1169769694 M * hardwire` siiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiigh
1169769699 N * hardwire` hardwire
1169769702 M * Bertl hmm?
1169769721 M * hardwire gar.
1169769736 M * hardwire just trying to figure out asterisk and /dev/zap* in a vserver
1169769738 M * hardwire and if its a good thing
1169769759 M * hardwire I need to make a monitoring application for 8 t1's, and I need to seperate the CPU usage on the encoder process quite a bit.
1169769776 M * hardwire haven't figured out if I should just hack cpu limiting into the program, or use two vservers
1169769834 M * Bertl just put them into two different contexts maybe?
1169769844 M * Bertl (doesn't have to be a full guest)
1169769848 M * hardwire true
1169769852 M * hardwire very true
1169769859 M * hardwire I am not all that familiar with doing that however
1169769874 M * Bertl chcontext or vcontext --help
1169770011 M * hardwire har
1169770020 M * hardwire I sent 3 image sizes to an exec to use in a powerpoint
1169770024 M * hardwire they used the lowest resolution
1169770026 M * hardwire made it fullscreen
1169770030 M * hardwire and blamed me for the quality
1169770031 M * hardwire sigh
1169770060 M * Bertl happens when you don't use vector graphics :)
1169770086 M * hardwire its a photos
1169770088 M * hardwire :(
1169770101 M * hardwire guess I get to remind them to read their email
1169770108 M * Bertl those look better in vector graphics too :)
1169770110 M * hardwire its always fun pointing out they suck
1169770117 M * Bertl just harder to accomplish :)
1169770131 A * hardwire started mucking with HDR photographs
1169770139 M * hardwire I do enjoy 32 bits of depth per channel
1169770377 M * bronson Sorry, was afk.
1169770388 M * bronson Bertl: no idea how Chandler is.  I was hoping to lurk and find out.  :)
1169770413 M * Bertl bronson: ah, thought you were a regular there :)
1169770506 Q * EvilDin Quit: AnacønÐa · "If anything can go wrong -- it will"
1169770586 M * Bertl daniel_hozac: still around?
1169770589 M * daniel_hozac yeah.
1169770600 Q * mnemoc Ping timeout: 480 seconds
1169770602 M * Bertl I think what happens in Hollow's setup is the following:
1169770619 M * Bertl - context is created, helper called
1169770636 M * Bertl - helper does a few funny things, like adding tasks to the context
1169770656 M * Bertl - after that, it exits with a failure
1169770679 M * Bertl we now end up doing an unhash on a populated context
1169770679 M * bronson Hm, most of the traffic in #chandler is from 5 people trying to get a conference call going.  So...  It's in about the state that one would expect of a large, bogged down groupware project.  Not particularly action-packed.
1169770682 M * bronson Not like here anyway.
1169770700 M * bronson fwiw.  :)
1169770702 M * daniel_hozac ah, yeah, that makes sense.
1169770712 M * Bertl bronson: I take that as compilemnt for the channel
1169770721 M * Bertl daniel_hozac: now the tricky part is this
1169770745 M * Bertl either we do not hash the context until the helper succeeded
1169770757 M * Bertl (which would render Hollows setup useless)
1169770775 M * Bertl or we do not unhash the context on failure of the helper
1169770798 M * Bertl (which might give funny results too, but actually should be cleaner than the current setup)
1169770863 J * kaner_ kaner@strace.org
1169770867 M * daniel_hozac couldn't we just kill it?
1169770982 M * Bertl what?
1169770983 Q * kaner Ping timeout: 480 seconds
1169771010 M * daniel_hozac the context, similar to what REBOOT_KILL does?
1169771073 M * Bertl we can do that, but that will open a can of worms I guess
1169771084 J * mnemoc ~amery@kilo105.server4you.de
1169771094 M * daniel_hozac hmm, yeah, i guess...
1169771216 M * Bertl do we have any userspace software which would rely on the context disposal on script failure?
1169771263 M * Bertl I mean, unless the context is persistent, it will go away ...
1169771264 M * daniel_hozac would have to be vcd, as util-vserver doesn't even implement the startup/shutdown helpers.
1169771299 M * daniel_hozac (as there's not really anything to do)
1169771305 M * Bertl now, if the script wants to kill a persistent context by returning an error 
1169771318 M * Bertl it could as well set it non persistent first
1169771404 M * daniel_hozac yeah, sure.
1169771432 M * Bertl so I'd suggest, we still return the error on create
1169771451 M * Bertl and avoid the task migration
1169771480 M * Bertl we could also auto clear the persistant flag if that makes sense
1169771551 M * Bertl but we leave the cleanup for the put/release
1169771575 M * daniel_hozac yeah, makes sense.
1169772206 M * Bertl daniel_hozac: do you remember what the reason was for making the context persistent _before_ we call the helper?
1169772243 M * daniel_hozac hmm, no. was there a reason?
1169772268 M * Bertl well, otherwise we would have moved that after the helper call, no?
1169772287 M * daniel_hozac i suppose...
1169772307 M * Bertl I think it had something to do with changing the persistant state in the helper
1169772341 M * Bertl we considered removing the flag from the flagword until after the helper or something like that
1169774096 J * dna ~naucki@63-232-dsl.kielnet.net
1169774455 M * Bertl daniel_hozac: http://vserver.13thfloor.at/Experimental/delta-helper-fix01.diff
1169774464 M * Bertl (after reverting the nsproxy-fix)
1169774524 M * Bertl there is also a small patch to improve the debugging
1169774543 M * Bertl (I would appreciate a test run of both of them on your setup)
1169774570 M * daniel_hozac i.e. util-vserver? sure.
1169774585 M * Bertl yes, TIA
1169774613 Q * dna Quit: Verlassend
1169774904 M * daniel_hozac i'm currently building http://people.linux-vserver.org/~dhozac/p/k/patch-2.6.16.37-vs2.0.3-rc1.diff
1169774923 M * Bertl ah, nice!
1169775004 M * daniel_hozac i think i got all of the important deltas, and hopefully i got them right... i'm a bit uncertain about the proc hunks though.
1169775058 M * Bertl well, you got the comma :)
1169775139 M * daniel_hozac (http://people.linux-vserver.org/~dhozac/p/k/delta-2.0.2.1-2.0.3-rc1.diff is the interdiff)
1169775227 M * daniel_hozac there were a few hunks i didn't merge, like find_proc_task_by_pid.
1169775237 M * Bertl hmm, the dynamic inode numebring for proc was ot backported to 2.6.16, or was it?
1169775252 M * Bertl s/ot/not/
1169775254 M * daniel_hozac hmm?
1169775302 M * Bertl nevermind, just saw that this is not related
1169775450 M * daniel_hozac testme and testfs succeeds.
1169775522 M * Bertl yeah, looks good here after a quick scan ... but the context creation looks very different ...
1169775548 M * Bertl did we move into a different direction on 2.0.3 there without adding that to 2.2.x?
1169775567 M * daniel_hozac humm, weird proc issues in the spectator...
1169775580 M * Bertl what kind of issues?
1169775646 M * daniel_hozac vps faux shows one of the with-init guests twice, and no real init.
1169775686 M * Bertl so you see two init processes?
1169775710 M * daniel_hozac well, yes and no.
1169775713 M * daniel_hozac i see the same process twice.
1169775722 M * Bertl that's what I meant
1169775745 M * Bertl it's caused by not removing the actual init process
1169775753 M * Bertl  	/* discard wrong fakeinit */
1169775777 J * zombiee ~Matlabi@tor-irc.dnsbl.oftc.net
1169775780 M * Bertl -	if (!vx_check(vx_task_xid(task), VX_IDENT))
1169775781 M * Bertl +	if (!vx_check(vx_task_xid(task), VX_IDENT|VX_WATCH))
1169775789 M * daniel_hozac yeah, that hunk made me very suspicious...
1169775801 M * Bertl it probably makes sense in newer kernels
1169775994 M * daniel_hozac yeah, reverting that makes it good again.
1169776008 M * daniel_hozac AFAICT, anyway.
1169776855 M * daniel_hozac it seems to be working alright, do you want to have a go at it or should i post to the ML?
1169776897 M * Bertl just post it, folks will test it I hope ...
1169776911 M * Bertl the interdiff looks promising
1169776939 M * daniel_hozac ? for the create stuff?
1169776971 M * Bertl I need to take a look at the 'final' version there
1169777022 M * Bertl but I think, once we have a final version for 2.2.x, we can largely use that for 2.0.x too
1169777057 M * Bertl i.e. it will probably make sense to check for differences between 2.0.x and 2.2.x regarding context creation and similar
1169777178 M * daniel_hozac indeed.
1169777420 Q * zombiee Ping timeout: 480 seconds
1169778197 M * daniel_hozac btw, interdiff -p1 patch-2.6.18.3-vs2.1.1.{1,2}.diff seems to be the have __create claim the [nv]xi hunks.
1169779043 J * Daniel15 ~dansoftau@server.daniel15.com
1169779097 M * Daniel15 Anyone know how to get rid of an 'undeletable' file?
1169779100 M * Daniel15 http://rafb.net/p/1sv8sr58.html :\
1169779116 M * Daniel15 That file seems to not want to get deleted for some strange reason
1169779132 M * Bertl lsattr svn/fs.pyc ?
1169779160 M * Daniel15 -----ad-cB--EjI-T- svn/fs.pyc
1169779164 M * Daniel15 What does all that mean? O.o
1169779177 M * Bertl well, you got quite some funny flags on that one :)
1169779191 M * Bertl one of those (the I) will prevent the removal
1169779211 M * Daniel15 So, chattr -I or something?
1169779214 M * Bertl hmm, actually it should be an -i IIRC
1169779231 M * Bertl ah, check the dir too
1169779255 M * Bertl with ls -lad svn and lsattr -d svn
1169779391 M * Daniel15 The chattr -i didn't seem to do anything o.O
1169779391 M * Daniel15 http://rafb.net/p/lBYOTF35.html
1169779423 M * Bertl what about the ls -lad svn ?
1169779443 M * Daniel15 drwxrwxrwx 2 root root 4096 2007-01-26 13:31 svn
1169779477 M * Daniel15 I'll remove all those attributes on the file and see if it does anything :)
1169779486 M * Bertl okay, let's get rid of all those funny flags by using chattr with -X where X is a,d,c,B,E,j,I and T
1169779515 M * Bertl a looks like append only, which could be the one biting you here
1169779531 M * Bertl btw, what filesystem is that?
1169779548 M * Daniel15 It's ext3
1169779560 M * Daniel15  chattr -b and -e didn't work
1169779570 M * Daniel15 They just show a "Usage: chattr [-RV] [-+=AacDdijsSu] [-v version] files..." message
1169779582 M * Daniel15 Yay
1169779585 M * Daniel15 I got rid of it :)
1169779586 M * daniel_hozac -B -E
1169779603 M * Daniel15 One of those attributes was causing it... I fixed it now
1169779607 M * Daniel15 Thanks for your help :)
1169779613 M * Bertl you're welcome!
1169779627 M * Bertl could that be one of the rare ext3 corruptions?
1169779662 M * Daniel15 I was asking for help on #linux on FreeNode, and they were like "CHMOD the directory", "CHMOD the file", "you don't have permission to delete that file" (even though I was root), etc...
1169779667 M * Daniel15 Possibly, I have no idea
1169779672 M * daniel_hozac i guess so...
1169779693 M * Bertl Daniel15: well, you know where to get _real_ soultions :)
1169779723 M * Daniel15 Yeah :D
1169779869 M * Daniel15 I'll be back later...
1169779870 Q * Daniel15 
1169780020 M * daniel_hozac well, i'm gonna get some sleep. i'll test the delta-helper-fix01+delta-debug-feat01 kernel tomorrow.
1169780036 M * Bertl okay, np, have a good one!
1169780040 M * daniel_hozac good night!
1169780389 J * Daniel15 ~dansoftau@server.daniel15.com
1169780443 M * Daniel15 I found out what was causing my high load averages as well...
1169780464 M * Daniel15 Whenever I stop a vserver, there's a Kernel Oops, and then the load goes really high
1169780476 J * ntrs_ ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com
1169780513 M * Daniel15 I think it's the network card (again), the Kernel Oops is straight after it removes the vserver's IP address (whatever it does)
1169780603 J * ntrs__ ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com
1169780603 Q * ntrs_ Read error: Connection reset by peer
1169780736 Q * ntrs Ping timeout: 480 seconds
1169780790 M * Bertl could you upload the oops to paste.linux-vserver.org please?
1169780791 Q * DEac- cation.oftc.net plasma.oftc.net
1169780861 M * Daniel15 http://paste.linux-vserver.org/926
1169780895 M * Daniel15 I have no idea what any of that means, but I assumed it was due to the network card (I've been having problems with it)
1169780912 M * Daniel15 I search Google for "Bad EIP value" and some sites said it was a hardware problem
1169780921 M * Bertl hmm, check with dmesg, it should show a better version with cutmarks
1169780932 J * DEac- ~deac@2001:6f8:1021::20
1169780982 M * Daniel15 :o I didn't notice that before
1169780985 M * Daniel15 One sec
1169781008 M * Daniel15 http://paste.linux-vserver.org/927
1169781019 M * Daniel15 The formatting went weird on the pastebin for some strange reason
1169781043 M * Daniel15 http://paste.linux-vserver.org/928
1169781046 M * Daniel15 There, all fixed :)
1169781096 M * Bertl yep, what kernel is that? distro kernel, but what distro?
1169781101 M * Daniel15 Debian
1169781121 M * Bertl okay, I'd suggest you upgrade, this one is known to be broken
1169781149 M * Daniel15 OK, is there a newer Debian one available, or do I need to compile one?
1169781155 M * Bertl either you compile a 2.6.19.2 yourself, or you get one of the newer debian kernels
1169781185 M * Bertl the fixed version is not in debian yet for unknown reasons
1169781197 M * Bertl but there is one, and it should be listed somewhere
1169781280 M * Bertl google found this one:
1169781282 M * Bertl http://people.debian.org/~nobse/kernel/linux-2.6/i386/
1169781367 M * Daniel15 So 2.6.18-4 should be fine?
1169781392 M * Bertl I assume so, at least according to: http://www.mail-archive.com/debian-kernel@lists.debian.org/msg23980.html
1169781424 M * Daniel15 OK, thanks, I'm downloading it now :)
1169781434 M * Bertl you're welcome!
1169781539 M * Daniel15 Ok, I'm going to reboot the server now and see if it works
1169781546 M * Bertl k
1169781690 Q * sannes Ping timeout: 480 seconds
1169781792 M * Daniel15 Yay, that fixed it :)
1169781841 M * Bertl excellent!
1169782496 Q * meandtheshel1 Quit: Leaving.
1169782631 M * Daniel15 Is there an easy way to resize ext3 partitions without losing data?
1169782642 M * Bertl yep, resize2fs
1169782647 M * Daniel15 OK, thanks :)
1169782661 M * Bertl np
1169782683 M * Daniel15 I made my /home and /var about 4.5 GB each, and now the /var is about 89% full, and the /home is only 6% fill
1169782684 M * Daniel15 full*
1169782703 M * Bertl what do you keep in /var ?
1169782716 M * Bertl ah, debian, so you have the guests there too :)
1169782725 M * Daniel15 I've got /var/www for the web stuff, and Debian keeps the guests at /var/lib/vserver
1169782729 M * Daniel15 I've fixed it temporarially
1169782756 M * Daniel15 I moved /var/lib/vservers to /vservers and symlinked it (not sure if this is the best approach, but it's only temporary)
1169782794 M * Bertl well, it will work ... it's in /vservers for all other distros :)
1169782794 J * sannes ace@har.sagt.no
1169782803 M * Daniel15 Just until I either get a bigger hard drive, or resize that partition. It's fine in the root, as / is about 7 GB :)
1169782808 M * Daniel15 Oh, OK
1169782817 M * Daniel15 I haven't tried VServer on any other distro, so I had no idea :P
1169783711 Q * Piet Quit: Piet
1169787750 J * Aiken_ ~james@ppp102-225.lns1.bne1.internode.on.net
1169788040 Q * sannes Ping timeout: 480 seconds
1169788073 Q * Aiken Ping timeout: 480 seconds
1169788696 J * FireEgl ~FireEgl@adsl-4-58-61.bhm.bellsouth.net
1169789156 J * sannes ace@har.sagt.no
1169789705 Q * FireEgl Quit: ...
1169790122 J * FireEgl ~FireEgl@adsl-4-58-61.bhm.bellsouth.net
1169790269 Q * MrX Quit: X
1169790429 J * wart_ ~wart@spray.anyhost.ru
1169790443 J * MWUtjGaum ~hollow@styx.xnull.de
1169790516 Q * wart Read error: Connection reset by peer
1169790553 J * MrX ~chaos@179.13.95.219.kmr01-home.tm.net.my
1169790637 J * Vudu 35d83f8466@217.20.138.14
1169790750 Q * Hollow Ping timeout: 480 seconds
1169790764 Q * trippeh Ping timeout: 480 seconds
1169790799 Q * Vudumen Ping timeout: 480 seconds
1169790803 N * MWUtjGaum Hollow
1169791732 J * trippeh atomt@x.vx.no
1169792565 Q * Aiken_ Quit: Leaving
1169793269 Q * ntrs__ Read error: Connection reset by peer
1169793414 J * ntrs__ ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com
1169793891 J * Aiken ~james@ppp102-225.lns1.bne1.internode.on.net
1169795266 Q * hardwire Quit: Coyote finally caught me
1169795627 J * hardwire ~hardwire@rdbck-5429.wasilla.mtaonline.net
1169799144 N * Daniel15 Daniel15[Busy]
1169799810 N * Daniel15[Busy] Daniel15
1169800047 M * id23 morning
1169800118 M * id23 Bertl - do you know why I am not able to switch from a running X (inside a guest) to tty1,2,3,... ?
1169800545 J * gab ~gab@158.36.45.236
1169801554 Q * shedi Quit: Leaving
1169802405 M * matti Morning.
1169802490 J * meebey_ meebey@booster.qnetp.net
1169802505 M * meebey_ morning
1169802551 Q * Daniel15 Quit: ( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )
1169802564 M * meebey_ I still have the VFS problem when I am root in a vserver, no idea if its a bug or feature of vserver
1169803813 Q * Aiken Quit: Leaving
1169804085 J * dlezcano ~dlezcano@blueice3n1.uk.ibm.com
1169804667 M * Hollow morning
1169804879 M * daniel_hozac meebey_: hmm?
1169805031 M * Hollow Bertl: helper-fix01 does not work with nsproxy-fix01?
1169805108 M * meebey_ smb_galilei:/home/samba# ls Profiles/
1169805108 M * meebey_ ls: Profiles/: Permission denied
1169805109 M * meebey_ daniel_hozac: that one
1169805130 M * daniel_hozac Hollow: no, you need to revert nsproxy first.
1169805155 M * Hollow ok
1169805225 M * meebey_ daniel_hozac: is vserver supposed to protect the VFS in any way?
1169805231 M * meebey_ besides chroot barrier
1169805263 M * daniel_hozac meebey_: no, AFAIK the chroot barrier and the xid tagging are the only things we add to the permission checks.
1169805297 M * meebey_ how can I check xid stuff? maybe there is somehting set
1169805314 M * daniel_hozac lsxid.
1169805317 M * meebey_ daniel_hozac: I am surprised that lsattr doesn't show --barrier btw
1169805325 M * daniel_hozac try showattr.
1169805341 M * meebey_ rau:/vservers# showattr 
1169805341 M * meebey_ ---Bui- .
1169805343 M * meebey_ ah :)
1169805347 M * meebey_ didnt know that one
1169805361 M * meebey_ ---bui- ./smb
1169805368 M * meebey_ small b?
1169805376 M * meebey_ No manual entry for showattr
1169805377 M * meebey_ grrr
1169805380 M * meebey_ I hate those programs
1169805413 M * meebey_ ---bui- ./Service-CD
1169805413 M * meebey_ ---Bui- ./Users
1169805413 M * meebey_ ---Bui- ./tmp
1169805413 M * meebey_ ---Bui- ./MapmasWinDokumente
1169805413 M * meebey_ ---Bui- ./Profiles
1169805415 M * meebey_ wtf
1169805415 M * daniel_hozac lowercase means not set, but available, uppercase means set, - means unavailable.
1169805432 M * meebey_ something/one set barrier attributes!
1169805435 M * meebey_ grrrr
1169805446 M * meebey_ ok this will be definetly the cause :)
1169805460 M * daniel_hozac yep.
1169805477 M * meebey_ nasty
1169805484 M * meebey_ I copied the files using rsync from the old server
1169805506 M * meebey_ could it be that the 2.4 or vs1.2 didnt honor the barrier attribute?
1169805507 M * daniel_hozac that's unlikely to have copied those attributes...
1169805522 M * meebey_ oh no, the files were not inside a vserver before! :)
1169805529 M * daniel_hozac 1.2 had an entirely different way of expressing barriers, with the 000 directory permissions, IIRC.
1169805556 M * meebey_ but now I have the cause, good :)
1169805561 M * meebey_ daniel_hozac: thanks for the hint ;)
1169805567 M * daniel_hozac you're welcome!
1169805580 J * Aiken ~james@ppp102-225.lns1.bne1.internode.on.net
1169805598 M * meebey_ I did always lsattr and thought it shows all
1169805657 M * daniel_hozac no, showattr is the util-vserver command...
1169805672 M * meebey_ ah
1169805677 M * meebey_ barrier is a vserver feature?
1169805678 M * daniel_hozac lsattr is from e2fsprogs.
1169805681 M * daniel_hozac yes, of course.
1169805688 M * meebey_ aaah, didnt know that :)
1169805693 M * meebey_ now it makes sense
1169805699 M * meebey_ so setattr is vserver too?
1169805702 M * daniel_hozac yes.
1169805703 M * meebey_ while chattr is e2fs?
1169805707 M * daniel_hozac exactly.
1169805712 M * meebey_ thats confusing ;)
1169805733 M * meebey_ I didnt notice the tools are from different projects
1169805751 M * meebey_ but I wondered why they have so similar names
1169805781 M * meebey_ so I have no chance to search for barrier attribute using find...
1169805802 Q * dlezcano Read error: Connection reset by peer
1169805871 M * daniel_hozac no.
1169805883 M * daniel_hozac or well, shouldn't be too hard to add to the findutils patch.
1169805891 M * meebey_ rau:/vservers# setattr -R -x --~barrier smb/
1169805894 M * meebey_ that should do it
1169805900 M * daniel_hozac yep.
1169805940 A * meebey_ adds that to his memory, if file access for root acts crazy inside vserver, check barrier attributes
1169805965 M * meebey_ and now I have to solve the ICMP brokeness of ISPs
1169805980 M * meebey_ I wish there would be a tool for that, since its so common problem
1169805998 M * meebey_ that some broken routers drop ICMP need-fragementation packets
1169806011 M * meebey_ and also big providers have this problems
1169806759 M * Hollow is it supposed that vx_kill(xid=6, pid=1, sig=9) fails to kill init, but vx_kill(xid=6, pid=<realinitpid>, sig=9) works?
1169806761 J * dlezcano ~dlezcano@blueice2n1.uk.ibm.com
1169806763 M * ard That's not completely correct: there are also icmp problems with big "security" companies...
1169806809 M * meebey_ true
1169806809 M * ard (which find icmp evil, and advice to block *all* icmp traffic)
1169806843 M * meebey_ right, icmp must be evil as it sends secretly data that is not visible to the user!
1169806863 M * meebey_ and its not encrypted, lol
1169806895 M * ard http://www.ubizen.com/ is one such company which we managed to kick out of this company a few years back... :-)
1169806928 M * waldi meebey_: if someone wants to kill itself out of the network it is there decision, i prefer to deliver LARTs to them
1169806978 M * ard waldi : the problem is that it actually keeps working for most people...
1169806982 M * daniel_hozac Hollow: 2.2.0-rc8?
1169806995 M * Hollow yep
1169807040 M * daniel_hozac hmm, that should work, AFAICT.
1169807106 M * daniel_hozac cat /proc/virtual/<xid>/info shows init's pid?
1169807143 M * Hollow yep
1169807166 M * daniel_hozac does the context have VXF_STATE_ADMIN?
1169807187 M * daniel_hozac and are there at least one other process in the guest?
1169807194 M * daniel_hozac +english
1169807316 M * daniel_hozac if the guest lacks VXF_STATE_ADMIN and there are processes left, the signal gets changed to SIGINT.
1169807339 M * Hollow uh.. strange things goigng on
1169807389 M * Hollow http://paste.linux-vserver.org/929
1169807444 M * daniel_hozac xid pid sig?
1169807454 Q * Aiken Quit: Leaving
1169807468 M * Hollow yep
1169807526 M * Hollow ls -l /proc/virtual/7/ works, but ls /proc/virtual does not show 7 :)
1169807558 M * daniel_hozac yeah, that's weird.
1169807558 J * dna ~naucki@118-196-dsl.kielnet.net
1169807603 M * Hollow all commands say "no such process" but i can also not create a new context with same xid ..
1169807619 M * Hollow probably now the context is never unhashed?
1169807651 M * daniel_hozac it should still be unhashed during normal circumstances.
1169807659 M * daniel_hozac did the startup helper die?
1169807672 M * daniel_hozac i.e. return non-zero.
1169807703 M * Hollow http://paste.linux-vserver.org/930
1169807760 M * daniel_hozac so we shouldn't hit that path...
1169808051 M * Hollow it reproducably happens if i try to kill init (which is the last process), which fails with vx_kill(xid, 1, 9) but succeeds with vx_kill(xid, <initpid>, 9)
1169808550 Q * dlezcano Read error: Connection reset by peer
1169808739 Q * gab Quit: Leaving
1169808789 M * daniel_hozac can you reproduce it with a simple vcmd sequence?
1169808881 M * daniel_hozac e.g. vcmd -i 100 -BC ctx_create .flagword=^34^33^32^4 -- vcmd -i -1 -BC set_cflags .mask=^33 -- sleep 60 & vcmd -i 100 -C ctx_kill .pid=<pid> .sig=9
1169809138 M * Hollow ok, sec
1169809354 M * Hollow daniel_hozac: yep, it's reproducable with this command
1169809379 M * daniel_hozac oh really? that exact command? what do you set pid to?
1169809385 M * Hollow $!
1169809409 M * daniel_hozac so the realpid
1169809466 M * daniel_hozac and this leads to the proc issue, right?
1169809471 M * Hollow yep
1169809492 M * Hollow i have the helper, proc and debug fix applied
1169809495 M * daniel_hozac hmm, i don't see it.
1169809507 M * daniel_hozac i wonder if i have proc...
1169809528 J * dlezcano ~dlezcano@blueice2n1.uk.ibm.com
1169809558 M * daniel_hozac delta-proc-fix04, right?
1169809565 M * Hollow yep
1169809579 M * daniel_hozac yeah, i didn't have that.
1169809602 M * daniel_hozac i probably should start fresh, this tree is some sort of mix of 2.2.0-rc6 through -rc8.
1169809654 M * Hollow yeah, i'm compiling a fresh one too atm ;)
1169809699 M * daniel_hozac oh hey, do you have http://vserver.13thfloor.at/Experimental/delta-vkill-fix02.diff?
1169809757 M * Hollow no, didn't had it
1169809776 M * Hollow ah wait
1169809783 M * Hollow think we have this in out rc8 patchset
1169809784 M * daniel_hozac that's probably it.
1169809788 M * Hollow *our
1169809792 M * daniel_hozac oh.
1169809793 J * Aiken ~james@ppp121-180.lns1.bne4.internode.on.net
1169809804 M * Hollow will double check ..
1169809814 M * daniel_hozac well, vkill-fix02 fixes vc_ctx_kill, so seems like a likely candidate...
1169809873 M * Hollow yeah, we already have vkill-fix0{2,3}, getsched and noforkns
1169809879 M * daniel_hozac okay.
1169809887 M * Hollow and i applied the other three on top of that
1169809903 M * Hollow but as i said, doing a fresh one currently
1169809924 Q * cdrx Ping timeout: 480 seconds
1169810103 M * daniel_hozac yeah, same here.
1169810173 Q * michal` Ping timeout: 480 seconds
1169810332 M * Hollow daniel_hozac: it's gone o.o
1169810340 M * daniel_hozac lol
1169810352 M * daniel_hozac can you kill init now?
1169810361 Q * FireEgl Read error: Connection reset by peer
1169810364 M * Hollow will try.. but the vmc line works ..
1169810380 M * Hollow *vcmd
1169810389 M * daniel_hozac that's strange...
1169810564 M * Hollow but i still cannot kill init with (xid, 1, 9)
1169810596 M * Hollow the proc issue seems to be gone though
1169810612 M * daniel_hozac well, that's good, i guess.
1169810739 J * michal` ~michal@www.rsbac.org
1169810859 M * daniel_hozac hmm. it's weird, bash says the process was killed, but only after sleep is done sleeping.
1169810883 M * daniel_hozac (this is with the previous messed up kernel)
1169810925 M * daniel_hozac so it's like the kill is scheduled somehow.
1169810985 M * Bertl morning folks!
1169810991 M * Hollow morning Bertl :)
1169810994 M * daniel_hozac morning Bertl
1169811027 M * daniel_hozac SEND_SIG_FORCED vs. SEND_SIG_PRIV maybe?
1169811063 M * Bertl how did the changes work out we established yesterday?
1169811177 M * Bertl so we can also add the 'new' nfs patch, as I see?
1169811203 M * daniel_hozac yeah, it seems to do the trick.
1169811216 M * Hollow Bertl: i didn't come across the unhash issue with the helper-fix yet ..
1169811237 M * Bertl excellent, and it also 'works' for your setup?
1169811258 M * Bertl i.e. it doesn't break any of your helper scripts or so?
1169811272 M * Hollow well, not sure yet ;)
1169811272 M * daniel_hozac hmm. i seem to have the proc issue now.
1169811282 M * Hollow there are some other issues around still
1169811283 M * daniel_hozac (still not a pristine kernel, will see if that one also shows it)
1169811320 M * Bertl note: there is a new error message on context migrate (EFAULT) which means that the context is in shutdown state but still hashed
1169811325 M * Hollow ah, how did you manage to reproduce it?
1169811337 M * daniel_hozac i just stopped one of the guests i had started.
1169811360 M * Bertl proc entry gone, tasks still running/sleeping?
1169811367 M * daniel_hozac other way around.
1169811377 M * Bertl proc entry there, tasks gone?
1169811386 M * Hollow proc entry gone, but still there
1169811387 M * daniel_hozac context gone, ls /proc/virtual doesn't show anything, but ls /proc/virtual/<xid> works
1169811389 M * Hollow ;)
1169811401 M * daniel_hozac cating the files results in nothing though.
1169811405 M * Bertl daniel_hozac: okay, please dump the stat/status
1169811409 M * Hollow heh
1169811453 M * Hollow daniel_hozac: can you still create a acontext with the same xid?
1169811487 M * daniel_hozac yeah.
1169811516 M * daniel_hozac what error message did you get?
1169811521 M * Hollow ok, iirc that worked here too, but things like bcaps_get said "no such process"
1169811538 M * daniel_hozac ah...
1169811558 M * daniel_hozac i'm rebooting to a pristine 2.2.0-rc8+patches now.
1169811593 J * shedi ~siggi@dsl-149-109-85.hive.is
1169811639 M * daniel_hozac yep, still does it.
1169811654 M * Bertl okay, some vcmd test case?
1169811656 M * daniel_hozac ./vcmd -i 100 -BC ctx_create .flagword=^34^33^32^4 -- ./vcmd -i -1 -BC set_cflags .mask=^33 -- sleep 10 & sleep 1; ./vcmd -i 100 -BC ctx_kill .pid=1 .sig=9
1169811662 M * Bertl excellent!
1169811670 M * daniel_hozac ls -l /proc/virtual/100 after 10 seconds still shows the directory.
1169811700 M * daniel_hozac (that's the other oddity, it doesn't die immediately, but when it dies it says it was killed)
1169811726 M * Bertl hmm, with immediately you mean?
1169811746 M * daniel_hozac i.e. right after the command.
1169811768 M * daniel_hozac the process is allowed to live for however long it wants...
1169811771 M * Hollow htat's probably why it never works for init ... because init does not exit ..
1169811800 M * daniel_hozac right.
1169811801 M * Bertl okay, first, let's break down the sequence into:
1169811818 M * Bertl ./vcmd -i 100 -BC ctx_create .flagword=^34^33^32^4 -- ./vcmd -i -1 -BC set_cflags .mask=^33 -- sleep 10
1169811831 M * Bertl ./vcmd -i 100 -BC ctx_kill .pid=1 .sig=9
1169811847 M * Bertl ls -l /proc/virtual/100
1169811875 M * Bertl that is (given that you pause a little between the commands) what you do, right?
1169811889 M * daniel_hozac right.
1169811910 M * Bertl and when does it report being killed?
1169811923 M * daniel_hozac after the ten seconds have passed.
1169811931 M * Bertl okay, checking now
1169811971 M * daniel_hozac seems it's not 100% reproducible, i didn't get the /proc directory to stick around this time.
1169811978 M * daniel_hozac the killed thing is though.
1169812304 M * daniel_hozac if you change the .pid=1 to .pid=$!, i.e. the real pid, the process dies right away.
1169812325 M * Bertl okay, that's a good hint
1169812399 M * Bertl that's forced vs priv
1169812409 M * daniel_hozac right.
1169812425 M * Bertl what really confuses me now is that we had a case (init?) which required the force, no?
1169812446 M * Bertl now we have a setup, where it requires the priv?
1169812472 M * Bertl I'm more than happpy to revert the force change
1169812483 M * Bertl as I think the priv should be enough/better :)
1169812532 M * daniel_hozac what was the forced for?
1169812562 M * Bertl we did that to fix a vkill not killing init issue, IIRC?
1169812571 M * daniel_hozac http://vserver.13thfloor.at/Experimental/delta-vkill-fix02.diff
1169812575 M * Bertl yep
1169812593 M * daniel_hozac that was just to fix vkill.
1169812617 M * Bertl so?
1169812638 M * Bertl it's early in the morning, you have to elaborate :)
1169812642 M * daniel_hozac i don't remember the init specific part...
1169812646 M * daniel_hozac yeah, it's very early.
1169812677 M * Bertl okay, personally I'd tend to remove the force part completely
1169812722 M * Bertl i.e. use priv for all cases
1169812729 M * daniel_hozac rather strange that FORCED doesn't actually tell the process to do anything...
1169812742 M * daniel_hozac yeah, seems like the best option.
1169812746 M * Bertl let's see what the code actually does, brb
1169813032 J * meandtheshell ~markus@85-124-206-169.dynamic.xdsl-line.inode.at
1169813144 M * Bertl the forced is supposed to send the signal directly, i.e. without queueing it
1169813164 M * Bertl maybe that does require some additional work though
1169813193 M * daniel_hozac seems like it...
1169813198 Q * ensc Ping timeout: 480 seconds
1169813199 M * Bertl so I'd suggest we switch the PRIV only, if that works for us? 
1169813217 J * Hurga ~user@h-213.61.155.114.host.de.colt.net
1169813227 M * Bertl wb meandtheshell! Hurga!
1169813230 M * Hurga Hi.
1169813369 M * meandtheshell Bertl: hello everybody :)
1169813539 M * daniel_hozac switching to priv makes it work as expected.
1169813786 M * Bertl okay, I'll add that and the nfs part now, then I'll check the proc issue
1169813943 M * Bertl why does serge's patch remind me of the early_exit we did :)
1169813971 M * daniel_hozac hehe yeah, i got flashbacks from that too.
1169813981 M * daniel_hozac especially since i just backported it yesterday ;)
1169814021 M * Bertl yeah, I can imagine ...
1169814204 M * brcc bertl! ltns
1169814340 M * Bertl hey brcc!
1169814468 M * brcc next month i iwll create iptables-save/iptables-restore, to finish the iptables stuff
1169814471 M * brcc just that is missing
1169814479 M * brcc btw, no one is trying it
1169814479 M * brcc hehehe
1169814545 M * Bertl btw, we now have peertag info on unix sockets
1169814568 M * Bertl (so you could figure the context from a socket via that)
1169814646 M * daniel_hozac oh yeah.
1169814742 M * brcc i dont remember how i get the context.. i think i am getting it from the source ip
1169814753 M * brcc easier to use unix sockets ?
1169814776 M * Bertl probably the next best to using a kernel interface :)
1169814778 M * brcc i am into some projets, as soon as i finish them i am going back to iptables. and after iptables i will make those quota tests scripts (if you sitll want them) 
1169814895 M * id23 Bertl, maybe you can give me a tip why i am unable to switch to a console via alt-strg-f1 while i am in a Xsession inside a vserver ?
1169814922 M * Bertl hmm, maybe it is disabled in your X?
1169814938 M * id23 it it std etch
1169814942 M * id23 it is
1169814967 M * id23 maybe the kdm grabs all the ttys
1169814972 M * Bertl does X recognize/log the special sequence?
1169814999 M * Bertl it probably executes something like 'switchto 1'
1169815001 M * id23 if i use alt-strg-f1 the keyboard does not function any longer
1169815006 M * id23 the mouse works
1169815011 M * id23 if i restart X
1169815013 M * Bertl try if that would work
1169815019 M * id23 the keyboard is working again
1169815022 J * paperboy ~stylistik@ks32003.kimsufi.com
1169815026 M * id23 ok thanx
1169815044 M * Bertl id23: what happens if you do not restart X, but press CTRL-ALT-F7 instead?
1169815055 M * Bertl (or whatever console the X is running on)
1169815062 M * Bertl welcome paperboy!
1169815078 M * id23 does not work
1169815080 M * paperboy hi everybody :-)
1169815100 M * paperboy may I ask here questions about installation problems ?
1169815116 M * Bertl as long as they are Linux-VServer related, yes of course :)
1169815120 M * id23 this is stange - the last time i used a vserver X - it was just smooth - yesterday i had to fiddle with udev 
1169815137 M * Bertl hmm? udev? in a guest?
1169815144 M * id23 nope
1169815161 M * id23 i had to copy some devices from /dev/.ststic/dev to /dev
1169815167 M * id23 .static
1169815175 M * id23 i guess this is debian related
1169815199 M * paperboy I have problems with creating a guest under debian etch
1169815216 M * paperboy I'm using http://linux-vserver.org/Installation_on_Debian for instructions
1169815224 M * paperboy kernel is running well, but after I do:
1169815226 M * Bertl paperboy: okay, what issues?
1169815232 M * paperboy newvserver --vsroot /home/vs/ --hostname vserver --domain example.com --ip 213.251.135.20 --dist etch --mirror http://ftp.fr.debian.org/debian/ --interface eth0
1169815241 M * paperboy deboostrap starts with
1169815242 M * paperboy I: Retrieving Release
1169815245 M * paperboy and then:
1169815249 M * paperboy I: Retrieving base-files
1169815249 M * paperboy W: Couldn't download package base-files
1169815257 M * paperboy and that for every further package after base-files
1169815268 M * paperboy no matter which debian mirror I use
1169815275 M * Bertl daniel_hozac: is newvserver obsoleted by now?
1169815299 M * daniel_hozac AFAIK.
1169815322 M * Bertl paperboy: okay, please try with the vserver <name> build method instead
1169815336 M * Bertl paperboy: and if that works as expected, please update the wiki
1169815343 M * paperboy hehe
1169815350 M * paperboy just another question:
1169815358 M * paperboy my server has only 1 public internet ip
1169815364 M * Bertl okay
1169815392 M * Bertl server means host here, I assume?
1169815396 M * paperboy somebody told me that I can add as "--ip " that ip too? so the guest doesn't need a lan ip and an own local net ?
1169815399 M * paperboy yes
1169815419 M * Bertl well, you can share that ip with the guest, but it has some disadvantages
1169815449 M * Bertl when you share an ip, both host and guest can bind ports on that ip
1169815476 M * Bertl this might easily lead to DoS if the guest e.g. binds your sshd port when the host restarts sshd
1169815489 M * id23 paperboy try this
1169815497 M * id23 vserver test23 build -m debootstrap --hostname test23.auuu.network --netdev eth0 --interface 213.251.135.20/24 --context 23 -- -d etch -m http://ftp.de.debian.org/debian 
1169815507 M * paperboy won't be an disadvantage for me - I would but the guest's sshd on another port
1169815511 M * Bertl usually it is better to assign a private ip and use D/SNAT to handle it
1169815516 M * paperboy better than forwarding lot of ports
1169815568 M * Bertl if you host is mostly unused (i.e. sshd only or so)
1169815579 M * Bertl you might just do it the other way round
1169815593 M * Bertl i.e. assign a private ip to the host and bind sshd there
1169815603 M * Bertl then DNAT a specific port to that
1169815612 M * paperboy I: Retrieving aptitude
1169815612 M * paperboy I: Validating aptitude
1169815612 M * paperboy I: Retrieving base-files
1169815612 M * paperboy W: Couldn't download package base-files
1169815617 M * paperboy same problem with vserver build
1169815618 M * id23 ?
1169815621 M * paperboy I just don't get it
1169815621 M * id23 hmm
1169815630 M * id23 which arch ?
1169815639 M * paperboy etch
1169815647 M * id23 i386 amd64...
1169815653 M * paperboy ah, sorry :)
1169815654 M * paperboy i386
1169815660 M * Bertl paperboy: can you reach the internet, especially ftp.de.debian.org quite fine?
1169815663 M * paperboy 2.6.18-3-vserver-686
1169815672 M * paperboy yes - I already tried several mirrors
1169815683 M * Bertl paperboy: ah, you might also want to upgrade to 2.6.18-4
1169815689 M * Hurga Bertl: It's fun to have two guests with the same IP, though. I once had two guests, one old and one new, both had the same partition mounted, and one was running postfix and the other one the rest of the services. Makes an easy migration.
1169815692 M * paperboy do you know the debootstrap cmd which vserver build calls in effect ?
1169815696 M * Bertl the 2.6.18-3 kernel is broken
1169815724 M * Bertl paperboy: just add --debug to the vserver command (right after vserver)
1169815736 M * paperboy I like the debian packages thing but I guess there's no new kernel for it out ?
1169815748 M * Bertl yes there is, it's just not included yet
1169815767 M * daniel_hozac http://kernel-archive.buildserver.net/
1169815797 M * Bertl should be in debian etch in 2-3 years^H^H^H^H^Hmonth
1169815828 M * daniel_hozac lol
1169815833 M * paperboy hehe, nah, etch is quite ok :>
1169815853 M * id23 :)
1169815859 M * paperboy .++ /usr/sbin/debootstrap etch /etc/vservers/.defaults/vdirbase/test23 http://ftp.de.debian.org/debian ''
1169815866 M * Bertl daniel_hozac: for how long is that bug fixed in Linux-VServer by now?
1169815868 M * paperboy ok, I'll try running this cmd on other shells
1169815875 M * daniel_hozac Bertl: since 2.0.2.2-rc9
1169815898 M * daniel_hozac so... a month and a half?
1169815900 M * id23 paperboy, just include sid in you sources.list
1169815902 M * id23 do a update
1169815907 M * id23 install the -4 kernel
1169815915 M * daniel_hozac it's not in sid.
1169815918 M * Bertl right, Dec 10th 2006
1169815930 M * id23 linux-image-2.6.18-4-vserver-amd64 - Linux 2.6.18 image on AMD64
1169815930 M * id23  ?
1169815932 M * waldi it will be in about 10 hours
1169815934 M * paperboy nah, I don't want to make a whole dist-upgrade
1169815935 M * id23 ?
1169815941 M * id23 just the kernel
1169815944 M * paperboy before I do that I compile the kernel on my own
1169815946 M * Bertl waldi: ah, great news!
1169815947 J * darkbyte ~usuari@87.219.167.237
1169815955 M * Bertl welcome darkbyte!
1169815956 M * id23 paperboy, do that
1169815964 M * darkbyte Hi all !
1169815983 M * darkbyte Anyone are using the vserver kernel that comes with etch ???
1169815988 M * paperboy yes id23, but you see that I already stuck at a userspace issue imo
1169816016 M * Bertl waldi: btw, is 'debian' interested in keeping an updated 2.6.18 branch (like we do for the 2.6.16 now)?
1169816042 M * paperboy darkbyte, I'm runnning 2.6.18-3-vserver-686 kernel here ;) but failing at creating guests by now
1169816078 M * ard heh
1169816082 M * darkbyte paperboy: Mine works fine, except a problem with apache2 installed guests...
1169816088 M * ard I only got oopses with the debian kernel
1169816092 M * harry what's the best way to enable cpu scheduling on a guest os?
1169816097 M * harry where are the docs on that
1169816101 M * Bertl ard: that's the -3 kernel :)
1169816102 M * harry i can't find any on the website
1169816103 M * ard I use 2.2.0-rc[678], and that works
1169816129 M * id23 i just have a shutdown issue with -3 - i just installed -4 and do a retry - else i will do a vanilla compile too
1169816131 M * ard (debian vserver kernel oopses when stopping a guest)
1169816135 M * daniel_hozac harry: ?
1169816141 M * Bertl ard: as I said, the -3 ekrnel
1169816149 M * darkbyte ard: At the moment I only get OOps with a apache2 installed guest (at shutdown)
1169816149 M * ard ah :-)
1169816152 M * harry daniel_hozac: i want the scheduler
1169816159 M * harry it used to be a file: schedule
1169816162 M * harry with numbers in it
1169816163 M * daniel_hozac harry: i.e. the token bucket?
1169816164 M * ard darkbyte : that also describes it :-)
1169816165 M * waldi Bertl: we are always interested in things which lowers our load. and the kernel team have currently too less active developers
1169816168 M * harry the last line was: dummy
1169816176 M * daniel_hozac harry: no it wasn't.
1169816177 M * harry but that had to change to 0
1169816179 M * harry or so
1169816188 M * ard anyway: I always use vanilla kernels, but I needed something for the initial boot :-)
1169816189 M * darkbyte ard: Other guest whithout apache2 installed works fine
1169816191 M * daniel_hozac yes, since a priority bias of "dummy" doesn't make any sense.
1169816193 M * harry but now, there was something about a directory sched or so
1169816203 M * id23 the problems are gone :)
1169816209 M * Bertl waldi: as it seems, there is some interest in up-to-date 2.6.16 kernels (Linux-VServer) to combine with stable Xen
1169816210 M * id23 Bertl, thanx for the -4 tip
1169816211 M * daniel_hozac yes...
1169816215 M * harry anyway, now i'm looking for documentation on it
1169816224 M * daniel_hozac the flower page explains it.
1169816230 M * daniel_hozac as usual.
1169816233 M * harry so that i can "fix" my vserver init script to correctly do all
1169816235 M * Bertl waldi: I think simmilar might be true for the 2.6.18 too
1169816243 M * harry there is no link on the wiki to the flowerpage, right?
1169816255 M * daniel_hozac there is.
1169816268 M * daniel_hozac http://linux-vserver.org/Frequently_Asked_Questions#What_is_the_.22great_flower_page.22.3F
1169816358 M * waldi Bertl: yep
1169816373 M * harry http://oldwiki.linux-vserver.org/Scheduler+Parameters <== that one clearly said: dummy
1169816378 M * harry in the example... but ok...
1169816415 M * daniel_hozac yes well, documentation is wrong all the time.
1169816530 M * harry so true ;)
1169816536 J * lilalinux ~plasma@dslb-084-058-200-112.pools.arcor-ip.net
1169816686 J * ensc ~irc-ensc@p54B4E326.dip.t-dialin.net
1169816717 P * thunder1 
1169816728 J * thunder1 ~thu@anonymisierungsdienst2.foebud.org
1169816756 M * matti Hi harry, how are you?
1169816803 M * Bertl daniel_hozac: does that look okay to you? http://vserver.13thfloor.at/Experimental/delta-nsproxy-fix02.diff
1169816908 M * Bertl paperboy: does the debootstrap work for you in a separate shell?
1169816936 M * matti Morning Bertl :)
1169816941 M * Bertl hey matti!
1169816991 M * daniel_hozac Bertl: yeah, looks good.
1169817043 M * Bertl proc seems to work fine here ...
1169817125 Q * darkbyte Read error: Connection reset by peer
1169817139 M * harry matti: goooooooood ;)
1169817585 M * Bertl daniel_hozac: maybe the proc issues were related to the nsproxy stuff?
1169817605 M * Bertl (or did you get them with the nfs fix applied?)
1169817643 M * matti harry: :)
1169817667 M * harry vserver avnetconvert build --force -m debootstrap --hostname convert.avnet.kuleuven.be --netdev eth0.7 --interface 192.168.0.2/24 --context 902 -- -d edgy -m http://be.archive.ubuntu.com/ubuntu
1169817671 M * harry this... doesn't work...
1169817684 M * harry http://ftp.debian.org/debian/pool/main/d/debootstrap/debootstrap_0.3.3_all.deb:
1169817685 M * harry 14:21:19 ERROR 404: Not Found.
1169817685 M * Bertl edgy in ubuntu?
1169817688 M * harry and... he's right
1169817695 M * harry there is no debootstrap_0.3.3_all.deb
1169817730 M * Bertl what debootstrap do they use ...
1169817731 M * id23 ok - Bertl i went one step backwards - i added tty8 and tty9 to my default vserver - the vserver is started automatically (kdm is now disabled) but i still unable to switch to tty8 or tty9
1169817740 M * id23 in inittab
1169817770 Q * Aiken Quit: Leaving
1169817771 M * harry Bertl: edgy is the latest ubuntu
1169817784 M * harry and: i don't know.... but it looks for that on ftp.debian.org
1169817790 M * id23 both devices are available in the host and vserver
1169817790 M * harry how/why/who...
1169817807 M * Bertl harry: what tool version?
1169817831 M * harry              util-vserver: 0.30.212; Dec 13 2006, 15:59:36
1169817872 M * daniel_hozac Bertl: could be, i suppose. but i don't really see how they relate to one another.
1169817886 M * Bertl well, seems they updated to 0.3.3.1 and removed the old one :(
1169817904 M * paperboy debootstrap works on another etch shell, fuck :(
1169817925 M * harry Bertl: its looking for debootstrap at DEBIAN, not UBUNTU
1169817938 M * Bertl daniel_hozac: my best guess would be that proc is backed in the namespace
1169817940 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at
1169817952 M * Bertl harry: and I checked the debian site for debootstrap
1169817959 Q * Roey Remote host closed the connection
1169817980 M * Bertl daniel_hozac: I think the debootstrap needs to be updated
1169817988 M * daniel_hozac it's updated in trunk already.
1169817993 M * Bertl ah, okay
1169818003 M * daniel_hozac (i.e. 0.30.213-preX all have it)
1169818020 M * Bertl harry: so please either update that yourself, or get a newer tool version
1169818024 M * harry hmpf... can i stick with the 212, and still make it work?
1169818029 M * harry maky
1169818029 M * daniel_hozac but as i recall, that error message is one of the most verbose messages.
1169818031 M * harry mkay :)
1169818040 M * Bertl I have no idea _why_ they did remove the older one though
1169818043 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net
1169818043 M * daniel_hozac i.e. with explicit instructions on how to update it.
1169818044 M * harry daniel_hozac: true, but it's not a normal one ;0
1169818045 M * harry ;)
1169818057 M * daniel_hozac it is.
1169818068 M * daniel_hozac happens whenever debootstrap is updated.
1169818267 M * harry ahm...
1169818280 M * harry can i install UBUNTU with the debian debootstrap?
1169818310 M * harry E: No such script: /var/tmp/debootstrap.YbE2V2/usr/lib/debootstrap/scripts/edgy
1169818436 M * harry W: Failure trying to run: chroot /usr/local/etc/vservers/.defaults/vdirbase/avnetconvert dpkg --force-depends --install var/cache/apt/archives/libc6_2.4-1ubuntu12_i386.deb
1169818558 M * Bertl I think it should work, but daniel_hozac has the details ...
1169818572 M * harry daniel_hozac: i used the ubuntu debootstrap now
1169818579 M * harry but the last line... stups :)
1169818582 M * harry stops :)
1169818625 M * Bertl okay, off for now ... back later, I uploaded the rc8.7 for 2.2.0 (current state here)
1169818633 N * Bertl Bertl_oO
1169818638 M * harry cya, Bertl_oO 
1169818788 M * id23 cya
1169818798 Q * michal` Ping timeout: 480 seconds
1169818802 M * harry daniel_hozac: can you help me here?
1169818996 M * harry PAX: From 192.168.28.2: execution attempt in: <anonymous mapping>, 5922d000-59243000 5922d000
1169819000 M * harry PAX: terminating task: /usr/sbin/iconvconfig(iconvconfig):22127, uid/euid: 0/0, PC: 59241c54, SP: 59241b34
1169819003 M * harry PAX: bytes at PC: b9 4c 1c 24 59 e9 f2 7d e0 ae 08 08 c8 ad 08 08 c8 ad 08 08
1169819009 M * harry PAX: bytes at SP-4: 000007c3 499bfce8 08083a70 00000000 00000000 59241c54 000007e9 0000002d 59241c80 08049e85 08083a70 59241c54 00000000 00000001 0806f598 00000003 7273752f 62696c2f 6f63672f 672f766e 766e6f63
1169819131 M * sid3windr pax schmax
1169819314 Q * dlezcano Quit: Leaving
1169819396 J * michal` ~michal@www.rsbac.org
1169819448 M * harry yeah... it gets me again
1169819452 M * harry but... this is a debian problem
1169819469 M * harry why the hell would you want to execute code there...
1169819471 M * harry idiots!
1169819564 J * cdrx ~legoater@blueice1n1.uk.ibm.com
1169820571 Q * shedi Quit: Leaving
1169822154 Q * Hunger Read error: Connection reset by peer
1169822219 J * Hunger Hunger.hu@Hunger.hu
1169822688 M * harry /etc/init.d/rc: 166: cannot open /dev/console: No such file
1169822691 M * harry is this fixable?
1169822700 M * daniel_hozac touch /dev/console?
1169822707 M * daniel_hozac or ln -s tty /dev/console
1169822809 M * harry tnx
1169823218 N * wart_ wart
1169824006 J * dna_ ~naucki@118-196-dsl.kielnet.net
1169824019 M * daniel_hozac PowerKe: ping
1169824103 J * shedi ~siggi@dsl-149-109-85.hive.is
1169824259 Q * cdrx Read error: Connection reset by peer
1169824389 Q * dna Ping timeout: 480 seconds
1169824869 Q * Hollow Remote host closed the connection
1169824923 J * Hollow ~hollow@styx.xnull.de
1169824950 Q * Loki|muh Remote host closed the connection
1169824959 J * Loki|muh loki@satanix.de
1169825292 M * id23 i almost tried everything - but switching to another virtual console via strg-alt-f1 ist not working
1169825326 M * id23 when running X inside a vserver
1169825376 M * daniel_hozac does the guest have /dev/ttyX?
1169825382 M * id23 yes
1169825388 M * id23 i want to use vt8
1169825395 M * id23 tty8 is available
1169825408 M * id23 no message in the logs
1169825417 M * id23 the keyboard just stops working - mouse is fine
1169825430 M * id23 if i restart x - the keyboard is working again
1169825436 M * id23 until i want to switch again
1169825532 M * id23 do i need to have other tty (those i want to switch to) as well in my vserver ?
1169825613 M * daniel_hozac i have no idea... i assume it all works fine in the host?
1169825653 M * daniel_hozac PowerKe: if you're interested, i've added an option to vhashify to ignore mtimes.
1169825660 M * id23 yes - host is fine
1169825847 M * id23 chvt is not working as well
1169825858 M * id23 in the guest
1169825874 M * daniel_hozac probably why... strace it and maybe we'll find out more.
1169825880 M * id23 inside X -. without X is it fine
1169825886 M * id23 k
1169827987 J * dlezcano ~dlezcano@blueice2n1.uk.ibm.com
1169828595 M * Radiance hmm guys, how do i specify the gateway for a vserver ? 
1169828623 M * Radiance i can set the ip, dev and prefix but where do i set the gateway ?
1169828642 P * thunder1 
1169828719 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1169828838 P * thunder1 
1169828852 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1169828918 P * thunder1 
1169829042 M * Hurga Radiance: on the host... there is only one kernel which does networking.
1169829158 M * Radiance ok so it means i'll have to add new routing for the subnet range
1169829513 Q * meebey_ Quit: leaving
1169829915 Q * dna_ Ping timeout: 480 seconds
1169830295 M * id23 back
1169830302 M * id23 daniel_hozac, i get this
1169830445 M * id23 http://pastebin.ca/329278
1169830469 M * id23 generally a permission denied
1169830474 Q * bronson Ping timeout: 480 seconds
1169830606 M * daniel_hozac id23: you _are_ doing chvt as root, right?
1169830660 M * id23 oh no - but if i run as root i get NO output
1169830668 M * id23 & the keyboard ist dead
1169830671 M * id23 again
1169830693 M * daniel_hozac does the command succeed?
1169830713 M * id23 just no
1169830725 M * daniel_hozac does the same commannd work if you run it in the host?
1169830728 M * id23 i see no error
1169830744 M * id23 i removed X from the host
1169830754 M * id23 i have to install X again there
1169830756 M * daniel_hozac so, chroot to the guest.
1169830772 M * id23 ok
1169831062 M * id23 when starting X i get konsole (kde) cannot open PTY
1169831705 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1169831904 M * Hollow Bertl_oO, daniel_hozac: FYI, kill/stop work fine now
1169831907 M * Hollow with 8.7
1169831944 Q * id23 Ping timeout: 480 seconds
1169831946 J * stefani ~stefani@tsipoor.banerian.org
1169832438 Q * dlezcano Read error: Connection reset by peer
1169832482 J * id23 ~id@p50813BEE.dip0.t-ipconnect.de
1169832518 M * id23 http://pastebin.ca/329307
1169832716 M * id23 daniel_hozac, working there - i dont know which messages reached the channel - i just got a new IP
1169833449 J * FireEgl Proteus@adsl-4-58-61.bhm.bellsouth.net
1169833914 Q * shedi Quit: Leaving
1169833950 J * dna_ ~naucki@103-246-dsl.kielnet.net
1169834193 N * Bertl_oO Bertl
1169834213 M * Bertl Hollow: so anything left for a 2.2.0 release?
1169834240 M * Hollow probably ..
1169834257 M * Hollow but have to check if it it vcd's fault first
1169834259 M * Bertl okay, will do an rc9 shortly (tomorrow or so)
1169834306 M * Bertl and if we get some positive feedback without any new issues, we have a release
1169834736 M * Hollow Bertl: it seems the vs_reboot_helper is invoked during startup..., i added a vxdprintk (ignored cmd) to the default case.. the result is: http://paste.linux-vserver.org/931
1169834768 M * Hollow also the assertion failed seems suspicious
1169834808 M * Hollow and every helper line appears twice.. no idea if that is intended ..
1169834868 M * Bertl well, I have the feeling your syslog is confused again
1169834895 M * Bertl or, you are actually doing some system calls twice too
1169834917 M * Bertl Jan 26 18:59:19 MiddleEarth vxD: nx_migrate_task(ffff81013f7f0100,ffff81013da424c0[#6.1.0])  
1169834923 M * Bertl Jan 26 18:59:19 MiddleEarth vxD: nx_migrate_task(ffff81013f7f0100,ffff81013da424c0[#6.1.0])  
1169834927 M * Bertl for example
1169834941 M * Bertl that would mean you migrated one and the same task twice
1169834957 M * Bertl I tend to believe that your syslog has weird entries
1169834965 M * Hollow hm, yeah.. seems a syslog issue... dmesg show it only once
1169834966 M * Bertl (or doesn't handle kernel logs properly)
1169834985 M * Bertl if you enable the prink timestamping
1169835000 M * Bertl you can at least get rid of the duplicates with sort -u
1169835024 M * Hollow yeah, i usually have that enabled on my boxes ..
1169835038 M * Bertl the assertion is interesting
1169835042 M * Hollow nevertheless the reboot with cmd=0 seems strange to me
1169835052 M * Bertl it basically means that you use a lock across contexts
1169835125 M * Hollow yeah, vcd uses a lock on startup to prevent another startup at the same time .. maybe the lock ist still held on migrate
1169835193 M * Bertl #define LINUX_REBOOT_CMD_CAD_OFF        0x00000000
1169835200 M * Bertl this is the cmd=0 function
1169835210 M * Bertl which is 'usually' called on init start :)
1169835215 M * Hollow aah... init does that, right?
1169835230 M * Bertl yep, it disables CTRL-ALT-DEL
1169835256 M * Bertl i.e. it makes it send SIGINT instead of rebooting
1169835269 M * Hollow yep, so, only the assertion is strange
1169835312 M * Bertl yes, I would suggest to make sure not to pass _any_ filehandles to the guests, except for a unix socket or so (stdin/out) via pipes and such
1169835336 M * Hollow ok, will check that
1169835381 M * Hollow dinner time now though.. will be back later
1169835471 J * danychouinard ~dchouinar@206.167.65.15
1169835552 M * Bertl welcome danychouinard!
1169835610 J * shedi ~siggi@ftth-237-144.hive.is
1169835759 Q * Hurga Remote host closed the connection
1169835760 M * danychouinard Hello Bertl :)
1169837301 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1169837308 M * Bertl wb thunder1!
1169837974 Q * FireEgl Read error: Connection reset by peer
1169838222 M * Bertl okay, translocating now ... will be back later ...
1169838227 N * Bertl Bertl_oO
1169838774 Q * ex Ping timeout: 480 seconds
1169839153 J * ex ex@valis.net.pl
1169839913 Q * lilalinux Remote host closed the connection
1169839961 J * Aiken ~james@ppp121-180.lns1.bne4.internode.on.net
1169840147 Q * rob-84x^ Quit: brb
1169840154 J * rob-84x^ ~rob@submarine.ath.cx
1169845453 P * stefani I'm Parting (the water)
1169845909 Q * DavidS Quit: Leaving.
1169846227 J * bonbons ~bonbons@83.222.37.103
1169846847 Q * danychouinard Quit: Quitte
1169847295 J * comfrey ~comfrey@70.91.185.84
1169847615 N * Bertl_oO Bertl
1169847621 M * Bertl back now ...
1169847670 M * daniel_hozac oh, while i remember it, i think testfs's test 116 is broken.
1169847693 M * Bertl no, not broken, just not updated yet
1169847715 M * daniel_hozac hmm?
1169847718 M * Bertl (at least I have it on my todo list here :)
1169847734 M * daniel_hozac it references a variable called $ER which, AFAICT, doesn't exist.
1169847751 M * daniel_hozac i assume the $ER is supposed to be ^^^.
1169847841 M * daniel_hozac or is ER meant to be ^^^ or ... depending on the availability of full COW?
1169847868 M * Bertl I guess I had something like that in mind
1169847921 M * Bertl but for now, I guess it is best to change that to ^^^ hardcoded
1169847937 M * daniel_hozac yeah, that's what i did to get it to succeed on the 2.6.16 kernel.
1169847940 M * Bertl as the 116 test is disabled for COW atm
1169847952 M * Bertl (full cow that is)
1169847957 M * daniel_hozac yeah.
1169847977 M * Bertl my notes here say, todo: make a separate CoW test section :)
1169848000 M * Bertl and indeed there is one, just without tests :)
1169848006 M * daniel_hozac hehe, ok :)
1169848050 M * Bertl my first idea was to simply invert the test result for 116
1169848080 M * Bertl but I figured soon, that with full cow, only the first result of 116 would make sense, the others would not operate on cow links anymore
1169848106 M * Bertl thus I decided to make that a separate section, the ER was probably left over
1169848687 J * rhodes ~rhodes@hc652a84b.dhcp.vt.edu
1169848710 Q * rhodes 
1169849752 Q * bonbons Quit: Leaving
1169850410 J * toom ~toom@coco.docisland.org
1169850420 M * toom hi !
1169850422 M * Bertl welcome toom!
1169850461 M * toom I found a bug in vserver when unionfs is used
1169850474 M * Bertl unionfs is in mainline now?
1169850504 M * toom /proc/self/exe refers real absolute path (in the root vserver)
1169850529 M * Bertl root vserver?
1169850565 M * toom the real host
1169850586 M * toom vserver-stat call it "root server"
1169850607 M * Bertl okay, /proc/self/exe should always point to the absolute path (relative to the root)
1169850685 M * toom I use 2.6.19.2-grsec2.1.10-vs2.2.0-rc8
1169850701 M * Bertl okay, and what's the problem?
1169850778 M * toom the rootfs of my vserver (h1) is /vservers/h1
1169850792 M * Bertl okay
1169850828 M * toom /vservers/h1 is a unionfs of /vservers/h2(ro) and /vservers/delta/h1(rw)
1169850907 M * toom when I'm in h1: ls -al /proc/self/exe 
1169850929 M * toom /proc/self/exe -> /vservers/h2/bin/ls
1169850993 M * Bertl i.c. interesting ...
1169851012 M * Bertl could you try without grsec?
1169851021 M * toom no
1169851056 M * toom you think grsec can be the cause ?
1169851069 M * Bertl could be, also the unionfs confuses me
1169851089 M * Bertl there is no unionfs in mainline or Linux-VServer AFAIK
1169851125 M * Bertl nevertheless, I'm taking that serious and will check it shortly
1169851181 M * toom I patched my kernel to add unionfs support
1169851203 M * Bertl maybe that messed up the pathes?
1169851221 J * FireEgl ~FireEgl@adsl-4-58-61.bhm.bellsouth.net
1169851277 M * toom maybe
1169851316 M * Bertl okay, I'd suggest the following:
1169851319 Q * shedi Ping timeout: 480 seconds
1169851332 M * Bertl you give it a try with mainline Linux-VServer (without unionfs)
1169851345 M * Bertl and I'll check what might cause the wrong pathes ...
1169851388 M * toom there are no pb without unionfs
1169851404 M * Bertl ah? so there the pathes are fine?
1169851418 M * toom when I start h2 vserver /proc/self/exe is fine
1169851430 M * Bertl btw, what do you use/need the unionfs for?
1169851517 M * toom I've configured a main vserver and all other is a derived from it to ruspecific task (service)
1169851550 M * toom when I update the main vserver, all my vservers are updated
1169851580 M * Bertl i.c., you might be able to do that with --bind mounts too, no?
1169851590 M * toom when I install a packages in the main vserver, it is available for all other
1169851633 M * Bertl another idea
1169851634 M * toom I need different configurations in vserver
1169851649 M * Bertl try to do the unionfs mount _inside_ the guest namespace
1169851667 M * toom and when I remove a file in vserver I don't want the file disapear in all vserver
1169851683 M * toom ah !
1169851692 M * Bertl i.e. add it to each guest config (fstab)
1169851697 M * toom is it possible ?
1169851705 M * toom fstab.remote ?
1169851716 M * Bertl the normal fstab should suffice
1169851726 M * toom ok
1169851736 M * Bertl note: I don't think it was ever tested with unionfs
1169851742 M * toom I try it now
1169851753 M * Bertl so it might be necessary to add the mount to one of the pre-init scripts
1169851762 M * Bertl probably the initialize script
1169851782 M * toom pre-init scripts ?
1169851795 M * daniel_hozac fstab should do it, IMHO.
1169851814 M * daniel_hozac unless unionfs requires some additional setup.
1169851851 J * shedi ~siggi@ftth-237-144.hive.is
1169851972 M * toom I can't start my vserver
1169852018 M * toom chroot-sh: open("/etc/mtab"): No such file or directory
1169852035 M * daniel_hozac i don't think that's a fatal error.
1169852050 M * toom fakerunlevel: open("/var/run/utmp"): No such file or directory
1169852059 M * toom Failed to start vserver '...'
1169852061 M * daniel_hozac but what did you put in the fstab?
1169852090 M * toom ho ... I do a typo 
1169852195 M * toom I get the same errors
1169852221 M * toom I add in the first line of my fstab : unionfs /vservers/h1       unionfs dirs=/vservers/delta/h1=rw:/vservers/h2=ro 0    0
1169852241 M * daniel_hozac the second path is inside the guest, so that should be /.
1169852287 M * toom I don't understand 
1169852302 M * daniel_hozac /vservers/h1 should be /.
1169852309 M * toom ah yes !
1169852312 M * daniel_hozac it's relative to the guest root.
1169852314 Q * dna_ Ping timeout: 480 seconds
1169852381 M * toom I get weird error messages
1169852385 M * toom /etc/init.d/rc: line 44: /dev/null: Permission denied
1169852465 M * daniel_hozac oh right, you need dev in the options
1169852593 M * toom which options ?
1169852610 M * daniel_hozac the fourth column are options.
1169852613 M * daniel_hozac add ,dev.
1169852675 M * toom this is an option of unionfs ?
1169852682 M * toom ah ok
1169852685 M * daniel_hozac no, of the mount.
1169852730 M * toom good ! this solve the pb !
1169852744 M * toom and /proc/self/exe is fine !
1169852752 M * Bertl good!
1169852754 M * toom thanks a lot !
1169852760 M * Bertl you're welcome!
1169853109 Q * paperboy 
1169853981 P * toom 
1169854944 M * Bertl okay, I guess I call it a day ... off to bed now ... have a good one everyone! cya!
1169854951 N * Bertl Bertl_zZ