1226621722 Q * hparker Quit: Quit
1226622076 Q * pisco Ping timeout: 480 seconds
1226622226 Q * dowdle Remote host closed the connection
1226622428 J * dowdle ~dowdle@scott.coe.montana.edu
1226622495 J * pisco ~pisco@86.59.118.153
1226622788 Q * independence Ping timeout: 480 seconds
1226623647 J * independence independen@titan.blinkenshell.org
1226623679 J * hparker ~hparker@2001:470:1f0f:32c:212:f0ff:fe0f:6f86
1226624759 Q * xdr Ping timeout: 480 seconds
1226624981 Q * dowdle Remote host closed the connection
1226630249 Q * geb Remote host closed the connection
1226634075 J * aj__ ~aj@p5B23CFE4.dip.t-dialin.net
1226634366 Q * pisco Ping timeout: 480 seconds
1226634506 Q * derjohn_foo Ping timeout: 480 seconds
1226635139 J * pisco ~pisco@86.59.118.153
1226637873 J * xdr ~xdr@gote2.179.cust.blixtvik.net
1226638300 N * quinq qzqy
1226645078 Q * FireEgl Ping timeout: 480 seconds
1226645746 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com
1226645839 M * fb hello :)
1226646093 N * Bertl_zZ Bertl
1226646098 M * Bertl morning folks!
1226646159 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1226646663 M * Bertl okay, off for now .. bbl
1226646668 N * Bertl Bertl_oO
1226647298 J * mtg ~mtg@vollkornmail.dbk-nb.de
1226648734 Q * aj__ Ping timeout: 480 seconds
1226649875 M * pisco moin folks, Moin Bertl.
1226650023 M * pisco it seems that i solved or understand my problem with non working 'chage' command inside a guest
1226650082 M * pisco it seems selinux related. I disabled selinux on the host machine.
1226650230 M * pisco am i right?
1226651787 J * dna ~dna@201-244-dsl.kielnet.net
1226652147 J * davidkarban ~david@193.85.217.71
1226653871 Q * pisco Ping timeout: 480 seconds
1226654220 J * pisco ~pisco@86.59.118.153
1226654478 N * Bertl_oO Bertl
1226654482 M * Bertl back now ..
1226654549 M * Bertl pisco: if you say so ... I don't see how chage is selinux related, but maybe it is
1226654571 M * Bertl in any case, if you enable selinux, you need to configure it properly for guest usage
1226655952 Q * kir Ping timeout: 480 seconds
1226656293 M * arekm can rss mem limit be changed while guest is running?
1226656299 M * Bertl  sure
1226656317 M * arekm oh 8)
1226656322 M * Bertl vlimit --help
1226656343 M * arekm was there some command to reapply limits from config btw? 
1226656363 M * Bertl not that I know, that is on the feature request list for a while now ...
1226657270 M * pisco Bertl: anybody who can configure selinux corretly is suspicious ;)
1226657313 M * Bertl probably, to me it always seemed that selinux was not designed to be used at all :)
1226657481 Q * dna Quit: Verlassend
1226660581 M * bluegene_  
1226660612 M * bluegene_ Anyone knows whether vserver is compatible with openssi, or similiar?
1226660871 Q * nkukard Quit: Leaving
1226660888 M * Bertl in theory it is compatible, but in practice it will need a few minor adjustments
1226660911 M * bluegene_ What kind of impact would guests see?
1226660963 M * Bertl hmm?
1226660988 M * bluegene_ Unified process space, etc. How would this affect a running guest?
1226661252 M * Bertl well, most likely you would want to move guests around as units, but in theory, you could even distribute a guest between nodes
1226661261 M * Bertl s/units/entities/
1226661290 M * bluegene_ It would probably suck on GBit Ethernet, though.
1226661311 M * bluegene_ Migrating running guests is kinda planned with vserver, right?
1226661365 M * bluegene_ Oh, and while you're there: do you see any problems in running guests of a PVFS2 store? It probably would support vhashify & Co, though.
1226661397 M * Bertl not really, no point in doing so atm .. but mainline is working on snapshot mechanisms, so as soon as that works, it will be supported on Linux-VServer too
1226661422 M * bluegene_ That's good, so the functionality will be there eventually.
1226661436 M * Bertl no idea what PVFS2 is, but I have the feeling it is lacking Linux-VServer support
1226661457 M * bluegene_ It's a cluster file system. Used the local file system to store things in.
1226661521 M * bluegene_ What must a file system support in order to host vserver guests?
1226661547 M * Bertl usually you want the xattr stuff and the taggin (although you can do without that)
1226661568 M * Bertl if you are using very recent kernels and util-vserver, you do not need the xattr stuff (for barrier)
1226661589 M * bluegene_ Apparently it does xattr, according to a brief Google search.
1226661612 M * Bertl yes, but I doubt it knows about the Linux-VServer specific attributes :)
1226661651 M * bluegene_ Is that a lot of code? In one piece, or scattered?
1226661663 J * An|ma ~Anima@host58-32-dynamic.52-79-r.retail.telecomitalia.it
1226661688 M * Bertl welcome An|ma!
1226661693 M * Bertl bluegene_: not much, but as I said, you might get away without it (with recent kernels)
1226661708 M * bluegene_ I think I'll try that once I get a couple of new boxes.
1226661719 M * bluegene_ Thanks, Bertl. Most helpful, as always.
1226661763 M * Bertl you're welcome!
1226662008 J * Alekz ~W1z@218.248.255.193
1226662021 P * Alekz 
1226663526 J * kir ~kir@swsoft-msk-nat.sw.ru
1226664641 Q * An|ma Ping timeout: 480 seconds
1226665381 Q * Aiken Quit: Leaving
1226667212 M * Hawq hello
1226667221 M * Bertl hello :)
1226667290 M * Hawq Bertl: how about little X + console switching debugging? I'm about to boot 2.6.27.5 with newest patch and debug enabled
1226667378 M * Bertl well, not too much time atm ... but enable signal related debugging and give it a try
1226668170 N * qzqy quinq
1226668462 M * Hawq hm. nothing appeared in dmesg after trying to switch consoles. which setting is for signal debugging? I've turned on all of them.
1226668529 M * Bertl sec
1226668803 M * Bertl try sysctl -w vserver.debug_misc=144
1226669293 M * Hawq still nothing. just this appears 1 entry per second vxD: e4a66aa0: check_kill_permission(0,e4a51f24,f7b4d280[#102,6720])
1226669422 M * Bertl well, then Linux-VServer doesn't interfere with signals 
1226669440 M * Bertl i.e. we have to look somewhere else for the misbehaviour
1226669535 M * Hawq so what should I try now?
1226669636 M * Bertl try to analyze what actually happens when you switch
1226669652 M * Bertl i.e. does the keyboard focus switch to the new console?
1226669715 M * Bertl I also get the feeling, this might be related to way we hide pids
1226669841 M * Bertl this could be verified by creating a new text console from the same context your X is in
1226669852 M * Bertl and then checking if that one can be switched to
1226669875 M * Bertl it might also make sense to try first with text consoles only
1226669915 M * Hawq well, after pressing ctrl-alt-f1 there are no visible changes. it looks like keyboard stops working while actually its output goes to tty1. in the same time X are working normally, I can do anything in it using mouse.
1226669960 M * Bertl okay, so you are seeing the X console, but at least the keyboard input has already switched to a different one, yes?
1226669970 M * Hawq yes
1226670006 M * Hawq I'm now creating console from guest. just a moment...
1226670513 M * Hawq ok. with just text consoles it works as it should
1226670560 M * Hawq with x and text console running in same context it doesn't work. behaves same way as with x in guest and text console on host
1226670632 M * Hawq btw, while on host I now see /proc with fix suggested by daniel_hozac, in guest I still don't have access to /proc. just letting you know.
1226670652 M * daniel_hozac that's because you're hiding it.
1226670765 M * Hawq daniel_hozac: evil setattr was removed but I didn't upgraded util-vserver packages yet.
1226670948 M * Bertl okay, so text console switch works ... interesting
1226670991 M * Bertl I don't know how the terminal switching is implemented in X
1226671005 M * Bertl but I presume (from what I saw) that it does this with
1226671031 M * Bertl VT_ACTIVATE and VT_WAITACTIVE)
1226671047 M * Bertl what confuses me, is that the trace shows both to succeed
1226671083 M * Bertl but I suspect there is some kind of signalling involved too, which is missing for whatever reason
1226671107 M * Bertl (and X would act upon that signal and actually finish the console switch)
1226671127 M * Hawq just encountered one more missbehavior. when I manually stop X and then stop guest I'm getting back to tty1. when I'm stopping guest without stopping X first X display is gone but left in graphics mode and there is no way to get text console even with chvt. I guess I'm getting text console but display is hanged.
1226671393 M * Bertl try to find some more docu or information on the net (maybe from the Xorg folks) about the actual switching procedure
1226671412 M * Bertl (also interesting is, does switching back to X work/succeed)?
1226671494 J * chI6iT41 ~chigital@services.mivitec.net
1226671583 M * Hawq after trying ctrl-alt-f1 I would normally go back to X with alt-f9 or whatever console x is running on
1226671610 M * Hawq but that doesn't work, I have to blindly type chvt 9 to get my keyboard back to X session
1226671622 M * Bertl ah, so that works then
1226671676 M * Bertl what if you switch (from the host) to a text console (after strating X)
1226671686 M * Bertl and then try to reach the 'new' X session?
1226671921 M * Bertl the following test come to my mind:
1226671962 M * Bertl - start X from with a chroot to the guest dir (make sure to mount /proc there) and see if that makes X switching work
1226671987 M * Bertl - disable proc security, avoid running the vprocunhide or any other setattr script on proc, and see if that helps
1226672049 J * tramjoe_merin ~tramjoe@193.41.238.151
1226672744 M * Hawq when run from within a chroot switching works
1226672756 M * Hawq now I'm building kernel with proc security disabled
1226672765 M * Bertl okay
1226672848 M * daniel_hozac isn't setattr --~hide -Rx /proc equivalent?
1226672937 M * Bertl should be, but better be sure there
1226672985 M * daniel_hozac well, with initscripts that do --hide, i don't think the kernel's default really matters :)
1226673012 M * Bertl as I said, all init scripts doing setattr should be disabled
1226673034 M * Hawq daniel_hozac: no more --hide in initscripts :) I just upgraded util-vserver packages
1226673066 M * Hawq Bertl: yes, none of vserver init script is run. disabled them all
1226673146 M * Bertl if that fails, we try to hit the guest with all bcapabilities we have
1226673167 M * Bertl (if that makes it work, we are definitely missing some cap check)
1226673406 M * Hawq booted kernel w/o proc restriction, no initscripts run, started guest, switching doesn't work
1226673914 M * Hawq now trying running guest with all bcaps from http://svn.linux-vserver.org/svn/util-vserver/trunk/lib/bcaps-v13.c
1226673936 M * Hawq and nope, still no luck
1226674803 M * Bertl very strange ... okay, I have to leave now, but we can try a little later with the process stuff reverted (i.e. without process isolation, but with everything else on)
1226674810 N * Bertl Bertl_oO
1226674942 M * Hawq Bertl_oO: actually, I have to leave too but will be back later :)
1226676988 J * pmenier ~pmenier@ACaen-152-1-12-17.w83-115.abo.wanadoo.fr
1226677015 Q * mtg Quit: Verlassend
1226678216 Q * chI6iT41 Ping timeout: 480 seconds
1226678298 Q * tramjoe_merin Quit: using sirc version 2.211+KSIRC/1.3.12
1226678525 Q * davidkarban Quit: Ex-Chat
1226678895 J * dowdle ~dowdle@scott.coe.montana.edu
1226679101 J * geb ~geb@4.4.82-79.rev.gaoland.net
1226680163 Q * doener_ Read error: Connection reset by peer
1226680313 J * doener ~doener@i577AE40B.versanet.de
1226680979 N * morrigan_oO morrigan
1226682668 N * Bertl_oO Bertl
1226682672 M * Bertl back now ...
1226682942 J * hparker ~hparker@2001:470:1f0f:32c:215:f2ff:fe60:79d4
1226683420 Q * kir Quit: Leaving.
1226685733 M * Hawq wb Bertl 
1226686118 J * hijacker ~hijacker@87-126-142-51.btc-net.bg
1226686947 J * ktwilight_ ~ktwilight@37.87-66-87.adsl-dyn.isp.belgacom.be
1226686947 Q * ktwilight__ Read error: Connection reset by peer
1226686973 M * Bertl daniel_hozac: didn't you have a patch to remove the pid related checks from a recent kernel?
1226686988 M * Bertl (in preparation for the pid space change)
1226687092 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-vspid-revert03.diff seems to be the most recent diff, but i think i've got a tree.
1226687112 M * daniel_hozac a 2.6.27 tree, that is.
1226687237 M * daniel_hozac hmm. i can't find it.
1226688782 M * Hawq uhm, I must go :( cu tomorrow.
1226689324 M * Bertl k, cya!
1226690444 J * nkukard ~nkukard@196.212.73.74
1226690587 J * dkg ~dkg@lair.fifthhorseman.net
1226690617 M * Bertl welcome dkg!
1226690632 M * dkg hey folks --  are y'all aware of the bad interaction between vservers and the -H argument to GNU ps ?
1226690652 M * Bertl hmm? please elaborate!
1226690659 M * dkg compare the output of "ps -e"
1226690663 M * dkg with "ps -eH"
1226690672 M * dkg try it on a vserver, and on a non-vserver
1226690673 M * Bertl inside a guest, on the host or what?
1226690683 M * dkg inside a guest, the -H doesn't list everything
1226690691 M * dkg i think -H tries to calculate the forest from pid 0.
1226690704 M * dkg but there is no pid 0, so it just builds the parent tree of the current process.
1226690705 M * Bertl works fine here
1226690713 M * dkg really?
1226690715 M * dkg hrm.
1226690721 M * Bertl sec, let me upload the output
1226690745 M * dkg dkg@smoke:~$ ps -e | wc -l
1226690745 M * dkg 35
1226690745 M * dkg dkg@smoke:~$ ps -eH | wc -l
1226690745 M * dkg 1
1226690745 M * dkg dkg@smoke:~$ 
1226690763 M * Bertl http://paste.linux-vserver.org/12603
1226690878 M * dkg interesting.
1226690907 M * dkg ok, i'm not using GNU ps -- i take it back.  i'm using /bin/ps from procps.sourceforge.net
1226690942 M * dkg Bertl: is this within a guest?
1226690944 M * Bertl I think I can test compile that too .. give me a few minutes
1226690948 M * daniel_hozac dkg: known bug.
1226690949 M * Bertl dkg: yep
1226690967 M * dkg daniel_hozac: can you point me toward the bug?  i'd like to fix it.
1226690978 M * dkg (or at least comment on it in the right place)
1226691019 M * daniel_hozac plain initstyle, right?
1226691026 M * dkg yea, plain initstyle.
1226691031 M * dkg is that the problem?
1226691046 M * daniel_hozac yeah.
1226691053 M * daniel_hozac fs/proc/array.c:task_state
1226691093 M * daniel_hozac you need to map the pids and set ppid to 0 if p is the guest's init.
1226691136 M * daniel_hozac EasyFix really, but i didn't want to spend time on code we're supposed to remove anyway.
1226691158 M * dkg why is it scheduled for removal? (sorry i'm not up-to-speed on this stuff)
1226691178 M * daniel_hozac pid namespaces make it obsolete.
1226691239 M * Bertl dkg: anyway, a proper fix (read patch) will be accepted and included
1226691308 M * dkg Bertl: OK, thanks.  It'll likely take me a while to learn my way around the code, but i'll look into providing that.
1226691328 M * Bertl I think we can help there (to some extend :)
1226691353 M * Bertl a simple 'workaround' for you might be to switch to sysv init
1226691357 M * Bertl (style)
1226691371 M * dkg i'll take you up on the offer of help once i get a chance, but i'm afraid it's probably more time-limited on my end (not your fault)
1226691390 M * dkg Bertl: is sysv init the recommended style?
1226691409 M * dkg I'd really like to have my vservers monitored by a supervising process (runit)
1226691423 M * dkg and it seemed to me like plain init was the way to do that.
1226691432 M * Bertl no, there is no 'recommended style' it was just a suggestion
1226691456 M * dkg Bertl: OK, makes sense.
1226691495 M * dkg and i feel more confident patching code that's gonna be removed anyway ;)
1226691502 M * dkg thanks for the pointers, y'all.
1226691513 M * Bertl you're welcome! feel free to hang around!
1226691908 J * chI6iT41 ~chigital@tmo-100-43.customers.d1-online.com
1226695303 Q * dowdle Remote host closed the connection
1226695484 J * dowdle ~dowdle@scott.coe.montana.edu
1226696035 Q * hijacker Quit: Leaving
1226696236 J * mrfree ~mrfree@host253-180-dynamic.13-79-r.retail.telecomitalia.it
1226696768 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1226697374 J * yarihm ~yarihm@77-56-182-18.dclient.hispeed.ch
1226699364 J * Piet ~piet@86.59.118.153
1226701248 Q * mrfree Quit: Leaving
1226701591 Q * bonbons Quit: Leaving
1226702366 J * Aiken ~Aiken@ppp118-208-13-1.lns1.bne1.internode.on.net
1226704114 Q * chI6iT41 Ping timeout: 480 seconds
1226704253 Q * hparker Quit: Quit
1226704754 J * chI6iT41 ~chigital@tmo-096-167.customers.d1-online.com
1226704997 J * hparker ~hparker@2001:470:1f0f:32c:212:f0ff:fe0f:6f86
1226705249 Q * FireEgl Ping timeout: 480 seconds
1226705750 J * aj__ ~aj@p5B23CFE4.dip.t-dialin.net
1226705851 Q * SpComb Ping timeout: 480 seconds
1226706264 J * SpComb terom@zapotek.paivola.fi
1226706310 Q * aj__ Ping timeout: 480 seconds