1116982330 Q * monrad Quit: Leaving 1116982953 J * eXplasm2 explasm@p549FF2FB.dip.t-dialin.net 1116983334 Q * eXplasm Ping timeout: 480 seconds 1116986351 Q * _mountie Remote host closed the connection 1116988265 Q * Johnsie Quit: G'bye! 1116989270 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1116991497 Q * tchan Quit: leaving 1116991596 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1116994090 Q * Loki|muh Ping timeout: 480 seconds 1116994374 J * Loki|muh loki@satanix.de 1116997961 M * eyck 1117000141 Q * DaPhreak Ping timeout: 480 seconds 1117001089 N * Bertl_zZ Bertl 1117001095 M * Bertl morning folks! 1117001139 M * eyck morning 1117001157 M * eyck is it normal that the sploit worked under ext2/reiserfs? 1117001187 M * eyck it seems that they ignore iunlink? 1117001209 M * Bertl sure? 1117001221 M * eyck I'm sure I got out, 1117001229 M * Bertl withe the new patch? 1117001235 M * eyck and sure that when I reformatted to xfs I couldn't get out 1117001237 M * eyck yes. 1117001266 M * Bertl hmm .. okay, I'll check it ... 1117001270 M * eyck but the error message seems to be more related to iunlink then the barrier itself 1117001320 M * eyck I got permission denied when sploit tried to change permissions on ../ 1117001776 M * Doener` morning folks 1117001822 M * Doener` eyck: iunlink is part of the barrier... the tools just don't know that (bug has been filed) 1117001848 M * Bertl morning Doener`! 1117001865 M * Bertl eyck: how did you setup the barrier on ext2/reiser? 1117001891 M * Doener` hey Bertl! forgot to change my nick yesterday... seems i only manage to either change my nick or update my status in my instant messenger ;) 1117001923 M * Bertl hehe, np 1117001956 M * Bertl discovered another funny issue with 2.0-rc1++ 1117001976 M * Bertl (but it was easy to fix, just leaves another question ;) 1117001989 M * eyck Bertl: setattr --barrier --iunlink vservers 1117002044 M * eyck Bertl: the same for xfs-formatted partition stopped the sploit. 1117002086 M * eyck (but without iunlink it hapilly got out and chmodded my whole /var/lib/vservers path to 000 :( ) 1117002119 M * Doener` 000? i'd expect 001... 1117002126 M * eyck nope, 0000, 1117002151 M * eyck with following effects: ls / - permission denied. sudo - can't read /etc/sudoers etc... 1117002186 M * Doener` yep, like the well known "chmod 000 /vservers/myvserver" ;) 1117002268 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1117002271 Q * DaPhreak Quit: 1117002301 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1117002328 M * eyck that too 1117002343 M * DaPhreak morning eyck, Bertl .. 1117002416 M * eyck morning phreak 1117002698 M * Bertl morning DaPhreak! 1117002855 M * Psy0rz hi all 1117002889 M * Bertl hey Psy0rz! 1117003174 M * Bertl eyck: okay, tested here in this way: 1117003190 M * Bertl ifconfig lo 127.0.0.1 1117003190 M * Bertl ifconfig eth0 10.0.0.2 1117003190 M * Bertl mount /dev/discs/disc1/part1 /vservers 1117003190 M * Bertl chmod 000 /vservers/XXXX/.. 1117003190 M * Bertl chattr +t /vservers/XXXX/.. 1117003196 M * Bertl vserver XXXX start 1117003213 M * Bertl then ssh to 192.168.0.2 (vserver IP) and run any of those exploits 1117003264 M * Bertl (rootesc2 returns chmod: Operation not permitted, cd ..: Permission denied several times) 1117003274 M * Bertl this is with ext2 .. now testing ext3 1117003958 M * eyck Bertl: hmm, I haven't tried chattr +t / 1117003965 M * Doener` got to go... back in the evening... have a nice day! 1117003967 N * Doener` Doener|gone 1117003978 M * eyck Bertl: so what should be the correct barrier for vs1.2.x ? 1117004070 M * Bertl it 'should' be --barrier (obviously it isn't :/) 1117004171 M * eyck brb, I'll try chattr this time. 1117004633 M * eyck hmm, but it IS set: lsattr vservers/ -d 1117004633 M * eyck ----i----------t- vservers/ 1117004643 M * eyck ls -al vservers/ -d 1117004643 M * eyck d--------- 5 root root 120 May 24 08:20 vservers/ 1117004663 M * eyck /dev/mapper/system-temp on /var/lib type reiserfs (rw) 1117004685 M * eyck and then I get out of chroot 1117004746 M * eyck OTOH chattr +t vservers/ 1117004746 M * eyck chattr: Operation not permitted while setting flags on vservers/ 1117004843 M * Bertl reiser need a mount flag 1117004865 M * Bertl (otherwise it doesn't do xattrs at all) 1117004893 M * Bertl mount -o attrs 1117004938 M * eyck /dev/mapper/system-temp on /var/lib type reiserfs (rw,attrs) 1117004948 M * eyck and I still can't change it 1117004973 M * eyck oh! 1117004979 M * eyck but exploit stopped working :) 1117005047 M * eyck but it's interesting that cp -a managed to copy the attribute 1117005154 M * Bertl it's interesting that the chmod is blocked even for too ... 1117005154 M * Bertl (host root that is) 1117005304 Q * stupidawy jupiter.oftc.net arion.oftc.net 1117005304 Q * mugwump jupiter.oftc.net arion.oftc.net 1117005304 Q * SNy jupiter.oftc.net arion.oftc.net 1117005304 Q * pusling jupiter.oftc.net arion.oftc.net 1117005313 Q * nox jupiter.oftc.net neutron.oftc.net 1117005401 Q * DaPhreak jupiter.oftc.net keid.oftc.net 1117005401 Q * Loki|muh jupiter.oftc.net keid.oftc.net 1117005401 Q * tchan jupiter.oftc.net keid.oftc.net 1117005401 Q * eXplasm2 jupiter.oftc.net keid.oftc.net 1117005401 Q * eyck jupiter.oftc.net keid.oftc.net 1117005401 Q * FaUl jupiter.oftc.net keid.oftc.net 1117005401 Q * Psy0rz jupiter.oftc.net keid.oftc.net 1117005401 Q * virtuoso jupiter.oftc.net keid.oftc.net 1117005401 Q * maharaja jupiter.oftc.net keid.oftc.net 1117005401 Q * romke jupiter.oftc.net keid.oftc.net 1117005401 Q * ndim jupiter.oftc.net keid.oftc.net 1117005401 Q * Vudumen jupiter.oftc.net keid.oftc.net 1117005401 Q * hvd jupiter.oftc.net keid.oftc.net 1117005401 Q * sebd jupiter.oftc.net keid.oftc.net 1117005401 Q * Doener|gone jupiter.oftc.net keid.oftc.net 1117005401 Q * janra jupiter.oftc.net keid.oftc.net 1117005401 Q * albeiro jupiter.oftc.net keid.oftc.net 1117005401 Q * mep_ jupiter.oftc.net keid.oftc.net 1117005401 Q * mcp jupiter.oftc.net keid.oftc.net 1117005401 Q * gaba jupiter.oftc.net keid.oftc.net 1117005419 Q * BWare jupiter.oftc.net quasar.oftc.net 1117005419 Q * bro jupiter.oftc.net quasar.oftc.net 1117005419 Q * gregster jupiter.oftc.net quasar.oftc.net 1117005419 Q * sith jupiter.oftc.net quasar.oftc.net 1117005419 Q * Seraph jupiter.oftc.net quasar.oftc.net 1117005419 Q * case jupiter.oftc.net quasar.oftc.net 1117005419 Q * micah jupiter.oftc.net quasar.oftc.net 1117005419 Q * matti jupiter.oftc.net quasar.oftc.net 1117005419 Q * Hunger jupiter.oftc.net quasar.oftc.net 1117005446 J * SNy ~mfr@bmx-chemnitz.de 1117005446 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1117005446 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1117005446 J * pusling ~pusling@195.215.29.124 1117005454 J * BWare ~bware@office.intouch.net 1117005454 J * matti matti@linux.gentoo.pl 1117005454 J * bro ~vanity@lanparty.lv 1117005454 J * micah micah@micha.hampshire.edu 1117005454 J * case ~case@donpanic.faveve.uni-stuttgart.de 1117005454 J * gregster ~gregor@greart.de 1117005454 J * sith sith@aaronp.com 1117005454 J * Seraph kk@projects.verfaction.de 1117005454 J * Hunger Hunger.hu@Hunger.hu 1117005461 M * Bertl anybody still here? 1117005488 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1117005488 J * Loki|muh loki@satanix.de 1117005488 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1117005488 J * eXplasm2 explasm@p549FF2FB.dip.t-dialin.net 1117005488 J * mep_ mep@p5091E1EF.dip.t-dialin.net 1117005488 J * Doener|gone ~doener@p54877F87.dip.t-dialin.net 1117005488 J * albeiro albeiro@albeiro.usercloak.oftc.net 1117005488 J * sebd ~sebd@lesdeveloppementsdurables.org 1117005488 J * hvd ~takeagues@fw-grz.hollomey.com 1117005488 J * gaba ~gaba@protest.net 1117005488 J * Vudumen vudumen@perverz.hu 1117005488 J * ndim hun@helena.bawue.de 1117005488 J * romke ~romke@procyon.romke.net 1117005488 J * maharaja maharaja@ipax.at 1117005488 J * virtuoso ~s0t0na@80.253.205.251 1117005488 J * mcp ~hightower@wolk-project.de 1117005488 J * Psy0rz ~psy0rz@195.169.61.234 1117005488 J * FaUl ~immo@ip88.164.1211G-CUD12K-01.ish.de 1117005488 J * eyck eyck@81.219.64.71 1117005488 J * janra janra@paradox.homeip.net 1117005508 J * micah_ micah@micha.hampshire.edu 1117005571 M * eyck welcome back 1117005583 M * eyck how was it on the other side of the split? 1117005624 Q * micah Ping timeout: 480 seconds 1117005649 Q * tarantul Ping timeout: 480 seconds 1117005710 J * nox ~nox@noxlux.de 1117005872 M * Bertl eyck: lonely :/ 1117005914 M * Bertl is the net that shitty nowadays? 1117006701 M * eyck net rox, haven't you heard? 1117007479 J * prae ~prae@ezoffice.mandriva.com 1117007497 M * Bertl welcome prae! 1117007527 M * prae Hi 1117007768 M * DaPhreak morning prae 1117007785 M * prae matin DaPhreak :p 1117007815 M * DaPhreak hmmm would it not be bonjour ?! :P 1117007857 M * SiD3WiNDR a l'aise! 1117007862 M * Bertl btw, we write martin with 'r' ;) 1117007873 M * Loki|muh lol 1117007892 M * DaPhreak heh 1117008061 M * Bertl btw, how is vs2.0-rc1 ? 1117008071 M * Bertl do you like it, does it work fine? 1117008084 M * DaPhreak here nice .. runs like a charm .. 1117008100 M * DaPhreak didn't spot again this vserver stop issue (kill hangs) 1117008146 M * DaPhreak but i guess this is related to the utils, right ? 1117008179 M * Bertl yup .. 1117008961 M * id morning #vserver 1117008983 M * eyck morning, evening, #vserver is timezone-agnostic. 1117009114 M * id =) 1117009216 M * Loki|muh yeah, since Bertl is timezone-agnostic, too ;) 1117009298 M * Bertl hey id! 1117009307 M * id Bertl, i found a bug in util-vserver .207 - where is enrico ? 1117009326 M * Bertl lost ... no probably busy working ... 1117009334 M * id i see 1117009347 M * Bertl best way is to check with the bug database on savannah 1117009357 M * Bertl and file it there if it isn't already filed ... 1117009378 M * id ok 1117009394 M * Bertl btw, has anybody an idea why enrico doesn't show up on the channel anymore? 1117009415 M * id i dont know. i hoped you know why ,... 1117009532 M * Bertl anybody good with uml around? 1117009591 M * eyck isn't enrico italian? 1117009601 M * Bertl the name? 1117009611 M * eyck summer started, blondes from all the europe started flocking to italy 1117009623 M * Bertl enrico is german ;) 1117009716 M * id Bertl, do you run 2.6 vs2.0-rc1 + uml skas ? 1117009734 M * Bertl no, why? 1117009770 M * id thats all i can offer for uml 1117009779 M * id but i guess 1117009787 M * id the kernel is somewhat broken 1117009791 M * Bertl ah, so you are running uml? 1117009810 M * id i tried to install the c 't ipcop uml inside 1117009815 M * id a vserver 1117009827 M * id uml mixed 1117009828 M * id host 1117009843 M * id but tun/tap ¨ & 2.6 sucks 1117009852 M * id in combination 1117009865 M * id alone there are all fine 1117009878 M * Bertl hmm, I'd need somebody to verify that vserver inside uml works and vice versa 1117009879 M * id or uml with 2.4 1117009924 M * Bertl with 2.0-rc* that is ... 1117010113 M * id sorry i currently dont have such an environment @ the moment 1117010134 M * id (a working one) 1117010176 M * Bertl meaning 2.0-rc* or uml or both? 1117010183 M * id i can try to set it up this night - if no one else volunteers 1117010202 M * id i currently run vs2.0-rc1 & drbd 1117010224 M * Bertl k, no problem, and thanks! 1117010243 M * id the uml thing was a test - i like to have a 2.4 kernel beside 1117010250 M * id 2.6 for IPsec VPNs 1117010257 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc2, ng9.4 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1117010260 M * id in UML but this is still a dream 1117010278 M * id -rc2 =) 1117010437 Q * rs Quit: rs 1117010464 M * Bertl id: shouldn't be too hard to run 2.4 as uml, no? 1117010525 M * id thats right but the host is a amd64 - and i like to run the host with full speed ;) - 2.6 is quite faster with 10+ vservers 1117010546 M * id compared to 2.4 1117010560 M * Bertl thought so ... what's the issue with a 2.4 uml instance though? 1117010670 Q * aba Ping timeout: 480 seconds 1117010739 M * id the umls did not start the way i expected - the mem parameter was quite confusing . the times i added -mem switch to ./linux the uml started and after 1,5 seconds it gave a panic - the times i did not use the -mem switch there was no kernel panic but it complained about the root device (rest of the settings where there same) i invested not more then 2 hours on this topic so far ;) 1117010778 M * id including compiling ... so this was a pretty basic shit so far 1117010796 M * id i had no problems ever with uml 2.4 host 1117010871 J * aba ~aba@eos.turmzimmer.net 1117010896 M * Bertl wb aba! 1117010904 M * Bertl id: hmm, interesting ... 1117010963 M * id i will do a new compile run tonight 1117010967 M * id with -rc2 1117010984 M * Bertl only if you find the time ... don't worry to much ... 1117011359 M * id j #vlan 1117011437 M * Bertl hmm ... 1117011441 M * id vlan trunking with debian is .... 1117011451 M * Bertl easy? 1117011453 M * id i dont find a valid doku 1117011461 M * Vudumen hi allz 1117011468 M * Bertl hey Vudumen! 1117011494 M * Vudumen id: it's easier than you think. just install "vlan" package and then use vconfig to create your vlan interfaces 1117011505 M * id i have vlans 1117011511 M * id vconfig eth1 2 1117011520 M * id adds vlan 2 to eth1 1117011539 M * id if i now do vconfig eth1 3 1117011555 M * id is there alreay a trunk with vlan 2 & 3 on eth1 ? 1117011556 M * Vudumen well this weekend on a LAN party i had a debian router (dual-xeon) with ~100 vlan interfaces on 2 intel gigabit ethernet controllers bonded :) 1117011570 M * Vudumen id: yes 1117011611 M * Vudumen id: configuring eth1 is something like native vlan on ciscos and all the vlan interfaces are trunked by default on the interface 1117011728 M * id ok - i still wait for the linksys switch to do further testing - the lucent switch i am currently working on is evil ;) 1117011736 M * Vudumen :) 1117011775 M * Vudumen i use cisco switches here and on this party we used SMC switches. SMC has a very small CPU (only used when configuring) but in switching it's very fast. 1117011781 M * id lucent == lucifer enterprises 1117011786 M * Vudumen :))) 1117011807 M * matti id: LOL 1117011834 M * eyck and plan9 = ??? 1117011841 M * eyck HellOS 1117011863 M * matti Yh. 1117011973 M * eyck eh, now there is only one problem I've got with 2.4.30, 1117012336 N * id id_werk 1117012418 M * Bertl eyck: you are going to tell us? 1117012516 J * rs ~rs@staff.lycos.fr 1117012582 M * Bertl morning rs! 1117012692 M * eyck Bertl: oh, it's just that radeon doesn't work with 2.4.30, only with 2.4.28. 1117012860 M * Bertl hmm, radeon as in Xorg radeon? 1117012885 M * eyck yes, XFree86 though 1117012902 M * eyck which is interesting because there is nothing related to radeon in 2.4.28-2.4.29 diff 1117013092 M * Vudumen Bertl: how ngnet is stable nowadays? 1117013110 M * Bertl very ngnet, as a matter of fact ;) 1117013118 M * Vudumen :) 1117013140 M * Bertl no seriously, you need some 'tricks' to get it working 1117013149 M * Bertl but it is stable ... 1117013167 M * albeiro like no advanced routing ? 1117013177 M * Bertl still the whole thing is considered highly experimental 1117013194 M * Bertl albeiro: yes, like this and manual arp entries 1117013233 M * Bertl but it's unlikely that it will panic for example ... 1117013334 M * rs hi 1117013364 M * albeiro hi rs ! 1117013379 M * Bertl rs: new release candidate is out, do you see any chance to test it under 'real world' conditions? 1117013394 M * rs what are changes ? 1117013403 M * rs I still have reboot issue with the current one 1117013406 M * Bertl cleanups, plenty of cleanups 1117013416 M * Bertl which kind of issues? 1117013448 M * rs sometime vservers won't stop, because a process isn't dead 1117013462 M * Bertl any details? 1117013471 M * rs and the only way to kill it is to enter the context and kill -9 it, you can't kill it from context 1 1117013490 M * Bertl vkill doesn't work on that one? 1117013494 M * rs the process is often zombie and really often a mysqld process but not always 1117013501 M * rs yeah it doesn't work 1117013518 M * Bertl but entering the context and killing it works? 1117013519 M * rs hmm didn't tried vkill actually 1117013536 M * rs just tried kill from differents context (1 and the vserver context) 1117013548 M * Bertl okay, vkill is the way to go for those ... 1117013566 M * rs is it a tool issue ? 1117013589 M * Bertl not sure, nobody did report such issues yet ... 1117013596 M * rs really ? 1117013604 M * rs so it's maybe NFS related 1117013628 M * rs or my setup is really bad :) 1117013635 M * Bertl might be, but I don't know how many folks really tested rc1 ;) 1117013658 M * rs I have this problem for long time now 1117013662 M * rs before rc1 1117013664 M * Bertl when does it happen, the zombie mysql? 1117013667 M * rs I already reported it 1117013681 M * rs when ? 1117013686 M * rs it's not predictable 1117013697 M * Bertl I mean, while the guest is running 1117013703 M * Bertl or just on shutdown? 1117013703 M * rs seems that it happen more with some vservers than with others 1117013713 M * rs on vserver restart 1117013723 M * Bertl restart from inside or outside? 1117013724 M * rs restart=stop 1117013733 M * rs good question 1117013753 M * Bertl because the tools still don't get the sync stuff right ... 1117013755 M * rs I don't really know 1117013773 M * rs the reboot from inside vservers doesn't work as well 1117013774 M * Bertl (at least the latest release doesn't) 1117013828 M * rs even with the shutdown trick in vshelper config 1117013856 M * Bertl do you want to investigate/fix that? 1117013864 M * rs yeah if possible 1117013873 M * Bertl when do you have time? 1117013889 M * rs now for instance :) 1117013899 M * Bertl okay, then let's investigate it ;) 1117013944 M * Bertl can we use the rc2 for testing? 1117013969 M * rs yeah 1117013971 M * Bertl (it has some cleaned up debug messages, I'd prefer to have) 1117013991 M * rs I don't know how to reproduce the issue anyway 1117014002 M * rs and I never got it on devel environment 1117014007 M * Bertl well, the reboot from inside not working should be easy, no? 1117014014 M * rs as usual :) 1117014023 M * rs yeah this would be a good start :) 1117014041 M * Bertl okay, so then let's start there ... 1117014063 M * Bertl we need, a) the new kernel with debug/history turned on 1117014093 M * Bertl and b) the preliminary vwait command (probably best to get doener's test version) 1117014107 M * rs vwait ? 1117014156 M * Bertl yeah, well, it's in cvs .. kernel supports it since a long time 1117014212 M * Bertl let me prepare a vwait package for that while you compile the kernel 1117015909 M * Bertl http://vserver.13thfloor.at/Experimental/vwait-0.01.tar.bz2 1117016048 M * matti vwait - what is for? 1117016074 M * Bertl for proper waiting until a context is gone 1117016135 M * matti :) 1117016156 M * matti Bertl: Do you want t-shirt with tux? 1117016194 M * Bertl sure, always ... 1117016197 M * matti http://www.linuxstuff.pl/index.php?p=skl&idc=DP2TZECZda4tBR4Ng20zz1Qy7WopIA0 1117016229 M * matti ... or whatever you want from that store ;] 1117016567 M * rs bertl: my poor connexion didn't finished to get the kernel :( 1117016597 M * Bertl are you downloading the entire kernel? 1117016613 M * matti rs: You can download patchset only. 1117016630 M * rs yep but I don't have a clean base here 1117016640 M * matti I see. 1117016640 M * matti ;] 1117016652 M * Bertl matti: what about that 'Kubek z Tuxem'? 1117016662 M * matti Bertl: Which one? 1117016688 M * Bertl I only see one labeled like that, actually ;) 1117016699 M * matti Yep :P 1117016707 M * matti Hm, looks nice. 1117016716 M * matti It's not Think Geek. 1117016726 M * matti But I meet owner of this store personly. 1117016744 M * Bertl yeah, nice store, and thankfully it has pictures and sizes ;) 1117016752 M * matti :) 1117016769 M * Bertl guess I learned a few new words today ;) 1117016802 M * matti Kubek? :] 1117016802 M * eyck oczywi¶cie. 1117016809 M * matti eyck: LOL :P 1117016826 M * matti eyck: Chcesz tez cos z Linux stuff? 1117016924 M * eyck a maja koszulke z vserverem? 1117016927 M * rs odd language, what is it ? :) 1117016950 M * eyck swahili 1117016953 M * matti eyck: Niestety nie. 1117016978 M * matti rs: Polish. 1117016989 M * matti rs: And it's not so hard to learn ;] 1117016992 M * matti rs: Ask Bertl :P 1117017013 M * matti eyck: Jak cos, to daj znac. 1117017040 M * rs yeah maybe, seems to be easy to write, you just punch your keyboard and that's it :) 1117017048 M * Bertl lol 1117017072 M * eyck easier then german, 1117017103 M * Bertl really? 1117017108 M * matti Yep. 1117017119 M * matti And than french ;] 1117017124 M * matti :P 1117017141 M * eyck oui 1117017190 M * rs yeah, It's been 26 years that I've been trying to learn french... :) 1117017230 M * matti ;p 1117017544 M * rs Bertl: which debug options do you want ? 1117017558 M * rs vserver history tracing ? 1117017577 M * Bertl yup, verbose_debug, spinlock_debug 1117017589 M * Bertl and of course vserver_debug 1117017591 M * rs btw, what is the Tag NFSD User Auth and Files ? 1117017609 M * Bertl that's for making nfsd xid aware ;) 1117017643 M * rs ok for filers then 1117017652 M * rs good to know :) 1117017661 M * Bertl yup 1117017690 M * rs did you make changes in the server side nfs code ? 1117017924 M * Bertl yes, basically dis and enabling those code 1117017934 M * rs ok 1117017948 M * Bertl s/those/this/ 1117017952 M * rs and about swap emulation, did you start something ? 1117017966 M * Bertl page in/out is accounted, that's it so far 1117017993 M * rs ok 1117018075 M * rs it's the main limitation of linux-vserver today, many users are locked out their vserver because it can't fork new process like... sshd when their vserver is out of memory 1117018113 M * Bertl well, that's basically a question of raising the limits, no? 1117018172 M * rs not really, raising the limit will give them more memory (for free?) and then... their badly setup software will take more memory :) 1117018231 M * Bertl well, and we already talked about swap, the only useful method to hadnle this is penalization 1117018249 M * rs yeah and I think it's a good idea 1117018282 M * Bertl basically one could do that already from userspace 1117018320 M * rs and something confusing is that swap values in a vserver aren't 0 1117018375 M * Bertl it's the difference from vm to rss, clamped to the real swap space 1117018468 M * rs my kernel is ready 1117018487 M * Bertl excellent! 1117018819 M * rs ok, we can start to work :) 1117018839 M * Bertl good, first let's get a guest which fails to reboot 1117018849 M * Bertl (from inside) 1117018866 M * Bertl after that happened, please upload the output of 1117018872 M * Bertl - vps auxwww 1117018894 M * Bertl - cat /proc/virtual/%xid/* 1117019081 M * rs there is now /proc/virtual/xid/ directory after reboot from inside 1117019107 M * rs the reboot stop the vserver correctly 1117019111 M * Bertl s/now/no/ hmm, then it should have worked so far, no? 1117019111 M * rs but doesn't restart it 1117019130 M * Bertl good, so 'just' the startup is not happening ... 1117019138 M * rs yeah 1117019145 M * Bertl that's better than a context which is hanging around, no? 1117019151 M * rs yeah sure 1117019167 M * Bertl okay, any output in the logs? 1117019173 M * Bertl (vshelper logs) 1117019230 M * rs hmmm the log file doesn't exist 1117019235 M * rs could be the pb 1117019242 M * Bertl check that $__CONFDIR"/.defaults/apps/vshelper/logfile is configured 1117019253 M * rs logfile -> /var/log/vshelper.log 1117019257 M * rs but no file exists 1117019277 M * Bertl okay, touch it, try again please ... 1117019314 M * Bertl and check that your /sys entry is pointing to the right vshelpe 1117019377 M * rs Wed May 25 13:08:53 CEST 2005: vshelper startup 117 1117019377 M * rs No responsible vserver found for xid '117' (117); aborting... 1117019377 M * rs Wed May 25 13:09:20 CEST 2005: vshelper restart 117 1117019377 M * rs Restarting vserver '/etc/vservers/194.98.30.253' 1117019410 M * Bertl okay, the startup can be 'eliminated' by doing the same as for the shutdown 1117019423 M * rs ok 1117019428 M * Bertl but it should not really matter 1117019470 M * Vudumen hmmhmm. as i read freevps has virtual interfaces. does it the same as vserver + ngnet? 1117019498 M * Bertl no, as far as I know, they created 'virtual' ip stacks 1117019524 M * Bertl which is probably the more generic approach, but also has higher overhead 1117019533 M * Vudumen thanks 1117019699 M * Bertl but it would be interesting to compare current linux-vserver with freevps 1117019712 M * rs Wed May 25 13:14:50 CEST 2005: vshelper restart 117 1117019712 M * rs Restarting vserver '/etc/vservers/194.98.30.253' 1117019713 M * rs vserver '194.98.30.253' already running 1117019713 M * rs Wed May 25 13:14:56 CEST 2005: vshelper shutdown 117 1117019726 M * Bertl as I expected ;) 1117019784 M * Bertl could you try the following: 1117019830 M * Bertl modify the vshelper (restart action) so that it 1117019861 M * Bertl spawns a command which doesn't do anything but 1117019870 M * Bertl - sleep for 10 seconds 1117019885 M * Bertl - output something to the log (could be syslog) 1117019916 M * Bertl what I'm looking for is that this message should appear _after_ the shutdown call 1117019959 M * rs where should I put this command ? before or after the restart handling ? 1117019980 M * Bertl best, right after the logging, before the killContext 1117019987 M * rs ok 1117020124 M * BWare morning 1117020131 M * Bertl morning BWare! 1117020165 M * rs Wed May 25 13:22:20 CEST 2005: vshelper restart 117 1117020165 M * rs Restarting vserver '/etc/vservers/194.98.30.253' 1117020166 M * rs test 1117020166 M * rs vserver '194.98.30.253' already running 1117020166 M * rs Wed May 25 13:22:36 CEST 2005: vshelper shutdown 117 1117020176 M * rs test is the message 1117020235 M * BWare ah interesting... someone else with the same problem I had with vshelper ;) 1117020246 M * BWare s/had/have 1117020273 M * Bertl rs: hmm, so what happened to the sleep? 1117020319 M * rs hmm what do you mean ? 1117020355 M * Bertl sec trying to reproduce this on the quad 1117020414 M * rs I just did that : 1117020416 M * rs logging $"Restarting vserver '$VSERVER'" 1117020416 M * rs sleep 10 1117020417 M * rs logging $"test" 1117020417 M * rs spawn killContext "$XID" 1117020419 M * rs execute $_VSERVER --defaulttty "$VSERVER" restart 1117020514 M * rs food quest... brb 1117020523 M * Bertl ah, okay, ;) 1117020715 M * Vudumen Bertl: which is the latest one kernel i should use? (i'd like to test ngnet too) 1117020731 M * Bertl 2.6.11.10-vs2.0-rc2 1117020742 M * Bertl you have to patch ngnet ontop of that 1117020749 M * Vudumen ok. 1117020884 M * Vudumen Bertl: i'm looking in the Experimantal/NGNET dir but which is the patch i have to use? 1117020946 M * Bertl sec, I'll upload an up-to-date one 1117020980 M * Vudumen thanks :) 1117021036 M * Bertl ah, this one should work fine: 1117021037 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-vs2.0-pre4-ng9.5.diff 1117021062 M * Vudumen okay 1117021062 M * Bertl if you ahve issues applying it, let me know, I'll have a look at it then ... 1117021084 M * Vudumen it's applied fine :) 1117021089 M * Bertl but be careful with ngnet, you have to met certain conditions (see howto) 1117021094 M * Vudumen well currently is the latest util-vserver is .207? 1117021102 M * Vudumen i'm reading it :) 1117021106 M * Bertl i.e. no multiple tables and ipv6 for now 1117021165 J * jsambrook ~jsambrook@aelfric.plus.com 1117021169 P * jsambrook 1117021199 M * Vudumen oooo well, what doyou mean multiple tables? 1117021210 M * Vudumen arp/bridge/nat tables? 1117021213 M * Vudumen only filter? 1117021239 M * Bertl basically it's best you keep as close as possible to the given configs 1117021250 M * Vudumen ok 1117021257 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/kernel-ng8.config 1117021259 M * Bertl for example 1117021288 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/config-2.6.11-rc3-vs1.9.4-rc5-ng9.0 1117021294 M * Bertl (newer one ;) 1117021359 M * Vudumen :) 1117021404 M * Vudumen Bertl: do you think if it's possible to get this more general virtula ip stack solution from freevps and merge it to vserver to give users the possibility to choose from this 2 solutions? 1117021663 M * Bertl well, if somebody manages to understand the freevps code, sure ... 1117021704 M * Bertl first problem would be: freevp is still 2.4 IIRC 1117021712 M * Vudumen ahh :( 1117021723 M * Bertl then, the code is very entangled with the rest of freevps 1117021743 M * Bertl and finally, it is not easy to read freevps code ;) 1117022163 J * as as@openbug.org 1117022190 M * Bertl welcome as! 1117022194 M * as hi 1117022298 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com 1117022306 M * Bertl welcome ntrs! 1117022337 M * Bertl as: hmm, bsd? 1117022349 M * as Bertl: what version would u sugest for a vserver for some customers (small lamp servers) ? i have curently the stable running, but the new releases semms more secure. 1117022363 M * as thats a box of a friend :) 1117022459 M * as would u say the devel verison is stable enoug if i know what i do ? 1117022487 M * Bertl we are currently at vs2.0-rc2 (you know what the rc2 means?) 1117022510 M * as so when do u think ur releasing it ? 1117022539 M * Bertl pretty soon I guess ... 1117022556 M * as ok that sounds good 1117022763 M * Bertl of course, as usual there is no guarantee that it won't break, but it looks really good actually ... 1117022804 M * rs re 1117022810 M * Bertl wb, rs ... 1117022927 M * as i trust in ur work. a friend of mine has the devel running without any problems. and i had no probs with the stable branch. so i give the rc2 a try :) 1117022975 M * Bertl k, let me know how it works for you ... 1117022997 M * as ok 1117023155 M * Bertl rs: undo your changes, and use this one (you might have to adapt it a little) 1117023157 M * Bertl http://vserver.13thfloor.at/Experimental/vshelper-hack-01.diff 1117023180 M * rs ok 1117023222 M * Bertl note the vwait which is in /tmp ;) 1117023310 M * rs yeah, I saw it and fixed it 1117023316 M * rs ok then I retry ? 1117023357 M * Bertl then you start your vserver guest, and issue a reboot -f from inside 1117023403 M * rs -f ? 1117023425 M * rs I usually use reboot alone 1117023438 M * Bertl both should do, if you have init inside 1117023444 M * rs yeah 1117023556 M * rs Wed May 25 14:18:52 CEST 2005: vshelper restart 117 1117023556 M * rs Restarting vserver '/etc/vservers/194.98.30.253' 1117023556 M * rs Wed May 25 14:18:52 CEST 2005 restartContext started 1117023556 M * rs Wed May 25 14:18:52 CEST 2005 killContext started 1117023556 M * rs Wed May 25 14:18:52 CEST 2005 restartContext now restarting 1117023559 M * rs Wed May 25 14:18:52 CEST 2005 killContext finished 1117023561 M * rs Wed May 25 14:18:52 CEST 2005: vshelper shutdown 117 1117023564 M * rs WARNING: The hostname is now configured in 'uts/nodename' but not in 1117023566 M * rs 'hostname'. 1117023569 M * rs RTNETLINK answers: File exists 1117023571 M * rs Wed May 25 14:18:53 CEST 2005: vshelper startup 117 1117023574 M * rs the vserver is restarted correctly 1117023699 M * Bertl the RTNETLINK is funny 1117023707 M * Bertl ah, but expected ;) 1117023737 M * Bertl well, it's a gross hack anyway .. but it shows how it is supposed to work ... 1117024054 M * Bertl BWare: what problems did you encounter with the vshelper? 1117024153 M * BWare reboot within the vserver caused a shutdown 1117024228 M * Bertl okay, probably same issue then ... 1117024235 M * bro good afternoon all ;> 1117024243 M * BWare probably 1117024251 M * Bertl welcome bro! 1117024252 M * BWare noon bro 1117024296 M * BWare Bertl: is it normal that setting rlimits still requires the host to set ulimits prior to the vserver start ? 1117024329 M * Bertl rlimits or ulimits and from inside or outside? 1117024335 M * BWare heh 1117024338 M * BWare rlimits 1117024341 M * BWare 2.6.11 1117024359 M * BWare so /etc/vservers/NAME/rlimits/nofile 1117024362 M * Bertl doesn't have anything to do with the ulimits 1117024368 M * BWare hmmm 1117024382 M * BWare oke, so how do I specify ulimits in 2.6 then ? 1117024425 M * BWare as I understood from the util-vserver docs, the tools don't support /etc/vservers/NAME/ulimits anymore 1117024435 M * BWare on a 2.6 based host 1117024449 M * BWare or is that my misinterpretation of the docs ;) 1117024471 M * Bertl well, it is most likely a 'bug' in the tools ... 1117024512 M * rs is enrico still working on the util-vserver project ? 1117024525 M * Bertl http://savannah.nongnu.org/bugs/?func=detailitem&item_id=12840 1117024549 M * Bertl rs: well, I hope so ... 1117024576 M * Bertl but he has basically _zero_ presence here ... unfortunately 1117024579 M * BWare Bertl: Oke that clears the fog :) 1117024622 M * Bertl it doesn't make _so_ much sense with 2.6 1117024622 M * Vudumen Bertl: do it really necessary to use the legacy apis? (kernel api, networking api) 1117024646 M * Bertl BWare: because with the proper ccaps you can change the rlimits as root inside the guest 1117024668 M * Bertl (i.e. the context 'rlimits' did take over the function of the ulimits) 1117024692 M * Bertl Vudumen: no for the kernel api, yes for the network api without ngnet 1117024709 M * Vudumen Bertl: fine. thanks. 1117024736 Q * ntrs Quit: Leaving 1117024751 M * BWare Bertl: It is 'impossible' for the root in the guest to change his ulimits 1117024787 M * Bertl no, on 2.6 it is quite possible to change them ... 1117024828 M * BWare So I to specify the ulimits before starting the guest and the guest 'inherits' the host's ulimits 1117024838 M * Vudumen Bertl: does it have any sense if i enable other targets in iptables than what is enabled in the example config you gave me? 1117024846 M * BWare Bertl: With the proper ccaps 1117024848 M * Bertl BWare: yes, unless some scripts inside change it 1117024880 M * Bertl Vudumen: you can enable most of them, as long as they are not related to connection tracking or multiple tables 1117024897 M * Vudumen so anything what is realted to conntection tracking should not be used. am i right? 1117024901 M * Vudumen related 1117024922 M * Bertl let me put it this way, if it compiles, you can use it ;) 1117024988 M * Vudumen okay :) 1117024992 M * Vudumen i try it :) 1117025023 M * rs bertl: maybe we could try to search someone to co-maintain the project 1117025046 M * Bertl well, we did put up the svn repository for folks to contribute 1117025065 M * eyck what svn repository? 1117025083 M * rs savannah is using subversion ? 1117025084 M * Bertl and if you look at the ML thread, you will see that enrico was not too happy about it 1117025096 M * rs why that ? 1117025101 M * Bertl is anybody reading the ML, bt? 1117025109 M * Bertl s/bt/btw/ 1117025113 M * rs not those days unfortunatly 1117025205 M * Bertl sec, looking for the thread id 1117025248 M * Bertl http://www.paul.sladen.org/vserver/archives/200505/0022.html 1117025270 M * Bertl http://list.linux-vserver.org/archive/vserver/msg09638.html 1117025276 M * Bertl (whatever you prefer ;) 1117025293 M * rs subversion does support changset 1117025327 M * Bertl well, svn not, svk does IIRC 1117025344 M * rs it depends what you call changeset then 1117025370 M * Bertl anyway, everybody except enrico was happy about it ;) 1117025608 N * Doener|gone Doener 1117025616 M * Bertl wb Doener! 1117025618 M * Doener greetings 1117025664 M * Vudumen hmmhmm 1117025688 M * Vudumen Bertl: is it possible that it doesn't compile because the conntrack? a moment and i show you the output 1117025715 M * Vudumen http://inferno.interware.hu/1.txt 1117025791 M * Bertl hmm, seems to be a missing include ... sec 1117025843 M * Bertl this is vs2.0-rc2 ontop of 2.6.11.10, yes? 1117025850 M * Vudumen yes 1117025852 M * Vudumen plus ngnet 1117025904 M * Bertl hmm ... the patch doesn't apply quite that fine here, did it apply cleanly for you? 1117025920 M * Bertl (patch-vs2.0-pre4-ng9.5.diff) 1117026001 M * Bertl the failed hunks are basically: 1117026013 M * Bertl in ./net/core/sock.c, remove 1117026016 M * Bertl - // BUG_ON(sk->sk_nx_info); 1117026017 M * Bertl - clr_nx_info(&sk->sk_nx_info); 1117026017 M * Bertl - sk->sk_nid = -1; 1117026029 M * Bertl (around line 674) 1117026034 M * Vudumen it was fine here 1117026051 M * Vudumen vanilla 2.6.11.10 + vs2.0-rc2 + ngnet9.5 1117026064 M * Bertl then you did something wrong, I guess 1117026069 M * Vudumen maybe :) 1117026074 M * Vudumen i check out this part. a moment. 1117026081 M * Vudumen ahh 1117026082 M * Vudumen yes 1117026091 M * Vudumen i just didn't see the error message :( 1117026116 M * Bertl I'll upload a modified patch in a minute 1117026123 M * Vudumen ok. thanks :) 1117026370 M * Vudumen until this i go and drink a coffee :) 1117026378 M * Bertl yes, make it so ;) 1117026386 M * Bertl test compiling now 1117026538 J * yarihm ~yarihm@80-218-5-16.dclient.hispeed.ch 1117026543 M * Bertl welcome yarihm! 1117026564 M * yarihm yo Bertl 1117026810 M * Bertl Vudumen: http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.11-vs2.0-rc2-ng9.5.diff 1117027260 M * Bertl rs: IIRC, you had some perl magic to hog memory ... 1117027816 M * FaUl yippeee 1117027827 M * FaUl finaly get behind this fscking scsi-controller-problems 1117027858 M * FaUl note to self: direct pci-access method is not always save 1117027876 M * Bertl hmm .. how unexpected ;) 1117027904 M * FaUl Bertl: i'd never problems with that before 1117027914 M * FaUl in 2.6.9 it works with direct-access btw 1117027933 M * Bertl really? interesting ... 1117027948 M * Bertl maybe you should report that on lkml? 1117028003 M * FaUl Bertl: maybe i'll report that on linux-scsi 1117028038 M * FaUl anyway - now i'm going to t est 2.6.11.10-vs2.0-rc2-ng9.5 :-) 1117028118 M * Bertl cool! 1117028447 M * FaUl which options should I enable in 2.6.11.10? 1117028505 M * FaUl do you need vserver debugging code? 1117028561 M * Vudumen Bertl: thanks i try it 1117028813 M * Bertl FaUl: if you test ngnet, you should stick close to the ngnet configs 1117028840 M * Bertl FaUl: you do not _need_ vserver debugging, but as you are testing, it might come handy sooner or later 1117028857 M * Doener Bertl: memory hog, for a short time, or a persistant hog? 1117028865 M * FaUl is this only valid for the entire kernel or only for the vserver part of the config? 1117028901 M * Bertl Doener: something which hits the limit would suffice (maybe with output how much memory it got) 1117028925 M * Bertl FaUl: with ngnet for all network related options 1117028947 M * FaUl Bertl: btw: do i need policy-routing with ngnet for multiply gateways or is this only valid for the legancy chbind? 1117028965 M * Doener to just hit some value i usually do: head -c100m /dev/full | sed "s/.//" > /dev/null 1117028999 M * Doener sed will use about _twice_ the memory you get from head, i.e. 200m in this case... 1117029027 M * Doener annoying side effect: it's also a cpu hog ;) 1117029059 M * Bertl sounds good to me ... 1117030133 M * FaUl Bertl: topic: s/ng9.4/ng9.5/ :) 1117030450 M * FaUl CC net/ipv4/fib_frontend.o 1117030450 M * FaUl net/ipv4/fib_frontend.c:65: error: conflicting types for `__fib_new_table' 1117030450 M * FaUl include/net/ip_fib.h:204: error: previous declaration of `__fib_new_table' 1117030450 M * FaUl net/ipv4/fib_frontend.c: In function `fib_flush': 1117030451 M * FaUl net/ipv4/fib_frontend.c:90: error: too many arguments to function `fib_get_table 1117030458 M * FaUl hmpf 1117030557 M * Bertl multiple tables? 1117030574 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc2, ng9.5 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1117030665 M * Bertl FaUl: here is my test config: http://vserver.13thfloor.at/Experimental/NGNET/config-2.6.11.10-vs2.0-rc2-ng9.5 1117030845 M * FaUl ah, I should *NOT* set multiple tables ;-) 1117030866 M * Bertl yup :) 1117030897 M * FaUl Bertl: do you have time? than i'll steal your time with more compiler-failures if i include v6-support.. 1117030926 M * Bertl yeah, go ahead, I'm used to it ... 1117030955 M * FaUl mh, but i don't know when a friend of mine comes to fetch me for learning maths :) 1117030968 M * FaUl (well, and i've to fix his computer of course :) 1117030989 M * FaUl CC net/ipv6/af_inet6.o 1117030989 M * FaUl net/ipv6/af_inet6.c: In function `inet6_bind': 1117030989 M * FaUl net/ipv6/af_inet6.c:285: error: too few arguments to function `inet_addr_type' 1117031014 M * FaUl Bertl: is it importent to know that i'm building an smp-kernel btw? 1117031021 M * Bertl seriously, ipv6 is not suppoted yet ... 1117031045 M * Bertl but we can add/address that once you got ipv4/ngnet working 1117031051 M * daniel_hozac FaUl: wasn't that fixed in DaPhreak's patch? 1117031062 M * FaUl daniel_hozac: where is DaPhreaks patch? 1117031087 M * Bertl the patch only papered over the issues ... 1117031109 M * Bertl i.e. a 'partial' implementation which will compile but not work 1117031128 M * daniel_hozac ah, ok. 1117031154 M * FaUl Bertl: so you suggest to switch ipv6-support off and get my vservers running with ngnet first? ok, ill do that 1117031164 M * Bertl yep, definitely ;) 1117031171 M * Bertl SMP/SMT is fine by the way 1117031209 M * FaUl thats good in any case :) 1117031226 M * Bertl you can also use s390 if you like ;) 1117031265 M * FaUl Bertl: i would like if i had one 1117031320 M * Bertl don't have one either, but look at the bottom of http://linux-vserver.org/Syscall+Switch+Info 1117031328 M * rs bertl: I have a blocked vserver for you 1117031336 M * rs if you want try to investigations 1117031342 M * Bertl rs: good, details? 1117031354 M * rs vps aux|grep 84.244.4.236 1117031355 M * rs root 7562 1213 84.244.4.236 0.0 0.0 0 0 ? Z May24 0:00 [java] 1117031358 M * rs root 7569 1213 84.244.4.236 0.0 0.0 0 0 ? Z May24 0:00 [java] 1117031371 M * rs there is two process blocked in it 1117031386 M * Bertl okay, this is with unmodified vshelper I guess? 1117031393 M * rs yes 1117031396 M * rs and rc1 1117031409 M * rs status report 16 tasks 1117031436 M * Bertl did you notice the changed order in killContext ? 1117031453 M * rs no 1117031458 M * Bertl (in my hack patch) 1117031462 M * rs ah ok 1117031463 M * rs yes 1117031469 M * rs PROC: 16 126 1000 0 1117031469 M * rs VM: 103656 351033 2560000 0 1117031469 M * rs VML: 0 0 -1 0 1117031469 M * rs RSS: 3715 65536 65536 54 1117031469 M * rs ANON: 1308 24686 -1 0 1117031472 M * rs FILES: 55 984 5000 0 1117031475 M * rs OFD: 6 448 -1 0 1117031475 M * Bertl it _could_ be related to this ... 1117031477 M * rs LOCKS: 0 13 -1 0 1117031480 M * rs SOCK: 0 56 -1 0 1117031482 M * rs MSGQ: 0 0 -1 0 1117031485 M * rs SHM: 0 2729 -1 0 1117031496 M * Bertl what does /proc/virtual/%xid/status show? 1117031507 M * rs UseCnt: 19 1117031507 M * rs Tasks: 16 1117031507 M * rs Flags: 00100000030b0310 1117031507 M * rs BCaps: ffffffffd44c04ff 1117031508 M * rs CCaps: 0000000000010103 1117031510 M * rs Ticks: 0 1117031524 M * Bertl okay, which context id is that? 1117031530 M * rs 1213 1117031550 M * Bertl do you see any tasks 'hanging' around on the host which refer to that xid? 1117031569 M * Bertl ps auxwww | grep 1213 1117031584 M * rs no nothing 1117031683 M * Bertl okay, what does 'vkill --xid 1213 -s 9 -- 0' do? 1117031705 M * rs nothing :) 1117031745 M * rs hanging processes are still there 1117031761 M * Bertl but you can fix that by entering the context and killing them? 1117031768 M * rs yeah 1117031778 M * Bertl really interesting ... 1117031796 M * rs a simple vcontext --migrate --xid 1213 kill -9 7562 7569 should fix this 1117031809 M * Bertl try with one of them first 1117031813 M * rs ok 1117031825 M * rs worked 1117031833 M * rs the other one is still there 1117031842 M * Bertl okay, yeah, makes sense ... 1117031858 M * Bertl the vkill only delivers to 'running' processes 1117031873 M * rs hehe :) 1117031883 M * rs and those process are already dead 1117031890 M * rs but nobody is there to clean them 1117031891 M * Bertl let's try with a direct vkill 1117031905 M * Bertl vkill --xid 1213 -s 9 -- 7562 1117031911 M * Bertl (or the other one) 1117031932 M * rs didn't work 1117031941 M * Bertl okay, that confirms my theory 1117031953 M * rs vkill sucks ? :P 1117031973 M * Bertl :) 1117032008 M * rs do you have something else to do on this one, I have to fixe it, the customer is waiting or his vserver :) 1117032017 M * rs ? 1117032031 M * Bertl yeah, please save the 'new' limits 1117032039 M * rs new limits ? 1117032054 M * rs there is something new in limites ? 1117032057 M * Bertl /proc/virtual/1213/{limits,status} 1117032106 M * rs PROC: 8 126 1000 0 1117032107 M * rs VM: 51727 351033 2560000 0 1117032107 M * rs VML: 0 0 -1 0 1117032107 M * rs RSS: 1703 65536 65536 54 1117032107 M * rs ANON: 543 24686 -1 0 1117032110 M * rs FILES: 26 984 5000 0 1117032112 M * rs OFD: 3 448 -1 0 1117032115 M * rs LOCKS: 0 13 -1 0 1117032117 M * rs SOCK: 0 56 -1 0 1117032120 M * rs MSGQ: 0 0 -1 0 1117032122 M * rs SHM: 0 2729 -1 0 1117032125 M * rs UseCnt: 11 1117032127 M * rs Tasks: 8 1117032130 M * rs Flags: 00100000030b0310 1117032132 M * rs BCaps: ffffffffd44c04ff 1117032135 M * rs CCaps: 0000000000010103 1117032137 M * rs Ticks: 0 1117032140 M * rs sorry for the flood :) 1117032246 M * Bertl np 1117032265 M * Bertl okay, kill them 1117033133 M * matti LOL 1117033184 M * matti I just saw store, with t-shirts "European users piss-off 18 U.S.C. 2319". 1117033187 M * matti LOLLLLLLL 1117033188 M * matti ;D 1117033237 M * Loki|muh öhm? could you explain? 1117033273 M * matti The "18 U.S.C. 2319" is USA copyright act. 1117033283 M * matti Ya know. 1117033293 M * Loki|muh ah yeah 1117033312 M * matti If you've some mp3, then in USA you make violation those statement ;] 1117033329 M * matti ;] 1117033330 M * Loki|muh nice *g* 1117033332 M * matti Hehehe. 1117033707 M * Doener Bertl: hmm... the "16:31:47 rs Tasks: 16" is probably caused by threading, right? 1117033720 M * Bertl yup 1117033730 M * Bertl 8 threads each java thingy 1117033747 M * Bertl so it's probably somehow related to multithreading 1117033764 M * Bertl I was expecting the folowing sequence to leave a zombie behind: 1117033779 M * Bertl vcontext --create --xid 101 -- vcontext --xid 101 --migrate-self --initpid --endsetup -- bash -c "sleep 100 &" 1117033792 M * Bertl but it doesn't ... 1117033798 M * Doener hm, i have some vservers with apache2(worker).... never had a problem 1117033818 M * Bertl rs: could you try this sequence (with sleep 10 or so) on your server? 1117033900 J * monrad ~monrad@213083190130.sonofon.dk 1117033908 M * Bertl evening monrad! 1117033927 M * monrad evening 1117033935 M * rs Bertl: sure 1117033941 M * rs and then ? 1117033959 M * Bertl well, either the sleep zombifies or not ;) 1117033996 M * rs I guess I have to wait 100 seconds :) 1117034015 M * Bertl that's why I said with sleep 10 or so ;) 1117034021 M * rs the fact is that it's always a threaded process that is blocked 1117034052 M * rs ok I jumped the thing inside () :) 1117034101 M * rs the process died cleanly 1117034113 M * Bertl almost expected ... 1117034128 M * Bertl okay, I'm off for now ... back later ... will think about it ... 1117034134 M * rs ok 1117034138 M * rs see you 1117034139 M * matti Uh. 1117034156 M * matti albeiro: My new link works perfectly. 1117034158 M * Bertl Doener: interesting details: chcontext+kill works, vkill fails ... 1117034167 M * Doener yep, seen that... 1117034178 M * rs hello doener btw :) 1117034181 M * Doener got some vague idea... need to check code... 1117034182 M * Doener hi rs 1117034188 M * matti albeiro: SAF CFM-8-MUX, 6 Mbps radio link at 22.7 GHz ;) 1117034199 M * Doener rs: 11.9-vs2.0-rc1? 1117034203 M * Bertl yeah, kill_something_info != send_sig_info 1117034204 M * rs yeah 1117034211 M * Doener k 1117034215 M * Bertl okay, off now ... 1117034218 M * matti Wait! 1117034221 M * matti Just a sec ;] 1117034223 M * Bertl hmm? 1117034265 M * matti Bertl: Look at query.. 1117034352 M * Bertl k, cya 1117034355 N * Bertl Bertl_oO 1117035628 J * ntrs ntrs@Dardeene-68.188.50.87.charter-stl.com 1117035761 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it 1117036398 J * DuckMaster ~duckx@195.75.27.158 1117036709 J * Doener` ~doener@p54877BD0.dip.t-dialin.net 1117036947 J * mep__ mep@p5091EC6E.dip.t-dialin.net 1117037126 Q * Doener Ping timeout: 480 seconds 1117037386 Q * mep_ Ping timeout: 480 seconds 1117037960 M * albeiro matti: nice :) 1117038444 J * knoppix_ ~knoppix@dsl-082-082-080-120.arcor-ip.net 1117038783 Q * cryo Remote host closed the connection 1117040049 Q * ntrs Quit: Leaving 1117041531 Q * prae Quit: Client exiting 1117042539 Q * DuckMaster Quit: Leaving 1117043440 Q * rs Quit: rs 1117044362 M * daniel_hozac TheSeer: i updated the FC3 howto with a reference to my kernel RPMs. 1117044541 M * TheSeer kewl :) 1117044559 M * TheSeer if you have more additions, feel free to add them ;) 1117044611 M * TheSeer you should make that apt-able 1117044636 M * daniel_hozac apt is evul ;) 1117044637 M * TheSeer then we can provide a subsection on how to add your resource into apt and install the packages from there 1117044651 M * daniel_hozac any reason why you don't use yum? 1117044652 M * TheSeer apt is one of the best tools ever invented ;> 1117044658 M * TheSeer because yum suxx 1117044663 M * TheSeer it's slow like hell 1117044680 M * TheSeer and because vserver-utils has a vapt-get wrapper 1117044713 M * daniel_hozac util-vserver also have a vyum wrapper. 1117044734 M * daniel_hozac (and the apt on the page is a release old ;)) 1117044775 M * TheSeer well.. i used yum for a while on my desktop maschine 1117044790 M * daniel_hozac when? 1117044805 M * TheSeer but after it took like ages to remove something using yum because it for whatever reason kept checking for updates packages i removed it 1117044818 M * TheSeer fc2 and "early" fc3 1117044827 M * daniel_hozac FC3 wouldn't do that. 1117044842 M * TheSeer well.. i like apt 1117044854 M * TheSeer does yum have a dist-upgrade? 1117044858 M * TheSeer i never tried ,> 1117044868 M * daniel_hozac what's the difference? 1117044890 M * TheSeer basically an upgrade from one distro-version to another 1117044896 M * daniel_hozac dist-upgrade processes obsoletes? 1117044905 M * TheSeer guess so.. 1117044911 M * daniel_hozac yum always does that. 1117044924 M * TheSeer i apt-get dist-upgraded one of my servers from fc1 to fc3 1117044927 M * daniel_hozac (well, unless obsoletes is set to 0 ;)) 1117044942 M * TheSeer worked like a charm 1117044956 M * daniel_hozac i yum update'd one of mine from FC2 to FC3, which also worked like a charm. 1117044975 M * TheSeer what i didn't like at yum was the time it took for downloading package lists 1117044983 M * TheSeer apt downloads one gz file 1117044990 M * TheSeer yum downloaded a trillion small files 1117044997 M * TheSeer taking forever to load 1117045005 M * TheSeer maybe the newer yum is better 1117045006 M * daniel_hozac that's yum-2.0 1117045017 M * TheSeer but the version i tried sucked big time 1117045021 M * daniel_hozac yum-2.1+ uses the repomd format. 1117045054 M * daniel_hozac (one tiny XML file and one compressed larger XML file) 1117045063 M * TheSeer sounds good 1117045070 M * TheSeer well.. maybe i'll try yum again with FC4 ;> 1117045107 M * TheSeer if you make your rpms apt-able/yum-able we can rewrite the howto for FC4 ,) 1117045117 M * daniel_hozac they're already yum-able ;) 1117045123 M * TheSeer then add apt 1117045153 A * daniel_hozac yum's apt 1117045161 M * TheSeer hehe 1117045182 M * daniel_hozac also, why aren't you using the apt from Extras? 1117045198 M * daniel_hozac i think Panu is the one maintaining apt-rpm these days. 1117045213 M * TheSeer because by the time i wrote that howto, extras wasn't offically available 1117045220 M * daniel_hozac well, fedora.us was ;) 1117045245 M * TheSeer well.. freshrpms.net just happend to ever work for me 1117045260 M * TheSeer fast and reliable downloads 1117045260 M * daniel_hozac are regular Fedora mirrors even apt-able? 1117045270 M * daniel_hozac or is it just freshrpms and fedora.us? 1117045276 M * TheSeer dunno.. 1117045288 M * TheSeer dag's stuff is too ;) 1117045299 M * Doener` at least the mirror at some german university at apt-able 1117045319 M * daniel_hozac is that a Fedora mirror or a fedora.us mirror? 1117045398 M * Doener` how do i know? i installed the whole fc3 vserver from that mirror 1117045414 M * daniel_hozac well, the directory structure is different. 1117045451 M * Doener` ftp://sunsite.informatik.rwth-aachen.de/pub/Linux/fedora-core/ 1117045486 M * Doener` eek, wrong one... 1117045499 M * Doener` http://sunsite.informatik.rwth-aachen.de/ftp/pub/Linux/fedora/fedora/ 1117045540 M * daniel_hozac yep, that's a fedora.us mirror. 1117045545 M * daniel_hozac the first one is a Fedora mirror ;) 1117045553 M * Doener` ok 1117048275 M * daniel_hozac TheSeer: updated to 2.0-rc2 and it should be apt-able now. 1117048343 M * FaUl re 1117048492 M * FaUl hmm, how can i figure out which version of vserver i am running 1117048512 M * daniel_hozac the patch? 1117048517 M * daniel_hozac uname usually tells you. 1117048552 M * FaUl no, vserver-utlis 1117048570 M * FaUl i'm quite sure that i'm using 2.0-rc2-ngnet9.5 :) 1117048644 M * Doener` hm? 1117048658 M * Doener` 2.0-rc2-ngnet9.5 would be a kernel version suffix... 1117048667 M * Doener` thus uname should do... 1117048696 M * FaUl Doener`: i want to find out which version of vserver-userland-tools i'm running 1117048740 M * Doener` vserver-info - SYSINFO 1117048774 A * Doener` thought of "I'm quite sure I'm running X but this and that doesn't behave like it would on X" ;) 1117048811 M * FaUl hmm, i don't have vserver-info 1117048821 M * FaUl so my version is obvously too old 1117048832 M * FaUl .oO( ok - its debian so i expected this ) 1117048840 M * Doener` chcontext --version 1117048860 M * Doener` debian is supposed to have pretty up2date packages atm... 1117048885 M * FaUl 0.30 1117048897 N * Bertl_oO Bertl 1117048908 M * Doener` http://packages.debian.org/cgi-bin/search_packages.pl?searchon=names&version=all&exact=1&keywords=util-vserver 1117048921 M * Bertl evening folks! 1117048934 M * Doener` 0.30.204-5 in Sarge, 0.30.207-6 in sid... 1117048942 M * Doener` evening Bertl! 1117048963 M * FaUl Doener`: oh, ok, then i'll do apt-get update :) 1117048965 M * DaPhreak evening Bertl, Doener`, FaUl, daniel_hozac ... and any other who's here ;) 1117048970 M * FaUl hey DaPhreak 1117048985 M * Bertl stupid question: will Sarge become stable? 1117048998 M * DaPhreak haha, ... once upon a time ;) 1117049024 M * DaPhreak shouldn't it be stable since *think* march or so ? 1117049025 M * Bertl okay, and how will the upcoming release be called? debian release I mean? 1117049032 M * FaUl Bertl: just a few days before "GNU/HURD" is released 1117049055 M * FaUl Bertl: Debian GNU/Linux 3.1 (Sarge) I guess 1117049083 M * Bertl okay, so they basically keep the names, yes? 1117049096 M * Doener` yep 1117049115 M * DaCa except that sid always stays sid 1117049121 M * Doener` sid is always unstable, and once a new testing is created, it gets a name that it will keep forever 1117049129 M * DaCa when sarge releases, etch becomes the new testing 1117049151 M * Bertl who is etch? 1117049167 A * Doener` only watched Toy Story in german :/ 1117049216 M * FaUl DaCa: how is 'experimental' called again? 1117049245 M * Doener` scud 1117049246 M * DaCa isnt that the thing where you make drawings on and erase (etch-n-sketch?) 1117049261 M * Bertl ah, yeah, that's a good point 1117049395 M * Doener` ah, rc5... time for a port ;) 1117049427 M * FaUl hmm, vnet isn't available in sarge, is it? 1117049456 M * Bertl guess not, the whole ngnet should not be ... 1117049478 M * FaUl ok, so i'll compile this on my own 1117049484 M * Bertl Doener`: any ideas regarding the zombie-threads? 1117049486 M * FaUl does it link against dietlibc? 1117049590 M * Doener` Bertl: no... but a) I wonder if we could simply use send_sig instead of creating the sig_info and using send_sig_info and b) ehrm... I just fell out of my brain... *goes searching* 1117049651 M * Bertl FaUl: probably but not necessarily ... 1117049814 M * FaUl Bertl: it does, so i don't have to care about different libc-versions between my gentoo (desktop) and my debian-sarge (server) :-) 1117049832 M * FaUl because i've neither gcc nor libc-dev nor everything like that on the server 1117049955 M * Bertl since doener updated it to use the new interface ... it's even portable and platform agnostic ;) 1117049967 M * FaUl ok, where is the ultimative manual for setting up a vserver other than 1.* - i've never done that before 1117049988 M * FaUl Bertl: very nice 1117050041 M * Bertl http://linux-vserver.org/alpha+util-vserver 1117050049 M * FaUl thx 1117050068 M * Doener` Bertl: i guess you _love_ the first changelog entry for rc5 *g* 1117050080 M * Bertl hmm? 1117050141 M * Doener` " Fix a c99ism." 1117050191 M * Doener` i remember you talking about c99 stuff and portability... never really looked at the c99 stuff yet *hides* 1117050218 M * Bertl ah, well, it's only about 5 and a half year now ;) 1117050255 M * FaUl i like this webdevel-extension of firefox. just switch off css on the 'flower-page' makes it readable again ;) 1117050264 M * Doener` ah, well, I'll do like companies do with windows, only use every second version 1117050282 M * Doener` FaUl: you could as well just use one of the other styles... 1117050286 M * Doener` gras1 is nice 1117050309 M * Doener` (View->Page Style->...) 1117050361 A * Doener` .oO( there _must_ be an hidden option to save the style one wants to be chosen for a page...) 1117050361 M * FaUl i like weedpage 1117050368 M * FaUl Doener`: :-) 1117050387 M * Doener` hey, that's a new one :) 1117050422 M * Doener` yes, that's even better than gras1 1117050551 M * Bertl except for the broken title/heading 1117050603 M * Bertl Doener`: hmm, my galeon automagically stores the last one, it seems 1117050834 M * Bertl hmm, how does debian call the openssh server? 1117050862 M * Bertl (I mean the package) 1117050953 M * Doener` ssh 1117051020 M * Bertl ah, thanks ... 1117051390 A * FaUl is to stupid 1117051423 M * FaUl *read* 1117051917 M * FaUl ah, now 1117052501 M * Bertl hmm, just discovered that debian has a package called lcap ... 1117052518 M * Bertl just to find out that it seems to be broken :/ 1117052800 M * Bertl ah, no, it seems to rely on /proc/sys/kernel/cap-bound 1117053519 M * FaUl diva:/etc/vservers/foobar# vserver foobar start 1117053519 M * FaUl chbind: vc_set_ipv4root(): Function not implemented 1117053533 M * FaUl .OO( of course - i want to use ngnet ) 1117053540 M * Bertl ;) 1117053549 M * FaUl any hint how to start this vserver anyway? 1117053564 M * Bertl you have to disable the interfaces 1117053589 M * Bertl i.e. remove /etc/vservers/%name/interfaces/* 1117053607 M * FaUl i removed even /etc/vserers/%name/interfaces :) 1117053617 M * FaUl but doesn't help either 1117053623 M * Bertl hmm? 1117053649 Q * locksy Ping timeout: 480 seconds 1117053689 M * FaUl diva:/etc/vservers/.defaults/vdirbase/foobar/etc# ls /etc/vservers/foobar/ 1117053689 M * FaUl apps context fstab name run uts vdir 1117053727 M * Bertl hmm, creating a disabled file should be sufficient ... 1117053788 M * Bertl removing interfaces completely is probably not the best idea ;) 1117053789 M * FaUl so a file called /etc/vservers/foobar/interfaces/disabled / 1117053798 M * FaUl so a file called /etc/vservers/foobar/interfaces/disabled ? 1117053807 M * Bertl well, no actually it would be 1117053812 M * Bertl /etc/vservers/foobar/interfaces/*/disabled 1117053823 M * FaUl ok, i'll try this 1117053827 M * Bertl but as you already removed them 1117053836 M * Bertl I would try with an empty interfaces 1117053862 M * FaUl that was my first try - doesn't work either 1117053899 M * FaUl diva:/etc/vservers/foobar/interfaces/*# ls 1117053899 M * FaUl disabled 1117053907 M * FaUl and that problem still exist 1117053926 M * Bertl could you show me a --debug trace (via pastebin.com or so) 1117053975 M * FaUl vserver foobar start --debug? 1117053985 M * Bertl vserver --debug foobar start 1117053996 M * FaUl ah, ok, mom 1117054106 M * Doener` hm, not sure if the tools handle an actual * in the path... 1117054184 M * Bertl did he really? OMG, yes he did! 1117054198 M * Doener` *lol* 1117054222 M * Bertl I'm giving up using wildcards and placeholders ... 1117054291 M * Doener` heh, i also gave up on /path/to/vservers and stick to /vservers now... it ended up in being interpreted as /path/to/vservers/ too often ;) 1117054308 M * FaUl http://faul.bundessicherheitsministerium.de/starttrace 1117054341 M * FaUl Bertl: urgs ;) 1117054373 M * FaUl Bertl: ok. you must say that i have to interpret this wildcard :) 1117054410 M * Bertl FaUl: you're an executive (Beamter)? 1117054511 M * Doener` I somehow doubt that officials would like such a (sub-)domain ;) 1117054563 M * Bertl probably not ;) 1117054613 M * FaUl Bertl: no, i'm not 1117054630 M * FaUl it's just my domain :) 1117054664 M * Bertl well, no idea why it uses chbind ... 1117054710 M * FaUl Doener`: i offer free subdomains for 'bundessicherheitsministerium.de', just say where i should delegate and which 1117054732 M * FaUl Bertl: /me neither do 1117054752 M * FaUl anyway, ill replace chbind with a symlink to /bin/true to work around this problem i guess 1117054771 M * Bertl no, which tool version is that? 1117054786 M * Bertl /bin/true won't work ;) 1117054819 M * FaUl diva:/usr/sbin# vserver-info 1117054819 M * FaUl Versions: 1117054819 M * FaUl Kernel: 2.6.11.10-vs2.0-rc2-ng9.5 1117054819 M * FaUl VS-API: 0x00010025 1117054821 M * FaUl util-vserver: 0.30.204; Apr 10 2005, 21:35:54 1117054846 M * Bertl okay, please upgrade to 0.30.207 1117054863 A * FaUl hates debian 1117054887 A * FaUl upgrades his debian to sid :) 1117054984 M * aba FaUl: just use sarge now. 1117055079 M * FaUl aba: i do 1117055237 M * Doener` Bertl: .207 also seems to unconditionally run chbind AFAICT 1117055376 M * Bertl hmm, I don't understand why chbind without any args would do anything at all? 1117055415 M * Bertl I mean, what is the purpose calling vc_set_ipv4root() with no ips? 1117055451 M * Bertl but yes, you seem to be right on that ... 1117055462 M * Bertl but what version did we use then for ngnet? 1117055646 M * FaUl Kernel: 2.6.11.10-vs2.0-rc2-ng9.5 1117055646 M * FaUl VS-API: 0x00010025 1117055646 M * FaUl util-vserver: 0.30.207; May 22 2005, 21:21:24 1117055647 M * FaUl ok :-) 1117055666 M * FaUl diva:/usr/sbin# vserver foobar start 1117055666 M * FaUl /proc/uptime can not be accessed. Usually, this is caused by 1117055670 M * FaUl procfs-security. *faqread* 1117055679 M * Bertl vprocunhide 1117055886 M * FaUl where do i find vprocunhide? 1117055896 M * Bertl usually in /etc/init.d 1117055913 M * FaUl oh, thx 1117055925 M * FaUl narf 1117055944 M * FaUl now it says the same chbind-foo again 1117055997 M * Bertl Doener`: did we patch the tools back then? 1117056014 M * Doener` no, don't think so... 1117056252 M * Bertl hmm, cap-bound is bound to CAP_SYS_MODULE??? somebody must have been smoking something ;) 1117056393 M * FaUl so any idea how to start this vserver? 1117056435 M * Bertl you probably have to write a small script 1117056455 M * Bertl which ignores args until it finds -- 1117056462 M * Bertl then execs the rest .... 1117056483 A * Bertl is still wondering how we did avoid that back then ... 1117056554 M * Bertl anyway, I'm off to bed now ... pretty tired ... 1117056561 M * Bertl good luck! 1117056564 M * Doener` good night Bertl! 1117056571 M * DaPhreak night Bertl 1117056576 N * Bertl Bertl_zZ 1117056581 M * Bertl_zZ night everyone! 1117056709 M * FaUl n8 Bertl_zZ 1117056729 M * Doener` FaUl: as a workaround, you could also disable the syscall in chbind... 1117056810 M * FaUl Doener`: how? comment it out in the source or what is your suggestion 1117056834 M * Doener` sed -i "s/vc_set_ipv4root(bcast,nbaddrs,ips)!=0/0/" chbind.c 1117056864 M * FaUl Doener`: i guess i'll prefer bertls method ;) 1117057208 M * FaUl chbind is now: #!/bin/sh\n`echo $@|sed 's/^.*-- //'` 1117057222 M * FaUl vshelper.init: can not determine xid of vserver 'foobar'; returned value was '' 1117057236 M * FaUl any suggestions? 1117057275 M * daniel_hozac FaUl: you should probably exec the result. 1117057284 M * FaUl daniel_hozac: i do 1117057287 M * FaUl se backticks 1117057319 M * daniel_hozac that's what generates the result, yes. 1117057328 M * daniel_hozac you'd want exec in front of them. 1117057365 M * FaUl uhm? 1117057389 M * daniel_hozac as in exec `echo $@ | sed 's/^.*-- //'`. 1117057600 M * Doener` without exec -> runs as child of the bash, with exec -> replaces the bash 1117058024 M * FaUl ah, ok 1117058044 M * FaUl but this should not solfe that 'vshelper.init' problem, should it? 1117058603 M * Doener` it actually should... 1117058699 M * FaUl nope 1117058860 A * FaUl will put himself to bed 1117058860 M * FaUl n8 1117058866 M * Doener` night FaUl 1117059316 Q * monrad Ping timeout: 480 seconds 1117060131 J * rs ~rs@imhotep.rhapsodyk.net 1117060147 M * Doener` nighty night folks 1117060151 N * Doener` Doener_zZz 1117060368 J * locksy ~locksy@mrtg.sisgroup.com.au 1117060478 Q * knoppix_ Quit: Verlassend 1117061194 J * tb ~tb@pha-84-242-95-4.nat.karneval.cz 1117061201 M * tb Hi 1117062597 A * mugwump decides to build the new vserver RC, because his immutable linkage invert script `immucp' isn't working for him for some strange reason 1117063081 Q * tb Remote host closed the connection 1117064397 M * mugwump Bah, it was directory immutability again. That always gets me. Wonder if it's a bug in my immucp 1117064406 M * mugwump or just me :) 1117065015 M * mugwump Hmm, no there is definitely some weirdness going on via the legacy API at least... 1117065039 M * mugwump in my 2.6.9-rc3-vs1.9.3-rc2; building a recent one now