1102291509 M * brc do the vserver need to be stoped, so i can unify it ? 1102291535 M * Bertl basically no, but it would be better ... 1102291559 M * brc i think i fucked up everything 1102291571 M * brc i have context quotas 1102291578 M * brc and i did a vunify on some vservers 1102291581 M * brc and now i get permission denied 1102291587 M * brc and i see the xid of the files have changed 1102291642 M * Bertl identify the 'unified' files and force them back into xid=0 1102291659 M * Bertl then everything should be fine again 1102291746 M * brc may i chxid -R -c 0 /vservers/* ? 1102291765 M * Bertl only if you want to remove all xid tagging 1102291766 M * Doener Bertl: hm, we did the lock/private one more or less together and the CAP_CONTEXT one looks like an enhanced version of mine 1102291790 M * brc Bertl: i will do that just to maek everything work again 1102291793 M * brc people will start complaining soon 1102291794 M * brc hehehe 1102291806 M * Bertl it's easy to identify the unified files 1102291850 M * Bertl start with a find on files with link > 1 1102291877 M * Doener so i like it ;) 1102291881 M * brc good idea 1102291883 M * brc :) 1102291899 M * Bertl find /vservers/* -type f ! -links 1 1102291915 M * Bertl Doener: thanks! ;) 1102291938 M * Bertl brc: then for those files, check the flags 1102291952 M * brc ok 1102291955 M * brc thnx 1102291960 M * Bertl immutable files are almost certainly unified files 1102292013 M * Bertl all can be done in a single (simple) script, you could then post to the ml ;) 1102292046 M * brc the vunify just unifies vservers based on reference vservers 1102292056 M * brc wouldn't it be better to have it to unify everything ? 1102292070 M * Bertl sure, IIRC sam had a script for that ... 1102292072 M * brc i mean, vserver A and B have bind 1102292075 M * brc vserver C and D have apache 1102292085 M * brc i dont know about that, so for me it would be better to unify everything 1102292089 M * brc I was doing tha tby hand 1102292094 M * brc but everything stoped working 1102292095 M * brc heheh 1102292095 M * Bertl you just want to avoid config files and such ... 1102292099 M * brc hm 1102292114 M * Bertl they for certain will break once you unify them ... 1102292116 M * brc what would be the problem to unify them if they are equal ? 1102292121 M * mugwump brc: a good balance is to unify only /sbin, /bin, /lib, /usr 1102292130 M * mugwump everything else is small 1102292133 M * Bertl brc: that they do not stay equal ... 1102292150 M * brc but what would i lose unifying everything? 1102292156 M * mugwump and, generally speaking, programs that put files in those places do it "properly", ie renaming new files on top of stuff 1102292160 M * brc i mean, cfg files will be changed, but would that cause a negative impact? 1102292167 M * Bertl the ability to append or overwrite unified files 1102292188 M * Bertl (you have to remove an unified file before you replace it) 1102292209 M * brc so it would be slower 1102292221 M * mugwump find . -type f -links +1 -print0 | xargs -0 lsattr | grep -- -i- 1102292229 M * mugwump that gets you close :) 1102292235 M * mugwump anyway, lunch for me ... bbl 1102292239 M * Bertl brc example: echo "test" >unified_file 1102292247 M * Bertl will result in an error 1102292264 M * Bertl you have to do: rm -f unified_file; echo "test" >unified_file 1102292281 M * brc ah ok 1102292284 M * brc so i fuckd up everything 1102292290 M * Bertl (this will go away with the CoW stuff) 1102292302 M * brc cause users will probrablyu want to upgrade packages 1102292303 M * Bertl well, you can un-unify vservers easily 1102292305 M * brc and on be able 1102292309 M * brc wont be able 1102292321 M * Bertl for package updates, exactly this sequence happens 1102292336 M * Bertl (at least for rpm and pkg based distros) 1102292341 M * Loki|muh Bertl: what is CoW? 1102292349 M * Bertl Copy on Write ... 1102292398 M * Bertl should be CoW links (one of Jörn's projects) 1102292409 M * Loki|muh ok 1102292444 M * Bertl I'll probably start integrating the stuff right after the ngnet stuff has stabilized somewhat 1102292497 M * brc Bertl: i will re-do the process only for sbin bin lib usr as mugwump said 1102292526 M * Bertl brc: yes, but you have to un-unify them first 1102292897 M * Bertl Doener: I've found the issue, it's so ashaming ... 1102292935 M * Doener where is it? 1102292950 M * Bertl include/net/ip_fib_ngnet.h have a look at the info struct 1102293036 M * Doener 23:31:22 Doener hm, you got nfxid in ip_fib.h, but fib_xid in ip_fib_ngnet.h ... so according to your explanation it shouldn't even compile, right? 1102293038 M * Doener that one? 1102293059 M * Bertl nope, fixed that in the previous version ... 1102293123 M * Bertl what is defined immediately before that? 1102293207 M * Bertl you're doing that to make me feel better, right? 1102293260 M * Bertl struct fib_nh fib_nh[0]; 1102293268 M * Bertl that is a zero size array ... 1102293282 M * Doener d'oh, didn't notice... 1102293298 M * brc Bertl: trying to deunify 1102293818 M * Bertl Doener: okay uploaded latest version, any code cleanup is really appreciated 1102293824 M * brc cant deunify 1102293828 M * brc i unified everything manually 1102293849 M * brc i am doign the same thing i did before, but with -R 1102293852 M * brc and it does nothing 1102293872 M * Bertl did you make the list of 'unified' files we talked about before? 1102293907 M * brc yes i have the list 1102293915 M * brc not using it 1102293918 M * Bertl okay, you know how to write a simple script? 1102293925 M * brc yes 1102293937 M * Bertl okay, then do the following for every file in that list: 1102293952 M * Bertl mv $file $file.unlikely 1102293962 M * Bertl cp -va $file.unlikely $file 1102293969 M * Bertl rm -f $file.unlikely 1102293987 M * brc so i will have to stop the vservers 1102293989 M * Bertl where file is the full path ... 1102293990 M * brc i wouldn't deunify it now 1102294003 M * brc but i am having problems with the xid thing 1102294006 M * brc i chxid one file 1102294010 M * Bertl doesn't matter if you stop them or not 1102294013 M * brc and it seems like the other hard link gets chxided 1102294034 M * Bertl yes, the xid is per inode 1102294063 M * Bertl hardlinks are just pointers to the same inode 1102294108 M * brc so how are the files upgraded ?? 1102294114 M * brc if their xid belong to other context ? 1102294699 M * Bertl unified files are never changed 1102294710 M * Bertl they are just removed (which dissolves the hardlink) 1102294730 M * Bertl so unified files always have to belong to xid=0 1102294731 M * brc how will they be removed if they belong to anoter xid 1102294734 M * brc ahhhhhhh ok 1102294737 M * brc understood 1102294738 M * brc :) 1102295439 J * monrad ~monrad@213.83.190.130 1102295444 M * monrad hi 1102295448 M * Bertl welcome monrad! 1102295957 M * Bertl Doener: I need a good idea how to handle vserver to vserver traffic in ngnet 1102295997 M * Bertl inter-vserver traffic that is ... 1102296108 M * mugwump What you need is the dummy equivalent of a hub 1102296119 M * monrad is it still possible to run a bind9 server in a vserver without recompiling it 1102296152 M * Bertl monrad: if you add enough caps, I'd say yes ... 1102296223 M * monrad and i got an error when mysql tried to renice itself on debian 1102296261 M * brc Finally 1102296264 M * brc Bertl: deunified :) 1102296324 M * Bertl congrats! 1102296346 M * Bertl monrad: change the config file 1102296392 M * Bertl no idea why debian 'thinks' that mysql must be reniced ... 1102296452 M * Bertl mugwump: I was more thinking of propagating local addresses to all vservers and handling the loopback case special 1102296462 M * Doener Bertl: any details on the problem? 1102296482 M * Bertl do you have the ngn running in a qemu or so? 1102296519 M * Doener no, never managed to get any console in qemu excapt for runlevel 1 1102296549 M * Bertl *sigh* why don't you use my disk image ... 1102296605 M * Bertl okay, I basically us the setup done by vnet3_setup.sh 1102296622 M * Bertl chcontext --ctx 100 ping -c 1 192.168.0.1 1102296628 M * Bertl works fine, of course 1102296631 M * Bertl so does: 1102296636 M * Bertl chcontext --ctx 100 ping -c 1 192.168.0.2 1102296649 M * Bertl all accounting is fine for the interfaces ... 1102296750 M * Bertl now that doesn't work (expectedly) for 1102296754 M * Bertl chcontext --ctx 100 ping -c 1 192.168.0.3 1102296783 M * Bertl because the context does not know of 192.168.0.3 and therefor send out an unanswered arp request to the net 1102296825 M * Bertl Doener: if you like I can upload a debug log for that 1102296968 M * Doener hm, what should i pass as init to the kernel with your disk image? 1102297023 M * Bertl qemu-fast -nographic -L qemu-0.6.1/pc-bios -m 64 -snapshot -hda IMGs/TESTn_32M.img -kernel linux-2.6.10-rc3-vs1.9.3.9-ng6-P1/arch/i386/boot/bzImage -append "rw root=/dev/hda1 devfs=mount" 1102297174 M * Doener the rw seems to do the trick 1102298686 M * Bertl my best idea so far is to add a special type in ip_route_output_slow() 1102298714 M * Bertl including the appropriate check for all vservers, so that the route basically becomes: 1102298744 M * Bertl RTN_VNET/RTCF_VNET 1102298968 M * Bertl or maybe we should do a separate hash for all vserver addresses with the xid/vnet number in it 1102298984 Q * tchan Remote host closed the connection 1102298999 M * Bertl and check for those on the vnet xmit ... 1102299410 M * Doener how do i get iptables into the disk image? i.e. what distro is that? 1102299534 M * Bertl sec, I'll uplaod the binary 1102299551 M * Bertl (it's basically an old mandrake) 1102299615 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/iptables 1102300195 M * Doener iptables v1.2.9: Unknown arg `--vnet' 1102300219 M * Doener vc_add_vndev: Invalid argument 1102300223 M * Bertl hmm, forgot something ... 1102300318 M * Doener /proc/vnet is listed twice by ls 1102300331 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/lib_iptables.tar.bz2 1102300345 M * Bertl unpack this into the root of the qemu image 1102300643 M * Bertl the vnet twice issue is interesting ... 1102300737 M * Bertl but not hard to explain actually ... 1102300751 M * Bertl I should not use the same name for different entries ;) 1102300755 M * Doener hehe 1102301222 J * tchan ~tchan@c-24-13-81-164.client.comcast.net 1102301250 M * Bertl wb tchan! 1102301333 M * Doener Bertl: which interface has address 192.168.0.1 in your setup? 1102301392 M * Doener just noticed that you used 'ping -c 1 192.168.0.1', but vnet3_setup.sh doesn't define that address 1102301416 M * Bertl that is outside of qemu 1102301437 M * Bertl 2: dummy0: mtu 1500 qdisc noqueue 1102301437 M * Bertl link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff 1102301437 M * Bertl inet 192.168.0.1/24 brd 192.168.0.255 scope global dummy0 1102301445 M * Bertl (on the host) 1102301670 M * Doener hm, how do you do that? i can't setup any routes within qemu 1102301675 M * Doener # route add -net 192.168.0.0/24 gw 10.0.0.1 1102301675 M * Doener fib_create_info() set 87d33a60,#0 1102301675 M * Doener SIOCADDRT: Network is unreachable 1102301682 M * Doener same with "ip r a ..." 1102301756 M * Bertl please show me your 'ip route ls' 1102301762 M * Bertl and 'ip addr ls' 1102301770 M * Bertl (or upload it somewhere) 1102301795 M * Doener http://doener.homeip.net/doener/ip 1102301838 M * Bertl that is inside qemu? 1102301846 M * Doener yes 1102301856 M * Bertl hmm, why did you configure dummy0 ? 1102301883 M * Doener was just a test... 1102301888 M * Bertl my dummy0 is on the host running qemu 1102301910 M * Doener was the same before i configured dummy0 1102301913 M * Bertl and the 'route' is not added on the qemu, but inside the vserver and automatically by the vnet3 script 1102301935 M * Bertl (it's a link type of route) 1102301963 M * Bertl the only thing you ahve to configure on the qemu host is the ip of the tun0 device 1102301975 M * Bertl (which is automatically set to 10.0.0.1 here) 1102301983 M * Doener the same here 1102302006 M * Bertl so then chcontext --ctx 100 ping -c 1 192.168.0.1 1102302010 M * Bertl should work fine 1102302027 M * Bertl (if you configure any ip on the host to 192.168.0.1) 1102302098 M * Doener fails for 192.168.0.101 (can't use .1 since that is my routers address ;) 1102302199 M * Bertl ifconfig dummy0:1 192.168.0.101 (on the qemu host) 1102302214 M * Bertl chcontext --ctx 100 ping -c 1 192.168.0.101 1102302221 M * Bertl 64 bytes from 192.168.0.101: icmp_seq=0 ttl=64 time=15.018 msec 1102302259 M * Bertl try to tcpdump the tun interface 1102302276 M * Bertl maybe your host isn't answering the arp requests ... 1102302281 M * Doener hm, no... i guess the host tries to reply on eth0 ;) 1102302301 M * Bertl okay, just switch to 192.168.1.x 1102302313 M * Bertl (or whatever is unused in your lan) 1102302459 M * Doener 04:07:32.333106 arp who-has 192.168.2.2 tell 192.168.2.1 1102302462 M * Doener never answered 1102302562 M * Bertl did you change the vnet3 script? 1102302597 M * Doener yep, reads 192.168.2.(2|3) now 1102302663 M * Bertl okay, add a route for 192.168.2.x on the host to the tun device 1102302669 M * Bertl 192.168.0.0 10.0.0.2 255.255.255.0 UG 0 0 0 tun0 1102302838 M * Doener that works... had forgotten to add the gateway to the route 1102303390 M * Doener *lol* 1102303395 M * Doener 04:23:05.995943 arp who-has 10.0.0.2 tell 10.0.0.2 1102303415 M * Bertl who am I? ;) 1102303445 M * Bertl arp is still handled by the host ... 1102304008 M * Doener i'm too tired to think about that now... will continue tomorrow... 1102304031 M * Bertl okay .. found a few issues .. will update the stuff tomorrow I guess ... 1102304053 M * Bertl thanks for helping, have a good sleep ... 1102304097 M * Doener Please press Enter to activate this console. ??? profile done. 1102304116 M * Doener / # 1102304116 M * Doener Please press Enter to activate this console. profile done. 1102304120 M * Doener damn 1102304125 M * Doener / # !!! cvirt: 87d8e2cc[nr_uninterruptible] = -1 on exit. 1102304156 M * Bertl hmm .. have to look into that too I guess ... 1102307181 M * Doener night! 1102307184 N * Doener Doener_zZz 1102308501 J * _no_x ~vps@c150064.adsl.hansenet.de 1102308532 M * Bertl welcome _no_x! 1102308600 Q * no_x Ping timeout: 480 seconds 1102308610 N * _no_x no_x 1102309282 M * matti Bertl: Why "util-vserver" uses both C and C++ code? 1102309315 M * Bertl legacy reasons ... 1102309323 M * Bertl the original tools where c++ 1102309344 M * matti Hmm... Can I add gettext support in future? 1102309396 M * Bertl if you like to, why not ;) 1102309398 M * matti :] 1102309531 M * matti Bertl: Hm, what is "vflag"? Some experimental new tool? 1102309560 M * Bertl well, an old experimental tool I'd say ... ;) 1102309569 M * matti ;-P 1102309578 M * Bertl it's already included in util-vserver 1102309593 M * Bertl vattribute --help 1102309614 M * matti Yep, I see. 1102309652 M * Bertl but you're right, it was a testing tool 1102309807 M * matti Hmm... Can I use separate Linux distribution in vServer context? For example - Gentoo and Debian in same time... 1102309871 M * Bertl not in the same context, but on different contexts, yes 1102309883 M * matti Great! 1102309883 M * matti ;] 1102310070 M * matti Hm... Traffic hidding is still in development state? 1102310256 M * Bertl what is traffic hiding? 1102310305 M * matti Do you remember my question about tcpdump? 1102310388 M * Bertl hmm, not exactly .. but tcpdump doesn't work in vservers ... 1102310576 M * matti But if I enable capability, that allows tdpdump working, it will see traffic from all other vservers - not only for his context? 1102310591 M * Bertl without ngnet, yes 1102310637 M * matti What is "ngnet"? 1102310676 M * Bertl ngnet is the next generation networking, are you not subscribed to the mailing list (ml)? 1102310704 M * matti Yes :/ 1102310728 M * Bertl well, then get subscribed ;) 1102310733 M * matti Aye aye! 1102310745 M * matti How ngnet works? 1102310764 M * Bertl ngnet uses partial stack virtualization 1102310776 M * matti Use some magic skb_* buffers or stack...? 1102310778 M * Bertl every context can bind any address 1102310788 M * matti ;-) 1102310803 M * matti Eh... 1102310813 M * matti OOM Killer just kill Firefox :] 1102310831 M * matti I love my Pentium :-] 1102310872 M * Bertl check the ml archives for the announcement 1102310893 M * matti OK 1102310955 M * matti OK, I hate Macromedia :] 1102310983 M * Bertl those are the guys doing those strange plugins, right? 1102310988 M * matti Yep. 1102311004 M * matti And on my Pentium 166 MMX Flash 7 almost don't work at all... 1102311005 M * Bertl never got one for my lynx ;) 1102311019 M * matti Hahha :] 1102311061 M * matti Some site have stupid flash menus only :/ 1102311189 M * Bertl well, just ignore them ... 1102311254 M * matti I want to buy memory for my new machine... But polish GeIL site is in Flash only :/ 1102311257 M * matti Hm... 1102311261 M * matti GeIL sucks :] 1102311589 T * * http://linux-vserver.org/ | latest stable 1.29, devel 1.3.9, 1.9.3 1102311589 T * Bertl - 1102311652 J * ravenedge click@dsl-84-161.aal.tiscali.no 1102311663 J * Bertl ~herbert@janus.mc.tuwien.ac.at 1102311665 J * ndim U2FsdGVkX1@helena.bawue.de 1102311671 Q * tchan Remote host closed the connection 1102311680 J * Snow-Man ~sfrost@snowman.net 1102311685 J * matti matti@linux.gentoo.pl 1102311713 J * Zoiah Zoiah@matryoshka.zoiah.net 1102311735 M * matti Hmm... 1102311738 J * Loki|muh loki@satanix.de 1102311789 Q * UnionPivo Ping timeout: 480 seconds 1102311854 M * matti Bertl: WOW! NGNET is really awsome! :-)))))))))) 1102311855 J * UFOczek ufoczek@hood.openbug.net 1102311876 M * Bertl ;) 1102312221 J * chand ~chand@cancun.aspic.com 1102312221 A * chand is gone. Gone since Thu Dec 2 15:41:59 2004 1102312237 J * UnionPivo ~union@clj8-137.dial-up.arnes.si 1102312352 J * monrad ~monrad@213083190130.sonofon.dk 1102312591 M * Bertl okay, latest ng6.4 uploaded for testing ... 1102312601 M * Bertl I'm off to bed now .. cya tomorrow ... 1102312617 M * matti Sleep well! 1102312618 M * matti :] 1102312625 N * Bertl Bertl_zZ 1102314751 J * grecea ~grecea@h-195-22-237-74.mdl.net 1102314755 J * brc bruce@201008071179.user.veloxzone.com.br 1102319784 Q * UnionPivo Ping timeout: 480 seconds 1102322499 M * Eyck hmm, I'm worried about Bertl... 1102322503 M * Eyck he's sleeping all day... 1102322538 M * Eyck this will cause 'spring depression' to kick in soon... 1102322806 J * BWare ~bware@212.26.196.41 1102324279 J * rs rs@ice.aspic.com 1102324285 M * rs hi dudes 1102325056 J * jsambrook ~jsambrook@aelfric.plus.com 1102325157 M * chand yo 1102325426 Q * jsambrook Remote host closed the connection 1102325706 J * jsambrook ~jsambrook@aelfric.plus.com 1102327245 J * anonymous-coward ~nwalsh@shaggy.internode.com.au 1102332200 Q * chand Ping timeout: 480 seconds 1102332200 J * chand ~chand@cancun.aspic.com 1102332201 A * chand is gone. Gone since Thu Dec 2 15:41:59 2004 1102335373 J * hesus darksoul@pingu.ii.uj.edu.pl 1102335999 J * Hollow ~bene@home.xnull.de 1102336484 Q * Hollow Ping timeout: 480 seconds 1102338677 P * jsambrook 1102340578 J * tchan ~tchan@c-24-13-81-164.client.comcast.net 1102341328 N * Doener_zZz Doener 1102344745 N * ravenedge click 1102347949 Q * cereal Remote host closed the connection 1102348171 N * Bertl_zZ Bertl 1102348737 M * Bertl morning folks! 1102349289 M * chand hey bertl 1102349404 M * Bertl hey chand! you are back! ;) 1102349496 M * chand hehe yeah 1102349501 M * chand i'm always off and on :0 1102349505 M * chand here and there :) 1102350154 M * Bertl i.c. 1102350256 M * rs hey Bertl! 1102350266 M * Bertl hey rs! how is life? 1102350285 M * rs pretty fine thx:) 1102350290 M * rs what about yours ?:) 1102350304 M * Bertl yeah, fine too, had a great weekend! ;) 1102350321 M * lilo hi Bertl! 1102350327 M * rs really ? did you go somewhere ? 1102350334 M * Bertl hey lilo, how are you? 1102350386 M * Bertl rs: yes, but it wasn't necessary to go somewhere ;) 1102350489 M * rs you're right :) 1102350528 M * Bertl so, already switched to 2.6.10-rc3? 1102350579 M * rs hey I heard that you stole the NGN code to freeVPS :) 1102350591 M * rs bad guy! =) 1102350642 M * Bertl yeah, I ripped it out from the RHEL kernel and plugged it into the 2.6.10, not! 1102350642 M * rs rc3 ? no, still on 2.6.9, and still trying to figure out where is this f***g Herbert Xu's fix :) 1102350683 M * rs freeVPS is still 2.4, isn't it ? 1102350695 M * Bertl AFAIK yes ... 1102350723 M * rs so how his code could work on a 2.6? 1102350732 M * rs fud fud fud 1102350774 M * Bertl ah, well, alexey probably had a bad day ... so what ... 1102351556 M * Bertl okay, dinner, back later ... 1102351561 N * Bertl Bertl_oO 1102352375 Q * albeiro Ping timeout: 480 seconds 1102352412 J * albeiro albeiro@linux.gentoo.pl 1102352947 J * jsambrook ~jsambrook@aelfric.plus.com 1102353271 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1102354266 N * Bertl_oO Bertl 1102354285 M * Bertl okay, back now ... 1102354661 M * rs re Bertl 1102354713 M * Bertl tx rs! 1102354946 J * ensc ~ircensc@ultra.csn.tu-chemnitz.de 1102355058 M * Bertl welcome ensc! 1102355575 J * mask3345 ~mask3345@pD9EB3E87.dip.t-dialin.net 1102355598 M * Bertl mask3345: please stop that! 1102355859 P * mask3345 1102355860 Q * sannes Read error: Connection reset by peer 1102355904 M * Doener wow, you're getting paranoid ;) 1102355922 M * no_x ? 1102355950 M * no_x what one have 2do to make Bertl that angry ??? 1102355955 M * Doener .*[0-9]+@.*\.dip\.t-dialin\.net = evil! *g* 1102356005 M * no_x *g* 1102356099 N * click ravenedge 1102356136 M * Bertl hmm, is this paranoia or educated guess? 1102356149 M * Bertl welcome no_x! and Doener! 1102356169 M * Bertl no_x: scan the logs for Doener's pattern ;) 1102356178 M * no_x hi Bertl ... ok 1102356529 M * no_x when there will be regex 4 google ??? *g* 1102357965 Q * Plug Ping timeout: 480 seconds 1102358036 J * Plug ~plug@datadot.net 1102358055 M * Bertl wb Plug! 1102358124 Q * grecea Quit: Leaving 1102358735 Q * albeiro Ping timeout: 480 seconds 1102358756 J * albeiro albeiro@linux.gentoo.pl 1102359098 J * weasel ~weasel@weasel.noc.oftc.net 1102359513 J * mask3345 ~mask3345@pD9EB3E87.dip.t-dialin.net 1102359623 P * mask3345 1102359874 Q * rs Quit: home 1102360492 J * DuckKing ~Duck@dyn-83-154-126-25.ppp.tiscali.fr 1102360568 M * Bertl welcome DuckKing! 1102360919 Q * DuckMaster Ping timeout: 480 seconds 1102361155 M * Doener Bertl: does you cscope find fib_find_info as a global symbol? 1102361311 M * Bertl ah, my hero! 1102361339 M * Bertl I'm just about filing a bug report, but I could not remember the symbol 1102361373 M * Doener ah good, so i'm not alone! 1102361407 M * Bertl do you want to narrow it down? i.e. make a simple file which fails too for testing? 1102361425 M * Doener i could try... 1102361455 M * Bertl don't know it could be a has issue too .. but I would look for a parser issue first 1102361676 M * Doener hm, if i just try to start to ncurses interface in the kernel tree, it fails... 1102361697 M * Bertl huh? 1102361698 M * Loki|muh which error message? 1102361737 M * Doener none 1102361753 M * Doener "cscope -R" works, "cscope" does not 1102361774 M * Doener linux-2.6.10-rc3-vs1.9.3.9-ng6.3 kernel tree 1102361797 M * Bertl works fine here, do you have enough disk space? 1102361823 M * Bertl ah, maybe cscope doesn't find anything without -R ? 1102361923 M * Doener right 1102362150 M * Bertl okay, I narrowed it down ... 1102362163 M * Doener 2fast4me 1102362176 M * Bertl http://vserver.13thfloor.at/Stuff/test.c 1102362181 M * Bertl (failing testcase) 1102362243 M * Doener probably because cscope does not try to be smart, i.e. does not know what #define means 1102362251 M * Bertl yep 1102362505 M * Bertl okay, next issue .. did you get any good to excellent ideas for ngnet? 1102362547 M * Doener nope, didn't even have 'breakfast' (a.k.a dinner) yet 1102362572 M * Bertl so what you are waiting for? christmas? 1102362582 M * Bertl get some dinner ;) 1102362961 J * sannes ~ace@home.skarby.no 1102362984 M * Bertl welcome sannes! 1102363099 M * albeiro hi all :) 1102363103 M * albeiro hi Bertl :) 1102363110 M * albeiro hi matti ;p 1102363145 M * albeiro Bertl: what is current state of virtual server networking ? 1102363164 M * Bertl testing but looks fine ... needs a lot of testing tough 1102363199 M * albeiro good :) 1102363215 M * Bertl so any help is appreciated ... 1102363273 M * albeiro sure. /me is waiting for new hardware to arrive and hope to help after... 1102363287 M * albeiro i hope that vserver does not lack support for PPC 1102363322 M * Bertl well, I hope too ;) 1102363338 M * albeiro we will se 1102363339 M * albeiro e 1102366904 Q * jsambrook Ping timeout: 480 seconds 1102366951 A * mugwump builds linux-2.6.10-rc3-vs1.9.3.9-ng6 1102367000 M * Bertl try ng6.4 1102367038 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-rc3-vs1.9.3.9-ng6.4.diff 1102367307 M * mugwump ooo, failed hunk 1102367341 M * mugwump kernel/vserver/switch.c.rej 1102367348 M * Bertl probably the same issue as yesterday ... yeah, give me a second 1102367378 M * mugwump I've been using immutable links for my kernel source, so no real chance of corrupted dir :) 1102367440 M * Bertl no, it's a patch I added to 1.9.3.9 1102367443 M * Doener mugwump: now patch with -R and then again without -R ... kind of workaround... 1102367462 M * Bertl I'm uploading an updated version of 1.9.3.9 1102367528 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.10-rc3-vs1.9.3.9.1.diff 1102367977 J * mask3345 ~mask3345@pD955EFDF.dip.t-dialin.net 1102368070 P * mask3345 1102368162 M * mugwump http://nopaste.snit.ch/1259 # failed hunk 1102368203 M * Bertl okay, what patch against what vs version? 1102368215 M * mugwump linux-2.6.10-rc3-vs1.9.3.9.1 1102368276 M * mugwump I don't know why that hunk failed actually 1102368298 M * Bertl hum, just the linux-2.6.10-rc3-vs1.9.3.9.1 ? 1102368311 M * mugwump nothing else 1102368330 M * Bertl *sigh* .. checking please hold on ... 1102368383 M * mugwump it looks like it's just a whitespace problem or something stupid like that 1102368439 M * Bertl patch applies cleanly against 2.6.10-rc3 here .. 1102368446 M * Bertl now checking 2.6.10-rc3 ... 1102368487 M * mugwump btw, in case it wasn't clear, it was the diff-..ng6.4 patch that had the failed hunk 1102368500 M * Bertl *deep sigh* 1102368570 A * mugwump builds linux-2.6.10-rc3-vs1.9.3.9.1-ng6.4 1102368673 M * mugwump hmm, HZ of 1000 on an old PII-166 ... maybe I should drop that down a bit 1102368679 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-rc3-vs1.9.3.9.1-ng6.4.diff 1102368699 M * Bertl stop 1102368704 A * mugwump stops 1102368704 M * Bertl that is definitely wrong ... 1102368893 M * Bertl okay, please continue ;) 1102368941 M * mugwump I'm going forward with the manual merge of that final hunk 1102368951 M * Bertl not necessary 1102368959 M * Bertl all patches are updated/uplaoded 1102368990 M * Bertl *vs1.9.3.9.1* should be fine ... 1102369007 M * mugwump do I need to re-fetch it? 1102369027 M * Bertl the base patch should be fine 1102369033 M * Bertl the ng6.4 is new 1102369133 M * mugwump right, I've got another one ... I'm restarting from scratch to be sure 1102370591 M * Bertl okay, I guess I got the basic arp stuff virtualized too (but not the arp itself) 1102370800 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/delta-ng6.4-ng6.5.diff 1102370849 M * Bertl (might contain one or the other debug message) 1102371601 M * mugwump Makefile EXTRAVERSION bug with ng6.4 diff... 1102371758 M * Bertl okay, I give up, can you please take over linux-vserver ;) 1102371822 M * mugwump well, the little server I have to play with could have "reference" kernel trees, and the patches could be extracted dynamically & cached 1102371888 M * mugwump ie, the "reference" trees immulinked copies of the kernel source 1102371947 M * mugwump maybe I'm being dumb and this is something arch or svn could provide infrastructure for 1102371960 M * Bertl ah, no the 1.9.3.9.1 wasn't uploaded ... 1102372034 M * Bertl okay, I don't like those nines anyway, let's get 1.9.3.10 out ... 1102372111 J * Nik ~Nik@cable-153-130.online.bg 1102372121 M * Bertl welcome Nik! 1102372518 M * Nik hi Bertl 1102372524 M * Nik hi all 1102372618 M * Bertl how are you nik? 1102372669 M * Nik sleepy.. pretty tired.. trying to keep up with the updates ;-) what about you? 1102372708 M * Bertl mostly fine ... with what updates do you try to keep up with? 1102372741 M * Nik yours :-) 1102372751 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1102372754 M * Nik i mean vserver 1102372762 M * Bertl welcome SiD3WiNDR! 1102372765 M * SiD3WiNDR moo :) 1102372789 M * SiD3WiNDR I just stopped a vserver, tarred its directory, scp'd to other box, untarred and started it again... network doesnt seem to work. any clues? 1102372795 M * SiD3WiNDR it's to another box on the same subnet 1102372801 M * SiD3WiNDR so taking the ip with it should work fine 1102372803 M * Bertl Nik: hmm, are you referring to ngnet or just vserver ... 1102372827 M * Bertl SiD3WiNDR: network doesn't work means? 1102372836 M * Nik Bertl no idea what exactly ngnet is 1102372845 M * SiD3WiNDR Bertl: well, I cant reach the machine from outside 1102372853 M * SiD3WiNDR only from the host box 1102372859 M * Bertl can you reach teh outside from inside? 1102372863 Q * anonymous-coward Quit: adios 1102372869 M * SiD3WiNDR I can ping but I cant ssh out 1102372870 M * Nik Bertl: new generation net? 1102372886 M * Bertl close, next generation networking ... 1102372903 M * Bertl SiD3WiNDR: what does fail on the ssh out? 1102372919 M * Nik SiD3WiNDR arp caches? 1102372920 M * SiD3WiNDR the connection simply times out 1102372929 M * SiD3WiNDR Nik: they dont last 10 minutes, do they :) 1102372940 M * Bertl what does tcpdump on the host tell you? 1102372944 M * Nik SiD3WiNDR not really 1102373036 M * SiD3WiNDR Bertl: I only see outgoing to the box I'm ssh'ing to, no incoming :o 1102373055 M * SiD3WiNDR I see in&out pings from local subnet though 1102373055 M * SiD3WiNDR hmhm 1102373087 M * Bertl firewall? 1102373104 M * SiD3WiNDR nope 1102373135 M * Bertl okay, router issues? 1102373162 M * Bertl maybe wrong default gateway? 1102373175 M * SiD3WiNDR nope 1102373180 M * SiD3WiNDR host box works fine 1102373193 M * SiD3WiNDR and the vserver comes from another box on same subnet with same gw and everything 1102373210 M * Bertl well, if there are no incoming packets, you can hardly blame linux-vserver code for that, right? 1102373216 M * SiD3WiNDR yea, apparently :) 1102373227 M * SiD3WiNDR and locally on the subnet it works :) 1102373229 M * SiD3WiNDR odd odd. 1102373251 M * SiD3WiNDR I'm going to stop it and start it on the old box again to see if that makes stuff at least work again for now :) 1102373256 M * Bertl maybe the 'other' interface is still up? 1102373260 Q * _sebd Ping timeout: 480 seconds 1102373274 M * SiD3WiNDR nah, I checked, it's not 1102373310 M * SiD3WiNDR hmpf, starting it on the other box makes it ping again... I'll investigate further tomorrow 1102373313 M * SiD3WiNDR thanks for the help :) 1102373364 M * Bertl okay, your're welcome! and good luck! 1102373409 M * SiD3WiNDR bah wth :) 1102373416 J * anonymous-coward ~nwalsh@shaggy.internode.com.au 1102373417 M * SiD3WiNDR I now stopped the old one and started the new one again and it works fine now 1102373420 M * SiD3WiNDR odd networking thingy ;) 1102373423 M * SiD3WiNDR thanks and good night :) 1102373439 M * Bertl night! 1102375494 A * Doener is back from dinner... 1102375506 M * Bertl excellent! :) 1102375506 M * Doener well, and some other stuff was done as well ;) 1102375532 M * Bertl anybody any idea for a quick uninterruptible testcase? 1102375685 M * Bertl ah, thanks, found one ... 1102375700 M * Doener hehe 1102375709 M * Bertl chcontext --ctx 100 bash -c "dd if=/dev/zero bs=1M | dd of=/dev/null bs=1M" 1102375756 M * Loki|muh hu, whats that for? 1102375790 M * Bertl testing for the 'now-fixed' uninterruptible issue ;) 1102375802 M * Loki|muh ah ;) 1102375858 M * Bertl okay, 1.9.3.10 is out, a few issues fixed ... 1102376643 M * mugwump My test server can't compile kernels as quickly as you can release them Herbert! 1102376662 M * Bertl hmm, sorry about that! ;) 1102376681 M * mugwump is there an ngn patch for it? 1102376695 M * Bertl in a few seconds ... (now test compiling) 1102376864 M * Bertl okay, compile done, a quick boto test now ... 1102377008 M * Bertl okay, works fine ... 1102377026 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-rc3-vs1.9.3.10-ng7.diff 1102377048 M * Bertl mugwump: you might find the following useful 1102377054 M * Bertl echo 255 >/proc/sys/vserver/debug_net 1102377059 M * albeiro Bertl's work speed amazes me 1102377080 M * Bertl thanks albeiro! 1102377094 M * Bertl now I'm tired .. so I guess I go to bed ... 1102377111 M * Bertl have fun and a good whatever, everyone ... 1102377120 M * albeiro good night Bertl 1102377135 M * Bertl cya 1102377140 N * Bertl Bert_zZ