1121040208 M * Bertl hmm, probably worth a bug report ...
1121040419 M * Aiken commented out the offending include in kernel/dlimit.h so the build would finish
1121040419 Q * renihs Read error: Connection reset by peer
1121040427 M * Aiken still no problem with chbind
1121040439 M * Bertl ah, wait ...
1121040445 J * renihs ~renihs___@193.170.52.70
1121040456 M * Bertl somehow I missed this detail ...
1121040475 M * Bertl could you try to use the vserver headers from the kernel?
1121040490 M * Bertl i.e. replace the dlimit and friends with them
1121040512 M * Bertl (the include/linux/vserver/* files are for kernel and userspace)
1121040606 M * Aiken so use them to replace the headers in the kernel directory in the tool src tree?
1121040670 M * Bertl yep, sec, let me check how they are layed out ...
1121040826 M * Bertl yeah, 'should' work ...
1121040968 M * Aiken :(  http://pastebin.com/310927
1121041006 Q * renihs Remote host closed the connection
1121041147 M * Bertl okay, give me a minute, I got my kernel ready ... booting now
1121041324 M * Aiken it is building now, I used the include/linux/vserver/* from 2.6.11.11 rc4 
1121041483 M * Bertl k
1121041573 M * Aiken using those headers to rebuild the tools, chbind still works with no unaligned problem
1121041636 M * Aiken also no errors when building
1121042564 M * Rushmoom Hmm.. Starting sshd on the Gentoo vserver results in ERROR: "/etc/init.d/net" has syntax errors in it; not executing...
1121042577 M * Rushmoom net.eth0, which should provide net, is existent, but empty.
1121042585 M * Rushmoom equery says, net.eth0 is provided by baselayout-vserver. It also says, net.eth0 is linked to dummy (on the broken machine), whereas on a "proper" system it is linked to net.lo
1121042661 M * Rushmoom Is this a normal cause of the baselayout-vserver, or did something go wrong somewhere?
1121042664 M * Bertl hmm, well, net should not be initialized at all
1121042703 M * Rushmoom But then how is the dependency on net being resolved?
1121042735 M * Bertl no idea, I'm not very good with the gentoo dependancy system
1121042759 J * eXplasm explasm@p549F4295.dip.t-dialin.net
1121042792 M * Rushmoom Hm... re-symlinking the files helps, but this can't be a good solution, can it? ;-)
1121042814 M * Bertl no idea, I use some kind of mandrake :)
1121042827 M * Rushmoom I think I have to get back on Hollow on this issue tomorrow :)
1121042841 M * Bertl is probably the best idea ...
1121042895 M * Rushmoom But thanks for your help anyway, knowing that net shouldn't be initialized already helps me a bit...
1121043150 M * Aiken the rearanging of namespace.h and the new namespace_cmd.h seem to have something to do with why the tools compile with the 2.6.12.2 vserver headers
1121043166 M * Bertl Aiken: http://vserver.13thfloor.at/Experimental/patch-0.30.207-vs2.0.diff
1121043181 M * Bertl this updates the tool to vs2.0 kernel ehaders :)
1121043569 M * Aiken looks simillar to what I was getting trying to compile against the 2.0rc6 headers http://pastebin.com/310949
1121043599 M * Bertl very interesting ...
1121043627 M * Bertl compiled flawlessly ehre .. but maybe my 'original' isn't original anymore
1121043664 J * eXplasm2 explasm@p549F7CE7.dip.t-dialin.net
1121043665 M * Aiken vx_vhi_name_field is defined in namespace.h
1121043675 M * Aiken I'll try a rm -rf and start again
1121043686 M * Bertl give me a minute to double check it ..
1121043698 M * Aiken too late   :)
1121043702 M * Bertl :)
1121043751 M * Aiken the headers that introduce namespace_cmd.h also has a different namespace.h
1121043888 M * Bertl my 'original' sources are broken ...
1121044102 Q * eXplasm Ping timeout: 480 seconds
1121044217 M * Bertl Aiken: okay, updated the patch, please reload
1121044547 M * Bertl tools seem to work fine, but I still get the unaligned trap ...
1121044586 M * Aiken still building
1121044619 M * Bertl maybe it's the gcc, you use some 3.3.6 gcc, mine is the debian 3.3.5
1121044650 M * Aiken mine is a std gcc downloaded from a .au mirror of the gnu stuff
1121044739 M * Bertl we should swap the binaries ...
1121044775 M * Aiken compiled find and no unaligned trap error
1121044789 M * Aiken ping on the other hand is a different story  :(
1121044810 M * Aiken google -> ping unaligned trap gives many hits
1121044821 M * Aiken we could
1121044863 M * Bertl http://vserver.13thfloor.at/Stuff/chbind
1121044887 M * Bertl chbind --ip 127.0.0.1 true
1121045000 M * Aiken http://members.optusnet.com.au/~boddingt/chbind
1121045059 M * Aiken not a problem, your chbind works fine
1121045077 M * Bertl yours too, I get my alignment traps :)
1121045097 M * Aiken so it is a difference in our systems, not just the tool chain
1121045099 M * Bertl so I'd say it's a kernel option .. no?
1121045137 M * Bertl don#t think that those systems are _so_ different, but might be
1121045139 M * Aiken possible
1121045238 M * Bertl CONFIG_CC_ALIGN_FUNCTIONS=0
1121045238 M * Bertl CONFIG_CC_ALIGN_LABELS=0
1121045238 M * Bertl CONFIG_CC_ALIGN_LOOPS=0
1121045238 M * Bertl CONFIG_CC_ALIGN_JUMPS=0
1121045267 M * Bertl but that is kernel space, so it should not hurt userspace
1121045291 M * Aiken CONFIG_CC_ALIGN_FUNCTIONS=0
1121045292 M * Aiken CONFIG_CC_ALIGN_LABELS=0
1121045292 M * Aiken CONFIG_CC_ALIGN_LOOPS=0
1121045292 M * Aiken CONFIG_CC_ALIGN_JUMPS=0
1121045311 M * Bertl /proc/cpuinfo
1121045317 M * Bertl kernel unaligned acc: 2022 (pc=fffffc000039e8e4,va=20000002024)
1121045317 M * Bertl user unaligned acc: 53 (pc=120000870,va=2000000006a)
1121045336 M * Bertl do you have such entries?
1121045374 M * Aiken kernel unaligned acc    : 2091 (pc=fffffc000039ee88,va=20000002024)
1121045374 M * Aiken user unaligned acc      : 24 (pc=120000870,va=2000000006a)
1121045378 M * Bertl looks like the last unanligned access (which was fixed up)
1121045422 M * Bertl so that looks similar
1121045436 M * Bertl maybe your debug level is lower than mine?
1121045455 M * Aiken your chbind does cause unaligned traps
1121045486 M * Aiken so I was wrong above
1121045491 M * Bertl ah, wait, me too 
1121045510 M * Bertl so it is userspace ... probably a compiler issue
1121045527 M * Aiken it does not always cause a trap
1121045558 M * Bertl well, I'm not too worried, it will probably be fixed soon ...
1121045596 M * Aiken ok, so your chbind mostly causes an unaligned trap but not always
1121045609 M * Aiken and I have not seen a trap message from my chbind yet
1121045633 M * Bertl I'd assume it is a struct, which isn't properly aligned by my gcc
1121045639 M * Aiken so that explains why I said not trap when I first ran your chbind
1121045670 M * Bertl well, I have no such good explanation, I just did run the wrong one :)
1121045673 M * Aiken install gcc 3.3.6 under say /opt/gcc-3.3.6 and see what happens
1121045700 M * Bertl yes, will take ages to compile though ...
1121045744 M * Aiken I am think athlon compile speed here, whoops
1121045752 M * Aiken I would expect maybe 4 hours
1121045848 M * Bertl yeah, I will start a run and let it compile over night ...
1121046026 M * Aiken just saw one suggestion for getting rid of the the unaligned trap messages, remove the print from the kernel :)
1121046087 M * Bertl ah, very interesting approach :)
1121049530 M * Aiken line 212 in chbind.c, that is where I get traps using your binary and maybe sometimes line 220
1121049648 M * Bertl how do you tell the line?
1121049683 M * Aiken step; wait; step; wait; step
1121049687 M * Aiken with gdb
1121049702 M * Bertl ah, okay, thought there is a simpler way ...
1121049710 M * Aiken and I get the unaligned trap errors after executing line 212
1121049728 M * Aiken don't know, using this as an exercise to learn how to use gdb
1121050588 M * Aiken src/chbind.o(.text+0x640): In function `readIP':
1121050589 M * Aiken src/chbind.c:200: warning: warning: gethostbyname() leaks memory.  Use gethostbyname_r instead!
1121050596 M * Aiken do you get that when compiling chbind?
1121050744 M * Bertl sec
1121050820 M * Bertl src/chbind.c:200: warning: warning: gethostbyname() leaks memory.  Use gethostbyname_r instead!
1121050823 M * Bertl yup
1121050867 M * Aiken I am wondering what you happen if you tried the _r function instead
1121050874 M * Aiken what you = what would
1121050935 M * Bertl hmm probably has different syntax, but yeah I can try
1121051005 M * Aiken I get the same man page for each
1121051042 M * Bertl k
1121051071 M * Aiken very different
1121051109 M * Bertl yeah, the idea is to not allocate the returned struct in the library
1121052577 Q * virtuoso Read error: Connection reset by peer
1121053875 Q * monrad Quit: Leaving
1121053935 M * Aiken I can make mine produce an align trap just by putting a 1 line comment before line 179: struct hostent  *h;
1121053965 M * Bertl lol
1121055385 M * Bertl okay, I'm off to bed now ... thanks for your time Aiken!
1121055400 M * Bertl night everyone!
1121055405 N * Bertl Bertl_zZ
1121055431 M * Rushmoom Good night Bertl! :)
1121055533 Q * Hollow Remote host closed the connection
1121055551 J * Hollow ~Hollow@home.xnull.de
1121055565 M * Rushmoom Guten morgen :)
1121055588 M * Rushmoom Sorry, fallen back to german :)
1121056479 J * Aiken_ ~james@tooax6-050.dialup.optusnet.com.au
1121056795 Q * Aiken Ping timeout: 480 seconds
1121058579 J * jkl_ eric@c-67-165-222-86.hsd1.co.comcast.net
1121059036 M * DaPhreak lo Rushmoom ;P
1121059114 M * Hollow tztztz
1121059115 M * Hollow ;)
1121059181 M * DaPhreak ah .. lo Hollow ;)
1121059231 M * Hollow morning DaPhreak 
1121061572 Q * jkl_ Ping timeout: 480 seconds
1121061614 M * Rushmoom Just another Gentoo related question, before I finally go to bed :)
1121061620 M * Rushmoom Starting sshd on the guest results in ERROR: "/etc/init.d/net" has syntax errors in it; not executing...
1121061631 M * Rushmoom equery says, net.eth0 is provided by baselayout-vserver. It also says, net.eth0 is linked to dummy (on the broken machine), whereas on a "proper" system it is linked to net.lo
1121061637 M * Rushmoom Is this a normal cause of the baselayout-vserver, or did something go terribly wrong somewhere?
1121061829 M * Rushmoom BTW, linking /etcTinit.d/net.lo to /etc/init.d/net helps, but this can't be a proper solution :)
1121062248 M * DaPhreak Rushmoom: it is ;) since the all the basic startup scripts (like net) are symlinks to the dummy ..
1121062292 M * Rushmoom DaPhreak: So how do dependencies get resolved, when daemons depend on net?
1121062382 M * DaPhreak net.eth0 should be named net since, if its named net.eth0 the scripts get it wrong
1121062402 M * Rushmoom DaPhreak: Yep, they do, obviously :)
1121062439 M * Rushmoom DaPhreak: So /etc/init.d/net *should* exist, but be linked to dummy?
1121062459 M * DaPhreak so linking to net to dummy (ln -s /etc/init.d/dummy /etc/init.d/net) resolves the issue with this init-dependency
1121062479 M * DaPhreak since this net symlink now "provides net"
1121062509 M * DaPhreak which is need by other apps/daemons (like sshd or apache)
1121062536 M * Rushmoom DaPhreak: Hmm, I actually solved the problem that way (linked net to net.eth0, which itself is linked to dummy)
1121062567 M * Rushmoom DaPhreak: But I still wonder why this happened, and how I can prevent it in the future :)
1121062586 M * DaPhreak yeah .. also a possibility, but in the future (i hope so) that net.eth0 is useful again (if ngnet gets some more integration)
1121062609 M * DaPhreak Rushmoom: what do you want to prevent ? the missing symlink ?
1121062617 M * Rushmoom DaPhreak: exactly
1121062659 M * DaPhreak untar the stage{whatever} to a temporary space, fix the symlink, tar it back and in the future, install from this tar ;)
1121062669 M * Rushmoom DaPhreak: lol
1121062675 M * DaPhreak *G*
1121062688 M * Rushmoom DaPhreak: Ok, but then it's kind of a bug in the baselayout-vserver, isn't it?
1121062750 M * Rushmoom DaPhreak: I just want to make sure, I didn't miss anything, which also broke other things that hunt me down some day ;-)
1121062783 M * DaPhreak nah, you didn't do that, since i must had done the same here :) it's already known ;)
1121062798 M * Rushmoom -> *relieved* :)
1121063027 M * Rushmoom Thank you for your help, now I can go to bed and dream... about many many vservers, vcopy-ing themselves over and over again ;-)
1121063046 M * DaPhreak Rushmoom: i guess the stages were produced with an older ebuild, since the current one, fixes this
1121063053 M * DaPhreak lol
1121063063 M * DaPhreak yeah and eating all your space .. *G*
1121063082 M * Rushmoom DaPhreak: I guess the day that happens will come soon enough :)
1121063090 M * DaPhreak heh :)
1121063103 M * DaPhreak sweet dreams ;P
1121063131 M * Rushmoom Thanks! And to you a good nigh^W day! ;-)
1121063139 M * DaPhreak yeah .. day *g*
1121063144 N * Rushmoom Rushmoom_zZ
1121064597 Q * lilo Remote host closed the connection
1121065490 Q * DaPhreak Quit: leaving
1121065505 J * DaPhreak ~phreak@styx.xnull.de
1121066309 J * Doener_ ~doener@p54875C33.dip.t-dialin.net
1121066313 J * jkl_ eric@c-67-165-222-86.hsd1.co.comcast.net
1121066736 Q * Doener` Ping timeout: 480 seconds
1121067257 J * lilo ~lilo@lilo.usercloak.oftc.net
1121068132 J * Atcom ~atcoms@220.227.236.20
1121068137 M * Atcom hi all
1121068173 J * jsambrook ~jsambrook@aelfric.plus.com
1121068184 M * Atcom I have a problem with vserver which I recently configured based on 2.4x kernel
1121068230 M * Atcom I have only one NIC on server (eth0) when I create a VPS it creates it on same interface 
1121068254 J * BWare ~bware@office.intouch.net
1121068255 M * Atcom now I when trying to connect to VPS using shell, it is failing
1121068261 M * Atcom what can be the reason ?
1121068322 M * Atcom can anyone help me?
1121068451 M * maharaja mhm
1121068493 M * Atcom does vserver allow remote ssh connections to VPS?
1121068537 Q * Aiken_ Quit: Leaving
1121068596 J * prae ~prae@ezoffice.mandriva.com
1121068906 Q * gin_lit Ping timeout: 480 seconds
1121069306 Q * jkl_ Ping timeout: 480 seconds
1121069442 P * jsambrook 
1121069537 J * jsambrook ~jsambrook@aelfric.plus.com
1121069551 P * jsambrook 
1121071801 Q * Atcom Quit: 
1121072826 Q * Pazzo Quit: ..
1121074536 J * virtuoso ~s0t0na@80.253.205.251
1121074804 Q * zimbo Quit: 
1121074830 Q * mcp Read error: Connection reset by peer
1121075067 Q * cryo Remote host closed the connection
1121075273 J * cryo ~say@212.86.243.154
1121076123 J * jkl_ eric@c-67-165-222-86.hsd1.co.comcast.net
1121076999 J * Jani ~mail@G9109.g.pppool.de
1121087207 Q * DaCa-fi Ping timeout: 480 seconds
1121087499 M * Jani Is it possible to see as a non-rooted user the vserver-stat informatins?
1121087607 M * kestrel use sudo or equivalent
1121088878 J * jonsmel_zZ ~jscottorn@209.33.206.3
1121089289 M * Jani Ok, I will look into it. Thanks
1121089397 N * Bertl_zZ Bertl
1121089410 M * Bertl morning folks!
1121089431 M * DaPhreak lo Bertl 
1121089436 M * Jani Morning
1121089447 M * Medivh hi
1121089472 M * Bertl Jani: in theory, yes
1121089476 M * Medivh how can i specify which packages util-vserver shoudl install during vserver build (apt-rpm)?
1121089485 M * Medivh most recent util-vserver alpha her
1121089486 M * Medivh e
1121089508 M * Bertl IIRC, it supports arguments/package lists ...
1121089630 J * DaCa-fi ~danny@daca.info
1121089686 M * Medivh hrm, but it needs to take the default list from somewhere, and i'm searching for that somewhere ;)
1121089707 M * Medivh problem is that the build process for fc3 doesn't complete because it cannot find some commands which would need to be installed before anything else
1121089737 M * Bertl sec
1121089925 M * Bertl #  /etc/vservers/.distributions/dist/pkgs  
1121089931 M * Bertl (from the flower page)
1121089941 M * Bertl #
1121089941 M * Bertl Contains files with packagenames.
1121089941 M * Bertl     * list
1121089941 M * Bertl       File which contains the name of packages. On top of file the special keywords '--reinstall' and '--can-fail' are possible.
1121089973 M * Medivh [root@netz-stein fc3]# pwd
1121089973 M * Medivh /etc/vservers/.distributions/fc3
1121089973 M * Medivh [root@netz-stein fc3]# ls -al pkgs
1121089973 M * Medivh ls: pkgs: Datei oder Verzeichnis nicht gefunden
1121089975 M * Bertl where 'dist' is fc3 in your case
1121089975 M * Medivh mhmm ;>
1121090243 M * Medivh hm, well, another question then - i would rather use a preconfigured image anyway, but i've been using the 2.4 kernel patch til now, first time i'm doing 2.0-rc6 now... what configs do i need to create after i installed my image?
1121090322 M * Bertl best way is to use the skeleton build method, and _after_ that, copy over the image
1121090345 M * Bertl (just make sure that your fstab and /dev is correct)
1121090455 M * Medivh skeleton build? hm, how do i start that then? ;)
1121090487 M * Bertl http://linux-vserver.org/alpha+util-vserver
1121090506 M * Bertl     *  a basic skeleton (configuration entries, /dev entries); the command would be
1121090509 M * Bertl     # vserver <vserver-id> build -m skeleton
1121090737 M * Medivh ah great... i guess i'll make it from here... ;) thx!
1121090804 M * Bertl yw
1121091150 M * Medivh working perfectly... ;-))
1121091167 M * Bertl excellent!
1121091194 M * Medivh i really should move my 2.4 boxes to 2.6 when i got some more overview about how things work with the current patch ;)
1121091222 M * Bertl what do you need? :=
1121091259 M * Medivh basically an overview about how the configuration works... all the new stuff in /etc/vservers ;)
1121091302 M * Bertl hmm, the flower pages is very detailed about the what, and the vserver paper describes the why and how ...
1121091327 M * Bertl but basically you can use your old config with 2.6 quite fine ...
1121091327 M * Medivh yeah that's what i meant by getting an overview... go and read docs ;)
1121091391 M * Bertl (because the whole reason for backwards compatibility is to allow for lazy migration)
1121092790 J * Val ~val@v41.ath.cx
1121092807 M * Val Hi
1121092865 M * Bertl evening Val!
1121092876 M * Val evening Bertl :)
1121092898 M * Bertl somebody reported an issue with percpu data recently, is that somebody here now?
1121092989 M * Val Got an 'vxW: xid=XX did lookup hidden XXXX,XXXX "/proc/bus"' kernel message... What is it ? (vs2.0rc4 - util 0.30.207)
1121093020 M * Val when using vnamespace to add automount /home directory inside a vserver
1121093040 M * Val from the main host (using a post-start script)
1121093099 M * Bertl the vxW messages are 'warnings'
1121093113 M * Val ok
1121093122 M * Bertl in this case, some context (xid=XX) did try to lookup an inode
1121093131 M * Val ok
1121093150 M * Bertl (in this case XXXX,XXXX) in the procfs, which was intentionally hidden (by procfs security)
1121093178 M * Val ok, thx Bertl :)
1121093182 M * Bertl yw
1121093200 N * jonsmel_zZ jonsmel
1121093874 M * Val hum...
1121093876 M * Val well...
1121093913 M * Val when i use autofs to mount my home directory inside a vserver got a kernel panic :-(
1121093971 M * Val lot's of "clr_vx_info / set_vx_info / claim_vx_info / release_vx_info / set_vx_info"
1121093974 M * Val ...
1121093996 M * Bertl please upload!
1121094002 M * Val what ?
1121094007 M * Bertl the kernel panic
1121094007 M * Val the logs ?
1121094009 M * Val ok
1121094015 M * Val hum
1121094028 M * Val where to find the kernel panic logs ?
1121094046 M * Val hum /var/log/syslog may record this
1121094063 M * Val wait, rebooting :-(
1121094094 M * Bertl hmm, if it is a panic, your syslog will not be able to capture it :)
1121094116 M * Val note: i use loop-aes for accounts partition and autofs to get accounts automouted in /home inside vservers
1121094130 M * Val Bertl : well, how to get these logs ?
1121094161 M * DaPhreak heh, via serial console or typing them by hand ;)
1121094178 M * DaPhreak s/typing them/typing them off/
1121094193 M * Val ouch
1121094197 M * Val too late :)
1121094245 M * Bertl yeah, a photo is a good idea too
1121094501 M * Val i compiled in vserver debug code
1121094526 M * Val ...where to find the vservers debug logs ?
1121094539 M * Bertl in the kernel log
1121094547 M * Val hum
1121094549 M * Val fine =)
1121094624 M * Val "vxW: xid=67 messing with the procfs." in syslog, then nothing else :(
1121094690 M * Bertl well, the debug isn't enabled by default, it would flood your logs
1121094745 M * Bertl (but you can enable relevant parts via /proc/sys/vserver/debug_*)
1121094756 M * Val ok, thx 
1121094776 M * Bertl but if you get a panic, you should really record it, and upload it somewhere
1121094981 M * Val i'll try it
1121094991 M * BWare Bertl: I am experiencing occasional lockups on vserver hosts... I am guessing it has to do something with memory starvation and it seems vserver releated, but I have not got a clue on how to debug it 
1121095093 M * Bertl please define 'lockups'
1121095172 Q * Val Quit: testing with a "real" address (non private one) ... i'll be back
1121095389 M * BWare Complete hang of the server
1121095398 M * BWare It won't even ping :(
1121095413 M * BWare It seems completely random
1121095668 M * BWare I must admit I'm using 'old' versions of the code
1121095745 M * BWare Versions:
1121095745 M * BWare                    Kernel: 2.6.11.6-grsec-vs1.9.5
1121095745 M * BWare                    VS-API: 0x00010025
1121095746 M * BWare              util-vserver: 0.30.207; May 18 2005, 17:47:45
1121095769 M * BWare CC: i686-pc-linux-gnu-gcc, i686-pc-linux-gnu-gcc (GCC) 3.3.4 20040623 (Gentoo Linux 3.3.4-r1, ssp-3.3.2-2, pie-8.7.6)
1121095786 M * DaPhreak oO .. *g* the evil returns ... BWare, could you try the vs without grsec ?
1121095807 M * BWare I did... didn't help
1121095824 M * DaPhreak also the new rc6 ?
1121095841 M * BWare Not yet
1121095865 M * BWare I just don't know if it will fix the problems I am having
1121095878 M * Bertl depends on what causes your lockups
1121095890 M * BWare :)
1121095903 M * Bertl if it is broken hardware for example, it won't be fixed easily with software :)
1121095913 M * BWare hehehe
1121095932 M * Bertl okay, anyway, off for dinner now, let's continue later ...
1121095943 M * BWare well... It's all the 'same' hardware... (about 30 machines) 
1121095964 N * Bertl Bertl_oO
1121095964 M * DaPhreak on all occurs the same lockup ?!
1121095969 M * BWare yeah
1121095975 M * BWare occasionaly
1121095985 M * BWare and totaly random
1121096006 M * DaPhreak then this must be something related to the kernel ... since hardware is nearly unpossible :)
1121096024 M * DaPhreak s/unpossible/impossible/
1121096075 M * DaPhreak BWare: could you please try (if its possible, maybe on a single machine) the -rc6 ebuild ?
1121096085 M * BWare I can give a go
1121096089 M * BWare +it
1121096107 M * BWare but the stuff is pretty much production
1121096132 M * DaPhreak yeah, that why I said only a single machine ;) (i guess thats a cluster *G*)
1121096155 M * DaPhreak s/that/that's/
1121096181 M * BWare I'll see if I can confiscate some hardware and update to latest release
1121096200 M * BWare But I still don't have a clue on how to debug the host
1121096240 M * DaPhreak hmm any serial console there ?! or an (ip)kvm-switch ?
1121096281 M * DaPhreak so you could attach an monitor (if its an option) or a laptop / other pc to the serial ..
1121096287 M * BWare Serial console only through software ;)
1121096308 M * BWare Blades don't come with serial ports :(
1121096327 M * BWare I could try USB->Serial converter 
1121096373 M * BWare and capture serial with minicom or something
1121096508 M * DaPhreak maybe Bertl knows a trick (since I don't know another option than those I/Bertl said)
1121096521 N * DaPhreak DaPhreak|aw
1121097143 J * Val ~val@v41.ath.cx
1121097270 M * Val ... still get a kernel panic :(
1121097336 N * Bertl_oO Bertl
1121097348 M * Bertl back now ...
1121097367 M * Val i'm not sure that i did the right things in the right order to add autofs binding main host filesystem...
1121097382 M * BWare quick diner :)
1121097385 M * Val :)
1121097409 M * Bertl yeah, it was good ... :)
1121097432 M * BWare very tasty too I suppose :D
1121097448 M * Bertl yep
1121097508 M * BWare I'll see if I can setup a box to do some debugging... 
1121097570 M * BWare One other thing I probably should mention ... there is 2GB RAM in the machine using a split that I can't seem to remember ;) not using highmem though
1121097612 M * Bertl well, that is easy to figure ...
1121097628 M * BWare :)
1121097803 Q * Val Quit: test 2
1121098200 M * BWare back tomorrow
1121098318 M * Bertl k, cya!
1121099627 M * jonsmel Hi Bert, how's it going today
1121099743 M * jonsmel when you get a sec, I am wondering if I need to worry about this message or how do I get rid of it. When I enter a vserver I get a line that says mesg: /dev/pts/0: operation not permitted
1121099776 M * jonsmel I can enter the vserver and do what ever  I want so it works but I just want that message to disappear
1121100302 M * Bertl hey jonsmel! fine, thanks!
1121100340 M * Bertl jonsmel: well, you probably enter the guest with /dev/pts/0 being allocated/used on the host
1121100371 M * Bertl this is known and often reported as the 'screen doesn't work in a guest' issue, which is just wrong :)
1121100391 M * Bertl it does work, either inside or outside the guest, but not when half in ...
1121100409 M * FaUl re
1121100442 M * jonsmel thanks for responding bert, I have to run for now I will figure it out in a bit, thanks.
1121100456 N * jonsmel jonmel|out
1121100460 M * Bertl k, cya
1121100471 M * FaUl hi Bertl 
1121100513 M * FaUl i read lot of kernel-networking-sources a few days before, but i was not able to find that bug in ngnet ;-)
1121100551 M * Bertl hehe, bug exactly being?
1121100578 M * FaUl that routing-tables bug
1121100648 M * Bertl well, what your plans for tonight?
1121100782 M * FaUl my girlfriend doesn't like if i'm on my computer while she is asleep
1121100794 M * FaUl but i may use my laptop
1121100816 M * FaUl i think i'm back at 10PM or something
1121100827 M * Bertl CET?
1121101697 Q * prae Quit: Execute Order 69 !
1121102144 N * Rushmoom_zZ Rushmoom
1121102156 M * Rushmoom *yawn*... Morning! ;-)
1121102197 M * Bertl morning Rushmoom!
1121102254 M * Rushmoom Note to self: Change bio-rythm ;)
1121103262 M * baggins hi Bertl 
1121103278 M * baggins Bertl: testfs.sh results http://sith.mimuw.edu.pl/~baggins/testfs/
1121103285 M * FaUl Bertl: CEST
1121103300 M * baggins Bertl: vs2.0rc6 on x86_64
1121103326 M * baggins Bertl: looks worse than rc5 :(
1121103337 M * Bertl great! looks good, most of the 'false' errors are fixed in the latest version
1121103342 Q * cryo Ping timeout: 480 seconds
1121103405 M * Bertl baggins: I didn't check with different xattr tools, and depending on the version they have more or less flags
1121103466 M * Bertl and some errors are caused by a 'to-be-resolved' mainline kernel issue
1121103679 M * Bertl testing jfs filesystem ...
1121103679 M * Bertl [000]# failed.
1121103694 M * Bertl this means that you did not install the jfs utils (most likely)
1121103708 M * Bertl testing ext2 filesystem ...
1121103711 M * Bertl [001]# failed.
1121103726 M * Bertl this is the kernel issue I discovered yesterday ...
1121103743 M * Bertl (it is currently worked around by doing a sleep 6 :/
1121103787 M * Bertl testing reiser filesystem ...
1121103787 M * Bertl [000]# failed.
1121103801 M * Bertl this suggests that the reiser tools are not installed either
1121103820 M * Bertl  verify /test/file_1519: ----i-------- = ----i------------
1121103850 M * Bertl those are the additional flags I did not ahndle in vers=2
1121103863 M * baggins that's right, I didn't have jfsutils and reiserfstools
1121103896 M * baggins attr-2.4.20 here
1121103928 M * Bertl yeah, I made it a pattern match in 0.03
1121103954 M * baggins what bothers me is that almost all xid tests failed on XFS
1121103980 M * Bertl [001]# failed.
1121103988 M * Bertl means that the filesystem wasn't mounted
1121104027 M * Bertl http://vserver.13thfloor.at/Stuff/x.sh
1121104034 M * Bertl http://vserver.13thfloor.at/Stuff/x.sh.txt
1121104045 M * Bertl (this is with vanilla 2.6.12.2 :)
1121104072 M * Bertl the results are different every run .. each time a different fs can not be mounted
1121104176 M * baggins that's freaking weird, I just run a test on XFS and it's all green
1121104186 M * baggins except attr 103 and 113
1121104190 M * Bertl yep
1121104208 M * Bertl that's why I added the sleep in the last version
1121104231 M * baggins I wonder WHY it did not mount...
1121104255 M * Bertl it seems that the mkfs isn't written back before the mount happens
1121104268 M * Bertl and once it _is_ written back, it corrupts the filesystem
1121104289 A * Bertl already filed it on lkml ...
1121104389 M * baggins ah, yes sleep after mkfs helped, now all green except jfs, but that's expected?
1121104624 M * Bertl jfs has no notion of xattrs, so the xattr tests fail ...
1121104632 M * Bertl (some of them)
1121104645 M * Bertl the xid tests should be fine for jfs
1121104717 M * baggins my fault - mount: unknown filesystem type 'jfs'
1121104734 M * Bertl ah :)
1121104758 M * Bertl yeah, the verbose output was improved too (in the latest script version)
1121104950 M * baggins all green with latest testfs.sh :D
1121104970 M * baggins UID24/GID24 vs2.0rc6
1121105076 Q * xcyborg_ Read error: No route to host
1121105172 M * baggins I'll check all other combinations in a few minutes
1121105190 M * Bertl runtime and none are expected to fail
1121105210 M * baggins ok
1121105216 M * Bertl but all the other taggings should work on ext2,ext3, reiserfs and xfs
1121105226 M * Bertl jfs tagging works, xattrs probably fail
1121105233 M * baggins I wont check jfs - no time to compile it now
1121105241 M * Bertl okay, np
1121105265 M * Bertl did forget it too on my first testrun, just built it as module :)
1121105280 M * baggins be back in 20 minutes
1121105289 M * Bertl k, thanks, and cya!
1121105382 J * cryo ~say@212.86.243.154
1121105395 M * Bertl wb cryo!
1121105421 M * Bertl cryo: say? is it really you?
1121106446 M * Medivh hmm, i guess i got another question about the 2.6 patch... ;)
1121106459 M * Medivh how can i set a limit for the maximum space a vserver can consume?
1121106485 M * Bertl space as in disk space?
1121106592 M * Medivh yep
1121106635 M * Bertl vdlimit
1121106747 Q * cryo Remote host closed the connection
1121106764 M * Medivh space_total for limiting the max disk space of a vserver, and value in KB?
1121106898 M * Bertl vdlimit act as accounting system
1121106919 M * Bertl you have to add a new limit for your xid/fs combo
1121106932 M * Bertl and set the current as well as max values for blocks and inodes
1121106938 M * Bertl (blocks are now 1k IIRC)
1121106939 M * Medivh so basically i would...
1121106942 M * Medivh ... vdlimit --xid 200 --set space_total=2048000 /vservers/
1121106946 M * Medivh to limit xid 200 to 2 gb total?
1121106960 M * Bertl yes, but this alone would not work
1121106972 M * Bertl you also need to set the currently used space
1121106998 M * baggins Bertl: new results at the same place, everything looks as expected :)
1121107006 M * Bertl great!
1121107095 J * cryo ~say@212.86.243.154
1121107188 M * Bertl baggins: anny problems if I add you to our Hall'o'Fame?
1121107208 M * baggins Bertl: I'll be happy :)
1121107247 M * Medivh hmm
1121107247 M * Medivh [root@netz-stein vservers]# vdlimit --xid 200 --set space_used=`du -skx /vservers/uwe/ | awk '{print $1}'` /vservers/
1121107247 M * Medivh [root@netz-stein vservers]# vdlimit --xid 200 --set space_total=2048000 /vservers/
1121107248 M * Medivh [root@netz-stein vservers]# vserver uwe enter
1121107248 M * Medivh [root@uwe /]# df
1121107248 M * Medivh Dateisystem 1K-Blöcke Benutzt Verfügbar Ben% Eingehängt auf
1121107250 M * Medivh none 16384 0 16384 0% /tmp
1121107255 M * Medivh /dev/hdv1              2048000    451780         0 100% /
1121107261 M * Medivh shouldn't that be working tho...?
1121107298 M * Bertl check the current settings 
1121107303 M * Bertl (from the host)
1121107333 M * Medivh [root@netz-stein vservers]# vdlimit --xid 200 /vservers/
1121107333 M * Medivh 200 /vservers/
1121107333 M * Medivh space_used=451780
1121107333 M * Medivh space_total=2048000
1121107333 M * Medivh inodes_used=0
1121107334 M * Medivh inodes_total=0
1121107336 M * Medivh reserved=100
1121107340 M * Medivh looks quite fine to me, doesn't it?
1121107359 M * Bertl reserved is a little hight, and inodes_total quite low
1121107375 M * Bertl except for that, it is somewhat fine :)
1121107388 M * Medivh heh... it does need the inodes set?
1121107457 M * Bertl you can set them to unlimited
1121107500 J * monrad ~monrad@213083190130.sonofon.dk
1121107502 M * Medivh now that i changed reserved it works fine tho, even without inodes set
1121107524 M * Bertl baggins: your irc account info is correct?
1121107530 M * Medivh so do i need to recalculate current usage after reboots and set the limits, or are they kept?
1121107564 M * baggins Bertl: yes
1121107622 M * baggins Bertl: can you use latin2?
1121107674 M * Bertl hmm, you can improve it yourself (too late for me :)
1121107694 M * Bertl http://linux-vserver.org/Hall+of+Fame
1121107721 M * baggins thanks :)
1121107739 M * Bertl Medivh: either you recalculate it, or you save/restore it
1121107750 M * baggins how can I improve it?
1121107767 M * Bertl baggins: if I got some extended characters wrong :)
1121107768 M * Medivh Bertl, hmyeah saving would save quite some overhead i guess... ok, learning... ;-) thx again!
1121107793 M * Bertl Medivh: I have a hack tool (vdlimit) which can write it's own format
1121107813 M * Bertl i.e you can use the output to set it on startup
1121107873 M * baggins Bertl: s/e/ê/ ;)
1121107882 M * baggins ok time to go home
1121107889 M * baggins see you later
1121107922 M * Bertl cya!
1121111334 M * maharaja hi bertl
1121111337 M * maharaja i did as you asked
1121111351 M * maharaja right now, im getting a continous flow of backtrace information via my serial console
1121111396 M * maharaja it seems to be xfs related
1121111645 M * maharaja MHM
1121111649 M * maharaja or maybe not
1121111652 M * maharaja ill up the oops
1121111723 M * maharaja http://raoul.bhatia.at/~raoul/oops_no_smt_preemp_highmem.txt
1121111730 M * maharaja thats the information i got a zillion times
1121111731 M * Bertl hey maja!
1121111785 M * Bertl well, there is a lot of info missing before and after, no?
1121111920 M * maharaja i got back home and received a continous flow of this information
1121111925 M * maharaja i had no logging on - sry about that
1121111951 M * Bertl np, well, do you know what triggers it by now?
1121111978 M * Bertl did the non-preempt, non-smp kernel work longer?
1121112045 M * maharaja no, no noticable changes
1121112187 Q * brc Remote host closed the connection
1121112207 J * brc bruce@201008088244.user.veloxzone.com.br
1121112228 M * Bertl maharaja: okay, so you are interested in narrowing down the issue now?
1121112297 M * maharaja as stated yesterday, getting a stable setup is my primary concern right now. but i'm willing to help finding the issue
1121112345 M * Bertl okay, stable setup: avoid xfs, use ext2/ext3 and change the hardware
1121112447 M * maharaja do you think that the harddisk might be defective?
1121112490 M * Bertl it could be everything, but if you are looking for stability, a change of the hardware is a good way to go
1121112522 M * maharaja ok, ill try switching to ext3
1121112533 M * maharaja how may i help you?
1121112550 M * maharaja do you think that its a vserver related issue or a generic linux problem?
1121112570 M * Bertl I'm not sure it _is_ a linux issue at all
1121112591 M * Bertl I have no indication that it might be linux-vserver related
1121112619 M * Bertl but I also ahve no information to work on yet ...
1121112628 M * Bertl (almost no information)
1121112687 M * maharaja mhm
1121112735 M * maharaja is badblocks a sufficient non-destructive harddisk testing tool?
1121112739 M * maharaja so i can check the harddisks?
1121112753 M * Bertl not really
1121112764 M * Bertl if you do write tests, it's destructive
1121112776 M * Bertl and with read only tests, you will not get much
1121112795 M * maharaja mhm
1121112939 M * maharaja i missed a couple of lines during my c/p - but i think that they are not that important as there are in the middle of the calltrace
1121112970 M * Bertl the trace is probably not the result of the issue
1121112992 M * Bertl it is very likely a side effect once the kernel is messed up
1121113018 M * maharaja mhm
1121113048 M * maharaja and what about the trace from this morning: http://raoul.bhatia.at/~raoul/kernel_panic_new.txt
1121113098 M * Bertl Process °ý/°ý/°ý/Øý/ùþ/°ý/°ý/°ý/°ý/ùþ/°ý/°ý/°ý/°ý/ÿ/ÿ/%ÿ/°ý/°ý/qÿ/°ý/Qþ/°ý/°ý/Lþ/Í0à0à0à0à
1121113107 M * Bertl this is supposed to list the process
1121113116 M * Bertl (no real process is called so)
1121113127 M * Bertl and we are missing the registers
1121113148 Q * brc Quit: Connection reset by beer
1121113370 M * maharaja mhm
1121113397 M * maharaja damn damn damn it - ill change the cables tomorrow. but changing the hd would be tricky
1121113703 M * maharaja my problem is that i cannot reproduce this issue
1121113704 M * maharaja :-/
1121113725 M * Bertl you could try with a bunch of iozones
1121113732 M * maharaja what are iozones?
1121113742 M * Bertl an i/o test benchmark
1121113752 M * Bertl it really stresses the i/o system
1121113819 M * maharaja ill do that
1121113825 M * maharaja can you provide me with a link?
1121113832 M * maharaja ah, got it
1121113834 M * maharaja iozone.org
1121113850 M * Bertl yep, distro should provide it too
1121113964 M * maharaja do you recomand a way to test?
1121113980 M * Bertl let it run several times in parallel
1121113993 M * Bertl if possible on different filesystems
1121114012 M * Bertl put it into different contexts to stress this subsystem too
1121114114 M * maharaja i use "chcontext iozone -a"
1121114117 M * maharaja hopefully thats ok
1121114137 M * Bertl well, I'd prefer a static context id, but yes, it should work
1121114160 M * maharaja ok, testing it on xfs and ext3 on top of lvm
1121114275 M * maharaja tests are running
1121114309 M * maharaja why do you prefere static context? is there any _real_ reason or do you simply feel better by doing that?
1121114361 M * Bertl dynamic contexts are a) harder to track, and b) by design flawed ...
1121114409 M * maharaja mhm
1121114437 M * maharaja well, its working right now, if the server crashes, ill ask for help ;)
1121114582 M * Bertl do you have a serial or networ console now?
1121114628 M * Bertl (because it's important to capture the first oops and to capture it completely)
1121114649 M * maharaja still my serial console
1121114656 M * maharaja with non working magic sysrq key
1121114665 M * maharaja i today got another serial cable
1121114671 M * maharaja maybe ill get minicom to work with that
1121114697 M * Bertl hm maybe, there is also a wiring diagram available ...
1121114721 M * maharaja i got it from a friend of mine
1121114728 M * maharaja he never used it thou
1121114952 J * beichti ~beichti@p54AC86E3.dip0.t-ipconnect.de
1121114970 M * beichti Hallo zusammen
1121115005 M * beichti ich habe Fedora4 mit vserver-util 0.30 installiert
1121115023 M * beichti kann mir einer sagen wie ich dabei ram und cpu begrenze ???
1121115032 M * beichti den space habe ich schon begrenzt
1121115085 M * Bertl greetings beichti! channel language is english ...
1121115098 M * Bertl (you think you can rephrase that?)
1121115115 M * beichti hello , oh ok , than i must get my schoolenglish .
1121115172 M * beichti i have install fedora4 with linux-vserver and vserver-util 0.30 
1121115181 M * beichti i habe limit the space. 
1121115191 M * Bertl disk space, I presume
1121115193 M * beichti but i want to limit ram and cpu 
1121115199 M * beichti how i can do this ???
1121115206 M * Bertl you need newer tool for that, best 0.30.207
1121115220 M * Bertl the 0.30 (aka stable tools) are only for 2.4 kernels
1121115234 M * beichti i have kernel 2.6 
1121115234 M * Bertl (they work because of backwards compatibility with 2.6)
1121115249 M * Bertl but you cannot use the 'new' features with them
1121115251 M * maharaja ok, got some information 
1121115252 M * FaUl re
1121115259 M * beichti i habe install the version 0.30
1121115260 M * Bertl wb FaUl!
1121115260 M * FaUl oh, about 1 hour to late
1121115266 M * FaUl sry for that
1121115283 M * maharaja bertl: uploading the current oops, system is still up and running
1121115322 M * maharaja bertl: http://raoul.bhatia.at/~raoul/main_minicom.log
1121115360 M * Bertl hmm, that looks interesting ...
1121115369 M * maharaja ok
1121115371 M * maharaja i did a sysctl -a
1121115375 M * beichti berti i want only a little howto to limit the ram . 
1121115390 M * maharaja and got this error:
1121115391 M * maharaja error: "Operation not permitted" reading key "net.ipv6.route.flush"
1121115391 M * maharaja error: "Operation not permitted" reading key "net.ipv4.route.flush"
1121115411 M * Bertl beichti: get the new tools, and have a look here: http://linux-vserver.org/alpha+util-vserver
1121115457 M * maharaja Bertl: its crashed now
1121115486 M * maharaja the only thing i did is switching window inside a screen session
1121115508 M * maharaja i can provide the full logfile if you want to
1121115515 M * Bertl maharaja: you're still with -rc5, could you compile rc6 please?
1121115522 M * Bertl (just to make sure)
1121115539 M * Bertl but keep the 'original' vmlinux from your source tree
1121115543 M * maharaja Bertl: of course i can do
1121115686 M * maharaja i uploaded the whole logfile i got from minicom
1121115689 M * Bertl maharaja: which contexts did you use for the iozone?
1121115706 M * maharaja as said some random contexts and once 4444
1121115728 M * maharaja and context 0 (root system) for 2/3 tests (both xfs and ext3)
1121115747 M * Bertl so 200 and 500 and those are not from the tests, right?
1121115846 M * maharaja (#8c6d,*0):803ca3cc clr_vx_info b669c000[#900,171.46] @9f07b5a0
1121115851 M * maharaja you mean #900, don't you?
1121115858 M * beichti berti my be im to stupid , but i can set the limit for the space ... i don´t can get the limit for the ram , can you give me an example
1121115862 M * maharaja in this example
1121115880 M * Bertl beichti: you found the flower page?
1121115885 M * maharaja bertl: these are my vservers
1121115961 M * beichti flower page ??? 
1121115973 M * Bertl on the url I pointed you too
1121115988 M * Bertl (it describes the configuration options)
1121116006 M * beichti ah yes i have this page 
1121116016 M * beichti to smoke :-)
1121116061 M * beichti .30.207is the version on my vserver
1121116062 M * maharaja bertl: my vservers: http://pastebin.com/311493
1121116104 M * Bertl maharaja: ah, yes, good ...
1121116140 M * maharaja #200 is a mailserver running postfix
1121116180 M * Bertl okay, that oops is very informative ...
1121116215 M * Bertl maharaja: and it looks like a vserver issue (well indirectly)
1121116243 M * maharaja #102 is running some php shell scripts wich query gameserver for their current status and generate rrdtool stats
1121116250 M * Bertl do you have some usepsace tool polling the sysfs?
1121116315 M * beichti the userspace is ok .. 
1121116348 M * maharaja Bertl: not that i know of
1121116381 M * maharaja Bertl: the vservers run apaches, openldap and all the tools for pam_ldap, mysqls, postfixes and some shell scripts
1121116393 M * Bertl beichti: good, the relevant config limits are in rlimits, just set a limit for RSS(resident set size) or AS(virtual memory/adress space)
1121116487 M * Bertl maharaja: Process sysctl
1121116495 M * Bertl who is running sysctl?
1121116520 M * maharaja i called it by hand
1121116528 M * maharaja to see if the sysrq is enabled
1121116538 M * Bertl well, it definitely caused the creash/oops :)
1121116549 M * maharaja in ctx 0
1121116556 M * beichti sorry berti , but i find on the flowerpage nothing for limit ram like this 
1121116557 M * beichti vdlimit --xid 100 --set space_total=1024000 /vservers/
1121116577 M * maharaja true
1121116580 M * maharaja i reproduced it
1121116584 M * maharaja right now
1121116598 M * maharaja calling "sysctl -a" hangs the server
1121116602 M * maharaja mhm
1121116606 M * maharaja no, it does not hang the server
1121116608 M * maharaja it simply oopses
1121116612 M * Bertl beichti: yeah, because it is the vlimit tool you want, and very likely you set it in the config ...
1121116615 M * maharaja and i can continue working
1121116624 M * Bertl maharaja: which is good, as it should not happen :)
1121116646 M * maharaja ok
1121116649 M * maharaja still compiling rc6
1121116727 M * Bertl beichti: of course, you can set them at runtime too ... current settings and values are visible in /proc/virtual/<xid>/limits
1121116970 M * Medivh Bertl, would vlimit --xid 200 --as 128000 do for limiting the vserver to 128 MB?
1121116990 M * Medivh for context id 200 that is, of course
1121117022 M * Bertl it would limit the address space to 128000 pages
1121117039 M * Medivh hrm, what exactly does that mean again... ;P
1121117040 M * Bertl (page size should be 4-16k depending on the arch)
1121117061 M * Medivh arch is i686
1121117086 M * Medivh or, well, how do i find out the page size of my arch? ;)
1121117306 M * Medivh hm, as an example, let's say page size is 4k... then vlimit --xid 200 --as 32000 would do what i want (limit to 128 MB)?
1121117335 M * maharaja does sound ok to me
1121117616 M * Bertl yep, check current usage, max observed and limits in /proc/virtual/<xid>/limit
1121117658 M * Bertl maharaja: in kernel/sysctl.c
1121117666 M * Bertl (in the new kernel you are compiling)
1121117701 M * Bertl please change the following line
1121117702 M * Bertl if (table->virt_handler)
1121117705 M * Bertl to
1121117715 M * Bertl if (0 && table->virt_handler)
1121118053 M * Hollow evening folks
1121118062 M * Bertl evening Hollow!
1121118063 M * Medivh Bertl, what parameters does vlimit --ctx 200 --cpu expect?
1121118097 M * Bertl it's not active, but it would take seconds 
1121118105 M * beichti vlimit --xid 200 --cpu 50 means user can user 50 % from cpu ???
1121118116 M * Hollow Bertl: the segfault is gone by using the syscall() function from glibc... is this likely to cause problems?
1121118136 M * Medivh Bertl, oh, not active? so no way to limit cpu usage per vserver as of now?
1121118167 M * Bertl cpu 'usage' is a not very precise term ...
1121118175 M * Medivh yeah, well, true...
1121118189 M * Bertl the cpu 'limit' which would limit the total cpu time a context can get is not active
1121118218 M * Medivh well, let's put it this way: how can i prevent one context from eating 99% cpu and leaving nothing for the others? ;p
1121118220 M * Bertl the cpu 'usage' (often referred to as %cpu) can be configured via the scheduler
1121118235 M * maharaja berl: i'm done and currently scp the kernel
1121118249 M * Bertl maharaja: did you make the suggested changes?
1121118319 M * Medivh Bertl, scheduler sounds like what we need then... could you point me to some docs on the scheduler perhaps?
1121118360 M * maharaja Bertl: which changes?
1121118380 M * maharaja the smt/preemt/ht in bios/2-2gb split/etc from yesterday?
1121118397 M * Bertl maharaja: no, the changes in kernel/sysctl.c (please scroll up :)
1121118449 M * Bertl Medivh: http://linux-vserver.org/Linux-VServer-Paper-06
1121118533 M * Bertl http://linux-vserver.org/Scheduler+Parameters
1121118547 M * maharaja Bertl: mhm, damn...
1121118674 M * maharaja Bertl: changing and compiling
1121118703 M * Bertl will eliminate the issue with sysctl (until the cause it located)
1121118857 N * jonmel|out jonsmel
1121118872 M * Bertl wb jonsmel!
1121118879 M * jonsmel thanks
1121118993 M * jonsmel Hey bert, ?
1121119014 M * Bertl yep?
1121119016 M * jonsmel I am wondering if this error has to do with something I have messed up with a vserver config
1121119035 M * jonsmel when I try to tab complete a command it says
1121119068 M * jonsmel shell-init: error retrieving current directory: getcsd: cannot access parent directories: No such file or directory
1121119097 M * Bertl where do you get that?
1121119121 M * jonsmel just when I try to type in a command then tab complete it within a vserver
1121119193 M * Bertl sounds strange ...
1121119239 M * Bertl btw, what's getcsd?
1121119273 M * jonsmel oops, forgive me I had to type that whole line, it's actually getcwd
1121119291 M * Bertl okay, what does 'pwd' report there?
1121119309 M * jonsmel it does report my cwd
1121119337 M * Bertl okay, try to follow that up, by removing one entry after the other
1121119455 M * jonsmel hmm, said couldn't find cd
1121119459 M * jonsmel that's weird
1121119474 M * jonsmel I exited out of vserver and reentered back in and now it worked
1121119488 M * jonsmel I still get that error I had before though
1121119545 M * jonsmel mesg: /dev/pts/0: operation not permitted
1121119572 M * Bertl don't enter and expect the pts to follow you
1121119581 M * Bertl use ssh to reach the guest
1121119646 M * jonsmel oh, ok so i can't enter that way I have to ssh
1121119661 M * Bertl well, you can, but you should not make assumptions
1121119679 M * Bertl (like that every open filedescriptor, or terminal will follow you inside :)
1121119757 M * jonsmel ok, gotcha
1121119801 M * maharaja bertl: rebooted with your sysctl patch
1121119823 M * maharaja "sysctl -a" is working
1121119829 M * maharaja without oops
1121119830 M * Bertl okay, great .. let's restart the iozones
1121119843 M * maharaja k
1121119851 M * maharaja ill use fixed ctx
1121119862 M * Bertl could be that some tool called/polled the sysctl ... and this caused your issues ...
1121119875 M * Bertl (that is definitely a bug I'm investigating right now)
1121120122 M * maharaja 4x xfs and 4x ext3 iozones are running
1121120160 M * Bertl excellent ...
1121120233 J * brc bruce@201008088244.user.veloxzone.com.br
1121120240 M * Bertl welcome brc!
1121120247 M * brc bertl
1121120252 M * brc is there a vserver start debug ??
1121120271 M * Bertl start debug?
1121120272 M * Jani vserver --debug <name> start
1121120288 M * Bertl ah, thanks! yep!
1121120293 M * brc ok :)
1121120299 M * Jani *g*
1121120318 P * beichti 
1121120344 M * Jani Bertl: Btw. I have make a whole new vserver, it really seems that the vserver user have fiddle around with a library and this have caused the init.d problems
1121120359 M * Bertl ah, i.c.
1121120368 M * Jani i.c.?
1121120375 M * Bertl I see :)
1121120409 M * Jani Oh, right...
1121120409 M * Jani  :|
1121120417 M * maharaja damn, i'm tired
1121120418 M * Bertl it's kind of onomatopoetic :)
1121120442 M * Jani hehe
1121120782 Q * Hollow Read error: Connection reset by peer
1121120995 M * maharaja Bertl: the system is still up and running the tests
1121121013 M * Bertl looks good so far ...
1121121028 M * Bertl how long did it take last time? (with the iozones in place?)
1121121069 M * maharaja approximatly 15mins
1121121131 M * Bertl looks like there is some kind of sysctl corruption going on, which indirectly triggers the issue ...
1121121437 M * maharaja mhm - lets see
1121121440 M * maharaja i'm going to bed now
1121121451 M * jonsmel well, night all
1121121457 M * Bertl maharaja: night!
1121121459 N * jonsmel jonsmel_zZ
1121121478 M * Bertl hmm, night jonsmel_zZ too!
1121121499 M * maharaja feel free to check in on http://raoul.bhatia.at/ once in a while. when its unavailable then the kernel crashed again ;)
1121121511 M * maharaja good night bertl!
1121121517 M * Bertl k, night!
1121124392 J * Aiken ~james@tooax8-217.dialup.optusnet.com.au
1121126392 M * Bertl FaUl: hmm, didn' realize that you are back ...