1161129603 M * ntrs what signalling?
1161129621 M * Bertl host-context signalling
1161129631 M * ntrs How long until 41 is out?
1161129641 M * Bertl probably less than an hour
1161129669 M * ntrs hmm, ok. I can wait. Please let me know as soon as it's ready. I'll try rc41
1161129680 M * Bertl ok, will do so
1161129811 M * rgl_ daniel_hozac, ping :D
1161129994 M * rgl_ so later here!
1161130004 M * rgl_ gtg.  g'night guys!
1161130010 M * Bertl k, have a good one!
1161130093 M * ntrs Bertl, another guy on a x86_64 with 2.6.18 and rc38 complains of a mysql problem.
1161130128 M * ntrs Warning: mysql_connect(): Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (11)
1161130139 M * ntrs From what I can see the mysql.sock file is there.
1161130152 M * Bertl maybe the process died away
1161130157 M * Bertl check the limits
1161130201 M * ntrs no limit problems. all limits are ok.
1161130213 M * Bertl okay, any mysql logs?
1161130228 M * Bertl I'd suspect nothing is connected to the socket anymore
1161130245 Q * rgl_ Quit: Fui embora
1161130248 M * Bertl (what does /etc/init.d/mysql status say, e.g.?
1161130261 M * ntrs any connection attempt results in that error message. let me try tcp/ip socket
1161130274 M * ntrs sec
1161130390 M * ntrs mysqld stop cannot stop the service.
1161130568 M * ntrs after a killall -9 the processes are killed. After I start the service it works for a while
1161130601 M * ntrs this is with mysql 4.1.11
1161130607 M * Bertl define while and more important, what changes when it stops working?
1161130618 M * Bertl obviously the socket was not connected anylonger
1161130625 M * ntrs about a day before it stops
1161130646 M * ntrs and the socket is no longer connected, and the init script cannot stop the service.
1161130669 M * ntrs I think at some point with another guest an upgrade to 4.1.21 resolved the issue. I will try that now.
1161130674 M * Bertl sounds like a mysql bug to me ... 
1161130755 M * ntrs Well, that's what I remember, but I may be somewhat wrong.
1161130762 M * ntrs So, I'll try the upgrade first.
1161130803 M * Bertl okay
1161132006 Q * gerrit Ping timeout: 480 seconds
1161132845 J * dmax ~semaj@81.193.57.215
1161132851 M * Bertl welcome dmax!
1161132857 Q * s0undt3ch Ping timeout: 480 seconds
1161132970 J * s0undt3ch ~s0undt3ch@81.193.57.215
1161134422 J * rob-84x^ rob@submarine.ath.cx
1161134701 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.1-vs2.1.1-rc41.diff
1161134707 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.1-vs2.0.2.2-rc2.diff
1161134792 T * Bertl http://linux-vserver.org/ <- new and shiny | latest stable 2.02.1, exp 2.02.2-rc2, devel 2.1.0, exp 2.1.1-rc41, stable+grsec 2.0.2.1 | util-vserver-0.30.211 | libvserver-1.0.2 & vserver-utils-1.0.3 | 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 ;)
1161134860 M * Bertl ntrs: ping!
1161134874 M * ntrs Bertl, yes?
1161134985 M * ntrs is rc41 against 2.6.18 or 2.6.18.1?
1161135030 M * Bertl 2.6.18.1
1161135040 M * ntrs ok
1161135425 M * Bertl ntrs: you might get some warnings (in dmesg) which look like stack traces (oopses) at first glance
1161135442 M * ntrs from what?
1161135451 M * Bertl ntrs: those are intentional and should not happen too often (unless you do strange things :)
1161135475 M * ntrs Is that really needed?
1161135478 M * Bertl the only thing you are supposed to do with them is to report them back to me
1161135510 M * ntrs ok, got it.
1161135895 M * Bertl derjohn2: rc41 is there, JFYI
1161137218 M * Bertl ntrs: hmm, please revert the following patch (which was not supposed to be included)
1161137231 M * Bertl http://vserver.13thfloor.at/Experimental/delta-lback-feat01.diff
1161137248 M * ntrs Will there be a full rc42 shortly?
1161137278 M * Bertl no, but I can update the rc41 in place, okay?
1161137423 M * Bertl okay, rc41 was updated in place, please reload (or alternatively revert the patch)
1161137438 M * Bertl sorry for the inconvenience ...
1161138044 Q * bronson Quit: Ex-Chat
1161140095 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1161140100 M * neuralis Bertl: what's the recommended way of doing upgrades to the host system which need to propagate into CoW guests?
1161140104 J * ensc ~irc-ensc@p54B4FF99.dip.t-dialin.net
1161140136 M * Bertl neuralis: hmm, probably the best way is to do the upgrade in both places, then reunify
1161140157 M * Bertl but this process could be combined, if you _know_ what you update where
1161140201 M * neuralis in what sense?
1161140208 M * Bertl let me give an example:
1161140225 M * Bertl - let's assume your host/guest system is rpm based
1161140259 M * Bertl - in this case, it is _a lot_ simpler to upgrade the rpm in both, host and guest
1161140280 M * Bertl - then, right after that, run vhashify and 'reunify' common files
1161140301 M * Bertl now another example, where a different approach could be taken:
1161140337 M * Bertl - let's assume you have a 'known' database file on both, host and guest (readonly of course, otherwise it would not stay unified for long)
1161140352 J * FireEgl FireEgl@Sebastian.Atlantica.US
1161140371 M * Bertl - let's further assume, you do the update while the guest is 'down' or in a controlled state
1161140413 M * Bertl - you could now 'remove' the immutable flag on the link, update the linked file in place, and reestablish the immutable flag
1161140422 M * neuralis okay, that's what i figured.
1161140475 M * Bertl note, while this is _not_ secure if you consider the guest admin potentially hostile, the re-hashification is secure
1161140491 M * neuralis there's no reason why i can't do the mutable/in-place update/immutable on a guest that's running, is there?
1161140504 M * Bertl (except for the security, no :)
1161140628 M * Bertl but keep in mind, that packaging system often have side effects attached to package isntalls, which _might_ require to be run in the 'proper' context
1161140718 M * neuralis right
1161140743 M * neuralis i'm not sure i see the security issue with the in-place update. what is it?
1161140771 M * Bertl an evil guest admin could write to the file while it is unprotected
1161140787 M * Bertl which would cause the host file to contain guest data/code
1161140832 M * Bertl usually the simple, straight forward approach (install it twice, reunify afterwards) is the best
1161140842 J * _mcp ~hightower@wolk-project.de
1161140843 Q * mcp Read error: Connection reset by peer
1161140876 M * Bertl also has two advantages:
1161140896 M * Bertl - files will be reunified (which might otherwise be not)
1161140900 M * neuralis oh, that race condition. i considered that, but figured it wouldn't be an issue since i'd do a verification run after the updates.
1161140910 M * Bertl - scripts and side effects run in the proper contexts
1161140938 M * Bertl let me give two examples here:
1161140957 M * Bertl a) guest admin decides to modify /etc/profiles
1161140970 M * Bertl   shortly afterwards, he decides to change it back
1161140985 M * Bertl   although the files are identical now, they are _not_ unified
1161141003 M * Bertl b) guest admin changes /etc/ssh/sshd_config
1161141020 M * Bertl   now sshd get's updated and adds a few lines with a sed script
1161141042 M * Bertl   the dual update would work quite fine here, while the 'in place' approach fails
1161141112 M * neuralis right. the correct way to go about things is a bit non-trivial since i'm using things in a strange way :)
1161141166 M * Bertl you are free to do so :)
1161142630 M * Bertl have a good one everyone! I'm off to bed ...
1161142638 N * Bertl Bertl_zZ
1161144352 J * mcp ~hightower@wolk-project.de
1161144429 Q * _mcp Read error: Connection reset by peer
1161144431 Q * FireEgl Read error: Connection reset by peer
1161149210 J * bronson ~bronson@c-71-198-75-160.hsd1.ca.comcast.net
1161149702 Q * bronson Ping timeout: 480 seconds
1161149716 J * bronson ~bronson@adsl-64-161-106-11.dsl.snfc21.pacbell.net
1161150155 Q * ruskie Remote host closed the connection
1161150197 J * FireEgl FireEgl@Sebastian.Atlantica.US
1161151142 J * ruskie ~ruskie@ruskie.user.oftc.net
1161151262 Q * ruskie 
1161151456 J * ruskie ~ruskie@ruskie.user.oftc.net
1161152043 Q * bronson Ping timeout: 480 seconds
1161152788 J * bronson ~bronson@c-71-198-75-160.hsd1.ca.comcast.net
1161153874 P * anonc adios
1161153947 J * _mcp ~hightower@wolk-project.de
1161153967 Q * mcp Read error: Connection reset by peer
1161153974 N * _mcp mcp
1161153982 Q * SNy Read error: Operation timed out
1161154028 J * SNy b32cba1b6e@bmx-chemnitz.de
1161154332 Q * Aiken Ping timeout: 480 seconds
1161154596 Q * bronson Ping timeout: 480 seconds
1161154608 J * bronson ~bronson@adsl-64-161-106-11.dsl.snfc21.pacbell.net
1161157171 Q * _are_ Ping timeout: 480 seconds
1161157536 Q * mountie Ping timeout: 480 seconds
1161157580 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1161158673 J * dna_ ~naucki@102-218-dsl.kielnet.net
1161158807 J * rgl Rui@217.129.151.190
1161158811 M * rgl good morning
1161158846 M * doener morning rgl
1161159314 J * meandtheshell ~markus@85-124-38-146.dynamic.xdsl-line.inode.at
1161159467 J * _are_ ~are@62.112.159.81
1161160486 Q * shedi Quit: Leaving
1161160748 Q * bronson Ping timeout: 480 seconds
1161160928 J * mire ~mire@7-167-222-85.adsl.verat.net
1161160964 Q * mire 
1161160987 J * mire ~mire@7-167-222-85.adsl.verat.net
1161161635 M * rgl vserver-info 0 VERIFYCAP fails here :(
1161161640 M * rgl 'VERIFYCAP' can be executed as root only
1161161645 M * rgl any ideia why?
1161161929 J * bronson ~bronson@c-71-198-75-160.hsd1.ca.comcast.net
1161164691 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at
1161165240 M * renihs hmm strange, i have here a vserver (www) which has a bindmount from another vserver (ftp), i can create directories and files, they also show up in the www vserver but *only* with ls 
1161165252 M * renihs if i try ls -l i get no such file or directory for some files
1161165272 M * renihs i dont really get that?, ls works, ls -l doesnt? i cannot access or stat the files neither
1161165340 M * daniel_hozac rgl: your uid isn't 0
1161165370 M * daniel_hozac renihs: you're using xid tagging. you're not able to access the files in any way from the other guest.
1161165391 M * renihs daniel_hozac, that was my first thought too (its not my box) but there is no xid tagging in place
1161165401 M * renihs  lsxid /vservers/
1161165401 M * renihs                    0  /vservers/
1161165401 M * renihs                    0  /vservers/dns
1161165403 M * renihs all 0
1161165410 M * daniel_hozac and the same for the files?
1161165443 M * renihs hmm nope, you are right
1161165446 M * renihs gruml
1161165447 M * daniel_hozac note that 0 actually means xid tagging is enabled. if it's disabled, you'll get !!ERR!!.
1161165447 M * renihs :
1161165461 M * renihs ya but 0 is host vserver?
1161165469 M * daniel_hozac yes, 0 is the host's xid.
1161165559 M * renihs hmm can i disable automatic xid tagging for some dirs?
1161165584 M * renihs this is a reiserfs, it xid tagging wasnt enabled until now (so no prob arose) 
1161165604 M * renihs grml, i dont like this box
1161165621 M * daniel_hozac no, xid tagging is a filesystem property.
1161165629 M * renihs hmm 
1161165656 M * daniel_hozac you're going to have to unmount it and then mount it again without tagxid.
1161165661 M * renihs ah ok
1161165674 M * renihs hmm i guess i will make a small image which i bindmount
1161165699 M * renihs daniel_hozac, thx alot!
1161165836 J * debugger Rui@217.129.151.190
1161166307 Q * rgl Ping timeout: 480 seconds
1161166487 Q * debugger Ping timeout: 480 seconds
1161166855 M * matti :)
1161166889 Q * Nei Remote host closed the connection
1161167454 M * renihs hmm daniel_hozac this time i made a 500M file and created a ext3 fs in it, and mounted it to both vserver with /vservers/image /var/www/upload        ext3 loop,rw
1161167463 M * renihs however if i touch a file in 1 vserver it doesnt show up in the other
1161167483 M * daniel_hozac do both guests have it mounted?
1161167486 M * renihs i could mount it in the host system and binmount it but
1161167489 M * renihs yes
1161167501 M * renihs but lemme doublecheck procedure
1161167612 M * renihs (ftp vserver) /dev/loop1 /home/pub ext3 rw,nodev,data=ordered 0 0 
1161167642 M * renihs (www server) /dev/loop3 /var/www/upload ext3 rw,nodev,data=ordered 0 0
1161167669 M * renihs says cat /proc/mounts
1161167783 M * daniel_hozac wouldn't it be better to use the same loop device?
1161167850 M * renihs sounds like a plan but hmm
1161167865 M * renihs i dunno howto do that, if i use /etc/vservers/fstab
1161167894 M * daniel_hozac probably won't work quite as automatically.
1161167911 M * renihs i mean, it would work, but i guess i am going to use alot loops in future so it would be better if it would find it itself
1161167932 M * renihs hmm never mind, i will create a loopmount on the host and bindmount that to the guests
1161167946 M * renihs that should work either way
1161172024 J * Hollow ~hollow@i577B6FF6.versanet.de
1161172032 M * Hollow yay
1161172047 A * Hollow is online again :D
1161172163 M * daniel_hozac yay! hey Hollow1
1161172244 M * Hollow feels so good :)
1161172341 M * daniel_hozac i can imagine.
1161172403 M * Hollow hehehe... 1.5MB/s
1161172489 M * FaUl in both directions?
1161172498 M * Hollow no, just downstream
1161172509 A * daniel_hozac has 100/10 Mbps ;)
1161172553 M * FaUl daniel_hozac: thats nice ;-)
1161172576 M * Hollow well, upstream is "just" 600kbps
1161172579 M * daniel_hozac indeed.
1161172661 M * Borg- bah.. who cares about downstream.. upstream thats what is important
1161172796 M * nox and expensive for isp
1161172985 M * renihs bah, me wants internet at home too
1161173464 M * morfoh wb Hollow :D
1161174242 Q * dna_ Ping timeout: 480 seconds
1161174334 J * dna_ ~naucki@196-216-dsl.kielnet.net
1161175007 J * IceGuest_5 ~icechat5@pD9E04105.dip0.t-ipconnect.de
1161175012 M * IceGuest_5 hi
1161175016 N * IceGuest_5 gudi
1161175033 M * gudi anyone can tell me howto acces an usb port from a vserver guest ?
1161175046 J * Piet hiddenserv@tor.noreply.org
1161175052 M * gudi i want to use gnokii to send an sms via usb handy
1161175056 M * trippeh Borg-: Newsflash, downstream is what most users care about ;)
1161175077 M * trippeh I'm on 20Mbps down/1.3Mbps up ADSL
1161175143 A * phedny has 3000/512 ADSL
1161175165 M * trippeh (and no monthly cap of course) :)
1161175293 Q * Hollow Read error: Operation timed out
1161175294 J * Hollow ~hollow@i577B7503.versanet.de
1161175301 A * phedny wonders .. would it be feasible to run an X server on a VPS on a colo server and access it over my ADSL line? :)
1161175327 M * gudi noone ?
1161175350 M * phedny gudi: I think all you need to do is create the appropriate device node in /dev
1161175363 M * phedny gudi: which can only be done from outside the guest, but can afterwards be used inside
1161175380 M * phedny but then, I've not much experience with usb 
1161175390 M * gudi so i have do do a mknod .. /dev/ttyUSB0 
1161175434 M * phedny from the host and then use /etc/vservers/<name>/vdir/dev/ttyUSB0 or something like that
1161175459 M * gudi huh ? why i need to add config ?
1161175504 M * phedny gudi: what do you mean?
1161175533 M * gudi why i nee dto add stuff the etc/vservers ?
1161175545 M * phedny because you cannot use mknod from inside the guest
1161175726 M * gudi hmm seems not to work
1161176002 M * gudi for the mknot command ?
1161176591 M * Borg- trippeh: s/users/lusers/
1161177874 M * Hollow daniel_hozac: btw, congrats for 0.30.211 :)
1161178061 J * debugger Rui@217.129.151.190
1161178067 N * debugger rgl
1161178076 M * rgl daniel_hozac, are you there?
1161178599 M * rgl which APIs should I enable when using 2.0.2.1?
1161178608 M * rgl ALL?  as in Available APIs: legacy,compat,v11,fscompat,v13obs,v13,net,oldproc,olduts
1161178624 M * rgl or leave the default, wich is v13,net ?
1161178734 M * rgl oh well, using legacy it doesn't build at all :(
1161178739 J * id23 ~id@p50813468.dip0.t-ipconnect.de
1161178741 M * rgl any tip, pulease?
1161178747 M * id23 greetings #vserver
1161178763 M * rgl hello
1161179131 M * Hollow rgl: --enable-apis=NOLEGACY is the preffered way
1161179138 M * Hollow ALL will probably not work
1161179150 M * rgl hi Hollow :)
1161179153 M * Hollow hi rgl :)
1161179161 M * Borg- who
1161179166 M * rgl good to see you here :D
1161179171 M * Borg- ups ;) ECONSOLE
1161179173 M * Hollow yeah, good to be online again :)
1161179370 M * rgl :D
1161179392 M * rgl oh, I still get the same error, and have no ideia how to fix this :(
1161179401 M * rgl 'VERIFYCAP' can be executed as root only
1161179401 M * rgl capabilities are not enabled in kernel-setup
1161179403 M * rgl :(((
1161179469 A * harry installing a 64 bit vserver!
1161179507 M * rgl harry, which dist are you using?
1161179536 M * harry debian
1161179594 M * rgl harry, the oficial x86_64 port?
1161179613 M * harry think so
1161179621 M * harry is there another?
1161179627 M * harry it's the debian_amd64
1161179655 M * rgl if you are getting it from debian.org its the "final" one
1161179660 M * rgl I think ;)
1161179669 M * harry i get it from a mirror
1161179676 M * harry ftp.belnet.be/debian_amd64 iirc
1161179689 M * harry deb http://ftp.belnet.be/debian-amd64/debian/ stable main
1161179700 M * rgl tell me how it went, cause I also want to install it :D
1161179710 M * harry it sucks
1161179712 M * rgl (I don't have the box yet, thou ...)
1161179728 M * rgl why?  too old?
1161179738 M * harry it installs debsig-something
1161179745 N * Bertl_zZ Bertl
1161179748 M * harry which makes it check packages signing
1161179749 M * Bertl morning folks!
1161179751 M * harry which... doesn't work
1161179757 M * harry so it doesn't want to install shit
1161179761 M * rgl dunno about that :(
1161179766 M * phedny morning Bertl 
1161179792 M * rgl hey Herbert
1161179800 M * harry so you just have to remoe 
1161179807 M * harry remove that one
1161179809 M * harry hey Bertl 
1161179825 M * Hollow heya Bertl
1161179836 M * Bertl hey Hollow! back for good?
1161179840 M * Hollow indeed :)
1161179843 Q * rob-84x^ Quit: That's it for today
1161179863 J * rob-84x^ rob@submarine.ath.cx
1161179864 M * Bertl Hollow: we already missed you desperately ...
1161179866 M * rgl Hollow, how is vcc? :D
1161179876 M * Hollow Bertl: yeah, i guessed so ;)
1161179908 M * Bertl - nobody pushing the kernel interfaces
1161179908 M * rgl Hollow, can I already use it instead of util-vserver? (that I can't get to work :()
1161179918 M * Hollow rgl: well, i did not code the last 5 weeks... ain't no fun without inet ;)
1161179930 M * matti Bertl: :))
1161179931 M * matti Hollow: :))
1161179937 M * Bertl hey matti!
1161179940 M * harry ld: BFD 2.15 assertion fail ../../bfd/linker.c:619
1161179942 M * matti What's up?
1161179944 M * harry how bad is that?
1161179948 M * matti harry: :)
1161179960 M * Hollow Bertl: which interfaces would you like to be pushed? ;)
1161179973 M * matti Eh.
1161180001 M * Hollow rgl: well, it's not production-ready yet, and i'm a bit out of date what's working and what's not... have to catch up 5 weeks..
1161180070 M * Hollow can someone tell me if http://home.xnull.de is reachable from outside?
1161180090 M * daniel_hozac Hollow: thanks.
1161180095 M * daniel_hozac Hollow: yeah, works here.
1161180098 M * Hollow good :)
1161180149 M * daniel_hozac rgl: as i already told you, that means your uid isn't 0 and thus you're not root.
1161180170 M * Bertl Hollow: all of them, of course :)
1161180200 M * Bertl Hollow: seems to work here too :)
1161180232 M * Hollow Bertl: i guess you're talking about the stat interfaces mainly? or just the general legacy pushing? :P
1161180242 M * Hollow *punting
1161180243 M * Hollow ;)
1161180346 Q * Greek0 Ping timeout: 480 seconds
1161180558 M * Bertl Hollow: yeah, both ...
1161180592 M * Hollow Bertl: ok.. so... let's start with the missing stat interfaces then :)
1161180594 M * Hollow ETA?
1161181161 M * daniel_hozac Bertl: about the stricter pid_task, was there anything in particular you wanted me to test? seems to break vkill for instance.
1161181368 M * rgl daniel_hozac, but it is root
1161181391 M * daniel_hozac rgl: so why doesn't getuid() return 0? :)
1161181407 M * rgl daniel_hozac, I have no ideia
1161181419 M * rgl but "id" saus I'm root
1161181428 M * daniel_hozac strace might help you there.
1161181447 M * rgl and doing stuff that only root cat do, I can.
1161181486 M * rgl daniel_hozac, funny is, if I strace vserver, it never returns!
1161181500 M * daniel_hozac _never_ strace vserver.
1161181506 M * daniel_hozac bad things happen :)
1161181511 M * rgl oh! hehe
1161181524 M * rgl so what I have to strace?
1161181532 M * daniel_hozac vserver-info 0 VERIFYCAP?
1161181604 M * rgl daniel_hozac, getuid()                                = -1 ENOSYS (Function not implemented)  ?!
1161181638 M * rgl does it have something to do by not having uid16 syscalls?
1161181647 M * daniel_hozac rgl: uh, well... i guess that explains it.
1161181764 M * Bertl daniel_hozac: ah, okay, but the version I uploaded yesterday should be fine?
1161181878 M * daniel_hozac Bertl: why was the PIDTYPE_PID check moved into the if? do we want to hide other pid 1s?
1161181946 M * Bertl I just felt like PID type info would be sufficient for now
1161182014 M * Bertl Hollow: what data are you missing with rc41?
1161182205 J * rf ~rf@ro.risc.uni-linz.ac.at
1161182212 Q * gudi Quit: Friends help you move. Real friends help you move bodies.
1161182213 M * Bertl welcome rf!
1161182233 M * Hollow Bertl: well, tbh, i didn't look at the new patches yet, but will do so shortly
1161182243 M * Bertl np *G*
1161182295 P * rf 
1161182302 M * Bertl Hollow: btw, I would appreciate a few minor changes to the wiki (if you find the time)
1161182331 M * rgl daniel_hozac, that was it, vserver-util for some reason was using syscalls that are using uid of 16 bit
1161182338 M * rgl err util-vserver
1161182339 M * Hollow Bertl: sure, will have to do some updates anyway, since i got some emails about migrated pages etc...
1161182371 M * Bertl yep, I also migrated the VServer_Users page, for example
1161182396 M * Bertl (which should get a link from the menu)
1161182397 M * daniel_hozac rgl: what arch is that?
1161182405 M * daniel_hozac Bertl: can't you edit the menu?
1161182408 M * rgl daniel_hozac, nor x86
1161182413 M * rgl normal x86 that is
1161182426 M * Bertl daniel_hozac: hmm, haven't tried, how would I do that?
1161182464 M * Hollow Bertl: http://linux-vserver.org/MediaWiki:Sidebar
1161182497 M * Bertl ah, great, tx!
1161182500 M * daniel_hozac rgl: kernel?
1161182505 M * Hollow format is URL|Title
1161182513 M * rgl daniel_hozac, 2.6.17.13
1161182528 M * Bertl Hollow: hmm, no edit option there :(
1161182547 M * Hollow sec..
1161182607 M * Hollow Bertl: ok, you should be in the sysops group now
1161182644 M * daniel_hozac rgl: ah, seems getuid is indeed the 16-bit, getuid32 is the 32-bit one.
1161182799 M * Bertl Hollow: okay, works
1161182859 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1161182951 M * Bertl wb cdrx!
1161183003 M * cdrx he bertl !
1161183218 M * Hollow Bertl: ah, the VCMD_virt_stat was missing the last time i checked, but looks ok to me now, will test it asap
1161183733 M * harry ld: BFD 2.15 assertion fail ../../bfd/linker.c:619
1161183737 M * harry what does that mean?
1161183935 M * Bertl that your linker failed some assertion
1161183945 M * Bertl (i.e. an internal compiler/linker error)
1161183985 M * harry is it bad?
1161183998 M * Bertl depends .. do you need the linked result?
1161183999 M * harry the kernel works... but i don't know what the consequences are..  :S
1161184010 M * harry probably , why else would it link it?
1161184027 M * Bertl well, then it is probably bad ...
1161184046 M * harry and how can i fix it?
1161184056 M * Bertl upgrade binutil/gcc
1161184079 Q * id23 Read error: Connection reset by peer
1161184080 M * harry i'm up to date, it's debian stable :S
1161184087 M * Bertl lol
1161184112 M * harry grmbl... no dietlibc for x86_64 :S
1161184174 M * harry hahaa... but there is a dietlibc-dev :)
1161184241 M * matti Eh.
1161184264 M * rgl harry, have you consideres using ubuntu?
1161184269 J * id23 ~id@p50813468.dip0.t-ipconnect.de
1161184277 M * harry no
1161184305 M * Bertl welcome id23!
1161184372 M * renihs ubuntu...pff, try gentoo
1161184384 M * daniel_hozac yay, distro war! try Fedora!
1161184398 M * Bertl nah! mandriva! :)
1161184437 M * Bertl ah, maybe even suse, now they dropped reiserfs *evil-grin*
1161184455 M * harry use openbsd!
1161184465 M * harry (moehaha... try to beat that in absurdiness!)
1161184483 M * Bertl RMS-Linux/Hurd?
1161184499 M * harry hm... getting close :)
1161184510 M * harry maybe minix? ;)
1161184605 M * harry where is daniel_hozac 's patch for my vserver-functions problem?
1161184607 J * Hollow_ ~hollow@hollow.dialup.fu-berlin.de
1161184628 M * Hollow_ damn
1161184637 M * Hollow_ my vpn connection kicked off everything else
1161184655 M * daniel_hozac harry: http://people.linux-vserver.at/~dhozac/p/uv/experimental/ns-cleanup.patch
1161184685 M * daniel_hozac which reminds me, Hollow_: didn't we discuss putting people on the FTP? what happened to that?
1161184689 M * harry tnx
1161184710 M * Hollow_ daniel_hozac: well, i guess a symlink would do?
1161184731 N * Hollow Guest1898
1161184742 M * daniel_hozac Hollow_: from the FTP space? that doesn't work, we'll have to do it the other way.
1161184760 M * daniel_hozac or reconfigure the ftpd to follow symlinks.
1161184790 Q * Guest1898 Ping timeout: 480 seconds
1161184804 N * Hollow_ Hollow
1161184836 M * Hollow daniel_hozac: ok, will take a look
1161184894 M * id23 Hi Bertl
1161184944 M * harry tnx
1161184977 M * Hollow arr, this cisco vpn client fucks up everything
1161184979 Q * Hollow Remote host closed the connection
1161185002 J * Hollow ~hollow@i577B7503.versanet.de
1161185017 M * Hollow proprietary shit
1161185040 M * daniel_hozac Hollow: hmm, maybe it did work... i can't remember what i did to test it (i.e. if i actually did test it properly).
1161185087 M * daniel_hozac Hollow: hehe, isn't there an open source alternative? vpnc or something like that?
1161185088 M * Hollow no it doesn't, i just tested the symlink approach
1161185092 M * daniel_hozac ok.
1161185124 M * Hollow yes, there is... but it does not work with our vpn, in fact the original cisco software doesn't either... they provide their own modified version of the cisco software
1161185138 M * daniel_hozac oh, lol.
1161185164 M * Hollow but maybe there are some config options, so it does not completely disable my local network, will have to investigate ;)
1161185265 J * debugger Rui@217.129.151.190
1161185296 M * Hollow daniel_hozac: well, we could use bind mounts for the ftpd, but sounds like overkill
1161185542 M * Hollow ok, testing vpn again, brb
1161185545 Q * Hollow Remote host closed the connection
1161185731 Q * rgl Ping timeout: 480 seconds
1161185808 J * Hollow ~hollow@i577B7503.versanet.de
1161185816 M * Hollow :(
1161185834 M * daniel_hozac Hollow: yeah, that was my opinion too.
1161185837 J * stefani ~stefani@tsipoor.banerian.org
1161185849 M * daniel_hozac Hollow: what about symlinks from public_html to the FTP space?
1161185861 M * daniel_hozac alternatively reconfiguring httpd to just use the FTP space directly.
1161185883 M * Hollow the latter one sounds ok to me
1161185896 M * Hollow maybe a combination of both
1161185900 M * Hollow for convenience
1161185914 M * daniel_hozac yeah
1161185940 M * Bertl hola stefani!
1161185948 M * Bertl okay, off for dinner now ... back shortly
1161185953 N * Bertl Bertl_oO
1161186002 M * matti harry: Minix 3 is not so bad ;]
1161186409 J * hardwire` ~hardwire@89-208-58-66.gci.net
1161186540 M * Hollow btw, 69.94.66.94 floods our sshd with login try&failures
1161186695 M * doener Hollow: heh, sounds like our vpn at university ;)
1161186711 Q * hardwire Ping timeout: 480 seconds
1161186724 M * Hollow doener: any hints?
1161186765 M * doener I could give you the jid of someone who has it running (I still didn't setup my laptop ;)
1161186796 Q * bronson Ping timeout: 480 seconds
1161186802 M * Hollow yeah, that would be nice
1161187014 J * mkhl ~mkhl@200-153-153-189.dsl.telesp.net.br
1161187017 P * mkhl 
1161187524 J * bonbons ~bonbons@83.222.36.111
1161187755 Q * id23 Ping timeout: 480 seconds
1161187935 N * Bertl_oO Bertl
1161187940 M * Bertl back now ...
1161187954 M * _are_ Hi Bertl
1161187979 M * _are_ is some linux-vserver presentation at linuxworld expo?
1161188246 M * Bertl yep
1161188288 M * Bertl http://www.linuxworldexpo.de/vkf_06.php?ID=122
1161188518 M * eyck how do you find time for coding in between those linuxtags/expos and beerfests?
1161188553 M * _are_ Bertl: any booth as well?
1161188591 M * Bertl _are_: nope, just a short visit
1161188611 M * _are_ k, ordered a ticket and will try and find the time to take a visit
1161188651 N * debugger rgl
1161188655 M * _are_ anyway, have to run
1161188686 M * rgl what is the best way to boostrap a CentOS install?  should I use http://rpmstrap.pimpscript.net/ ?
1161189087 M * Bertl latest tools should support centos network install
1161189092 M * Bertl (again)
1161189130 M * Bertl daniel_hozac: btw, have you thought about rpmstrap yet? what's your opinion?
1161189261 Q * _are_ Ping timeout: 480 seconds
1161189394 J * lilalinux ~plasma@dslb-084-058-206-218.pools.arcor-ip.net
1161190703 J * bronson ~bronson@66.160.177.202
1161190897 M * meandtheshell Bertl: do you know a website where all the help messages taken from all Kconfig files in the kernel source tree can be found - of course in their tree representation as a user might recognize them during make menuconfig if he invokes the help
1161190926 M * Bertl hmm, nope, but you can easily create that yourself
1161190927 J * shedi ~siggi@inferno.lhi.is
1161190958 M * meandtheshell Bertl: that means hacking  - how would you do that?
1161190979 M * Bertl meandtheshell: the data is in the Kconfigs
1161190985 A * meandtheshell is pretty sure there's a URL readily made somewhere
1161191012 M * Bertl well, I'd suggest to google for a config option with some words of the help text then :)
1161191020 M * meandtheshell Bertl: :)
1161191285 J * marcfiu ~mef@targe.CS.Princeton.EDU
1161191297 M * marcfiu hello
1161191309 M * rgl Bertl, humm, indeed there is a centos 4 dist in there.  but where are the docs?  do I have to dig into the sources
1161191715 J * Rich_Estill ~restill@c-24-11-195-139.hsd1.mi.comcast.net
1161191863 J * shuri ~shuri@64.235.209.226
1161191895 M * Hollow daniel_hozac: ftp://ftp.linux-vserver.org/pub/people
1161192010 M * rgl is the oldwiki down?
1161192068 M * Hollow it seems like DNS is f*cked up (again?)
1161192068 M * Bertl rgl: probably jacques is playing with the host .. sec
1161192081 M * Bertl Hollow: no, we switched to the original wiki
1161192088 M * Hollow ?
1161192090 M * Bertl Hollow: the copy was unfixable :)
1161192101 M * Bertl wrong history, wrong contents .. 
1161192120 M * Hollow ah ok... i thought you replaced the new wiki with the old one again
1161192122 M * Hollow :)
1161192156 Q * derjohn2 Ping timeout: 480 seconds
1161192157 J * derjohn2 ~aj@dslb-084-059-028-251.pools.arcor-ip.net
1161192160 M * Hollow yeah, the whole encoding/collation whatever crap in mysql sucks
1161192345 M * Bertl rgl: lets give the old wiki an hour or so ... if it still fails then we will have to reactivate the new old wiki :)
1161192362 M * rgl Bertl, oki
1161192374 M * rgl Bertl, and about the other question? :)
1161192408 M * Bertl should be straight forward ... vserver - build help
1161192428 M * rgl Bertl, ah thx :)
1161192445 M * Bertl vserver - build --help
1161192448 M * Bertl actually
1161192466 J * debugger Rui@217.129.151.190
1161192471 M * Bertl centos should be yum, IIRC, so that's what you want
1161192483 M * Bertl add the --interface and --context values and that's it
1161192517 J * fosco fosco@213.161.192.244
1161192521 N * debugger rgl_
1161192533 M * Bertl welcome fosco!
1161192596 M * rgl_ Bertl, do you have an example.  the --help is not enough for me to understand :(
1161192677 M * Bertl vserver karli build -m yum --context 42 --hostname karli.test.org --interface 192.168.42.1/24 --initstyle sysv -- -d centos4
1161192753 M * rgl_ Bertl, why do I need the context?
1161192770 M * Bertl because static contexts are the future!
1161192931 Q * rgl Ping timeout: 480 seconds
1161192932 M * rgl_ Bertl, ah, so I should always build a guest with a static context ok :D
1161192957 M * Bertl until the tools assign static contexts at build time, yes :)
1161192973 M * Bertl (automatically that is)
1161193168 M * rgl_ ok thx :)
1161193231 M * rgl_ Bertl, is there a way to see which packages will be installed?
1161193246 M * daniel_hozac it should tell you while it's doing it.
1161193272 M * rgl_ can I tell it to install the packages from the cdrom instead of the network?
1161193302 M * daniel_hozac you'd have to modify the .repo's then.
1161193318 M * rgl_ inside /usr/lib/util-vserver/distributions/centos4 correct?
1161193335 M * daniel_hozac you ought to copy the yum.repos.d to /etc/vservers/.distributions/centos4 first.
1161193387 M * rgl_ hummm inside that directory there is no centos4, so I just copy the entire tree?
1161193424 Q * lilalinux Remote host closed the connection
1161193446 N * rgl_ rgl
1161193503 M * rgl oh, and the redhat too, cause centos has some symlinks inside it :|
1161193732 M * rgl daniel_hozac, the centos4.4 is a bit diferent than the one that is shipped with util-vserver, maybe you should update it
1161194075 M * daniel_hozac rgl: hmm?
1161194099 M * rgl daniel_hozac, ops, the centos 4.4 .repo file
1161194116 M * daniel_hozac you should copy the yum.repos.d to /etc/vservers/.distributions/centos4/yum.repos.d, creating directories if necessary.
1161194132 M * daniel_hozac rgl: what CentOS 4.4 .repo file?
1161194146 M * daniel_hozac there is just a CentOS 4 .repo file in util-vserver.
1161194156 M * rgl daniel_hozac, its not only the yum.repos.d directory that is missing, there is no /etc/vservers/.distributions/centos4/ at all.
1161194163 M * daniel_hozac right.
1161194172 M * daniel_hozac but you only want to change the repo config, no?
1161194204 M * rgl daniel_hozac, humm, yes.   the files I add into /etc will override the ones in /usr/lib/?
1161194264 M * daniel_hozac yes, but on a file/directory by file/directory basis.
1161194343 M * rgl I see. it will first look into /etc, then in /usr/lib/. ok :D  thx
1161194669 M * rgl is there a way to tell vserver build to where it should save the rpms?   and to where install centos?
1161194732 M * Bertl well, it installs the rpms into the guest :)
1161194760 M * rgl it doesn't get them from the repo?
1161194772 M * rgl (and saves them on some temporary directory somewhere)
1161194800 M * Bertl it probably saves them somewhere temporarily, but it definitely isntalls it into the guest
1161194811 M * Bertl (details from the source or daniel_hozac :)
1161194885 M * rgl Bertl, yes, it should, but how can I change that location? :)
1161194902 M * Bertl it#s the guest, why do you want to change it?
1161194940 M * Bertl you can change where the guest will be located with --rootdir
1161194990 J * yang ~yang@cpe-213-157-253-172.dynamic.amis.net
1161194997 M * Bertl welcome yang!
1161195019 M * yang hello Bertl ! If you will need access to Indigo, just tell me
1161195039 M * yang I am re-locating it
1161195060 M * Bertl okay, just send me an email how to reach it (or you :)
1161195180 J * id23 ~id@p50813468.dip0.t-ipconnect.de
1161195206 M * Bertl wb id23!
1161195535 M * rgl Bertl, I just want to "make" the guest install inside a dedicated directory/partition and not inside /usr/lib/... :D
1161195589 M * rgl Bertl,  --rootdir sounds exactly what I want :D
1161195631 M * Hollow Bertl: you still need to tell me which software/script you need for your people.l-v.org dir, so you could finally move your stuff..
1161195639 M * yang Bertl: did you manage to compile the sgi kernel on x86 machine?
1161195811 Q * weeble Ping timeout: 480 seconds
1161196043 J * weeble ~weeble@81.52.144.1
1161196360 M * Bertl rgl: be aware, that the guest config will _assume_ that the guest will stay there
1161196401 M * rgl Bertl, where is "there"? hehe
1161196403 M * Bertl yang: I can do that, but I didn't do it yet
1161196423 M * rgl Bertl, in --rootdir /path/to/my/guest/root ?
1161196429 M * Bertl yep
1161196438 M * Bertl btw, the default is /vservers
1161196442 M * rgl an yes, it will :D
1161196496 M * rgl maybe there is another way of doing this, but I want to mount a partition, maybe I can just have it already mounted on /vservers/ before calling vserver ... build ... ?
1161196512 M * daniel_hozac yep, that should work fine.
1161196615 M * Bertl Hollow: hmm, wouldn't it be simpler to mirror vserver.13thfloor.at?
1161196832 M * rgl daniel_hozac, cool :)
1161196872 M * rgl daniel_hozac, btw, is there a way to get yum showing the download progress stuff when we run vserver ... build?
1161196921 M * rgl doener, I mean, it just said which packages it will install, but no progress is displayed to the user.
1161197047 M * daniel_hozac YUM="yum -d2" vserver .... should do it, i think.
1161197091 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1161197099 M * Bertl wb Blissex!
1161197117 M * Hollow Bertl: sure, i could just mirror it... but http is not the best choice to mirror something IMO... is there ftp access to /Experimental and /Stuff?
1161197146 M * Blissex hi!
1161197168 M * Blissex Hollow: HTTP and FTP nowadays are vastly equivalent. RSYNC would be much better.
1161197209 M * Hollow Blissex: well, i don't want to follow the links generated by the directory listing to get all the files, this will most likely result in chaos
1161197249 M * Hollow especially i don't like to tinker around...
1161197272 M * rgl daniel_hozac, ah cool.  I'll try that on my next test :)
1161197282 M * Blissex Hollow: well, all HTTP mirroring program I know have a ''just under this, not above or laterally'' option...
1161197343 M * Hollow but i don't want to mirror things like http://foo/?C=M&O=D and friends e.g.
1161197375 M * Blissex Hollow: yes that requires a bit too much config...
1161197389 M * Blissex so Bertl, what about RSYNC access for Hollow?
1161197392 M * Hollow also would a move (in contrary to mirror) be a good point to remove old cruft
1161197434 M * Bertl Hollow: 'old cruft' gets removed on a regular basis
1161197448 M * Hollow holy sh...
1161197449 M * Hollow :)
1161197480 M * Bertl will see, maybe rsync is an option, maybe ftp too ... for now I have other stuff to do
1161197514 M * Bertl so maybe the Q&D solution for you could be to redirect from that location to vserver.13thfloor.at
1161197517 M * Hollow ok, that was my primary question anyway
1161197530 M * Hollow nah, no tinkering...
1161197800 M * rgl daniel_hozac, humm, the vserver...build does no show a "success" kind of message?
1161197891 Q * Piet Remote host closed the connection
1161197964 M * rgl daniel_hozac, here is the log http://paste.linux-vserver.org/507  maybe something is not right?
1161198038 M * daniel_hozac looks fine to me.
1161198057 J * Piet hiddenserv@tor.noreply.org
1161198135 J * kevinp ~kevinp@ny.webpipe.net
1161198157 M * rgl daniel_hozac, cool thx :)
1161198175 M * kevinp Looks like oldwiki.linux-vserver.org is down :(
1161198187 M * Hollow indeed
1161198200 A * Hollow is innocent
1161198220 M * Bertl yep, probably jacques didn't manage to pay the bills for the host too
1161198267 M * kevinp any discussion on setting up mirrors for the site yet?
1161198274 M * Bertl Hollow: so shall we switch to the broken copy of the old wiki?
1161198346 M * Hollow there has to be a way to get a working copy .. but i'm not sure how, i could give it another try if you send me the sqldump again
1161198380 M * Bertl no chance, as the host is down
1161198405 M * Hollow too bad, i don't have a copy anymore...
1161198420 M * kevinp I don't know how much bandwidth the site uses, but I might be able to provide some hosting, whether it is a mirror or the actual site
1161198428 M * Hollow well.. neither one is a good solution imo
1161198457 M * Bertl Hollow: what did you do with the last three copies I sent you=
1161198501 M * Hollow the last you sent me was when we did the switch iirc
1161198509 M * Bertl yup, so?
1161198515 M * Bertl I can send it to you again?
1161198516 M * Hollow i don't have it anymore..
1161198521 M * Hollow rm
1161198581 M * Bertl okay, you should get it in your email
1161198592 M * Bertl do you need the tavi source/changes/config too?
1161198595 M * Hollow ok, good
1161198601 M * Hollow i don't think so..
1161198615 J * _are_ ~are@62.112.159.81
1161198626 M * Hollow the borkage is definitely in mysql
1161198635 M * Bertl well, why not get an older version?
1161198655 M * Bertl (where you simply import the db without changing it=
1161198667 M * Bertl s/=/)
1161198670 M * Hollow that would imply a conversion of mediawiki then i guess
1161198676 M * Hollow Bad Thing
1161198677 M * Bertl huh?
1161198690 M * Hollow i can't run multiple mysql servers..?
1161198693 M * Bertl never heard of Linux-VServer before?
1161198709 M * Hollow yeah ok... i could just run another vserver for the db
1161198757 M * Hollow will have to setup an internal network then...
1161198790 M * Hollow tinkering again
1161198813 M * Hollow i guess i will try to convert the sqldump to 4.1 ;)
1161198840 M * Bertl make it so :)
1161198876 M * Bertl I'll switch the dns now ...
1161198893 M * kevinp any idea on how close 2.1.x is to going stable?
1161198900 M * rgl humm, the package management is done on the host by default, how can I make it inside the guest instead?
1161198906 M * Bertl kevinp: very close
1161198912 M * kevinp excellent
1161198921 M * rgl cool :D
1161198924 M * Bertl rgl: externalize/internalize
1161198941 M * rgl Bertl, howto? :D  /me doesn't have acess to the wiki :(
1161198955 M * marcfiu Has anyone used the CoW-links support for anything other than to reduce the disk footprint of a guest's filesystem?
1161198966 M * Bertl rgl: vserver --help
1161198981 M * rgl Bertl, ah great :)
1161199029 M * daniel_hozac marcfiu: such as?
1161199029 M * marcfiu e.g., in coordination with something like rdiff-backup to reduce the footprint between incremental and full backups?
1161199053 M * Bertl doesn't rdiff-backup do hard links by default?
1161199089 M * marcfiu bertl: yes it does... though I am not sure whether it does so between multiple hosts.
1161199118 M * daniel_hozac how would COW-links help?
1161199122 M * Bertl Hollow: ok, dns entry is changed
1161199184 M * marcfiu daniel_hozac: suppose you had, say, 100 clients and a single backup server.  And the 100 clients each use rdiff-backup to dump stuff from their fs to the servers /backs/machine-id/ filesystem.
1161199233 M * marcfiu daniel_hozac: using CoW-links one could now "unify" the amount of space that is used between the various /backups/machine-id/ without having to modify the rdiff-backup server directly.
1161199244 M * daniel_hozac marcfiu: ok, but how do COW-links help? wouldn't regular hardlinks work fine?
1161199281 M * marcfiu daniel_hozac: no... I don't think so.
1161199284 M * marcfiu here's the example:
1161199356 M * marcfiu suppose both machine A and B had the same version of /bin/bash
1161199385 M * marcfiu now machine A updates /bin/bash with a new version
1161199432 M * marcfiu you wouldn't want the backup of machine B's /bin/bash to be modified.
1161199452 M * daniel_hozac rdiff-backup doesn't break hardlinks?
1161199463 A * daniel_hozac only uses rsync for backups.
1161199573 M * marcfiu I think you'd have to modify the rdiff-backup server to be aware of hardlinks that span across directories that belong to different machines.
1161199604 M * marcfiu If you modify rdiff-backup to keep track of this, then I suppose it could do the right thing just with hardlinks.
1161199666 J * debugger Rui@217.129.151.190
1161199674 M * marcfiu With CoW-links it seems that there wouldn't be a need to modify rdiff-backup.
1161199708 M * daniel_hozac are you sure rdiff-backup doesn't work as is?
1161199798 M * marcfiu AFAIK, rdiff-backup works just fine for incremental and full backups of a directory to another directory.
1161199835 M * marcfiu however, suppose I had 100 relatively identical machines each with a 500MB FS.
1161199859 M * marcfiu A full backup of these 100 machines to a single server would require 50GB of disk space.
1161199890 M * marcfiu This is because rdiff-backup wont hardlink identical files across the directories dedicated to the 100 machines.
1161199928 M * marcfiu E.g., there would be a /backup/machine-1, /backup/machine-2, and so on.  rdiff-backup wont unify files between the various machine-X directories.
1161199982 M * marcfiu But that's precisely what CoW-links was designed for: unifying files from different "guests" in an application-oblivious manner.
1161200018 M * marcfiu any way, was just one idea of how to use CoW-links in a way that was slightly different from its normal use.
1161200025 M * daniel_hozac CoW-links on their own doesn't help. you'll still have to unify them.
1161200039 M * marcfiu Right... a combo of CoW-links and unification.
1161200039 M * daniel_hozac and even then, i'm not sure you need CoW.
1161200113 M * kevinp Bertl: Let me know if you ever want to start mirroring the site, I've got several mirrors that I just rsync whenever they want.  Distributing the bandwidth should help with the bills.
1161200129 M * marcfiu daniel_hozac: ok... worth for me to look into this.
1161200136 Q * rgl Ping timeout: 480 seconds
1161200371 M * marcfiu bertl: is the stuff you want to distribute over different servers just stuff that one can access via http?  If so, why not use one of the popular CDNs that operate on PlanetLab?
1161200639 N * debugger rgl
1161200643 M * Bertl marcfiu: actually I do not want to distribute anything atm :)
1161200679 M * Bertl marcfiu: it just seems that jacques has serious troubles and/or is already gone
1161200721 M * Bertl marcfiu, kevinp: but thanks for the offer
1161200853 M * Hollow Bertl: btw, do you have any bandwidth limits on vserver.13thfloor.at?
1161200982 M * Bertl not that I'd know of, why?
1161201012 M * Hollow i only get ~100KB/s
1161201028 M * Hollow in fact it varies between 50-100
1161201256 M * Bertl well, from US that should be fine, I guess
1161201262 M * Hollow <- Berlin
1161201271 M * Hollow but it wasn't any better in munich too
1161201279 M * Bertl yeah, but vserver.13thfloor.at is in us
1161201302 M * Hollow oh..
1161201354 M * kevinp I got 126.91K/s downloading the latest patch and I'm in the US
1161201402 M * kevinp 109.08K/s downloading the utils from ftp.linux-vserver.org
1161201446 M * Hollow ftp.* is in germany
1161201536 M * Rich_Estill I am in US and got 400+ KB  It took a second or two to download the patch.
1161201582 M * Bertl so I'd say all observed rates are quite fine :)
1161201668 J * dna___ ~naucki@196-216-dsl.kielnet.net
1161201702 M * Hollow i don't consider 100KB/s fine, but ok :)
1161201725 M * Hollow if it's from the us it is fine actually
1161201781 M * Bertl so a megabit line isn't enough for you? eh?
1161201850 M * Hollow everything below 500KB/s is considered slow :)
1161201941 M * Hollow i was used to 6MBit line, and since today i'll get used to a 20MBit line, so.. yep, 1MBit is not enough
1161201942 M * Hollow ;)
1161201975 J * Greek0 ~greek0@85.255.145.201
1161201980 M * Bertl well, I don't want to comment on that ... :)
1161201987 M * Bertl wb Greek0!
1161202095 Q * dna_ Ping timeout: 480 seconds
1161202161 Q * bronson Quit: Ex-Chat
1161202320 J * bronson ~bronson@66.160.177.202
1161202322 Q * bronson 
1161202395 J * DreamerC_ ~dreamerc@61-217-226-241.dynamic.hinet.net
1161202419 M * Hollow Bertl: while configuring rc41 i still see the Kconfig options for legacy have not changed yet.. didn't we agree that we should either say "Disable ..." or "Enable ..."?
1161202472 M * Bertl did we? maybe .. patch?
1161202490 M * Hollow hm, there was one... iirc doener did it?
1161202492 Q * dna___ Quit: Verlassend
1161202583 M * Bertl url?
1161202586 M * Hollow Bertl: ah no.. daniel_hozac did.. http://people.linux-vserver.org/~dhozac/p/k/delta-config-clean01.diff
1161202689 M * Bertl okay, looks fine to me, will add it ...
1161202694 M * Hollow great :)
1161202721 M * Bertl ah, btw, daniel_hozac: could you check the includes for rc41 too please?
1161202735 Q * DreamerC Read error: Operation timed out
1161202929 M * daniel_hozac hmm, didn't we decide that patch was broken?
1161202949 M * Bertl ah, did we? I simply can't remember .. let's check it once again
1161203010 M * daniel_hozac CONFIG_VSERVER_NGNET needs to be default y
1161203043 M * Bertl ah, right
1161203073 M * Bertl so, it wasn't tested back then?
1161203086 M * daniel_hozac well, i found that when i tested it :)
1161203116 M * Bertl so we basially change the last line and that is fine then, yep?
1161203140 M * daniel_hozac i'd say so.
1161203151 M * Bertl hehe, so was that tested or not? :)
1161203167 M * daniel_hozac i can't remember exactly, but i'd assume so :)
1161203170 M * Hollow would compile tested be enough?
1161203194 M * daniel_hozac diff'ing .configs should be enough, no?
1161203196 M * Bertl yes, I think so, you should compare before and after with similar options
1161203214 M * Hollow ok, i'm compiling now anyway, just a sec
1161203221 M * Bertl okay, tx!
1161203227 M * kevinp slightly OT: it might be nice to include the testme.sh script in with the utils for easy access
1161203250 M * kevinp I always have to hunt for the URL to get it, maybe include it on the downloads page?
1161203285 M * kevinp Do you want results from rc41 that I just compiled?
1161203311 M * Bertl of course, paste.linux-vserver.org please
1161203318 Q * FireEgl Read error: Connection reset by peer
1161203325 M * kevinp which parameter?
1161203381 M * Bertl with -L would be nice
1161203426 M * kevinp nice to see we have our own pastebin: http://paste.linux-vserver.org/509
1161203432 M * Hollow ok, compilation works, new config looks like http://paste.linux-vserver.org/508
1161203544 M * Hollow reboot, brb
1161203548 Q * Hollow Quit: Konversation terminated!
1161203570 M * Bertl hmm, a 20Mbit line but no irc proxy/server :)
1161203589 M * daniel_hozac hehe.
1161203678 M * daniel_hozac Bertl: no unneeded/missing headers in 2.1.1-rc41 according to the script.
1161203682 J * Hollow ~hollow@i577B7503.versanet.de
1161203691 M * daniel_hozac (well, some, but only false positives :))
1161203708 M * Bertl okay, great!
1161204028 M * Hollow Bertl: i get a bunch of failed tests with rc41: http://paste.linux-vserver.org/510
1161204065 M * Bertl 116 is actually okay (i.e. false positive)
1161204069 M * Hollow ok..
1161204081 M * Bertl 031 is not, can you redo that one with -vv
1161204102 M * Bertl 34,35,37 depend on your settings
1161204112 M * Hollow  /usr/sbin/vuname: invalid option -- -
1161204131 M * Bertl huh?
1161204131 M * Hollow for 031
1161204149 M * Hollow http://paste.linux-vserver.org/511
1161204166 M * Bertl so some tool incompatibility?
1161204171 M * Hollow hm..
1161204173 Q * meandtheshell Quit: exit (0);
1161204176 M * Hollow let's reinstall util-vserver
1161204229 M * Bertl could be that it would prefer a -- now before the uname -a
1161204247 M * Bertl could you check chcontext --xid 49151 --hostname zaphod.10884 uname -a  vs
1161204251 M * Bertl chcontext --xid 49151 --hostname zaphod.10884 -- uname -a 
1161204301 M * Hollow will do, but most likely one of my packages overwrote vuname...
1161204326 M * Hollow still not uptodate with all packages yet
1161204393 M * Hollow yep, that did the trick... 031 works
1161204417 M * Hollow also the testfs tests beside 116 work now...
1161204438 M * Hollow (too much chaos :)
1161204671 Q * Blissex Remote host closed the connection
1161204694 J * Aiken ~james@tooax8-128.dialup.optusnet.com.au
1161204902 M * Bertl morning Aiken!
1161204906 M * kevinp what is the new ip address for oldwiki? I want to add it to my hosts file so I can access it now...
1161204925 M * kevinp or is it not up yet at the new ip address either?
1161204947 M * Bertl dig oldwiki.linux-vserver.org @ns1.linux-vserver.org
1161204967 P * Rich_Estill Leaving
1161205104 M * rgl is there a wiki page that talks about the difference between 2.0.x and 2.1.x?
1161205107 M * kevinp that worked, thanks!
1161205262 M * Bertl rgl: http://oldwiki.linux-vserver.org/Release+FAQ
1161205291 M * Bertl needs to be updated, but what's there is correct
1161205322 M * rgl ah cool :)
1161205483 M * fs Bertl: is there something yet for vs2.0.2.2 I can test regarding the X11 mouse issue?
1161205491 M * Bertl sure
1161205511 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.1-vs2.0.2.2-rc2.diff
1161205531 M * Bertl IIRC, I 'pinged' you yesterday (regarding this)
1161205535 M * fs oh great
1161205561 M * fs I guess it got lost in the trouble...
1161205567 M * fs I'll fire up the compiler :)
1161205577 M * Bertl k, let us know
1161205771 M * rgl Bertl, how hard was it to port from 2.6.17.13 to 2.6.18.1?
1161205780 M * rgl I mean, port the patches
1161205806 M * Bertl well, the port was not too hard, but the 'changes' were intrusive
1161205888 M * rgl you have to maintain a patch for each kernel?
1161205914 M * Bertl yep, 2.6.17.3 is quite different
1161205938 M * rgl thats sounds like a full time job hehe
1161205958 M * Bertl yep, it almost is :)
1161206026 M * rgl :D
1161206048 M * kevinp So Hollow, how do I migrate a page from the oldwiki to the new one, if I try and hit the new page and am redirected?
1161206065 M * daniel_hozac kevinp: wiki.linux-vserver.org
1161206098 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de
1161206284 M * kevinp ok, missed the wiki subdomain before
1161206425 M * kevinp so is the wiki language compatible and is there a way to get to the source of the old wiki for faster migration?
1161206450 M * kevinp I click edit on the old and get and error that it is locked
1161206455 M * daniel_hozac i don't think it's compatible.
1161206475 M * Bertl a lot of stuff is quite similar
1161206479 J * Piet_ hiddenserv@tor.noreply.org
1161206508 M * kevinp I still think it would be easier to start with the original source then migrate from the html version
1161206512 Q * Hollow Remote host closed the connection
1161206537 M * Bertl kevinp: yep, maybe Hollow can adjust that
1161206583 M * kevinp ok, let me know.  I was going to migrate the CentOS page, since I need to make a change on it
1161206612 M * Bertl well, best contact him immediately when he returns ...
1161206630 M * kevinp I'll try and keep an eye out for him :)
1161206815 M * daniel_hozac the CentOS page needs a lot of changes, IIRC.
1161206831 M * rgl I'm reading http://oldwiki.linux-vserver.org/Step-by-Step+Guide+2.6 and I don't quite understand why do we need a "<path-to-mounted-lvm>  /  none   bind" in fstab, can you tell me why?
1161206878 J * debugger ~Rui@217.129.151.190
1161206880 Q * Piet Ping timeout: 480 seconds
1161206884 N * debugger rgl_
1161206897 M * rgl_ omg... my net is crappy :(
1161206956 M * Bertl where is that?
1161206979 M * rgl_ Bertl, search for <path-to-mounted-lvm>  /  none   bind 
1161206988 M * rgl_ or /etc/vservers/<vserver-name>/fstab
1161207003 M * rgl_ its near the end of the wiki page
1161207030 M * Bertl ah, found it .. well, it 'suggests' to bind mount stuff already mounted on the host
1161207050 M * rgl_ why do I want to do that?
1161207051 M * Bertl you can as well mount it directly there
1161207086 M * Bertl no idea, I guess the 'author' did it this way ...
1161207097 M * rgl_ I don't have any entry on the fstab and it works fine (or maybe its missing something)
1161207118 M * Bertl do you have your guest placed on an lvm volume?
1161207126 M * rgl_ yes
1161207131 M * Bertl each guest on a separate one?
1161207150 M * rgl_ yes I will place each on each lvm volume
1161207177 M * Bertl so you would probably be better off to do the entire mount inside the guest
1161207179 M * rgl_ I shouln't?
1161207193 M * Bertl (to avoid inheriting the host mounts to your guests)
1161207203 M * Bertl although recent tools should take care of that
1161207221 M * rgl_ how do I known they are inherited?
1161207227 M * Bertl daniel_hozac: btw, how 'flexible' are 'we' regarding new namespaces?
1161207235 M * daniel_hozac hmm?
1161207245 M * daniel_hozac oh, you mean like IPC namespaces?
1161207247 M * Bertl rgl_: if you start a guest, it inherits all host mounts
1161207258 M * Bertl daniel_hozac: and pid spaces and utsname etc ...
1161207278 M * Bertl daniel_hozac: we would need some 'options' to add clone() flags
1161207288 M * daniel_hozac right, well, i haven't really looked in to it yet. shouldn't be too hard.
1161207320 M * Bertl okay, no need to hurry right now, my todo list still has a few other things ontop before the 2.6.19 migration
1161207324 M * rgl_ Bertl, what does that mean? :(
1161207331 Q * rgl Ping timeout: 480 seconds
1161207337 M * daniel_hozac clone is already available, it'd just need some new utils that set the new flags.
1161207352 M * rgl_ Bertl, you mean that the mount command is going to list all mounts I have on the host?
1161207358 M * Bertl rgl_: that means, everything mounted on the host, is 'copied' into the guest namespace
1161207372 M * Bertl rgl_: recent tools unmount most mounts right after that
1161207395 M * rgl_ Bertl, I still don't get it :(
1161207424 M * Bertl hmm, IIRC we had a wiki page about namespaces :)
1161207441 M * Bertl http://oldwiki.linux-vserver.org/Namespaces
1161207444 M * rgl_ Bertl, because the mount command on the guest only shows the normal guest mounts, that is, it doesn't show up all the host mounts
1161207454 M * daniel_hozac the mount command lies.
1161207460 M * Bertl the mount command is dumb
1161207463 M * daniel_hozac and the guest is additionally limited.
1161207478 M * fs Bertl: looks great
1161207486 M * rgl_ "cat /proc/mounts" lies too?
1161207514 M * fs starting the cpuhog only inside, outside and in both the guest and the host has no effect on the mouse and keyboard
1161207615 M * daniel_hozac rgl_: no.
1161207638 M * Bertl fs: that's how it is supposed to be :)
1161207655 M * Bertl fs: tx for testing!
1161207687 M * fs :)
1161207746 M * rgl_ daniel_hozac, humm, but in the guest I don't have for example the /sys mount, so is util-vserver unmounting stuff?
1161207760 M * daniel_hozac not by default.
1161207769 M * daniel_hozac but as i said, a full guest is further limited to the chroot.
1161207786 M * daniel_hozac you'll have to enter just the guest's namespace to get an accurate reading.
1161207788 M * Bertl rgl_: check with vnamespace -e <xid> -- cat /proc/mounts
1161207789 M * rgl_ humm, so if the mounts are inherited, howcome that /sys mount in not in the guest?
1161207804 M * fs Bertl: do you have an ETA for 2.0.2.2-final?
1161207842 M * Bertl fs: unless somebody brings any new issues, rc2 will be released as 2.0.2.2 final tomorrow
1161207859 M * fs good to know =)
1161207873 M * rgl_ Bertl, what is xid?  its the vserver context?
1161207873 M * Bertl so keep testing :)
1161207880 M * Bertl rgl_: yep
1161207898 M * rgl_ Bertl, why are the tools calling it xid?
1161207912 M * rgl_ x is from conte*x*t ?
1161207919 M * Bertl conteXt ID, yep
1161207931 M * rgl_ ok hehe
1161207987 M * rgl_ Bertl, woah, indeed that shows all the mounts in there
1161208012 M * rgl_ Bertl, woah, they are duplicated even!
1161208042 M * Bertl that's the result of the rbind mount
1161208121 M * rgl_ how can the guest access the lvm mount if I place it inside the vserver fstab?   do I have to use mknod to create the lvm device?
1161208149 M * Bertl nope, the 'mount' from the fstab happens with access to the host
1161208170 M * daniel_hozac the source is in the host, the destination is in the guest.
1161208172 M * rgl_ I see thx! :)
1161208355 J * Hollow ~hollow@i577B7503.versanet.de
1161208409 A * Hollow sighs
1161208430 M * Bertl hmm?
1161208472 M * Hollow the VPN used at my uni sucks
1161208481 M * Hollow everytime i connect to it, the rest of my LAN is dead
1161208489 M * Hollow and i can't change routes afterwards
1161208515 M * Hollow and it only works with some proprietary cisco software
1161208530 M * Bertl do you need that?
1161208550 M * Hollow well, i can only reach some services from inside...
1161208561 M * Hollow although i can use a http proxy as workaround
1161208563 M * Bertl tinc? openvpn?
1161208640 M * Hollow i tried vpnc and pptp for now, but they don't work... even the original cisco software doesn't work... they provide a special adapted version of it ..
1161208693 M * Bertl what was it you are studying there?
1161208707 M * Hollow info
1161208718 M * Bertl computer sciences?
1161208721 M * Hollow yup
1161208732 M * Hollow but the vpn is used for the whole university
1161208739 M * Bertl so you should figure a way to 'avoid' annoying vpns :)
1161208756 M * Hollow well... http proxy as a workaround would do it as well
1161208860 M * rgl_ daniel_hozac, how can I umount the uneeded stuff when I start a guest?
1161208879 M * daniel_hozac rgl_: with util-vserver 0.30.211, touch /etc/vservers/.defaults/namespace-cleanup
1161208894 M * rgl_ daniel_hozac, thx :)
1161208903 M * rgl_ daniel_hozac, why isnt that enabled by default?
1161208921 M * daniel_hozac because it breaks in some setups.
1161208935 M * daniel_hozac (should be better in 0.30.212)
1161208945 M * daniel_hozac but still lots of breakage potential.
1161208986 M * rgl_ what are the symtoms?   the vserver fails to start, just that?  or are they more severe?
1161208994 Q * bonbons Quit: Leaving
1161209005 M * daniel_hozac "just that", hehe.
1161209070 M * rgl_ alright :D
1161209111 M * rgl_ and if I wanted to do that on a specific guest only?  a touch /etc/vservers/example/namespace-cleanup does the trick too, right?
1161209119 M * daniel_hozac yep.
1161209135 M * rgl_ cool thx :)
1161209254 M * rgl_ daniel_hozac, odd, it didn't clean anything at all
1161209267 M * daniel_hozac hmm?
1161209292 M * rgl_ oh I lie... the /sys is not there anymore
1161209329 P * marcfiu 
1161209373 M * rgl_ daniel_hozac, http://paste.linux-vserver.org/512
1161209392 P * stefani I'm Parting (the water)
1161209395 M * rgl_ daniel_hozac, the /proc is in the guest namespace twice, is that normal?
1161209413 M * rgl_ and /dev/shm is there once too (and I think its not needed)
1161209430 M * daniel_hozac yes, as Bertl said, that's a result of the rbind.
1161209448 M * daniel_hozac /dev/shm has /dev in it, so it's kept.
1161209473 A * rgl_ doesn't known that is rbind hehe
1161209478 M * Bertl daniel_hozac: hmm, interesting ...
1161209486 M * rgl_ has /dev/shm ?
1161209491 M * rgl_ err
1161209509 M * rgl_ /dev/shm has /dev?  I don't understand why :|
1161209531 M * rgl_ what shm has to do with dev?
1161209536 M * daniel_hozac you see /dev at the beginning of /dev/shm, no?
1161209555 M * rgl_ yes
1161209561 M * daniel_hozac right, that's why.
1161209565 M * Bertl I think, the question is more .. why do we need/want /dev ?
1161209571 M * daniel_hozac so you can mount things.
1161209587 M * daniel_hozac otherwise vnamespace -e <guest> mount /dev/sda1 /vservers... wouldn't work anymore.
1161209604 M * Bertl ah, udev! okay ... now I got it :)
1161209617 M * daniel_hozac or devfs.
1161209629 M * rgl_ daniel_hozac, still don't understand why.  /dev/shm is mounted above /dev, so why isn't that "umounted"?
1161209643 M * daniel_hozac rgl_: because it starts with /dev.
1161209652 M * rgl_ daniel_hozac, ah!
1161209671 M * rgl_ daniel_hozac, util-vserver is skiping those, ok ok :)
1161209676 M * daniel_hozac (well, has /dev in it is more true)
1161209708 M * rgl_ huh?  now I don't see how you can say that :(
1161209739 M * daniel_hozac the string /dev is in the string /dev/shm.
1161209753 M * rgl_ I get you now :)
1161209804 M * Bertl Hollow, daniel_hozac: btw, rc41 supports building userspace headers
1161209832 M * daniel_hozac yeah, i saw that in one of the interdiffs, cool!
1161209944 M * daniel_hozac well, i gotta get some sleep, good night everyone!
1161209961 M * Bertl have a good one!
1161209996 M * Hollow me too, cu tomorrow
1161210022 M * Bertl u2 then!
1161210032 Q * Hollow Remote host closed the connection
1161210415 M * rgl_ Bertl, when doing the lvm mount on the guest fstab I'll loose the ability to access the guest fs from host, so are there disadvantages on mount on host besides the extra stuff that is inherited?
1161210468 M * derjohn2 Bertl, with the 2.6.18 kernel i found "TCP receive copy offload" - what is it?
1161210523 M * Bertl derjohn2: what does the help say?
1161210586 M * derjohn2 mmm ... /me in oldconfig mode. but maybe it's a good idea to look there first. (usually too short for me to get it right)
1161210588 M * Bertl rgl_: not really, just if you want to change it, it has to be unmounted in all namespaces
1161210613 M * Bertl derjohn2: oldconfig has '?' too, you know?
1161210650 M * derjohn2 Just wanted to prevent that some option breaks the xid tagging (/me imagines tco offloading to some place where are no xids or such ,,)
1161210684 M * derjohn2 oh  cool! yes ... ? works .. but fir the option mentioned above its too late ;)
1161210689 M * Bertl I'd assume it is a special network mode requiring special cards with TCO
1161210700 M * rgl_ Bertl, I see.  If I want to say, resize the volume I have to shutdown every vserver.  ok gotcha :)
1161210732 M * derjohn2 jffs2 has xattr support now - but I dont need it as we dont supoort jffs2 ?
1161210736 M * Bertl derjohn2: just remove it with sed or vi and rerun?
1161210748 M * derjohn2 jup . i'll so
1161210764 M * Bertl xattr != ext2 attrs (usually)
1161211198 M * derjohn2 it's a part of the DMA engine suport .. NET_DMA is he kernek option. what the heck is a dma engine? Or: which NIC does not do DMA?
1161211229 M * Bertl hmm, all of those cheep realtek ones?
1161211291 M * derjohn2 really? wow. But why is the option new in the kernel? I feel a dma engine might be something special ?
1161211305 M * derjohn2 anyway, I complie it in ;)
1161211313 M * derjohn2 *compile
1161211475 M * id23 good night
1161211490 Q * id23 Quit: Leaving
1161211635 M * derjohn2 Bertl, if I use NFS TAG xid as option and dont mount an nfs share with .o tagxid ... then there should ne no difference in using nfs from the "non vserver case" ?
1161211657 M * derjohn2 or is there some hidden danger when I enable it?
1161211836 M * Bertl there is, a tagged nfs uses a slightly different protocol
1161211840 Q * Piet_ Ping timeout: 480 seconds
1161211866 M * derjohn2 so I cant mount non-tagged NFS shares anymore?
1161211885 M * Bertl you have to test, IIRC I designed it 'compatible' :)
1161211886 M * derjohn2 Or do you speak about theoretcial problems?
1161211899 M * derjohn2 THATs what I wanted to hear :)
1161211978 J * Piet_ hiddenserv@tor.noreply.org
1161212224 M * derjohn2 Bertl: help says "Enable this if you do want the in-kernel NFS". Thats only partly true or? It only vaild for in Kernel NFS v3 ?
1161212272 M * Bertl it only affects v3 atm, but in any case, only kernel side .)
1161212273 M * derjohn2 furthermore it says that need patched clients. In which way? A client with a patched kernel only? Or also patches userspace (lockd, mount or whatever)?
1161212321 M * derjohn2 well, you should mention the v3 fact in the help ... maybe there are ppl who use v2 or v4 and dont get "it doesnt work".
1161212366 M * Bertl patches to make the help more comprehensible are always welcome!
1161212435 M * derjohn2 I which file resides the help?
1161212457 M * Bertl Kconfig
1161212477 M * Bertl there are severals of them, most vserver related are in /kernel/vserver/KConfig
1161212566 M * derjohn2 if I export a share via nfs and tag NFS on, will a non-patched client work? (I mean without tagging)
1161212576 M * derjohn2 file found BTW
1161212608 M * Bertl which version, what kind of tagging? :)
1161212644 M * rgl_ gtg.  g'night guys
1161212656 Q * rgl_ Quit: Fui embora
1161212770 M * derjohn2 well, the server exports /share and vserver NFS XID Tagging is on. Client Foo mounts the share but isnt even vserver patched (unlikely, but possible :)) Can Foo use the files?
1161212803 M * Bertl possible :)
1161212845 M * Bertl try it out, especially with different nfs versions and tagging options and let us know
1161212859 Q * shuri Remote host closed the connection
1161212878 M * derjohn2 uhm .... :)
1161213029 M * derjohn2 Bertl, we usually write the project's name as "Linux-VServer" ?
1161213063 M * derjohn2 the kernel menu says "Linux VServer" without the dash 
1161213074 M * Bertl yep, that's the canonical name Linux-VServer
1161213120 M * derjohn2 grrrr
1161213168 M * derjohn2 squashfs too few arguments to function get_sb_bdev
1161213181 M * derjohn2 hm, lets search for a new squashfs ...
1161213183 M * Bertl is that in the kernel now?
1161213216 M * derjohn2 no, I patch several stuff in. squash and drbd
1161213244 M * derjohn2 (an think about unionfs ... but dropped it due to the bme is not in kernel yet stuff)
1161213278 M * Bertl hmm?
1161213293 M * Bertl I mean, how's that related?
1161213319 M * derjohn2 I once patched the Unionsfs by hand and NULLed some function calls
1161213332 M * Bertl ah, okay :)
1161213336 M * derjohn2 i,e: It compiled but the BME stuff didnt probably work
1161213352 M * derjohn2 so I decided tio wait until BME get in vanilla ;)
1161213372 M * Bertl only a matter of a few years now :)
1161213383 M * derjohn2 I guess so :)
1161213405 M * derjohn2 IMVHO there are still some unresolved issues in
1161213492 M * derjohn2 like rsync's optiom "stay on oje filesystem", and moving files around from a BME mount to the real mountpoint
1161213524 M * Bertl a bind mount does not affect the filesystem
1161213619 M * derjohn2 move a file from a BME mount to the real mountpoint and it gets "moved by copying". I did that once with a 150 GB directory.
1161213643 M * derjohn2 I remeber also the "df -h" counter not updated properly
1161213649 M * derjohn2 (on a XFS)
1161213652 M * Bertl well, that's abug/missing feature of rsync/move/whatever
1161213692 M * derjohn2 mv bug? I dont think to. mv knows nothing about differnt FSes ?
1161213707 M * Bertl i.e. it is not bme specific, you get that with 'normal' 0815 bind mounts too
1161213721 Q * dreamind Quit: dreamind
1161213749 M * derjohn2 oh , lart /me but I thought thats the same ... I didnt know bind mounts != BME
1161213766 M * Bertl bind mounts are in mainline since 2.4
1161213785 M * Bertl (same with  rbind and move mounts)
1161213818 M * Bertl nevertheless, userspace tools do not really cope well with them (regarding recognizing that the underlying fs is same/different when required)
1161213852 M * derjohn2 why doesnt VFS find out what to do ?
1161213860 M * Bertl VFS is not involved
1161213893 M * derjohn2 but e.g. "mv" does call VFS to move the files ?
1161213903 M * Bertl let's make a simple example/test shall we?
1161213922 M * derjohn2 well, maybe that get too offtopic ?
1161213929 M * Bertl nah :)
1161213932 M * derjohn2 ok ;)
1161213954 M * Bertl dd if=/dev/zero of=/tmp/x.dat bs=1M count=1
1161213978 M * Bertl mount --bind /tmp /mnt/tmp
1161213978 M * derjohn2 done
1161213993 M * Bertl (might require creating the mountpoint)
1161214023 M * derjohn2 how do I create a mountpoint?
1161214028 M * derjohn2 </joke>
1161214037 M * Bertl stat /tmp/x.dat /mnt/tmp/x.dat
1161214066 M * derjohn2 all the same
1161214068 M * Bertl strace -fF /tmp/x.dat /tmp/y.dat
1161214092 M * Bertl strace -fF mv /tmp/x.dat /tmp/y.dat
1161214095 M * Bertl *sorry*
1161214120 M * Bertl strace -fF mv /mnt/tmp/y.dat /tmp/x.dat
1161214132 M * mnemoc :D
1161214183 M * Bertl hey mnemoc! how's going?
1161214314 M * derjohn2 Bertl, Invalid cross-device link
1161214350 M * derjohn2 it tries to "move", but cannot?
1161214369 M * Bertl right, maybe the kernel should handle that 
1161214421 M * derjohn2 well, thats what I meant with "VFS should know". But maybe I was on the wrong hotel room ...
1161214456 M * Bertl nah, actually I would have suspected that this is handled properly
1161214477 M * Bertl but obviously the kernel is playing cheap here too
1161214493 M * Bertl (maybe I'll add that to the next BME round?)
1161214509 M * derjohn2 doener looked it up for me once. he said something like "thats done by intention".
1161214530 M * Bertl ah? what was the intention?
1161214578 M * derjohn2 well ... I dont know, he didnt really have an explaination. but at that time i just wanted to know if my sys is misbehaving
1161214619 M * derjohn2 even more funny: when moving this 150 GB directory i pressed CTRL-C as it took too long.
1161214667 M * derjohn2 one part was moved, the other not. and the disk got 150 G space left, so it seems the inodes counter wasnt freed properly.
1161214683 M * derjohn2 (or: whatever entity counts the space).
1161214715 M * Bertl hmm, xfs?
1161214719 M * derjohn2 yes
1161214730 M * Bertl well, they do strange things there ...
1161214747 M * derjohn2 even xfs_rerair wasnt able to recover the space. I reformatted.