1139788812 M * Bertl wb zobel! 1139789118 J * FireEgl Atlantica@Atlantica.Tcldrop.US 1139789767 M * daniel_hozac is vs_base.h automatically included somehow? 1139789802 M * Bertl hmm, no, but some .h files pull it in IIRC 1139789810 J * grant_ mep@p50918A0D.dip0.t-ipconnect.de 1139789837 M * daniel_hozac linux/sched.h. 1139789940 M * daniel_hozac and then everything includes linux/sched.h? 1139789959 M * Bertl yes, sched.h is extremely abused 1139789978 M * daniel_hozac ok. 1139789987 M * daniel_hozac just wondered where vx_flags came from :) 1139789989 M * Bertl because of the task struct 1139789998 M * daniel_hozac oh, right. 1139790229 Q * grant Ping timeout: 480 seconds 1139790460 M * daniel_hozac is current->pid correct? i.e. does it handle initpid and such? 1139790503 M * Bertl probably not 1139790528 M * Bertl well, it depends on where, if it goes userspace, then it's wrong 1139790545 M * Bertl if it is a kernel message or kernel function, then it's fine 1139790547 M * daniel_hozac log messages, mostly. 1139790559 M * Bertl kernel log is fine 1139791016 M * daniel_hozac all files that use find_task_by_pid need #include , right? 1139791033 M * Bertl yup 1139791148 M * daniel_hozac http://daniel.hozac.com/tmp/delta-vs_pid-clean01.diff 1139791221 M * Bertl the fs/proc/base.c is already covered, no? I mean with the wrong cvirt :) 1139791256 M * daniel_hozac oh damn, did i do the ptrace.h change in that tree? 1139791272 M * Bertl np, 1139791324 M * Bertl wh does signal.c suddenly need it? 1139791347 M * Bertl and do we need cvirt.h in sys.c? 1139791364 M * Bertl same for timer.c? 1139791379 M * Bertl rest looks good! 1139791388 M * daniel_hozac timer i was unsure of. 1139791396 M * daniel_hozac the rest i checked, IIRC. 1139791400 M * Bertl rip it out and test compile :) 1139791437 M * daniel_hozac kernel/sys.c has find_task_by_pid. as does kernel/signal.c 1139791458 J * DukeGangsta Gangsta@i577B8109.versanet.de 1139791482 M * Bertl welcome DukeGangsta! 1139791484 M * DukeGangsta hi 1139791504 M * Bertl daniel_hozac: yeah, I was just wondering, where did the vs_cvirt.h come from (in signal.c) before? 1139791553 M * daniel_hozac yeah, it seemed strange to me too. 1139791653 M * DukeGangsta i'm searching for some info to get 2 network-interfaces running... the vserver (webserver) should have 2 IPs: one official internet-IP, and one 192.168.0.*.... the last interface should only "see" some other vservers with 192.168.0.* IPs, and should not be visible to outside (e.g. the other machines in that room)... 1139791693 M * DukeGangsta so it's a kind of localloopback, and shouldn't be bound to eth0 of the host... 1139791703 M * Bertl well, just assign two ips to the guest, first the external, then the internal 1139791714 M * Bertl you can put one on eth0 the other on eth1 1139791831 M * DukeGangsta but then both IPs are visible to other machines in that LAN? from "outside" of the machine there should only be the official IP-adresse, not the 192.168. adress... also should the mysql-servers not be accessible from outside, only from the other web-vserver 1139791869 M * derjohn DukeGangsta, solve that like you would solve it if you had no vserver, i.e. iptables. 1139791939 M * DukeGangsta isn't it easier to bind those interfaces to a loopback-device or so? 1139791969 M * derjohn DukeGangsta, you could even reach an IP on the loopback, if you bind a real ip to "lo". 1139791984 M * derjohn (you would need a route though, but there is no "block" 1139791986 M * daniel_hozac Bertl: hmm, the infiniband parts of http://daniel.hozac.com/tmp/delta-2.6.14-vs2.0.1-pre3-accounting-fix.patch seems to not be present in 2.1.0.11. any particular reason? 1139791986 M * derjohn but ... 1139792007 M * derjohn You can give the guests 127.0.0.2, 127.0.0.3 etc 1139792046 M * derjohn instead of 192.xx 1139792054 M * derjohn but: YMMV :) 1139792121 M * DukeGangsta hmm.. what if i bind the IP to eth0, and there is another guy in that LAN and does the same? ;) 1139792157 M * derjohn DukeGangsta, that would provoke the same fun as it would, if you had no vservers ... :) 1139792174 M * derjohn you man the 127.0.0.2 ? 1139792190 M * DukeGangsta yes.. or the 192.* 1139792201 M * derjohn well, the other hosts dont care much, because they usually have 127.0.0.0/8 on localhost 1139792226 M * derjohn I dunno if linux would progagate 127er in arp ... does anyone know? 1139792268 M * DukeGangsta intelligent switches should block that address-range i think... 1139792268 M * derjohn DukeGangsta, if you bind 192er on lookback, the outside world will no see that without 'tricks' 1139792292 M * derjohn ie, ip ro add 192.168.0.0/24 via 1139792300 M * derjohn from the other hosts side 1139792326 M * derjohn so: an iptables -j DROP is the best way 1139792345 M * DukeGangsta yes, thats my question, if that is possible, so that those adresses are not visible to the ourside... i think i have to try to bind it to loopback.. 1139792395 M * derjohn DukeGangsta, iptables is the more secure way 1139792410 M * derjohn afraid of iptables? 1139792456 M * DukeGangsta I didnt use it in the last years (security of obscurity doesnt work ;)) 1139792467 M * derjohn ipatbles -I INPUT -d 192.168.0.0/24 -J DROP and maybe ipatbles -I FORWARD -d 192.168.0.0/24 -J DROP 1139792498 M * derjohn then you vservers could still cimmunicate 1139792505 M * derjohn with each other 1139792520 M * Bertl daniel_hozac: hmm, doesn't seem to apply here, will check ... 1139792547 M * daniel_hozac Bertl: which one? 1139792555 M * Bertl infiniband part 1139792569 M * daniel_hozac oh, well, it is two kernel versions old. 1139792597 M * daniel_hozac hmm, how could i remove vs_cvirt.h from kernel/sched.c... 1139792996 M * Skram what class is .2 => .63 ? 1139793021 M * Bertl Skram: huh? 1139793026 M * Skram ip addresses 1139793031 M * Skram like /28... /29? 1139793033 M * Skram whatever. 1139793038 M * Bertl class Skram :) 1139793051 M * Skram whats it called. 1139793053 M * Skram meh. 1139793063 M * Bertl well, .0 - .63 is /26 1139793070 M * Bertl no special name for that 1139793099 M * Skram so /26.. Okay. 1139793108 M * Skram how do they come up with these names. 1139793118 M * Bertl you mean /26? 1139793123 M * Skram yeah. 1139793134 M * Bertl that means 26 bits are the network address 1139793141 M * Bertl leaves 32-26 bits for ips 1139793170 M * Skram /24 is one ip, right? 1139793176 M * Skram or something like that. 1139793184 M * Bertl no, thats 256 ips 1139793196 M * Skram Oy, I must be going crazy 1139793223 M * ebiederm /32 is one ip. 1139793235 M * Skram Alrigjt 1139793241 M * Bertl Skram: play around with http://jodies.de/ipcalc 1139793258 M * Skram :) kk 1139793293 M * mnemoc http://svn.exactcode.de/t2/trunk/package/network/iproute2/ipnm2nwbc.c :) 1139793298 M * DukeGangsta arg...i forgot that the vserver must have access to the internet (updates etc).. so i have to configure natting ;( 1139793302 M * daniel_hozac Bertl: arch/x86_64/ia32/syscall32.c:syscall32_setup_pages is missing a vx_vmpages_add. 1139793421 M * derjohn Skram, apt-get install sipcalc ; sipcalc 192.168.0.1/24 1139793441 M * derjohn DukeGangsta, yes, I put the rule you need on the wiki 1139793442 M * DukeGangsta i bound 192.168.0.101 to lo... i can ping all hosts ip-adresses from the guest, and from host i can ping the guest.... now it is interesting how to setup nat from eth0 to lo on the host 1139793449 M * derjohn look for hint 1139793456 M * derjohn hints from john .... 1139793458 M * Bertl derjohn: how is sipcalc different from ipcalc? 1139793475 M * derjohn Bertl, IMO sipcalc was the coolest one. 1139793492 M * derjohn Bertl, I think it differs in the output. 1139793504 M * Bertl ah, okay, so it's the coolnes factor :) 1139793633 M * ebiederm Coolness is definitely the most important characteristic in the winter! 1139793646 M * ebiederm Or is that in the summer? 1139793649 M * ebiederm :) 1139793660 M * derjohn Bertl, just tried ipcalc, very practical so see the binary form := 1139793660 M * derjohn ) 1139793698 M * derjohn ebiederm, of course, pengiuns love the coolness :) 1139793715 M * derjohn *penguins 1139793756 M * ebiederm That does explain why linux is so cool! 1139793826 M * derjohn ebiederm, you should tell Linus, maybe he doesnt know :) [if you have a T-level if 1] 1139793832 M * derjohn if=of 1139793856 M * DukeGangsta derjohn: do you have some minutes in query, i dont want to spam here ;) 1139793879 M * derjohn uf, I have to lease for bed ... 1139793880 M * ebiederm T-level? 1139793891 M * derjohn but you may ask 3 Questions :) 1139793893 M * derjohn hurry ! 1139794006 M * derjohn ebiederm, The "Torvalds Level" .. Linus is T-Level=0, Alan C. at al work wirk Linus directly -> T-Level=1 and so on ... 1139794036 M * ebiederm Ah.. I had not heard of that classification. 1139794102 M * derjohn ebiederm, I have been told that the large IT companies look for T-Level 0-2 employees .... 1139794159 M * ebiederm Given how Andrew works it is hard to get a T-level below 2..... 1139794463 M * derjohn a human firewall ? 1139794542 M * ebiederm s/below/above/ ... Sorry. 1139794720 M * derjohn ebiederm, that depends on the point of view :) below in a sense of a tree/radix or in numers :) 1139794725 M * derjohn numbers 1139794787 M * ebiederm Basically I am asserting that largely tree height == 2 1139794831 M * derjohn ok :) 1139794976 M * Bertl ebiederm: counting the node or edge levels *G* :) 1139794988 J * Aiken_ ~james@tooax6-208.dialup.optusnet.com.au 1139795042 A * derjohn imagines Linus' edge hangung down on ... *argh* 1139795052 M * ebiederm Usually the worst case is developer talks to Andrew. Andrew talks to Linus. Or it is developer talks to subsystem maintainer. Subsystem maintainer talks to Linus. 1139795145 M * Bertl no, the worst case is: developer talks to subsystem maintainer. subsystem maintainer ignores it :) 1139795146 M * derjohn ah, so gets rid of whiners? 1139795155 M * daniel_hozac Bertl: http://daniel.hozac.com/vserver/delta-infiniband-fix01.diff 1139795167 M * Bertl daniel_hozac: tx, but I already fixed that 1139795173 M * daniel_hozac ok. 1139795181 M * daniel_hozac got the x86_64 thing too? 1139795187 M * Bertl yup 1139795349 Q * Aiken Ping timeout: 480 seconds 1139795352 M * daniel_hozac did you see http://daniel.hozac.com/vserver/delta-meminfo-fix01.diff yesterday? 1139795357 M * Bertl here are the duplicate #includes from mainline: 1139795359 M * Bertl https://plm.osdl.org/plm-cgi/getpatch?id=4980 1139795372 M * Bertl daniel_hozac: looking now 1139795375 M * derjohn n9 everyone 1139795377 M * derjohn n8 1139795379 M * derjohn *lol* 1139795391 M * Bertl night derjohn! 1139795430 M * Bertl daniel_hozac: okay, looks fine, will add it 1139795666 M * Skram can i just get a final answer: on a soft limit, if no tokens and everything is defined, what happens when they go over? 1139795688 M * Bertl nothing 1139796001 M * Bertl daniel_hozac: could you rebase your vs_pid.h against http://vserver.13thfloor.at/Experimental/delta-include-clean01.diff 1139796026 M * daniel_hozac sure. 1139796344 J * click_ click@ti511110a080-4978.bb.online.no 1139796450 Q * click Ping timeout: 480 seconds 1139796677 M * daniel_hozac Bertl: missing #include in kernel/fork.c? 1139796710 Q * brc Ping timeout: 480 seconds 1139796743 M * daniel_hozac http://daniel.hozac.com/vserver/delta-vs_pid-clean02.diff 1139797132 M * Bertl daniel_hozac: right, tx! 1139797861 M * Bertl nox: btw, forgot to mention that the latest stable patch (exp) already contains bme now 1139797897 M * daniel_hozac 2.0.1.3? 1139798081 M * Bertl hmm, no :/ 1139798114 M * Skram free -m shows the vps using 1 or 2 more megabytes then /proc/virtual, weird? 1139798160 M * daniel_hozac what does /proc/virtual say while free -m is running? 1139798171 M * Skram That's probaby why. 1139798172 M * Skram ;) 1139798227 M * Skram how do we give a user their own swap if they go over the hard limit? 1139798287 M * daniel_hozac soft limit == start of swap, hard limit == end of swap, AIUI. 1139798381 M * Skram okay 1139798383 M * Skram uhmm 1139798389 M * Bertl ram + sawp = hard 1139798394 M * Bertl ram = soft :) 1139798398 M * Skram okay that makes more sense 1139798417 M * Skram hard - soft = swap? 1139798426 M * Bertl yup :) 1139798430 M * Skram okay 1139798431 M * Skram sorry. 1139798446 M * Bertl np 1139798447 M * ebiederm hmm. If anything gets seriously swapping doesn't hard == swap? Or am I missing something? 1139798487 M * daniel_hozac and if nothing is swapping, hard == RAM ;) 1139798512 M * ebiederm ? nothing is swapping or no swap is configured? 1139798513 M * Bertl hmm, no, actually it's virtual swap and virtual ram 1139798541 M * ebiederm Bertl: Ok. For virtual swap and virtual ram I haven't a clue how it is put together. 1139798557 M * Bertl ebiederm: no real swap handling nor other stuff, just faked numbers for the guests) 1139798575 M * Skram so rlimits/rss.soft == soft, .hard == hard? 1139798594 M * Bertl rlimits/rss.soft = soft rlimits/rss.hard = hard 1139798594 M * Skram RSS == hard? 1139798623 M * Bertl if you just set rss, then you will set hard and soft to the same value 1139798640 M * Skram so if i want to do soft, i cant do vlimit --xid XX --rss? 1139798649 M * Bertl (which basically looks like turning off swap for the guest) 1139798669 M * Skram ill just have to restart it 1139798683 M * Bertl you'd want vlimit --xid -S --rss ... 1139798786 M * Skram kk one moment 1139798797 M * Skram not good, taking a long time to stop this sucker. 1139798803 M * Skram hopefully i wont get error galores. 1139798816 M * Skram hercules / # vserver jtungate01-vs1-01 start 1139798816 M * Skram hercules / # vserver jtungate01-vs1-01 enter 1139798816 M * Skram Killed 1139798817 M * Skram eh? 1139798829 M * Skram VM: killing process bash 1139798832 M * Skram hmm. 1139798984 M * Skram free -m shows the hard limit. 1139799000 M * Skram how can i make it show the soft limit, and under swap it show the hard limit. 1139799021 M * Bertl that's what it does, given that you use a recent kernel patch 1139799040 M * Skram you mean given i didnt use a recent patch? 1139799064 M * Bertl well, depends on what you relate my answer to :) 1139799071 M * Skram eh 1139799073 M * Bertl but to clarify_ 1139799091 M * Bertl recent patches do show the soft as memory available 1139799099 M * Bertl and hard-soft as swap available 1139799112 M * Bertl s/available/total/ 1139799113 M * Skram hmm okay 1139799119 M * Skram does a patch require a reboot? 1139799133 M * Skram # Linux kernel version: 2.6.14-vs2.0.1-gentoo 1139799136 M * Bertl to some extend, yes, at least a kexec 1139799137 M * Skram thats what I am on. 1139799158 M * daniel_hozac that doesn't even support soft limits 1139799164 M * Skram What! 1139799172 M * Bertl yeah, you probably want to wait until hollow updates that ... 1139799199 M * Skram so if i set the soft to 64 and hard to 128, they are getting 128 of the RAM? 1139799211 M * Bertl yup 1139799216 M * Skram CRAP. 1139799233 M * Skram a server upgrade requires a reboot or what? 1139799257 M * Bertl as I said, usually, but at least a kexec (which is basically a soft-reboot) 1139799303 M * Skram well... the other admin changed some things in this kernel and i am afraid that it will not boot up. 1139799306 M * Skram :( 1139799333 M * Bertl well, won't hurt to contact him before ... 1139799342 M * Skram contact whom? 1139799350 M * Bertl the 'other' admin 1139799355 M * Skram the old admin? not around anymore, and kernels arent my forte, at all. 1139799400 M * Bertl hmm, well, I'd suggest to talk to hollow, he probably can tell you how to upgrade the kernel in a reasonably secure way 1139799446 M * Skram okay 1139799449 A * Skram sings for Hollow.. 1139799732 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.0.2-rc1.diff 1139799737 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.1.1-rc1.diff 1139799754 M * Bertl (next rc will be against 2.6.15.4 (or whatever is latests) 1139799805 M * Skram Right... 1139799813 M * Skram well, Ill talk to Hollow about upgrading. 1139799846 M * Bertl yup, should be pretty easy if you know how ... 1139799946 M * Skram *shrug* 1139799947 M * Skram yeah 1139799952 M * Skram as i said, i dont like kernel stuff 1139800904 M * Bertl daniel_hozac: I'll cancel 4977 and 78, as I do not expect any new clues from there, okay? 1139801277 M * daniel_hozac ok. 1139801363 M * Bertl updated the cross diffs first ... 1139801595 M * Bertl daniel_hozac: 4981,82 are the new rc1 diffs ... 1139801740 M * daniel_hozac ok, cool. 1139801896 M * daniel_hozac hmm 1139801899 M * daniel_hozac ./fs/nfs/dir.c is missing linux/mount.h 1139801899 M * daniel_hozac ./fs/hfsplus/ioctl.c is missing linux/mount.h 1139801910 M * daniel_hozac (2.0.2-rc1) 1139801973 M * Bertl well, it's an integral process :) 1139801987 M * daniel_hozac (output from http://daniel.hozac.com/vserver/find-problems.sh it's really slow, and has some bogus output) 1139802003 M * Bertl argl .. just deleted them 'accidentially' 1139802023 M * daniel_hozac hehe. 1139802033 M * Bertl (the osdl interface really sucks) 1139802113 M * Bertl okay, as nobody else is using it :) they will be 4983 and 84 1139802252 M * daniel_hozac heh. 1139802436 M * ebiederm osdl interface? 1139802446 M * Bertl plm 1139802464 M * ebiederm I'm not familiar with that. 1139802495 M * Bertl basically a cross compiling test-bed for patches against mainline 1139802507 M * ebiederm ok. 1139802525 M * Bertl I use it when I do not want to hog my machine with cross compiling ... 1139802685 M * daniel_hozac well, i'm going to get some sleep now. good night! 1139802688 M * Bertl daniel_hozac: hmm, it seems I also missed the atomic_long stuff in 2.6.15-vs2.1.1* 1139802710 M * Bertl somebody had a patch there, was it you or Doener? 1139802723 M * Bertl daniel_hozac: anyway, have a good night! 1139802729 M * daniel_hozac hmm, well, it wasn't me. 1139802759 M * Bertl np, thanks for your help today! 1139802768 M * daniel_hozac my pleasure! 1139802770 M * Skram going already? 1139802772 M * Skram heh 1139804237 M * mugwump ok, I'm not convinced by pg yet, but it's promising 1139804292 M * mugwump its rebase command failed on me :) 1139804536 M * Skram do images need to be redone when upgrading kernel? 1139804557 M * Bertl images == guests ? 1139804592 M * mugwump the bzImage does ;) 1139804888 M * ebiederm It is a well know fact that the implementation of /proc has a few problems?? 1139804918 M * ebiederm I just discovered on readdir the termination condition of the loop is to do a noop read through of the entire process list. 1139804932 M * ebiederm Insane! 1139804951 M * Bertl ebiederm: well, yeah, proc is broken :) 1139804984 M * Bertl dunno if it is well known though ... 1139805001 M * ebiederm Maybe it is just well known that /proc is ugly... 1139805254 A * Skram wished that /proc/virtual/*/cvirt said the vserver-name, not the hostname that the user sets.. 1139805507 M * Bertl we will add the 'context' utsname field in /proc/virtual//info 1139805540 M * Bertl but the kernel doesn't know the guest name, only userspace knows 1139805675 M * Bertl try vserver-info NAME 1139807992 M * Skram hercules / # vserver-info 10009 NAME 1139807992 M * Skram hercules / # 1139807993 M * Skram that? 1139808008 M * Bertl is it running? 1139808012 M * Skram indeed. 1139808030 M * Bertl legacy guest? 1139808038 M * Skram Vserver 'jtungate01-vs1-01' is running at context '10009' 1139808038 M * Skram Number of processes: 20 1139808038 M * Skram Uptime: 02:31 1139808154 M * nox morning Bertl! 1139808164 M * Bertl hey nox! 1139808186 M * nox got u right? so bme is already included or is there a newer? 1139808207 M * nox gonna try (; 1139808213 M * Bertl 2.0.2-rc1 has bme 1139808224 M * Bertl but there will be an rc2 shortly 1139808251 M * nox ahh great rc2 today? 1139808268 M * Bertl I'd say so 1139808282 M * SuperLag Bertl: I'm coming into the conversation late. What is BME? 1139808299 M * Skram Hiya, SuperLag 1139808314 M * SuperLag Skram: good evening, sir. 1139808315 M * nox bind mount extension SuperLag 1139808318 M * Skram ;) 1139808350 M * nox so mount - bind behave like you would expect 1139808367 M * nox s/-/-o/ 1139808630 M * Bertl yup, basically fixes a mainline issue 1139808870 Q * DukeGangsta Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1139811619 M * Skram Yeah. 1139811620 M * Skram heh 1139811627 M * Skram Arg, bio test tomorrow 1139812102 M * Hollow morning folks 1139812109 M * Bertl morning Hollow! 1139812139 M * Skram Hollow: !!! 1139812142 M * Skram hey, man. 1139812173 M * Hollow Bertl: would be nice if you could fix the double entry in /proc/virtnet for the highest nid.. 1139812185 M * Hollow (wrt to rc*) 1139812197 M * Bertl ah, okay, will look into it 1139812224 M * Hollow Skram: what's your problem with kernel upgrading? 1139812238 M * Hollow make menuconfig; make; cp bzImage /boot/foo; reboot 1139812257 M * Skram uhmm, just need to make sure its all working, and upgrading to a new kernel, im new to kernels 1139812276 M * Skram i have a feeling the old admin screwed this old kernel or something and it wont boot on reboot. 1139812279 M * Skram i dont know. 1139812280 M * Hollow well, then you probably go best with your old kernel config 1139812286 M * Bertl Hollow: is that really the gentoo way of updating a kernel? 1139812293 M * Skram heh 1139812300 M * Skram its 12:30am over here. 1139812305 M * Skram i dont know how much longer ill be on 1139812314 M * Hollow Bertl: well, you install new sources, point /sur/src/linux to it, build it and install it in grub/lilo 1139812345 M * Hollow (you can skip the /usr/src/linux symlink of course) 1139812353 M * Hollow just build it anywhere 1139812397 M * Hollow Skram: so, i'd suggest you start with: cd /new/sources; zcat /proc/config.gz > .config; make oldconfig; make 1139812435 M * Skram Hollow: can we talk about this later maybe? 1139812436 M * Skram yeah. 1139812441 M * Skram i get the basic point however. 1139812454 M * Skram i just was thinking about getting some cash and paying to make sure it would work. 1139812457 M * Skram :( 1139812462 M * Hollow http://gentoo-wiki.com/HOWTO_Compile_a_Kernel_Manually 1139812465 M * Skram I do kernels on my laptop, but a server is another thing 1139812469 M * Skram Hollow: yeah.. 1139812487 M * Hollow what's your current version? 1139812548 M * Skram lemme look 1139812561 M * Skram # Linux kernel version: 2.6.14-vs2.0.1-gentoo 1139812571 M * Skram id really like soft.rss support. 1139812577 M * Hollow guess you should be fine with make oldconfig then 1139812588 M * Skram hrmm 1139812592 M * Skram ill see when i have more energy 1139812600 M * Skram school sucks, man,. 1139812614 M * Hollow yeah, have to leave soon too ;) 1139812629 M * Skram i have school in 6 hours 1139812632 M * Bertl well, let me say it again, school is the best chance to learn something :) 1139812636 M * Skram well need to be ready in 6 hours 1139812643 M * Skram Bertl: I agree, and understand ;) 1139812676 A * Hollow sighs :) 1139812680 M * Hollow Bertl is always pro-school 1139812719 M * Skram I hear ya. 1139812725 M * Skram Hollow, in college? 1139812727 M * Bertl trust me, I know what I'm saying .. I wish I had learned more in school, would have saved me a lot of time later ... 1139812749 M * Bertl (and I'm pretty sure I already leared a lot there :) 1139812760 M * Skram Bertl: Yeapps. I go to the hardest school in central texas, so I have my plate full ;) 1139812821 M * Hollow well, in my opinion, you should only learn it, when you need it, and currently i really don't need much of the crap i got told in school 1139812849 M * Hollow if i'm into literature in 40 years, i would have forgotton everything anyway 1139812854 M * Bertl Hollow: the thing is, when you need it, it will 10 times as hard to learn than now :) 1139812894 M * Hollow well, that's life 1139812895 M * Hollow :) 1139812901 M * Hollow you get slower and slower every day 1139813053 M * Hollow Bertl: btw.. the start command is almost working, in C, hooray 1139813063 M * Bertl cool 1139813070 M * ebiederm Hollow :for what it is worth I believe I have used everything I learned in school. 1139813146 M * Skram biology? 1139813154 M * Skram im taking college level biology.. and im not in college 1139813203 M * Hollow well, i don't say school is useless, any many things would even be interesting to me, but if someone in front just talks to you, i'm off 1139813297 M * Hollow and, additionally, i'm not going very often to school, and i'm still as good as many of those who learn their ass of, so i guess it's ok ;) 1139813325 M * SuperLag Hollow: quit blowing your horn :) 1139813351 M * Skram heh 1139813378 M * Hollow SuperLag: well, i'm not good in school, i just said as good as those who learn.. ;) 1139813402 M * SuperLag Hollow: I have a different theory on that. 1139813407 M * SuperLag I'm VERY pro-education. 1139813419 M * Hollow ic.. i'm really in the wrong channel here :) 1139813420 M * Skram OY. 1139813426 M * Skram hah. 1139813435 M * Skram So, SuperLag, wanna help me study for this bio test? 1139813448 M * SuperLag I think the problem, more often than not, lies with outmoded/obsolete teaching methods. That, and the fact that EVERY student *can* be good in school. 1139813462 M * Hollow yeah, i agree 1139813576 M * SuperLag Hollow: the old and stale... "teacher lecture > student listen" model is just that. old and stale. It definitely has its place. But not 100% of the time. And get those #$$##@$#@!!$@#$#!@ out of their research rooms and BACK into the class room, and get RID of the teacher assistants. They're crap. 1139813607 M * SuperLag Hollow: most of it has to do with how people learn. And the teacher needs to learn and master how to engage EVERY type of student. 1139813659 M * SuperLag Hollow: the problem is... they either don't understand the idea of multiple intelligences... or they do understand it, and just don't care. If I were the dean of such an institution, those profs would be the first ones I'd throw out on their asses 1139813686 M * Skram I totally agree. 1139813697 M * SuperLag some people learn best by hearing a lecture. 1139813700 M * SuperLag Not me. 1139813707 M * Hollow yeah, completely right.. as i said.. many things would interest me more, if they weren't told me like "blah blah blah, ah.. who cares anyway.. let's go into break" 1139813711 M * SuperLag I want you to draw me a picture. 1139813723 A * Skram draws a picture. 1139813733 M * SuperLag Skram: you know what I mean, right? 1139813740 M * Skram Indeed. 1139813745 M * SuperLag conceptualizing things makes sooooo much more sense to me 1139813757 M * SuperLag if I can see a big picture, things start to make sense 1139813817 M * Skram Understandable. 1139813835 M * Hollow SuperLag: like http://home.xnull.de/misc/vudesign.jpg ? ;) 1139813863 M * Skram nice. 1139813914 M * Hollow yeah, and yesterday vcc got its start command 1139813921 M * Hollow with working rbind, woho 1139813923 M * SuperLag Hollow++ 1139813932 M * SuperLag yes. just like that. 1139813947 J * cauchy ~Administr@61.235.82.18 1139813949 A * Skram claps. 1139813954 M * SuperLag some of those concepts are way over my head though. That's very low level. Stuff that I've never touched. 1139813965 M * SuperLag just httpd :) 1139813970 M * Skram haha 1139813971 M * Hollow :) 1139813977 M * SuperLag and all the stuff on the remote client 1139814005 M * Hollow well, vcd does not exist yet 1139814016 M * Skram yeah, i was about to say, eh? 1139814022 M * SuperLag but that kind of illustration really helps a person like me learn things MUCH better than hearing a dry lecture 1139814037 M * Skram i think a combo is best. 1139814048 M * Skram i mean, you obviously dont understand it without words. 1139814052 M * Hollow yeah, and it helps me to not lose my development concept, and even make it clear before starting with it 1139814058 M * Skram Right on. 1139814062 M * Skram Hollow++; 1139814078 M * SuperLag Skram: agreed. 1139814099 M * Hollow the same for source code documentation *ahem* 1139814121 M * Hollow i really try to document my source, even if it's boring 1139814146 M * Skram we appreciate it Hollow 1139814147 M * Skram I do. 1139814183 M * Hollow http://home.xnull.de/work/vserver/vsevrer-utils/src/vcc/start.c btw, if someone is interested 1139814188 M * Hollow not in svn yet 1139814192 M * Hollow and typo 1139814205 M * Hollow http://home.xnull.de/work/vserver/vserver-utils/src/vcc/start.c 1139814227 J * Snarfles Snarfles@ESS-p-144-138-69-186.mega.tmns.net.au 1139814240 P * Snarfles 1139814456 P * cauchy 1139815618 M * Hollow ok, off to school now.. cu later 1139815627 M * Bertl cya! 1139815635 M * Skram Tootles. 1139815642 M * Skram School in 5-6 hours. 1139816365 M * SiD3WiNDR tsk tsk 1139816379 M * SiD3WiNDR Sam's mail did not include CET times! 1139816390 M * SiD3WiNDR Noone important lives on london however! :> 1139816395 A * Skram shakes his head at a cause he knows nothing about. 1139816396 M * Skram hrh 1139816399 M * Skram *heh 1139817317 J * meandtheshell ~markus@85-124-36-202.dynamic.xdsl-line.inode.at 1139817731 J * tudenbart ~willi@xdsl-81-173-174-37.netcologne.de 1139818150 Q * dothebart Ping timeout: 480 seconds 1139819037 J * burge lucifer@foocode.net 1139819048 M * Bertl welcome burge! 1139819058 M * burge hi there 1139819093 M * burge i just stumbled across the website and thought vserver sounded like the coolest thing around 1139819107 M * Bertl well, it is :) 1139819124 M * burge hah I'm sold 1139819164 M * burge got a question though - can all the virtual server instances utilise the same network device with different IPs? 1139819179 M * Bertl yep, networking happens on the host 1139819203 M * Bertl i.e. you assign some ips to each guest, and it will be limited to them 1139819204 M * burge I have a headless fc4 box that I would want to run a second vserver OS on and would like to be able to access them both with VNC/NX 1139819211 M * burge ah ok 1139819333 J * Smutje_ ~Smutje@xdsl-87-78-6-153.netcologne.de 1139819367 M * Bertl burge: there is a page discussing the various X variants on the wiki, you might have a look there 1139819459 Q * Smutje Ping timeout: 480 seconds 1139819459 N * Smutje_ Smutje 1139820342 M * ebiederm yeah!!! I have removed the giant enumeration from fs/proc/base.c 1139820524 M * ebiederm Have a good night folks. 1139820531 N * ebiederm ebiederm_zZ 1139820621 Q * shedi Quit: Leaving 1139820753 M * Bertl ebiederm_zZ: night! 1139821076 M * burge http://vserver.strahlungsfrei.de/tiki-index.php?page=FrequentlyAskedQuestions 1139821091 M * burge am I looking on the right forums? also, that page seems to have been defaced :( 1139821122 M * Bertl http://linux-vserver.org/ 1139821213 M * Bertl sec, searching ... 1139821276 M * Bertl http://linux-vserver.org/Vservers+and+X 1139821306 M * Bertl it seems not linked yet, or at least I didn't find the link 1139821401 Q * Smutje Quit: leaving 1139821558 J * prae ~prae@ezoffice.mandriva.com 1139821660 Q * harry Ping timeout: 480 seconds 1139823455 J * harry ~harry@d515321D1.access.telenet.be 1139823839 J * shedi ~siggi@tolvudeild-204.lhi.is 1139824133 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15.4-vs2.0.2-rc2.diff 1139824138 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15.4-vs2.1.1-rc2.diff 1139824143 M * Bertl (please test) 1139824167 M * Bertl okay, folks, I'm off to bed now ... have a good whatever everyone! cya tomorrow! 1139824175 N * Bertl Bertl_zZ 1139824192 M * nox sleep well and thx Bertl_zZ 1139824350 M * nox are there no more incremental patches on kernel.org? 1139825291 J * NetAsh ~foo_bar@193.219.160.108 1139825298 M * NetAsh hello 1139825817 Q * NetAsh Quit: 1139826083 Q * prae Quit: Execute Order 69 ! 1139826090 J * anonc ~anonc@staffnet.internode.com.au 1139826710 J * michal ~michal@www.rsbac.org 1139826720 N * michal Guest2153 1139826725 N * Guest2153 michal_ 1139827152 J * h01ger ~holger@bone.digitalis.org 1139827354 J * wally ~homebase@62.116.83.242 1139827416 M * h01ger huhu - on my vservers (running debian/sarge on a debian host running 2.6.15 with kernel-patch-vserver 2.0.1-1) every morning at 06:25:02 i get a logcheck mail "su[3113]: pam_open_session: Permission denied" like in http://list.linux-vserver.org/archive/vserver/msg10043.html - but opposed to that mail, i dont have setlimit problems and su works on the commandline 1139827442 M * h01ger at 06:25 cron.daily is run, if i run it manually, i get no errors... ?! 1139827444 M * h01ger any ideas ? :) 1139827582 J * Hmmmm ~Hmmmm@221.135.51.19 1139828028 M * Hmmmm hi, anyone /home? 1139830374 Q * Aiken_ Ping timeout: 480 seconds 1139831188 Q * burge Quit: Sleep? Oh boy! That's where I'm a viking! 1139831663 Q * Hmmmm Quit: Leaving 1139832568 J * s25 ~s25@86.42.17.163 1139832927 J * mkhl mkhl@200-148-41-143.dsl.telesp.net.br 1139832981 M * nox Bertl_zZ: 1. feedback 2.6.15.4-vs2.0.2-rc2 running and bme seems to work 1139833066 M * s25 hi all im about to try and isntall Vserver on an ubuntu box i have to use to host my site and a IRC server :D i am real excited :D 1139833144 M * nox s25: that both works fine 1139833207 M * s25 :) 1139833225 M * nox Bertl_zZ: is there a possibility to realize bind mount without the need of restarting the vserver? 1139833308 M * s25 how do you add things to sources.list? i did a search on the wiki and got nothing and googles results didnt seem aplicable 1139833407 M * nox s25: with vi (; 1139833560 M * nox s25: i have an extra vserver running apt-cache to disburden to mirrors 1139833571 M * nox s/to/the 1139833930 M * s25 how do i save the file once i hvae done that? 1139833959 M * s25 and edited them ect 1139834008 M * h01ger s25, vserver foo exec "echo 'deb http://foo ba ba' >> /etc/apt/sources.list" 1139834024 M * nox s25: you should be able to use an editor before you start with linux-vserver 1139834049 M * s25 i can just i dont see how you save the files after you have added the sources 1139834201 M * s25 if i just exit will the file be saved? 1139834423 M * s25 wait it wont let me save there says permision denied 1139834433 M * s25 will i ahve to sudo and edit again? 1139834591 M * nox s25: this is not vserver related look at the filepermissions 1139834651 M * s25 ok sorry lads will keep my questions strictly vserver from now on 1139834742 M * Hollow nox: vnamespace -e -- mount --bind /foo /vservers//bar 1139835724 J * Duckx ~duckx@195.75.27.158 1139835733 M * nox Hollow: great! 1139835814 J * DuckMaster ~duckx@195.75.27.158 1139837077 M * s25 ok i did apt-get update and it all seemed to go fine but then at 99% on connecting to ubuntu.uni-klu.ac.at i got GPG error "buntu.uni-klu.ac.at breezy Release:The following signatures couldnt be verified because the public key is not available 1139837085 M * s25 will this affect my installation? 1139837633 J * BartVB ~BartVB@84.35.54.120 1139837775 M * BartVB vserver seems to be working like a charm but I have a problem with apache2, I can't get it to start up on a vserver because it can't bind to the specified IP/port 1139837793 M * BartVB the weird thing is that I can't find anything else on the entire server that's doing something with port 80 on any address? 1139837814 M * BartVB I get: Starting web server: Apache2(98)Address already in use: make_sock: could not bind to address 82.161.230.66:80 1139838066 M * eyck hmmm 1139838067 M * eyck cat /proc/sys/kernel/random/entropy_avail 1139838067 M * eyck 0 1139838080 M * eyck any idea what to do with that? 1139838129 M * BartVB not really but doesn't seem to be normal 1139838138 M * BartVB it returns 3203 on my vserver 1139838158 M * BartVB what provides that service? IIRC it's an optional part of the kernel 1139838166 M * Loki|muh eyck: yeah, common problem with >=2.6.12 1139838170 M * Loki|muh sucks :( 1139838225 M * eyck Loki|muh: I'm on 2.4, but how can I feed it some more entropy? 1139838237 M * eyck Loki|muh: suppose I've got some in a file? 1139838271 M * Loki|muh "find . >> /dev/null" generates some entropy 1139838297 M * Loki|muh and if its a workstation grab the keyboard an hammer it a while ;) 1139838303 M * Loki|muh or move the mouse 1139838329 M * eyck find doesn't seem to do anything about feeding entrophy :( 1139838347 M * Loki|muh ups "find /" 1139838357 M * Loki|muh disk activity should generate entropy 1139838378 M * Loki|muh "du /" should work, too 1139838420 M * eyck hmm, hmm, there are rumours that scsi disks won't work with entrophy gathering.. :( 1139838730 M * cehteh there is a audioentropyd gathering entropy from cheap noisy microphones .. nice for servers since you doesnt usually use the micro anyways .. just put it beneath a fan 1139838755 M * cehteh plus: you can monitor whats going on at ur server :P 1139838790 M * cehteh and some modern chipsets have hardware random generators 1139839091 M * eyck cehteh: and how would I feed my entrophy to /dev/random? 1139839098 M * eyck suppose I have some in a file? 1139839113 M * cehteh cat file >/dev/random 1139839165 M * eyck hmm, tried that, but it doesn't work.. entrophy is at '0' 1139839219 M * cehteh strange 1139839236 M * cehteh do you have something running which is sucking entropy out of the pool? 1139839298 M * eyck yeah,...well..network for one, and alot of java instances... 1300 when I counted with 'lsof|grep /dev/random|wc' 1139839331 M * cehteh mhm even when i cat /dev/random >/dev/null it doesnt go down to 0 here 1139839388 J * lilalinux ~plasma@80.69.35.186 1139839405 M * eyck well, I did a little checking..and I've got 3 machines with random_avail at '0'... 1139839406 M * cehteh dunno .. looks like something is wrong there . well i have 2.6 1139839416 M * eyck all had a lot to do with encryption.. 1139839479 M * cehteh tried to stop this apps and see if the pool fill up then? 1139839501 M * eyck I took /dev/random away from them... 1139839507 M * eyck still '0' 1139839515 M * cehteh some other problem them 1139839517 M * cehteh then 1139839522 M * cehteh dunno what 1139839728 N * Bertl_zZ Bertl 1139839732 M * Bertl morning folks! 1139839782 J * x ~Angel@Cable-82-216.topallnet.ro 1139839786 M * x hi 1139839791 N * x angel4u 1139839831 M * eyck Bertl: how you got any idea what to do when random_avail goes to '0' ? 1139839833 M * Bertl welcome angel4u! 1139839844 M * Bertl eyck: make some noise :) 1139839861 M * Bertl eyck: well, you have to refill the entropy pool 1139839901 M * Bertl - keyboard/mouse activity 1139839914 M * Bertl - interrupts and disk elevator 1139839930 M * Bertl - write to /dev/random (with random data) 1139839945 M * cehteh Bertl: having it constantly at 0 is somewhat strange i think 1139839962 M * Bertl means something is constantly draining random data 1139839996 M * cehteh well .. i tried here the same with a handful processes but at least you can see that it fills up 1139839998 M * Bertl could be heavy ipv6 or strong cryptography (tunnels) 1139840066 M * cehteh cat /dev/urandom >/dev/random .... not optimal, but that should at least fill the random pool up fast enough to check if nothing else is broken 1139840100 M * angel4u hi Bertl 1139840108 M * Bertl cehteh: eh, well, I doubt that will work :) 1139840110 M * angel4u hi all 1139840111 M * cehteh mhm no does not 1139840116 M * h01ger morning Bertl :-) 1139840125 M * Bertl morning h01ger! 1139840126 M * h01ger vserver rock my world :) but... 1139840127 M * eyck cehteh: cat /dev/urandom drains /dev/random...NOT GOOD when you're at 0 1139840136 M * h01ger huhu - on my vservers (running debian/sarge on a debian host running 2.6.15 with kernel-patch-vserver 2.0.1-1) every morning at 06:25:02 i get a logcheck mail "su[3113]: pam_open_session: Permission denied" like in http://list.linux-vserver.org/archive/vserver/msg10043.html - but opposed to that mail, i dont have setlimit problems and su works on the commandline 1139840136 M * h01ger at 06:25 cron.daily is run, if i run it manually, i get no errors... ?! 1139840136 M * h01ger any ideas ? :) 1139840137 M * cehteh yeah 1139840137 M * angel4u i`m happy to see linux-vserver project workin 1139840178 M * daniel_hozac Bertl: hmm, changes from 2.0.2-rc1 to 2.0.2-rc2? 1139840225 M * Bertl daniel_hozac: will break them down into patches in a minute 1139840280 M * angel4u someone know where i can buy one dedicated server ? i prefer german line 1139840484 M * Bertl h01ger: hmm, well, some su/pam obviously fails ... because something is denied, so, it might be interesting to a) figure which user it tries to su to, and b) what stuff is done by pam ... 1139840856 M * h01ger Bertl, at 06:25 cron.daily is run, if i run it manually, i get no errors... ?! 1139840869 M * h01ger su also works 1139840885 M * Bertl IIRC, cron is usually run as nobody or root (depending on the setup) 1139840906 M * Bertl also, cron runs without terminal, while you have one (right now) 1139841089 M * h01ger hmmm... ok, will see / investigate further. thanx anyway! 1139841120 M * Bertl as I said, providing the pam seetup and/or testing with the specific user might give some clues 1139841123 M * Hollow Bertl: any fix for the /proc/virtnet bug yet? 1139841129 M * Bertl Hollow: nope 1139841136 M * Hollow and, regarding the new schuler.. is it likely to change again in 2.1.x? 1139841146 M * Hollow *scheduler 1139841149 M * Bertl should not 1139841155 M * Hollow ok, thanks :) 1139841277 M * Bertl daniel_hozac: check /Experimental the delta (as one patch) and the plm diffs are there now 1139841333 M * daniel_hozac ah, thanks. 1139841636 M * daniel_hozac you still have the atomic_long's ;) 1139841639 M * Bertl a few regressions (not unexpected after the huge patch) 1139841651 M * Bertl daniel_hozac: yes, unfortunately :) 1139841840 M * Bertl the INET stuff is still a todo too 1139841936 A * h01ger leaves (ETOOMANYCHANNELS) - did i say that vserver is great ? :-)) 1139841940 P * h01ger Leaving 1139842081 M * daniel_hozac you didn't include the vs_pid.h patch? 1139842592 M * Bertl I probably missed a lot of stuff in 2.6.15 ... 1139842986 M * nox Bertl: wb! is the bme in 2.15.4 also high experimental? 1139842990 M * Roey EHLO * 1139842999 M * Roey what's bme? 1139843066 M * nox [06:25:15] < nox>bind mount extension SuperLag 1139843067 M * nox [06:25:50] < nox>so mount - bind behave like you would expect 1139843078 M * nox s/SuprLag/Roey/ 1139843147 M * Bertl nox: hmm, no, otherwise we would not put it in a stable rc :) 1139843163 M * Roey nox: oh? 1139843166 M * Roey oh 1139843168 M * Roey bind mount extension 1139843169 M * Roey alright. 1139843184 M * Roey so you can do mount -o bind within a vhost?? 1139843205 M * Roey Bertl: nice to know it has passed into Stable. 1139843209 M * Bertl well, that works fine ... 1139843217 M * Roey Bertl: I vant my openvpn ;) 1139843224 M * Roey it would be /great/ for our setup. 1139843237 M * Bertl but the main thing BME gives is support for ro,noatime,nodiratime,noexec on --bind 1139844587 M * s25 on the http://linux-vserver.org/ it says the latest stable kernel is kernel [2.6.14.3] so i put into ubuntu CLI apt-get install linux-image-2.6.14.3 and i get a couldnt find package linux-image-2.6.14.3 1139844639 M * Bertl s25: well, it's mainline kernel not ubuntu kernels 1139844670 M * Bertl s25: I'm sure ubuntu provides a bunch of kernels (some probably with linux-vserver too) 1139844678 M * s25 ok 1139844714 M * Bertl but in case you do not find one, you can always take the mainline ekrnel and patch it 1139844931 M * s25 hmm 1139844938 M * s25 just gotta find some now :) 1139845058 M * Bertl https://wiki.ubuntu.com/VServer (just what google gave me :) 1139845134 M * s25 i typed in apt-get install linux-image and got a list :D 1139845141 M * s25 sorry did not see youre list 1139845321 M * s25 yay its working :D 1139845337 M * Bertl great! 1139845638 J * Doener doener@i5387D124.versanet.de 1139845646 M * Bertl welcome Doener! 1139845652 M * Doener hi Bertl! 1139845739 M * Bertl daniel_hozac: care to remind me what the vs_cvirt.h:107:2 duplicate inclusion issue was? (I know it's fixed by the vs_pid stuff, but I'm not sure we should put that into stable yet) 1139845749 M * s25 yep now just rebooting all seems to be going well 1139845998 M * s25 it says it needs 1139846014 M * Bertl well, then give it! 1139846021 M * s25 libbeecrypt6 but it is not installable and so vserver is a broken package 1139846028 M * s25 hmm 1139846043 M * Bertl even debian has libbeecrypt IIRC :) 1139846065 M * s25 hmm :) 1139846098 M * s25 i tryed to do pkg add toit and it says that it is missing or obsoleted or only available via source 1139846109 M * s25 *from another source 1139846139 M * Bertl well, strictly speaking you not not need libbeecrypt, but you probably have to rebuild the package 1139846156 M * s25 how should i do that? 1139846175 M * s25 and should i add libbeecrypt to the system ifi can find it? 1139846178 M * Bertl I'm no ubuntu guy, so no idea there ... 1139846196 M * Bertl on mandrake I do rpm --rebuild 1139846240 M * tudenbart hm. chcontext --xid 1 netstat -lnp tells me (Not all processes could be identified, non-owned process info will not be shown, you would have to be root to see it all.) 1139846267 M * Bertl on 1.2.11 ? 1139846288 M * tudenbart hm, don't know. i've installed the box on early december. 1139846305 M * tudenbart it was latest and brightest way then. 1139846327 M * Bertl testme.sh will tell us (run and upload output to pastebin.com) 1139846363 M * tudenbart hm, where should i find that? 1139846409 M * Bertl google says here: http://vserver.13thfloor.at/Stuff/SCRIPT/ shall we trust him? :) 1139846422 M * tudenbart ok. 1139846485 M * tudenbart http://pastebin.com/552735 1139846514 M * tudenbart i thought it would be somewhere in the mainline distribution... 1139846550 M * Bertl no, as we update it continously .. and it can be used for very old versions too 1139846575 M * tudenbart hm, wrap some wget magic arround it? 1139846600 M * Bertl maybe in the future ... but that depends on enrico 1139846607 M * tudenbart ok. 1139846622 M * Bertl anyway, looks fine, dumb question, you are running it as root? 1139846644 M * tudenbart um. because of i'm not paranoid enough. 1139846657 M * tudenbart so, i did just execute it. 1139846687 M * Bertl ah, no, not the script, that was fine, the netstat 1139846710 M * tudenbart yep, did run it as root. 1139846777 M * Bertl hmm, well, I remember something but I cannot put my finger on it right now ... might have been fixed since ... 1139846945 M * tudenbart so it works at your box? 1139846957 M * Bertl yup, definitely 1139846979 M * daniel_hozac Bertl: that was the ptrace.h issue. 1139846985 M * s25 how do you exit VI editor on the command line? 1139846989 M * daniel_hozac Bertl: http://vserver.13thfloor.at/Experimental/delta-include-clean01.diff 1139846990 M * tudenbart how powerfull is the spectator? may it just watch? or may it kill etc. too? 1139847003 M * Bertl s25: ESC : q 1139847031 M * s25 nope not for me my pc jsut beeps 1139847036 M * Bertl tudenbart: usually spectator can only watch, but host context can kill/signal too 1139847066 M * s25 ctrl+x wont work eitjher 1139847093 M * Bertl s25; press escape, then ':' you should get a colon in the bottom line 1139847111 M * s25 yes i do 1139847121 M * Bertl s25: now type q! and press enter 1139847123 M * tudenbart Bertl: so would you name it a good idea to let something like zabbix or bigbrother run beyound it? 1139847145 M * tudenbart the agent/collecting part? 1139847158 M * s25 hmm what do i do to save my work? 1139847169 M * s25 s wont work 1139847176 M * Bertl tudenbart: depends, a lot of things can be gathered from the procfs (from the host context too) 1139847197 M * Bertl s25: ESC : w or ESC : x 1139847214 M * s25 thasnk you :D 1139847220 M * Bertl s25: but better get a vi/vim reference manual (or a different editor :) 1139847246 M * s25 heheh :) 1139847258 M * s25 damnit still cant get that librypt 1139847266 M * s25 i made the other repositories available ya see 1139847326 M * s25 hmm cya in about 20 minutes gotta do some cleaning >( 1139847353 M * Bertl daniel_hozac: ah, tx! 1139847440 M * tudenbart hm. i screwed it. 1139847446 M * daniel_hozac Bertl: will probably need http://daniel.hozac.com/vserver/delta-vs_cvirt-fix01.diff as well for 2.6.15. 1139847504 J * _mountie ~mountie@trb229.travel-net.com 1139847512 M * tudenbart oh, that spectator zabbix won't be visible in netstat of the root box. 1139847559 M * tudenbart and the 0.0.0.0 bind it makes isn't reachable from the outside :( 1139847573 J * djudko ~davej@129.33.1.37 1139847594 M * Bertl welcome djudko! 1139847620 M * djudko hi 1139847878 M * tudenbart hm. for some reason the zabbix agent won't start. 1139847879 J * stefani ~stefani@superquan.apl.washington.edu 1139847908 M * Bertl welcome stefani! 1139847908 M * Bertl tudenbart: strace? debug log? 1139847918 M * Roey Bertl: so lemme understand, you guys do not intend to ever merge into the kernel, right? 1139847956 J * dearaujo ~dan@pixpat.austin.ibm.com 1139847960 M * Bertl Roey: hmm, well, hard to say now that mugwump is merging it :) 1139847970 Q * mountie Ping timeout: 480 seconds 1139847973 M * Bertl welcome dearaujo! 1139847976 M * Roey heh 1139847978 M * Roey Bertl: oh?? 1139847987 M * Roey Bertl: is it going to be hard living with xen under the same roof? 1139848006 M * Bertl no, Linux-VServer supports xen (once it's in mainline) 1139848015 M * Roey oh, ok. 1139848036 M * Roey Bertl: what about that other system that's being pushed by some company; the one we've discussed before 1139848062 M * Bertl OVZ (from the sw folks?) 1139848071 M * Roey yes. 1139848073 M * Roey openvz. 1139848079 M * Roey exactly that. 1139848087 M * Bertl well, what about it? 1139848130 M * Roey so that's going to be merged too, right? 1139848144 M * Roey an article somewhere suggested something to that effect. 1139848169 M * Bertl well, no, but the thing is, we will try to get something in mainline which can be used by all parties 1139848179 M * Roey ok, cool. 1139848248 M * tudenbart Bertl: http://tar.gzipped.org/out 1139848305 M * Bertl funny host :) 1139848305 J * Viper0482 ~Viper0482@p549776C3.dip.t-dialin.net 1139848362 M * Bertl tudenbart: you have to put the strace inside the chcontext 1139848376 M * tudenbart ah, ok. 1139848508 M * Bertl wb _mountie! 1139848532 M * Bertl Vudumen: ping! 1139848671 M * tudenbart <- was to stupid. 1139848700 M * tudenbart forked processes get the same context? 1139848736 M * Bertl yup 1139848750 Q * shedi Quit: Leaving 1139849031 M * Bertl daniel_hozac: okay, the atomic stuff seems fixed ... will upload changes soon 1139849878 J * jeeves ~Bob@c-24-11-171-10.hsd1.mi.comcast.net 1139849889 M * Bertl welcome jeeves! 1139849897 M * jeeves Hey B! 1139849971 M * jeeves Whats new? 1139849980 M * mnemoc B :) 1139850032 M * Bertl yo man! :) 1139850036 P * dearaujo 1139850043 M * Bertl jeeves: well, *-rc* is new 1139850051 M * jeeves Is someone still maintaining RPMs with the new releases? 1139850054 M * jeeves Cool 1139850074 M * Bertl kernel or userspace and what distro? :) 1139850094 M * jeeves kernel, Fedora 1139850101 M * daniel_hozac Fedora is very unspecific. 1139850114 M * daniel_hozac FC4 got an update not too long ago... 1139850120 M * daniel_hozac 2.0.2-rc2 is already in CVS. 1139850173 M * jeeves Ok, I will look. Thank You. 1139850197 M * daniel_hozac "09-Feb-2006 15:23" ;) 1139850218 M * jeeves Is it on your Yum repos? 1139850222 M * daniel_hozac yep. 1139850273 M * jeeves Ok, so in a couple weeks I am going from an old Athlon to a Athlon 64 X2 3800. Which should I be prepared to install? 1139850369 M * daniel_hozac you'd have to rebuild the source RPM if you want RPMs, as i have no x86_64 and qemu kept failing on me... 1139850425 M * jeeves ok, I will try it. 1139850441 M * jeeves It will be a few weeks b4 I have the upgrade installed. 1139850453 M * daniel_hozac if you could send me the built RPMs, that would be great. 1139850478 M * Bertl daniel_hozac: hmm? @ qemu 1139850480 M * jeeves I just got my dual 19" LCDs working right. Xorg was a pain. 1139850482 M * daniel_hozac currently the x86_64 repo has some 2.0.1-pre2 version. 1139850513 M * daniel_hozac Bertl: it was odd. i could never get an install finished before it crashed (i can't remember the details...). 1139850519 M * daniel_hozac for ppc nor x86_64. 1139850525 M * daniel_hozac this was with the qemu binaries. 1139850548 M * jeeves How buggy is kernel-smp-2.6.13-1.1532_FC4.vs2.0.1.0.pre2.1.x86_64.rpm 1139850576 M * daniel_hozac well, there are a number of mainline bugs and security issues. 1139850625 M * daniel_hozac but with a machine that fast, rebuilding the kernel RPM should only take a few hours. (and i would _really_ appreciate it :)) 1139850681 M * jeeves lol!. I will try it. I don't compile the kernel often. And I have never made an RPM from it. But, there is always a first 1139850705 M * daniel_hozac it should just be a matter of rpmbuild --rebuild kernel-....src.rpm 1139850745 J * ub ~ub@p5484BA05.dip0.t-ipconnect.de 1139850746 M * jeeves Forgive me if I ask you that again in a couple weeks. 1139850765 M * Bertl welcome ub! 1139850771 M * daniel_hozac heh, anything to get updated x86_64 RPMs ;) 1139850805 M * Bertl daniel_hozac: I might try to compile it on mandrake 1139850846 M * daniel_hozac got a Fedora vserver? :) 1139850867 M * Bertl not yet, but you might tell me how to install that, no? :) 1139850913 M * daniel_hozac i assume mandrake lacks yum? 1139850942 M * Bertl yup, but I guess I can get that, sec 1139850962 M * Bertl yum-2.2.2-3mdk.noarch good enough? 1139850983 M * daniel_hozac should be. 1139850993 M * Bertl probably needs the virt patch .. but I'll find that too 1139851068 J * NITZAN nitzangur@bzq-82-81-196-60.cablep.bezeqint.net 1139851074 M * NITZAN who knows about botnet 1139851074 M * NITZAN ? 1139851092 M * TheSeer wtf? 1139851093 M * Bertl NITZAN: how is that #vserver related? 1139851131 M * NITZAN i dont know i just looking for.. 1139851147 M * Bertl NITZAN: well, check out #offtopic 1139851159 M * NITZAN k tanks 1139851160 P * NITZAN 1139851205 M * SNy I would have suggested the internet storm center, maybe 1139851217 M * SNy ;p 1139851218 M * michal_ ;p 1139851222 M * TheSeer hehe 1139851550 Q * ub Remote host closed the connection 1139851922 M * jeeves daniel_hozac: If I download the SRPM from your repos and compile it on my soon to come X2 64, will it compile properly? 1139851937 M * daniel_hozac jeeves: that's the idea. 1139851966 M * daniel_hozac if it doesn't, i'll just have to fix it. 1139851981 M * jeeves Ok. Are the kernel options about the same to Fedora, or do I have to choose the right options? 1139851990 Q * meandtheshell Quit: a brave heart left ... 1139852030 M * daniel_hozac it's the same as Fedora's, i've only added the vserver options. 1139852061 M * daniel_hozac (for i686-smp there are HIGHMEM changes too, due to some bizarre bug) 1139852064 M * jeeves Sweet. 1139852500 Q * Viper0482 Ping timeout: 480 seconds 1139852521 J * bonbons ~bonbons@83.222.39.180 1139852670 M * daniel_hozac oh, and audit is disabled. i had forgotten about that. 1139852859 M * jeeves I don't even know what that is. 1139852878 M * daniel_hozac it causes problems with Fedora guests. 1139852885 M * Bertl daniel_hozac: *-rc3 on Experimental 1139852895 M * daniel_hozac damn, you work fast! 1139852908 M * Bertl daniel_hozac: includes prelim INET fix for 2.0.2 1139852944 M * daniel_hozac oh, nice. 1139852985 M * eyck hmmm 1139853027 M * eyck there seems to be no way of making sure to vserver user can empty the entropy pool... 1139853036 M * eyck vide cat /dev/urandom > /dev/null.... 1139853054 M * Bertl well /dev/random is probably better here ... 1139853067 M * Bertl but yes, maybe we should add an entropy limit :) 1139853080 M * eyck yes, but as I'm cautious and all, I removed /dev/random from this particular user... 1139853140 Q * kilian Ping timeout: 480 seconds 1139853142 M * eyck and then, in his vserver he's got /dev/random == /dev/urandom... but he can still very easily practically disable all encryption going on on all vservers (empty entropy pool) 1139853248 M * Bertl well, I guess we can add a limit there ... 1139853285 M * daniel_hozac token bucket style? 1139853290 M * Bertl yep 1139853292 J * kilian kk@projects.verfaction.de 1139853294 M * eyck woa, talk about overkill ;) 1139853351 M * eyck that would be fun - you configure vserver with 256M ram, 20G of hard disk, ~500Mhz of cpu, AND 16k/minute of randomness :) 1139853360 M * daniel_hozac hehe. 1139853379 M * daniel_hozac got any other ideas? ;) 1139853379 M * Bertl 8k plus 20k burst :) 1139853388 M * eyck golden pack - a whopping 32k/s of randomness, 1139853445 M * Bertl that's something folks can sell, I'm sure about that :) 1139853461 M * eyck hehe 1139853511 J * baggins baggins@kenny.mimuw.edu.pl 1139853520 M * Bertl the question is, how is that handled on a normal linux system, where joe user can use /dev/urandom via bash? 1139853545 M * baggins Hi Bertl 1139853549 M * eyck well, on normal system, any joe shmoe user can disable entropy pool 1139853555 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1139853562 M * Bertl hey baggins! flock! 1139853564 M * baggins Bertl: there is a bug in patch-2.6.15.4-vs2.1.1-rc2.diff 1139853568 M * eyck as usual, vserver would be light years ahead of competition.. 1139853571 M * daniel_hozac baggins: -rc3 ;) 1139853614 M * baggins ah, I see, with atomic_long_ fixed ;) 1139853633 M * Bertl baggins: but keep reporting! 1139853992 Q * flock Quit: Expert, n.: Someone who comes from out of town and shows slides. 1139854044 J * Viper0482 ~Viper0482@p5497762D.dip.t-dialin.net 1139854131 Q * mkhl Ping timeout: 480 seconds 1139854372 Q * Viper0482 Remote host closed the connection 1139854382 N * ebiederm_zZ ebiederm 1139854459 M * Bertl morning ebiederm! 1139854465 M * ebiederm morning Bertl 1139854757 J * Viper0482 ~Viper0482@p5497762D.dip.t-dialin.net 1139854776 M * Bertl ebiederm: the OVZ folks are worried ... because they haven't got feedback from you since yesterday or so ... 1139854800 M * ebiederm Bertl: They sent me mail right after I went to bed. 1139854806 M * ebiederm And I just woke up. 1139854811 M * ebiederm It's just bad timing. 1139854861 M * ebiederm Do you know if anyone cares about the chroot hack in /proc? 1139854868 M * baggins include/linux/vserver/limit_int.h:23: error: `__rlim_val' undeclared (first use 1139854871 M * baggins in this function) 1139854878 M * baggins that's rc3 1139854889 M * Bertl ebiederm: well, we disable it on the host :) 1139854915 M * Bertl baggins: x86_64 ? 1139854923 M * baggins yes 1139854930 M * Bertl give me a few minutes 1139854942 M * ebiederm Given that kill is not modified... I don't see the point in hiding things in /proc 1139854964 J * Smutje ~Smutje@xdsl-87-78-6-153.netcologne.de 1139854968 M * ebiederm kill -0 pid will tell me if the pid exists anyway. 1139855077 M * Bertl ebiederm: can go yesterday as far as I'm concerned 1139855112 M * ebiederm As we are in the process implemneting much strong mechanisms anyway, sounds sane here. 1139855125 M * ebiederm I'm trying to reduce /proc to something that is readable and correct. 1139855175 M * Bertl sounds too good to be true :) 1139855209 J * shedi ~siggi@inferno.lhi.is 1139855218 M * ebiederm Hey I killed the enum. The biggest reason for tying crap too closely together. 1139855890 J * liquid3649_ ~Viper0482@p5497762D.dip.t-dialin.net 1139855917 M * dhansen ebiederm: you're being very patient with Kirill :) 1139856122 M * ebiederm dhansen: Patience is not that hard. 1139856162 M * ebiederm Of course I still need to see what this weeks discussion is about. 1139856171 Q * Viper0482 Ping timeout: 480 seconds 1139856291 Q * cryo Quit: Покидаю 1139856333 M * dhansen ebiederm: I noticed your git tree hasn't been updated in a few days. Are you still working on it, or have you simply not pushed your work back to that tree? 1139856351 M * ebiederm dhansen: I haven't pushed my work back lately. 1139856412 M * dhansen ebiederm: do you have a list of Kirill's issues with your code? 1139856420 M * dhansen even the ones you don't really think are issues? 1139856476 M * ebiederm Not really. 1139856521 M * dhansen ebiederm: OK, how about this: 1139856531 M * dhansen we'll get your patches in a state where they can be posted as a nice patch series 1139856543 M * dhansen we'll get Kirill (and everyone else) have another go at them 1139856569 M * dhansen fix the issues that are raised this week, and try to get a higher-level look (Linus?) next week 1139856660 M * ebiederm Sounds roughly like a plan. 1139856683 M * dhansen BTW, I have stopped worrying about your code not having an explicit 'struct container' 1139856689 M * dhansen +struct pspace 1139856689 M * dhansen +{ 1139856689 M * dhansen + struct pspace *parent; 1139856689 M * dhansen + int last_pid; 1139856690 M * dhansen ... 1139856693 M * dhansen struct container 1139856693 M * dhansen { 1139856694 M * dhansen atomic_t count; 1139856696 M * dhansen struct pspace pspace; 1139856698 M * dhansen } 1139856700 M * dhansen it's as simple as that 1139856704 M * dhansen if it comes down to it 1139856710 M * ebiederm One of the things I want todo is fix all of the lower level issues. 1139856726 M * ebiederm dhansen: Exactly. 1139856745 M * dhansen for what it is worth, I worry whenever someone says "I need to think about it", especially when they're the author ;) 1139856768 M * dhansen With the flags, I'd really suggest using some of the existing atomic primitives that we have already 1139856785 M * dhansen even with those, we're always coming up with new issues with memory ordering, and they've had the snot tested out of them 1139856809 M * ebiederm dhansen: I already have done the thinking, renamed the variable state, and added a wmb after I set it. 1139856877 M * ebiederm For an aligned word sized variable reads/writes are atomic (they fit in a single cache line) 1139856908 M * ebiederm You can never observe a variable with different values in the system or else your cache coherency is broken. 1139856940 M * RoadRunnR ebiederm: what about non cache coherent systems? ;-) 1139856941 M * ebiederm The only real issue is can you observer a kill happening before you observe PSPACE_EXIT. 1139856965 M * ebiederm RoadRunnR: They don't run a single kernel so no problem. 1139856972 M * dhansen ebiederm: I'm less worried that it works, and more worried that few people can debug it :( 1139856974 M * ebiederm A wmb fixes the last issue. 1139856983 M * dhansen we _know_ the atomic_t's (for instance) work 1139857014 M * dhansen ebiederm: Let me put it another way: using the atomic_t is dumber, and less efficient 1139857023 M * dhansen but, it gets your code in with less fuss :) 1139857032 M * Bertl I have to agree with dhansen, especially since atomic_long made it into mainline 1139857043 M * dhansen (atomic_t/set_bit, etc...) 1139857075 M * ebiederm dhansen: When one of the kernel maintainers hollers to me about it I will worry. 1139857103 M * dhansen ebiederm: fair enough. I just try to avoid having them holler at me :) 1139857136 M * ebiederm Using something that allows a read shared cache line instead of cache line exclusivity is much nice on SMP. 1139857146 M * ebiederm Especially to close a race that is practically impossible to trigger anyway. 1139857159 M * ebiederm As for maintainers hollering at me, I have been known to holler back. 1139857171 M * dhansen "practically impossible" == will happen at the worst possible time :) 1139857196 M * dhansen ebiederm: IBM is very good at making machines that break "practically impossible" to break code 1139857197 M * ebiederm dhansen: There is a reason I was closing it. :) 1139857206 M * Bertl baggins: http://vserver.13thfloor.at/Experimental/delta-2.6.15.4-vs2.1.1-rc3.1.diff (ontop of rc3) 1139857243 M * baggins Bertl: thenks, I'll try it now 1139857258 M * ebiederm dhansen: Currently Linux Networx makes hardware and runs software that on a routine basis finds practically impossible to trigger hardware problems. 1139857269 M * Bertl dhansen: maybe you could get a few of them to test linux-vserver then :) 1139857285 M * ebiederm So I know what you are talking about. 1139857308 M * dhansen Bertl: if we end up using the same code, it will get lots of testing 1139857343 M * dhansen Bertl: the current Meiosys product testing is basically a copy of LTP that they checkpoint, restore and migrate continuously 1139857352 M * dhansen finds lots of bugs 1139857400 M * ebiederm Hmm. I gues that would catch the easy cases. 1139857428 M * ebiederm Especially for infrastructure style changes. 1139857482 M * dhansen ebiederm: I really want _somebody's_ code to succeed. I've seen too many patches derailed over little, silly issues like atomics. 1139857514 M * dhansen I think code that 99% of people can read and understand is much more important than an optimized version that 20% can read and understand. 1139857626 M * ebiederm dhansen: If there are bumps I have no problem going back and changing my code. 1139857675 M * ebiederm For most reading simply looking at the code and accepting that it does what it does what it tries to do, is sufficient. 1139857702 M * ebiederm Anyway I have I mentioned what the plan I am working on is? 1139857720 M * dhansen I don't think so. Remind me :) 1139857777 M * ebiederm Ok. My currently plan right now is to get the bugs fixed, and the cleanups made that my patches suggest. 1139857810 M * ebiederm Which at the moment seems to include running /proc through the wringer. 1139857819 M * ebiederm And then submit my patches on top of that. 1139857870 M * dhansen I'm contemplating how we can make the development a little more transparent 1139857877 M * Bertl if you are on it, you might consider fixing the proc security issues too 1139857879 M * dhansen basically, how can we get other people looking at the code and helping you out 1139857888 M * ebiederm BTW: Has anyone else noticed that unshare needs additional locks? 1139857900 M * dhansen ebiederm: First of all, does that sound like a good idea? 1139857929 M * ebiederm dhansen: During development I have no problems with people looking. It is just a matter of fairly regularly pushing my git tree. 1139857976 M * ebiederm The reason I am acting as a primary is that things can be much more cohesive if funneled through a single brain. At least during design. 1139858021 M * ebiederm Now that it looks like we have rough consensus I don't mind supporting my work to some extent. 1139858034 M * dhansen ebiederm: what I'm thinking of is that you freeze a copy of your patches and send them to somebody (I'd volunteer for now) 1139858042 M * dhansen they keep the set of patches consistent, and distribute them 1139858061 M * dhansen when you need to add or fix something, separate out the patch and send it 1139858074 M * dhansen that somebody publishes a new set every day or two 1139858111 M * dhansen The downside for you is that it requires some level of consensus and understanding from more than yourself when you want to make changes. 1139858129 M * dhansen But ... it bring some other people along for the ride, and brings them on your team 1139858221 M * ebiederm Well for things that are basically ready for kernel inclusion I don't have a problem. As they have largely settled. 1139858245 M * ebiederm For parts that are still in flux I do have issues. 1139858261 M * ebiederm I really think pushing thing up to my git tree is the easiest way you will get code out of me though. 1139858284 M * ebiederm I also think I want to do sysvipc instead of pids first. 1139858296 M * dhansen at this point, I'm less concerned about getting code out of you (that's not a problem :) 1139858307 M * dhansen and more concerned about gaining consensus 1139858320 J * brc_ bruce@20151170196.user.veloxzone.com.br 1139858332 M * dhansen I know it slows down development a bit 1139858371 M * ebiederm dhansen: We largely already have design consensus. 1139858394 M * dhansen Kirill might disagree ;) 1139858417 M * dhansen I think the "patches accepted" argument is a lot stronger than the "your code sucks one" :) 1139858439 M * ebiederm Agreed. And that is where I am focusing now. 1139858550 M * Bertl okay, updated to 2.6.15.4-vs2.1.1-rc4, and the latest 2.6.15.4 versions are running on plm (4989/90) 1139858556 Q * wally Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1139858573 M * Bertl I'm off to bed for now .. a little tired, but will probably be back later ... 1139858591 M * ebiederm Sleep well Bertl 1139858592 M * Bertl Hollow: will look at the nid list issue tomorrow! 1139858601 M * Bertl ebiederm: tx, have fun! 1139858616 M * ebiederm dhansen: How are you comming with understanding my patches? 1139858618 M * Bertl daniel_hozac: thanks a lot for your help! 1139858629 N * Bertl Bertl_zZ 1139858723 M * dhansen ebiederm: I have a better grasp of where you were 5 days ago :) 1139858742 M * dhansen I hope you understand my insistence on understanding where you are _now_ 1139858752 M * ebiederm dhansen: Well except for /proc the overall structure hasn't been changing. 1139858754 M * ebiederm dhansen: Of course. 1139858875 M * ebiederm dhansen: When I get something working properly I stop messing with it. 1139859731 J * dearaujo ~dan@pixpat.austin.ibm.com 1139859794 M * mugwump hey folks 1139859812 M * ebiederm mugwump: Hey. 1139862003 M * BartVB I'm trying to start apache2 on my vserver but it keeps complaining about being unable to bind to port 80 :\ 1139862025 M * daniel_hozac BartVB: chcontext --xid 1 netstat -pnlt 1139862029 M * BartVB the strange thing is that there isn't anything on port 80 on the whole box as far a I can see? 1139862077 M * BartVB daniel_hozac, ran that in the host and it shows some processes but nothing on port 80 1139862091 M * BartVB just 111, 112, 22, 695 and 25 1139862127 M * BartVB hmm: 1139862128 M * BartVB (Not all processes could be identified, non-owned process info 1139862128 M * BartVB will not be shown, you would have to be root to see it all.) 1139862136 M * BartVB but I am root when running that command? 1139862293 M * BartVB brought that down to just ports 22 and 25, didn't help 1139862309 M * BartVB (removed portmap/nfs and inetd/ident) 1139862314 M * BartVB I get: 1139862323 M * BartVB Starting web server: Apache2(98)Address already in use: make_sock: could not bind to address 82.161.230.66:80 1139862323 M * BartVB no listening sockets available, shutting down 1139862337 M * BartVB the IP should be correct: 1139862338 M * BartVB pegasus:/# ip a 1139862338 M * BartVB 2: eth0: mtu 1500 qdisc pfifo_fast qlen 1000 1139862338 M * BartVB link/ether 00:e0:81:55:10:8f brd ff:ff:ff:ff:ff:ff 1139862338 M * BartVB inet 82.161.230.66/24 brd 82.161.230.255 scope global secondary eth0 1139862675 M * BartVB hmm, seems like I fixed it 1139862686 M * BartVB I also had a Listen 127.0.0.1:80 directive 1139862696 M * BartVB that didn't work because that IP isn't available in the vserver 1139862707 M * daniel_hozac 127.0.0.1 is rewritten to the primary IP address. 1139862707 M * BartVB I removed that line and now it's working 1139862716 M * BartVB aaah :) 1139862717 M * daniel_hozac (primary == first) 1139862719 M * BartVB that explains it :D 1139862729 M * BartVB then it tries to bind to the same IP/port twice 1139862731 M * BartVB thanks! 1139862734 M * BartVB didn't know that 1139862751 M * BartVB then I also don't need to add something like 127.0.0.2 to the server for localhost stuff :) 1139862759 M * daniel_hozac nope. 1139862767 M * BartVB great, thanks a lot! 1139862782 M * daniel_hozac you're welcome! 1139862818 J * Blissex ~Blissex@82.69.39.138 1139862885 M * dearaujo hello - I am trying to run a vsftp server within a vserver guest. 1139862889 Q * lilalinux Remote host closed the connection 1139862928 M * mugwump > Uhhh... I see that you have no real arguments. Nice. 1139862933 M * mugwump *bzzt* 1139862950 M * dearaujo if I weren't using a vserver I could simply add a line to xinetd to spawn the service when a request for ftp was received 1139862982 M * dearaujo can you do that with a vserver? 1139862994 M * mugwump of course 1139863002 M * mugwump just make sure you are running xinetd inside the vserver, too 1139863034 J * dothebart ~willi@xdsl-213-196-242-183.netcologne.de 1139863045 M * dearaujo so I'd have to add a line to xinetd to spawn a vserver that ran xinetd and vsftp 1139863053 M * dearaujo that what you're saying 1139863094 M * mugwump not quite, 1139863113 M * mugwump xinetd is started when you start the vserver 1139863124 M * mugwump then, xinetd gets the connections, and starts the service for you 1139863171 M * dearaujo on the guest 1139863171 M * mugwump what you said sounded a little back to front, and while technically possible, is not the normal way. 1139863185 M * mugwump yes, xinetd runs in the guest 1139863318 M * dearaujo ok, ill give it a shot - thanks. 1139863343 M * dearaujo And I'll remember next time to type the whole question on one line.... 1139863348 T * daniel_hozac http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.1-rc4, 2.0.2-rc3 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1139863450 Q * tudenbart Ping timeout: 480 seconds 1139863653 Q * FireEgl Ping timeout: 480 seconds 1139864445 M * daniel_hozac Bertl_zZ: http://daniel.hozac.com/vserver/delta-vs_pid-fix01.diff for 2.1.1-rc4, same as delta-vs_cvirt-fix01.diff but for vs_pid. 1139864458 N * ebiederm ebiederm_zZ 1139864913 J * FireEgl Atlantica@Atlantica.DollarDNS.Net 1139865014 M * michal_ daniel_hozac: may i have a question ? 1139865036 M * michal_ (offtopic a bit, regarding sys_clone(), wait*() and CLONE_NEWNS) 1139865046 M * daniel_hozac i doubt i'll be able to answer, but sure ) 1139865080 M * michal_ basicaly i am trying to use sys_clone() directly 1139865089 Q * brc_ Quit: [BX] I see your BitchX is as big as mine! 1139865134 M * michal_ so i need to have some way to spawn a process (ok, doing it via clone), waiting for it beeing running, exit parent process, leave child running 1139865160 M * michal_ now the question is - how i do wait for a children and exit after it is running ? 1139865179 M * daniel_hozac hmm, do you want to wait for it to start or end? 1139865192 M * michal_ to start only 1139865198 M * michal_ than parent can safetly exit 1139865229 M * michal_ now i am waiting for it to exit (and i hate it this way because it is completly unnecesary) 1139865251 M * daniel_hozac so you are clone'ing and then execve'ing some other program? 1139865251 M * michal_ wait4(pid, &status, 0,0) 1139865257 M * michal_ daniel_hozac: yes 1139865267 M * michal_ exactly like this - clone, than child does execve 1139865277 M * michal_ (in a new namespace) 1139865363 M * daniel_hozac and you just want to make sure the execve succeeds, right? 1139865373 M * michal_ hm, i think so 1139865462 M * daniel_hozac well, i can't think of any way to do that. 1139865484 M * michal_ vserver utils are doing somethink similar i think 1139865496 M * michal_ somehow i cannot reverse the algoritm ;] 1139865498 M * daniel_hozac possible. 1139865532 M * michal_ ok, thx anyway, i'll wait for Bertl to wake up :] 1139865539 M * daniel_hozac hehe. 1139865539 J * Smutje_ ~Smutje@xdsl-87-78-16-225.netcologne.de 1139865545 P * dearaujo 1139865588 M * Doener hm, IIRC that's just a big execv() chain and that fact that, e.g. the start, is asynchronous for plain initstyle is because of the init process detaches 1139865600 M * Doener when you use sysv style, it is synchronous as well... 1139865624 J * brc_ bruce@20151170196.user.veloxzone.com.br 1139865626 M * Doener uuh, broken english, too much wine.. 1139865630 M * michal_ ;p 1139865649 Q * Smutje Ping timeout: 480 seconds 1139865649 N * Smutje_ Smutje 1139865677 M * michal_ vserver-utils / src/vnamespace.c 1139865678 P * Blissex Konversation terminated! 1139865683 M * michal_ mayby there will be sth interesting ;] 1139865755 J * shuri ~shuri@64.235.209.226 1139865969 Q * Doener Quit: Leaving 1139866022 M * Hollow michal_: signal(SIGCHLD, SIG_IGN); pid = fork(); if (pid > 0) exit(EXIT_SUCCESS); 1139866061 M * michal_ nah 1139866063 M * michal_ i cannot use fork 1139866075 M * Hollow why? 1139866077 M * michal_ look up - i need sys_clone() directly 1139866086 M * michal_ 1. CLONE_NEWNS 1139866089 M * Hollow then replace fork with sysclone 1139866093 M * michal_ 2. have no idea about stack size 1139866096 M * Hollow 0 1139866101 M * michal_ yep 1139866116 M * Hollow for simple clone; execve, this should be ok 1139866126 M * michal_ hm, tell me two things... 1139866136 M * michal_ why (and where ?) are you sending that signal ? 1139866164 M * Hollow because else you would get zombies if the child exits and no process does collect the exit status 1139866175 M * Hollow (via wait4) 1139866190 M * Hollow so, you tell it to just ignore it 1139866198 M * daniel_hozac Hollow: that just tells you that fork() succeeded. 1139866206 M * daniel_hozac rather than the following execve. 1139866232 M * Hollow well, then use VFORK 1139866258 M * Hollow it stops the parent until the child releases its virtmem via execve 1139866312 M * daniel_hozac ah, that's clever. 1139866338 Q * bonbons Quit: Leaving 1139866339 M * michal_ but 1139866372 M * michal_ say i have a nice switch there - for pid - case -1, 0 (call child function here), default 1139866396 M * michal_ in default i need to wait, or i'm getting something strange 1139866413 M * Hollow no, with the SIG_IGN you don't need to wait 1139866421 J * mkhl mkhl@200-148-41-143.dsl.telesp.net.br 1139866445 M * michal_ i see. 1139866464 M * michal_ so, got test code, where child is trying to run a shell in a new namespace 1139866487 M * michal_ this shell is something out there, but out of my reach 1139866509 M * Hollow yeah, you forked it away :) 1139866515 M * Hollow (cloned actually) 1139866520 M * michal_ and it's taking 95% cpu ;p 1139866561 M * michal_ ok, i'll try to clone away portmap ;p 1139866945 Q * liquid3649_ Remote host closed the connection 1139867118 M * michal_ ok, looks like this way i am doing what it was suposed to do (cloned away mocp -S lol, i like this name ;) 1139867122 M * michal_ thank you :) 1139867124 Q * jeeves Quit: Leaving 1139867510 J * dearaujo ~dan@pixpat.austin.ibm.com 1139868135 M * cehteh hey derjohn, i want to test if the scratchpad sends a notify hint 1139868318 Q * zobel Ping timeout: 480 seconds 1139868682 J * zobel zobel@zobel.irc.ftbfs.de 1139869060 J * Aiken ~james@tooax6-064.dialup.optusnet.com.au 1139869139 J * julian_tf ~julian_tf@dsl-201-129-245-103.prod-infinitum.com.mx 1139869505 Q * FireEgl Ping timeout: 480 seconds 1139869579 M * julian_tf hi, my host environment has internet connectivity and ip_forwarding enabled, how do i make my guests OSes internet connected 1139869657 M * daniel_hozac you won't need ip_forward for vserver. 1139869676 M * julian_tf what do i need, my host is in a 192.168.1 lan 1139869678 M * daniel_hozac assuming your host has a public IP address, and your guests have private ones, a simple SNAT should do the trick. 1139869695 M * julian_tf my guests are in 10.77.1.0 1139869705 M * julian_tf that would be double nat, right? 1139869713 M * daniel_hozac yep. 1139869732 M * daniel_hozac any particular reason you can't just put the guests on the same network as the host? 1139869748 M * julian_tf there is none 1139869755 M * julian_tf well 1139869773 M * julian_tf it might be possible that we are running short with IPs 1139869794 M * julian_tf but i will check that 1139869819 M * daniel_hozac might want to think about moving to a larger network then ;) 1139869823 M * Skram Hi. 1139869829 M * daniel_hozac hello. 1139869829 M * Skram dearaujo: work for ibm, eh? 1139869833 M * julian_tf that would help 1139869835 M * julian_tf thanks 1139869844 M * Skram The plant up north? 1139869849 M * dearaujo no 1139869851 M * dearaujo austin 1139869867 M * Skram (north austin) 1139869872 M * Skram I am in Austin too. 1139869895 M * dearaujo cool 1139869899 M * Skram Yes. 1139869947 M * dearaujo Skram: what company 1139869971 M * Skram State of Texas.. I am in High School. 1139869987 M * dearaujo lol 1139869990 M * dearaujo cool 1139869994 M * Skram I suppose. 1139870048 M * dearaujo heh 1139870053 M * dearaujo i know what you mean 1139870057 M * Skram Did you grow up in Austin? 1139870074 M * dearaujo no 1139870079 M * Skram Okay. 1139870085 M * dearaujo phx 1139870088 M * Skram Righ on. 1139870093 M * Skram *Right 1139870103 M * derjohn sad that the bat as mascot is already used up. would fit fine to that project - all noctural (just ran over this picture of a bat: http://www.spiegel.de/wissenschaft/erde/0,1518,grossbild-551145-388908,00.html ) 1139870112 M * Skram I know someone who works at the accessibility center at AustinIBM. 1139870124 M * Skram Have done work with a local non profit relating to technology. 1139870142 M * Skram Possibly looking for a summer job, preferrably paying though. 1139870226 M * daniel_hozac derjohn: lol. 1139870242 M * derjohn daniel_hozac, sad that you cant read German .... 1139870280 M * derjohn daniel_hozac, brain weight: 2,5 gr., testicles: 3.5 gr - that's a quota ! :) 1139870282 M * SNy bwahaha 1139870291 M * daniel_hozac haha. 1139870314 M * SNy hey, derjohn, warum fragtest Du eigentlich neulich nach dem LinuxTag? 1139870322 M * derjohn daniel_hozac, incredible animals those bats ... 1139870327 M * derjohn SNy, ja klaro! 1139870347 M * SNy ich mein, ich lurke ja hier nur rum 1139870356 M * derjohn SNy, http://linux-vserver.org/linuxtag2006 <-- trag mal ganz schnell hier ein :) 1139870358 M * Skram Meh, I speak English, some Japanese and some Hebrew, shall I add German to my list? 1139870361 M * SNy ich fahr da schon hin, aber mit dem Projekt nebenan 1139870367 M * SNy ... denk ich mal ... 1139870382 M * Skram I want to go/ 1139870383 M * derjohn Skram, only if you want to join us at Linuxtag/Germany :) 1139870393 M * derjohn SNy, which project? 1139870404 M * SNy Wiesbaden nervt, glaub 1139870406 M * Skram Yeah.. not gonna make it to Germany yet 1139870406 M * Skram geg 1139870409 M * Skram *heh 1139870410 M * SNy KA war so geil immer 1139870417 M * SNy derjohn: xbox-linux 1139870429 M * derjohn SNy, we didnt get a booth yet -> no one did for sure 1139870439 M * derjohn SNy, Nils mailed me ... 1139870448 M * derjohn SNy, our booth will be cooler ;) 1139870467 M * derjohn SNy, join the dark side and together ..... we'll rule the internet :) 1139870468 M * cehteh .. rather hotter with all that hardware 1139870469 M * SNy is that a fact? 1139870470 M * SNy ;p 1139870487 M * derjohn SNy, yes, its fact, we are hotter! 1139870488 J * FireEgl Atlantica@Atlantica.Tcldrop.US 1139870492 M * derjohn :) 1139870515 M * derjohn harry will brings his quadway opteron from belgium 1139870532 M * derjohn 1139870569 M * derjohn SNy, before I post to the ML, I want a (positive) Feedback from LT 1139870585 M * derjohn SNy, then I will mobilze more people.... 1139870607 M * derjohn SNy, on the wiki tere are already some, i am pretty sure the list will grow 1139870615 M * harry hehe 1139870616 A * derjohn doenst want to be alone on LT 1139870640 M * derjohn harry, oops, I paged you ... sry :) 1139870690 M * dhansen ebiederm_zZ: are you really Zz? 1139870690 M * dearaujo Hi - I've got a system running a vserver guest and I modified xinetd in the guest to spawn a ftp daemon when a ftp request is received. For some reason, however, I'm getting this when I try to access the server: 500 OOPS: could not bind listening IPv4 socket - any ideas? 1139870718 M * derjohn bertl said he tries to livecode an a beamer ... n loadsharing algo that would it make possible to run linux-vserver on a beowulf 1139870757 M * daniel_hozac dearaujo: when you connect or when you do something like ls? 1139870771 M * daniel_hozac dearaujo: could you strace it? 1139870780 M * dearaujo daniel_hozac: when I connect 1139870794 M * derjohn :) 1139870818 M * daniel_hozac dearaujo: sure you have configured vsftpd correctly? 1139870874 M * harry derjohn: np 1139870881 M * daniel_hozac dearaujo: i.e. let it know it's running in inetd mode? 1139870892 M * julian_tf guys can i duplicate a vserver by doing this: cp -a /vservers/existingvserver/* /vservers/newskeletonvservercp -a /vservers/existingvserver/* /vservers/newskeletonvserver ?? 1139870906 M * dearaujo it might be comfigured to run in standalone mode 1139870910 M * dearaujo let me check 1139870913 M * julian_tf i meant to paste the command once 1139870923 M * derjohn julian_tf, -p (preservce) or better rsync -azv 1139870927 M * derjohn (symlinks!) 1139870955 M * derjohn julian_tf, and be warned that you neec to chxid after that 1139870958 M * julian_tf derjohn you mean copy with option -p? 1139870960 M * daniel_hozac derjohn: -a includes -p. 1139870962 M * dearaujo daniel_hozac: that was it...... 1139870970 M * dearaujo listen=NO 1139870977 M * julian_tf derjohn: what is chixd? 1139870982 M * dearaujo thanks 1139870985 M * daniel_hozac dearaujo: cool. 1139870990 M * derjohn julian_tf, and: /etc/vservers/ ... dont forget to dup! 1139871014 M * daniel_hozac julian_tf: it's a command to change the xid tagging of files. 1139871032 M * julian_tf i saw in the faq that i needed to create a dummy vserver then cp-a into that vserver, 1139871051 M * derjohn julian_tf, hm, xid is a context for files ... it should be the same as the context id for that giest 1139871054 M * daniel_hozac that is better, yes. 1139871061 M * julian_tf seems easier 1139871061 M * daniel_hozac for creating the configuration. 1139871086 M * daniel_hozac if you have already created the configuration, the cp -a would be enough. 1139871107 M * julian_tf daniel_hozac: if i went that route, do i still need to do that xid thing? 1139871126 M * derjohn julian_tf, depends ... (if you have lots of custom codeschnipsel in you /etc/vserver/foo .... but builing a skeleton is ID 10 T proof. 1139871144 M * julian_tf i need idiotproff 1139871147 M * julian_tf proof 1139871148 M * daniel_hozac julian_tf: only if you have enabled tagging on the file system. 1139871148 M * julian_tf heh 1139871161 M * daniel_hozac (i.e. by default, you won't need it) 1139871170 M * julian_tf daniel: great, cuz, i dont know if i have 1139871192 M * daniel_hozac grep tag /proc/mounts would tell you. 1139871316 M * cehteh derjohn: http://www.myextralife.com/archive.php?date=2006-02-10 1139871372 Q * djudko Quit: Leaving 1139871526 M * derjohn cehteh, added to the wiki. and that one http://people.freebsd.org/~wpaul/daemonette/Dcp00473.jpg 1139871619 M * derjohn (das we would have to go for virtual pengiun dresses) 1139871631 M * derjohn s/das/sad/ 1139871736 M * derjohn cehteh, so more like this? http://www.nylug.org/images/adler_images/linux_nylug_booth.jpg 1139871736 M * julian_tf i dont have /vservers, i have debian, is it the same as /var/lib/vservers ? 1139871744 M * derjohn julian_tf, yes 1139871755 M * derjohn julian_tf, debian puts that in /var/lib/vservers 1139871793 M * derjohn julian_tf, you may symlink it, if you like 1139871800 N * ebiederm_zZ ebiederm 1139871822 M * ebiederm dhansen: Took a quick nap I was up too late last night. 1139871829 M * julian_tf cool 1139871835 M * julian_tf thanks 1139871857 M * julian_tf the copy was succesful 1139871862 P * dearaujo 1139871885 Q * _mountie Ping timeout: 480 seconds 1139871905 J * _mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1139873121 M * julian_tf thanks all, i now have 6 vservers running asterisk/debian, on one box. i will use this box to teach asterisk to a group of people, each of whom will have their own virtual asterisk 1139873162 M * julian_tf for any daemon, i have to prevent that it binds to 0.0.0.0? even if its running in a guest environment? 1139873339 Q * cemil Read error: Connection reset by peer 1139873649 M * Skram asterisk rocks! 1139873681 M * Skram julian_tf: no, if its inside a vps, 0.0.0.0 will bind to the ips or eth0 deviced given to that VServer 1139873686 M * Skram I do the same thing, more or less 1139873694 M * julian_tf cool 1139873709 M * julian_tf well i have spent some 15 minutes editing conf files 1139873713 M * Skram For example, a customer with three IPs can bind to 0.0.0.0 in iax.conf and make it work with all three of their ips. 1139873727 M * julian_tf thats cool to know 1139873771 M * julian_tf i was thinking about if i was a vps provider people will be messing with each others ips 1139873772 M * Skram Yes'sir 1139873777 M * julian_tf but if vserver handles that 1139873778 M * Skram julian_tf: Right. 1139873783 M * Skram Heh, that would be carazy. 1139873809 M * julian_tf i believe i am sharing ztdummy, and since it is virtual theres no problem 1139873818 M * Skram Right 1139873823 M * julian_tf i believe you can still seize others peoples channels 1139873826 M * Skram just make sure its modprobbed on the host. 1139873831 M * Skram I dont think so. 1139873833 M * Skram Im not sure. 1139873836 M * Skram never tried. 1139873857 M * julian_tf the top rated article in google for vserver asterisk discusses it i believe 1139873863 M * Skram Yes 1139873863 M * julian_tf telephreak 1139873868 M * Skram I know the people who wrote it 1139873871 M * julian_tf cool 1139873875 M * Skram I am an administrator of telephreak as well ;) 1139873887 M * Skram Used to do work with John, but he had to leave for personal issues. 1139873889 M * Skram Yes, Yes. 1139873890 M * Skram :) 1139873895 M * julian_tf and you are in high school? 1139873903 M * Skram Yes. 1139873906 M * Skram Busy :) 1139873924 M * julian_tf thats amazing to say the least.. im 25 i feel old hehe 1139873937 M * Skram Heh. 1139873940 M * Skram Welp, Yeah 1139873947 M * Skram I hope to run with it somehow in the future.. 1139873968 M * Skram Me and john were working on some pretty bug business plans, but issues I cannot discuss came up 1139873977 M * julian_tf i see 1139873991 M * julian_tf there is quite a lot of money to be made in asterisk 1139873997 M * Skram Yeapps. 1139874062 M * julian_tf well i gotta run, nice talking to you 1139874072 M * julian_tf thanks all 1139874139 Q * julian_tf Quit: Quitting! 1139874196 M * Skram Tootles