1257639074 J * friendly ~friendly@ppp118-209-136-134.lns20.mel6.internode.on.net 1257641911 J * kjj ~kjj@pool-74-107-128-126.ptldor.fios.verizon.net 1257642437 Q * bonbons Quit: Leaving 1257642748 M * daniel_hozac jrklein: i'm still working on the upgrades. 1257642914 M * fLoo Bertl 1257642915 M * fLoo mostly done 1257644648 J * hparker ~hparker@208.4.188.201 1257646559 Q * AmokPaule Quit: Nettalk6 - www.ntalk.de 1257647122 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1257648231 Q * Piet Remote host closed the connection 1257648878 J * Piet ~piet@04ZAAB3Z8.tor-irc.dnsbl.oftc.net 1257649227 J * derjohn_foo ~aj@e180195239.adsl.alicedsl.de 1257649411 J * hparker ~hparker@208.4.188.201 1257649657 Q * derjohn_mob Ping timeout: 480 seconds 1257649662 J * manana mayday_780@84.17.25.144 1257650239 Q * SubZero 1257651086 J * aj__ ~aj@e180195120.adsl.alicedsl.de 1257651457 Q * derjohn_foo Ping timeout: 480 seconds 1257652746 J * saulus_ ~saulus@c135164.adsl.hansenet.de 1257653157 Q * SauLus Ping timeout: 480 seconds 1257653161 N * saulus_ SauLus 1257653198 Q * fLoo Ping timeout: 480 seconds 1257654701 J * Hurga ~hurga@dslb-088-078-062-088.pools.arcor-ip.net 1257654790 M * Hurga Hiya. I just noticed that mysql inside a vserver either binds to 127.0.0.1 or to the vserver IP... but not to both. I need both, though... 1257655368 M * Bertl you sure you have both in your guest? 1257655518 M * Hurga definetely. 1257655542 M * Bertl then binding to 0.0.0.0 will do the trick 1257655543 M * Hurga and i can give either as the bind-address in mysql 1257655561 M * Hurga With 0.0.0.0 it binds to the vserver IP. 1257655579 M * Bertl then you don't have both in your guest :) 1257655590 M * Hurga same if I comment out bind-address completely. 1257655593 M * Bertl i.e. single IP assigned to the guest, and single_ip special casing active 1257655654 M * Hurga single_ip special casing ? Sorry you lost me... 1257655718 M * Hurga I mean, how can I have them more in my guest than seeing both IPs in ifconfig? And using them alternately. 1257655736 M * Bertl usually the kernel keeps a list of IPs assigned to a guest, to check at bind time, but for performance reasons, there is an optimization when only a single IP is assigned to the guest 1257655783 M * Bertl seeing is not a good criterion, as the loopback address is 'faked' 1257655808 M * Bertl check in /proc/virtnet//* what is assigned to your guest 1257655918 M * Hurga Lback: 127.0.20.1 1257655918 M * Hurga 0: [85.199.154.53-0.0.0.0/255.255.255.240:0010] 1257655924 M * Hurga hmm I see. 1257655952 M * Bertl so, simply put ~single_ip in your guest's nflags, and restart it 1257656035 M * Hurga that's a new one... especially other daemons both used 127.0.0.1 and the vserver IP. - will do. 1257656063 M * Bertl well, they actually used the same IP, the only one assigned to the guest 1257656432 M * Hurga and magically, everythings starts to work again... 1257656445 M * Hurga Thanks Bertl. :) 1257656476 M * Bertl you're welcome! 1257656691 Q * balbir_ Read error: Connection reset by peer 1257657127 Q * manana Ping timeout: 480 seconds 1257657630 J * balbir_ ~balbir@122.172.45.248 1257658500 Q * Hurga Remote host closed the connection 1257659146 J * root ~root@124.111.76.233 1257659153 M * root hello 1257659282 M * root hello 1257659282 M * root hello 1257659283 M * root hello 1257659579 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1257659624 M * root what ' up 1257659707 N * root korea 1257659711 M * korea dfdfzzz 1257659713 M * korea zzzz 1257659719 M * korea ha ha ha ~~ 1257659782 F * ChanServ +o Bertl 1257659798 M * Bertl korea: what's up? 1257659814 M * korea hello 1257661967 M * Bertl off to bed now .. have a good one everyone! 1257661973 N * Bertl Bertl_zZ 1257662648 M * korea zzz 1257663410 Q * korea Quit: 전 이만 갑니다. 1257666834 J * ntrs__ ~ntrs@77.28.7.199 1257667155 J * uva_ bno@118-160-171-183.dynamic.hinet.net 1257667188 Q * uva_ Read error: Connection reset by peer 1257667193 Q * uva Read error: Connection reset by peer 1257667243 J * uva bno@114-45-226-1.dynamic.hinet.net 1257667690 M * jrklein daniel_hozac: Thanks for the update! I've managed to compile 2.6.22.19 with vs patch 2.2.0.7. I just finished installing utils 0.30.215 and am copying a few existing guest systems so that I can test the mix in the morning. If things go poorly, I'll wait for your binaries. Again, thanks for your work on maintaining them! 1257667721 M * daniel_hozac you can always just install the older 2.6.22 kernel. 1257667768 M * jrklein Yum must have been picking up the new kernel already? I saw both out there, but it didn't occur to me to try the old one. 1257668609 Q * friendly Quit: Leaving. 1257668927 J * uva_ bno@114-45-226-1.dynamic.hinet.net 1257669326 Q * uva Ping timeout: 480 seconds 1257670244 J * uva bno@114-45-226-1.dynamic.hinet.net 1257670576 Q * uva_ Ping timeout: 480 seconds 1257671262 J * uva_ bno@114-45-226-1.dynamic.hinet.net 1257671661 Q * uva Ping timeout: 480 seconds 1257671692 J * uva ~bno@114-45-226-1.dynamic.hinet.net 1257671846 Q * uva_ Ping timeout: 480 seconds 1257673557 J * ghislainocfs2 ~Ghislain@adsl2.aqueos.com 1257674131 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1257674269 Q * aj__ Remote host closed the connection 1257674926 Q * scientes_ Ping timeout: 480 seconds 1257677442 J * ciprian_craciun ~ciprian@79.114.43.84 1257677606 M * ciprian_craciun Hello all! I want to play with Vserver, but when I try `vcontext --create -- /bin/bash` it gives me the following error: `vcontext: vc_ctx_create(): Invalid argument`. 1257677786 M * ciprian_craciun Any ideas? 1257678401 M * ciprian_craciun It seems that `vcontext --create --xid 5 -- /bin/bash` works. 1257678488 M * ciprian_craciun Anyway. What I'm trying to do (in the end): first understand how vserver works (I mean the low-level tools), secondly to be able to create some custom vserver-management scripts, based on the low-level tools. 1257678508 M * ciprian_craciun Any pointers? (The documentation is scarce in the low-level area. Or?) 1257680933 J * manana mayday_780@84.17.25.144 1257680997 M * ciprian_craciun Is anyone actively online now? 1257681150 J * geb ~geb@earth.gebura.eu.org 1257681203 N * Bertl_zZ Bertl 1257681207 M * Bertl morning folks! 1257681230 M * Bertl ciprian_craciun: did you read the pape? 1257681239 M * Bertl *paper 1257681260 M * ciprian_craciun Nop. I really should, shouldn't I? :) 1257681330 M * geb hi 1257681364 M * Bertl ciprian_craciun: yeah, probably would help to understand how things work ... at least a little :) 1257681493 M * ciprian_craciun Just a quick question. (I promisse I'll read the paper.) Is vserver-start equivalent (in initialization) with the following: `vnamespace --new -- vcontext --create --xid 5 -- vattribute --set --flag fakeinit -- vcontext --migrate-self --initpid -- /sbin/init` 1257681556 M * Bertl nope, it's a lot more complicated nowadays ... more contexts and name spaces, more attributes and setup 1257681586 M * Bertl but you can see what happens when you use --debug on 'vserver start' 1257681595 M * ciprian_craciun Aha. 1257681632 J * doener ~doener@i59F555E5.versanet.de 1257681676 M * ciprian_craciun The problem with `vserver start` is that I need to build a vserver root. (I just want to reuse the current root and run /bin/bash.) 1257681725 M * Bertl building a debian guest via network is quite simple, and you can then see all what happens ... doesn't sound interesting? 1257681730 M * ciprian_craciun But about the missing things from my line above. I know it misses all the security stuff. But will it create a working vserver instance? (I've tried it and it seems that /sbin/init thinks so.) 1257681737 Q * doener_ Ping timeout: 480 seconds 1257681749 M * ciprian_craciun Ok. I'll try to build a debian root. 1257681758 M * Bertl it will create a process isolated init, yes 1257681786 M * ciprian_craciun (The other problem is that I'm using Qemu for emulation and the bare system is on cramfs.) 1257681860 M * Bertl it's rather easy to add a disk to qemu/kvm :) 1257681875 M * ciprian_craciun :) Yup. 1257682934 M * geb morning Bertl :) 1257683017 J * AmokPaule ~amokpaule@brsg-4dbb0ac4.pool.mediaWays.net 1257683413 M * Guy- Bertl: I still have the lmhostid issue on 2.6.31.5-vs2.3.0.36.22 1257683471 J * jop__ ~jop@home.gaulier.info 1257683484 M * jop__ howdy 1257683530 M * jop__ get some trouble with IP address on vserver when I stop it for the first time, I get a "RTNETLINK answers: Cannot assign requested address" 1257683557 M * jop__ and when I want to start again the vserver, I get a poorly SIOCSIFADDR: File exist 1257683558 M * jop__ :( 1257683567 M * jop__ and I didn't find anything on google 1257683572 M * jop__ or the documentation 1257683580 M * jop__ any help is welcome 1257683581 M * jop__ :) 1257683600 J * ntrs_ ~ntrs@77.28.8.148 1257683744 M * Guy- jop__: are you trying to assign the same address to multiple vservers, or a vserver and the host? 1257683771 M * jop__ Guy-, no, not at all, I'm sure the address is free 1257683830 M * Guy- jop__: can you diff -u the output of 'ip addr sh' before starting the vserver for the first time and after starting it? 1257683844 M * jop__ yep, sure 1257683846 M * jop__ wait please 1257683978 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk 1257684025 M * jop__ Guy, at this time, I'm with the case which is not working : http://pastebin.com/m25d3c54d 1257684026 Q * ntrs__ Ping timeout: 480 seconds 1257684038 M * jop__ Guy-, I can reboot the server to be in the right configuration 1257684047 M * Guy- hold it 1257684103 M * Guy- this looks like your loopback interface didn't have an ipv4 address until you started the vserver 1257684106 M * Guy- is that possible? 1257684156 M * jop__ awaiting the end of the reboot, 1 minute more 1257684301 M * jop__ Guy http://pastebin.com/d5b496f8c 1257684305 M * jop__ the default config 1257684312 M * jop__ without starting the vserver 1257684322 M * jop__ after a cleaning reboot 1257684327 M * jop__ now, I start the bad one 1257684546 M * jop__ Guy-, so, the ip of the loopback seems good 1257684606 M * Bertl Guy-: that's interesting ... 1257684819 M * jop__ BenG, :) 1257684993 Q * ciprian_craciun Quit: leaving 1257685119 M * Bertl jop__: your 'original' problem is most likely related to using the wrong netmask/prefix for the guest, and having a device assigned there, for an IP using a host assigned IP 1257685152 M * jop__ I think I found it, thanks to ip addr sh 1257685234 M * Guy- Bertl: I can send you an strace after lunch 1257685281 M * Bertl k 1257685340 M * jop__ ok boys, a nasty file from an import. found it 1257685349 M * jop__ really thank you for the clue 1257685354 M * jop__ have a nice day 1257685360 M * Bertl u2 1257685377 Q * jop__ Quit: Quitte 1257687207 J * blues ~blues@ccb38.neoplus.adsl.tpnet.pl 1257687326 Q * blues_ Ping timeout: 480 seconds 1257689958 Q * BenG Quit: I Leave 1257690913 J * scientes_ ~scientes@174-21-207-238.tukw.qwest.net 1257691009 M * Guy- Bertl: http://pastebin.ca/1662105 1257691124 P * doener 1257692176 M * Bertl Guy-: hmm, yeah, not unexpected, now that I look at my patch :) 1257692287 M * Bertl let me roll a new one :) 1257693050 J * dna_ ~dna@175-195-103-86.dynamic.dsl.tng.de 1257693295 J * hparker ~hparker@208.4.188.201 1257695014 M * Bertl Guy-: please try with patch-2.6.31.5-vs2.3.0.36.23.diff 1257695169 Q * dna_ Quit: Verlassend 1257695284 Q * bonbons Quit: Leaving 1257695671 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1257695884 M * Guy- Bertl: I will, thanks 1257695919 M * Bertl thanks for testing! 1257695927 M * Guy- least I can do :) 1257696176 J * SubZero ~SubZero@chello089076140236.chello.pl 1257698269 J * ghislainocfs21 ~Ghislain@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr 1257698611 Q * ghislainocfs2 Ping timeout: 480 seconds 1257699380 J * pmenier ~pmenier@ACaen-152-1-47-143.w83-115.abo.wanadoo.fr 1257704811 Q * pmenier Quit: Konversation terminated! 1257710437 M * jrklein Bertl: Compiling kernel 2.6.22.19 with the vs 2.2.0.7 patch on CentOS 5 went well. I was able to boot test copies of various CentOS 4 guests fine. I'm trying to work out a number of problems booting CentOS 3 guests. Ideally we'd be rid of CentOS 3 by now, but we still have a few hanging out there. :( I couldn't find anything talking about migrating 2.4 guests to 2.6 hosts, so I'll... 1257710439 M * jrklein ...collect notes along the way as I work out the boot problems. 1257710512 M * jrklein Bertl: Seeing issues like occasional failure to bind an IP to a guest, /proc security warnings, startup sequence (rc 3) warnings, etc. Fun stuff. 1257711036 M * Bertl 2.4 guests should be fine on 2.6 hosts 1257711074 M * Bertl IPs are not bound to guests, they are assigned, what 'failures' do you see there? 1257711102 M * Bertl /proc security warnings are kind-of expected, when the guest is not cleaned up properly 1257711189 M * jrklein These CentOS 3guests came from a CentOS 3 vserver host. I was partially impressed that things worked at all .. and then partially impressed that I had problems when I started reading some more about running 2.4 guests within a 2.6 host. 1257711221 M * jrklein err .. I meant "partially surprised that I had problems" 1257711280 M * jrklein During boot, I see "RTNETLINK answers: Cannot assign requested address" when it tries to bind the IP. Still reading up on a solution, though probably pretty simple. The other errors are also reported by many others. Just a matter of finding out what I need to do to resolve them. 1257711324 M * Bertl the problem there is you told util-vserver (dev entry) to add that IP to the interface, but it's already there :) 1257711325 M * daniel_hozac sounds liks you already have that IP address assigned. 1257711765 Q * uva Quit: Leaving 1257711873 Q * balbir_ Ping timeout: 480 seconds 1257712204 M * AmokPaule wasnt there a list somewhere with programs that have issues when they are run under a vserver? 1257712365 M * jrklein You are correct. I had added the IP binding to the host during troubleshooting. I removed the binding on the host and the error during guest boot disappeared. When I enter the guest, an "ifconfig eth0" does not show an IP. However, a "netstat -an" does show that services are binding to the IP. 1257712374 M * Bertl AmokPaule: http://linux-vserver.org/Problematic_Programs 1257712405 M * Bertl jrklein: did you assign that IP to eth0? 1257712476 M * AmokPaule Ty 1257712506 J * balbir_ ~balbir@122.172.23.197 1257712573 M * jrklein Bertl: The problematic programs was a nice review. I hadn't read through that for a while. I've never had an issue with screens inside a VS. strange. That page was handy when I the pleasure of configuring BIND and NTP in various VS guests. :) 1257712590 M * jrklein I did ass 1257712603 M * jrklein isgn the IP to eth0 1257712634 M * jrklein wow, that extra CR/LF really changes the delivery. 1257712641 M * Bertl :) 1257712655 M * Bertl so, what does 'ip a ls' show on the host and guest? 1257712676 M * Bertl (or if you are missing it on the guest, just on the host :) 1257712776 M * jrklein Host "eth0" shows "inet xx.34.173.168/26 brd xx.34.173.191 scope global eth0" and Guest returns no output at all. 1257712839 M * jrklein The command is present within Guest. When I run "ipconfig" within guest, I also receive no output. I can run "ipconfig eth0" in guest and see MAC of host eth0. 1257712902 M * Bertl and what is in /proc/virtnet//* (where is the guest's network context id)? 1257713085 M * jrklein BTW, I'm suspecting this issue has to do with my /proc problems. Host's /proc/virtnet// contains files "info" (ID: 10, Info: ffff81012a67xxxx, 0: xx.34.184.166/255.255.255.0) and "status" (UseCnt:82, Tasks:7, Flags:0000000402000000, NCaps:0000000000000100). 1257713151 M * Bertl .166 != .168 :) 1257713167 M * Bertl 255.255.255.0 != /26 :) 1257713171 M * jrklein I'm making the assumption that the host IP and guest IP do not have to be in the same block. 1257713200 M * jrklein We have a similar configuration on CentOS 4 servers with VS guests hosted at Server Beach. It's screwy, but has been working. 1257713207 M * Bertl that's correct, but you need to configure util-vserver to assign it to eth0 or assign it yourself 1257713247 M * Bertl if the info you gave me is correct and complete, neither is true 1257713278 M * daniel_hozac it sounds like you were probably bitten by the lack of secondaries promotion 1257713312 M * Bertl now or back then :) 1257713352 P * mugwump 1257713488 M * jrklein Bertl: Not sure I follow. You are suggesting that the guest hasn't been properly assigned the .166 address via util-vserver? I configured the IP within the guest OS (/etc/sysconfig/network-scripts/ifcfg-eth0), but it fails to bind to the IP when I issue "service network restart". 1257713524 M * jrklein Bertl: I also provided the IP when creating the guest, but I could have configured it incorrectly. Just not sure what you are saying/suggesting. 1257713548 M * daniel_hozac the guest's /etc/sysconfig/network-scripts doesn't matter. 1257713571 M * daniel_hozac /etc/vservers//interfaces on the host is where you want to look 1257713654 M * Bertl jrklein: the guest doesn't have permission (by default) to mess with the networking, it is done on the host _for_ the guest 1257713685 M * jrklein daniel_hozac: On the host, "/etc/vservers//interfaces/0/" contains file "dev" (eth0), file "ip" (xx.34.184.166), file "prefix" (24). 1257713689 M * Bertl jrklein: so, you want to actually disable all hardware related (runlevel) scripts inside your guest 1257713698 M * jrklein got it. 1257713774 M * Bertl so, assuming your util-vserver is working properly, it should add that IP (on the host) to eth0 when you start the guest 1257713827 M * Bertl i.e. you should then be able to see it (as secondary if you have other IPs in /24, or as primary) with 'ip a ls' (on the host) 1257714064 M * jrklein I'm not seeing the guest's secondary IP when I issue an "ip a ls" on the host. Do I need to install any guest-side util-vserver code? I used to install something within the guest as part of FreeVPS, but I can't tell if that was just related to the hosting automation side of things. 1257714091 M * Bertl nope, nothing guest side, what util-vserver version do you have? 1257714150 M * jrklein Running "vserver --version" on host returns "... util-vserver 0.30.215 ..." 1257714205 M * Bertl okay, that should be fine, could you run 'vserver --debug start' and upload that (e.g. paste.linux-vserver.org, anonymized is fine)? 1257714473 M * Mr_Smoke Ok this is bad 1257714479 M * Mr_Smoke Defunct process in a vserver ... 1257714487 M * Mr_Smoke Can't stop it, can't kill it, can't shutdown the vserver 1257714490 M * Bertl anything in dmesg? 1257714526 M * Mr_Smoke PAX: suspicious general protection fault 1257714529 M * Mr_Smoke Hooray. 1257714569 M * jrklein Bertl: upload complete 1257714608 M * Mr_Smoke Lots of ext3_orphan_cleanup: deleting unreferenced inode 241419 1257714614 A * Mr_Smoke is panicing 1257714781 M * Mr_Smoke What can I do, short of rebooting the host ? 1257714841 M * Bertl change the xid, and restart the guest 1257714862 M * Bertl you can't really get rid of a process which diet inside the kernel 1257714876 M * Bertl (and it will keep the context around too) 1257715170 M * Mr_Smoke change the xid ? 1257715178 M * Mr_Smoke You mean rename the guest ? 1257715182 M * daniel_hozac no. 1257715197 M * daniel_hozac change the number in /etc/vservers//context 1257715204 M * Bertl nah, shutdown the guest, vkill all remaining processes, change the 'context' entry and start it up again 1257715211 M * Mr_Smoke I can't shut it down 1257715215 Q * micah Remote host closed the connection 1257715224 M * Mr_Smoke If I do that, the shell from which I do it gest stuck (defunct) 1257715351 M * Bertl daniel_hozac: where does the broadcast '+' come from? 1257715376 M * Mr_Smoke Shit, I've got to reboot 1257715383 M * Mr_Smoke One of the rsnapshot processes is stuck 1257715385 M * Mr_Smoke Damnit 1257715880 Q * mEDI_S Quit: mEDI_S 1257715932 J * mEDI_S ~medi@255.255.255.255.li 1257716300 Q * bonbons Quit: Leaving 1257716359 M * Bertl jrklein: your runlevel script starts a few services, but then suddenly fails 1257716396 M * Bertl as result, the guest is shut down right after the startup, which brings down the IP again 1257716410 M * jrklein Yes. This is one of several problems. 1257716418 M * jrklein Oh, that would explain a lot of things. 1257716436 M * Bertl well, the question is, why does your rc script suddenly return failure? 1257716449 M * jrklein Should I disable all my startup services, or do you think this is an OS rc script failure? 1257716490 M * Bertl the services I see started look fine to me, just right after that, it seems to exit with non-zero 1257716591 M * jrklein Each of the enabled services appear to startup fine (syslog, snmpd, sshd, postfix, crond). No other services are enabled. I'll disable them to confirm this is not a service issue. 1257716617 M * Bertl well, klog will not start up fine, but that should be disabled anyways 1257716642 M * Bertl try to add an 'exit 0' to the end of the rc script 1257716658 M * Bertl (or look where it exits non-zero) 1257716834 M * Bertl for example, you could add -x to the hashbang line 1257717387 Q * ghislainocfs21 Quit: Leaving. 1257717502 M * daniel_hozac Bertl: broadcast + is always there for IPv4, unless a different broadcast address is specified in the config? 1257717542 M * Bertl ah, so that's a special for 'iproute'? didn't know 1257717547 M * daniel_hozac yeah 1257717558 M * daniel_hozac + lets ip calculate it based on the IP and prefix. 1257719370 M * jrklein Amazing! Simply adding "exit 0" to the end of "/etc/rc.d/rc" and "/etc/rc.d/rc.local" seemed to fix that issue. 1257719489 M * jrklein Bertl: Is this the correct solution, or is this just a temporary fix that happens to hide an actual error? 1257719600 M * daniel_hozac no. 1257719611 M * daniel_hozac it's as correct of a solution as you'll get. 1257719756 M * Bertl I still find it strange that rc scripts exit with arbitrary exit codes :) 1257719825 M * Bertl but yeah, the solution is fine 1257720040 J * micah ~micah@micah.riseup.net 1257720665 M * jrklein Thanks for your help! I usually do fine solving these things on my own, but I'll have to admit that it would have taken a while for me to come up with that one unless I came across the same suggestion while reading/researching. You saved me a ton of time. Thanks!! 1257720682 M * jrklein brb 1257720745 Q * pmjdebruijn Quit: leaving 1257720745 Q * jrklein Read error: Connection reset by peer 1257720752 J * pmjdebruijn pascal@jester.pcode.nl 1257720769 J * jrklein ~jrklein@75.34.93.82 1257720788 M * Bertl jrklein: you're welcome! 1257720860 Q * pmjdebruijn 1257720888 M * daniel_hozac vserver ... start will suggest the exit 0 thing. 1257720895 J * pmjdebruijn pascal@jester.pcode.nl 1257721597 Q * geb Ping timeout: 480 seconds 1257722466 Q * ntrs_ Ping timeout: 480 seconds