1218585680 J * gbp ~gbpowell@ip-203.191.174.34.dsl.sta.onestream.com.au 1218585990 Q * dowdle Remote host closed the connection 1218585998 M * nebuchadnezzar I have some problem migrating vserver to one host to another 1218586123 M * Bertl how so? 1218586142 M * nebuchadnezzar I use rsync to copy the vservers and at destination my uid/gid are screwed up 1218586158 M * Bertl well, do you use the proper rsync options? 1218586178 M * Bertl (I'd suggest -axHPSD --numeric-ids :) 1218586185 M * nebuchadnezzar yes, it preserve the uid, gid, timestamp, and so one 1218586205 M * nebuchadnezzar arf 1218586218 M * nebuchadnezzar there is a --numeric-ids.... :-/ 1218586630 M * nebuchadnezzar thanks Bertl 1218587329 M * Bertl you're welcome! 1218587400 J * doener ~doener@i577BB318.versanet.de 1218587503 Q * doener_ Ping timeout: 480 seconds 1218588050 Q * loddafnir Ping timeout: 480 seconds 1218588266 Q * balbir Ping timeout: 480 seconds 1218588925 J * balbir ~balbir@122.167.240.196 1218589105 Q * nebuchadnezzar Remote host closed the connection 1218589118 J * nebuchadnezzar ~dad@zion.asgardr.info 1218589333 Q * nebuchadnezzar Remote host closed the connection 1218589336 J * nebuchadnezzar ~dad@zion.asgardr.info 1218589577 Q * nebuchadnezzar Remote host closed the connection 1218589580 J * nebuchadnezzar ~dad@zion.asgardr.info 1218589774 Q * nebuchadnezzar Read error: Connection reset by peer 1218589798 J * nebuchadnezzar ~dad@zion.asgardr.info 1218590688 Q * Mojo1978 Read error: Connection reset by peer 1218593543 Q * quiksilv 1218594497 P * gbp 1218597286 Q * FireEgl Ping timeout: 480 seconds 1218598146 Q * balbir Ping timeout: 480 seconds 1218598680 Q * derjohn_mob Ping timeout: 480 seconds 1218599407 M * Bertl off to bed now ... have a good one everyone! cya! 1218599412 N * Bertl Bertl_zZ 1218601811 J * cryptronic ~oli@p54A3B525.dip0.t-ipconnect.de 1218603454 Q * cryptronic Quit: Leaving. 1218604453 Q * xdr Ping timeout: 480 seconds 1218604713 J * xdr ~xdr@gote2.23.cust.blixtvik.net 1218604742 Q * xdr 1218604971 J * xdr ~xdr@gote2.23.cust.blixtvik.net 1218605087 J * sharkjaw ~gab@149-67-194.231210.adsl.tele2.no 1218605647 J * derjohn_mob ~aj@e180195159.adsl.alicedsl.de 1218606702 J * ntrs_ ~ntrs@77.29.71.56 1218607137 Q * derjohn_mob Ping timeout: 480 seconds 1218607465 Q * larsivi Quit: Konversation terminated! 1218607795 Q * fosco Remote host closed the connection 1218607797 J * fosco fosco@konoha.devnullteam.org 1218607817 Q * ag- Remote host closed the connection 1218607829 J * ag- ~ag@fedaykin.roxor.cx 1218610339 Q * nkukard Ping timeout: 480 seconds 1218610773 J * larsivi ~larsivi@85.221.53.194 1218611131 N * DoberMann[ZZZzzz] DoberMann 1218613141 Q * ntrs_ Ping timeout: 480 seconds 1218617282 M * larsivi I was able to install a ubuntu hardy guest, but when starting it, it seems to "hang" after Starting kernel deamon done, Ctrl-C makes it go on to after Running local boot scripts done, and yet another ctrl-c ends the process. It is now possible to enter and use the guest (at least update apt and install packages), but it seems a bit daunting 1218617299 M * larsivi also postgres seems to not actually start/run 1218617515 M * larsivi ah, must check something 1218617674 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1218617683 M * renihs good morning 1218618522 M * larsivi hmm, no, it still hangs in the start script - any way I can debug it? 1218619457 J * dna ~dna@245-227-dsl.kielnet.net 1218619692 J * nkukard ~nkukard@41.6.128.202 1218619747 M * larsivi ah, should kernel log deamon really be started when I start a vserver? 1218620441 M * larsivi removing klogd from inside the guests init setup fixed the startup hanging 1218621198 J * loddafnir ~mike@193.170.48.107 1218622420 J * yarihm ~yarihm@guest-docking-nat-1-103.ethz.ch 1218622542 Q * fluor Ping timeout: 480 seconds 1218622848 Q * xdr Read error: Connection reset by peer 1218623115 J * xdr ~xdr@gote2.23.cust.blixtvik.net 1218623149 J * kir ~kir@swsoft-msk-nat.sw.ru 1218623416 M * renihs i got something half working, i am not sure if it fails now because of my policy routing scenario or if its vserver feature 1218623501 M * renihs i have one vserver guest with a real ip, and i use policy routing to forward packets from this guest to a different gatway over a different interface 1218623524 M * renihs according to tcpdump it works fine (sniffing on vserver host shows packets from guest leaving and responses entering) 1218623536 M * renihs but in the guest i dont receive any packet 1218623566 M * renihs currently i only would like to know if that *should* work or if there might be vserver security feature which might interfere 1218623943 J * ntrs ~ntrs@77.29.76.247 1218623976 Q * infowolfe Quit: infowolfe 1218624248 J * infowolfe ~infowolfe@c-67-160-149-42.hsd1.or.comcast.net 1218624608 J * _nkukard_ ~nkukard@196.212.73.74 1218624658 Q * nkukard Ping timeout: 480 seconds 1218624669 Q * infowolfe Quit: infowolfe 1218627250 J * friendly ~friendly@ppp59-167-89-100.lns2.mel6.internode.on.net 1218627648 M * daniel_hozac renihs: we don't mess with networking stuff, other than to limit source address selection and bind. 1218629035 A * ard sighs 1218629053 M * ard we have heartbeat 0.*, 1.* and 2.* running 1218629065 M * ard And all versions contain the same time(2) wraparound bug :-( 1218629944 Q * Aiken Quit: Leaving 1218631040 M * renihs daniel_hozac, hmm k, thanks already 1218631283 Q * john Ping timeout: 480 seconds 1218631307 J * john ~john@glou.nurvnet.org 1218631596 Q * friendly Quit: Leaving. 1218631659 J * Darkglow ~pdesnoyer@208.71.184.41 1218631803 M * Darkglow Hi, I have a problem here :( (appart from the fact that my keyboard does 2 spaces instead of 1 ;-) Disk space on HDV volumes in vservers is wrong most of t he time... and now it shows 2.3G free (and the underlying filesystem show 9G free) and it says filesystem full when I try to write to it ?! Im running 2.6.22 from Etch backports 1218632216 M * Darkglow (found the answer(partly). inodes number...) But it still does not explain why most filesystem sizes are wrong (I use dlimits) 1218632526 M * daniel_hozac "wrong"? 1218632594 M * Darkglow well, for example : in the vserver I see 2.3G out of 50G available. BUT in reality, when I check from the host, I have 36G used... 1218632654 M * daniel_hozac which means you've messed up the disk limits somehow. 1218632678 M * Darkglow "messed up " ? 1218632684 M * daniel_hozac stop the guest, remove the caches, start it back up again. 1218632711 M * Darkglow cache ? 1218632731 M * renihs daniel_hozac, hmm any obvious reason why a vserver guest would not pick up packets destined for it? (eth1 has no ip adresse, eth1:guest has a public ip), packets do reach the vserver...but seem to get ignored 1218632757 M * Darkglow do you mean .hash directory ? 1218632763 M * daniel_hozac as i said, we don't mess with any of that. 1218632806 M * renihs hmm must be something linux specific then, hmm 1218632809 M * daniel_hozac if there's a service listening on the IP:port, it should get picked up. 1218632818 M * renihs daniel_hozac, its jusst a echo reply 1218632819 M * daniel_hozac check iptables/hosts.deny/etc. 1218632826 M * daniel_hozac i.e. ping? 1218632839 M * renihs if i try to ping from guest = no replies...but if i sniff i see echo requests leaving and echo replies entering 1218632860 M * renihs just inside the client there is no "response" 1218632872 M * daniel_hozac and ping -I google.com on the host works fine? 1218632875 M * renihs nope 1218632879 M * renihs i see request and reply 1218632888 M * renihs but according to "ping" its not pingable (from guest) 1218632903 M * daniel_hozac do it on the host. 1218632909 M * renihs on the host it works 1218632918 M * daniel_hozac with -I ? 1218632921 M * renihs (where i sniff) 1218632927 M * renihs lemme try 1218632948 M * renihs hmm not with guest ip hmm 1218632960 M * renihs interesting must be something i borked 1218633203 Q * sharkjaw Remote host closed the connection 1218633376 J * dna_ ~dna@107-230-dsl.kielnet.net 1218633487 J * RenOS ~renos@athedsl-4416950.home.otenet.gr 1218633515 M * RenOS hello world! 1218633537 M * RenOS does anyone have any tips for starting vservers from an NFS mount point? 1218633556 M * RenOS I have the vdir on the NFS server but I get this error: 1218633565 M * RenOS secure-mount: flock(): No record locks available 1218633565 M * RenOS Failed to update mtab-file 1218633615 M * daniel_hozac use util-vserver 0.30.215+. 1218633645 M * RenOS daniel_hozac, can I find it as a debian package? 1218633661 M * daniel_hozac i'd assume so. 1218633668 M * daniel_hozac micah knows best though. 1218633700 M * RenOS ok, let me have a look around 1218633711 Q * dna Ping timeout: 480 seconds 1218633744 M * Darkglow it is in etch-backports 1218633760 M * Darkglow 0.30.21502~bpo40+1 1218633787 M * RenOS Darkglow, thanks 1218634016 M * RenOS I get the same error: 1218634016 M * RenOS http://mirror.realroute.net/backports.org/pool/main/u/util-vserver/util-vserver_0.30.215-2~bpo40+1_i386.deb 1218634022 M * RenOS secure-mount: lockf(): No record locks available 1218634022 M * RenOS Failed to update mtab-file 1218634027 M * RenOS ^ this is the error.. 1218634130 M * daniel_hozac and you are mounting with -o lock, right? 1218634150 M * RenOS daniel_hozac, probably not. 1218634271 M * Darkglow Ok, I have found the vserver I was talking about... show 2.3G used and vdu shows 36G used... my /etc/vservers/name/cache is symlink and the dlimits folder in it is empty... 1218634304 M * daniel_hozac you have to stop the guest first. 1218634429 M * Darkglow it points to /var/cache/vservers/name and the dlimits directory is empty... it should contain files ? none of my vservers contain file in /var/cache/vserver/ !!? 1218634444 M * daniel_hozac because they're running. 1218634486 M * Darkglow is there docs about this ? I would read that instead of bothering you :-) 1218634531 M * Darkglow so to fix this I need to stop V, then delete cache info, then start ? 1218634550 M * daniel_hozac as i said half an hour ago, yes. 1218634856 M * RenOS OK. It seems I managed to hit a bug concerning the statd on debian/ubuntu 1218634884 M * RenOS I was able to mount -o nolock 1218634889 M * RenOS and the vserver started ok 1218634915 M * RenOS finally I have a vserver running off NFS, not bad.. 1218634918 M * Darkglow :( 1218635015 M * Darkglow automatic calculation of used disk space should occur if I delete the cache and then s tart the vserver ? It did not... I even changed the disk_space and that worked fine, but still showing 2.5G used. 1218635060 M * renihs daniel_hozac, i found the cause...it was partially my policy routing setup....disabling rp_filter for the interface helped :) 1218635113 M * daniel_hozac Darkglow: and that's "wrong"? 1218635120 M * daniel_hozac renihs: ah, yes. 1218635245 J * infowolfe ~infowolfe@c-67-160-149-42.hsd1.or.comcast.net 1218635246 M * Darkglow yes... it should show 37G used 1218635268 M * daniel_hozac why? 1218635295 M * daniel_hozac you obviously only have 2.5 GiB worth of files that belong to the guest. 1218635331 M * Darkglow ok, I see... so I would need to tag them 1218635380 M * Darkglow chxid can be run while online right ? 1218635395 M * daniel_hozac sure, but you'll have to update the disk limit manually. 1218635893 M * Darkglow humm... I stopped the V, did chxid -URx name /vservers/name, removed the cache a nd then started the V... still showing 2.5G... http://pastebin.com/d4a4093fe 1218635927 Q * larsivi Ping timeout: 480 seconds 1218635987 M * daniel_hozac you've hashified it, right? 1218635993 M * Darkglow yes 1218635998 J * docelic ~docelic@78.134.193.208 1218636001 M * daniel_hozac hashified files have to belong to the host. 1218636069 M * Darkglow I just saw that doing a lsxid -R... some (most) of the files do not belong to guest.. :( Is there a way to use hash and have proper disk usage display? 1218636110 M * daniel_hozac well, no. 1218636128 M * daniel_hozac if hashified files were to belong to a guest, other guests wouldn't be able to access it. 1218636136 M * daniel_hozac which would kind of suck. 1218636161 M * Darkglow ok. I understand. I will add an exclusion to hash for my main data directory... that should do the trick. 1218636965 N * Bertl_zZ Bertl 1218636975 M * Bertl morning folks! 1218637107 M * hijacker afternoon Bertl 1218637313 M * RenOS hello Bertl 1218637565 J * ntrs_ ~ntrs@77.29.76.247 1218637566 Q * ntrs Read error: Connection reset by peer 1218637634 M * alex___ hi 1218638162 Q * john Ping timeout: 480 seconds 1218638188 J * john ~john@glou.nurvnet.org 1218638284 Q * dna_ Quit: Verlassend 1218638342 Q * john 1218638528 J * dowdle ~dowdle@scott.coe.montana.edu 1218639644 J * dna ~dna@107-230-dsl.kielnet.net 1218639813 Q * Darkglow Remote host closed the connection 1218640718 J * dna_ ~dna@60-221-dsl.kielnet.net 1218641014 Q * dna Ping timeout: 480 seconds 1218641034 J * dna ~dna@15-226-dsl.kielnet.net 1218641148 J * cryptronic ~oli@p54A3B525.dip0.t-ipconnect.de 1218641203 J * dna__ ~dna@140-229-dsl.kielnet.net 1218641316 Q * dna_ Ping timeout: 480 seconds 1218641335 Q * renihs Quit: Leaving 1218641524 Q * dna Ping timeout: 480 seconds 1218641937 J * ntrs__ ~ntrs@77.29.72.137 1218641938 J * balbir ~balbir@122.167.240.113 1218642174 J * hparker hparker@linux.homershut.net 1218642376 Q * ntrs_ Ping timeout: 480 seconds 1218642757 J * larsivi ~larsivi@221.80-202-217.nextgentel.com 1218644060 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1218644595 Q * kir Remote host closed the connection 1218645635 J * NetAsh ~arturas@193.219.160.108 1218645652 M * NetAsh hello 1218645712 J * fluor ~fluor@silentio.us 1218645725 M * NetAsh I would like to force 2.3.x vserver to act like 2.2.x for the loopback device 1218645740 J * kir ~kir@swsoft-msk-nat.sw.ru 1218645743 Q * kir Remote host closed the connection 1218645755 M * daniel_hozac why? 1218645763 J * kir ~kir@swsoft-msk-nat.sw.ru 1218645764 Q * kir Remote host closed the connection 1218645788 M * daniel_hozac just disable CONFIG_VSERVER_AUTO_LBACK in the kernel configuration. 1218645802 M * NetAsh portmap tries to bind to * and actualy binds on dummy0 1218645823 M * NetAsh then nfs-user-server tries to find portmap on localhost 1218645843 J * kir ~kir@swsoft-msk-nat.sw.ru 1218645843 Q * kir Remote host closed the connection 1218645855 M * NetAsh I would like to do so per guest not globaly 1218645879 J * kir ~kir@swsoft-msk-nat.sw.ru 1218645879 Q * kir Remote host closed the connection 1218645897 J * kir ~kir@swsoft-msk-nat.sw.ru 1218645928 M * daniel_hozac so, disable hide_lback, lback_remap and enable lback_allow. 1218645960 M * NetAsh in where? 1218645983 M * daniel_hozac nflags. 1218645990 M * NetAsh aha 1218645990 M * daniel_hozac as per http://linux-vserver.org/Capabilities_and_Flags. 1218645994 M * NetAsh thanks 1218646035 M * NetAsh I remembered I saw something somether, but filed to re google this info :D 1218646074 M * Bertl daniel_hozac: any known issues (not missing freatures) with the 2.6.26.2-vs2.3*? 1218646300 P * kir Leaving. 1218646742 M * Bertl off for now ... bbl 1218646746 N * Bertl Bertl_oO 1218646748 M * bragon Bertl: and daniel_hozac i want to test vserver+sparc 1218646757 M * Bertl_oO bragon: go ahead 1218646774 M * bragon do you have any return for this 1218646780 M * Bertl_oO return? 1218646842 M * bragon hum 1218646850 M * bragon any feedback 1218646852 M * bragon sorry :) 1218646931 M * Bertl_oO well, it is supposed to work on sparc, if you encounter any issues, let us know, and we'll fix them ASAP 1218646955 M * Bertl_oO best way to test is to start with building a vanilla (mainline) kernel for your sparc 1218646981 M * Bertl_oO once that works (boots and all), apply the Linux-VServer patch and run 'make oldconfig' 1218646997 M * bragon ok 1218647001 M * Bertl_oO answer the new config questions (use the defaults if unsure) 1218647025 M * Bertl_oO then build and boot the kernel, after that, make sure util-vserver is built with dietlibc 1218647047 M * bragon do you have a sparc machine to do your test in the vserver team ? 1218647047 M * Bertl_oO then run testme.sh (and optionally testfs.sh) and check if everything is fone 1218647058 M * Bertl_oO *fine 1218647073 M * Bertl_oO I have a sparc, but it is currently packed away (so not accessible) 1218647115 M * bragon i can give a sparc to the projet if you want, if you have a devel in Paris (France) 1218647124 M * bragon a Netra T1 1218647138 N * DoberMann DoberMann[PullA] 1218647153 M * bragon i have another question 1218647158 M * bragon i recently test openvz 1218647165 M * bragon it's a good project to 1218647182 M * bragon but i never manage to makes work the ipv6 stack 1218647187 M * Bertl_oO well, it's probably better (for the project) if you test the recent Linux-VServer kernels every now and then, and let us know if something breaks 1218647190 M * bragon with vserver it's roulz 1218647235 M * Bertl_oO I'm not very fond of OVZ, otherwise I wouldn't do Linux-VServer ... but probably I'm biased :) 1218647256 M * bragon but with openvz you can iptables, 1218647278 M * RenOS any ideas if Vserver will support iptables in the container ? 1218647293 M * Bertl_oO with one of the upcoming mainline kernel releases, that will be automatically part of Linux-VServer 1218647304 M * Bertl_oO (i.e. with working network namespaces) 1218647314 M * RenOS that is great! 1218647318 M * Bertl_oO note: the faster ip-isolation will still be an option then 1218647336 M * Bertl_oO i.e. you can mix and match to your likings 1218647381 M * Bertl_oO so, now I'm really off ... bbl 1218647398 M * bragon ok 1218647456 Q * hparker Ping timeout: 480 seconds 1218647863 Q * RenOS Quit: later.... 1218647924 M * NetAsh daniel_hozac: curently portmap in the guest does not beheve (files to bind to eth0 and lo), this leads to nfs-user-server/unfs3 files to find one on loopback, or if I force portmap with -i 127.0.0.1 - nfs is unvisible outside 1218648005 M * NetAsh daniel_hozac: as nfs is unvisible - portmap looses its use, and I found simply disabling portmap, and "unfs3 -p" solves the problem 1218648047 J * hparker hparker@linux.homershut.net 1218648149 M * NetAsh daniel_hozac: I gues somebody will apresiate if my findings are documented (my broken english sucs, so I hope you will do :) 1218648164 M * NetAsh ./ 1218648467 Q * yarihm Quit: Leaving 1218648541 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1218649292 N * pmenier pmenier_off 1218649598 N * Bertl_oO Bertl 1218649600 M * Bertl back again 1218649648 M * Bertl NetAsh: sidenote: portmap works fine inside a guest, if configured properly 1218649922 M * Supaplex are there any plans to allow guests their own routing table? help me understand what's involved. 1218649961 J * _gh_ ~gerrit@67.170.155.50 1218649965 M * Supaplex or is this on the wiki? :) 1218650221 M * Bertl first, with multiple routing tables, you can use up to 251 different routing tables 1218650238 M * Bertl i.e. you can assign one per guest assumed that you have <= 251 guests 1218650268 M * Supaplex okay. is there someway to configure a guest for its own RT? 1218650294 M * Bertl second, once network namespaces in mainline are stable/working, you will be able to do a completely virtualized networking setup for your guests, at the expense of some performance 1218650319 M * Supaplex less of a hit than xen tho. right? 1218650333 M * Bertl no need to configure anything in the guest, you just setup a separate routing table for the guest IPs 1218650351 M * Bertl well, probably the same overhead (network wise) as Xen or Qemu has 1218650369 M * Supaplex that's okay. 1218650403 M * Supaplex soon our networking lab will have a real linux server in the rack :) 1218650428 M * Supaplex just for lab router tests and playground. 1218650485 M * NetAsh Bertl: sugestions for portmap config? i guess there is a need to play with vserver capability flags/interface config 1218650491 M * Supaplex I'm not sure how to setup a routing table for the guest ips. I guess I'll look around when I get to work. 1218650522 M * NetAsh Bertl: but I filed to find a working set of configs to make portmap bind to both: lo and eth0 1218650971 J * docelic_ ~docelic@78.134.198.146 1218651126 M * Bertl NetAsh: the question is, why do you want it to bind to both? 1218651160 M * Bertl but putting that aside, what setup does your guest have? 1218651169 M * NetAsh plain debian 1218651173 M * NetAsh lenny 1218651180 M * Bertl kernel/patch version of the host? 1218651245 M * NetAsh Versions: 1218651245 M * NetAsh Kernel: 2.6.25-2-vserver-amd64 1218651245 M * NetAsh VS-API: 0x00020302 1218651245 M * NetAsh util-vserver: 0.30.215; Aug 1 2008, 13:20:16 1218651287 M * Bertl okay, so you should have a vs2.3 kernel there, what is the guest's network config? 1218651300 M * NetAsh default 1218651326 M * Bertl means? i.e. a single ip assigned or so? 1218651332 M * NetAsh yes 1218651338 M * NetAsh single ip 1218651359 M * Bertl okay, check what /proc/virtnet//* contains (upload to paste.linux-vserver.org) 1218651366 Q * docelic Ping timeout: 480 seconds 1218651496 M * NetAsh one moment 1218651535 M * Bertl take your time 1218651564 M * NetAsh by the way - kernel uploaded 12 Jun, so definetely not the most recent 2.3.x pach :D 1218651595 M * Bertl yes, I expected that 1218651744 M * NetAsh :::::::::::::: 1218651744 M * NetAsh info 1218651744 M * NetAsh :::::::::::::: 1218651744 M * NetAsh ID: 10 1218651744 M * NetAsh Info: ffff810208462a80 1218651746 M * NetAsh Bcast: 255.255.255.255 1218651746 M * NetAsh Lback: 127.0.10.1 1218651748 M * NetAsh 0: [192.168.0.10-0.0.0.0/255.255.255.0:0010] 1218651748 M * NetAsh :::::::::::::: 1218651750 M * NetAsh status 1218651750 M * NetAsh :::::::::::::: 1218651752 M * NetAsh UseCnt: 4 1218651752 M * NetAsh Tasks: 2 1218651754 M * NetAsh Flags: 0000000406000300 1218651754 M * NetAsh NCaps: 0000000000000100 1218651764 M * Bertl did I say flood or did I say pastebin? 1218651794 M * NetAsh I guess I chousen the wrong one :D 1218651847 M * Bertl http://linux-vserver.org/Capabilities_and_Flags 1218651889 M * Bertl you see the Flags? if you compare them to the nflags 1218651907 M * Bertl you see that the sinlge ip special case is on 1218651913 M * NetAsh yes 1218651931 M * Bertl you want to disable that, if you want to have 127.0.0.1 _and_ your ip inside a guest 1218651967 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1218651976 M * NetAsh ok 1218651992 M * Bertl add ~SINGLE_IP to the nflags config file 1218652001 M * NetAsh trying 1218652056 M * Bertl the flags should now read 0000000406000200 1218652141 M * NetAsh Flags: 0000000406000200 1218652145 M * NetAsh yeap 1218652147 M * NetAsh thanks 1218652154 M * Bertl okay, inside the guest, you should now see lo and eth0 1218652159 M * NetAsh now portmap binds to 0.0.0.0 1218652202 M * NetAsh I saw both than and now 1218652213 M * NetAsh just the bind behavior difers 1218652222 M * NetAsh I like the later one more :D 1218652226 M * NetAsh thanks a lot 1218652258 M * Bertl you're welcome! 1218652339 M * NetAsh by the way, would'n it be logical to set nflags 0000000406000200 behavior as a default ? 1218652409 M * Bertl in my opinion, no :) 1218652449 M * NetAsh compatibility with the older realeses or what? 1218652468 M * Bertl no, I think that the single ip special casing is preferable 1218652474 M * Bertl (better performance) 1218652493 M * NetAsh thats a good reason 1218652558 M * NetAsh by the way 1218652566 M * NetAsh I had a cuestion 1218652579 M * NetAsh debian is in frease 1218652605 M * NetAsh and I gues 2.6.25 is the kernel witch will stay 1218652633 M * NetAsh I see most progress (aka bug fixes and so on migrated on 2.6.26) 1218652680 M * NetAsh is there a posibility someone knowledgable will port these to 2.6.25 1218652705 M * Bertl do you mean mainline fixes or Linux-VServer related changes? 1218652750 M * NetAsh lets say debian uses patch-2.6.25.4-vs2.3.0.34.10.diff fo lenny 1218652776 M * NetAsh I guess you are nearing a realease on 2.6.26 1218652852 M * NetAsh is it wery dificult to port patch-2.6.26.2-vs2.3.0.35-pre5.diff (or older ones) to work on 2.6.25 1218652867 M * Bertl we have a few experimental patches for 2.6.26 1218652897 M * Bertl and there should (so far) be no differences between the 2.6.25 and the 2.6.26 patches 1218652905 M * Bertl (given you take the lates one :) 1218652908 M * Bertl *latest 1218652930 M * NetAsh I simply did not tried to apply ones on 2.6.25 1218652970 M * Bertl 2.6.25.11-vs2.3.0.34.14 1218653001 M * Bertl should have the same fix/feature state as the vs2.3.0.35-pre5 1218653036 M * Bertl but IIRC, last time the debian folks complained about a missing 2.6.26 patch for the 'new' stable release 1218653080 M * NetAsh http://packages.qa.debian.org/linux-2.6 1218653092 M * NetAsh somebody uploaded 2.6.26 to unstable 1218653096 M * NetAsh but vserver disabled 1218653113 M * Bertl well, you have to complain to the debian folks for that 1218653119 M * NetAsh na 1218653147 M * Bertl na as in 'you don't really care'? 1218653158 M * NetAsh I simply tried to ask the bugfix/fieture level of diferent patches 1218653162 M * NetAsh but filed to formulate 1218653207 M * NetAsh I am a litle tired, and my english explisitly sux today :D 1218653209 M * Bertl well, you always can use a mainline kernel if Linux-VServer is dropped from debian 1218653218 M * NetAsh i doubt this 1218653246 M * NetAsh folks at debian kernel usualy does good job, just there is a lack of resourses 1218653454 M * trippeh NetAsh: It simply didnt make the first upload in time. There is a pretty big chance it will be back soon. 1218653566 M * NetAsh trippeh: I think so 1218653782 M * Supaplex Bertl: as in "ip route show table 255" right? :) 1218653885 M * Bertl yep, for example, except that 255 is not too good for a guest 1218654152 M * Supaplex right =) 1218654241 Q * larsivi Ping timeout: 480 seconds 1218654862 J * dna_ ~dna@178-232-dsl.kielnet.net 1218654904 M * NetAsh debian uses patch-2.6.25.4-vs2.3.0.34.10.diff 1218654940 M * NetAsh submited a bug report to update to patch-2.6.26.2-vs2.3.0.35-pre5.diff or newer 1218655182 M * NetAsh Bertl: is this 22 month patch usefull? http://svn.debian.org/wsvn/kernel/dists/trunk/linux-2.6/debian/patches/features/all/vserver/bindmount-dev.patch?op=file&rev=0&sc=0 1218655240 M * NetAsh ... 22 month old patch ... 1218655271 Q * dna__ Ping timeout: 480 seconds 1218655289 M * Bertl doesn't show me any code (the url) 1218655311 M * Bertl ah, sec 1218655336 M * Bertl I think that was incorporated by now 1218655396 M * NetAsh If I am not wrong its safe to remove this patch, am I? 1218655457 M * Bertl does it apply? 1218655465 M * NetAsh by the way I am not shure it is enabled :D 1218655483 M * NetAsh simply it poked into my eaa :D 1218655510 M * Bertl hehe, well, let me have a look if that is in 2.6.25/26 then 1218655524 M * NetAsh :D 1218655539 M * pmenier_off Hi all 1218655553 M * Bertl hi pmenier_off! 1218655557 M * pmenier_off Actually testing 2.6.26.2 .. 1218655558 M * micah NetAsh: see #489387 1218655567 M * pmenier_off i got a strange message : http://paste.linux-vserver.org/12357 1218655572 N * pmenier_off pmenier 1218655648 M * Bertl pmenier: interesting .. sec 1218655762 M * Bertl pmenier: you got that from running testme.sh? 1218655780 M * pmenier yes 1218655804 M * pmenier but the vservers runs fine for the moment 1218655812 M * Bertl is that mainline 2.6.26.2 or debian 2.6.26.2 flavor? 1218655833 M * pmenier mainline from kernel.org 1218655853 M * Bertl NetAsh: nah, the patch is not part of Linux-VServer yet, but it should be okay to apply it (if it works) 1218655865 M * NetAsh thanks 1218655877 M * Bertl pmenier: which testme.sh version? 1218655900 M * pmenier oucch.. good question :) just a minut 1218655929 M * Bertl best upload the output somewhere (of the testme.sh) 1218655935 M * pmenier V0.17 1218655969 M * Bertl just realized that I tested with 0.30.214, give me a second to update that 1218655990 M * NetAsh micah: both of us asks for the same solution, just the reasoning is different #495009 1218656017 M * pmenier info ... : the host is running in a vmware ... but i always test with that method before applying to real machines... 1218656046 J * Slydder1 ~chuck@dslb-088-075-197-152.pools.arcor-ip.net 1218656128 M * Bertl that's fine 1218656140 Q * pisco Ping timeout: 480 seconds 1218656304 M * micah NetAsh: great, the more the better 1218656355 M * pmenier Bertl: config used to compil is here : http://www.pmenier.net/vserver/config26262 1218656628 M * pmenier output of testme.sh : http://www.pmenier.net/vserver/testme26262 1218656688 M * Bertl hmm, util-vserver is from debian? 1218656770 M * pmenier no from daniel_hozac 1218656776 M * Bertl okay 1218656914 J * pisco ~pisco@tor.noreply.org 1218656932 M * Bertl while I rebuild the utils for my test environment, could you test if 0.30.214 (debian or not) triggers that too? 1218656966 M * pmenier yes sure 1218657104 M * Bertl daniel_hozac: 0.30.215 complains about a missing TUNSETLINK, I worked around by simply adding that ioctl definition (on an older mandriva) 1218657315 M * pmenier do i stop running vservers before installing 0.30.214 ? 1218657343 M * Bertl doesn't matter, but make sure to uninstall the other utils properly 1218657355 M * pmenier okay 1218657453 M * pmenier it's done and.. same message 1218657485 M * Bertl okay, that's strange and interesting ... because I don't get that with my setup here (and 0.30.214) 1218657493 M * Bertl will test with 0.30.215 in a minute 1218657624 M * Bertl I don't have it with 0.30.215 either ... checking kernel options now 1218657668 M * Bertl ah, you enabled PID namespaces 1218657680 M * Bertl but that shouldn't hurt either, sec 1218657746 M * Bertl probably sysfs (and the deprecated options) 1218657786 M * pmenier i can recompil without these options.. 1218657793 M * Bertl let me check here first 1218657820 Q * NetAsh Quit: Leaving. 1218657857 Q * pisco Remote host closed the connection 1218658226 Q * __gh__ Ping timeout: 480 seconds 1218658228 Q * _gh_ Ping timeout: 480 seconds 1218658307 Q * Slydder1 Quit: Leaving. 1218658529 J * larsivi ~larsivi@221.80-202-217.nextgentel.com 1218658555 M * Bertl hmm, no, still can't recreate it here 1218658616 M * pmenier i recompil now... will reboot in a few minuts 1218658620 J * _gh_ ~gerrit@67.170.155.50 1218658624 M * Bertl nah 1218658634 M * Bertl please run a the addresses of your tracew 1218658640 J * __gh__ ~gerrit@67.170.155.50 1218658644 M * Bertl through addr2line first 1218658677 M * pmenier ??? hmm how do i do that :) 1218658684 M * Bertl addr2line -e vmlinux c01244ea 1218658689 M * Bertl (in your kernel build tree) 1218658717 M * pmenier arghh too late... it was already compiling when you say no... 1218658732 M * Bertl okay, we definitely need that on the next trace 1218658822 M * Bertl I suspect it is trying to create some kind of uid entry somewhere (we had that some time ago) and fails, because it is already there 1218658895 M * Bertl we added an unique identifier for that, to work around it 1218658967 M * pmenier ok reboot done and compil without sysfs and PID namespace 1218659017 M * pmenier same message 1218659027 M * Bertl okay, with trace? 1218659036 M * Bertl if so, please run through addr2line for me 1218659037 M * pmenier result of : addr2line -e vmlinux c01244ea => user.c:0 1218659058 M * Bertl take the addresses from your new trace :) 1218659068 M * pmenier oops sorry... 1218659092 M * Bertl please upload the trace and the mapped addresses 1218659111 M * Bertl I noticed you are emulating x86 ... will check now if that makes a difference 1218659280 M * pmenier http://paste.linux-vserver.org/12358 1218659328 Q * bonbons Quit: Leaving 1218659493 J * FireEgl FireEgl@adsl-17-147-88.bhm.bellsouth.net 1218659503 M * Bertl what does your /sys/kernel/uids/ contain? 1218659606 J * pisco ~pisco@tor.noreply.org 1218659676 M * pmenier http://paste.linux-vserver.org/12359 1218659855 M * Bertl aha, on x86 I can recreate it 1218659883 M * pmenier so.. it would be an option to disable... 1218659909 M * Bertl it's harmless btw, you should be able to block it by disabling sysfs completely 1218659926 M * Bertl but I'll look into it and will fix it ASAP 1218659959 M * Bertl what does your /proc/virtual contain atm? 1218659992 M * pmenier okay thanks a lot... but don't care for me... it was just for testing and hope to help... 1218659995 M * Bertl note that this is a mainline bug, btw, not a Linux-VServer one :) 1218660033 M * Bertl i.e. user name spaces and sysfs are not friends :) 1218660062 M * Bertl but let me try a slightly different workaround 1218660079 M * pmenier with one vserver running : http://paste.linux-vserver.org/12361 1218660113 M * Bertl as I expected, the root uid is not removed from sysfs inside a context 1218660839 Q * cryptronic Quit: Leaving. 1218660874 N * DoberMann[PullA] DoberMann 1218660993 M * Bertl hmm ... seems that is more serious than expected ... and most likely a Linux-VServer issue ... context structures keep hanging around, investigating right now 1218661054 M * pmenier ah okay... 1218661120 M * Bertl hmm, no, they disappear properly as it seems 1218661162 M * daniel_hozac Bertl: no known issues as far as i know. 1218661205 M * Bertl we have one now :) 1218661235 M * Bertl actually it seems like there are two issues 1218661268 M * Bertl one is that the fs structures grow constantly (could be an accounting issue) 1218661326 M * Bertl the second one is the user space not going away properly 1218661327 M * Bertl #NSProxy: 8 507 0 0 0 8 0 1218661362 M * Bertl that is after running testme.sh twice, but this could be related to inproper cleanup in the sysfs error case 1218661403 M * Bertl have to check that, in general, the question is if we should create the sysfs uid entries at all for user spaces != init space? 1218661408 Q * ntrs__ Ping timeout: 480 seconds 1218661675 M * daniel_hozac i think so. 1218661690 M * daniel_hozac we might just want to grab the userns sysfs patches though. 1218661894 M * Bertl do you have an url for them? 1218662267 M * Bertl okay, the user ns leakage goes away when I avoid the sysfs error 1218662284 M * Bertl this means, that we do not handle that specific error properly 1218662308 M * daniel_hozac http://lkml.org/lkml/2008/6/18/233 1218662315 M * daniel_hozac is, i think, the latest set. 1218662369 M * Bertl sorry, is there something attached somewhere or am I missing the obvious? 1218662385 M * Bertl ah, the later one, nevermind 1218662392 M * daniel_hozac it's a set of 11 patches, they're all replies to that one. 1218662400 M * Bertl yeah, just realized that 1218662416 M * Bertl hmm, looks quite intrusive 1218662429 M * daniel_hozac it is. 1218662446 M * Bertl I guess we wait for mainline to incorporate it first, or are there any good reasons to do that now? 1218662457 M * daniel_hozac i think it's already in 2.6.27, 1218662640 J * ktwilight ~ktwilight@225.99-66-87.adsl-dyn.isp.belgacom.be 1218662712 J * Aiken ~Aiken@ppp118-208-53-42.lns4.bne1.internode.on.net 1218662910 Q * ktwilight_ Ping timeout: 480 seconds 1218663455 M * Bertl okay, the fs issue is an accounting bug I just fixed 1218663478 M * Bertl pmenier: expect a fix/updated version in a minute 1218663502 M * pmenier okay :) 1218663799 M * Bertl okay, either you apply the two deltas, or you use the vs2.3.0.35 diff 1218663841 M * pmenier okay but in the same i did a new compil with differents options than the previous and i didn't get any more errors messages ??? 1218663850 J * yarihm ~yarihm@84-74-147-84.dclient.hispeed.ch 1218663870 M * pmenier the diff : http://paste.linux-vserver.org/12362 1218663880 M * Bertl yep, that's expected, you can now revert to your old options (if you like) 1218663903 M * pmenier i try with new patch now... 1218663925 M * Bertl thanks for spotting this! 1218664911 N * DoberMann DoberMann[ZZZzzz] 1218664943 J * dna__ ~dna@178-232-dsl.kielnet.net 1218664953 Q * pisco Remote host closed the connection 1218665059 M * pmenier compil done, reboot , vserver started : all is right 1218665139 M * Bertl excellent! thanks again! 1218665183 M * pmenier no problem :) ... I must leave now: have a good night 1218665189 M * Bertl you too! 1218665204 N * pmenier pmenier_off 1218665321 Q * dna_ Ping timeout: 480 seconds 1218667023 Q * yarihm Quit: Leaving 1218667091 M * Bertl daniel_hozac: btw, how does util-vserver cope with --bind mount + remount combos in fstab? 1218667114 J * pisco ~pisco@tor.noreply.org 1218667119 M * daniel_hozac it doesn't, yet. 1218667121 M * Bertl i.e. does it work to specify two lines (one for the mount, and a second one for the remount) 1218667129 M * daniel_hozac well, yeah. 1218667154 M * Bertl okay, I think that is enough, as mainline prefers this syntax, obviously 1218667162 M * daniel_hozac hehe 1218667190 M * Bertl I wonder how 'secure' ro bind mounts look like, they probably need 3 mount lines 1218667213 M * Bertl two to get the ro behaviour and a third one to move the mount 1218667268 M * Bertl (otherwise you have a small window when the mount is rw) 1218667664 Q * dna__ Quit: Verlassend 1218667929 Q * docelic_ Quit: http://www.spinlocksolutions.com/ 1218668411 M * Bertl okay, off to bed now ... have a good one everyone! 1218668417 N * Bertl Bertl_zZ 1218669806 Q * infowolfe Ping timeout: 480 seconds 1218671036 J * infowolfe ~infowolfe@c-67-160-149-42.hsd1.or.comcast.net 1218671362 Q * loddafnir Read error: Connection reset by peer