm, in lib_internal/sys_clone.h, could we have "pid_t ret = syscall(...); return (ret == getpid()) ? 0 : ret;" there to solve that? 1144262826 M * doener_ (in sys_clone that is) 1144262837 M * daniel_hozac getpid() needs to be called before the clone for that to work. 1144262893 M * doener_ ah, right, read the output in the wrong order 1144262907 M * Hollow daniel_hozac: hm, what's terminal_end for? i understand what it does, but not why it does.. 1144262940 M * daniel_hozac Hollow: when SIGCHLD arrives, there's nothing guaranteeing that all output to the terminal has been sent to stdout. 1144262953 M * daniel_hozac Hollow: i was getting truncated output. 1144262962 Q * Viper0482 Ping timeout: 480 seconds 1144262988 M * Hollow ah.. as i said: i hate signals :P 1144262999 M * daniel_hozac hehehe, yeah, i agree. 1144263293 N * Bertl_oO Bertl 1144263305 M * doener_ wb Bertl 1144263309 M * Bertl back now ... what did the testing bring? 1144263321 M * daniel_hozac clone returns the parent's pid in the child process. 1144263342 M * Bertl hmm .. funny stuff 1144263412 M * daniel_hozac bertl around? Someone please tell bertl that the test enviroment for user quota is done. 1144263428 M * Bertl ah, good ... 1144263445 J * Viper0482 ~Viper0482@p54974EE8.dip.t-dialin.net 1144263665 M * Bertl brc: can you show me _what_ kind of test you currently do? 1144263769 M * Bertl brc: the first step will be to disable the default quuota hash assignment, which basically disables quota, but we need to test that in a quota environment, as we want to catch all issues with that first 1144263965 M * Bertl daniel_hozac: any workaround/fix for that yet? did nebuchadnezzar test something? 1144264089 M * daniel_hozac no, not yet. should be pretty simple though. 1144264115 M * Bertl is this a mainline bug, i.e. does 0.30.210 have it? 1144264128 M * daniel_hozac well, i'd say it's a kernel bug. 1144264134 M * pollux is there a way to tell a vserver to use a physical interface (without specifying the ip address) ? 1144264151 M * Bertl pollux: no, as the networking is ip based 1144264157 M * pollux hum 1144264167 M * pollux ok, so maybe there is a workaround 1144264195 M * Bertl probably, what do you want to work around? 1144264198 M * pollux i have an interface in trunk (8021q) mode, transporting all VLANs 1144264210 M * pollux I would like to "bind" a vserver to a specific VLAN 1144264213 M * Bertl I hope more of them 1144264234 M * Bertl i.e. more interfaces, otherwise trunking doesn't make sense, no? 1144264242 M * pollux yes :) 1144264263 M * pollux eth1 is the physical interface, in trunk mode 1144264268 M * Bertl okay, vlan is no problem, how is that ip related? 1144264274 M * pollux other interfaces are named vlan50 etc. 1144264292 M * Bertl so be it ... 1144264307 M * pollux I do not really have a fixed ip in the other vlans, i wanted to reduce the number of config files to be changed when the ip changes 1144264313 M * daniel_hozac nebuchadnezzar: just to rule out that it's a more serious bug, could you test http://daniel.hozac.com/vserver/wait-test.c ? 1144264323 M * Bertl pollux: (usually you name them bond0.50 or so nowadays) 1144264376 M * pollux I named it vlan50 so the ifup script automatically finds the vlan id :) 1144264384 M * nebuchadnezzar daniel_hozac: Child: pid 31005 1144264384 M * nebuchadnezzar pid = 31005 1144264384 M * nebuchadnezzar Parent: pid 31005 exited with status 0 1144264417 M * daniel_hozac sorry, forgot to copy it, refresh. 1144264427 M * Bertl pollux: okay, usually the ifup scripts deduce that from the .xy naming 1144264441 M * nebuchadnezzar Child: pid 31019, clone returned 31018 1144264441 M * nebuchadnezzar Parent: pid 31019 exited with status 0 1144264441 M * nebuchadnezzar Child: pid 31020, clone returned 0 1144264441 M * nebuchadnezzar Parent: pid 31020 exited with status 0 1144264447 M * pollux Bertl: ok, i can rename it, but what would it change ? 1144264461 M * Bertl nothing, I just mentioned it, go one with the real issue :) 1144264462 M * daniel_hozac nebuchadnezzar: ok, thanks. 1144264467 M * pollux :) 1144264518 M * pollux IIRC, vserver use vlans - is there a way to use a specific vlan on the trunk interface ? (that would solve the problem, no ?) 1144264734 M * Bertl well, linux-vserver do not use vlans explicitely, they are solely ip based 1144264752 M * Bertl of course, you can also create and/or use ips assigned to vlans 1144264787 Q * doener_ Quit: leaving 1144264867 M * pollux hum, not sure i understand 1144264889 M * Bertl you assign IPs to the guests, not interfaces or vlans 1144264900 M * Bertl the guest is then restricted to those IPs 1144264945 M * pollux so i can set up the virtual interface vlan50 without assigning it an ip ? (and give the ip to the vserver, using this interface ?) 1144264996 M * pollux ouch seems not 1144264997 M * pollux creation of VLAN_PLUS_VID_NO_PAD devices is not supported; please 1144264997 M * pollux create them before starting the vserver and use the 'nodev' flag then 1144265053 M * Bertl yes, you can, but you probably want the novlandev entry 1144265062 M * Bertl usually the tools would create the vlan for you 1144265079 M * Bertl (if it was named in the standard way, i.e. eth0.50) 1144265085 M * pollux I'm reading http://vserver.13thfloor.at/Stuff/VServer-IP-Setup-0.1.txt 1144265103 M * pollux should I use the point 3.1 ? (existing interface) ? 1144265178 M * Bertl the question is, what do you want to accomplish, and what should the tools do for you 1144265194 M * Bertl for example, I have the following setups in use: 1144265220 M * Bertl a) bond0.20 already exists, we only want to add the ip (and remove it) when the guest is started (stopped) 1144265238 M * Bertl b) eth0 exists, the guest will get a new ip on a private vlan (.10) 1144265273 M * pollux a) would be ok 1144265286 M * Bertl the text you are referring to is more than two years old and doesn't apply to current tools 1144265296 M * pollux :/ 1144265310 J * michal` ~michal@www.rsbac.org 1144265363 M * Bertl pollux: okay, in this case you need to put ip, prefix, dev and novlandev into the config 1144265411 M * Bertl the dev would contain vlan50 1144265427 M * Bertl the ip and prefix the usual, e.g. 192.168.0.1 and 24 for 192.168.0.1/24 1144265439 M * Bertl the novlandev just has to exist as file 1144265443 M * pollux I create files in /etc/vservers/vs50/interfaces/0/ 1144265445 M * pollux ? 1144265452 M * Bertl yep precisely 1144265465 M * Bertl e.g. echo 192.168.0.1 >/etc/vservers/vs50/interfaces/0/ip 1144265583 M * pollux ok, I set the prefix to 23. Do I have to put something in the 'novlandev' file, or leave it empty ? 1144265620 M * Bertl just touch it, i.e. empty 1144265628 J * Smutje_ ~Smutje@xdsl-87-78-19-188.netcologne.de 1144265678 M * pollux again the same error 1144265686 M * daniel_hozac but it will work. 1144265699 M * daniel_hozac (IIRC) 1144265720 M * pollux ah, touch nodev make things a bit better (but another error now) 1144265728 Q * Smutje Ping timeout: 480 seconds 1144265728 N * Smutje_ Smutje 1144265773 M * pollux ok, seems to work now 1144265792 M * pollux daniel_hozac: i'm not sure, but there is maybe another error in the debian package 1144265828 M * pollux daniel_hozac: vserver vs50 start was complaining that the directory /etc/vservers/.defaults/run.rev/ didn't exist 1144265874 M * pollux Bertl: thanks a lot ! 1144265885 M * daniel_hozac pollux: that would seem like a package issue. micah? 1144266003 M * pollux works like a charm, nice work 1144266031 M * Bertl pollux: you're welcome! 1144266033 M * pollux i still miss IPv6, but I'll wait ;) 1144266048 M * Bertl pollux: it will come, bonbons started working on it already 1144266060 M * pollux great 1144266116 M * brc bertl 1144266116 M * brc bertl around? Someone please tell bertl that the test enviroment for user quota is done. 1144266186 M * Bertl yeah, got that :) tx 1144266249 M * micah pollux: if you moved your vserver run.rev directory you will need to adjust those symlinks 1144266264 M * brc OK :) 1144266268 M * pollux micah: i didn't moved anything 1144266392 M * micah pollux: ls -l /etc/vservers/.defaults/run.rev 1144266401 M * micah err 1144266404 M * micah pollux: ls -ld /etc/vservers/.defaults/run.rev 1144266421 M * pollux drwxr-xr-x 2 root root 120 2006-04-05 21:42 /etc/vservers/.defaults/run.rev 1144266444 M * pollux i had to create the directory, it didn't exist 1144266456 M * daniel_hozac pollux: it should be a symlink... 1144266583 M * micah the debian post installation checks to see if that symlink exists, and if it doesn't it does mkdir /var/run/vservers.rev; ln -s /var/run/vservers.rev /etc/vservers/.defaults/run.rev 1144266860 M * Bertl brc: brc: can you show me _what_ kind of test you currently do? 1144266909 M * brc Bertl: i just created the boxes for the tests , don't know which tests should be run 1144267003 M * Bertl okay, then please try to figure some things (tests) like starting quota, allocating files from root and user accounts and such 1144267035 M * Bertl if you can script some of them, the better 1144267092 M * brc ok goign to do it in some minutes 1144267267 M * bonbons Bertl: in 2.1.1rc15 the "error" file on CoW (write; IUNLINK,~IMMUTABLE) has gone and now write succeeds 1144267410 M * bonbons if IUNLINK or IMMUTABLE is set alone no CoW is happening, if set together chown, chmod and touch fail 1144267411 M * Bertl yep, fixed that, tx for testing 1144267503 J * restill ~restill@c-24-11-171-10.hsd1.mi.comcast.net 1144267519 M * bonbons shall I upload the results? 1144267731 M * Bertl yes please ... 1144267735 M * Bertl welcome restill! 1144267766 M * restill Hey B. How are things? 1144267785 M * Bertl fine thanks, and for you? 1144267839 J * Hollow_ ~hollow@home.xnull.de 1144267846 Q * Hollow Read error: Connection reset by peer 1144267879 M * bonbons Bertl: here they are: http://homepage.internet.lu/brunop/vserver/CoW-test-vs2.1.1rc15 1144267962 M * restill good. I am lovin the vserver. My distributed app was tested successfully on vservers and now I am installing my app on other computers and is working well. I attribute it to being able to develope it using vservers. 1144268082 M * daniel_hozac Bertl, nebuchadnezzar: line 676 of arch/sparch64/kernel/process.c 1144268089 M * nebuchadnezzar yes 1144268089 M * Bertl restill: great! 1144268120 M * daniel_hozac nebuchadnezzar: could you try changing current->pid to 0?' 1144268127 M * daniel_hozac (have Bertl verify this first though) 1144268160 M * daniel_hozac restill: FYI, i've got my own x86_64 box now for building. thanks a lot for building them up until now. 1144268212 M * Bertl daniel_hozac: means we should add restill to the Hall'o'Fame, no? 1144268242 M * daniel_hozac sure. 1144268269 M * Bertl could you do that with a few word what he provided? 1144268270 M * nebuchadnezzar daniel_hozac: ok, I'll do this 1144268304 M * daniel_hozac building the Fedora Core 4 kernels and utils for x86_64. 1144268306 M * restill no prob. If you need more, let me know. 1144268369 M * restill It seems that with the dual cores, I have processor to spare. 1144268458 M * daniel_hozac hehe, i've been meaning to run the x86 builder on one processor, and the x86_64 on the other. 1144268580 M * restill I am still not sure how my motherboard adjusts the FSB, but it switches between 1 and 2 GHz alot. 1144268695 M * nebuchadnezzar hm, include/linux/vs_context.h is a vserver specific header ? 1144268702 M * daniel_hozac yes, it is. 1144268706 M * nebuchadnezzar I have: 1144268707 M * nebuchadnezzar include/linux/vs_context.h: In function '__leave_vx_info': 1144268707 M * nebuchadnezzar include/linux/vs_context.h:195: warning: value computed is not used 1144268713 M * nebuchadnezzar and for line 210 1144269442 M * ray6 ree 1144269498 M * nebuchadnezzar daniel_hozac: I look at 2.6.17-rc1 code, the current->pid is used here too 1144269679 M * nebuchadnezzar daniel_hozac: look at the comment line 614, it seems to mimik sunOS behavior if I understand correctly 1144269733 M * daniel_hozac nebuchadnezzar: hmm, ah. 1144269964 Q * restill Quit: Leaving 1144270078 M * nebuchadnezzar daniel_hozac: I'll wait before installing and booting this kernel 1144270129 M * nebuchadnezzar I ask to the sparclinux mailing list to see if the current code is what everybody expect or not 1144270277 M * daniel_hozac it seems to be what SunOS expects, but it's not what Linux expects... 1144270288 J * macsim ~macsim@88-106-20-57.dynamic.dsl.as9105.com 1144270295 M * Bertl welcome macsim! 1144270319 M * macsim hi Bertl 1144270333 M * nebuchadnezzar Bertl is a bot in fact, it welcome everybody few microsecond after the arrival 1144270333 M * Bertl daniel_hozac: what really makes me wonder is, that a year ago or so, I tested it on sparc64 and the tools just worked fine 1144270334 M * nebuchadnezzar :-) 1144270365 M * Bertl yeah, but I'm a freindly bot :) 1144270368 M * daniel_hozac Bertl: before namespaces were used? 1144270374 M * Bertl *friendly even 1144270386 M * Bertl daniel_hozac: could be, could be .. 1144270483 M * nebuchadnezzar 2.4.18 use the current->pid daniel_hozac saw 1144270496 M * nebuchadnezzar how old is 2.4.18 ? 1144270509 M * Bertl pretty, but eyck has details :) 1144270511 M * daniel_hozac very :) 1144270522 M * waldi nebuchadnezzar: 5 years? 1144270541 M * nebuchadnezzar so, this behavior is in place since so long... 1144270573 M * daniel_hozac i guess user space is expected to work around it. 1144270581 M * nebuchadnezzar yes 5years, I begin linux with this 1144270599 M * nebuchadnezzar potato powa 1144270630 Q * macsim Remote host closed the connection 1144270857 Q * ourseven Remote host closed the connection 1144270884 J * ourseven ~ourseven@tor-irc.dnsbl.oftc.net 1144270944 M * Bertl wb ourseven! 1144271077 Q * Viper0482 Remote host closed the connection 1144271129 M * nebuchadnezzar ok, Time To Sleep now, I work tomorrow 1144271163 M * daniel_hozac nebuchadnezzar: http://daniel.hozac.com/vserver/util-vserver-0.30.210-sparc64.patch 1144271205 M * nebuchadnezzar daniel_hozac: ok, I'll give this a try before going :-) 1144271309 M * Bertl daniel_hozac: btw, this is sparc32 :) 1144271357 M * nebuchadnezzar I get a reject :-/ on my --dry-run 1144271391 M * daniel_hozac Bertl: right :) 1144271427 M * daniel_hozac nebuchadnezzar: i guess you have the s390 fix too? i don't have that patch. 1144271446 M * nebuchadnezzar I use debian patch 1144271502 M * daniel_hozac if you could get vanilla util-vserver, it should apply. 1144271518 M * Bertl try -l 1144271549 M * nebuchadnezzar I may resolv the conflict 1144271771 M * nebuchadnezzar micah: I think you maintain the debian package 1144271805 M * nebuchadnezzar the return is is not good :-/ __NR_sys_clone instead of the defined __NR__sys_clone 1144271820 M * daniel_hozac i agree, that patch is broken. 1144271827 M * daniel_hozac i think waldi authored that one though ;) 1144271869 M * Skram hi all 1144271872 M * daniel_hozac although, it won't matter for the Debian package. 1144271887 M * nebuchadnezzar there's some problem with sys_clone and _sys_clone, like call _syscall2 with _sys_clone but sys_clone is defined 1144271911 M * daniel_hozac remove those #define's. 1144271922 M * waldi daniel_hozac: no, it is not 1144272000 J * yzenezy ~tom@83.70.151.186 1144272008 M * Bertl welcome yzenezy! 1144272014 M * micah nebuchadnezzar: I am working with Ola on it, yes 1144272014 M * daniel_hozac waldi: how is it not broken to define __NR__sys_clone and then use __NR_sys_clone? 1144272026 M * waldi daniel_hozac: is uses __NR__sys_clone 1144272030 M * yzenezy Bertl: hi :) 1144272047 M * daniel_hozac waldi: not in the traditional syscall case. 1144272057 M * micah + return syscall(__NR_sys_clone, stack, flags); 1144272061 M * waldi daniel_hozac: which is used where? 1144272075 M * daniel_hozac waldi: nowhere for the Debian package. 1144272095 M * micah waldi: that is in the s390syscall patch 1144272117 M * nebuchadnezzar http://www.asgardr.info/~nebu/vserver/sys_clone.h.orig <-- that the debian version 1144272157 M * waldi bah, okay 1144272169 M * daniel_hozac waldi: and you use #error instead of #else there too... 1144272180 M * Bertl in general, I'd opt for __NR_clone, but probably I'm alone with that :) 1144272214 M * daniel_hozac __NR_clone means that _syscall2 will want to create the function clone. 1144272228 M * Bertl which is fine, no? 1144272252 M * daniel_hozac i guess not, Enrico must've had some reason to use sys_clone instead. 1144272334 M * nebuchadnezzar http://www.asgardr.info/~nebu/vserver/sys_clone.h.patched <-- that do you think about this ? 1144272352 M * Bertl s/that/what 1144272371 M * nebuchadnezzar I do not understand why #include get out of #define ENSC_SYSCALL_TRADITIONAL but I follow... 1144272372 M * nebuchadnezzar :-/ 1144272375 M * nebuchadnezzar ;-) 1144272396 M * daniel_hozac getpid is in unistd.h 1144272405 M * nebuchadnezzar ok 1144272420 M * daniel_hozac line 41 should be #else 1144272422 M * nebuchadnezzar I'll be less idiot tonight 1144272463 M * daniel_hozac and the !ENSC_SYSCALL_TRADITIONAL branch lacks the s390 patches. i guess that won't matter for your testing though. 1144272477 M * nebuchadnezzar ok 1144272821 Q * mkhl Quit: 1144272888 M * nebuchadnezzar good, I learn how dpatch works :-) 1144272983 M * micah nebuchadnezzar: dpatch-edit-patch 1144273016 M * nebuchadnezzar well, I remove the 06_s390 in debian/patches/00list :-) 1144273019 M * nebuchadnezzar by hand 1144273031 M * micah nebuchadnezzar: if you fix the s390 patch and/or get a sparc64 patch working, submit a bug report with the patches attached and I'll include them in the package 1144273043 M * micah nebuchadnezzar: you weren't able to fix the patch? 1144273078 M * nebuchadnezzar micah: well, I just copy the sys_clone.h on a fresh debian tree 1144273084 M * nebuchadnezzar patch embedded 1144273095 M * nebuchadnezzar so I don't need the s390 patch 1144273112 M * nebuchadnezzar but I will make this correct and nice if it works :-) 1144273279 M * nebuchadnezzar daniel_hozac: it works 1144273290 M * nebuchadnezzar my sarge01 is started 1144273301 M * nebuchadnezzar and I can enter it 1144273306 M * daniel_hozac nebuchadnezzar: that's great. 1144273307 M * nebuchadnezzar many thanks 1144273335 M * nebuchadnezzar I'll supply patch for s390 and sparc64_sys_clone 1144273417 M * daniel_hozac it's sparc_sys_clone, as Bertl pointed out ;) 1144273428 M * daniel_hozac (meaning it's for sparc32 as well) 1144273435 M * nebuchadnezzar ok ok 1144273481 M * micah nebuchadnezzar: please do supply patches in a bug report, I will integrate them in the package and we'll provide them inthe upstream BTS 1144273491 M * nebuchadnezzar ok 1144273555 M * nebuchadnezzar I send a corrected s390 patch or a patch to patch the patch ? I supposed a corrected version ;-) 1144273635 M * nebuchadnezzar daniel_hozac: http://marc.theaimsgroup.com/?l=linux-sparc&m=114427223414160&w=2 1144273716 Q * bonbons Quit: Leaving 1144273725 M * daniel_hozac nebuchadnezzar: that's sort of the answer i expected. 1144273875 M * nebuchadnezzar so, how did it work before as Bertl said 1144273876 M * nebuchadnezzar ? 1144274203 M * daniel_hozac well, he wasn't sure when he tested it. if it was before namespaces were introduced, it would've worked fine. 1144274435 J * Dr4g ~Dr4g@80-195-133-167.cable.ubr06.uddi.blueyonder.co.uk 1144274468 Q * yzenezy Ping timeout: 480 seconds 1144274497 M * micah nebuchadnezzar: yes a corrected version is less confusing :) 1144274670 M * nebuchadnezzar ok 1144274682 M * nebuchadnezzar how can I insert a patch between 05 and 06 ? 1144274696 M * nebuchadnezzar with dpath ? 1144275207 J * dearaujo ~dan@pixpat.austin.ibm.com 1144275218 M * Bertl welcome dearaujo! 1144275280 M * dearaujo quick question - for resources limits, I notice the following keywords on the "flower" page - fsize, data, stack, core, rss, nproc, nofile, memlock, as and locks. What is rss, nofile, memlock, as, and core? 1144275317 M * Bertl various limits ... 1144275347 M * dearaujo i assume fsize == filesize, nproc == # of process 1144275349 M * daniel_hozac lol 1144275361 M * Bertl dearaujo: give me a minute to find the wiki page 1144275365 M * dearaujo thx 1144275371 M * dearaujo couldnt find anything 1144275436 M * Bertl http://linux-vserver.org/Resource+Limits 1144275459 M * dearaujo ah thanks 1144275495 M * brc bertl 1144275505 M * brc the script is gonna set some pre-defined values to quota and check if they are right 1144275506 M * brc is that ok ? 1144275513 M * brc should i try filling up that space ? 1144275513 P * dearaujo 1144275537 M * Bertl yes, that would be good, also check that adding so and so many disk blocks get accounted 1144275548 M * Bertl and same for the inodes of course 1144275560 M * brc you mean, after adding each file check if that was accounted 1144275562 M * brc is that it ? 1144275585 M * Bertl yep 1144275602 M * brc ok, finishing the script. 1144276178 Q * Dr4g Read error: Connection reset by peer 1144276845 M * brc bertl, almost :) 1144276846 M * brc hehe 1144276853 M * derjohn hey, are here any folks from openvps? or does anyone them? I would like to invite them to the linuxtag booth ... 1144277039 M * Bertl derjohn: you might dig through the irc logs, IIRC we know who they were :) 1144277072 M * Bertl (hint: there is a search page :) 1144277125 M * derjohn well, there is even a ML, but I hope some would hang out here... (grrr .. again a ML I have to register ...) 1144277147 M * nebuchadnezzar micah: ok, patch sends 1144277155 M * derjohn Bertl, BTW: you have concrete ideas about what to do a the linuxhotel? 1144277157 M * daniel_hozac nebuchadnezzar: bug number? 1144277193 M * Bertl derjohn: I expect something like a lan party with a lot discussion going on :) 1144277197 M * daniel_hozac nebuchadnezzar: i want to see if you solved the conflicts in the same way i did :) 1144277227 M * nebuchadnezzar daniel_hozac: well, I'm waiting the aknowlegement of BTS 1144277239 M * micah daniel_hozac: you should subscribed to the util-vserver package so you get emails on all bugs 1144277325 M * nebuchadnezzar I set the sparc issue to important 1144277329 M * nebuchadnezzar I hope it's correct 1144277338 M * derjohn Bertl, hehe, yes ... so all the discussion about hardware is useless ;) [except if you bring a 'big mamma' oct athlon or so ] 1144277539 M * nebuchadnezzar ok, now time to sleep 1144277556 M * nebuchadnezzar good night 1144277588 M * daniel_hozac good night! 1144277731 M * daniel_hozac micah: how would i do that? 1144278089 Q * f_ Quit: This computer has gone to sleep 1144278112 Q * Hollow_ Remote host closed the connection 1144278120 J * Hollow ~hollow@home.xnull.de 1144278139 M * Bertl wb Hollow! 1144278178 M * Hollow hey Bertl 1144278200 M * Bertl so it's not a gentoo bug :) 1144278204 M * Hollow hm? 1144278242 M * Bertl it's an OVZ .config issue :) 1144278259 M * Hollow -v? 1144278277 M * Bertl seems you missed the other channel :/ 1144278289 M * Hollow #openvz? 1144278373 M * brc BERTL! DONE 1144278379 M * Hollow Bertl: btw, all new syscall versions are now in libvserver and i'll soon release 2.0 (usable for >=2.{0.2,1.1}) 1144278383 M * brc Bertl: Wanna check it out ? 1144278410 J * lilalinux_ ~plasma@dslb-084-058-202-082.pools.arcor-ip.net 1144278411 M * Bertl brc: let's see test results or so ... 1144278432 M * brc the script does the following: 1144278447 M * brc loop from I=1 to N, at each iteration it removes quota and shows the output from "quota" 1144278492 M * brc then it sets quota blcoksoftlimit to I*512 and HARDLIMIT to I*1024 1144278495 M * Bertl okay, user and group quota? 1144278496 Q * matta Ping timeout: 480 seconds 1144278522 M * brc creates a file of size I*512 with that users and shows quota output again, and then creates a file with size I*1024 and shows quota output 1144278530 M * brc (the last one must return a warning) 1144278545 M * brc Hmm. shouldn't we leave group quotas as the next step? :D 1144278547 M * brc :) 1144278561 M * Bertl okay, you might actually want to try to go over limit too, and check that you cannot cross it 1144278584 M * Bertl also, important is that you test for root uid/gid=0 and some other uid/gid too 1144278626 M * Bertl but you can refine your script while we test/work 1144278789 M * brc ok made it go over the hard limit 1144278802 M * brc Ok. so i should set a quota for root too ? 1144278804 M * Bertl ah, and we start with ext2 1144278823 M * Bertl (because it has no journa) 1144278826 M * Bertl +l 1144278852 Q * lilalinux Ping timeout: 480 seconds 1144278866 M * brc ok so i will make this ext3 go back to ext2 1144278885 M * Bertl yeah, just remove the yournal and mount it ext2 1144278904 M * Bertl *journal 1144278911 M * brc doing it 1144279058 M * Bertl btw, would be good if the output could be easily diffed, so that we can detect changes 1144279123 M * Bertl what is your current kernel, btw? 1144279340 M * brc sorry for the delay, was answerring an e-mail 1144279358 M * brc Linux bertl 2.6.15.4 #1 Thu Feb 16 14:33:55 BRT 2006 i686 GNU/Linux 1144279368 M * brc Now we have ext2 1144279372 M * brc /dev/hda2 on / type ext2 (rw,errors=remount-ro,usrquota,grpquota) 1144279398 M * brc You mean, the scritp will generate an output , then we make changes and run again and DIFF 1144279399 M * brc is that it ? 1144279415 M * brc It is ok to be diffed the way it is right now i think :) 1144279496 M * Bertl okay, good, let's update the kernel to 2.6.16 (or 16.1) with vs2.1.1-rc15 1144279509 M * Bertl keep the built kernel source tree around 1144279520 M * Bertl you will get patches from me to apply ontop of that 1144279559 M * Bertl then rebuild and test. make sure that you can change the kernel easily, and also enable kernel debugging, and vserver debugging 1144279582 M * Bertl especially the history stuff and spinlock checks, we don't need slab or memory debugging 1144279798 M * brc ok 1144279814 M * brc Downloading 2.6.16.... 1144279909 M * brc It will take sometime until thekernel is compiled :) 1144279912 M * brc and downloaded 1144279921 M * brc On the next compiles it will be faster because of ccache 1144279931 M * brc changing kernels wont be a problem 1144279953 M * Bertl okay, as I said, leave the build tree around 1144279961 M * Bertl we need it for checking symbols and such 1144280208 M * brc ok! 1144280385 M * daniel_hozac nebuchadnezzar, micah: hmm, that patch won't do it. 1144280462 M * brc what about the root stuff? i ddint include root on the tests :) 1144280545 M * daniel_hozac nebuchadnezzar, micah: i merged the two patches, http://daniel.hozac.com/vserver/util-vserver-0.30.210-clone-arch.patch 1144280743 M * brc bertl wher can i get vs2.1.1-rc15 ? 1144280773 M * daniel_hozac http://vserver.13thfloor.at/Experimental/patch-2.6.16-vs2.1.1-rc15.diff.bz2 1144280779 M * brc 10x 1144281007 M * derjohn what exactly happens to a request for 127.0.0.1 within a guest? it's mapped to the primary Ip the guest? are there maybe patches out that create a 127.0.0. as dummy? that would make configuration lost easier, becuase many apps have 127.0.0.1 in the config by default or even hardcoded 1144281035 M * daniel_hozac why is the hardcoded 127.0.0.1 a problem? 1144281044 M * daniel_hozac as you said yourself, it's rewritten to the guest's first IP address. 1144281114 M * derjohn well, if the particular port is already bound? if you dont want that services accessible from the outer world? 1144281137 M * derjohn (e.g. amavisd-new) 1144281166 M * daniel_hozac why on earth would you have two different services listening on the same port on different IP addresses? 1144281170 M * daniel_hozac that would be extremely confusing. 1144281200 M * derjohn agreed, but the amavis-question? 1144281200 M * daniel_hozac and if you don't want the it accessible, firewall, or add another IP address that is not public. 1144281212 M * daniel_hozac i've never setup amavis :) 1144281224 M * derjohn how do you filter spam? bogo ? 1144281234 M * daniel_hozac i don't. :) 1144281258 M * derjohn daniel_hozac, hr hr :) 1144281355 M * derjohn well, two days ago i moved a physical host to a guest ... i costed hours until it ran smoothly.... I dunno why the beast didnt what to work with the 127.0.0.1 mapping. if i find that in a config, i simple read it as ? 1144281399 M * daniel_hozac if is the first IP address of the guest, yes. 1144281513 Q * mugwump Ping timeout: 480 seconds 1144281539 M * brc didnt find the history stuff on the kernel config. the spinlock was selected 1144281551 M * brc is bertl still around? 1144281570 J * matta ~matta@71.224.125.126