1124324652 J * tchan_ ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1124324711 Q * tchan Read error: Connection reset by peer 1124325450 Q * stephenM Quit: 1124325636 Q * tchan_ Ping timeout: 480 seconds 1124325796 P * litage Leaving 1124325824 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net 1124326222 M * Bertl okay, I'm off to bed now ... have a nice whatever everyone ... cya tomorrow ... 1124326251 N * Bertl Bertl_zZ 1124327428 N * micah__ micah 1124327562 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1124328073 Q * obi Ping timeout: 480 seconds 1124329536 J * Nero ~nero@i154-035.nv.iinet.net.au 1124329827 Q * Tank Quit: Leaving 1124331879 M * daniel_hozac Bertl_zZ: it seems to be related to your fix01 patch. 1124331905 M * daniel_hozac Bertl_zZ: if i reverse it, make clean, make, i get a working showattr. 1124331916 M * daniel_hozac Bertl_zZ: apply it again, make clean, make, a broken showattr. 1124333811 Q * VooDooMaster Ping timeout: 480 seconds 1124340206 J * Hunger Hunger.hu@Hunger.hu 1124343453 Q * Nero Quit: A yawn is a silent shout. 1124346409 P * stefani parting (is such sweet sorrow) 1124346433 Q * eugenesan Ping timeout: 480 seconds 1124348657 J * monrad ~monrad@213083190134.sonofon.dk 1124349601 Q * Hollow Ping timeout: 480 seconds 1124349730 J * saddi ~sven@GKMCCXVII.dsl.saunalahti.fi 1124350565 J * Hollow ~Hollow@home.xnull.de 1124351126 Q * Doener Ping timeout: 480 seconds 1124351171 J * Doener ~doener@p548763AF.dip.t-dialin.net 1124351187 M * wibble morning all 1124351359 J * erwan_taf ~erwan@81.80.43.77 1124351395 J * jsambrook ~jsambrook@host-62-69-64-93.bsve.net 1124351461 P * jsambrook 1124351620 M * saddi morning 1124351805 M * saddi I am a newbie to vserver and have one problem 1124351830 M * saddi whenever I start a virtual machine, Ill lose keyboard control to host 1124351846 M * saddi I hope someone of you could help me with this problem 1124351852 M * wibble hmm 1124351864 M * saddi I am using Arch Linux 1124351868 M * wibble i am fairly new so please forgive. I have never come across that before 1124351875 M * saddi ok 1124351881 M * wibble what happens if you are remotely logged in to the host via ssh ? 1124351901 A * Pazzo is away: busy 1124351902 A * Pazzo is back (gone 00:00:02) 1124351918 M * saddi actually I am at test phase now, I just try to get it work 1124351930 M * saddi I am using vserver modded kernel on laptop 1124351953 M * Hollow saddi: remove /dev/console in the guest 1124351956 M * saddi I followed the directions here: http://linux-vserver.org/ArchVserver 1124351970 M * saddi ok,Ill see how it goes 1124352003 M * Hollow Bertl_zZ: is terminal virtualization planned in the near future btw.? 1124352030 M * saddi btw I made the /vservers as a soft link to another dir 1124352034 M * saddi is this ok? 1124352060 M * Hollow should work, but you could set the /etc/vservers/.default/vdirbase symlink instead 1124352073 M * Hollow which should point to /vservers atm 1124352078 J * Aiken ~james@tooax6-229.dialup.optusnet.com.au 1124352084 Q * erwan_taf Quit: Leaving 1124352121 M * saddi ok 1124352128 M * saddi now I removed /dev/console 1124352136 M * saddi lets sew now... 1124352206 M * saddi theres always another problem, too 1124352215 M * saddi on "vserver test start" 1124352220 M * saddi I'll get this: The configured vshelper '/sbin/vshelper' does not match the 'vshelper' 1124352220 M * saddi script of the util-vserver package 1124352255 M * saddi 'echo "/usr/lib/util-vserver/vshelper" >/proc/sys/kernel/vshelper' seems to work, though 1124352258 M * Hollow echo 'kernel.vshelper = /usr/lib/util-vserver/vshelper' >> /etc/sysctl.conf && sysctl -p 1124352305 M * saddi then I always must start vprocunhide 1124352323 M * Hollow always, i.e. after reboot? 1124352331 M * saddi or I'll get this error: /proc/uptime can not be accessed 1124352337 M * saddi yes 1124352345 M * saddi I guess I should add it to daemons 1124352358 M * Hollow add it to your runlevel.. yep 1124352457 M * saddi yeah :) 1124352460 M * saddi now it started 1124352469 M * saddi and I still got keyboard 1124352474 M * Hollow :) 1124352492 M * saddi thanks, Hollow :) 1124352497 M * Hollow your welcoe 1124352499 M * Hollow m 1124353768 M * wibble dpkg: error processing /var/cache/apt/archives/bind9-host_1%3a9.2.4-1_i386.deb (--unpack): 1124353771 M * wibble unable to create /var/lib/dpkg/updates/tmp.i: No space left on device 1124353772 M * wibble hmm 1124353775 M * wibble E: Sub-process /usr/bin/dpkg received a segmentation fault. 1124353779 M * wibble running on ext3 with vdlimit 1124354005 M * wibble anyone got the 0.30.208 utils in .deb format ? 1124354007 M * wibble for sarge... 1124354291 Q * saddi Ping timeout: 480 seconds 1124354525 J * saddi ~sven@GKMCCXVII.dsl.saunalahti.fi 1124354525 Q * saddi Quit: 1124355742 Q * nokoya Ping timeout: 480 seconds 1124356056 J * martin ~multiplex@pc70-c512.uibk.ac.at 1124356440 J * Aiken_ ~james@tooax8-115.dialup.optusnet.com.au 1124356562 J * Neubix ~brian@p54B03D17.dip.t-dialin.net 1124356574 M * Neubix morning 1124356768 Q * Aiken Ping timeout: 480 seconds 1124356952 J * nokoya young@hi-230-82.tm.net.org.my 1124358371 Q * martin Remote host closed the connection 1124358824 T * * http://linux-vserver.org/ | latest stable 2.0, 1.2.10, devel 2.1.0-pre1 -- 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 ;) 1124358824 T * Bertl - 1124360555 M * wibble hmm 1124360565 M * wibble I think ext3 & vdlimits has an issue 1124362953 Q * Aiken_ Ping timeout: 480 seconds 1124365651 J * hansi33 ~hansi33@193.80.108.229 1124365751 Q * nox arion.oftc.net neutron.oftc.net 1124365751 Q * SNy arion.oftc.net neutron.oftc.net 1124365905 J * SNy db59d9d6e6@bmx-chemnitz.de 1124366043 J * _mcp ~hightower@wolk-project.de 1124366044 Q * mcp Read error: Connection reset by peer 1124366346 J * mef ~mef@pcp09895218pcs.ewndsr01.nj.comcast.net 1124366655 Q * mef Quit: using sirc version 2.211+KSIRC/1.3.10 1124366888 N * Bertl_zZ Bertl 1124366893 M * Bertl morning folks! 1124366938 M * hansi33 hi bertl 1124366978 M * Bertl hey hansi33! how are you? 1124366987 M * hansi33 do you have a fast hint why my vserver does not start, i am just reading the output of --debug 1124367020 M * hansi33 i am not so well, because the trouble is on a semi-productive server 1124367020 M * Bertl hmm, maybe you could upload that one somewhere, so I could have a peek too? 1124367041 M * Bertl (e.g. pastebin.com) 1124367044 M * hansi33 one moment. 1124367210 M * hansi33 its on http://brunner.at/aus1 1124367244 M * wibble morning Bertl!! 1124367252 M * wibble sorry about yesterday. wireless went down :( 1124367339 M * Bertl hansi33: and could you also upload the output of testme.sh? 1124367368 M * wibble brb 1124367371 M * Bertl wibble: np 1124367400 M * hansi33 Bertl, which testme.sh? 1124367431 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/ 1124367515 M * Bertl what distro is the guest? 1124367526 M * hansi33 it 1124367534 M * hansi33 its in http://brunner.at/aus2 1124367544 M * hansi33 gentoo 1124367564 M * Bertl hmm, yeah, kindof expected that ... 1124367596 M * Bertl daniel_hozac: you around? 1124367678 M * Neubix Hi Bertl 1124367690 M * Bertl hey Neubix! 1124367727 M * Bertl hansi33: I'm no gentoo person, but the issue is the following: 1124367753 M * Bertl the runlevel services write some files telling the system that the service is started/running 1124367771 M * Bertl when those files are present, the service will not start 1124367809 M * Bertl you have to remove those files (somewhere in /var/run or /var/log or whatever) 1124367829 M * hansi33 Bertl: thanks, that was the hint! 1124367843 M * hansi33 i did "rm -rf /vservers/serv7a12/var/lib/init.d/*" 1124367843 M * Bertl okay, you know where to find them? 1124367849 M * Bertl perfect! 1124367874 M * Bertl hansi33: and you should schedule an update to vs2.0 ... 1124367892 M * Bertl (and also update the tools to 0.30.208) 1124367893 M * hansi33 Bertl: i had this line in my old init-scripts, but this has gone, thanks again 1124367906 M * Bertl you're welcome! 1124368125 M * Neubix Bertl, with the old version and the debian-newserver.sh after installation I jump into the new guest and must set password, configure locales and apt. with the vserver build command I get only I: Base system installed successfully. is this right? 1124368222 M * Bertl well, I don't know the debian-newserver.sh, with util-vserver you get an install process (depending on the method) 1124368247 M * Bertl after the installation finishes successfully, you can start the guest 1124368270 M * Neubix yes .. all right .. but nothing is configured .. 1124368311 M * Neubix normany (with cd installation) after reboot the config is running .. root asswor etc .. 1124368538 J * CosmicRay ~jgoerzen@2002:4463:7269:1:20e:a6ff:fe66:c5a3 1124368592 M * CosmicRay hi all. I've had a number of issues with networking after upgrading to vs2.0. The one I can't fix is postfix. It now says "mail loops back to myself" when trying to send to another vserver on the same host. I *suspect*, but cannot prove, that this is because /sbin/ifconfig now lists all aliases on the host, whereas in vs1, it listed only the one that the config showed it as binding to 1124368599 M * CosmicRay any suggestions on how to fix that? 1124368704 M * Bertl Neubix: yes, you can either run such stuff manually, or add some config scripts to the postinstall scripts 1124368750 M * Bertl CosmicRay: your choice of words suggests that you had the very same guest running on pre vs2.0 kernel? 1124368765 M * CosmicRay correct 1124368786 M * CosmicRay I'm using the same (legacy) config now, as well 1124368798 M * CosmicRay previously I was running 2.4.26-vs1.27 1124368817 M * CosmicRay now at 2.6.12.4-vs2.0 1124368830 M * Bertl could you upload the legacy config file somewhere? 1124368862 M * CosmicRay sure 1124368870 M * Bertl do you use a nameresolver which might have changed since the update? 1124368896 M * Bertl and finally, could you also upload the output of the testme.sh script? 1124368905 M * CosmicRay http://www.complete.org/~jgoerzen/webmail_external.conf 1124368924 M * CosmicRay no change to the DNS, and anyway I'm routing directly to a smarthost given by IP 1124368955 M * CosmicRay btw I have util-vserver 0.30.208 with fix1 applied 1124368968 M * CosmicRay where can I find this testme.sh? 1124368972 M * Bertl okay, and you did not change any of the config entries (with the update) 1124368977 M * CosmicRay correct 1124368988 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/ 1124368998 M * CosmicRay I have had odd behavior with the network interfaces on every virtual host. 1124369017 M * CosmicRay it seems that binding to a port on 127.0.0.1 also meant binding to the same port on the server's bound IP address 1124369027 M * CosmicRay that was fixible by tightening up the interfaces that things bind to, but odd. 1124369044 M * CosmicRay also I noticed the same ifconfig behavior on other vservers (listing every alias on the system, where previously it didn't) 1124369051 M * Bertl well, yes and no, actually binding to 127.0.0.1 is rewritten to binding to the first ip 1124369066 M * CosmicRay that makes sense (all of these have only one IP) 1124369082 M * Bertl but this should be the case with the 2.4 kernel too 1124369100 M * CosmicRay even in 1.27? 1124369111 M * Bertl for the ifconfig shows stuff, that is because the legacy config doesn't set the hide_netif flag 1124369129 M * Bertl CosmicRay: I'm going to check that right now ... (the 1.27) 1124369150 M * CosmicRay all tests succeeded from testme.sh 1124369168 M * Bertl good could you upload the first 4 lines (or paste them here)? 1124369180 M * CosmicRay Linux-VServer Test [V0.13] Copyright (C) 2003-2005 H.Poetzl 1124369180 M * CosmicRay chcontext is working. 1124369180 M * CosmicRay chbind is working. 1124369180 M * CosmicRay Linux 2.6.12.4-vs2.0 i686/0.30.208/0.30.208 [Ea] (0) 1124369195 M * Bertl hmm, okay, another one please :) 1124369202 M * CosmicRay VCI: 0002:0001 273 00000026 1124369205 M * Bertl tx 1124369208 M * CosmicRay np 1124369260 M * CosmicRay I was fairly certain that loopback was (correctly) distinct in 1.27 1124369282 M * wibble right, im back 1124369285 M * CosmicRay in any case, I didn't get an error when a server tried to bind to, say, port 25 both there and on its eth0 alias 1124369290 M * Bertl wibble: wb! 1124369308 M * CosmicRay however, now that I think about it, it's possible that one of the problems was because it was trying to bind to every interface it saw on the system 1124369308 M * wibble thanks! 1124369319 M * Bertl CosmicRay: hmm, that sounds interesting ... 1124369322 M * wibble I appear to have an issue with ext3 and vdlimits now 1124369336 M * Bertl wibble: did you apply both patches? 1124369344 M * wibble nope 1124369352 M * wibble haven't seen where these patches are? 1124369361 M * CosmicRay Bertl: that would have been dhcpd, which was throwing out weird errors. however I also had issues with exim, bind, and a few other processes, which were readily fixed by clamping down the interface to which they bound 1124369374 M * Bertl wibble: http://vserver.13thfloor.at/Experimental/FOR-2.0.1/ 1124369409 M * CosmicRay so if I add hide_netif to S_FLAGS in my legacy config, I'd get the old behavior? 1124369420 M * wibble Bertl: three patches..? 1124369432 M * Bertl CosmicRay: well, no, because the legacy tools do not understand that flag 1124369459 M * Bertl CosmicRay: but in a perfect world, the legacy tools would set it automagically 1124369473 M * Bertl wibble: the dlimit patches are of interest to you, but you can add the vroot too 1124369491 M * wibble presuming i just add to my kernel source with the standard vs2.0 ? 1124369498 M * Bertl yep 1124369513 M * Bertl CosmicRay: but you can for a test add the flag with vattribute 1124369517 M * CosmicRay Bertl: is there a doc somewhere that describes exactly how to convert a legacy config to a modern one? and can I convert a few vservers at a time? (so part of my installation is legacy, part is modern) 1124369543 M * Bertl CosmicRay: yes, legacy and modern can coexist perfectly 1124369554 M * CosmicRay hmm, well I do have that utility but I have no clue how to run it ;-) 1124369572 M * CosmicRay I'm almost wondering if postfix is seeing the destination IP in the list of interfaces 1124369591 M * Bertl vattribute --set --xid --flag hide_netif 1124369593 M * CosmicRay personally I think that servers that enumerate the host's interfaces like that are broken, but oh well. 1124369601 M * CosmicRay what's the xid? 1124369602 M * Bertl replace with your context id ... 1124369628 M * Bertl (the one you didn't specify with S_CONTEXT= :) 1124369636 M * wibble what does hide_netif do? 1124369637 M * CosmicRay but I think I have it in /var/run/vservers 1124369637 M * CosmicRay heh 1124369666 M * Bertl wibble: hide network interfaces which do not carry IPs assigned to the guest 1124369673 M * wibble nice 1124369676 M * wibble I noticed that in vs2.0 1124369677 M * CosmicRay Bertl: that fixed it. 1124369752 M * wibble Bertl: what kernel source did you apply this to? 1124369789 M * Bertl 2.6.12.3-vs2.0 but it should work with others too 1124369800 M * Bertl (just keep the correct order) 1124369894 M * wibble hmm 1124369909 M * Bertl CosmicRay: from 2.4.26-vs1.27: 1124369915 M * Bertl + if (s_addr == 0) { 1124369915 M * Bertl + s_addr = ipv4root; 1124369921 M * Bertl + } else if (s_addr == 0x0100007f) { 1124369921 M * Bertl + s_addr = s_addr1 = ipv4root; 1124369929 M * Bertl + } else if (s_addr != v4_bcast) { 1124369945 M * Bertl so if you specify IP_ADDR_ANY to a bind 1124369951 M * CosmicRay is 0x0100007f localhost? 1124369965 M * Bertl it will bind the first ip address, and similar for IP_ADDR_LOCAL 1124369974 M * CosmicRay and uhm, shouldn't it be 0x7f000001? 1124369974 M * Bertl yep, it is ... 1124369997 M * CosmicRay at the very least, that will break on other archs ;-) 1124370019 M * CosmicRay I guess it depends on whether s_addr is in host or network byte order 1124370093 M * wibble Bertl: building kernel now 1124370125 M * CosmicRay Bertl: anyway, what is the preferred way to set this flag with my configuration? 1124370173 M * Bertl probably the best hoice is to set it in the startup script (after you fixed the context id) 1124370187 N * _mcp mcp 1124370192 N * mcp _mcp 1124370219 M * Bertl CosmicRay: you have a point with the byteorder, but IIRC that is always network byteorder .. but I will investigate this ... 1124370249 M * Bertl (anyway, it didn't change between 1.2 and 2.0, at least not in a way I could see) 1124370253 Q * hansi33 Quit: ChatZilla 0.9.61 [Mozilla rv:1.7.10/20050811] 1124370255 M * CosmicRay Bertl: I wonder, could this have unintentionally not bound to ipv4root all along? 1124370256 M * CosmicRay ah. 1124370392 M * wibble so where is Enrico? 1124370423 M * Bertl I guess somewhere in Germany, probably chemnitz ... 1124370460 M * wibble lol 1124370468 M * wibble I mean, how do I contact him! 1124370473 M * wibble need that vlan fix too hehe 1124370481 M * wibble kernel built, let me install and reboot 1124370490 M * Bertl it seems he is only communicating via email nowadays ... 1124370509 M * Bertl back then, when we started, he was a regular visitor on the channel too 1124370530 M * wibble right rebooting 1124370539 M * wibble so he looks after the utils, and you do the kernel hacking? 1124370556 M * Bertl CosmicRay: do you have a system to test with? 1124370581 M * Bertl (i.e. kernel install/reboot/etc) 1124370837 M * Doener good afternoon! 1124370848 M * Bertl afternoon Doener! 1124370933 M * Doener Bertl: i just saw the network code bit and the on-the-fly ip address stuff came to mind... did you spend any thoughts on whether the 0.0.0.0 to real address conversion is necessary/a big gain in terms of performance? 1124370950 M * Doener (the conversion happens when a vserver with a single ip address binds to 0.0.0.0) 1124370959 M * CosmicRay Bertl: no 1124370969 M * CosmicRay Bertl: well, I could probably set one up, I suppose 1124370980 M * Doener cause that's (API-wise) the only thing that prevents 'full' on-the-fly adding ip address 1124370994 M * CosmicRay this particular machine is in production use, so reboots there are discouraged and usually have to be scheduled in advance 1124371007 M * Doener (works fine now when the vserver already has 2+ addresses, otherwise services need to be restarted) 1124371014 M * Bertl Doener: yes, thought so ... we probably need a few tests, but I guess we can 'drop' that feature if we see a real benefit ... 1124371068 M * Bertl CosmicRay: well, some test setup would be required if you want to track this down ... 1124371082 M * CosmicRay you think it's more than a simple matter of adding that flag? 1124371103 M * Bertl hmm, I'm talking about the changes in network behaviour ... 1124371107 M * CosmicRay Bertl: I'll be happy to do that, but I leave for a few days' vacation tomorrow. 1124371108 M * CosmicRay ah 1124371114 M * Bertl (which might or might not be correct :) 1124371125 M * CosmicRay Bertl: do you think you could drop an e-mail to jgoerzen@complete.org describing what you want me to test, and I could take a look at it next week? 1124371153 M * Bertl sure, I'd say just come back here when you return from your vacation ... no? 1124371166 M * CosmicRay sure, I can just do that 1124371169 A * CosmicRay makes a note 1124371199 M * Bertl anyway, thanks for giving feedback, we could use more of that! 1124371226 M * CosmicRay Bertl: here's some more. vserver is great, and I'm a huge fan. I have no idea why so many people use xen or uml when vserver is here. 1124371232 M * CosmicRay so thanks for all your work on it. 1124371241 M * Bertl you're welcome! 1124371255 M * CosmicRay I use it at work for our internal file/mail server, and at home on my (600MHz Alpha) firewall/router/server 1124371270 M * CosmicRay I am pleased that it supports Alpha ;-) 1124371289 M * Bertl hehe, a lot of folks actually are ... 1124371296 M * wibble well thats a start Linux backup 2.6.12.5-vs2.0-fix01-p4-raid-3ware-e100x #1 Thu Aug 18 13:04:42 BST 2005 i686 GNU/Linux 1124371312 M * CosmicRay really, interesting. I don't run into a whole bunch of alpha users anymore these days 1124371325 M * CosmicRay wibble: that is an evil uname you have there 1124371337 M * CosmicRay just thought you should know. 1124371343 M * wibble lol 1124371353 M * wibble well it tells me everything about that kernel! 1124371357 M * wibble i.e. is vs2.0 with fix01 1124371368 M * Bertl wibble: why didn't you use dlimit-feat01-fix01 instead of just fix01? 1124371372 M * wibble its a p4, with software raid, 3ware modules and intel drivers! 1124371389 M * CosmicRay wibble: uhm, aren't those features built into the kernel? ;-) 1124371399 M * wibble Bertl: couldn't be botherd to type any further lol 1124371411 M * wibble CosmicRay: i remove everything bar what I need! 1124371423 M * Bertl CosmicRay: sure, but not everybody stores the .config in /proc :) 1124371451 M * CosmicRay Bertl: lamers ;-) 1124371475 M * CosmicRay Bertl: it beats storing .config in the uname ;-) 1124371476 M * Bertl btw, no offence wibble, I really like the ideas folks come up with ... 1124371494 M * wibble heh 1124371506 M * CosmicRay yeah, me neither, just kidding 1124371534 M * wibble well its just that I store all my kernel images on my mirror and apt-get :) 1124371547 M * wibble since I have a number of customers with standard dells etc 1124371557 M * Bertl makes perfect sense ... 1124371589 M * wibble right, building test13 again on reiser and lets see what happens 1124371593 M * CosmicRay perhaps you ought to do 2.6.12.5-vs2.0-fix01-pentium4-2gb-ide-scsi-adaptec-usb-audigy-raid-lvm-3ware-e100x-netfilter-ipv4-ipv6-ext2-ext3-reiser4-iso9660-iso8859-1-proc-multimedia-initrd 1124371596 M * CosmicRay wibble: ah ha! 1124371607 M * CosmicRay now see, I use that sort of naming scheme for the *debian* revision of my .debs 1124371613 M * CosmicRay so I can just grab one and install it 1124371618 M * CosmicRay but it doesn't go in the uname of the installed kernel 1124371623 M * CosmicRay (--revision= parameter to make-kpkg) 1124371726 M * CosmicRay I mean, what's the point of building a kernel with e100x but no ipv4? ;-) 1124371957 M * Bertl Doener: do you have a minute for me? 1124371971 M * Doener yeah, i'm annoyed of math atm anyways ;) 1124371999 M * wibble lol 1124372003 A * wibble slaps CosmicRay 1124372015 M * wibble you missed smp too! 1124372022 M * Bertl CosmicRay brought up the issue that the hard coded loopback address might not work on other archs ... 1124372041 M * CosmicRay wibble: heh 1124372056 M * Bertl Doener: and I agree, while it is probably the correct byteorder for most archs ... we might have an issue there ... 1124372081 M * CosmicRay for most? wouldn't it be wrong everywhere except x86 and amd64? 1124372090 M * Bertl Doener: now I'm inclined to change that to htonl(INADDR_LOOPBACK); 1124372092 M * CosmicRay or, correct everywhere except x86 and amd64, depending? ;-) 1124372169 M * Bertl CosmicRay: well, IIRC it's not that easy to say .. because a lot of structures are kept in network byteorder, where others are in host byteorder ... 1124372216 M * wibble right 1124372228 M * wibble time to test 1124372319 M * wibble dpkg: error processing /var/cache/apt/archives/libsasl2_2.1.19-1.5_i386.deb (--unpack): 1124372322 M * wibble unable to create `./usr/share/doc/libsasl2/macosx.html': No space left on device 1124372325 M * wibble hmm 1124372358 M * wibble houston, we have a problem 1124372371 M * Bertl well, probably you are hitting the limit, no? 1124372376 M * wibble nope! 1124372388 M * wibble allocated 4.7GB 1124372412 M * wibble and I cant even enter the vserver guest 1124372416 M * Bertl vdlimit --xid shows? 1124372429 M * Bertl vdlimit --xid /path/to/guest shows? 1124372450 M * wibble backup:~# vdlimit --xid 102 /var/lib/vservers/test13 1124372450 M * wibble 102 /var/lib/vservers/test13 1124372450 M * wibble space_used=46070 1124372450 M * wibble space_total=5000000 1124372450 M * wibble inodes_used=100 1124372453 M * wibble inodes_total=100 1124372455 M * wibble reserved=5 1124372461 M * Bertl see, out of inodes :) 1124372466 M * wibble doh 1124372470 A * wibble fixes 1124372525 M * wibble altough vserver test13 enter hangs 1124372530 M * wibble I should really stop using "test13" 1124372531 M * wibble lol 1124372578 M * wibble ok well I think we might have some stablility issues 1124372590 M * wibble vserver test13 stop hangs, vserver test13 enter hangs 1124372597 M * wibble cant ctrl-c out of it 1124372601 M * Bertl check for kernel dumps ... 1124372614 M * wibble where would they be? :) 1124372629 M * Bertl dmesg -s 1000000 1124372656 M * wibble ReiserFS: dm-1: warning: PAP-5660: reiserfs_do_truncate: wrong result -1 of search for [1 24589 0xffffffff DIRECT] 1124372659 M * wibble ReiserFS: dm-1: warning: clm-2100: nesting info a different FS 1124372668 M * wibble not the last message though 1124372671 M * wibble nothing obvious 1124372721 M * wibble root 11745 0.0 0.1 3308 1872 pts/3 S+ 13:41 0:00 | \_ /bin/bash /usr/local/sbin/vserver test13 stop 1124372725 M * wibble root 11753 0.0 0.0 100 20 pts/3 S+ 13:41 0:00 | \_ /usr/local/lib/util-vserver/lockfile /var/lock/vserver.etcvserverstest13.startup /tmp/vserver-lock.Xydkpo 1124372805 M * wibble i see the vwait process, and after 30 secs it disappears 1124372814 M * wibble but the vserver is still running 1124372816 M * wibble how odd 1124372830 M * Bertl strange .. maybe some stopped process inside? 1124372869 M * wibble 11778 102 test13 pts/3 D 0:00 /usr/local/sbin/vcontext --silent --migrate --chroot --xid 102 -- /etc/init.d/rc 6 1124372929 M * Bertl hmm, this doesn't look too nice ... 1124372944 M * Bertl it's with reiserfs? 1124372946 M * wibble indeed. do you want a login? 1124372947 M * wibble yeah 1124372956 M * Bertl do you have magic sysreq enabled? 1124372985 M * wibble dont even know what it is! 1124373008 M * Bertl okay, is there a sysrq-trigger file in /proc (on the host)? 1124373051 M * wibble yes there is 1124373059 M * Bertl and please make some notes what you did to trigger this :) 1124373080 M * wibble lol, by specifying 100 inodes and then installing something :) 1124373085 M * Bertl okay, then try 'echo t >/proc/sysrq-trigger' 1124373107 M * Bertl should add a stack trace for each process 1124373115 M * Bertl (to the dmesg output) 1124373256 M * wibble http://www.wibblewobb1e.co.uk/vserver/process-dump 1124373449 M * Bertl hmm, is your log buffer too small or didn't you use the -s 100000 option? 1124373472 M * wibble oops! 1124373473 M * wibble sorry 1124373501 M * Bertl np, you hit the inode limit, right? 1124373549 M * wibble hmm 1124373553 M * wibble cant get it any bigger 1124373569 M * wibble yes 1124373586 M * wibble i tried to install samba (rather big package) and its blown up on dpkg --unpack 1124373612 M * Bertl but we are not 100% sure it was reiser, no? because you are using ext3 too, right? 1124373637 M * wibble which was when I reached the inode limit. I did this earlier (as I used the same set command) on ext3 and it survived 1124373652 M * wibble I can try it with ext3 again if you want? 1124373670 M * Bertl hmm, let's try to reproduce it with reiser first ... 1124373691 M * wibble ok 1124373707 M * wibble I will reboot this box and start again - is ther anything you want me to set before I try to reproduce? 1124373738 M * Bertl you could increase the message ring buffer in the ekrnel, but that would require a recompile 1124373752 M * wibble well whatever you need! 1124373757 M * Bertl it would also be beneficial to enable the vserver debugging options when you do that 1124373765 M * Bertl (if not already done so) 1124373770 M * Doener Bertl: ok, after learning how to use cscope again, and figuring out, that i used the wrong search terms, i agree to the proposed change ot htonl() 1124373795 A * Doener .oO( how could i forget all that stuff so fast?! ) 1124373806 M * Bertl okay, funny part is, we use both a define IPI_LOOPBACK and the address/number itself 1124373855 M * wibble do you need vserver history tracing? 1124373861 M * Bertl can't hurt 1124373901 M * Bertl CONFIG_LOG_BUF_SHIFT <-- that's the buffer option 1124373940 M * wibble set to 14 atm 1124373946 M * wibble what you wnat me to set it to? 1124373950 M * Bertl Doener: what if we remove the single ip special case conditionally on the LEGACYNET? 1124373966 M * Bertl wibble: check the help text, use the largest buffer possible 1124373999 M * wibble brb 1124374005 M * Doener Bertl: hm, you need legacynet for util-vserver (or did that change?), but still might want to use some other tool for otf ip address changes 1124374029 M * Bertl only the network setup part of util-vserver (chbind) requires that 1124374208 M * Bertl (so that would need a minor fix then) 1124374262 M * Doener hm... folks would need a new kernel anyways... so they could as well upgrade uv... ok 1124374703 M * daniel_hozac morning. 1124374734 M * Bertl morning daniel_hozac! 1124374861 M * daniel_hozac morning Bertl. 1124374900 M * Bertl Doener: http://vserver.13thfloor.at/Experimental/FOR-2.0.1/delta-loopback-fix01.diff 1124375042 M * Bertl hmm, scratch it, I forgot the non macro case :/ 1124375076 M * Doener heh :) 1124375299 M * Bertl Doener: okay, updated in place ... 1124375477 M * Doener looks good 1124375652 M * Bertl hmm, regarding the 'change ips on the fly with old style networking' 1124375672 M * wibble right 1124375680 M * wibble rebooting with the new debug kern 1124375694 M * wibble got a feeling the box is going to hang 1124375700 M * wibble since its stuck on vserver stop 1124375718 M * Bertl Doener: I think removing the single ip special case isn't really helpful ... 1124375731 M * Bertl wibble: hmm, you could do a sync and reboot -f ? 1124375793 M * wibble nope 1124375801 M * wibble your patches owned my box :P 1124375837 M * Doener Bertl: why is that? 1124375861 M * Bertl Doener: because what happens to services which intentionally bound to a single ip, if , for example, you change that ip? 1124375894 M * Doener hm, what would change for those services? the check doesn't apply to them 1124375925 M * Bertl neither would it to services which bind after the change? 1124375942 M * Bertl so basically we get this scenario: 1124375959 M * Bertl service binds to 0.0.0.0 or single ip 1124375988 M * Bertl then we add another two ips to the v4root, and remove the original one 1124376011 M * Bertl after that, every service has to be restarted to work again 1124376055 M * Bertl I agree that the service binding to 0.0.0.0 would not need that with the single ip hack removed 1124376108 M * Bertl but I think it would get more complicated with for example tcp connections active on a 'removed' ip and such ... 1124376229 M * Doener hm, yeah... might be a problem 1124376310 J * obi ~obi@asus.saftware.de 1124376320 M * Bertl welcome obi! 1124376333 M * daniel_hozac Bertl: so, any ideas why your util-vserver patch would cause the EFAULT from showattr/setattr? 1124376334 M * Bertl Doener: maybe a special flag might be an option? 1124376340 M * obi hi all! 1124376364 M * Doener Bertl: sounds reasonable 1124376369 M * Bertl daniel_hozac: you mean the fix/update? 1124376399 M * daniel_hozac patch-0.30.208-fix01.diff 1124376451 M * Bertl it changes the syscall wrapper ... so if the gcc miscompiles that, it might be the cause ... 1124376466 M * wibble Bertl: right, new kernel in place 1124376504 M * Bertl daniel_hozac: if you tracked it down to that, I'd be interested in the .E and .s output of the file in question (showattr in this case?) 1124376540 M * Bertl wibble: excellent ... 1124376650 M * daniel_hozac Bertl: well, reversing that patch makes showattr work again. i've been unable to reproduce the issue with #include "lib/syscall-alternative.h"\n_syscall3(int, write,...) though. 1124376669 M * wibble so basically /var/lib/vservers is reiser. I am building a new vserver. Will set some vdlimits and then attempt to install some packages 1124376704 M * Bertl okay 1124376786 M * Bertl daniel_hozac: could you upload your broken showattr binary please? 1124376856 M * daniel_hozac http://daniel.hozac.com/tmp/showattr 1124376921 M * Neubix puuu... ready .. bertl, I have 12 sites of documentation.. how to install vserver with quota support. now I search for some errors 1124376945 M * wibble ok just being thick - i thought reiser didnt have inodes? 1124376970 M * Neubix when I'm redy, can you read this and correct or add something ? 1124377096 M * Bertl daniel_hozac: 1124377097 M * Bertl 8048b54: b9 00 00 00 00 mov $0x0,%ecx 1124377097 M * Bertl 8048b59: bb 01 00 01 26 mov $0x26010001,%ebx 1124377097 M * Bertl 8048b5e: b8 11 01 00 00 mov $0x111,%eax 1124377097 M * Bertl 8048b63: cd 80 int $0x80 1124377132 M * Bertl and I forgot: 1124377133 M * Bertl 8048b52: 89 e2 mov %esp,%edx 1124377199 M * Bertl so this is a perfect syscall here, 273 = 0x111, 0x26 = 38 = inode cat 1124377231 M * Bertl the id argument is zero and the struct argument is the stack 1124377334 M * wibble Filesystem Size Used Avail Use% Mounted on 1124377334 M * wibble /dev/hdv1 4.8G 600K 4.6G 1% / 1124377334 M * wibble none 16M 0 16M 0% /tmp 1124377334 M * wibble test12:/# df -i 1124377334 M * wibble Filesystem Inodes IUsed IFree IUse% Mounted on 1124377336 M * wibble /dev/hdv1 0 0 0 - / 1124377339 M * wibble none 129305 1 129304 1% /tmp 1124377365 M * wibble ok - so we have vdlimits, but I have set the inode max to be 100 (silly but should affect reiser no?) 1124377369 M * Bertl the inode = 0,0,0 is expected, it seems reiserfs does not support inode counts 1124377388 M * wibble backup:~# vdlimit --xid 120 /var/lib/vservers/ 1124377388 M * wibble 120 /var/lib/vservers/ 1124377388 M * wibble space_used=605 1124377388 M * wibble space_total=5000000 1124377388 M * wibble inodes_used=28 1124377389 M * Bertl but the dlimits will support it :) 1124377391 M * wibble inodes_total=100 1124377393 M * wibble reserved=5 1124377420 M * wibble so if i do apt-get install samba (or watever I like) this will own the box lol 1124377446 M * wibble and wont do it gracefully 1124377451 M * wibble which is part of the prob no? 1124377526 M * Bertl does it happen again? 1124377535 M * Bertl i.e. can you reproduce it? 1124377586 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net 1124377592 M * Bertl welcome stefani! 1124377611 M * stefani hullo. 1124377617 M * Bertl daniel_hozac: this is on x86, right? 1124377618 M * stefani question on networking. 1124377620 M * wibble hmm 1124377621 M * wibble no 1124377634 M * wibble it ran out of inodes, but didnt hang the box 1124377690 M * daniel_hozac Bertl: yeah. 1124377720 M * Bertl hmm, actually I see three places where you could get an EFAULT ... 1124377734 M * Bertl let's see if your binary can tribber it in my test system :) 1124377751 M * Bertl *trigger of course ... 1124377770 M * stefani can a vserver guest have an IP addr/gateway on a subnet different from host ? 1124377777 M * Bertl yep 1124377778 M * stefani or is that part of NGNET ? 1124377821 M * Bertl http://archives.linux-vserver.org/200311/0470.html 1124377851 M * Bertl wow, didn't see the distro specific howto section until now ... 1124377903 M * wibble ah-ha 1124377911 M * wibble I think I know what is causing part of the vlan problem 1124377920 M * wibble on shutdown, the vconfig command is incorrect. 1124377933 M * wibble hes trying to do vconfig rem eth0 2000 1124377939 M * wibble where it should be vconfig rm eth0.2000 1124377943 M * Bertl sounds like an upcoming solution/fix to me :) 1124377949 M * wibble I wonder if I could fix that 1124378089 M * Bertl daniel_hozac: any specific file it fails on? does it fail with /proc entries too? 1124378111 M * daniel_hozac it fails on just about everything, but just the first entry it tests. 1124378124 M * Bertl ah, yes, cool! 1124378286 M * wibble whats enrico's email addy? 1124378298 M * wibble if I can have it that is! 1124378354 M * Bertl you're not subscribed to ml yet? 1124378359 M * wibble i am 1124378363 M * wibble is that the best place to send it to? 1124378367 M * wibble I wnat to bug him re: vlan :) 1124378373 M * Bertl he posted yesterday, so you can get his email from there ... 1124378377 M * wibble ah ok 1124378379 M * wibble so 1124378387 M * wibble i believe that reiser dlimits works 1124378393 M * Bertl but please CC to the ML 1124378408 M * Bertl wibble: well, yes and no, I guess there is still an issue, after your ahng 1124378430 M * wibble yeah ok. I think i figured out why it doesnt destroy the vlan int - but cant make head nor tail of his code 1124378438 M * wibble i.e. where the shutdown occur 1124378443 M * Bertl wibble: the problem is to trigger it ... because those issues usually arise from a badly handled error path ... 1124378468 M * wibble Bertl: i just tried it and appears to be OK, when I ran out of inodes, dpkg exited gracefully 1124378481 M * wibble but I want to try it a few times to see what happens 1124378495 M * Bertl yeah, please also try with ext3, maybe the ext3 caused it ... 1124378574 M * wibble ok will do! 1124378593 M * Bertl and if something hangs, please do the stack trace thingy 1124378628 M * Bertl daniel_hozac: aha! 1124378636 M * Bertl # /tmp/showattr.daniel /proc/cmdline 1124378636 M * Bertl [ 47.430710] right before copy_from_user 1124378636 M * Bertl [ 47.431559] right before user_path_walk_link 1124378636 M * Bertl [ 47.435344] after user_path_walk_link ret=-14 1124378636 M * Bertl vc_get_iattr(): Bad address 1124378638 M * Bertl ERR /proc/cmdline 1124378700 M * daniel_hozac so the filename is corrupted somehow? 1124378709 M * daniel_hozac (the pointer, that is) 1124378715 M * Bertl yes, it seems like, going to add more info ... 1124378741 Q * lownoize Ping timeout: 480 seconds 1124378813 M * Bertl stefani: does the url help you/answer your question? 1124378837 N * _mcp mcp 1124378846 M * Bertl hey mcp! LTNS! 1124378863 M * eyck what is LTNS? 1124378863 M * wibble ok well it does appear to work 1124378878 M * stefani Bertl: only partly. 1124378887 M * Bertl eyck: Long Time No See 1124378902 M * Bertl daniel_hozac: [ 32.843002] user data = 7fbca37c, name = 00000000 1124378909 M * stefani it tells me it is possible. but how to put it into configs' 1124378915 M * Bertl daniel_hozac: no wonder that blows up :) 1124378938 M * daniel_hozac Bertl: indeed... 1124378956 M * Bertl now the really interesting question is, why is the name argument NULL? 1124378991 M * Bertl stefani: you put it into the host config ... 1124379006 M * stefani y. 1124379012 M * Bertl stefani: i.e. whatever tool you use to configure routing/networking on your host 1124379048 M * Bertl the ip (iproute2) utility should do fine ... 1124379350 J * lownoize ~lownoize@p54ACC0FA.dip0.t-ipconnect.de 1124379379 M * daniel_hozac Bertl: utilvserver_checkCompatVersion seems to overwrite the pointer... 1124379409 M * Bertl welcome back lownoize! 1124379429 M * Bertl daniel_hozac: interesting ... but that can only be triggered by my patches, not caused, no? 1124379445 M * daniel_hozac well, that uses a syscall to get the version... 1124379454 M * daniel_hozac (but only on the first run) 1124379556 M * wibble Bertl: is there any way to disable the inode limiting on reiser? 1124379577 M * wibble or do I just set it to a very large number? 1124379995 M * wibble interestingly df -i doesn't display inode usage 1124380050 M * daniel_hozac Bertl: correct me if i'm wrong (i suck at assembly), but isn't the filename stored in %ebx in vc_get_iattr, which is overwritten by the syscall in utilvserver_checkCompatVersion? 1124380191 M * Bertl sec 1124380387 M * Bertl yep, but it seems that gcc is not detecting that it clobbers the register ... 1124380433 M * Bertl we had that with pic code, and the gcc folks told me that this is proper behaviour with pic code, but gcc should notice with non pic code 1124380471 M * Bertl now the syscall code should do the correct thing with PIC enabled ... 1124380486 M * Bertl maybe they changed tis in recent releases? 1124380584 M * daniel_hozac likely... 1124380697 M * Bertl ah, no, it's a bug in my code ... 1124380719 M * Bertl could you change the syscall-alternative.h in the following regard: 1124380732 M * Bertl #define __sysc_regs "ecx", "edx", "esi", "edi" 1124380743 M * Bertl #define __sysc_regs "ebx", "ecx", "edx", "esi", "edi" 1124380774 M * Bertl and try with that? 1124380788 M * Bertl (of course, this will need an #ifdef regarding PIC 1124380817 M * daniel_hozac yep, that makes it work. 1124380832 M * Bertl excellent, thanks for tracking that down! 1124380843 M * Bertl could you try the following define 1124380857 M * Bertl nah, forget it, doesn't work :) 1124380879 M * Bertl okay, I'll update the patches asap ... 1124380897 M * daniel_hozac great, thanks. 1124380916 M * Bertl no, thank you! :) 1124381119 M * wibble backup:/var/lib/vservers# lsxid /var/lib/vservers/test13/export 1124381119 M * wibble lstat(): No such file or directory 1124381133 M * wibble should that not be lxid instead of lstat ? 1124381140 M * wibble lsxid rather 1124381145 M * daniel_hozac it's the lstat function that fails. 1124381146 M * wibble actually 1124381152 M * wibble im talking shit 1124381155 M * wibble as usual. 1124381179 M * wibble hmm 1124381181 M * wibble ok heres a q 1124381186 M * Bertl don't worry ... better ten useless comments, than one pretty useful missed ... 1124381214 M * wibble when I build the vserver with tagxid enabled, it doesn't create the files with xid ? 1124381218 M * wibble test20 /var/lib/vservers/test20/export 1124381218 M * wibble 0 /var/lib/vservers/test20/initrd 1124381220 M * wibble for example 1124381229 M * wibble i created export inside the vserver once I used vdlimits ? 1124381253 M * wibble is there a way to tag all the other files, so they are included in the dlimit ? 1124381262 M * Bertl yes, the creation happens on the host, in xid=0 ... 1124381284 M * Bertl the proper way to set the 'current' values would be to check for the proper xid 1124381294 M * Bertl remove the second proper :) 1124381317 M * wibble confused. 1124381368 M * Bertl files with xid != the guest context id should not be accounted to the 'current' setting 1124381384 M * Bertl (because they do not belong to the context) 1124381412 M * Bertl so if such a file (with xid=0) is changed from within the context, they will be 'moved' into the context (accounting wise) 1124381442 M * Bertl okay, dinnertime .. back in 20 I guess ... 1124381450 N * Bertl Bertl_oO 1124381554 M * stefani Bertl_oO: ok . that was not too painful 1124381731 Q * _ag_ Quit: BBL 1124381976 Q * Neubix Quit: Verlassend 1124383190 J * keyser_soze ~cimarron@host196.201-252-24.telecom.net.ar 1124383291 N * Bertl_oO Bertl 1124383310 M * Bertl okay, back for a few minutes, then translocating :) 1124383328 M * Bertl stefani: so you ahve your multi-gateway setup? 1124383391 M * stefani Bertl: actually did absolutely nothing special 1124383434 M * stefani in the host i did vserver build --interface eth1:/ blah-blah 1124383441 M * stefani and it started up and works 1124383490 M * Bertl excellent! that's the way it's supposed to be :) 1124383510 M * stefani bien sur. 1124383822 J * Matthew-1 ~mlong@grm-nx.grmims.com 1124383845 M * Matthew-1 hello 1124383858 M * Bertl welcome Matthew-1! 1124383882 M * Matthew-1 is it possible to run nfs-userspace-server with in vserver? 1124383902 M * Matthew-1 i tried and it didn't seem to work 1124383904 M * Bertl yes, somebody reported unfsd working quite fine 1124383925 M * daniel_hozac but only when using the debian patched version. 1124383932 M * Matthew-1 oh 1124383933 M * Bertl correct 1124383933 M * Matthew-1 ok 1124383936 M * Matthew-1 no 1124383945 M * Matthew-1 this is suse version 1124383956 M * Bertl you just need one of the patches 1124383968 M * Matthew-1 which one 1124383968 M * Bertl maybe the cvs version already has that one or similar in 1124383973 M * Matthew-1 do you know 1124383974 M * Matthew-1 ok 1124383984 M * Matthew-1 let me build it 1124383995 M * Matthew-1 i'll see if that works 1124384022 M * Matthew-1 thanks 1124384337 M * Bertl daniel_hozac: could you give the following patch a try? http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/delta-0.30.208-shiny7.diff 1124384345 M * Bertl (you have to revert the shiny6) 1124384361 M * daniel_hozac sure 1124384827 M * daniel_hozac Bertl: works great. 1124384880 M * Bertl excellent ... thanks again! 1124385018 J * Johnsie John@acs-24-154-53-74.zoominternet.net 1124385330 M * Bertl welcome Johnsie! 1124385348 M * Bertl daniel_hozac: okay, updated the fix on 13thfloor as well as the UV patches ... 1124385364 M * Bertl (and of course the mandrake packages) 1124385377 M * Johnsie Hello. 1124385433 M * daniel_hozac Bertl: great, thanks again. 1124385468 M * Bertl my pleasure ... so I'm translocating now ... 1124385479 M * Bertl (back in about two hours ...) 1124385489 N * Bertl Bertl_oO 1124385512 M * Matthew-1 i don't think unfsd can be considered stable for production use. 1124385523 M * daniel_hozac why? 1124385530 M * Matthew-1 i compiled it and tried to run it again 1124385546 M * Matthew-1 and portmap is still giving problems 1124385591 M * Matthew-1 it seems like i have to do too many hacks to get it working 1124385605 M * Matthew-1 at least in vserver 1124385611 M * Matthew-1 am i wrong about this? 1124385652 M * Matthew-1 portmap[3341]: cannot find any active local network interfaces 1124385676 M * Matthew-1 portmap[3346]: connect from 192.168.1.231 to unset(nfs): request from non-local host 1124385719 M * Matthew-1 this is a portmap issue maybe not a unfsd issue but still 1124385734 M * daniel_hozac portmap is broken in vservers. 1124385751 M * Matthew-1 ok 1124385769 M * Matthew-1 then it seems unfsd is alos 1124385784 M * Matthew-1 i can use it without portmap 1124385812 M * Matthew-1 but then that's just silly. 1124385836 M * daniel_hozac you can workaround the portmap brokenness by giving your vserver an alias on the interface, IIRC: 1124385846 M * daniel_hozac i have a patch for it as well. 1124385861 Q * Johnsie Ping timeout: 480 seconds 1124385913 M * Matthew-1 a patch for vserver? 1124385914 Q * keyser_soze Quit: Abandonando 1124385930 M * Matthew-1 or for portmap? 1124385933 M * daniel_hozac portmap. 1124385976 M * Matthew-1 how stable would you say the patch is? 1124386016 M * daniel_hozac well, it's a pretty simple patch. 1124386028 M * daniel_hozac http://daniel.hozac.com/stuff/portmap-4.0-getifaddrs.patch 1124386056 M * daniel_hozac i'd say it actually makes portmap more stable than anything else. 1124386064 M * Matthew-1 ha ha 1124386092 M * daniel_hozac hmm, there does seem to be a missing line in that patch... 1124386125 M * Matthew-1 oh man... now i have to download protmap and compile that too : ( 1124386169 M * Matthew-1 well it sucks i don't know c 1124386253 M * daniel_hozac there, that patch should do it. 1124386301 M * Matthew-1 ok.. i'll test it 1124386331 M * Matthew-1 make sure it work ; ) 1124387778 J * _ag_ ag@caladan.roxor.cx 1124387952 M * Matthew-1 ok.... unfsd seems to work but without portmap 1124387977 M * Matthew-1 i didn't recompile portmap i can't find the sources 1124388022 M * Matthew-1 but unfsd is working in vserver 1124389941 M * lownoize unfsd was very broken here 1124389955 J * MajorMajor ~blah@ip24-254-95-67.dc.dc.cox.net 1124389985 M * lownoize worked for most stuff but had some strange errors when applications tried to write to nfs server 1124390144 M * MajorMajor hi, does anyone know of a good documented comparison of VServer versus something like Solaris Zones? 1124390948 Q * MajorMajor Quit: Download Gaim: http://gaim.sourceforge.net/ 1124391321 Q * TheSeer Remote host closed the connection 1124392165 J * TheSeer ~theseer@border.office.salesemotion.net 1124394049 J * saddi ~sven@GKCDVI.dsl.saunalahti.fi 1124395313 P * fobi by 1124397906 J * yarihm ~yarihm@80-218-5-17.dclient.hispeed.ch 1124398320 Q * CosmicRay Quit: Client exiting 1124400397 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-194-10.adsl.proxad.net 1124400472 M * mess-mate is util-vserver-0.30.208 buggy ??? 1124400540 M * daniel_hozac what do you mean? 1124400544 M * mugwump mess-mate: There's a "fix" for it available, check the wiki or vserver.13thfloor.at 1124400560 M * daniel_hozac that's more of an update though, no? 1124400575 M * mugwump dunno. .206 works for me :) 1124400677 M * mess-mate sorry, but couldn't start a vserver. Looked at /usr/sbin/vserver and found a first bug about VSERVER_DIR. 1124400791 M * daniel_hozac what error did you get when starting the vserver? what bug? 1124401081 Q * Pazzo Quit: bye 1124401121 M * mess-mate (*) VSERVER_DIR=$__CONFDIR/$vserver changed to (*) VSERVER_DIR=/vservers/$vserver otherwise i've an error like 'WARNING: couldn't find serverdir....user the legacy...' 1124401168 M * daniel_hozac you don't have the configuration directories on /vservers, do you? 1124401193 M * mess-mate But there is more because can't start a vserver: find: /vservers/test/proc/16799/fd/4: No such file or directory 1124401239 M * daniel_hozac how did you create the vserver? 1124401313 M * mess-mate daniel_hozac: created a test.conf in /etc/vservers as test.conf and a dir as /vservers/test 1124401360 M * daniel_hozac ... that is a legacy configuration. 1124401483 M * mess-mate Ok i'll reset the original /usr/sbin/vserver 1124401609 M * mess-mate Starting the virtual server test; Server test is not running; find: var/run: No such file or directory; etc.. 1124401651 M * daniel_hozac you do have a proper chroot in /vservers/test, right? 1124401715 M * mess-mate capchroot: execvp("/etc/rc.d/rc"): No such file or directory 1124401746 M * mess-mate Of course i haven't a /etc/rc.d = Debian 1124401770 M * daniel_hozac you'll need to change the init type then. 1124402458 M * mess-mate you mean the file /usr/lib/util-vserver/legacy/vserver ? 1124402804 M * daniel_hozac well, i don't know how to do it with the legacy configuration, but with new style configurations, you have /etc/vservers//apps/init/style.init/type 1124402822 M * daniel_hozac s/init\/type// 1124403477 Q * yarihm Quit: Leaving 1124403611 J * keyser_soze ~cimarron@host196.201-252-24.telecom.net.ar 1124403805 M * mess-mate daniel_hozac: i haven't a dir as /etc/vservers/test, but there is a defaults without any 'init' file. 1124403824 M * daniel_hozac because you have a legacy configuration. 1124403887 M * daniel_hozac using vserver ... build would have created a new style configuration as well a guest for you. 1124403952 M * mess-mate daniel_hozac: i've tryed a vserver test build and have this: No build-method specified 1124403973 M * daniel_hozac you need to read the --help output and at least specify -m. 1124404118 M * mess-mate daniel_hozac:man page said: build Create a virtual server by copying the packages of the root 1124404169 M * daniel_hozac the man page is vastly outdated. 1124404209 Q * zobel Ping timeout: 480 seconds 1124404393 M * mess-mate daniel_hozac: Ok.. i'll reading/finding what more doc tomorrow. Go sleep, 0:32 here :) Thanks for your help. Come back. 1124404409 M * daniel_hozac 0:33 ;) 1124404433 M * mess-mate daniel_hozac: what... european ? 1124404444 M * daniel_hozac yep. 1124404466 M * daniel_hozac Sweden. 1124404565 Q * mess-mate Quit: leaving 1124404568 J * zobel zobel@zobel.irc.ftbfs.de 1124404916 J * renihs ~renihs___@193.170.52.70 1124405820 M * renihs hmm is it normal that /vservers/ shows for example 500mb with du -hs /vservers/, but a rsync copy of it is larger? (~+150mb) 1124405994 N * Bertl_oO Bertl 1124406013 M * Bertl evening folks! took a little longer than expected ... 1124406072 M * Bertl renihs: depends on the setup and probably on broken links? 1124406113 M * renihs Bertl, sounds reasonable 1124406117 M * renihs hmm 1124406141 M * Bertl could be that you used some options which expanded symlinks? 1124406162 M * Bertl that would multiply most libraries by 2-5 1124406277 M * renihs u mean when rsyncinc? simple rsync -a 1124406546 M * Bertl just -a? 1124406570 M * renihs yup 1124406579 M * Bertl what about -x ? 1124406605 M * renihs :( 1124406609 M * renihs thx! 1124406620 M * Bertl ah, and: 1124406624 M * Bertl Note however that -a does not preserve hardlinks, 1124406624 M * Bertl because finding multiply-linked files is expensive. 1124406624 M * Bertl You must separately specify -H. 1124406647 M * renihs k 1124406656 M * Bertl you're welcome! 1124406676 M * renihs thanx alot! 1124406908 M * Bertl evening zobel! up late? 1124409483 Q * Doener Ping timeout: 480 seconds 1124409494 J * Doener ~doener@p54877B68.dip.t-dialin.net