1196037705 J * tanjix ~tanjix@dslb-084-058-011-001.pools.arcor-ip.net 1196037708 Q * tanjix2 Ping timeout: 480 seconds 1196038094 Q * meandtheshell Quit: Leaving. 1196038466 J * gm_ ~gm@host193-107-dynamic.8-87-r.retail.telecomitalia.it 1196038474 Q * gm_ 1196038490 J * gm ~randall@host193-107-dynamic.8-87-r.retail.telecomitalia.it 1196039134 Q * mire Read error: Operation timed out 1196039623 Q * Johnnie Quit: G'bye! 1196039637 J * Johnnie ~jdlewis@c-67-163-142-234.hsd1.pa.comcast.net 1196039759 Q * Johnnie 1196040034 Q * FireEgl Quit: Bye... 1196043383 J * yarihm ~yarihm@84-75-119-160.dclient.hispeed.ch 1196043389 Q * yarihm 1196045429 J * shuri ~shuri@64.235.209.226 1196046292 Q * shuri Remote host closed the connection 1196048800 Q * _bjh_ Server closed connection 1196048808 J * _bjh_ ~bjh@84.112.154.154 1196049212 Q * hparker Remote host closed the connection 1196049230 J * hparker ~hparker@linux.homershut.net 1196050773 Q * balbir Ping timeout: 480 seconds 1196054542 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa 1196054719 Q * matti Remote host closed the connection 1196055375 J * balbir ~balbir@59.145.136.1 1196057722 J * sharkjaw ~gab@shell.ormset.no 1196058553 Q * dsoul Ping timeout: 480 seconds 1196059014 Q * nebuchadnezzar Server closed connection 1196059092 J * virtuoso_ ~s0t0na@ppp89-110-58-50.pppoe.avangarddsl.ru 1196059093 Q * virtuoso Read error: Connection reset by peer 1196059212 J * ntrs__ ~ntrs@79.125.228.243 1196060239 Q * tanjix Ping timeout: 480 seconds 1196060532 J * DavidS ~david@p54812D39.dip0.t-ipconnect.de 1196061035 Q * rorem- Ping timeout: 480 seconds 1196061810 J * rorem- ~roremtank@bzq-219-46-202.isdn.bezeqint.net 1196061958 J * JonB ~NoSuchUse@kg1-106.kollegiegaarden.dk 1196062556 J * larsivi ~larsivi@85.221.53.194 1196062938 Q * ntrs__ Quit: Leaving 1196063435 Q * JonB Quit: This computer has gone to sleep 1196063566 J * tanjix ~tanjix@office.star-hosting.de 1196064237 J * JonB ~NoSuchUse@kg1-106.kollegiegaarden.dk 1196064434 J * bzed_ ~bzed@devel.recluse.de 1196064512 Q * bzed Remote host closed the connection 1196065469 Q * JonB Ping timeout: 480 seconds 1196065495 Q * tokkee Server closed connection 1196065497 J * tokkee tokkee@ssh.faui2k3.org 1196065599 J * Alikus ~alikus@217.150.200.212 1196065843 Q * AndrewLee Server closed connection 1196065849 J * AndrewLee ~andrew@flat.iis.sinica.edu.tw 1196066811 J * dna ~dna@128-235-dsl.kielnet.net 1196067129 J * gebura ~gebura@77.192.186.197 1196067139 Q * almak Server closed connection 1196067140 J * almak ~almak@willers.employees.org 1196067186 M * gebura hi 1196067839 J * JonB ~NoSuchUse@130.227.63.19 1196068105 J * Punkie ~Punkie@goc.coolhousing.net 1196070549 J * lilalinux ~plasma@dslb-084-058-225-197.pools.arcor-ip.net 1196070650 Q * ensc Ping timeout: 480 seconds 1196070701 J * ensc ~irc-ensc@77.235.182.26 1196071075 Q * larsivi Quit: Konversation terminated! 1196072384 J * lilalinux_ ~plasma@80.69.41.3 1196072573 J * lilalinux__ ~plasma@dslb-084-058-225-197.pools.arcor-ip.net 1196072609 Q * lilalinux Read error: Connection reset by peer 1196072816 N * Bertl_zZ Bertl 1196072825 M * Bertl morning folks! 1196072872 M * gebura morning ! 1196073007 Q * lilalinux_ Ping timeout: 480 seconds 1196073041 Q * Aiken Remote host closed the connection 1196073146 M * DavidS hi bertl! 1196073223 J * Yvo ~yvonne@91.64.217.106 1196073237 P * Yvo 1196073439 J * larsivi ~larsivi@144.84-48-50.nextgentel.com 1196074167 M * Bertl okay, off for now .. back later ... 1196074171 N * Bertl Bertl_oO 1196075123 Q * zbyniu Remote host closed the connection 1196075125 J * zbyniu ~zbyniu@host13-188.crowley.pl 1196076876 Q * Punkie Quit: Odcházím 1196077220 J * Aiken ~james@ppp121-45-221-157.lns2.bne1.internode.on.net 1196077983 Q * balbir Ping timeout: 480 seconds 1196078123 Q * Aiken Quit: Leaving 1196078242 J * meandtheshel1 ~sa@85.127.109.80 1196078775 Q * AStorm Quit: ET calling home 1196078777 J * AStorm ~astralsto@tor-irc.dnsbl.oftc.net 1196079401 J * mire ~mire@115-168-222-85.adsl.verat.net 1196080492 Q * sharkjaw Remote host closed the connection 1196080722 J * yarihm ~yarihm@whitehead2.nine.ch 1196080865 J * mrfree ~mrfree@88.40.89.1 1196080880 M * mrfree hi all 1196080969 M * mrfree I tried to add a simple alias in .bashrc into a guest env (/root/.bashrc) but it seems it doesn't start 1196081102 J * Punkie ~Punkie@goc.coolhousing.net 1196081521 J * esa bip@ip-87-238-2-45.adsl.cheapnet.it 1196081576 M * nkukard hrmmmm 1196081594 M * nkukard if I've edited /etc/resolv.conf in a vserver, what could cause it from using the vserver IP for name resolution 1196081601 M * nkukard this is totally baffling me 1196081610 M * mrfree my bad... I need to use .bash_profile instead of .bashrc 1196081620 M * nkukard *cause it using not cause it from using 1196082222 M * mrfree sometime I have "vxW: xid=1345 did lookup hidden ffff810100055070[#0,4026531872] �/proc/bus�." in my logs ... what it means? 1196083273 N * ensc Guest327 1196083283 J * ensc ~irc-ensc@77.235.182.26 1196083293 N * Bertl_oO Bertl 1196083320 M * Bertl nkukard: what should use the guest? ip? 1196083340 M * nkukard well ... i set my nameserver to the hosts IP 1196083347 J * balbir ~balbir@59.145.136.1 1196083348 M * Bertl mrfree: means that a process in context 1345 did lookup a hidden entry (/proc/bus) 1196083372 M * Bertl nkukard: and your host is running a named? 1196083399 M * nkukard ah, fixed it .... was because the hosts nameserver was 127.0.0.1 1196083472 M * mrfree Bertl, is it bad? :) 1196083503 M * Bertl no, it just inform you, that some program is doing something you do not necessarily want 1196083552 M * nkukard seeing as my vserver doesnt' have 127.0.0.1, it used the first available IP, which was the vservers IP as the NS 1196083558 M * mrfree ok 1196083559 M * nkukard i think that was my problem 1196083586 M * Bertl yes, 127.0.0.1 is remapped 1196083604 M * Bertl (in recent kernels to a special loopback ip) 1196083624 M * nkukard so its mapped to the first IP for the vservers? 1196083627 M * nkukard *vserver 1196083694 Q * Guest327 Ping timeout: 480 seconds 1196083713 M * nkukard grr, remapped to the first IP on the vserver? 1196083721 M * Bertl for non devel kernels, yes 1196083796 M * nkukard and for devel kernel? 1196083840 M * Bertl it is mapped to the lback address 1196083852 M * nkukard on the host platform? 1196083870 M * Bertl hmm? 1196083901 M * nkukard sorry if I sound dumb ... in devel kernels does the vserver have its own loopback iface or does it remap 127.0.0.1 to the actual server and not the vserver? 1196083933 M * Bertl no, there are no _guest_ interfaces in Linux-VServer, as we do IP (layer 3) isolation not network stack virtualization 1196083947 M * Bertl but the guest 'sees' the lback address as 127.0.0.1 1196083963 M * Bertl so it 'looks' like it would have it's own loopback interface 1196083963 M * nkukard which is infact its own IP addy, right? 1196083994 M * Bertl which is a special loopback guest ip (lback address) by default 127.x.y.1 1196084011 M * Bertl where x and y is the higher and lower byte of the network context 1196084020 M * nkukard ah 1196084027 M * nkukard thats on devel kernels? 1196084034 M * Bertl yep 1196084086 J * mountie ~mountie@trb229.travel-net.com 1196084130 M * nkukard thanks Bertl :) 1196084132 Q * ag- Quit: Aiee, killing interrupt handler! 1196084145 J * ag- ag@fedaykin.roxor.cx 1196084300 M * Bertl you're welcome! 1196084427 Q * virtuoso_ Remote host closed the connection 1196084798 Q * zbyniu Quit: Lost terminal 1196084913 J * zbyniu ~zbyniu@host13-188.crowley.pl 1196085030 M * Bertl okay, off again ... bbl 1196085033 N * Bertl Bertl_oO 1196085227 J * Yvo yvonne@vpn052.rz.uni-mannheim.de 1196085241 P * Yvo 1196085464 Q * balbir Ping timeout: 480 seconds 1196086056 J * mark ~mark@ip-83-238-175-215.netia.com.pl 1196086305 M * mark hello everyone 1196086321 M * daniel_hozac hi 1196086334 M * mark has anyone seen this when using chxid: test/test/usr/share/locale/vi/LC_MESSAGES/ld.mo: Invalid argument 1196086335 M * mark hi DH 1196086343 M * mark I DO have tagxid on my fs 1196086371 M * mark on home type ext3 (rw,tagxid) 1196086384 M * daniel_hozac are you sure that's not a symbolic link to some other filesystem which doesn't have tagging enabled? 1196086396 M * mark positively sure 1196086406 M * mark that's just regular partition, on LVM 1196086431 M * mark architecture is 64bit though 1196086446 M * mark 2.6.22-3-vserver-amd64 #1 SMP Mon Nov 12 18:26:47 UTC 2007 x86_64 GNU/Linux 1196086458 M * daniel_hozac i meant the file... 1196086458 M * mark this is 64bit Debian Etch 4.0r1 1196086472 M * mark hold on 1196086526 M * mark chxid -c 147 smth 1196086526 M * mark smth: Invalid argument 1196086536 M * mark I just created this file with touch 1196086551 M * mark basically every file gets "invalid argument" 1196086865 J * Yvo yvonne@vpn051.rz.uni-mannheim.de 1196086884 P * Yvo 1196086888 M * daniel_hozac do you have the vserver debugging enabled in your kernel? 1196086899 M * mark I don't think so - I didn't enable it 1196086906 M * mark if it's not enabled by default, no 1196086917 M * mark I can turn it on, though if you tell me how :-) 1196086969 M * daniel_hozac CONFIG_VSERVER_DEBUG 1196087000 M * mark during kernel compilation, right? 1196087626 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1196088567 N * bzed_ bzed 1196088683 J * Yvo ~yvonne@91.64.217.106 1196088694 P * Yvo 1196089112 Q * Punkie Remote host closed the connection 1196089809 J * Yvo ~yvonne@91.64.217.106 1196089819 P * Yvo 1196089997 M * mark Daniel? 1196090056 M * mark everyone - if you're on 64 bit Etch do NOT use this kernel: http://packages.debian.org/etch-backports/linux-image-2.6-vserver-amd64 1196090061 M * mark chxid doesn't work 1196090089 M * daniel_hozac i doubt that's the problem... 1196090114 M * daniel_hozac cat /proc/mounts to paste.linux-vserver.org, please. 1196090119 M * mark now chxid works once I compiled the kernel 1196090143 M * daniel_hozac what kernel did you compile? 1196090148 M * daniel_hozac how did you configure it? 1196090148 M * mark http://paste.linux-vserver.org/9823 1196090156 M * daniel_hozac how does your configuration differ from the one in the package? 1196090172 M * mark First I installed 2.6.22-3-vserver kernel from backports 1196090174 M * mark then I d' 1196090188 M * mark then I d/l kernel source 2.6.22-12 1196090199 M * mark applied latest stable patch from vserver project 1196090213 M * mark copied the config from the backports kernel /boot folder 1196090226 M * mark configured vserver kernel debugging as you wrote 1196090231 M * mark compiled, rebooted 1196090235 M * mark now chxid works 1196090254 M * daniel_hozac diff the vserver portions of the configs. 1196090268 M * mark ok 1196090343 M * mark http://paste.linux-vserver.org/9824 1196091207 Q * mrfree Quit: Leaving 1196091413 M * mark ok, DH, thanks - CU 1196091419 Q * mark Remote host closed the connection 1196091670 J * dowdle ~dowdle@scott.coe.montana.edu 1196091997 M * dowdle Greetings. 1196092656 P * marcfiu 1196093974 J * arachnis1 ~arachnist@pool-71-174-118-56.bstnma.fios.verizon.net 1196094004 Q * arachnist Quit: brb 1196094004 N * arachnis1 arachnist 1196094469 Q * gebura Quit: Quitte 1196094965 J * misch ~misch@host-12-196-71-245.warpspeed1.net 1196094979 M * misch Good morning .. 1196095015 M * DavidS dowdle, misch: welcome :) 1196095030 M * misch Hey DavidS :) 1196095053 Q * fb Server closed connection 1196095054 J * fb ~fback@red.fback.net 1196095089 M * misch Question: I am running 2.6.20-1.2311.fc5.vs2.2.0.1 for months without problem. I installed ogg123 (including libao) in one client, now the machine reboots every 20-30 minutes. May that be a reason? 1196095114 M * daniel_hozac that's a very old kernel. 1196095125 M * harrydg very old??? not even 6 months 1196095132 M * harrydg mkay... maybe 1 year ;) 1196095148 M * harrydg those faults should never occur in the first place... ;) 1196095150 M * misch It was running just fine (it's only to have a semi-secured environment for an asterisk server) 1196095178 M * daniel_hozac you might want to get a more recent OS, one that is still supported. 1196095197 M * daniel_hozac FC6 is EOLed in like a week. 1196095233 M * misch Never touch a running system ;) It was working just fine without problems. I don;t know why it behaves that strange now .. 1196095253 M * daniel_hozac why does it reboot? 1196095277 M * misch daniel_hozac, Good question. No log entries, no nothing. Just *beep* and reboot. 1196095292 M * daniel_hozac overheating? 1196095334 M * misch It's like 68 in my office. And everything works ok (so far) but starting and stopping a client vserver triggers reboot. 1196095368 M * misch Only thing I did was vyum ogg-devel into the client. 1196095417 M * daniel_hozac does it become unresponsive for some time before it reboots, or is it more like you're typing a command and boom, reboot? 1196095462 J * mark_ ~IceChat7@gate12.kolornet.pl 1196095466 M * misch The latter. I just typed 'vserver asterisk start' and *beep* .. reboot. 1196095499 M * daniel_hozac what's your kernel command line? 1196095526 M * misch kernel /vmlinuz-2.6.20-1.2311.fc5.vs2.2.0.1 ro root=/dev/VolGroup00/LogVol00 rhgb quiet 1196095570 M * misch mpls .. phone call 1196095577 M * daniel_hozac so it shouldn't be a panic. 1196095622 M * misch No panic. I do Linux > 14 years. I wouldn;t have bothered you if I would have any log entry, panic, any thing I could work with .. 1196095650 M * daniel_hozac panics usually don't get you log entries :) 1196095664 M * misch They say one should not vyum into an active client ... I did anyway .. might that mess things up? 1196095665 M * fb daniel_hozac: playing oggs requires more power, thus leads to cpu temp raise. bios notices overheat and reboots the machine -- what do you think? 1196095680 M * misch fb: I didn;t play any oggs. 1196095693 M * daniel_hozac who says that? 1196095705 Q * yarihm Quit: Leaving 1196095706 M * daniel_hozac vyum should be perfectly safe. 1196095717 M * daniel_hozac (that's why it plays so many tricks) 1196095742 M * daniel_hozac misch: does testme cause the crash too? 1196095794 M * fb misch: i had simmilar issue with mu old server, and it was caused by a broken ram 1196095809 M * misch fb, I see .. 1196095828 M * misch daniel_hozac, haven't tried testme .. 1196095834 M * misch .. yet .. 1196095851 M * fb what was strange, it worked for a few years smoothly, then suddenly it started to reboot 1196095853 M * daniel_hozac yeah, it sounds like hardware to me... i'd check the CPU cooler and RAM. 1196095880 M * misch daniel_hozac, good idea .. i check .. 1196095892 M * mark_ fb: run memtest86, and extensively - not just a few minutes 1196095965 M * dowdle fb: Well, things generally work until they stop working. 1196095981 M * fb mark_: my father baked its mobo about a year ago :) 1196096006 M * mark_ sure, but have you ever seen a BURNING mobo, fb? :-) 1196096015 M * fb dowdle: i never thought ram can break during normal working 1196096052 M * mark_ sure it can, some dishonest manufacturers sell RAMs that are on the "verge" of breaking - their high (near the top) addresses are nearly broken 1196096054 M * fb dowdle: during reboot, power cycle, ok. but in normal work conditions? 1196096056 M * mark_ you can't detect it normally\ 1196096066 M * mark_ some usage and off RAM goes into the deep end 1196096131 M * mark_ that's why I always run exhaustive memtest86 on new systems 1196096203 M * fb this ram is still working, in another (not so important ;) system, with badram patch 1196096251 M * mark_ production system with bad ram? 1196096261 M * dowdle fb: I've had RAM work for years and then just break. 1196096263 M * DavidS what golden-plated RAM is worth the troubles of configuring a kernel-patch? 1196096267 M * fb mark_: i said, not so important :P 1196096338 M * mark_ my bottom line is, if it's in use, it's important enough not to have to reinstall it for a measly piece of ram - an hour of (competent) sysadmin time is worth more than average memory module 1196096340 M * fb DavidS: ram for machines based on intel's BX chipset is highly overpriced. 1196096350 M * daniel_hozac rambus? 1196096504 J * TheSeer ~theseer@border.office.nonfood.de 1196096508 M * TheSeer heya :) 1196096531 M * mark_ yo 1196096546 M * TheSeer did something change or do i do something wrong? 1196096559 M * mark_ both? 1196096565 M * fb mark_: and btw, this isn't me who have problems with sudden reboots :) 1196096573 M * TheSeer vnamespace -e mount -o bind,ro /path/in/host /path/in/guest 1196096583 M * mark_ fb: yeah, noticed after I wrote it 1196096587 M * TheSeer that correct? 1196096592 M * daniel_hozac TheSeer: was /path/in/host mounted when the guest was started? 1196096601 M * daniel_hozac TheSeer: is it excluded from namespace cleanup? 1196096602 M * TheSeer yes 1196096610 M * TheSeer how do i tell? 1196096634 M * TheSeer all i'm trying to do is avoid restarting the guest ;) 1196096641 M * daniel_hozac cat /etc/vservers/.defaults/namespace-cleanup-skip /etc/vservers//namespace-cleanup-skip 1196096685 M * TheSeer cat: /etc/vservers/.defaults/namespace-cleanup-skip: No such file or directory 1196096692 M * TheSeer okay.... what does that tell me? ;) 1196096701 M * TheSeer maybe i should just restart the vserver 1196096720 M * TheSeer since i did add the paths into the fstab for the guest already 1196096752 Q * JonB Ping timeout: 480 seconds 1196096752 M * daniel_hozac that would do the trick. 1196096773 M * TheSeer k.. 1196097013 Q * eyck Server closed connection 1196097015 J * eyck XBKDyl7I@nat05.nowanet.pl 1196097518 M * mark_ DH, I'm going to pester you again... 1196097527 M * mark_ vserver ti386 hashify 1196097527 M * mark_ Failed to initialize unification for vserver 1196097551 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1196097577 M * mark_ for some other vservers this works fine 1196097596 M * mark_ however, still no files are find by vhashify operation 1196097597 M * daniel_hozac and you followed the steps in http://linux-vserver.org/Frequently_Asked_Questions#How_do_I_manage_a_multi-guest_setup_with_vhashify.3F right? 1196097603 M * mark_ TWICE 1196097606 M * mark_ to the letter 1196097649 M * mark_ find .hash -type f -links 1 -print0 comes empty 1196097712 M * mark_ restarted vservers a few times, done chxid -c ctx -R /home/ several times 1196097744 M * mark_ lsxid in a directory of any vserver shows correct symbolic ctx of particular guest 1196097849 Q * opuk Server closed connection 1196097861 J * opuk ~kupo@c213-100-138-228.swipnet.se 1196097868 Q * Alikus Remote host closed the connection 1196098164 M * mark_ The dirs etc/vservers//apps/vunify are empty if this has any meaning 1196098182 M * mark_ even though I ran vserver name hashify a few times for each guest 1196099232 M * the-me is there still a limit of 16 ips per vserver in 2.2.0.5? 1196100111 Q * pmenier Quit: Konversation terminated! 1196100128 Q * gm Ping timeout: 480 seconds 1196100565 Q * lilalinux__ Remote host closed the connection 1196101251 M * daniel_hozac the-me: yes, but it's removed in 2.3. 1196101278 M * daniel_hozac mark_: uh, so what does /etc/vservers/.defaults/apps/vunify contain? 1196101483 Q * TheSeer Quit: Client exiting 1196102077 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com 1196102678 J * er ~yakker@aegis.CS.Princeton.EDU 1196102780 J * gm ~randall@host207-15-dynamic.4-87-r.retail.telecomitalia.it 1196104001 M * the-me is there still a limit of 16 ips per vserver in 2.2.0.5? 1196104009 M * the-me sorry for reasking, but I was offline 1196104040 M * daniel_hozac you don't have scrollback? 1196104078 M * nkukard the-me: yes, but it's removed in 2.3. 1196104098 M * the-me daniel_hozac, it was a crash *sigh* :( 1196104109 M * daniel_hozac your client didn't exit. 1196104117 M * the-me hmm, so in 2.2.0.5 I can use more than 16 ips for one vserver and in 2.3.x not? 1196104123 M * the-me daniel_hozac, bouncer :) 1196104133 M * daniel_hozac the-me: other way around.... 1196104162 M * the-me ehh yes, that was a typo, sorry 1196104165 M * the-me okay thanks 1196104349 Q * gm Ping timeout: 480 seconds 1196104624 N * Bertl_oO Bertl 1196104641 M * Bertl evening folks! 1196104658 M * daniel_hozac evening Bertl! 1196104815 M * Bertl how's going? everything fine in the basement? 1196104856 M * DavidS hey Bertl! 1196104926 M * daniel_hozac yep, everything fine here. how are you? 1196104948 M * er hi bertl and daniel_hozac 1196104961 M * daniel_hozac hello er 1196104963 M * Bertl hey er! LTNS! 1196105015 M * Bertl daniel_hozac: quite a number of things to manage with the relocation (electricity, insurance, internet ...) 1196105035 M * daniel_hozac ah, how's it going? how far along are you? 1196105069 M * Bertl it is already warm in the new place (which is a good thing, as it is quite cold outside) 1196105085 M * daniel_hozac hehe 1196105099 M * Bertl but in Austria, everything takes a _very_ long time to be processed 1196105132 M * er Bertl, which of course was your main motivation to make vserver fast, right? :) 1196105148 M * nkukard Bertl, Internet being the most important! 1196105159 M * Bertl for example, we had the electricity part ready (house side) in about a day, and it took about a week for the electricity comnpany to bring the meter (i.e. to provide power) 1196105173 M * Bertl nkukard: of course ;) 1196105192 M * nkukard Bertl, first thing i did moving here was organize internet here *before* i moved in ... haha 1196105248 M * Bertl hehe, what do you think is what I did :) 1196105277 M * Bertl there have to be priorities, water, heating .. not so important :) 1196105499 M * DavidS Bertl: how true, I'm already moving again (within Vechta, but still) and I already deposited, that I'll not leave the old flat (with internet access) unless there is networking in the new one... 1196105504 M * DavidS :) 1196105569 M * misch weird, weird. 1196105588 M * er Bertl, i was going to renew the NetNS+VServer tree we've been using for VINI (=PlanetLab with a per-container route table) 1196105618 M * er and thought I'd come around to ask you if this time I could do it in a way so that I could push something back 1196105640 Q * opuk Quit: leaving 1196105678 M * er or to see if I could hook into any plans you mighth ave to the same effect 1196105687 M * daniel_hozac 2.6.24 has parts of it merged. 1196105781 M * misch daniel_hozac: After replacing RAM, power supply the computer still reboots. I've giving up and changed to a different computer. 1196105791 M * Bertl daniel_hozac: 2.6.24 is out? 1196105827 M * daniel_hozac not yet. 1196105854 M * daniel_hozac rc3, i think. 1196105856 M * Bertl er: yes, network namespaces (mainline) will be an alternative for folks who require full virtualization and are willing to accept the overhead (compared to layer3 isolation) 1196105864 M * misch daniel_hozac: Everything is back to normal and runs happily. I wonder what might be wrong with the other motherboard ... 1196105868 M * DavidS is it a very bad idea to ln -s /proc/mounts /etc/fstab in a vserver? 1196105872 Q * AStorm Remote host closed the connection 1196105880 M * daniel_hozac DavidS: you mean mtab? 1196105891 M * daniel_hozac misch: you didn't change the CPU cooling? 1196105895 M * Bertl er: as far as I see it, most work will have to be done in userspace, the kernel just adds another namespace to unshare 1196105933 M * DavidS daniel_hozac: I'm having a tool failing on the empty default fstab in the container 1196105936 M * misch Well, after changing ram the computer boots right away .. uncompressing linux ,,, *beep* - the cpu has 34 degrees (c) that should be ok. 1196105983 M * DavidS of course I could just put some silly values in it for the fun of it .. I don't think this tool would notice the difference 1196105997 M * daniel_hozac DavidS: the fstab normally has at least an entry for the root filesystem. 1196106001 M * daniel_hozac i.e. when you build a guest. 1196106052 M * Bertl (with the proper tools, that is :) 1196106075 M * er Bertl, what work in user-space? you're talking about the netlink proxy we'd discussed briefly? 1196106113 M * er to do root context stuff involving namespaces from a context without CAP_NET_ADMIN? 1196106114 M * daniel_hozac more like adding the necessary support to unshare the namespace, set it up, etc. 1196106122 M * Bertl no, mainly setting up the network namespaces and such 1196106132 M * DavidS cat /etc/fstab says "# UNCONFIGURED FSTAB FOR BASE SYSTEM", which was built with vserver ... build -m debootstrap from the etch debian packages 1196106157 M * daniel_hozac i guess debootstrap overwrote the utils' version then. 1196106318 J * gm ~randall@host207-15-dynamic.4-87-r.retail.telecomitalia.it 1196106325 M * Bertl wb gm! 1196106327 M * er Bertl, daniel_hozac, i see, so you're thinking of unsharing namespaces at vserver startup time? 1196106342 M * Bertl we already do that for mainline namespaces :) 1196106384 M * Bertl (like the IPC or UTS namespace) 1196106404 M * daniel_hozac so the ugly way would be to just add CLONE_NEWNET to vc_get_space_mask (i love that name, sounds so astronauty). 1196106413 M * er Bertl, i see. so when NETNS gets into mainline, you'd just have an option in the guest config to unshare that as well? 1196106435 M * Bertl kind of, only the setup needs (a lot of) tweaking 1196106472 M * er ah ok, to get the guest connected to the host network and so on 1196106490 M * Bertl yes, and ips assigned and routes configured, etc 1196106499 M * Bertl all stuff you do not have to do atm 1196106665 M * er ok. i'll talk to daniel about that some in that case, since he's right across the corridor :) 1196107368 M * Bertl good idea :) 1196107653 Q * sladen Server closed connection 1196107663 J * sladen paul@starsky.19inch.net 1196108107 J * AStorm ~astralsto@tor-irc.dnsbl.oftc.net 1196108275 J * Yvo ~yvonne@91.64.217.106 1196108289 P * Yvo 1196109138 J * opuk ~kupo@c213-100-138-228.swipnet.se 1196110365 J * Aiken ~james@ppp121-45-221-157.lns2.bne1.internode.on.net 1196110459 Q * derjohn Ping timeout: 480 seconds 1196110475 J * derjohn ~derjohn@dslb-084-059-018-045.pools.arcor-ip.net 1196110886 Q * er Quit: Leaving 1196110891 J * Darkglow ~pdesnoyer@208.71.184.41 1196110950 M * Darkglow Hi all. I am trying to fix xid tags on some vservers... what is the quickest procedure ? (I tried chxid -R -c then hashify... 1196110981 M * daniel_hozac chxid -URx -c 1196111000 M * Darkglow damn... is it possible that the max character name for a vserver is 10 ? 1196111010 M * daniel_hozac no. 1196111025 M * daniel_hozac but the max name for an alias is 10, assuming an interface name of 4 characters. 1196111076 M * Darkglow when running or not ? 1196111087 M * daniel_hozac what? 1196111099 M * Darkglow when the vserver is running or not ? 1196111111 M * daniel_hozac the chxid? 1196111113 M * Darkglow running chxid I mean. 1196111117 M * daniel_hozac doesn't matter. 1196111150 M * Darkglow hum... 1196111204 M * mark_ hello DH 1196111205 M * Darkglow I did that on both my vservers... now on the first one, I see some file with the correct xid and some with the xid of the other vserver... it should be 0 right ? 1196111226 M * daniel_hozac that means you didn't use the -U option. 1196111231 M * daniel_hozac and that you had unified files. 1196111247 M * Darkglow is it possible to recup from that situation ;-)? 1196111278 M * daniel_hozac sure, chxid -R -c 0 /vservers/.hash 1196111317 Q * rob-84x^ Remote host closed the connection 1196111351 M * mark_ http://paste.linux-vserver.org/9830 1196111378 J * rob-84x^ rob@submarine.ath.cx 1196111474 M * mark_ DH, remember that vhashify problem? 1196111550 M * mark_ I have vserver debug turned on in kernel if that's necessary 1196111617 M * Bertl Darkglow: yes, alias names are limited to 15 chars 1196111637 M * Bertl (including the interface prefix) 1196111895 M * daniel_hozac mark_: remove method and blocks, i doubt they do anything... 1196111898 M * Darkglow alias as in network interface alias right ? not related to vserver name? 1196111907 M * mark_ ok 1196111935 M * Bertl Darkglow: yep 1196112142 M * mark_ DH: now, some unified files were found, but vhashify on one of vservers reports this: 1196112142 M * mark_ da2 (87.204.147.145) /etc/vservers/ti386/apps/vunify % vserver ti386 hashify 1196112142 M * mark_ vcontext: execvp("sh"): No such file or directory 1196112142 M * mark_ failed to determine configfiles 1196112198 M * daniel_hozac so your guest doesn't have sh installed. 1196112234 M * mark_ I can't even enter the damn thing 1196112241 M * mark_ vlogin: execvp(): No such file or directory 1196112255 M * daniel_hozac not unexpected, since you don't have a shell. 1196112279 M * mark_ in /bin it does have bash and 'sh' symlink to bash 1196112297 M * daniel_hozac check the libraries... 1196112332 M * mark_ DH: correct, libncurses.so.5 => not found 1196112423 M * mark_ vcontext: execvp("/etc/init.d/rc"): No such file or directory 1196112462 M * mark_ oh it's probably bash, right? scripts throw this kind of error when interpreter is not found 1196112499 M * daniel_hozac yes. 1196112540 M * mark_ anyway, what's more important is that very few links were found after hashify of two vservers, just 4 links 1196112547 M * mark_ is this normal or what? 1196112574 M * Darkglow ok, here's what I did : copy my 2 vservers to new names. remove old , then change names. Then I deleted .hash directory. I then ran chxid -URx -c xid path on both. on the first one, /etc/mailname is with the correct context. The other one has context 0... Any clues 1196112578 M * mark_ du -hs . in /home/.hash reports over 400M 1196112617 M * Darkglow oups, forgot to say I styarted the vservers and ran hashify after the chxids 1196112641 M * daniel_hozac mark_: what distributions? 1196112658 M * mark_ Debian Etch 64bit 4.0R1 1196112673 M * Darkglow do you use the backport for the utils ? 1196112674 M * mark_ Those two forced into 32bit mode according to FAQ 1196112677 M * mark_ yep 1196112678 M * daniel_hozac and how do you determine that only 4 links were found? 1196112693 M * mark_ da2 (87.204.147.145) /home/.hash % find . -type f -links 1 1196112734 M * mark_ Darkglow: I mean yes, I use backports 1196112754 M * Darkglow (just checkin :-) 1196112768 M * daniel_hozac mark_: what's that supposed to do? 1196112796 M * mark_ that's what FAQ says to find unused links 1196112803 M * mark_ find /vservers/.hash -type f -links 1 -print0 | xargs -0 rm 1196112803 M * mark_ Until you do this, the files still take up place even though no vservers need them. 1196112823 M * Darkglow it's the cleanup line for unused links in .hash... 1196112825 M * daniel_hozac yes, and after that, you only have 4 files left? 1196112828 M * Darkglow I have that too. 1196112832 M * daniel_hozac i know what it is, i wrote it ;) 1196112838 M * mark_ no I didn't delete those files 1196112840 M * Darkglow hehe 1196112845 M * Darkglow sorry about that ;-) 1196112858 M * mark_ I just checked how many files of this kind it finds 1196112869 M * daniel_hozac and it was every file but four? 1196112881 M * mark_ No, it found JUST 4 such links 1196112897 M * mark_ I can paste it in the pastebin if you like 1196112898 M * daniel_hozac that's a _good_ thing... 1196112912 M * daniel_hozac that means you only have 4 files in the hash that aren't used. 1196112939 M * daniel_hozac try find /vservers/.hash -type f -links 3 -print, if you have 2 guests. 1196112948 M * daniel_hozac that'll show you the files which are unified across both guests. 1196112949 M * mark_ meaning? I expected it would find a lot of such links - on the premise that those two guests are almost the same 1196112993 M * mark_ DH: it found milionz and milonz :-) 1196112996 J * shuri ~shuri@64.235.209.226 1196113010 M * daniel_hozac so, there you go. 1196113014 M * daniel_hozac those are the unified files. 1196113027 M * Darkglow if I start with a non-tagged v1 and v2 and an empty .hash... what are the steps to unify ? 1196113036 M * daniel_hozac the other 4 files are ones that were at one point unified, but both guests have since diverged and no longer have that file. 1196113037 M * mark_ OK, bc I don't get smth: do you mean that while setting the vserver up (vserver build...) those files are created as unified to begin with? 1196113047 M * daniel_hozac no. 1196113054 M * daniel_hozac hashification is a separate step. 1196113081 M * mark_ OK, so does the hashify op delete unused links? 1196113099 M * mark_ I thought it leaves them around in .hash 1196113100 M * daniel_hozac what do you mean when you say unused links? 1196113127 M * mark_ that unified file is created, and so some of unused copies of files are left around 1196113137 M * mark_ and they need to be cleaned up in .hash 1196113195 M * Darkglow looks like the unify process cannot unify my files correctly for some reason... on 1 of my 2 vservers... :( 1196113196 M * daniel_hozac right. 1196113285 Q * tanjix Ping timeout: 480 seconds 1196113307 M * mark_ OK, so if the two guests weren't unified, and all had their separate copies of files, vhashify should create a lot of unified files and leave around a lot of unused files around, right? or wrong? 1196113342 M * daniel_hozac no. 1196113390 M * Darkglow It will create a copy in .hash and a link in your vservers. (taking space for 1 file only). and leave your other files as is on the file system. 1196113397 M * Darkglow (right ?) ;-) 1196113482 M * Darkglow doing a chxid -Rx would give unlink all files for a particular vserver right ? 1196113532 M * mark_ Darkglow: I've deleted /home/.hash a few times and files were still there all right, so unless copy-on-write copied them back into fs on unlink, they couldn't have survived this op 1196113561 M * Darkglow I admit, I'm not sure I understand why that is :( 1196113593 M * Darkglow In my understanding, it should not have worked... except is cow did what you just said.. 1196113642 M * mark_ I admit I'm kind of in the dark when it comes to details of what _exactly_ hashify operation does 1196113653 M * daniel_hozac unused files are ones that no guest no longer links to. 1196113681 M * daniel_hozac i.e. you unify A, B and C. all three have a file called /usr/share/X that is identical among the three. 1196113701 M * Darkglow how does it decide not to link some files ? 1196113711 M * mark_ i.e. they share the hard link? 1196113719 M * mark_ no, wrong 1196113725 M * daniel_hozac yes, their hard links point to the same inode. 1196113727 M * mark_ there are 3 hard links, right? 1196113748 M * daniel_hozac so, A, B and C are subsequently updated, getting a new /usr/share/X file. 1196113748 M * Darkglow if you delete the files in .hash, you lose that inode ? 1196113812 Q * sid3windr Server closed connection 1196113814 J * sid3windr luser@bastard-operator.from-hell.be 1196113838 M * daniel_hozac at this point, the file in .hash isn't linked to anymore. 1196113850 M * daniel_hozac so it's down to a link count of 1. 1196113853 M * daniel_hozac this is an unused file. 1196113887 M * mark_ ok but what if I delete the file in .hash with link count of, say, 2 1196113887 M * Darkglow for a hard link to work, it point to a real file right ? and these files are in /vservers/.hash 1196113921 M * daniel_hozac there's no such thing as a "real file" with hardlinks. 1196113938 M * daniel_hozac there's an inode, and a bunch of references to it (hardlinks) 1196113960 M * daniel_hozac all entries in your filesystem have an inode. 1196113966 M * Darkglow Ok I get it. 1196114006 Q * AStorm Remote host closed the connection 1196114046 M * mark_ thanks, DH 1196114051 J * AStorm ~astralsto@tor-irc.dnsbl.oftc.net 1196114053 M * mark_ GTG, see you around 1196114058 M * Darkglow bye 1196114067 Q * mark_ Quit: Oops. My brain just hit a bad sector 1196114111 M * Darkglow now, I still have an issue here ;-) hashify does things the wrong way I think... what is the sequence to unify starting from empty .hash and freshly copied vservers ? 1196114146 M * Darkglow (It works great on many other servers (hosts). Only on that one with 2 vservers that I have problems... 1196114180 M * daniel_hozac vsrever X hashify; vserver Y hashify 1196114188 M * daniel_hozac (with correct spelling) 1196114212 M * Darkglow no chxid before that ? 1196114227 M * Darkglow I start with empty .hash and all files in xid 0 ? 1196114254 M * daniel_hozac if you want to chxid too, do that first. 1196114276 M * Darkglow lsxid -R v{1,2} shows xid 0 for all files when I use a fresh copy. 1196114345 M * Darkglow Ok, so I really have an oddity here... (I think). I did chxid -URx -c id for both, then ran unify on both. 1196114371 M * Darkglow v1 is ok. v2 is not. For example the file /etc/mailname is still xid 0 1196114433 M * Darkglow 0 shop-tomcat4/usr/share/man/man5/mailname.5.gz 1196114434 M * Darkglow shop-tomcat4 shop-tomcat4/etc/mailname 1196114476 M * Darkglow this one is ok... context is shop-tomcat4 for /etc/mailname 1196114480 M * daniel_hozac maybe one of them reports /etc/mailname is a configuration file (and thus excluded from hashification), while the other does not. 1196114506 M * Darkglow they are the same kind of server.... where does is decide what to hashify or not ? 1196114552 M * Darkglow both etch debootstraps 1196114565 Q * Vudumen Server closed connection 1196114566 J * Vudumen ~vudumen@perverz.hu 1196114604 M * daniel_hozac same packages installed? 1196114614 M * Darkglow 98% identical 1196114623 M * Darkglow 1 has tomcat 4 and the other tomcat 5 :-) 1196114654 M * daniel_hozac and they both depend on the exact same set of packages? 1196114663 M * Darkglow yes. 1196114694 M * Darkglow can I "see" what hasify thinks should be considered as a config ? 1196114734 M * daniel_hozac sure, just add -vvvvv and you'll see how it inspects all the files. 1196114735 Q * DavidS Quit: Leaving. 1196114802 M * PowerKe I'm having problems running an ftp client in a guest using active mode. When I do 'ls', I get: ftp: bind: Address already in use 1196114833 M * Darkglow I see the list and I see it is excluded from v1 and not from v2... 1196114914 M * Darkglow I don't see "where" or "how" it determines what is a config file (I see "Fetching configuration-file list from packagemanagement") 1196114951 M * Bertl PowerKe: maybe the port it tries to use is already bound on the host? (or another guest)? 1196114978 M * PowerKe I have the same problem on 2 guests on different hosts 1196114989 J * tanjix ~tanjix@dslb-084-058-011-001.pools.arcor-ip.net 1196115012 M * PowerKe bind(4, {sa_family=AF_INET, sin_port=htons(35178), sin_addr=inet_addr("172.16.0.17")}, 28) = -1 EADDRINUSE 1196115035 M * daniel_hozac Darkglow: /usr/lib*/util-vserver/vpkg get-conffiles 1196115037 M * PowerKe I'm guessing the 35178 is the port it's trying to bind to? 1196115044 M * daniel_hozac yes. 1196115045 M * Bertl yep, what does lsof tell you? 1196115062 M * Bertl PowerKe: and what kernel/patches do you use? 1196115079 M * PowerKe Bertl: 2.6.20-vs2.2.0-gentoo 1196115089 M * Darkglow DH : Unsupported tag 'get-configfiles' 1196115108 M * Darkglow typo... 1196115108 M * Darkglow sorry 1196115136 M * Bertl PowerKe: we fixed quite a number of issues since vs2.2.0 (if that is the version used there) 1196115145 M * Darkglow v1 has 327 config files. v2 has 254... :( 1196115147 M * PowerKe Should I run lsof on the host or the guest? 1196115175 M * Bertl best would be in network context 1 1196115433 Q * dna Quit: Verlassend 1196115454 M * PowerKe I have no idea what to look for in the lsof output 1196115486 M * Bertl for the beforementioned port 1196115543 M * PowerKe It doesn't show anything, and trying again gives another port number, but the same problem 1196115590 M * PowerKe Sorry, I mean, searching for the port number doesn't show anything 1196115658 M * PowerKe I also never had any problems running web services in the guests, the ftp client is the first program that seems to give trouble 1196115667 M * Bertl well, personally I'd suggest to try with vs2.2.0.5 and see if it works or not 1196115744 M * Bertl if you can recreate the issue there, we have a bug to fix ... 1196115783 M * PowerKe Ok, I'll unmask it and upgrade to 2.2.0.5 1196115805 M * daniel_hozac is it true for _all_ FTP clients? 1196115807 J * TheSeer ~theseer@e177148092.adsl.alicedsl.de 1196115810 M * TheSeer re... 1196115820 Q * cohan Server closed connection 1196115821 J * cohan ~cohan@koniczek.de 1196115823 M * Bertl wb TheSeer! 1196115837 M * PowerKe I just emerged 'net-ftp/ftp-0.17-r6', didn't try any others yet 1196115840 M * PowerKe Any suggestions? 1196115870 M * daniel_hozac lftp? 1196115884 M * PowerKe will try 1196115906 M * daniel_hozac do you have the strace somewhere? does it bind multiple times? 1196115928 M * PowerKe It seems to try and fail twice 1196115932 M * Darkglow your command /usr/lib*/util-vserver/vpkg get-conffiles does not get ALL config files... for example /etc/mailname is not there... but 1 vserver says it's a config file and the other not... :( (All my other servers (host/vservers) do not have this issue... I will try installing a new vserver to see what the new one sees... 1196115962 M * PowerKe No, that's not right... I'll post it on pastebin 1196116031 M * PowerKe http://paste.linux-vserver.org/9835 1196116071 Q * Radiance Server closed connection 1196116084 J * Radiance 78af56ead3@193.16.154.187 1196116104 M * TheSeer dang.. i'm too stupid to use the simpelst things :/ 1196116117 M * TheSeer how do i force a 32bit guest on a 64bit host? 1196116127 M * TheSeer export ARCH=i386 doesn't seem to have any effect :/ 1196116198 M * daniel_hozac -- --arch i386 for debootstrap. 1196116222 M * PowerKe any idea how I verify that lftp is in active mode ? 1196116222 M * daniel_hozac linux32 for yum-based installs. 1196116234 M * TheSeer ouhm.. 1196116240 M * TheSeer linux32 vserver build .... ? 1196116253 M * PowerKe nvm 1196116263 M * daniel_hozac yes. 1196116268 M * TheSeer okaaay... 1196116278 M * TheSeer that would have been to easy.. as again :/ 1196116296 M * TheSeer sometimes i wonder why i actually got up today.. i do the simpelst things wrong today.. 1196116329 M * TheSeer btw... 1196116336 M * TheSeer i keep getting those rpm-fake-resolver: vc_ctx_migrate(): No such process 1196116336 M * TheSeer 1196116341 M * TheSeer again .. 1196116355 M * daniel_hozac interesting. 1196116361 M * daniel_hozac it must be a lunar cycle thing. 1196116363 M * TheSeer retrying a couple of times sort of fixes it 1196116498 M * PowerKe lftp works 1196116538 M * daniel_hozac TheSeer: do you specify a context id when you build the guest? 1196116575 M * TheSeer daniel_hozac: doesn't make a difference.. i get that with or without it 1196116684 P * Darkglow Konversation terminated! 1196116810 Q * larsivi Quit: Konversation terminated! 1196116824 Q * blizz Server closed connection 1196116824 J * blizz_ ~stephan@evilhackerdu.de 1196117561 M * PowerKe 2.2.0.5 doesn't fix the problem with ftp 1196117600 M * daniel_hozac could you upload an strace -e bind -Ff ftp...? 1196117637 M * PowerKe ok 1196117839 M * TheSeer daniel_hozac: [root@chase vservers]# vserver fc6-32 pkgmgmt internalize 1196117839 M * TheSeer rpm-fake-resolver: vc_ctx_migrate(): No such process 1196117851 M * TheSeer something is really strange ;) 1196117860 M * daniel_hozac indeed... 1196117902 M * PowerKe daniel_hozac: http://paste.linux-vserver.org/9837 1196117967 Q * hparker Ping timeout: 480 seconds 1196118019 M * PowerKe If it's too much trouble, I don't mind using another ftp client. 1196118079 Q * arekm Server closed connection 1196118082 J * arekm arekm@carme.pld-linux.org 1196118084 M * daniel_hozac i don't really see anything wrong. 1196118130 M * PowerKe Well, when I use active ftp, I don't get a directory listing but only: 500 Illegal PORT command 1196118133 M * PowerKe ftp: bind: Address already in use 1196118144 M * daniel_hozac yeah, i see that. 1196118147 M * daniel_hozac but i don't see why. 1196118173 M * Bertl we have a bind with port 0 here 1196118575 M * Bertl but I would really appreciate to see the full trace :) 1196118585 M * PowerKe sec 1196118602 J * hparker ~hparker@linux.homershut.net 1196118716 M * PowerKe http://www.hacks.be/ftp_strace 1196118876 M * Bertl does it happen always or just every now and then? 1196118910 M * Bertl and do you use S/DNAT on the host? 1196118918 M * PowerKe It's the first time I've tried it, on 2 different hosts and guests and it fails every time 1196118938 M * Bertl it also fails on the host? 1196118939 M * PowerKe lftp seems to work with set ftp:passive-mode false 1196118942 M * PowerKe it works on the host 1196118956 M * Bertl could you get me an strace -fF from the host too? 1196118960 M * PowerKe But the host has both private and public ip's 1196119073 M * PowerKe Also, nf_conntrack_ftp isn't loaded automatically and lftp fails if I don't load it, so I'm quite sure it's using active ftp 1196119083 Q * Guy- Server closed connection 1196119090 J * Guy- ~korn@elan.rulez.org 1196119114 M * Bertl PowerKe: so you need the ftp connection tracking to get that error? 1196119137 M * PowerKe No, I get the error regardless. 1196119169 M * Bertl okay, let's take one step back and revalidate the data 1196119181 M * Bertl you are running an ftp server, yes? 1196119187 M * PowerKe http://www.hacks.be/ftp_strace_host 1196119195 M * PowerKe No, just using a public ftp server on the web 1196119218 M * Bertl okay, so you are using lftp to access a public server from a guest and from the host 1196119244 M * PowerKe I was using ftp, which works from the host, but fails from the guest (unless in passive mode) 1196119245 M * Bertl the guest access (with the same tool) doesn't work, while the host access does 1196119261 M * PowerKe Since that didn't work, daniel_hozac suggested to use lftp which does work from the guest (didn't try from the host) 1196119270 M * Bertl guest and host have been using the same ip for those tests? 1196119311 M * PowerKe no, the host has a public ip + 172.16.8.1 and the guest only has 172.16.8.11 1196119330 M * Bertl okay, so you are comparing apples and oranges here 1196119344 M * Bertl in the host case (public ip) no nat is necessary 1196119360 M * Bertl in the guest case (with private ip) you have to do some nat-ing 1196119385 M * Bertl this doesn't play well with your connection tracking setup, as it seems 1196119386 M * PowerKe true, which works for lftp but not for ftp 1196119417 M * Bertl I would suggest to try two things: 1196119432 M * Bertl a) provide a public ip to a test guest (see if that works) 1196119453 M * Bertl b) use/force the host to use the private ip and go through the nat 1196119485 M * TheSeer daniel_hozac: uname -r 2.6.22.10-vs2.3.0.29.1 1196119498 M * TheSeer daniel_hozac: that's the latest of your repo isn't it? 1196119571 M * daniel_hozac yes. 1196119640 M * daniel_hozac why? 1196119652 M * PowerKe Bertl: Using a public ip on the guest works 1196119658 M * TheSeer i really wonder where that instability comes from 1196119718 M * daniel_hozac me too. 1196119725 M * daniel_hozac i don't think it should be possible. 1196119743 M * Bertl PowerKe: now if b) fails (if that can be done to ftp) then we know it is purely nat related 1196119761 Q * ensc Ping timeout: 480 seconds 1196119806 M * PowerKe Bertl: would chbind work? 1196119834 J * ensc ~irc-ensc@p54B4D376.dip.t-dialin.net 1196119892 M * PowerKe hmm: chbind --ip 172.16.8.1 ftp ftp.belnet.be --> ncontext: vc_net_create(): Invalid argument 1196119910 M * daniel_hozac --nid 42 1196119941 M * PowerKe That gives the bind error on the host 1196119962 M * daniel_hozac the best way would be to instruct your client to use a different bind address though. 1196120013 M * TheSeer [root@chase etc]# linux32 vserver fc6-32 build --force -m yum --context 3206 --hostname=fc6-32.caska.netpirates.net --interface fc6-32=dummy0:10.32.0.6/24 -- -d fc6 1196120014 M * TheSeer rpm-fake-resolver: vc_ctx_migrate(): No such process 1196120014 M * TheSeer rpm-fake.so: failed to initialize communication with resolver 1196120014 M * TheSeer rm: cannot remove directory `/etc/vservers/.defaults/vdirbase/fc6-32/.rpmdb': Device or resource busy 1196120014 M * TheSeer rm: cannot remove directory `/etc/vservers/.defaults/vdirbase/fc6-32/proc': Device or resource busy 1196120028 M * daniel_hozac TheSeer: so you can reproduce this easily, right? 1196120035 M * TheSeer yes, 1196120056 M * PowerKe I don't see any options in the man page to bind to a specific ip 1196120057 M * TheSeer the scary thing is that leaves an fc6-32 dir which i cannot even remove because "device is busy" 1196120072 M * daniel_hozac does not using linux32 change the percentage at all? 1196120086 M * daniel_hozac i.e. the success/failure ratio. 1196120098 M * daniel_hozac that's due to a separate bug that i have not yet tracked down 1196120113 M * TheSeer nope 1196120114 M * daniel_hozac a namespace is being leaked somewhere when using rpm-fake. 1196120130 M * daniel_hozac (it's on my TODO) 1196120141 M * TheSeer as i had that very problem before and just started using linux32 for i need to build a 32bit guest 1196120162 M * daniel_hozac okay. 1196120171 M * daniel_hozac and nothing is using context 3206, right? 1196120226 M * TheSeer i freshly booted the box, no vserver is running and it is a 101% never been used ctx id 1196120249 M * daniel_hozac okay. 1196120268 M * daniel_hozac do you have a util-vserver source tree that you can easily build? 1196120275 M * TheSeer not yet 1196120288 M * TheSeer it's centos5 setup along with your repo 1196120308 M * daniel_hozac okay, so get the prereqs and a util-vserver tarball. 1196120320 Q * ensc Ping timeout: 480 seconds 1196120390 M * TheSeer any particular version to use? 1196120406 M * daniel_hozac 0.30.214 1196120622 M * TheSeer any particular configure option? 1196120677 M * daniel_hozac just --prefix=/usr --sysconfdir=/etc --localstatedir=/var 1196120688 M * TheSeer k.. 1196120688 M * daniel_hozac oh, and since it's 64-bit, --libdir=/usr/lib64 1196120839 M * daniel_hozac and before building, run curl http://people.linux-vserver.org/~dhozac/p/uv/experimental/rpm-fake-debug.patch | patch -p0 1196121005 M * TheSeer k.. working on some compile rpms first.. 1196121038 M * daniel_hozac oh, and you should enable the switch debugging too. 1196121088 M * TheSeer --enable-lib-debug? 1196121091 M * daniel_hozac echo 2 > /proc/sys/vserver/debug_switch 1196121096 M * TheSeer ah 1196121124 M * daniel_hozac that should hopefully get us information to figure out what is going on... 1196121137 M * daniel_hozac +enough 1196121227 J * ensc ~irc-ensc@77.235.182.26 1196121309 M * TheSeer okay.. patch applied 1196121309 M * TheSeer building now 1196121418 M * TheSeer okay.. completed and isntalled 1196121421 M * TheSeer debug switch enabled