1229563510 M * maharaja m_o_d: look into the .rej files 1229563521 M * maharaja and into the original files 1229563526 M * maharaja e.g. kernel/sched.c 1229563532 M * maharaja thos thines *might* be trivial to fix 1229563538 M * maharaja g'nite 1229564807 J * derjohn_mob ~aj@e180199232.adsl.alicedsl.de 1229565443 J * arapaho_ ~arapaho@213.223.114.206 1229565563 Q * arapaho Ping timeout: 480 seconds 1229569926 J * balbir_ ~balbir@122.167.212.247 1229571515 Q * dowdle Remote host closed the connection 1229575079 Q * derjohn_mob Ping timeout: 480 seconds 1229578324 Q * balbir_ Ping timeout: 480 seconds 1229578681 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1229578866 J * hparker ~hparker@linux.homershut.net 1229584350 J * balbir_ ~balbir@59.145.136.1 1229585094 J * mtg ~mtg@vollkornmail.dbk-nb.de 1229585420 J * kir ~kir@swsoft-msk-nat.sw.ru 1229586106 J * ktwilight_ ~ktwilight@214.102-66-87.adsl-dyn.isp.belgacom.be 1229586382 Q * ktwilight__ Ping timeout: 480 seconds 1229586496 J * dna ~dna@77-207-103-86.dynamic.dsl.tng.de 1229587133 N * pmenier_off pmenier 1229589484 N * arapaho_ arapaho 1229591191 J * chi6IT41 ~chigital@services.mivitec.net 1229592710 M * derjohn m_o_d, you can also try patch --fuzz=100 ... but I doubt it will be trivial to fix. 1229592751 M * pmjdebruijn derjohn: --fuzz=100? 1229592782 M * pmjdebruijn does even even do anything, I've never found anything above 4-5 useful 1229592887 M * derjohn pmjdebruijn, oh, wen merging patches I used up to fuzz 20 or so, if some patch introdues a new additionla function with 10 or 20 lines before the planned insertion point, it might help 1229592926 M * pmjdebruijn oh ok cool, good to know... 1229592955 M * derjohn but as maharaja already said: I doubt it will be trivial, as vserver heavily moddes the scheduler and RT linux also does so 1229592968 M * pmjdebruijn m_o_d: do you really require RT? 1229592976 M * pmjdebruijn m_o_d: seems like an odd thing to combine with vserver? 1229592986 M * pmjdebruijn m_o_d: what do you want to archive? 1229592992 M * pmjdebruijn achieve even... 1229593136 N * morrigan_zZ morrigan_uk 1229594122 J * whuji ~huji@jem75-8-88-170-103-188.fbx.proxad.net 1229594131 M * whuji hello ! 1229594197 M * pmjdebruijn lo 1229594257 M * whuji fine ? 1229594360 M * whuji I've some questions about vserver 1229594431 M * ktwilight_ whuji, go ahead 1229594446 M * whuji I'm trying to install dtc-postfix-courier. I've this error : 1229594447 M * whuji SIOCSIFADDR: Permission non accordée 1229594447 M * whuji bin: ERROR while getting interface flags: Aucun périphérique de ce type 1229594451 M * whuji so in english : 1229594471 M * whuji SIOCSIFADDR: Permission denied 1229594471 M * whuji bin: ERROR while getting interface flags: no device found 1229594478 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1229594488 M * whuji how can I fix it ? 1229594492 M * pmjdebruijn whuji: what does ifconfig tell you? 1229594508 M * pmjdebruijn whuji: does courier bind to the specific ip? or just 0.0.0.0 1229594517 M * whuji eth1 Lien encap:Ethernet HWaddr 00:15:17:27:E2:44 1229594517 M * whuji inet adr:88.191.91.19 Bcast:88.191.91.255 Masque:255.255.255.0 1229594517 M * whuji UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 1229594517 M * whuji RX packets:2688274 errors:0 dropped:359 overruns:0 frame:0 1229594517 M * whuji TX packets:626223 errors:0 dropped:0 overruns:0 carrier:0 1229594517 M * pmjdebruijn whuji: trying having it bind to your vserver specific ip 1229594518 M * whuji collisions:0 lg file transmission:1000 1229594518 M * whuji RX bytes:874679758 (834.1 MiB) TX bytes:100519327 (95.8 MiB) 1229594520 M * whuji lo Lien encap:Boucle locale 1229594520 M * whuji inet adr:127.0.0.1 Masque:255.0.0.0 1229594522 M * whuji UP LOOPBACK RUNNING MTU:16436 Metric:1 1229594522 M * whuji RX packets:2911993 errors:0 dropped:0 overruns:0 frame:0 1229594524 M * whuji TX packets:2911993 errors:0 dropped:0 overruns:0 carrier:0 1229594524 M * whuji collisions:0 lg file transmission:0 1229594526 M * whuji RX bytes:689982687 (658.0 MiB) TX bytes:689982687 (658.0 MiB) 1229594536 M * pmjdebruijn whuji: please use pastebin for such large pasten 1229594540 M * pmjdebruijn pastes* 1229594541 M * whuji sorry 1229594602 M * whuji where can I find pastebin ? 1229594607 M * whuji oups 1229594607 M * pmjdebruijn pastebin.com ? 1229594608 M * whuji found 1229594633 M * Supaplex google paste bin and you'll find a horde of them. :P 1229594642 M * pmjdebruijn whuji: btw, I'd recommmend Dovecot over Courier :) 1229594647 M * pmjdebruijn but that's another matter 1229594650 M * whuji or just paste.linux-vserver.org 1229594651 M * Supaplex it's like every php programmer kiddies dream to make their own paste site. 1229594699 M * whuji pmjdebruijn : I use DTC and courier is fully supported by default 1229594829 M * whuji in /etc/courier/imapd it's written ADRESS=0 1229594852 M * pmjdebruijn whuji: I'm not familiar with imap 1229594875 M * pmjdebruijn whuji: I'm sure the docs can tell you how to bind courier to 88.191.91.19 1229594967 M * pmjdebruijn that least that's my first guess what _could_ be a problem... 1229595085 M * whuji ok. Because courier is already installed and seems to work well 1229595136 M * morrigan_uk is this "normal" postfix? 1229595148 M * whuji yes. 1229595162 M * whuji DTC is a web and mail hosting pannel 1229595176 M * whuji but DTC uses normal postfix, courier, bind and apache 1229595196 M * morrigan_uk maybe this still applies: http://www.postfix.org/postconf.5.html#smtp_bind_address 1229595326 M * whuji still the error 1229595449 M * whuji ok now I have this : 1229595480 M * whuji authdaemonrc: ERROR while getting interface flags: no media of this type 1229595608 M * morrigan_uk anything interesting in dmesg or such? 1229595683 M * morrigan_uk i wonder whether it really tries to use eth1 or whether something there tries eth0 by default... 1229595687 M * derjohn Hello morrigan_uk .... are you in "UK" now ? 1229595702 M * morrigan_uk derjohn: yep :) BEA weblogic admin training :) 1229595793 M * whuji nothing in dmesg... 1229595907 M * whuji the installation works but there still is this error. Does it a problem ? 1229596604 M * whuji and I can't login in proftpd 1229596883 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net 1229597751 A * arekm is trying to sniff vserver "loopback" traffic on host. How 127.x.x.x are mapped per vserver? 1229597781 M * arekm for example I see 127.3.238.1 ip being used, how I know to which guest it belongs? 1229598574 M * ktwilight_ is it possible to insert a flag/caps while the guest is running? 1229598580 M * ktwilight_ or wouldi have to restart the guest? 1229598608 M * ktwilight_ arekm, check /etc/vservers/guest/interfaces/0/ip? 1229599118 M * arekm ktwilight_: vs2.3, loopback are there without need for interfaces/* 1229599147 M * ktwilight_ right 1229599159 M * ktwilight_ i was refering to how you know which ip is being used in which guest. 1229599203 M * arekm and what you say doesn't apply to loopback IP 1229599227 M * ktwilight_ of course 1229599243 A * ktwilight_ must have misunderstood that question 1229599255 Q * independence Ping timeout: 480 seconds 1229599280 M * arekm heh, traffic in quest to 127.0.0.1, which IP should I sniff on host to see that traffic? 1229600212 M * mnemoc i think you can't sniff that 1229600325 M * mnemoc afaik one sniffed interfaces, not IPs 1229600331 M * mnemoc sniffes* 1229600436 J * independence independen@titan.blinkenshell.org 1229600937 M * arekm mnemoc: I can filter per ip when sniffing on a interface 1229600967 M * arekm and this traffic is there, just I don't know what's 127.x.x.x in guest -> 127.y.y.y on host relation 1229602017 J * doener ~doener@i577BB746.versanet.de 1229602434 M * bonbons arekm: not sure, but check 127..0.1 or 127.0..1 (assuming your NID < 256 - otherwise probably a matter or getting low and high bytes in the right order) 1229602819 A * mnemoc didn't know lo was sniffable 1229602823 M * whuji is it possible to give a read only access to /dev and /proc to a guest ? 1229602828 Q * Aiken Remote host closed the connection 1229602869 M * mnemoc whuji: /dev, use bind mount, /proc is limited by the capabilities and visibility rules 1229602986 M * whuji mnemoc : to configure in the /etc/vservers/guest directory ? 1229604354 M * mnemoc /dev in /etc/vservers/guest/fstab should work 1229604446 M * whuji yep, but for de /proc, I need to use /etc/vservsers/guest/capabilities ? 1229604473 M * mnemoc no, /proc is restricted by default 1229605224 M * whuji ok 1229605241 M * whuji so it's impossible 1229605293 M * mnemoc what in /proc is giving you troubles? 1229605656 J * larsivi ~larsivi@92.62.36.7 1229606062 M * whuji I just need to get some infos like cpu infos or battery 1229606204 M * mnemoc whuji: look at `vprocunhide` for details about how to unhide other proc entries 1229606271 M * whuji ok I'll take a look after lunch. 1229608979 Q * larsivi Remote host closed the connection 1229609698 Q * mtg Read error: Connection reset by peer 1229609870 Q * kiorky Ping timeout: 480 seconds 1229610069 J * kiorky ~kiorky@cryptelium.net 1229610105 J * derjohn_mob aj@p57A6E587.dip.t-dialin.net 1229610520 Q * balbir_ Ping timeout: 480 seconds 1229611183 N * Bertl_zZ Bertl 1229611196 M * Bertl morning folks! 1229612681 M * maharaja hi there bertl 1229612873 M * Bertl whuji: a ro bind mount for /proc should be possible too :) 1229612893 M * whuji oh thanks :) 1229612921 M * Bertl but you won't get entries which are hidden (by the config) 1229613523 J * mtg ~mtg@dialbs-088-079-143-204.static.arcor-ip.net 1229613785 M * matti Bertl: ;]]] 1229615488 Q * jaqque Ping timeout: 480 seconds 1229615522 J * jaqque foobar@buddha.sbih.org 1229615756 Q * derjohn_mob Ping timeout: 480 seconds 1229616711 Q * chi6IT41 Remote host closed the connection 1229616811 M * whuji I could see cpuinfo meminfo version ? 1229617957 M * Bertl hmm? 1229618031 M * Bertl /proc/cpuinfo and /proc/meminfo are visible by default, meminfo can be virtualized 1229618603 J * dowdle ~dowdle@scott.coe.montana.edu 1229618705 M * ktwilight_ hm, i have VIRT_MEM on, and set as.hard and rss.soft, when i do apt-get update, i get a "FATAL -> Failed to fork." if i remove as.hard, it's fine. as.hard and rss.soft are set at 512MB. am using 0.30.216-pre2827 1229618756 M * ktwilight_ am pretty sure apt-get update doesn't exceed as.hard's memory limit 1229618913 Q * whuji Quit: Leaving. 1229619051 M * Bertl well, check with /proc/virtual//limits 1229619068 M * Bertl i.e. upload that for the guest, and we can take a look at it 1229619145 M * ktwilight_ gimme a mo 1229619227 M * ktwilight_ hm, strange. now that i add as.hard back again, restart, and it just works no errors. 1229619252 M * ktwilight_ gonna paste anywayz 1229619286 M * ktwilight_ Bertl, http://rafb.net/p/IJZ1Hj57.html 1229619324 M * Bertl yeah, as you can see, you are quite close to the vm limit 1229619336 M * Bertl 121054 was the max observed 1229619342 M * ktwilight_ aha 1229619346 M * Bertl 131072 is the limit 1229619348 M * ktwilight_ i see, and current VM 103187 1229619366 M * Bertl so one more bash or so, and you're over the limit 1229619379 M * ktwilight_ then i'm not understanding as.* at all :/ 1229619394 M * Bertl summed up address space across processes 1229619405 M * ktwilight_ oh! i thought it's per process 1229619425 M * Bertl the per process limit can be set too 1229619439 M * Bertl (with the ulimit entry) 1229619453 M * ktwilight_ "The Address Space (as) is the total amount of memory (pages) mapped in each process in the context. " 1229619467 M * ktwilight_ that's what stated in the wiki, does that still hold? or... 1229619478 M * ktwilight_ i guess it should be "The Address Space (as) is the total amount of memory (pages) mapped in all the processes in the context." 1229619479 M * Bertl yes, but it is ambiguous, I guess 1229619493 M * ktwilight_ where s/each process/all the processes/ 1229619507 M * Bertl in _all_ the processes (summed together) not in _each_ process, for all processes :) 1229619525 M * ktwilight_ k, i'll make the changes 1229619533 M * Bertl fell free to rephrase that in an unambiguos way :) 1229619543 M * ktwilight_ :) 1229619681 M * ktwilight_ Bertl, whta's a good use case for as.soft and/or as.hard? 1229619812 M * SpComb if apt-get mmaps some 1GB database, then it'll hit a 512MB as limit pretty quickly 1229619944 M * ktwilight_ hm, generally speaking, it's not really interesting to use as.* since it's difficult to know what the mmaps will be, unless it's a very controlled environment. 1229620125 M * Bertl it is good to avoid DoS via java vm for example :) 1229620373 M * ktwilight_ oh 1229620621 M * faheem I'm using Debian. If I run postgres on the main system and in a vserver, these appear to know the other exists. 1229620629 M * faheem Any way of removing the problem? 1229620653 M * faheem The issue is that Debian autoshifts postgres to a different port if the preferred port is in use. 1229620865 J * larsivi ~larsivi@9.80-202-30.nextgentel.com 1229620946 M * nox faheem: don't bind to all interfaces in the host 1229620982 M * faheem nox: yes, apparently that is the default. Just localhost, then? 1229620990 J * balbir_ ~balbir@122.167.204.221 1229620990 M * faheem for both the vserver and the host? 1229620999 M * nox just not * 1229621014 M * nox since the host would also bind guests ip 1229621056 M * nox in the guest * is fine since he only can bind its IPs 1229621142 M * faheem nox: it is localhost for both, I think. 1229621165 M * nox which vs version? 1229621167 M * faheem the postgres people recommend setting them to specific ips. at least the guest, I suppose. 1229621198 M * faheem 0.30.216~r2772-5 (util-vserver) 1229621220 M * nox and debian kernel? 1229621241 M * faheem 2.6.26-1-vserver-amd64 1229621327 M * nox if it is not a security issue by you i would suggest to bind to a rfc1918 ip 1229621356 M * nox or limit access by the hosts's firewall 1229621400 M * nox lo is only a good option with 2.3 vservers. Others am i right? 1229621436 M * faheem nox: is a 2.3 vserver more or less recent than what I'm running? 1229621471 M * nox i would assume debian uses stable aka 2.2 1229621496 M * faheem nox: Ok. 1229621497 M * nox there are a lot of enhancements in 2.3 but it is experimental 1229621680 M * faheem what is the schedule for a 2.3 release? 1229621731 Q * hparker Quit: Quit 1229621737 J * hparker ~hparker@linux.homershut.net 1229621934 M * nox schedule? 1229622007 J * transacid_home ~moarlulz@e176010082.adsl.alicedsl.de 1229622035 M * transacid_home the debian vserver-utils problem hit me :( 1229622044 M * transacid_home no ssh :( 1229622114 M * transacid_home i guess there is no exploit to break out of a guest 1229622114 A * ard uses 2.3 for production 1229622159 M * faheem nox: yes, when it will become stable. 1229622185 M * faheem if I want to use 2.3, i'll have to patch my own kernel? 1229622217 A * ard always advices to compile kernels and not to use distro kernels 1229622229 Q * snooze Ping timeout: 480 seconds 1229622235 M * ard make-kpkg is really easy :-) 1229622241 M * faheem ard: Why? 1229622255 M * faheem yes, I know how to use make-kpkg. 1229622257 M * ard because you can tune the kernels to your need... 1229622272 M * faheem ard: upgrades are a pain, though. 1229622274 M * ard And at least you know what's in there. 1229622281 M * maharaja indeed 1229622286 M * faheem and you can't report bugs. 1229622288 M * ard upgrades are more easy, because you know what's in it 1229622299 M * ard you can, thats what linux-kernel is for :-) 1229622308 M * faheem ard: well, you can't just do apt-get upgrade. 1229622315 M * ard hehe :-) 1229622330 M * ard we have our own debian mirror :-) 1229622337 M * ard and our own repository 1229622347 M * ard so yes, we can do what we want ;-) 1229622350 M * faheem linux-kernel? 1229622356 M * ard the mailing list 1229622357 M * maharaja thou vanilla kernels might be broken sometimes 1229622357 M * faheem ard: Well, that's fancy. 1229622362 M * maharaja or too conservative 1229622367 M * faheem ard: oh, right. 1229622385 M * faheem ard: who do you work for? 1229622396 M * ard sometimes you are just working on features that are not in the vanilla kernel f.i. 1229622407 M * ard I currently do most of my work for telegraaf 1229622414 M * ard biggest newspaper company in the netherlands 1229622421 M * faheem ard: you're a consultant? 1229622426 M * nox ard: yes i use also use it everywhere, but had probs sometimes with it 1229622429 M * faheem ard: amsterdam? 1229622430 M * ard also streaming for classic.fm and such 1229622435 M * ard yes 1229622444 M * ard eh, yes, amsterdam 1229622461 M * faheem lots of free software people in that area, apparently. 1229622462 M * ard almost 9 years (with a pause of 1.5 years ;-) ) 1229622463 M * nox transacid_home: i have no idea what could help there 1229622486 M * maharaja for "bleeding edge hardware" (e.g. some stupid new revision of some realtek chipsets) you're bound to the most recent kernels ... that kind of bugs me 1229622506 M * Bertl transacid_home: no known exploit :) 1229622517 M * ard heh, new servers that have broadcom ethernet cards instead of intel... 1229622522 M * transacid_home i mean that's great, but uhmmm 1229622529 M * ard you suddenly know why you want to buy intel nics :-) 1229622573 M * transacid_home i dun have a rescue system. i have to call the hoster and let the server be manualy but into a rescue system 1229622581 M * transacid_home that gets pretty expensive 1229622614 M * Bertl so, what exactly is the problem, you lost sshd on the host? 1229622616 M * transacid_home micah: does the version in unstable fix the problem? 1229622622 M * ard transacid_home : always make sure you have serial console or something like that 1229622634 M * transacid_home Bertl: yes, apparenty due to the problem with util-vserver 1229622644 M * ard Bertl : I guess his vserver was the way to get onto his host 1229622652 M * transacid_home the daemon wouldn't start anymore 1229622657 M * Bertl hmm, util-vserver has no effect on the host's sshd, AFAIK? 1229622665 M * transacid_home hmm ok 1229622678 M * Bertl only on the guests, and only if you have an older version 1229622680 M * ard ah: 1229622700 M * ard transacid_home : Have you turned on both auto-loopback and auto-single-ip ? 1229622701 M * transacid_home when i did the update last night there were both updates for util-vserver and openssh-server 1229622714 M * ard no new kernel? 1229622715 M * nox Bertl: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=506938 1229622764 M * Bertl yep, that confirms what I was thinking 1229622766 M * ard nox : that's only in a vserver I guess 1229622775 M * transacid_home nox: is there a bugreport for the initial problem? 1229622784 M * transacid_home oh ok 1229622793 M * nox i am still not sure what was broken in the -4 version 1229622794 M * transacid_home then i dunno where the problem with my ssh daemon is 1229622823 M * Bertl nox: it didnt give the required capabilities, like chroot 1229622848 M * Bertl nox: with SYS_CHROOT missing, the priv separation stuff/chroot thingy fails 1229622850 M * ard transacid_home : you did an upgrade, and rebooted? Only util-vserver and openssh-server, no kernel? 1229622873 M * transacid_home ard: i build my own kernels 1229622878 Q * pmenier Quit: Konversation terminated! 1229622884 M * transacid_home and no i didn't reboot sshd just didn'T work anymore 1229622886 M * ard good: [18:51] transacid_home : Have you turned on both auto-loopback and auto-single-ip ? 1229622890 M * transacid_home then i try to restart 1229622899 M * transacid_home and sshd didn't come up anymore 1229622906 M * transacid_home i search for the strace 1229622914 M * transacid_home http://phpfi.com/389649 1229622915 M * transacid_home here 1229622935 M * nox Bertl: ic thx 1229622937 M * transacid_home ard: where to turn that of/on? 1229622962 M * transacid_home i didn't touch my /etc/ssh/sshd_config 1229623023 Q * kir Quit: Leaving. 1229623042 M * ard in your kernel compile 1229623064 M * ard CONFIG_VSERVER_AUTO_LBACK and CONFIG_VSERVER_AUTO_SINGLE 1229623068 M * transacid_home it's a 2.6.22.19-vs2.2.0.7 build on Apr 12 11:08:19 CEST 2008 1229623077 M * ard ah 1229623081 M * ard that's ok 1229623108 M * transacid_home i can't even lookup what config i use cause i can'T access the host 1229623114 M * transacid_home only have ssh to one guest 1229623120 M * ard did you configure sshd on the host to only listen to a certain ip? 1229623137 M * transacid_home i had it listen on several 1229623163 M * ard Hmmmm... I didn't know sshd was able to listen to serveral ip's.... 1229623172 M * transacid_home my inner (10.1.1.1) and my (outer) 62.141.52.87 1229623190 M * ard With Listen in the /etc/sshd/sshd_config ? 1229623194 M * transacid_home yap 1229623208 A * ard has to try that again then 1229623217 M * nox worx fine here 2 1229623244 A * ard just puts in in /etc/inetd.conf 1229623245 M * transacid_home hmm before i tell my hoster to start a rescue system i have to gather as much information as i can get 1229623255 M * ard except for the out-of-band interface 1229623258 M * transacid_home cause i really have no clue what i should look for 1229623300 M * ard I am guessing that some other process (vserver?) is listening on *:22 1229623312 M * ard sshd in a normal debian install gets started way to late 1229623333 M * transacid_home i haven'T reboot since april 1229623334 M * ard and there are weird restarts now in the networking scripts 1229623338 J * jla ~jacob@green.cachevalley.org 1229623340 M * transacid_home why should it occure now 1229623382 M * jla What are the pre/post start/stop scripts called? I've forgotten the name and am having a hard time finding the documentation 1229623394 M * transacid_home jla: rc? 1229623407 M * jla for a vserver, to do stuff like add ip route rules 1229623412 M * jla not /etc/rc.d 1229623425 M * jla /etc/vserver//run 1229623429 M * ard transacid_home : usually the order in which things are started. 1229623462 M * transacid_home but how should a guest interfere with the host? 1229623478 M * ard heh /etc/network/if-up.d also contains an init script for openssh-server :-( 1229623581 M * ard if you made a mistake in the assigning ip addresses to the vservers, or if you had enabled both CONFIG_VSERVER_AUTO_LBACK and CONFIG_VSERVER_AUTO_SINGLE 1229623599 M * ard the latter is not possible since you use 2.2 1229623609 M * Bertl transacid_home: it could be, e.g. if the guest bind an IP, the host's sshd is configured for, then that could bail out with an error 1229623632 M * Bertl (result of a misconfiguration, btw :) 1229623648 M * transacid_home Bertl: that is not the case 1229623649 M * ard inetd only barfs, and retries :-) 1229623674 M * ard eh 1229623675 M * ard barks 1229623692 Q * mtg Quit: Verlassend 1229623693 M * ard (subtle difference between running and not running ;-) ) 1229623735 M * ard too bad you don't have an strace -f :-( 1229623766 M * jla /etc/vserver//scripts pre-start; post start. 1229623772 M * jla later/good luck 1229623773 Q * jla Quit: leaving 1229623910 J * snooze ~o@1-1-4-40a.gkp.gbg.bostream.se 1229623929 M * transacid_home ard: if i had known ... :/ 1229623996 M * ard transacid_home : for next time: 1) never ever reboot a system if you can't get sshd to work 2) never ever logout until you have assured with a second session that sshd works ;-) 1229624018 M * transacid_home ard: logout as in 24h disconnect 1229624055 M * ard and if you have the money or means: make sure you have a serial console. 1229624067 M * ard These problems are moot if you have a serial console 1229624077 M * transacid_home that's true 1229624093 M * maharaja Bertl: i just verified the sshd problem with my current setup and i too have the -4 version. is the SYS_CHROOT problem compile-time related? i guess, so, right? 1229624134 M * Bertl well, it is a broken util-vserver version in debian 1229624153 M * Bertl upgrade to a newer version and you should be fine 1229624162 M * maharaja k 1229624168 M * nox which is in sid right now 1229624170 M * maharaja i thought that maybe its like the cron issue in vserver build 1229624170 M * nox -6 1229624179 M * Bertl micah can probably explain why that version is still in debian :) 1229624181 M * maharaja so it can be fixed by setting some options 1229624201 M * maharaja indeed 1229624204 M * maharaja micah: wake up ;) 1229624205 M * Bertl yes, you could specify all the proper capabilities manually for each guest 1229624251 M * Bertl CRITICAL **: gtk_pizza_set_size: assertion `pizza != NULL' failed ... hmm, should I understand VLC developers? :) 1229624267 M * Bertl (or maybe gtk developers :) 1229624289 M * maharaja shouldn't it be sufficcient to add SYS_CHROOT to bcapabilities *testing* 1229624417 M * nox lol Bertl 1229624564 M * maharaja ok, i guess not :) 1229624660 M * Bertl okay, off for now .. bbl (~1 hour) 1229624666 N * Bertl Bertl_oO 1229624805 M * maharaja while we're at it, ill add another patch to util-vserver 1229624812 M * maharaja which has bugged me quite some time 1229625437 M * maharaja http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=509134 1229625598 M * maharaja error writing /proc/self/oom_adj: Permission denied 1229625599 M * maharaja mhm 1229625602 M * maharaja what does this tell me? 1229625734 M * maharaja what does this tell me? 1229625738 M * maharaja ups, sorry 1229626424 Q * xdr Remote host closed the connection 1229627147 M * maharaja have you seen that: 1229627148 M * maharaja CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME 1229627148 M * maharaja 97 22 186.3M 14.5M 0m00s40 0m00s96 26m29s20 vserver1 1229627148 M * maharaja 106 5 140.9M 7.3M 0m23s89 0m09s72 44m59s10 1229627162 M * maharaja CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME 1229627162 M * maharaja 97 22 186.3M 14.5M 0m00s40 0m00s96 26m29s20 vserver1 1229627162 M * maharaja 106 5 140.9M 7.3M 0m23s89 0m09s72 44m59s10 1229627163 M * maharaja # cat /etc/vservers/lenny64/context 1229627163 M * maharaja 106 1229627171 M * maharaja sorry for the c/p error 1229627548 N * Bertl_oO Bertl 1229627578 M * Bertl maharaja: that tells you that the debian/ssh folks try to cheat on the oom killer :) 1229627629 M * arekm so you won't get sshd killed in oom condition 1229627644 M * Bertl but IIRC, that is in the initscript, so can easily be commented out 1229627655 M * arekm and vserver had a bug in that 1229627666 M * Bertl no, not really, it is just not allowed :) 1229627673 M * arekm that file was shown as writtable while it's not writtable 1229627701 M * arekm admit it, it's a bug :P 1229627702 M * Bertl most permission checks (capability based) are not shown in the read/write attributes 1229627721 M * Bertl I'm fine with declaring it a bug, but then it is a bug in mainline :) 1229627736 M * arekm hehe 1229628395 M * maharaja btw 1229628402 M * maharaja if a vserver is running 1229628437 M * maharaja # vserver-stat 1229628437 M * maharaja CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME 1229628437 M * maharaja 106 2 53.5M 1.8M 0m00s24 0m00s76 1m59s79 lenny64 1229628442 M * maharaja and i manually exec 1229628451 M * maharaja /usr/lib/util-vserver/distributions/lenny/initpost /etc/vservers/lenny64/ /usr/lib/util-vserver/util-vserver-vars 1229628459 M * maharaja the NAMES is "dropped" 1229628461 M * maharaja resulting in 1229628473 M * maharaja 106 2 53.5M 1.8M 0m00s24 0m00s76 3m51s39 1229628479 M * maharaja any idea whats happening there? 1229628490 M * maharaja util-vserver r2772 debian lenny "of course" ;) 1229628510 M * maharaja maybe some pidfile messup 1229628563 M * maharaja ok, the pidfile (run) was removed 1229628576 M * maharaja but how do i now stop the vserver? :) 1229628588 Q * balbir_ Ping timeout: 480 seconds 1229628606 M * maharaja ah, echo 106 > /etc/vservers/lenny64/run 1229628609 M * maharaja mhm 1229628615 M * maharaja that should be easier 1229628806 M * Bertl vkill will do that 1229628828 M * Bertl but note that it isn't smart to mess with runtime information while the geusts are running :) 1229628879 M * maharaja nice, didn't note that about vkill 1229628893 M * maharaja well, as there is nothing essential in the run file 1229628896 M * maharaja such as a pid or so ... 1229628914 M * maharaja i guess it would be save to make a guess by parsing /etc/vservers/*/context 1229628968 J * Aiken ~Aiken@ppp118-208-39-59.lns3.bne1.internode.on.net 1229628971 Q * Aiken 1229629217 J * xdr ~xdr@118-173-96-87.cust.blixtvik.se 1229629421 J * Aiken ~Aiken@ppp118-208-39-59.lns3.bne1.internode.on.net 1229630605 J * whuji ~huji@bgl93-2-82-226-41-76.fbx.proxad.net 1229630617 M * whuji hello again :) 1229630715 M * Bertl wb 1229630801 M * whuji some questions again and again :D 1229630821 M * whuji is it possible for a guest and a host to share the same ip with differents ports mapping ? 1229630910 M * Bertl a host and guest can share one (or more) IPs, I don't know what you mean with ports mapping :) 1229630919 M * whuji and there is some tools to easily migrate vserver guests from a server to another ? (vserver-copy doesn't seem to work) 1229630965 M * whuji for exemple guest and host have the same ip : a.b.c.d but a.b.c.d:10022 is the ssh for the host and a.b.c.d:22 is the ssh for the guest ? 1229630967 M * Bertl there is no vserver-copy, except for debian 1229630974 M * whuji I'm on debian 1229630987 M * Bertl and there it should have been wiped out 5 years ago :) 1229631010 M * whuji ok 1229631012 M * Bertl use rsync or a shared medium to move guests around 1229631022 M * whuji ok so no tool :) 1229631032 M * Bertl (for rsync you want -NurpP --numeric-ids 1229631108 M * Bertl regarding the ip/prot, yes, just start sshd on the guest on port 10022 1229631119 M * Bertl (or the other wayround) 1229631134 M * whuji ok perfect :d 1229631153 M * Bertl note that it is not the best idea to do that this way though 1229631162 M * whuji it would be great to have migration and monitoring tools for vserver 1229631189 M * Bertl mainly because the guest is able to DoS the host with shared IPs 1229631204 M * Bertl it is preferable to use D/SNAT to map a private IP to the host IP 1229631232 M * Bertl whuji: write one and let us know, we'll add a link to the web page :) 1229631276 M * whuji sure, in python ? or bash ? 1229631291 M * Bertl whatever you prefer .. it's your tool ... 1229631312 M * whuji it's for better vserver integration 1229631357 M * whuji I don't really need such a tool I don't migrate vps often enough 1229631514 M * Bertl hmm, so you want to write a tool, which you do not need? unusual :) 1229631658 M * whuji I often read that vserver lacks such tools 1229631716 M * Bertl ah, okay, good point ... 1229631740 M * whuji it's said that it's a good point for openVZ over vserver 1229631908 M * Bertl sould be fairly simple, the usual process is to run rsync while the guest is running 1229631924 M * Bertl then shutdown the guest, re-run rsync, and then start it on the other end 1229631925 M * whuji yes it is easy 1229631956 M * whuji but it would be more easy to do vsmigrate guest1 remote_host:[name2] 1229632001 M * Bertl yeah, no problem with that, write it, test it, we can definitely provide a link/copy, and maybe daniel_hozac includes it in util-vserver 1229632023 M * whuji because from a server to another, we need to migrate the guest first, and its configuration, then. It's not complicated, but it could be easier 1229632447 J * Moo ~troy@shells195.pinchaser.com 1229632459 M * Bertl welcome Moo! 1229632473 M * Moo greetz 1229632482 N * Moo MooingLemur 1229632485 M * whuji and it could be possible to use lvm snapshot to clone guest without stopping it ? 1229632516 M * Bertl sure, why not ... note that some services might require a proper shutdown 1229632608 M * whuji ok 1229633260 Q * larsivi Read error: Connection reset by peer 1229633262 M * whuji there is some limitation about architecture when moving guest from server to another ? 1229633264 J * larsivi_ ~larsivi@9.80-202-30.nextgentel.com 1229633280 M * Bertl yes, they need to be ninary compatible 1229633283 M * Bertl *binary 1229633362 M * whuji so it should work between 2 i686 ? 1229633407 M * Bertl definitely, also moving an x86 guest to x86_64 host (with ia32 enabled) will work, as well as moving it back 1229633493 M * whuji ok because I've moved a guest from an intel xeon quad to an intel core 2 duo and the guest hangs with an error 'init failed (/etc/init.d/rc 3)' 1229633509 M * Bertl similar is true for 32/64 bit sparc, sh or powerpc 1229633520 M * Bertl how did you move it? 1229633571 M * whuji tar then ftp 1229633589 M * Bertl you probably lost msot of the flags and ownerships 1229633607 M * Bertl (unless you got the tar options right :) 1229633615 M * whuji tar with preserve 1229633638 M * Bertl --numeric-owner 1229633646 M * whuji ah... ok thank you. 1229633652 J * cga ~weechat@94.36.80.137 1229633716 M * Bertl you might also want --sparse 1229633741 M * whuji ok 1229633845 M * whuji is vserver automaticaly startup guest when reboot ? 1229633871 M * Bertl that depends on the config, the default is not to startup the guests 1229633881 M * whuji ok. 1229634071 M * whuji so : 1229634071 M * whuji tar cvjf -p -spare --numeric-owner guest1.tbz2 guest1 1229634074 M * whuji do u agree ? 1229634101 M * Bertl should work 1229634107 M * whuji something better? 1229634146 M * Bertl rsync is definitely better/simpler, but if that isn't an option :) 1229634176 M * Bertl also not sure that -j (bzip2) is worth the time, i.e. -z (gzip) is probably sufficient 1229634210 M * whuji ok. go for rsync :d 1229634301 M * whuji rsync -NurpP --numeric-ids guest1 host: 1229634304 M * whuji like that? 1229634354 M * Bertl ... path/to/guest/ host:/path/to/guest/ 1229634375 J * derjohn_mob ~aj@e180193048.adsl.alicedsl.de 1229634381 M * whuji yep of courge 1229634383 M * whuji course 1229634399 M * whuji it's about the arguments of rsync -_- 1229634465 M * Bertl yep, the arguments I listed above are fine :) 1229634478 M * whuji rsync error: syntax or usage error (code 1) at main.c(1318) [client=2.6.9] 1229634481 M * whuji o_O 1229634512 M * Bertl 21:10 < Bertl> (for rsync you want -NurpP --numeric-ids 1229634551 M * Bertl so what is your rsync line? 1229634572 M * Bertl oops, sorry 1229634581 M * Bertl that's the diff options :) 1229634583 M * whuji rsync -NurpP --numeric-ids /guests/dtc1 huji@88.191.97.155: 1229634588 Q * gnuk Quit: NoFeature 1229634608 M * Bertl use -axHPSD :) 1229634627 M * whuji rsync -axHPSD /guests/dtc1 huji@88.191.97.155: 1229634629 M * whuji ok 1229634632 M * Bertl together with --numeric-ids 1229634636 M * whuji pl 1229634637 M * whuji ok 1229634958 M * whuji how can i say to rsync to use the port 10022 instead of 22 ? 1229634980 M * whuji login@host:port: doesn't work ? 1229635036 M * Bertl -e 'ssh -p 10022' 1229635114 M * whuji hmm. well thank you again. 1229635121 M * Bertl np 1229635462 M * whuji same error... 1229635492 M * whuji I've posted it with pastebin 1229635551 M * Bertl can you upload the output of testme.sh and vserver-info - SYSINFo for both machines please? 1229635594 Q * xdr Remote host closed the connection 1229635600 M * whuji I need to restart the machine first 1229635659 M * whuji for the first machine 1229635716 M * whuji where can i find testme.sh 1229635747 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh 1229635837 M * whuji (where do you read what I post in pastebin ?) 1229635843 M * whuji and for the second machine 1229635894 M * whuji done 1229635894 M * Bertl yep, did you enable ia32 emulation for the 64bit machine (new one)? 1229635909 M * whuji the new one is 32 bit 1229635913 M * whuji the old one is 64bit 1229635927 M * Bertl ah, okay, well, then is the guest 32 bit? 1229635957 M * Bertl but I think I remember the new one was a core 2? 1229635969 M * Bertl (which should be capable to run 64 bit) 1229635985 M * whuji yep I must install a 64 bit kernel so 1229635990 M * whuji and it will work ? 1229636008 M * Bertl assuming that the guest is actually 64bit, then yes 1229636016 M * Bertl but the error looks like it is 1229636048 M * whuji ok I try 1229636059 M * Bertl if you create a 32bit guest on the 'old' machine, you can move it to the new (as is) 1229636089 Q * dna Quit: Verlassend 1229636247 M * whuji apt-cache search linux-image vserver gives me a kernel for AMD K7, no AMD64... 1229636273 M * Bertl probably debian folks know more :) 1229636290 M * whuji weird... on the old server it's amd64 :d 1229636469 M * Bertl IIRC, amd64 is the name debian uses for x86_64 1229636477 M * whuji ahh... I'm sorry. It thought it was a core 2 duo but it is a CELERON220 1229636487 M * whuji I know 1229636493 M * Bertl check for the 'lm' flag in /proc/cpuinfo 1229636539 M * whuji it is there 1229636549 M * Bertl then it is 64bit capable :) 1229636645 M * whuji ok i'll try something 1229636782 M * whuji ok. 1229636786 M * whuji doesn't work 1229636807 M * whuji I reinstall the server 1229636856 M * Bertl probably the best, I know that you can 'upgrade' but that seems to be non-trivial 1229637191 M * whuji and It's more clean :) 1229637243 M * whuji oups 1229637245 M * whuji it's cleaner 1229637248 M * whuji -_-" 1229638602 Q * hparker Remote host closed the connection 1229639985 M * whuji u were right bertl 1229640129 J * rangaparmastan ~bihar@84.78.128.255 1229640214 M * whuji but it still hangs on stop process 1229641031 M * Bertl in what way? 1229641106 M * whuji hmmm I've the same error... 1229641190 M * Bertl the same error on stop? 1229641282 M * whuji RTNETLINK answers: File exists 1229641282 M * whuji ncontext: vc_net_create(): Invalid argument 1229641287 M * whuji at the creation 1229641308 M * Bertl sounds like you have a bunch of config errors :) 1229641323 M * Bertl the first message suggests that the IP already exists 1229641342 M * whuji Yes the guest and the host share this ip 1229641344 M * Bertl the second messag could be caused by a broken/older util-vserver version (normal for debian) 1229641372 M * Bertl you definitely do not want to specify a 'dev' if you actually want to share the IP 1229641394 M * Bertl otherwise a guest shutdown will remove the host IP from the network stack 1229641430 M * whuji arf :/ 1229641436 M * whuji so I can't do this way 1229641458 M * whuji I must do a NAT on all ports except ssh one 1229641464 M * Bertl sure, but you specify 'nodev' instead of 'dev' 1229641503 M * whuji why ? 1229641519 M * whuji in /etc/vservers/guest/interface/0/dev ? 1229641562 M * Bertl dev just means that util-vserver handles the up/down of that IP 1229641605 M * whuji so I rename dev in nodev ? 1229641629 M * Bertl remove the 'dev' entry, touch a file called 'nodev' (as described on the flower page) 1229641641 M * whuji ok 1229641666 M * whuji so I try again 1229641707 M * whuji it works 1229641709 M * whuji ok :) 1229641716 M * whuji Thank you bertl 1229641738 M * whuji so now I don't have to d-nat ? 1229641822 M * Bertl you're welcome! 1229641862 M * Bertl no, you do not _have_ to, but as I said, this is a potential DoS scenario (from the guest) 1229641898 M * whuji yep of cours I undestand. 1229642313 M * whuji I think there still is a problem of permission or owner... 1229642357 M * Bertl in what way? 1229642415 M * whuji mysql can't start because : mysqld: File '/var/log/mysql/mysql-bin.index' not found (Errcode: 13) 1229642425 M * whuji and if I do a ls -l /var/log/mysql : 1229642435 M * whuji vs:/# ls -l /var/log/mysql 1229642435 M * whuji total 1656 1229642435 M * whuji -rw-rw---- 1 1000 1000 14539 2008-12-17 17:33 mysql-bin.000001 1229642455 M * Bertl did you use the --numeric-ids option? 1229642464 M * whuji for the rsync 1229642479 M * whuji yes 1229642513 M * whuji rsync -axHPSD --numeric-ids -e 'ssh -p 10022' guest host: 1229642520 M * whuji that's what I did 1229642538 M * Bertl and on the original, what does ls -l /var/log/mysql show? 1229642556 M * whuji vs0:/# ls -l /var/log/mysql 1229642556 M * whuji total 59212 1229642556 M * whuji -rw-rw---- 1 mysql adm 14539 2008-12-17 17:33 mysql-bin.000001 1229642616 M * Bertl and with -lan ? 1229642646 M * whuji total 59284 1229642646 M * whuji drwxr-s--- 2 105 4 4096 2008-12-18 12:27 . 1229642646 M * whuji drwxr-xr-x 9 0 0 4096 2008-12-18 12:26 .. 1229642646 M * whuji -rw-rw---- 1 105 4 14539 2008-12-17 17:33 mysql-bin.000001 1229642680 M * Bertl so that doesn't look like --numeric-ids to me :) 1229642744 M * whuji I try to rsync again 1229642991 M * whuji ok 1229643001 M * whuji so I do the rsync with --numeric-ids 1229643022 M * whuji but on the new server it's the bad permission... 1229643072 M * Bertl if you are 100% sure that you rsync to the right path (better check that) then your rsync is broken :) 1229643155 M * whuji what's the right path ? 1229643180 M * Bertl what isthe naem of your guest? 1229643183 M * Bertl *name 1229643185 M * whuji dtc 1229643214 M * Bertl and you have debian on both systems, yes? 1229643248 M * whuji yes 1229643494 M * Bertl then you want to do: rsync -axHPSD --numeric-ids root@oldhost:/var/lib/vservers/dtc/ /var/lib/vservers/dtc/ 1229643536 M * Bertl note the / at the end of the directory, otherwise you'll do a copy into a subdir 1229643661 M * whuji ok 1229643678 M * whuji and why the rsync from the old couldn't work ? 1229643713 M * Bertl it also works the other way round 1229643753 M * whuji the owner of /var/log/mysql is now sshd and not mysql... 1229643753 M * Bertl rsync -axHPSD --numeric-ids /var/lib/vservers/dtc/ root@newhost:/var/lib/vservers/dtc/ 1229643783 M * whuji creaktifserver:/home/pepere/dtc_bak/var/log/mysql# rsync -axHPSD --numeric-ids root@vs0.reseau2000.org:/backup/dtc_bak . 1229643792 M * whuji -rw-rw---- 1 sshd adm 17947 2008-12-18 13:18 mysql-bin.000024 1229643985 M * Bertl looks good (but check with -n) .. anyway off to bed now .. have a good one everyone! 1229643999 N * Bertl Bertl_zZ 1229644253 M * whuji it works. add og to args