1164585641 J * lilalinux_ ~plasma@dslb-084-058-193-142.pools.arcor-ip.net
1164585772 M * Bertl Aiken: ping?
1164585788 M * Aiken yes?
1164585822 M * Bertl could you verify on your alpha that the scheduler is broken there too?
1164585855 M * Bertl http://linux-vserver.org/VCMD_HowTo
1164585886 M * Bertl last line, should run the hog at 25% but I think it doesn't for 2.1.1.2
1164585889 M * Aiken which kernel? it currently has 2.6.18.3-vs2.2.0-pre4 and 2.6.18.1-vs2.1.1-rc47-gb10502b2-dirty
1164585891 M * Bertl (or even 2.1.1.1)
1164585903 M * Bertl probably even 2.1.1 is affected
1164585928 M * Bertl just try with what you have atm, and see if it gets 100% cpu
1164586071 Q * lilalinux Ping timeout: 480 seconds
1164586099 M * Aiken silly ?, where is vcmd from?
1164586112 M * daniel_hozac http://vserver.13thfloor.at/Experimental/TOOLS/
1164586465 M * Aiken 2.6.18.3-vs2.2.0-pre4 I get http://paste.linux-vserver.org/700
1164586492 M * Aiken nevermind
1164586500 M * daniel_hozac you'll need the cpuhog too ;)
1164586505 M * daniel_hozac same directory.
1164586557 M * Aiken 100% cpu
1164586581 J * anonc ~anonc@staffnet.internode.com.au
1164586588 M * Aiken nothing a while (4 < 5) won't do
1164586604 J * harry_ ~harry@d54C2508C.access.telenet.be
1164586604 Q * harry Read error: Connection reset by peer
1164586772 M * Bertl Aiken: okay, now let's see if 2.1.1.2.1 fixes that
1164586783 M * Bertl (and/or adds new issues :)
1164586924 M * daniel_hozac it shouldn't compile :)
1164586940 M * daniel_hozac arch/alpha/kernel/ptrace.c is the first match for the missing linux/vs_base.h check.
1164586941 M * Bertl right ...
1164586968 M * Bertl but I'm confident we can fix that up with a small patch from you :)
1164587073 M * daniel_hozac i think i've got the script finding all the missing/not required base and context includes now.
1164587120 M * daniel_hozac (http://people.linux-vserver.org/~dhozac/t/find-problems.sh)
1164587151 M * Bertl okay, did you already run it on the 2.1.1.2.1? if so, what are the results?
1164587158 M * daniel_hozac i'm running it now.
1164587164 M * Bertl k
1164587175 M * daniel_hozac takes a while to awk through a kernel tree, at least on my box :)
1164587197 M * Bertl yeah, don't tell me that with my amd k6/400 :)
1164587210 M * daniel_hozac hehe, ouch.
1164587299 Q * DavidS Ping timeout: 480 seconds
1164587405 M * Aiken what you were expecting? http://paste.linux-vserver.org/701
1164587438 M * daniel_hozac yep.
1164587568 Q * mariilusii autokilled: Spammer. Mail support@oftc.net if you feel this to be in error. We do have spamassassin there tho. (2006/11/27 00.32)
1164587570 M * Aiken adding #include <linux/vs_base.h>   enough?
1164587575 M * daniel_hozac yep, should be.
1164587589 M * Aiken it is compiling 
1164587882 M * daniel_hozac Bertl: while that's running, https://savannah.nongnu.org/bugs/index.php?13844 is a debootstrap/Debian bug, isn't it? (amd64 not being on the official mirrors/locations), https://savannah.nongnu.org/bugs/index.php?15448 should be completely fixed by now as vlan handling is disabled by default.
1164587957 M * daniel_hozac https://savannah.nongnu.org/bugs/index.php?13845 i'm not sure how to reproduce, i've always gotten a sources.list when i've built guests with debootstrap.
1164587967 M * doener daniel_hozac: IIRC yes, the Sarge release for AMD64 is inofficial, should work for etch and later
1164587973 M * Bertl 13844: yes I think so
1164587978 M * doener .. with the default repo that si
1164587980 M * daniel_hozac doener: right.
1164587991 M * Bertl 15448: how is the default now?
1164588012 M * daniel_hozac VLAN devices are assumed to exist already, and are not brought down when the guest is.
1164588024 Q * Carpenter Ping timeout: 480 seconds
1164588085 M * Bertl okay, that's fine
1164588108 M * Bertl 13845: IIRC, that was just after isntalling a new sarge guest
1164588125 M * Bertl i.e. it was not able to reach anything as the config/sources were missing
1164588180 M * Aiken another compile error http://paste.linux-vserver.org/702
1164588225 M * Bertl please add an #include <linux/vs_context> at the beginning of that file
1164588297 M * Bertl (the ptrace.c just needs <linux/vs_base.h>)
1164588303 M * Bertl *linux/vs_context.h
1164588355 M * Aiken I put vs_base,h in ptrace.c
1164588368 M * Bertl yes
1164588685 M * Aiken yet again http://paste.linux-vserver.org/703
1164588696 M * Aiken put in vs_context.h
1164588717 M * Bertl yep, correct
1164588720 M * Aiken http://paste.linux-vserver.org/704
1164588723 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-headers-fix88.diff
1164588738 M * daniel_hozac that's all the script found.
1164588746 M * Bertl yep, context for fib_hhash too
1164588763 M * daniel_hozac i went with just vs_context.h for those that needed both.
1164588768 M * Aiken already done
1164589136 M * Aiken finally compiled
1164589163 M * Bertl excellent!
1164589358 M * Aiken tried that vcmd again, 98.9% cpu
1164589385 M * Aiken vcmd -i 100 -BC ctx_create .flagword=^34^33^32^8 -- /var/tmp/a.out
1164589421 M * Bertl okay, but with the old kernel, I hope?
1164589434 M * Aiken 2.6.18.3-vs2.1.1.2.1
1164589443 M * Bertl uhm ...
1164589454 M * Aiken I need to go into town
1164589454 M * daniel_hozac seriously?
1164589466 M * Aiken should be back in an hour
1164589469 M * Bertl Aiken: okay, np
1164589474 M * Bertl Aiken: thanks
1164589492 M * Aiken Linux pebbles.bedrock 2.6.18.3-vs2.1.1.2.1 #1 Mon Nov 27 10:54:28 EST 2006 alpha Unknown Alcor GNU/Linux
1164589503 M * Bertl did I forget some __enter? ... *searching*
1164589522 M * daniel_hozac no...
1164589641 M * daniel_hozac Bertl: about the bugs, you don't mind if close the VLAN and sarge x86_64 ones?
1164589660 M * Bertl sure, go ahead ...
1164589947 J * hardwire` ~hardwire@12.110.75.141
1164590151 Q * hardwire Ping timeout: 480 seconds
1164590177 M * Bertl daniel_hozac: I have one hope left, that he has the ahrd cpu scheduler disabled in the config :)
1164590182 J * derjohn ~derjohn@80.69.37.19
1164590188 M * daniel_hozac hehe, yeah, that has to be it.
1164590198 M * Bertl welcome derjohn! wanna test something for us?
1164590268 M * derjohn not if it is amd64 related. the box is simply broken ;( some HW failure
1164590277 M * derjohn Bertl, yes.
1164590279 M * derjohn what ?
1164590304 M * daniel_hozac i'm rebooting into 2.1.1.2.1+headers-fix88 now.
1164590317 M * daniel_hozac derjohn: whether the scheduler problem has been fixed :)
1164590340 M * Aiken Berlt it was not enabled, I thought it was because I had been making use of it in the past
1164590343 M * derjohn *aaargh*
1164590351 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.3-vs2.1.1.2.2.diff
1164590363 M * Aiken Bertl new kernel will be done by the time I get home
1164590372 M * Bertl Aiken: glad to hear that *phew*
1164590390 M * derjohn http://blog.derjohn.de/snipsnap/space/start/2006-11-27/3 <-- I just finished the 2.1.1.2 package set.
1164590400 M * daniel_hozac derjohn: too bad! muhahah! ;)
1164590415 A * derjohn thinks about doing it again ... will it be 2.1.1.3 or 2.1.1.2.1.2.1.1.2 ?
1164590435 M * daniel_hozac 2.1.2.1.2.1.2.1.2.2.1.2.1.2 :)
1164590449 M * Bertl nah, 2.1.2.1.2.1.2.1.1.2.1.2.1.2
1164590454 M * daniel_hozac Bertl: it's just getting 25% now.
1164590462 M * daniel_hozac (well, 25.9% to be exact)
1164590462 M * Bertl excellent!
1164590475 M * derjohn Bertl, the diff is a "all included" diff ?
1164590483 M * daniel_hozac yes, it's a complete patch
1164590510 M * derjohn what was wrong with 2.1.xxx^W current devel ?
1164590530 M * Bertl a bunch of things :)
1164590531 M * daniel_hozac the hard CPU scheduler was broken under certain conditions.
1164590546 M * daniel_hozac it was not attributing ticks to the context.
1164590578 M * derjohn hey, more cpu power per guest!
1164590647 J * Piet_ hiddenserv@tor.noreply.org
1164590668 M * daniel_hozac Bertl: any reason VLIMIT_MAPPED isn't allowed?
1164590696 M * daniel_hozac i.e. am i supposed to set something else?
1164590701 M * Bertl we have no memory check there
1164590714 M * Bertl RLIMIT_RSS is the enforced 'sum' now
1164590715 M * daniel_hozac ah, i must've misread the delta then.
1164590717 M * daniel_hozac ok.
1164590730 M * daniel_hozac that makes more sense :)
1164590747 M * Bertl yeah, and actually I added a check to the pte mapper
1164590760 M * Bertl that should make OOM kill the correct task :)
1164590877 M * daniel_hozac ��� 5 limit hit @ mm/memory.c:2048
1164590878 M * daniel_hozac yay :)
1164590931 Q * Piet Ping timeout: 480 seconds
1164591070 M * Bertl still the vserver badness is missing, but that should be easy to add now
1164591492 M * derjohn applying patch 12-patch-2.6.18.1-vs2.1.1rc42-ipv6k to ./ ... failed. :(
1164591610 M * Bertl that's what I actually planned to work on that weekend
1164591627 M * Bertl but hey, we hopefully improved a lot of other things
1164591787 M * daniel_hozac hmm, i'm pretty sure bonbons' patch is wrong in one of the rejects.
1164591827 M * daniel_hozac +       if (family == PF_INET6 && task_get_nx_info(current) && task_get_nx_info(current)->nbipv6 == 0) {
1164591833 M * daniel_hozac without any puts that i can see.
1164591899 M * daniel_hozac i'm guessing that's supposed to be current_nx_info().
1164591992 M * derjohn well, I can leave that out ... especially for testing.
1164592023 M * daniel_hozac leave what out?
1164592028 M * daniel_hozac oh, IPv6.
1164592058 M * derjohn yes, I am not using it anyway, just want to be prepared ...
1164592065 M * derjohn now the complie run started ...
1164592086 M * Bertl great!
1164592118 M * derjohn *lol*
1164592124 M * derjohn *compile
1164592349 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-2.1.1.2.2-ipv6.diff should apply.
1164592592 M * daniel_hozac (note: i just fixed the conflicts and changed it to current_nx_info, i didn't check if it's also missing includes)
1164592608 M * daniel_hozac i guess the compile will tell me that though :)
1164592743 M * derjohn huh ... well ... I let my compile run go ...
1164592770 M * derjohn i'll add the v6 stuff l8ter ... of wait until Bertl includes it with 2.1.1.1.2.2.22..3
1164592788 M * derjohn or isnt the Plan to inlcude it in devel ??
1164592853 M * daniel_hozac it is, this is just an interim solution.
1164592930 M * Bertl ipv6 will be in 2.3.x and late 2.1.x I guess
1164592942 M * Aiken Bertl does 24.8% sound better?
1164592951 M * Bertl superb!
1164593002 M * derjohn daniel_hozac, : you should include bonbons "edition letter" in the patch name. Is it the "K" edition of the patch ?
1164593011 M * daniel_hozac yes.
1164593018 M * daniel_hozac i reuploaded a fixed version though.
1164593020 M * derjohn fine
1164593039 M * daniel_hozac with (hopefully) all the necessary includes and VS_ADMIN.
1164593047 M * daniel_hozac it compiles for me :)
1164593223 J * jack ~jack@xplr-ts-t11-208-114-141-115.barrettxplore.com
1164593234 M * daniel_hozac hello jack
1164593267 M * jack Just looking to see if the domain transfer did proceed
1164593270 M * Aiken the scheduling jumps around a little bit, if I have top sampling every 10 seconds it shows 25%, if I have top sampling every second it is anywhere from 17% to 35%
1164593362 M * Bertl yes, that's excpected
1164593369 M * Bertl unless you limit the idle task too
1164593458 M * daniel_hozac jack: i haven't seen ntrs today.
1164593469 M * Bertl hey jack!
1164593480 M * Bertl ntrs is around 
1164593691 M * jack No news from him by email. Anyway, it does not matter much since the two secondaries are updated. The transfer will be done nevertheless
1164593717 M * Bertl okay, great! sec he left me a message for you, IIRC
1164593790 M * Bertl <ntrs> Herbert, if jack connects ask him about the status of the domain transfer.
1164593828 M * Bertl so porbably waiting for some confirmation or so on your/provider side
1164593870 M * jack No news from gandi yet
1164593906 M * Bertl okay, the secondaries still have old entries for .net and .info, and do not handle .com yet (as it seems :)
1164593927 M * Bertl but that's not such an issue
1164593945 M * Bertl (the .org was the important one, tx)
1164594120 M * jack I can fix them all. Just a moment
1164594173 M * Bertl make the ns1.13thfloor.at the master for now
1164594349 Q * lilalinux_ Remote host closed the connection
1164594527 M * Bertl daniel_hozac: http://vserver.13thfloor.at/Experimental/delta-badness-feat01.diff
1164594585 M * Bertl (actually those are two patches, a fix to the existing rss, and the feature)
1164594621 M * daniel_hozac right.
1164594625 M * jack Hum... whois claims that only linux-vserver.{org,info} point to my dns. The .com and .net point elsewhere
1164594647 M * Bertl jack: ah, okay, then that's my fault .. please ignore
1164594715 M * daniel_hozac Bertl: looks good.
1164594869 J * Aiken_ ~james@tooax6-214.dialup.optusnet.com.au
1164594911 J * duckx ~Duck@tox.dyndns.org
1164594942 Q * DuckMaster Read error: Connection reset by peer
1164595101 M * Bertl okay, guess I'm off to bed for tonight ... 
1164595117 M * Bertl have a good one everyone! and cya tomorrow!
1164595123 N * Bertl Bertl_zZ
1164595134 Q * duckx 
1164595194 Q * Aiken Ping timeout: 480 seconds
1164595402 Q * jack Quit: Leaving
1164595478 J * duckx ~Duck@tox.dyndns.org
1164597014 Q * hardwire` Ping timeout: 480 seconds
1164597051 J * Smutje_ ~Smutje@xdsl-87-78-98-134.netcologne.de
1164597157 Q * Smutje Ping timeout: 480 seconds
1164597157 N * Smutje_ Smutje
1164597268 J * jayeola ~jayeola@87.236.198.30
1164598176 Q * yang Ping timeout: 480 seconds
1164598187 J * yang yang@2a01:b0:1114::19:2
1164604558 Q * Piet_ Remote host closed the connection
1164607448 J * Carpenter ~gt-blacks@p5086E2C1.dip.t-dialin.net
1164612319 Q * eyck Ping timeout: 480 seconds
1164613450 J * dna_ ~naucki@111-193-dsl.kielnet.net
1164614493 J * hardwire ~hardwire@rdbck-3625.palmer.mtaonline.net
1164616031 Q * dna_ Quit: Verlassend
1164616254 J * hardwire` ~hardwire@72.35.108.81
1164616594 Q * hardwire Ping timeout: 480 seconds
1164616953 J * DavidS ~david@vpn.uni-ak.ac.at
1164617531 Q * ruskie Ping timeout: 480 seconds
1164617959 J * Piet hiddenserv@tor.noreply.org
1164618197 J * Torsti76 tkurbad@gate.iwm-kmrc.de
1164619142 J * ruskie ~ruskie@ruskie.user.oftc.net
1164619278 Q * Piet Ping timeout: 480 seconds
1164619363 J * Piet hiddenserv@tor.noreply.org
1164619550 Q * ||Cobra|| Remote host closed the connection
1164619768 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1164620246 J * djrise ~djrise2b@194.3.30.212
1164620257 M * djrise hello everyone
1164620322 M * DavidS hi djrise!
1164620406 M * djrise i have a question about /tmp in vserver
1164620427 M * djrise does anyone can help me
1164620430 M * djrise ?
1164620452 J * prae ~benjamin@foxhound.sherpadown.net
1164620453 Q * prae 
1164620608 M * DavidS only if you tell what your problem is
1164620655 M * djrise ok
1164620699 M * djrise i want to have a /tmp partition more than 16Mo (originally is 16Mo)
1164620725 M * djrise so i have set dlimit for the vserver to 5Go
1164620740 M * djrise but i have also 16Mo on /tmp
1164620786 M * djrise does anyone know how to increase ?
1164620963 M * waldi /etc/vservers/$name/fstab?
1164621051 M * djrise yes it work cool
1164621081 M * djrise but i set dlimit to 5Go for one vserver, it is normal when 
1164621113 M * sid3windr (s/o/B/)
1164621126 M * djrise i 'df -h' i see the entire partition (/var of the host) in /dev/hdv1 
1164621127 M * djrise ?
1164621207 Q * brcc Ping timeout: 480 seconds
1164621211 J * brcc bruce@i.am.someasshole.com
1164621768 M * nox djrise: there is no "virualized" partitiontab
1164621825 M * djrise so it's normal
1164621859 M * djrise i see the entire partition but i have only what i write in dlimit
1164623367 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1164623849 Q * thunder1 Ping timeout: 480 seconds
1164623938 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164624269 Q * ruskie Ping timeout: 480 seconds
1164624323 J * ruskie ~ruskie@ruskie.user.oftc.net
1164624623 J * SNy_ 16f1f55f38@bmx-chemnitz.de
1164624665 Q * SNy Read error: Connection reset by peer
1164625229 Q * Aiken_ Ping timeout: 480 seconds
1164626079 Q * ruskie Ping timeout: 480 seconds
1164626742 J * ruskie ~ruskie@ruskie.user.oftc.net
1164627082 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1164627092 J * ensc ~irc-ensc@p54B4F5BA.dip.t-dialin.net
1164627244 Q * thunder1 Ping timeout: 480 seconds
1164627251 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164627660 M * Torsti76 hi everyone
1164627703 M * Torsti76 does anyone know, whether 2.0.2.1 on gentoo still has the 2.6.17 xfs bug?
1164627741 M * Torsti76 2.1.1-r1 really doesn't run stable with our zope3 instances - i can't figure why
1164628319 M * Torsti76 ah, ok - xfs has been fixed since 2.6.17.7
1164628803 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1164628813 J * ensc ~irc-ensc@p54B4D3E1.dip.t-dialin.net
1164628836 J * prae ~Benjamin@foxhound.sherpadown.net
1164629144 J * aurel42 ~marc@P5c58.p.pppool.de
1164629218 M * aurel42 hello folks... I have a problem with vservers running on Debian. I think something went wrong when migrating from the legacy configuration to the current configuration system, when I start the init.d-script "util-vservers" it seems to start every client system twice.
1164629338 M * aurel42 Under /etc/vservers, for every vserver I have a file <vservername>.conf AND a configuration directory ("<vservername>/"). Should I really have both?
1164629355 N * SNy_ SNy
1164629579 M * aurel42 No, aurel42, you shouldn't. Just remove the .conf files and you'll be fine.
1164629581 M * aurel42 Thanks. :)
1164629599 J * dna_ ~naucki@130-209-dsl.kielnet.net
1164630154 J * lilalinux ~plasma@dslb-084-058-193-142.pools.arcor-ip.net
1164630421 P * aurel42 
1164630444 Q * cdrx Ping timeout: 480 seconds
1164630644 Q * m4z Server closed connection
1164630985 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1164630996 P * marcfiu 
1164631008 J * Guest16 ~etosi@151.100.100.23
1164631185 P * Guest16 
1164632332 J * cdrx ~legoater@cimai.net4.nerim.net
1164632760 M * Torsti76 okay, folks, there really seems to be a problem in the conjunction of 2.1.1.x and zope3
1164632760 M * Torsti76 downgrading to 2.0.2.1 helped
1164632810 M * Torsti76 we have another "unusual" - java based - service running that shows the same symptoms:
1164632865 M * Torsti76 suddenly the client looses the connection - only a restart of the service fixes the problem
1164632900 M * Torsti76 the strange thing: tcpdump still shows packets in both directions for the port in question
1164632973 M * Torsti76 the problems only occur with 2.6.18-based vserver kernels
1164632999 M * Torsti76 2.0.2.2-rcX doesn't run stable either
1164633117 Q * tso Server closed connection
1164633120 J * tso ~tso@244-005.dyn-fa.pool.ew.hu
1164634341 Q * oo Server closed connection
1164634343 J * oo oo@ormset.no
1164634772 M * daniel_hozac does it work as expected on 2.6.17?
1164634837 M * daniel_hozac djrise: no, df should only show the disk limited values. how did you configure them?
1164634896 M * Torsti76 daniel_hozac: yes, it does
1164634910 M * daniel_hozac i meant on 2.6.18, of course :)
1164634915 M * daniel_hozac (i.e. vanilla)
1164634951 M * Torsti76 daniel_hozac: i didn't try
1164634980 M * Torsti76 daniel_hozac: should i try with 2.1.1 and vanilla?
1164634990 M * daniel_hozac does export LD_ASSUME_KERNEL=2.2.5 or 2.4.19 make it work?
1164635013 M * Torsti76 nope, that can't work
1164635023 M * Torsti76 since i'm using glibc-2.5
1164635068 M * Torsti76 ldconfig then searches for libpthread.so.0
1164635091 M * Torsti76 and for sure won't find it ;o)
1164635171 M * Torsti76 i had a look at the genpatches at sources.gentoo.org 
1164635200 M * daniel_hozac my glibc-2.5 host has libpthread.so.0.
1164635207 M * daniel_hozac so i don't see why not.
1164635230 M * Torsti76 is your python compiled against glibc-2.5?
1164635242 M * daniel_hozac well, umm, yeah.
1164635242 M * Torsti76 if yes, try the following:
1164635251 M * daniel_hozac _everything_ is compiled against glibc-2.5.
1164635257 M * daniel_hozac it's not like i'm using Gentoo :)
1164635258 M * Torsti76 LD_ASSUME_KERNEL=2.4.19 python
1164635287 M * Torsti76 haha ;o)
1164635314 M * Torsti76 does the above line work for you?
1164635329 M * Torsti76 for me it complains about missing libpthread
1164635409 M * daniel_hozac indeed.
1164635436 M * daniel_hozac the question is why, given that /lib/i686/nosegneg has lib*.
1164635482 M * Torsti76 yes, but something inside me struggles against such kind of "solution"...
1164635496 M * Torsti76 since it affects the whole system
1164635541 M * Torsti76 i don't bother to run 2.6.17 at the moment
1164635573 M * Torsti76 we should only investigate some things before vserver 2.2.x ;o)
1164635580 M * daniel_hozac if everyone has that attitude, the bug will never be fixed :)
1164635644 M * Torsti76 don't get me wrong - i really want this to get fixed, but i don't think that assumed a two year old kernelversion is the right way
1164635663 M * daniel_hozac that's not the fix, obviously.
1164635680 M * daniel_hozac that's a test to see if it fixes it for you.
1164635691 M * daniel_hozac if it does, it will have narrowed down the problem area.
1164635704 M * Torsti76 ok, i see
1164635740 M * Torsti76 i'll try vanilla + vanilla-patch as soon as my box finished updating the world (i.e. gcc-4.1.1-r3)
1164635742 M * daniel_hozac (and would also make (at least) me rather certain it's the same bug ntrs has reported)
1164635748 M * daniel_hozac that likely won't help.
1164635752 M * daniel_hozac try with just vanilla.
1164635809 M * Torsti76 yes, with vanilla-patch i meant the 2.1.1 from the vserver-hp
1164635844 M * Torsti76 and gcc is more than halfway through, so i won't press ctrl-c now ;o)
1164635892 M * daniel_hozac and i mean without any patch at all.
1164635903 M * daniel_hozac try it with _just_ 2.6.18.3.
1164635909 M * Torsti76 that works
1164635931 M * Torsti76 let me just describe my investigations a little further:
1164635946 M * Torsti76 i have a host with 3 vservers:
1164635990 M * Torsti76 all three are running ssh
1164635997 M * Torsti76 perfectly stable
1164636009 M * Torsti76 one is running a heavily loaded apache with ssi, php5, several wikis and blogs based on php, perfectly stable, too
1164636036 M * Torsti76 the other two are running zope3 with some addons we develop(ed) here
1164636059 M * Torsti76 as soon as some people access the sites they stop answering
1164636077 M * Torsti76 and this occurs from 2.0.2.2_rc upwards
1164636091 M * daniel_hozac and what does strace say?
1164636117 M * Torsti76 i still didn't try
1164636123 M * Torsti76 the thing is:
1164636137 M * Torsti76 the service itself doesn't stop working
1164636151 M * Torsti76 tcpdump still shows packets in both directions
1164636167 M * Torsti76 i'm a bit in fear of running an strace on zope3
1164636201 M * Torsti76 since it sets up 4 twisted threads at once, etc.
1164636223 M * Torsti76 so this might be an awful lot of junk
1164636234 M * Torsti76 should i try it anyway?
1164636282 J * m4z ~m4z@bastard-operator.from-hell.net
1164636310 M * daniel_hozac -ffF -o zope.trace should give you zope.trace.<pid> files.
1164636340 Q * michal` Ping timeout: 480 seconds
1164636347 M * Torsti76 ah, ok - I'll try that and get back to you
1164636355 M * djrise a little help please
1164636361 M * djrise about network in a vserver
1164636389 M * daniel_hozac what about it?
1164636410 M * djrise when a start my machine (and vserver) for each vserver i have
1164636413 M * djrise RTNETLINK answers: File exist
1164636433 M * djrise i have two
1164636434 M * daniel_hozac what IP address have you assigned to them?
1164636446 M * djrise interface configured to them
1164636458 M * djrise for the host : 10.0.0.38/24 and 127.0.0.1
1164636479 M * djrise for the guest : 10.0.0.39/24 for 1/ and 127.0.0.1 for 0/
1164636508 M * daniel_hozac is 127.0.0.1 set as nodev?
1164636517 M * daniel_hozac and why are you giving your guests 127.0.0.1?
1164636563 Q * cdrx Ping timeout: 480 seconds
1164636584 M * djrise no i have dev for lo and i have 127.0.0.1 because i have got tree vserver
1164636591 M * djrise who needs loopback
1164636609 M * djrise bind9, postfix and apache2
1164636611 M * daniel_hozac why doesn't the loopback rewriting work?
1164636630 M * daniel_hozac at least BIND and httpd work fine without assigning 127.0.0.1.
1164636662 M * djrise ok i have read something wrong
1164636674 M * djrise so i need loopback only with postfix
1164636677 M * daniel_hozac you realize your guests will also be able to interfere with eachother if you give them all 127.0.0.1?
1164636683 M * daniel_hozac do you? what makes you say that?
1164636710 M * djrise yes now i understand
1164636715 J * michal` ~michal@www.rsbac.org
1164636720 M * djrise it's why for postfix i am must say nodev ?
1164636721 M * daniel_hozac but anyway, if you for whatever reason _actually_need_ 127.0.0.1, you'll have to set it as nodev.
1164636733 M * djrise what nodev please ?
1164636742 M * daniel_hozac in the directory-
1164636761 M * daniel_hozac touch /etc/vservers/<guest>/interfaces/<lo dir>/nodev
1164636891 M * djrise i try now
1164636930 J * thunder18 ~thu@tor-irc.dnsbl.oftc.net
1164636982 M * djrise thank you daniel_hozac it's ok now
1164636985 Q * thunder1 Remote host closed the connection
1164636985 M * djrise another one!
1164637000 N * thunder18 thunder1
1164637013 M * djrise as you understant i hace dns, mail and http in vserver
1164637061 M * djrise does i really need xinetd in the host and can i remove inetd from vserver ?
1164637078 M * daniel_hozac are you using them for anything
1164637136 M * djrise no for both , i don't set config for them but maybe system yes
1164637173 M * daniel_hozac so if you're not using them, what would be the problem with removing them?
1164637318 M * djrise because when i start a vserver i have
1164637365 M * djrise "no service enabled" in init.d script with inetd
1164637442 M * daniel_hozac so, disable inetd entirely.
1164637501 M * djrise Starting internet superserver: no services enabled, inetd not started.
1164637595 M * djrise thank you for all your answer i try that now
1164638044 J * s0undt3ch_ ~s0undt3ch@81.193.58.116
1164638104 J * _dmax ~semaj@81.193.58.116
1164638254 J * marcfi1 ~mef@aegis.CS.Princeton.EDU
1164638264 M * marcfi1 hello
1164638274 M * daniel_hozac hi
1164638274 Q * marcfi1 
1164638317 J * marcfiu ~mef@aegis.CS.Princeton.EDU
1164638320 M * daniel_hozac wb.
1164638324 M * marcfiu hello
1164638327 M * marcfiu that's better
1164638333 M * marcfiu hi daniel_hozac
1164638365 M * marcfiu did Bertl fix that scheduling problem we talked about over the weekend?
1164638389 M * daniel_hozac yes, it should be fixed.
1164638398 M * daniel_hozac 2.1.1.2.2 has it.
1164638413 M * marcfiu oh... looking for the fix in 2.0.2.2-x
1164638430 M * marcfiu maybe I can figure out the delta from 2.1.1.2.2
1164638432 M * daniel_hozac it hasn't been backported yet.
1164638449 M * daniel_hozac well... it required a restructure of the includes.
1164638454 Q * dmax Ping timeout: 480 seconds
1164638460 N * _dmax dmax
1164638465 Q * s0undt3ch Ping timeout: 480 seconds
1164638465 N * s0undt3ch_ s0undt3ch
1164638521 M * daniel_hozac but if you backport it, please post your delta to -rc8.
1164639227 N * Bertl_zZ Bertl
1164639231 M * Bertl morning folks!
1164639263 M * Bertl ah, marcfiu is working on back porting the include fixes?
1164639266 M * daniel_hozac morning!
1164639326 M * Bertl I thought about ripping the entire __enter/leave out of 2.0.x, as it seems it worked before (well, reasonably well at least :)
1164639346 M * Bertl without adding the in_interrupt() checks at all
1164639346 M * DavidS wb Bertl!
1164639380 M * Bertl hey DavidS! chances are minimal that I'll make it this evening
1164639423 M * DavidS :-( too bad, I'd hoped to see you again
1164639428 M * DavidS can't be helped
1164639435 M * Bertl (my car is still at the workshop)
1164639448 M * ntrs Bertl, it seems someone else reported the same socket issue with 2.6.18 and java
1164639455 M * Bertl DavidS: and ~80km without car is hard ...
1164639483 M * Bertl ntrs: no Linux-VServer? or with Linux-VServer?
1164639496 J * cdrx ~legoater@cimai.net4.nerim.net
1164639501 M * Bertl wb cdrx!
1164639501 M * ntrs with Linux vserver
1164639519 M * cdrx hi !
1164639524 M * ntrs look at the discussion between Torsti76 and daniel_hozac above
1164639536 M * DavidS Bertl: aye ... and since i'm still bike-only I can't even fetch you ;)
1164640159 Q * ruskie Quit: Caught sigterm, terminating...
1164640211 M * Bertl Torsti76: ping?
1164640252 M * ntrs Bertl, Torsti76 seems to be away. His name is greyed out in my irc client
1164640271 M * ntrs but his symptoms look exactly like I have here.
1164640371 M * Bertl well, from the irc log I cannot tell that
1164640391 M * Bertl first, daniel_hozac 'suggested' the fix you used
1164640395 J * ruskie ~ruskie@ruskie.user.oftc.net
1164640405 M * Bertl second, we do not really know if that helped at all
1164640423 M * Bertl the only common thing so far is that 2.6.17 seems to work fine, while 2.6.18 fails
1164640473 Q * Piet Ping timeout: 480 seconds
1164640509 J * Piet hiddenserv@tor.noreply.org
1164640524 Q * cdrx Read error: Operation timed out
1164640647 M * daniel_hozac Bertl: the __enter/leave was added to fix the MySQL issues, right?
1164640663 M * daniel_hozac the weird lockup issue, that is,
1164640927 M * Bertl yes, but it didn't help there
1164640960 M * Bertl at least ntrs is reporting quite similar issues with all 2.6.18 kernels/patches
1164641140 M * Torsti76 bertl: hi!
1164641166 M * Torsti76 i'm just rebooting my host into 2.1.1-r1 and then try strace
1164641192 M * Bertl could you give me a short version of what you 'think' you're observing?
1164641215 M * Bertl (I'll ask a few questions to that, where appropriate)
1164641216 Q * Hollow Server closed connection
1164641227 J * Hollow ~hollow@styx.xnull.de
1164641240 M * Torsti76 for me it looks like the network data packets get somehow incomplete
1164641268 M * Torsti76 we have been trying several things for about 2 weeks now
1164641303 M * Torsti76 the result is always the same: suddenly the browser loads forever
1164641320 M * Torsti76 tcpdump shows req and ack packets
1164641321 M * Bertl okay, so it starts loading, then hangs ...
1164641333 M * Bertl does it timeout eventually?
1164641341 M * Torsti76 and the zope process is still up
1164641350 M * Torsti76 no, it just loads and loads
1164641368 M * Torsti76 migrating the same code to a real machine solved the problem immediately
1164641385 M * Torsti76 a machine, running 2.6.18.3, btw.
1164641386 M * Bertl okay, can you use wget on such a page?
1164641404 M * Torsti76 i can try ;o)
1164641434 M * Bertl what I actually want to know is, do you get the same packet/stream over and over again or does it actually send you data
1164641473 M * Torsti76 so, how do i find out with wget? just looking at the progress bar?
1164641535 M * Torsti76 ok, first of all i have to "construct" the failure state
1164641546 M * Torsti76 that means a lot of clicking
1164641549 M * Torsti76 ;o)
1164641559 M * marcfiu bertl: morning
1164641590 M * marcfiu I'll take a look at back porting things.
1164641597 M * marcfiu haven't looked at 2.1.1.x yet.
1164641648 M * Bertl marcfiu: well, actually you want to head towards 2.2.x (for planetlab, I mean)
1164641667 M * Bertl marcfiu: but don't let me stop you from backporting stuff :)
1164641703 M * marcfiu bertl: I know... but I have to first create a stable release based on 2.0.x that we can roll out ASAP.
1164641720 M * marcfiu after that I will branch our kernel and move towards 2.2.x
1164641757 M * Bertl i.c.
1164641782 M * marcfiu so where can I find the fix in 2.1.x or 2.2.x?
1164641800 M * Bertl daniel_hozac: btw, it seems that bind9.2.3 does have the same issues (on my name server) with 2.6.18 but not 2.6.17
1164641803 M * Torsti76 bertl: ok, i'm ready
1164641827 M * daniel_hozac Bertl: i guess that's "good"...
1164641840 M * Bertl daniel_hozac: with strace, I see absolutely nothing, it seems that the packets do not even arrive at the socket
1164641862 M * Bertl have to do some tcpdumping there, will take a little
1164641863 M * daniel_hozac socket tagging?
1164641878 M * Torsti76 bertl: wget says it sent a request
1164641883 M * Torsti76 waiting for answer
1164641890 M * Torsti76 and getting none
1164641904 M * Torsti76 let's see, if some timeout occurs
1164641909 M * daniel_hozac marcfiu: interdiff should give it to you 2.1.1.2{,.2}.
1164641927 M * Bertl marcfiu: give me a minute to break down the changes
1164641933 M * Bertl marcfiu: you can then 
1164641937 M * Bertl backport them
1164641989 M * marcfiu excellent.
1164642008 M * Torsti76 bertl: still no timeout
1164642009 M * daniel_hozac Bertl: don't we want all the changes backported?
1164642018 M * daniel_hozac (the RSS stuff, i mean)
1164642032 M * Bertl nope, I'm not going to put the RSS changes in 2.0
1164642047 M * Bertl it's basically untested and just a feature
1164642057 M * marcfiu what's the url to the experimental area again?
1164642057 M * Bertl IMHO definitive 2.2 material
1164642072 M * Bertl http://vserver.13thfloor.at/Experimental/
1164642105 M * daniel_hozac Bertl: so RSS is going to be unlimitable for 2.0?
1164642146 M * Bertl we'll move to limiting anon instead
1164642157 M * daniel_hozac ah, ok.
1164642225 M * Bertl probably we'll switch RSS and ANON
1164642240 M * Bertl i.e. account anon as RSS and call the other MAPPED
1164642275 M * daniel_hozac makes sense.
1164642333 M * Torsti76 bertl: my tcpdump now shows traffic to be completely dead
1164642356 M * Bertl okay, restart the wget if possible and try to capture the relevant parts
1164642587 M * Torsti76 ok, i get some protocol traffic
1164642595 M * Torsti76 and then the whole thing hangs
1164642609 M * Bertl could you upload that dump please?
1164642627 M * Torsti76 bertl: just wanted to ask, where to... ;o)
1164642634 M * Bertl paste.linux-vserver.org
1164642654 M * Torsti76 bertl: just a second
1164642742 M * Torsti76 bertl: it's up
1164642796 M * Torsti76 bertl: if you want a more verbose dump, just let me know
1164643049 A * Torsti76 is hungry and will be in the bakery for the next 10 minutes
1164643098 M * Bertl ok
1164643187 M * Bertl marcfiu: first one: http://vserver.13thfloor.at/Experimental/delta-enter-rev01.diff
1164643338 M * Bertl marcfiu: btw, access to the serial console is still missing :/
1164643431 M * marcfiu oh
1164643439 M * marcfiu I was just trying to create that delta
1164643451 M * marcfiu I'll walk downstairs in a few minutes to do the serial console
1164643518 M * marcfiu ok
1164643530 M * marcfiu that diff just gets rid of the __enter and __leave stuff.
1164643546 M * Bertl well, it's basically the ssh access to vici? which is not working
1164643555 M * marcfiu oh
1164643556 M * marcfiu root
1164643583 M * Bertl root@vici.cs.princeton.edu's password:
1164643600 M * Bertl so I guess either ssh as root is disabled, or my key is missing :)
1164643606 M * marcfiu your key is there
1164643611 M * marcfiu ssh as root works
1164643628 M * marcfiu Maybe I messed up adding your key
1164643634 M * Bertl check the logs
1164643663 M * marcfiu try again please
1164643682 M * Bertl yep, works now, thanks!
1164643687 M * marcfiu good
1164643718 M * marcfiu are there any other delta diffs?
1164643720 M * DavidS Bertl: http://www.13thfloor.at/consulting/overview/ -> "Why Linux?" -> 404
1164643731 M * DavidS "Basically this is an error on your side." ;=)
1164643738 M * DavidS Not good PR ;)
1164643757 J * eyck ewGAjt1H@nat.nowanet.pl
1164643758 M * marcfiu Maybe "Why Linux?" doesn't need to be answered any more. :)
1164643810 M * Bertl DavidS: hmm, yeah, I should spend more time on my pages than on Linux-VServer ... you're right!
1164643846 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1164643852 M * Bertl wb eyck!
1164643972 M * DavidS Bertl: thank you for your feedback too.
1164643976 M * DavidS *grmpf*
1164643986 M * Bertl btw, can somebody verify if interdiff -q works for you?
1164644000 M * Bertl here it simply gives the help/usage stuff
1164644012 M * marcfiu seems to work for me.
1164644020 M * marcfiu Just did interdiff on 2.1.1.2.1 and 2.1.1.2.2
1164644036 M * Bertl with '-q'?
1164644041 M * marcfiu yes
1164644069 M * Bertl what patchutils version?
1164644078 M * marcfiu  interdiff -q patch-2.6.18.3-vs2.1.1.2.1.diff patch-2.6.18.3-vs2.1.1.2.2.diff  > schedfix.diff
1164644080 A * Torsti76 is back
1164644089 M * marcfiu [mef@localhost tmp]$ rpm -q patchutils
1164644089 M * marcfiu patchutils-0.2.30-4
1164644111 M * marcfiu [mef@localhost tmp]$ interdiff --version
1164644111 M * marcfiu interdiff - patchutils version 0.2.30
1164644112 Q * Piet Remote host closed the connection
1164644114 M * Bertl hmm, 2.30 here too
1164644129 M * marcfiu you sure you typing "-q". :)
1164644140 M * marcfiu bertl: does the serial console work?
1164644156 M * marcfiu if not, then I've probably plugged the cable into the wrong machine.
1164644163 M * Bertl yeah, -q is there, but the second arg was missing .. so nevermind :)
1164644169 M * marcfiu that is, possibly pluster-19 rather than pluster-20.
1164644190 M * Bertl I'll check the console in the next few minutes
1164644193 M * marcfiu ok
1164644206 J * Piet hiddenserv@tor.noreply.org
1164644249 M * daniel_hozac marcfiu: 2.1.1.2.1 to 2.1.1.2.2 isn't the schedfix, that's just cleanup.
1164644256 M * daniel_hozac (just clarifying)
1164644264 M * marcfiu just noticed that too
1164644277 M * marcfiu daniel_hozac: thanks for pointing that out.
1164644279 M * daniel_hozac as i said, you want 2.1.1.2{,.2}.
1164644319 M * Bertl marcfiu: no data on ttyS0/1 on vici
1164644330 M * marcfiu oh... interdiff 2.1.1.2 and 2.1.1.2.2?
1164644336 M * daniel_hozac right.
1164644343 M * marcfiu Bertl: ok... my fault most likely.
1164644381 M * Bertl screen on vici would be nice too :)
1164644407 M * Torsti76 bertl: i have to go soon - just reverted to 2.0.2.1 for the time being, but i'm open to further experiments
1164644411 Q * matti Ping timeout: 480 seconds
1164644426 M * Bertl Torsti76: 2.0.2.1 on 2.6.17, right?
1164644441 M * Torsti76 bertl: right
1164644472 M * Bertl okay, thanks for your time, we'll continue when you're back
1164644486 M * Torsti76 bertl: that's the last one that is working for me
1164644513 M * Torsti76 bertl: nope, thanks for YOUR time - and daniel_hozacs, too
1164644554 M * Bertl DavidS: if you have some lengthy explanation why Linux is so much better than the rest, I'd appreciate a contribution (and I'd add it immediately)
1164644622 M * Torsti76 cu
1164644625 P * Torsti76 
1164644795 M * marcfiu bertl: yes, it was plugged into the wrong machine.  Sorry about that.  Please try again.
1164644847 M * Bertl works, but it is a cable without handshake .. shouldn't be a problem though
1164644892 M * Bertl (have to test the magic-sysrq though)
1164644916 M * DavidS Bertl: What about http://www.linfo.org/reasons_to_convert.html ?
1164644979 M * Bertl but I guess I cannot copy that (at least it says All Rights Reserved :( )
1164645024 M * DavidS yay for free information flow :-(
1164645115 Q * Adrinael Server closed connection
1164645116 J * Adrinael adrinael@hoasb-ff0edd00-43.dhcp.inet.fi
1164645142 M * Bertl wb Adrinael!
1164645153 Q * DavidS Quit: Leaving.
1164645197 J * matti matti@linux.gentoo.pl
1164645220 M * Bertl welcome matti!
1164645312 J * sebastian ~info@p54A95B63.dip.t-dialin.net
1164645325 M * Bertl wb sebastian!
1164645449 M * Bertl marcfiu: http://vserver.13thfloor.at/Experimental/delta-base-clean01.diff
1164645471 M * Bertl that's the second one, would be interesting if a kernel with the two patches applied compiles
1164645483 M * Bertl (IMHO it should)
1164645677 M * marcfiu ok
1164645721 M * marcfiu will glotz at the diff for a bit. 
1164646141 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1164646490 J * whm ~whm@61.144.54.55
1164646539 Q * whm Remote host closed the connection
1164646689 Q * bubulak_ Quit: leaving
1164646869 Q * ruskie Ping timeout: 480 seconds
1164646936 M * marcfiu bertl: I don't think that applying http://vserver.13thfloor.at/Experimental/delta-base-clean01.diff does the trick.
1164646956 M * daniel_hozac why not?
1164646961 M * marcfiu that is, I would get some funky mix of vs2.0.2.x and vs2.1.x
1164646980 M * marcfiu daniel_hozac: I am just stating that I don't think I can apply the patch...
1164647017 M * marcfiu Still looking for the critical changes that fix the scheduling problem.
1164647023 M * daniel_hozac well yes, i wouldn't expect either of the patches to apply to 2.0.2.2-rcX as is.
1164647035 M * daniel_hozac see the changes to vx_check and __vs_check.
1164647038 M * marcfiu the first one does... only one hunk rejected.
1164647210 M * daniel_hozac and you have no left-over __enter/leave_vx_admins in the tree?
1164647256 M * daniel_hozac Bertl: hmm, delta-base-clean01 doesn't include the IRQ changes?
1164647296 M * marcfiu the patch refactors __vx_check from <linux/vs_base.h> to __vs_check in <linux/vserver/base.h>.
1164647319 M * daniel_hozac right.
1164647340 M * daniel_hozac but __vs_check doesn't include the IRQ things, while the vx_check define tries to use them.
1164647342 M * marcfiu The new vx_check includes a m|VS_IRQ
1164647363 M * Bertl daniel_hozac: I missed the vs_check change there
1164647377 M * Bertl will fix that up in place
1164647382 M * daniel_hozac okay.
1164647496 M * marcfiu so is it fair to say that the IRQ related fixes are isolated down to the definition of the vx_check() and the implementation of __vx_check() handling the IRQ flag/mode?
1164647544 M * Bertl the irq fixes will be the last patch
1164647699 J * ruskie ~ruskie@ruskie.user.oftc.net
1164647734 J * bonbons ~bonbons@83.222.39.117
1164647869 Q * waldi Server closed connection
1164647871 J * waldi ~waldi@bblank.thinkmo.de
1164647895 M * Bertl wb bonbons! ruskie! waldi!
1164647935 M * Borg- uhmm .. ive read Bonobos ;)
1164647947 M * Borg- too much Africa influence ;)
1164648032 M * Bertl have you been there recently?
1164648175 M * Borg- no really.. just listening to some music..
1164648266 M * Borg- anyway... off to home :>
1164648321 M * bonbons Hey Bertl!
1164648440 Q * Osgiliath Quit: The Sign Of Wrath Awaked.
1164648492 M * daniel_hozac bonbons: were you around yesterday (re: fixes to your IPv6 patch + rebase on 2.1.1.2.2)?
1164648548 M * bonbons I didn't see it, and unless my nick wasn't mentionned I wasn't around at that time...
1164648569 M * bonbons an idea at what time it was?
1164648674 M * daniel_hozac 02:52 onwards.
1164648688 M * daniel_hozac sorry, 02:43.
1164648749 M * bonbons that's night time ;)
1164648802 N * hardwire` hardwire
1164648862 M * bonbons Ok, found in the logs, let's check...
1164648982 M * bonbons if I understand correctly: I have a race-condition in the patch (when fetching current nx_info)?
1164649002 M * bonbons if so, there are more occurences of it...
1164649041 M * daniel_hozac no, more like it messes up the reference counting.
1164649072 M * daniel_hozac you get() the nx_info without put()ing it, so it would have lots of references.
1164649093 Q * ruskie Remote host closed the connection
1164649134 M * bonbons ok
1164649226 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-2.1.1.2.2-ipv6.diff uses current_nx_info() instead (i guess that's what you meant, right?)
1164649246 Q * Bertl Server closed connection
1164649247 J * Bertl herbert@IRC.13thfloor.at
1164649286 J * ruskie ~ruskie@ruskie.user.oftc.net
1164649392 J * stefani ~stefani@tsipoor.banerian.org
1164649461 M * Bertl welcome stefani!
1164649470 M * stefani hola
1164649724 Q * Piet Quit: Piet
1164649815 M * bonbons daniel_hozac: yes, that's it
1164650041 M * Bertl marcfiu: http://vserver.13thfloor.at/Experimental/delta-inirq-feat01.diff (that's the one you have been looking for :)
1164650085 M * Bertl marcfiu: in theory, backporting that and the first one (i.e. removing the enter/leave) would suffice
1164650217 J * bon_ ~bon@193.148.46.20
1164650255 M * Bertl daniel_hozac: how does util-vserver discover the dietlibc version?
1164650306 M * Bertl thing is this:
1164650313 M * Bertl # diet 
1164650313 M * Bertl diet version dietlibc-0.30
1164650318 M * bon_ ok hello :)
1164650328 M * Bertl configure: WARNING: *** dietlibc<=0.27 is known to be broken for x86_64 systems     ***
1164650346 M * bon_ i am having this newbie issue, that is df -h show nothing, i would like to see my partition there
1164650364 M * Bertl what does /etc/mtab contain?
1164650389 M * bon_ the guest one
1164650389 M * bon_ ?
1164650392 M * Bertl yep
1164650405 M * bon_ just /proc
1164650419 M * Bertl there you go, add something else there and df will show it
1164650464 M * Bertl btw, a default config should contain a few more entries
1164650567 M * bon_ what do i say..
1164650569 M * bon_ slackware .)
1164650614 M * Bertl okay, have to leave for a moment, should be back in 20 minutes
1164650624 M * bon_ thanks
1164650640 M * bon_ what about network devices inside the guest?
1164650657 M * Bertl should be fine, why?
1164650665 M * bon_ when i list them with ifconfig, there seems to be if0 and if0:<$vsname>
1164650671 M * bon_ is that the proper behaviour?
1164650682 M * Bertl only if you use older configs/tools
1164650695 M * Bertl recent tools do not create aliases by default
1164650699 M * bon_ hm
1164650700 M * bon_ 2.6.18.2-grsec2.1.9-vs2.1.1
1164650705 M * bon_ util-vserver: 0.30.211; Nov 21 2006, 17:58:07
1164650710 M * bon_ that doesn't seem to be that old
1164650720 M * Bertl nope, so you probably 'requested' and alias :)
1164650727 M * Bertl *an
1164650743 M * bon_ ok, i will talk to you about this later since i am lost .)
1164650746 M * Bertl remove the 'name' entry in interfaces/0/ (after a shutdown) and restart
1164650753 M * bon_ mhm
1164650753 M * Bertl okay, off now ..
1164650758 N * Bertl Bertl_oO
1164650764 Q * bon_ Quit: leaving
1164650939 J * bronson ~bronson@66.160.177.202
1164651216 Q * prae Quit: Quitte
1164651268 J * DavidS ~david@217.19.46.18
1164651493 N * Bertl_oO Bertl
1164651496 M * Bertl back now ...
1164651547 M * bon hm
1164651557 M * bon ok so can i have it being added into mtab after guest startup?
1164651571 M * Bertl yes, actually that is the default
1164651580 M * Bertl how did you create the guest/config?
1164651607 M * Bertl btw, it might be that your guest removes the /etc/mtab in some startup script
1164651615 M * bon seems so:/
1164651621 M * bon i will re-check later on, have to go now
1164651629 M * Bertl okay, np, cya!
1164651730 M * bon does it mount ufs?
1164651742 M * bon /dev/hdv1             149G  2.2G  147G   2% /
1164651748 M * bon hdv1 seems like hd virtual 1 :)
1164651775 M * Bertl yes, but that's just to avoid issues with quota tools
1164651779 M * bon i see
1164651787 M * Bertl nothing virtual there, it's the filesystem you use
1164651789 M * marcfiu bertl: thanks... will do the back port now.
1164651801 M * Bertl bon: but OTOH, nothing is mounted inside the guest
1164651808 M * marcfiu bertl: does serial console work for you now?
1164651832 M * Bertl marcfiu: yes, except for the magic sysrq, I have to investigate that further
1164651840 M * Bertl daniel_hozac: ping?
1164651969 M * daniel_hozac pong
1164652096 M * Bertl any idea to the dietlibc issue?
1164652109 M * Bertl (or did I miss an answer?)
1164652110 M * marcfiu magic sysrq?
1164652151 M * daniel_hozac nah, i just got back...
1164652164 M * marcfiu The patch file you gave me also had VS_HARDIRQ and VS_SOFTIRQ, where are those used?
1164652176 M * Bertl daniel_hozac: ah, k, didn't know :)
1164652191 M * Bertl marcfiu: nowhere for now
1164652195 M * marcfiu ok
1164652196 M * marcfiu good
1164652228 M * daniel_hozac Bertl: m4/ensc_dietlibc.m4:ENSC_ENABLE_DIETLIBC does the checking.
1164652250 M * Bertl which means?
1164652271 M * Bertl I mean, why does it tell the 0.27 stuff, when it obviously has 0.30?
1164652276 M * daniel_hozac Bertl: did you specify --enable-dietlibc?
1164652291 M * Bertl very likely
1164652299 M * Bertl let me check the spec file
1164652332 M * daniel_hozac that makes it skip the version checking.
1164652340 M * Bertl yep
1164652349 M * Bertl okay, so that 'noise' is normal
1164652356 M * daniel_hozac i'd say so.
1164652366 M * Bertl okay, fine, tx
1164652397 M * marcfiu bertl: So the back port is straightforward (obviously).  Do you want me to submit it as a single patch to you so you can then release an udpated rc9 for 2.0.2.2?
1164652428 M * Bertl yeah, but please compile and run test it first
1164652436 M * marcfiu compiling now
1164653108 J * XuS ~soussa@p5087413F.dip0.t-ipconnect.de
1164653118 M * Bertl welcome XuS!
1164653233 M * XuS thx
1164653284 M * marcfiu Bertl: looks like you missed the __enter_vx_admin and __leave_vx_admin in arch/.../kernel/irq-xen.c
1164653296 M * marcfiu I'll take care of those.
1164653320 M * Bertl marcfiu: I do not have an arch xen :)
1164653345 M * marcfiu I guess I got those from daniel_hozac.
1164653353 M * marcfiu 'cause fc has arch xen.
1164653374 M * marcfiu We also have a pl kernel with vserver running on xen domains at Duke.
1164653381 M * marcfiu Which is why we need it.
1164653469 M * Bertl that's fine with me, but doesn't change the fact that I obviously didn't miss it :)
1164653480 M * marcfiu oh right...
1164654813 Q * sebastian 
1164654934 Q * tso Ping timeout: 480 seconds
1164655169 J * sebastian ~info@p54A95B63.dip.t-dialin.net
1164656194 Q * ruskie Ping timeout: 480 seconds
1164657717 M * marcfiu Bertl: it turns out that it is not enough for vs_base.h to just #include <linux/hardirq.h>.  When compiling arch/i386/kernel/i8237.c it fails to properly define the in_interrupt() macro.  I have not narrowed down why this is the case just yet.  If I add a #include <linux/sched.h> to i8327.c then it works just fine.
1164657884 M * Bertl hehe
1164658143 M * Bertl ah, btw, http://plm.testing.osdl.org/patches/show/Linux-VServer-2.6.18.3-vs2.1.1.2.3
1164658209 M * trippeh Ohh, Lifecycle Manager, how enterpricy :)
1164658896 M * marcfiu wacky
1164659240 J * ruskie ~ruskie@ruskie.user.oftc.net
1164659447 M * Bertl http://vserver.13thfloor.at/Experimental/crosscompile-2.6.18.3-vs2.1.1.2.3.diff
1164660005 M * marcfiu compiling the following two line C code:
1164660005 M * marcfiu #include <linux/hardirq.h>
1164660005 M * marcfiu int foo() { return in_interrupt(); }
1164660005 M * marcfiu will correctly macro replace the above in_interrupt(), but not the one that is included by #include/vs_base.h
1164660027 M * Bertl that's why we did the base cleanup :)
1164660040 M * marcfiu oh
1164660067 M * Bertl if you look closely, you'll see, the problem is a recursive include 
1164660075 M * marcfiu yes
1164660079 Q * sebastian 
1164660099 M * marcfiu that's what I realized once you mentioned the base cleanup.
1164660103 M * Bertl so we decided to move the vs* stuff out of sched.h to break that
1164660169 J * Osgiliath ~osgiliath@kurzweil.no-ip.org
1164660219 M * Bertl wb Osgiliath!
1164660305 M * marcfiu Bertl: and then you needed to explicitly #include vs_base.h in a bunch of c files?
1164660334 M * Bertl well, as stuff was moved out of the vs_context and vs_network too
1164660352 M * Bertl a new base.h was introduced, and vs_base is the wrapper
1164660447 M * Osgiliath : )
1164660614 Q * DavidS Ping timeout: 480 seconds
1164661140 Q * click Ping timeout: 480 seconds
1164661560 Q * bonbons Quit: Leaving
1164661699 Q * XuS 
1164661784 J * click click@ti511110a080-5677.bb.online.no
1164662020 J * Aiken ~james@tooax7-043.dialup.optusnet.com.au
1164662048 M * Bertl morning Aiken!
1164662054 M * Aiken hello
1164662348 J * tso ~tso@241-059.dyn-fa.pool.ew.hu
1164662359 M * Bertl wb click! tso!
1164662850 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de
1164662857 M * Bertl hey dreamind!
1164662865 M * dreamind Hi Bertl :D
1164662876 M * Bertl I have something to test for you ...
1164662876 M * dreamind Bertl: any news about that rss / map_anon problem? :)
1164662881 M * dreamind oh nice :)
1164662891 M * dreamind wait a sec, I need to restart my irc client 
1164662899 M * Bertl np
1164662903 Q * dreamind 
1164662914 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de
1164662922 M * dreamind re :)
1164662931 M * dreamind now this is the latest colloquy build ;)
1164662945 M * Bertl you have the choice between:
1164662949 M * Bertl http://vserver.13thfloor.at/Experimental/delta-rss-feat01.diff
1164662956 M * Bertl and the current:
1164662966 M * Bertl http://vserver.13thfloor.at/Experimental/crosscompile-2.6.18.3-vs2.1.1.2.3.diff
1164662968 M * dreamind hm, experimental is 2.2?
1164662984 M * Bertl 2.2.x is upcoming stable
1164662987 M * dreamind currently I'm running 2.1.1.1
1164663008 M * dreamind is there any new really important feature with 2.2?
1164663014 M * dreamind and do I need newer utils for it?
1164663032 M * Bertl no, 2.2.x is basically stabilizing 2.1.x
1164663039 M * dreamind ah ok :)
1164663081 M * dreamind hm, and so whats the difference between those 2 patches? one is for 2.1.1.2.3 and the other?
1164663107 M * Bertl one is against 2.1.1.x the other against mainline
1164663124 M * dreamind ah ok
1164663156 M * dreamind hm, but I guess you recommend running 2.1.x on a working server right? ;)
1164663159 M * dreamind not 2.2.x ;)
1164663206 M * Bertl well, 2.2.x will be my recommendation once it is out
1164663225 M * dreamind ok :D
1164663232 M * dreamind I'll upgrade then shortly :)
1164663367 M * dreamind so I'll use the 2.1.1.2 patch from the website and the second diff upon it right?
1164663400 M * Bertl nope, either the first one, or instead of both of them, the second one :)
1164663416 M * Bertl but I pasted the wrong url for the second one :)
1164663426 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.3-vs2.1.1.2.3.diff
1164663435 M * Bertl that's the correct one :)
1164663453 M * dreamind aah ok
1164663459 M * dreamind so I'll just use the second one :D
1164663467 M * Bertl yep
1164663479 M * dreamind and create a new git branch in my local git repo *grin*
1164663486 M * dreamind .oO( git )
1164663532 J * RichyF ~RIchy@82-32-120-219.cable.ubr04.hawk.blueyonder.co.uk
1164663555 M * Bertl welcome RichyF!
1164663564 M * RichyF hi :D
1164663652 M * RichyF ok to pm Bertl?
1164663681 M * Bertl sure
1164663972 Q * phreak`` Server closed connection
1164663973 J * phreak`` ~phreak``@styx.xnull.de
1164664175 M * marcfiu Bertl: should nx_check() also call __vs_check with the IRQ flag?
1164664257 M * marcfiu I guess not
1164664286 M * Bertl not atm, we have no context related network checks with irq relevance
1164664629 Q * lilalinux Remote host closed the connection
1164664745 Q * meandtheshell Quit: Leaving.
1164664820 J * lilalinux ~plasma@dslb-084-058-193-142.pools.arcor-ip.net
1164665191 J * yarihm ~yarihm@84-75-123-221.dclient.hispeed.ch
1164665696 M * dreamind Bertl: I got an error while compiling: http://paste.linux-vserver.org/706
1164665835 M * dreamind Bertl: could that be because I'm using xfs?
1164666026 Q * lilalinux Remote host closed the connection
1164666041 M * daniel_hozac dreamind: arch?
1164666045 M * dreamind daniel_hozac: amd64
1164666075 M * daniel_hozac i guess adding #include <linux/sched.h> to fs/xfs/quota/xfs_qm_syscalls.c makes it work?
1164666100 M * dreamind didn't try that yet :)
1164666107 M * dreamind but I'll try
1164666268 M * dreamind daniel_hozac: kernel build 's running again ;)
1164666484 M * daniel_hozac strange, it works fine for me...
1164666499 M * daniel_hozac i guess i must have a strange tree or it doesn't show up on x86.
1164666500 M * Bertl probably xfs quota disabled?
1164666503 M * dreamind seems to work now.
1164666509 M * daniel_hozac no, i specifically enabled it.
1164666512 M * daniel_hozac after seeing the PLM results.
1164666513 M * dreamind Bertl: think so
1164666530 M * dreamind well maybe I disabled xfs quota but I'm not sure
1164666539 M * dreamind I could upload my config somewhere...
1164666547 M * dreamind perhaps to the pastebin?
1164666568 M * daniel_hozac you must have XFS quota enabled.
1164666574 M * daniel_hozac otherwise that file wouldn't be compiled.
1164666589 M * dreamind daniel_hozac: ok true ;) - seems I wrote before I thought ;)
1164666625 M * Bertl did we remove the sched.h there, btw?
1164666640 M * daniel_hozac no.
1164666667 M * dreamind my config: http://paste.linux-vserver.org/707
1164666694 J * lilalinux ~plasma@dslb-084-058-193-142.pools.arcor-ip.net
1164666879 J * DavidS ~david@62.178.45.213
1164667198 M * marcfiu crap... just realized that I hacked on getting the patches to work on the head of my tree vs. vanilla kernel. :(
1164667229 M * marcfiu I'll have to validate that things work and then will have to port those changes back to vanilla, as my tree is based on an fc kernel. :(
1164667714 M * dreamind ok the kernel is ready now, so lets hope it works out of the box ^^
1164668081 M * dreamind works like a charm :D
1164668088 Q * eyck Ping timeout: 480 seconds
1164668110 M * dreamind Bertl: thanks a lot :D
1164668128 M * Bertl great, so it does the job for you, right?
1164668146 M * Bertl I'd appreciate if you could also test the OOM badness we added
1164668261 M * dreamind Bertl: it does the job for me :D
1164668271 M * dreamind hm, how could I test it?
1164668315 M * Bertl relatively simple
1164668341 M * Bertl the idea is, that guests which are over soft limit will have a greater 'badness' when OOM strikes
1164668378 J * eyck 8c7RGNGP@nat.nowanet.pl
1164668388 M * dreamind sounds cool...
1164668415 M * Bertl so, the idea would be to get a bunch of task which consume 'known' amounts of memory
1164668433 M * Bertl let's say each one has 50MB of RSS
1164668451 M * Bertl and you set up, let's say 5 guests with different soft limits
1164668461 M * Bertl e.g. 100M - 500M
1164668473 M * Bertl (adjust the limits to your real hardware)
1164668586 M * Bertl then start the exact same number of those memory hogs inside each guest
1164668604 P * marcfiu 
1164668620 M * dreamind Bertl: I'll try that tomorrow or so ok? :)
1164668624 M * Bertl and provoke an OOM by adding hogs on the host until all memory is full
1164668624 M * dreamind because I need some sleep now :D
1164668635 M * Bertl no problem ... good night then!
1164668646 M * dreamind thanks, nite :)
1164668655 Q * dreamind Quit: dreamind
1164668843 Q * yarihm Quit: Leaving
1164668984 Q * dna_ Quit: Verlassend
1164669401 M * jayeola hey Bertl :)
1164669410 M * Bertl hey jayeola!
1164669453 M * jayeola can can a vserver have a guest of a different OS? debian/gentoo/etc all mixed up?
1164669472 M * jayeola i could not determine that from the wiki
1164669473 M * Bertl different OS, no, different distro, of course!
1164669491 M * jayeola ya-ya.. all *nix!
1164669507 M * Bertl more precisely, all linux :)
1164669521 M * Bertl i.e. netbsd won't work :)
1164669544 M * jayeola :) yes of course. thanks for that ... will prolly build next week. need a new box
1164669592 M * jayeola ah. and can guests have access to devices like wireless cards?
1164669615 M * daniel_hozac wireless what?
1164669641 M * jayeola pcmcia (wireless) cards
1164669655 M * daniel_hozac yes, but wireless what.
1164669659 M * Bertl hehe, are those wireless cards? I don't think so :)
1164669660 M * daniel_hozac ethernet?
1164669682 M * Loki|muh so you want a guest to use iwconfig&co.?
1164669690 M * jayeola yah
1164669702 M * Bertl probably the question boils down to:
1164669716 M * Bertl can wireless network cards be used as fixed nics?
1164669721 M * jayeola may need some bridging magic?
1164669730 M * Bertl s/as/like/
1164669748 M * Bertl if that is/was the question, then the answer is simple
1164669773 M * Bertl there is no real difference between wireless and wired nics as far as Linux-VServer is concerned
1164669795 M * jayeola :) will have to read the wiki and the sources a bit more
1164671591 Q * DavidS Read error: No route to host
1164671647 J * DavidS ~david@62.178.45.213