[00:01] yes [00:01] is this still true? [00:01] yes [00:01] i need 2*2months [00:01] and the big one, for 6 months [00:01] I stumbled over a very interesting project or area, but YMMV [00:01] if it is bigger, i might get a friend to join it [00:01] you know I'm using QEMU for testing ... [00:02] yeah, what is it ? [00:02] well, have a look at hist page first ... http://fabrice.bellard.free.fr/qemu/ [00:02] (if you do not already know it ;) [00:03] i heard about it [00:03] okay, I stumbled over some other pages, which deal with emulation and testing/instrumentation ... [00:04] and a friend of mine has written the bintrans binary translator [00:04] a just in time arch independant translator ... [00:04] http://www.complang.tuwien.ac.at/schani/bintrans/ [00:04] how can it be arch independent ? [00:05] well, it uses something calle MLD machine language description [00:05] you basically specify one machine language, and that's it, the bintrans compiler knows how to translate ;) [00:06] so recently I stumbled over a project which does ia64 emulation (for testing) [00:06] http://www.pacificsites.com/~ccrayne/ia64.html [00:06] and I mailed Fabrice, what might be required to get this working with QEMU ... [00:07] and he said, well basically if you have a complete machine description, you are half done ... the other half is hardware simulation ... [00:08] which was it you wanted to combine ? [00:09] so, my idea was, maybe somebody could combine the technique used in bintrans and the emulation QEMU does, to build an emulator which only needs a machine description, and a bunch of hardware simulations ;) [00:10] btw. same issue with 3.3 :( [00:10] dont you think that might be slower than building it for a specific computer ? [00:10] we'll ahve to investigate ... [00:10] Jon, the bintrans is fast, trust me that guy knows what he is doing ... [00:10] yeah, but... i dont ;-P [00:10] I would say, it's about twice as fast as qemu ;) [00:11] well, this 'project' comes in three flavors: [00:11] what about that ia64 emulator then ? [00:12] a) add ia64 emulation to QEMU [00:12] probably really easy, not much hardware changes ... [00:12] why? i dislike the ia64 idea [00:13] i suppose that would give me an idea of qemu [00:13] b) combine QEMU and bintrans idea/code to make it faster and improve the code generation ... [00:13] c) improve one (or more) of the supported archs until it is capable of booting a kernel ;) [00:14] not just any kernel [00:14] a vserver kernel *grin* [00:14] well, THE kernel 8-) [00:15] hehe [00:15] it might be interresting [00:15] well, was just an idea .. so no expectations there ... [00:16] i'll look more into it, since i am planning to look at a projekt tomorrow [00:16] i need to get started soon if i want to finish [00:16] within one year [00:22] back [00:26] hmm, seems the connection is down ... [00:32] no, just slooooow [00:32] okay .. ;) [00:33] it is not on my own line yet [00:43] hum hum, interesting ... [00:44] ah, I know ... [00:47] oooh, fedora kernel with vserver and vquota [00:47] this was nice [00:48] on a laptop as well :/ [00:48] great! [00:48] JonB: you know now what caused the error? [00:48] Bertl: no? [00:49] tanjix (tanjix@pD904A149.dip.t-dialin.net) left irc: [00:49] the switch((long)v) [00:49] in your patch, or the kernel ? [00:49] the debian primitives seem not to support that [00:50] in my patch ... [00:50] and it will prvent compiliation of 2.6 kernels too, I guess ... [00:50] okay [00:51] but the 'workaround' is in place, and I'll post a message on the ml, how to 'fix' that ... [00:52] good [01:12] finally! found out why groff didnt format HTML propperly under solaris. seems by default it passes through all macros used by teh system troff trhough to groff and you have to use groff -gms -gme etc to use the GNU versions. and the system macros are incompatable with the HTML backend deiver. [01:13] took a long time to figure that one out. [01:13] hmm, another useless piece of information? [01:13] ;) [01:13] just thoguht you might be interested in what was making me pull my hair out for the last day and a half. [01:13] preventign me form continueing to work on the howto. [01:13] now that thats takne care of though. [01:14] ah, okay, we pity you? is that correct? [01:14] huh? [01:14] that you needed to spend so much time on that ... [01:15] later [01:15] JonB (~NoSuchUse@kg203.kollegiegaarden.dk) left irc: Quit: Leaving [01:15] talon: we grieve about the resources and amount of time lost for vserver .... [01:15] don't lsiten to me ;) [01:16] does vserver work on sparc? [01:16] last time i tried the testme.sh script it worked. [01:16] okay, thanks ... [01:16] rob is currently working on getting the new switch/vlan setup now. [01:17] once thats done we will plug in the U10. [01:18] no problem, didn't remember if you already tried the vserver stuff on a sparc ... [01:20] yeah. [01:20] one of the first things i tried when i got the machine up. [01:20] was curious myself. [01:20] serving (~serving@213.186.188.205) left irc: Read error: Connection reset by peer [01:20] led to ensc fixing util-vserver for sparc builds. [01:20] hehe ... [01:21] it's always easy to fix stuff, if you know what is broken ... [01:21] (and enrico is especially great at fixing stuff) [01:22] i wasnt trying to make you feel sorry for me. just figured you might be wondering why i havent been around. [01:22] forget about it, was a bad joke, and probably bad english anyway ... [01:23] sounded mroe like sarcasim to me. but i will forget about it. [01:24] I'm sorry if it was received that way ... [01:25] either wya now that my documentation tools are working propperly again. im going back to the howto. i will probably be popping in form time to time to have you take a look at diffrent sections to see if theres anything ive missed. etc. [01:26] np, make it so ... [01:27] i only pulled my hair out so much with groff because im already familiar with teh -ms and -me macro sets. and i can generate those directly into HTML, ascii text and PS/PDF. [01:28] well, I suggested to use the wiki, right? [01:28] the wiki pages are html, and can produce latex, which in turn, can be converted to pdf, and with the use of lynx, the html can be converted to ascii ... [01:29] but I'm sure the groff based docu will look nicer ... [01:31] thats the general idea yes. i hope that ends up being the case. ive never tried to write a document like this before. (aside from another internal document i wrote at a previous job in teh -me macros that documented the network which looked fairly nice) [01:37] im probably going to start with the examples first and then flesh out the background information on context quotas later. [01:37] whatever you prefer ... it's your howto, I'l comment on that, but that is all ... [01:55] fair enough. [02:00] i dont want you to do it all for me or anything, just want to me sure its as accurate and as useful as possible. [02:01] as I said, I'll comment on that, if you find my comments useful, you'll adapt the howto, if not ... so be it ... ;) [02:29] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) left irc: Remote host closed the connection [02:41] serving (~serving@213.186.188.205) joined #vserver. [02:42] hi serving! [02:43] Bertl: teh most recent patches are aganst rc2 now right? [02:43] as far as q0.13 and vs 1.26? [02:43] hmm, for 2.4 yes ;) [02:48] Hi Bertl :) [02:49] Last message repeated 1 time(s). [02:49] sladen (paul@starsky.19inch.net) left irc: Ping timeout: 492 seconds [02:52] so how are you, serving? [02:56] hi serving [02:56] :) [02:56] hi all [02:56] hi Cmaj! [02:56] not a robot :) [02:56] good idea for some music ... let me see, David Bowie ;) [02:57] well trying to get elvis costello for all days [03:01] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) joined #vserver. [03:01] hi rick! [03:01] HI! [03:02] whats new [03:02] hmm, let's see, a new 1.26 release for 2.4.25-rc2 [03:03] a new SMP test system (property of JonB) for vserver tests [03:03] awesome [03:03] i didn't see 2.4.25 on kernel.org [03:03] let me check again [03:03] some compilers for architectures, you ahvenÄt even heard from ... [03:04] cool [03:06] I have a question. If I create a vserver with only specific dev links, what stops the vserver root user from creating new ones? [03:07] maybe i need to read exactly how the dev's work [03:11] well, it's the capability system, which prevents this ... [03:12] #define CAP_MKNOD 27 [03:12] Allow the privileged aspects of mknod() [03:13] Is that a setting in the /etc/vserver/server.conf? [03:13] or is it already set [03:13] you can make some tests yourself, do you have a vserver at hand? [03:13] yes [03:13] okay even access to the vserver host I assume? [03:14] yes both [03:14] i am root on both [03:14] okay, on the host do [03:14] grep Cap /proc/self/status [03:14] those are the 'capabilities' encoded as 'bits' [03:15] I see [03:15] chcontext --secure grep Cap /proc/self/status [03:15] that is the subset a vserver gets as default [03:15] does chcontext come from the utils package? [03:16] yes, it does, in util-vserver it's part of the core utilities .. [03:18] Ok I ran that command [03:19] hmm new quota-tools release. [03:19] hum, newer than 3.11? [03:19] Action: talon tests the new quotacheck. [03:19] Bertl: no i mean 3.11 [03:19] was 3.10 last i looked at it. [03:19] ah okay, yeah honza released that one, should contain the fix ... [03:20] thats whast teh changelog says. [03:20] read it on lkml ... [03:20] should I look at this page http://www.linux-vserver.org/ for the latest software releases? [03:21] it seems like there are a lot of pages for the project [03:21] should always be up to date ... [03:21] ok [03:21] well it's a community project, what do you expect ;) [03:21] hehe i know :) [03:23] Bertl: just looking at how his fix differs from mine and if i can get them to disagree in any way. [03:23] okay, good thinking ... [03:23] vs ext2 direct io on a test fs. [03:32] at least i will after dinner anyway. i already tested it inside a vserver instance and it seems to be slightly different at counts than mine. need ot test it more in depth to see which one is truly correct. [03:33] if i do think ive found a real problem with 3.11 i will ofcourse let honza know about it. [03:50] hmm. heh [03:50] looks like it isnt a complete fix. [03:50] his fix still deosnt match up with what direct io reports. [03:51] Action: talon makes sure his patch does match up. [04:04] Action: talon plays with it a bit more [04:05] and, any insights so far? [04:07] Bert, could you send me that list of necessary devs again? [04:07] sure, you installed the util-vserver tools now? [04:07] yes .29 [04:08] okay, just for a test, start the 'vserver test2 build' again ... [04:08] ok [04:08] or 'test3' [04:09] does this finish now faster? [04:09] _shur1 (~shushushu@vserver.electronicbox.net) joined #vserver. [04:09] i am going to just put the necessary files in the vserver, because some of my host root directories are lange [04:10] large [04:11] crw-rw-rw- 1 root root 1, 7 Apr 6 2003 full [04:11] srw-rw-rw- 1 root root 0 Dec 14 17:29 log= [04:11] crw-rw-rw- 1 root root 1, 3 Apr 6 2003 null [04:11] crw-rw-rw- 1 root root 5, 2 Feb 15 02:12 ptmx [04:11] drwxr-xr-x 2 root root 0 Dec 14 17:27 pts/ [04:11] crw-r--r-- 1 root root 1, 8 Apr 6 2003 random [04:11] crw-rw-rw- 1 root root 5, 0 Feb 13 13:57 tty [04:11] crw-r--r-- 1 root root 1, 9 Apr 6 2003 urandom [04:11] crw-rw-rw- 1 root root 1, 5 Apr 6 2003 zero [04:11] the command should create a minimal server which contains only those entries ... [04:12] does the command "vserver create test3" do anything except create the directory and cp the directories from the host? meaning, can I do it manually [04:12] instead of using the vserver command [04:13] well it should do some things, but it should not copy the host ... [04:13] i mean build [04:13] at least not in the util-vserver version ... [04:14] when I ran vserver test3 build, I could see: 9725 pts/2 R 0:00 \_ cp -ax /sbin /bin /etc /usr /var /lib /vservers/temp/. [04:14] and that will take a long time [04:14] copying unnecessary files [04:14] hmm, let me verify that ... [04:14] ok [04:19] hmm, ah probably that was in the alpha branch then ... [04:19] should I use v.28? [04:19] sorry, obviously got that messed up, but I'm sure there is an option for it ... [04:19] let me check ... [04:19] ok [04:22] if test "$UTIL_VSERVER_AVOID_COPY"; then [04:22] that is the magic word for the stable branch ;) [04:23] ahh [04:23] should I be using v.29 then? [04:23] okay try: vserver build test3 UTIL_VSERVER_AVOID_COPY=1 [04:23] or even better [04:23] export UTIL_VSERVER_AVOID_COPY=1 [04:23] gotcha [04:24] that should give you an empty layout ... [04:24] export that and then vserver test3 build, correct? [04:24] yup [04:24] noel- (~noel@pD9E09E31.dip.t-dialin.net) joined #vserver. [04:25] nice that was a Lot faster [04:25] yeah, you can put those config options into the config file ... [04:25] : ${UTIL_VSERVER_VARS:=$(dirname $0)/util-vserver-vars} [04:25] . "$UTIL_VSERVER_VARS" [04:25] so that is sourced ... [04:25] /etc/vservers.conf [04:26] that file does not exist. I have a backup from when I was not using the utils, but utils did not cerate that file [04:26] well, yes that is read too, but util-vserver uses the util-vserver-vars for it's configs ... not for the vserver defaults ... [04:27] /usr/sbin/util-vserver-vars [04:27] /usr/lib/util-vserver/util-vserver-vars [04:28] got it [04:28] /usr/local/lib/util-vserver [04:28] cool [04:29] enrico is currently writing some docu, not finished yet ... [04:29] PKGLIBDIR='/usr/local/lib/util-vserver' [04:29] SBINDIR='/usr/local/sbin' [04:29] VROOTDIR='/vservers' [04:29] UTIL_VSERVER_AVOID_COPY=1 [04:29] yeah that should do the trick ... [04:30] the alpha tools provide some nice ways to get a new server installed too ... [04:30] http://www.linux-vserver.org/index.php?page=alpha+util-vserver [04:30] if you are interested ;) [04:30] you know I am [04:32] basically they create complete servers out of thin air ... [04:32] yes I see [04:32] noel_ (~noel@pD9E098E4.dip.t-dialin.net) left irc: Ping timeout: 504 seconds [04:32] because right now this doesnt copy any programs [04:32] and using the link you just showed me, it creates a distro [04:33] yes, as I said, those are the alpha (means development) branch [04:33] 0.28 and 0.29 are the stable utils ... [04:34] but one decent way to get servers nice and easy is to build a template server, and copy that for every new server you need ... [04:34] yes that is what I think I am going to do next [04:34] that, or use this alpha version [04:34] and once a server is working, you can move it from one host platform to the next, with no change ... [04:35] (well asumed that the underlying arch is the same) [04:35] yes [04:35] but you can move from a debian host to a redhat one and the other way around ... [04:35] yes I like the concept a lot [04:36] last time you mentioned testing quota too, right? [04:36] I am not sure where to start with that one. I was having trouble finding the right doc for quota limits [04:36] and IIRC I was explaining some basics about the file tagging ... [04:37] yes [04:37] is there some paper written on this [04:37] I would like to give you a short overview .. [04:37] grea [04:37] tt [04:37] there are some older documents ... [04:38] and talon is currently working on a complete howto ... you might be interested in testing that one? [04:38] yes definately [04:38] first let me give you a list of relevant docu ... [04:39] http://vserver.13thfloor.at/Linux2.6/index.php?page=Per+Context+Quota [04:39] http://vserver.13thfloor.at/Linux2.6/index.php?page=Per+Context+Disk+Limits [04:39] bookmark, bookmark [04:40] http://www.13thfloor.at/old/VServer/ [04:40] http://www.13thfloor.at/old/VServer/Concepts.shtml [04:40] http://www.13thfloor.at/old/VServer/HowTo.shtml [04:40] those are all outdated, but contain the basics and some additional info ... [04:41] the quota project is actually about a year and a half old ... [04:41] Ok I will read those [04:41] thank you [04:41] what you should know is the folowing: [04:42] a) there is context quota and per context disk limits [04:42] b) both requires your files to be tagged with the context (xid) information [04:42] c) both are per context _and_ per filesystem [04:43] ok [04:43] where the concept is simple, the context tagging does modify your filesystem so make sure that you do not change to much at a time ... [04:44] the methods used to tag files with xids fall in two categories ... [04:44] 1) those who use existing inode structures [04:44] 2) those who use yet unused disk space ... [04:45] the second category is simpler to use if you want to switch between tagged and untagged mode, the first is more 'compatible' and works on more filesystems ... [04:45] I see [04:46] reiserfs and ext3, are those compatile [04:46] the second only works for ext2/ext3 [04:46] ok [04:46] the first works for all fs [04:46] it basically reuses some parts of the uid/gid for the xid ... [04:47] that is the magic behind this tagging method ... there is for example UID24/GID24 which uses the upper quarter of both uid/gid for the xid (which is 16bit) [04:48] and you have to specify a special mount option (tagxid,tagctx) to get this feature ... [04:48] I saw those options in my kernel config, i wasnt sure which one to pick [04:48] psax (psax@huset.festnett.com) joined #vserver. [04:48] but currently I'd suggest to use a separate partition for the /vservers to avoid problems .... [04:48] hi psax! [04:49] ok I will work on that [04:49] thanks for the briefing [04:49] Hello Bertl :) [04:49] rick: you're welcome, have fun! [04:49] and talk to talon, if you want to testdrive his howto ... [04:50] ok I will [04:56] When i start a vserver, i only get segfault. Is this a known problem, or is it only my machine? ;) running linux-2.4.25-rc2-vs1.26 [04:56] that should not happen ... but it sounds like a bad setup or even hw trouble, interested in investigating? [04:58] Yeah, sure. :) [04:58] http://vserver.13thfloor.at/Stuff/testme.sh [04:58] okay give that a spin on the host [04:59] both failed [04:59] could you paste me the output, or make it available somewhere? [05:00] hmm, okay, this was on the host, not inside a vserver? [05:01] yeah, I cannot start the vserver.. when i try like vserver dragon start, it Segfaults.. [05:01] okay, lets check something ... [05:01] what does 'ldd `which chcontext`' output? [05:02] libc.so.6 => /lib/libc.so.6 (0x4001c000) [05:02] /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) [05:02] hmm, that looks good, okay, next test ... [05:02] chcontext --ctx 1 true [05:02] segfault [05:03] okay, do you have strace installed? [05:03] yup [05:03] does 'strace true' work? [05:03] (I need no output of that) [05:03] yes, it works [05:04] okay, lets try the same with the chcontext --ctx 1 true [05:05] http://prodigy.fadeworks.com/misc/vserver/output.txt [05:06] I/O error, hmm, do you use any security patches? [05:06] nope [05:07] not that i can remember applying at least [05:07] okay, lets try again with 'strace -fF chbind --ip 127.0.0.1 true' [05:08] http://prodigy.fadeworks.com/misc/vserver/output2.txt [05:09] same err it seems [05:09] yeah [05:09] what version of strace is this? [05:10] 4.2 [05:10] hmm [05:10] perhaps i should get a new one :) [05:11] Give me a sec [05:11] well, it might help to use strace >= 4.5 as this shows syscalls too ... [05:13] there, outputs updated [05:13] btw, does dmesg mention anything unusual? [05:14] whoa, yeah.. [05:14] lets see/hear [05:14] http://prodigy.fadeworks.com/misc/vserver/dmesg.txt [05:14] wow, cool! [05:16] okay, could you do the following: cat /proc/version [05:16] Linux version 2.4.25-rc2-vs1.26 (root@prodigy.fadeworks.com) (gcc version 2.95.3 20010315 (release)) #1 SMP lør feb 14 17:59:09 GMT+1 2004 [05:17] hmm, maybe the compiler is the issue, but maybe not ... [05:17] okay, do you have ksymoops installed? [05:18] or could you install it ;) [05:18] okay. Nope, do not think so, never heard of it anyway.. [05:18] Sure :) [05:18] ksymoops will help to decipher the kernel output ... [05:19] there, installed [05:19] okay, you use it this way: [05:19] ksymoops -v vmlinux -m System.map [05:19] where vmlinux is the kernel and System.map is the appropriate map [05:20] then you pipe in the dmesg output, and it outputs a decoded version ... [05:20] okay [05:20] that should look similar to the dmesg output, but with more information in it ... [05:21] http://prodigy.fadeworks.com/misc/vserver/out3.txt [05:23] could you give me some details about that machine? [05:25] Sure, what do you want to know? like athlon 1600+, [05:25] or software? [05:26] hmm, okay, that is a little strange ... [05:27] the system works 'normal' for other tasks (non vserver related)? [05:28] yeah, been stable since i don't remeber [05:28] hmm, well, you have a 'new' kernel now ,right? [05:28] yeah, installed today [05:29] okay, did you test anything with that kernel (except for vserver)? [05:29] been running on 2.4.22-ac2 for quite some time though [05:29] no, that i did not [05:29] okay, could you just test some apps/tools you normally use ... [05:29] ok, hang on [05:30] nice ipv4root is now 127.0.0.1 [05:30] is it this :) [05:30] i try to do the same here :) [05:30] well, for a moment yes, if you use the chroot ... true ;) [05:32] well i try to learn ,so never used any of these [05:32] those are the core tools .. basically chroot, chcontext and chbind ... [05:32] Seems okay. All server/desktop apps seems to work like before. [05:32] okay, what does cat /proc/self/status output? [05:33] http://prodigy.fadeworks.com/misc/vserver/status.txt [05:34] that looks sane to me ... [05:36] where do you have the vserver patches from? [05:37] http://www.13thfloor.at/vserver/s_release/v1.26/ [05:37] okay, it seems like ther userspace/kernel interface is somehow broken ... [05:37] but only for the vserver syscall ... which is a little strange ... [05:38] what distro is that, and does it enforce some security stuff via preloaded libraries? [05:39] The distro is called JBLinux, pretty old actually. I do not know if it does that, if you want, and if it may help your development, you may explore it if you like.. [05:40] well, I'm not sure that this is the reason for what we have here ... [05:40] what patch did you put first :) maybee [05:40] do me a favour and update the gcc/binutils to something more recent and recompile the tools (util-vserver-0.29) [05:41] okay, will do [05:41] (just to make sure that this isn't the cause) [05:41] did the 25-rc2 first, then the vs [05:42] won't work the other way round ... [05:42] oky mine is neat 2.4.99-rc2-vs1.26 [05:43] TheSeer (~theseer@border.office.salesemotion.net) left irc: Ping timeout: 492 seconds [05:43] TheSeer (~theseer@border.office.salesemotion.net) joined #vserver. [05:56] ouch, got some errors compiling gcc.. its 4am here now, so i'll think i'll look more into it in the morning :) thanks for your help! gnight! [05:56] good night! it's 4am here too ;) [06:02] true [06:02] night [06:02] ;) [06:02] have a good one! [06:03] you too [06:16] okay, guess I'll call it a day too ... [06:16] have a nice wossname, everyone ... cu tomorrow ... [06:16] Nick change: Bertl -> Bertl_zZ [08:05] Nick change: talon -> talon_zz [08:16] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) left irc: Ping timeout: 480 seconds [08:28] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) joined #vserver. [08:55] rmoriz (rmoriz@rmoriz.cpan.de) got netsplit. [08:55] maharaja (maharaja@213.158.115.2) got netsplit. [08:55] riel (~riel@riel.netop.oftc.net) got netsplit. [08:55] jap (~root@jap.globe.cz) got netsplit. [08:55] Medivh (ck@62.93.217.199) got netsplit. [08:55] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) got netsplit. [08:55] psax (psax@huset.festnett.com) got netsplit. [08:55] serving (~serving@213.186.188.205) got netsplit. [08:55] Doener (~doener@pD958818B.dip.t-dialin.net) got netsplit. [08:55] click (click@gonnamakeyou.com) got netsplit. [08:55] [HvD] (~guess@62.99.252.14) got netsplit. [08:55] talon_zz (talon@host-63-149-223-100.irwinresearch.com) got netsplit. [08:55] tom9 (~tom@pc-3741.ethz.ch) got netsplit. [08:55] _shur1 (~shushushu@vserver.electronicbox.net) got netsplit. [08:55] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) got netsplit. [08:55] noel- (~noel@pD9E09E31.dip.t-dialin.net) got netsplit. [08:55] mids (mids@mids.student.utwente.nl) got netsplit. [08:55] meebey (meebey@meebey.net) got netsplit. [08:55] cgone (~chrisw@fw.osdl.org) got netsplit. [08:55] mcp (~hightower@wolk-project.de) got netsplit. [08:58] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) returned to #vserver. [08:58] psax (psax@huset.festnett.com) returned to #vserver. [08:58] noel- (~noel@pD9E09E31.dip.t-dialin.net) returned to #vserver. [08:58] _shur1 (~shushushu@vserver.electronicbox.net) returned to #vserver. [08:58] serving (~serving@213.186.188.205) returned to #vserver. [08:58] rmoriz (rmoriz@rmoriz.cpan.de) returned to #vserver. [08:58] Doener (~doener@pD958818B.dip.t-dialin.net) returned to #vserver. [08:58] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) returned to #vserver. [08:58] click (click@gonnamakeyou.com) returned to #vserver. [08:58] [HvD] (~guess@62.99.252.14) returned to #vserver. [08:58] talon_zz (talon@host-63-149-223-100.irwinresearch.com) returned to #vserver. [08:58] maharaja (maharaja@213.158.115.2) returned to #vserver. [08:58] riel (~riel@riel.netop.oftc.net) returned to #vserver. [08:58] Medivh (ck@62.93.217.199) returned to #vserver. [08:58] jap (~root@jap.globe.cz) returned to #vserver. [08:58] tom9 (~tom@pc-3741.ethz.ch) returned to #vserver. [08:58] mids (mids@mids.student.utwente.nl) returned to #vserver. [08:58] meebey (meebey@meebey.net) returned to #vserver. [08:58] mcp (~hightower@wolk-project.de) returned to #vserver. [08:58] cgone (~chrisw@fw.osdl.org) returned to #vserver. [09:06] jap (~root@jap.globe.cz) got netsplit. [09:06] maharaja (maharaja@213.158.115.2) got netsplit. [09:06] rmoriz (rmoriz@rmoriz.cpan.de) got netsplit. [09:06] riel (~riel@riel.netop.oftc.net) got netsplit. [09:08] rmoriz (rmoriz@rmoriz.cpan.de) returned to #vserver. [09:08] maharaja (maharaja@213.158.115.2) returned to #vserver. [09:08] riel (~riel@riel.netop.oftc.net) returned to #vserver. [09:08] jap (~root@jap.globe.cz) returned to #vserver. [10:30] Medivh (ck@62.93.217.199) got netsplit. [10:30] tom9 (~tom@pc-3741.ethz.ch) got netsplit. [10:30] talon_zz (talon@host-63-149-223-100.irwinresearch.com) got netsplit. [10:30] [HvD] (~guess@62.99.252.14) got netsplit. [10:30] click (click@gonnamakeyou.com) got netsplit. [10:30] Doener (~doener@pD958818B.dip.t-dialin.net) got netsplit. [10:30] serving (~serving@213.186.188.205) got netsplit. [10:30] psax (psax@huset.festnett.com) got netsplit. [10:30] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) got netsplit. [10:30] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) got netsplit. [10:30] _shur1 (~shushushu@vserver.electronicbox.net) got netsplit. [10:30] noel- (~noel@pD9E09E31.dip.t-dialin.net) got netsplit. [10:30] meebey (meebey@meebey.net) got netsplit. [10:30] mids (mids@mids.student.utwente.nl) got netsplit. [10:30] cgone (~chrisw@fw.osdl.org) got netsplit. [10:30] mcp (~hightower@wolk-project.de) got netsplit. [10:31] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) returned to #vserver. [10:31] psax (psax@huset.festnett.com) returned to #vserver. [10:31] noel- (~noel@pD9E09E31.dip.t-dialin.net) returned to #vserver. [10:31] _shur1 (~shushushu@vserver.electronicbox.net) returned to #vserver. [10:31] serving (~serving@213.186.188.205) returned to #vserver. [10:31] Doener (~doener@pD958818B.dip.t-dialin.net) returned to #vserver. [10:31] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) returned to #vserver. [10:31] click (click@gonnamakeyou.com) returned to #vserver. [10:31] [HvD] (~guess@62.99.252.14) returned to #vserver. [10:31] talon_zz (talon@host-63-149-223-100.irwinresearch.com) returned to #vserver. [10:31] Medivh (ck@62.93.217.199) returned to #vserver. [10:31] tom9 (~tom@pc-3741.ethz.ch) returned to #vserver. [10:31] mids (mids@mids.student.utwente.nl) returned to #vserver. [10:31] meebey (meebey@meebey.net) returned to #vserver. [10:31] mcp (~hightower@wolk-project.de) returned to #vserver. [10:31] cgone (~chrisw@fw.osdl.org) returned to #vserver. [10:36] Medivh (ck@62.93.217.199) got netsplit. [10:36] tom9 (~tom@pc-3741.ethz.ch) got netsplit. [10:36] talon_zz (talon@host-63-149-223-100.irwinresearch.com) got netsplit. [10:36] click (click@gonnamakeyou.com) got netsplit. [10:36] Doener (~doener@pD958818B.dip.t-dialin.net) got netsplit. [10:36] psax (psax@huset.festnett.com) got netsplit. [10:36] [HvD] (~guess@62.99.252.14) got netsplit. [10:36] serving (~serving@213.186.188.205) got netsplit. [10:36] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) got netsplit. [10:36] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) got netsplit. [10:36] _shur1 (~shushushu@vserver.electronicbox.net) got netsplit. [10:36] noel- (~noel@pD9E09E31.dip.t-dialin.net) got netsplit. [10:36] mids (mids@mids.student.utwente.nl) got netsplit. [10:36] meebey (meebey@meebey.net) got netsplit. [10:36] cgone (~chrisw@fw.osdl.org) got netsplit. [10:36] mcp (~hightower@wolk-project.de) got netsplit. [10:36] jap (~root@jap.globe.cz) got netsplit. [10:36] maharaja (maharaja@213.158.115.2) got netsplit. [10:36] rmoriz (rmoriz@rmoriz.cpan.de) got netsplit. [10:36] riel (~riel@riel.netop.oftc.net) got netsplit. [10:36] cgone (~chrisw@fw.osdl.org) returned to #vserver. [10:36] mcp (~hightower@wolk-project.de) returned to #vserver. [10:36] meebey (meebey@meebey.net) returned to #vserver. [10:36] mids (mids@mids.student.utwente.nl) returned to #vserver. [10:36] tom9 (~tom@pc-3741.ethz.ch) returned to #vserver. [10:36] Medivh (ck@62.93.217.199) returned to #vserver. [10:36] talon_zz (talon@host-63-149-223-100.irwinresearch.com) returned to #vserver. [10:36] [HvD] (~guess@62.99.252.14) returned to #vserver. [10:36] click (click@gonnamakeyou.com) returned to #vserver. [10:36] Cmaj (~cmaj@3ffe:bc0:5f3:1:9999:911:c3d3:5431) returned to #vserver. [10:36] Doener (~doener@pD958818B.dip.t-dialin.net) returned to #vserver. [10:36] serving (~serving@213.186.188.205) returned to #vserver. [10:36] _shur1 (~shushushu@vserver.electronicbox.net) returned to #vserver. [10:36] noel- (~noel@pD9E09E31.dip.t-dialin.net) returned to #vserver. [10:36] psax (psax@huset.festnett.com) returned to #vserver. [10:36] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) returned to #vserver. [10:36] rmoriz (rmoriz@rmoriz.cpan.de) returned to #vserver. [10:36] maharaja (maharaja@213.158.115.2) returned to #vserver. [10:36] riel (~riel@riel.netop.oftc.net) returned to #vserver. [10:36] jap (~root@jap.globe.cz) returned to #vserver. [10:44] johnny (~johnny@ip68-10-185-29.hr.hr.cox.net) left irc: Ping timeout: 501 seconds [10:56] kestrel (athomas@home.swapoff.org) joined #vserver. [10:56] hi [11:13] jap (~root@jap.globe.cz) left irc: Remote host closed the connection [11:13] jap (~root@jap.globe.cz) joined #vserver. [11:13] johnny (~johnny@ip68-10-185-29.hr.hr.cox.net) joined #vserver. [11:31] johnny (~johnny@ip68-10-185-29.hr.hr.cox.net) left irc: Ping timeout: 501 seconds [12:12] Doener_zZz (~doener@pD9E12951.dip.t-dialin.net) joined #vserver. [12:14] JonB (~NoSuchUse@kg203.kollegiegaarden.dk) joined #vserver. [12:20] Doener (~doener@pD958818B.dip.t-dialin.net) left irc: Ping timeout: 480 seconds [12:31] rick (~rick@adsl-065-083-169-002.sip.mco.bellsouth.net) left irc: Remote host closed the connection [13:08] sladen (~paul@starsky.19inch.net) joined #vserver. [14:13] JonB (~NoSuchUse@kg203.kollegiegaarden.dk) left irc: Ping timeout: 488 seconds [14:52] JonB (~NoSuchUse@kg203.kollegiegaarden.dk) joined #vserver. [15:11] psax (psax@huset.festnett.com) left irc: Read error: Connection reset by peer [15:38] rick12321 (~rick12312@113.192.33.65.cfl.rr.com) left irc: Quit: Client Exiting [15:46] mhepp (~mhepp@r72s22p13.home.nbox.cz) joined #vserver. [16:07] mhepp (~mhepp@r72s22p13.home.nbox.cz) left irc: Remote host closed the connection [16:37] sim0n_ (~simon@cable-212.76.255.43.coditel.net) joined #vserver. [16:37] hi [16:37] hi [16:38] can someone please explain me what this exactly means "Unable to handle kernel paging request at virtual address 00473a10" ? [16:39] paging means swapping in or out [16:41] if my kernel outputs that error including some other stuff...does that mean that something is wrong with my swap partition, or ? [16:41] you take a page, or more of memory, and move it to or from the ram from or to the disk [16:41] so, this one tells you that it can not perform that paging request at a certain virtual address [16:41] sim0n_: no, it does not need to mean that [16:41] it could be that the address is in use, or already swapped, or ... [16:43] hmm... i got this brand new server, but the ram hasn't arrived yet. so i took an older one. whenever i compile stuff, i get random segfaults. sometimes they occur sometimes not. [16:44] the cpu is not overheating, and memtest ran without errors [16:44] but sometimes i get those kernel errors as well [16:44] do you have an idea what my problem might be? [16:44] bad ram [16:44] bad cpu [16:44] bad timings [16:44] try to change the timings [16:46] uh? what exactly should i change? i read things like downclocking the ram? [16:47] kramer (~kramer@80.86.103.47) joined #vserver. [16:47] kramer (~kramer@80.86.103.47) left irc: Client Quit [16:48] correct [16:48] kramer (~kramer@80.86.103.47) joined #vserver. [16:48] read some howto on memory tuning [16:48] i'm sure there are some out there [16:49] what you need to do is make your memory SLOW, by adding the ammount of cycles there is waited before the result can be red [16:50] i hope it's the ram :-/ [16:50] i set all delays to maximum [16:51] try that [16:51] there are other things as well [16:52] ok...thnx for your help! [16:52] no problemo [17:15] i've compiled some rather large programs now and there haven't been any segfaults so far [17:17] now does this mean the ram is broken? because i don't think i should have to set those timings manually, instead they should be autodetected, or am i wrong? [17:17] Nick change: Doener_zZz -> Doener [17:17] hi [17:17] hi [17:19] what kind of system is that? [17:22] p4 2.6ghz, 512ram (nonecc), 3ware-sata-raid 2x160gb, tyan tomcat i875p [17:24] sim0n_: timings are not always auto detected [17:24] sim0n: memtest86 [17:24] Zoiah: read your history [17:24] Sorry, I only have about an hour of history and didn't find memtest86 there. [17:24] Ahh. [17:24] memtest ran without errors [17:24] :-)...i did run a full memtest, no errors found [17:24] Then you can be 99% sure the RAM is fine. [17:25] Probably the CPU then. [17:25] sim0n_: read some documentation about how a computer system works [17:25] but increasing the ram delays to a max, seems to have solved the problem ? [17:25] so it's x86... does the ram match the fsb of the system? often pc266 modules, for example, run quite stable at 166mhz when timings are adjusted [17:25] yes [17:25] its a pc266 module [17:26] JonB: i DO know how a computer system works [17:26] sim0n_: then you would know about ram timings [17:26] i just havent encountered such problems yet.. [17:26] i know, but "normally" they are set automatically, and one doesnt have to bother about it [17:26] Doener: of course they do, they make ram just like cpu's, they test them at a faster rate, and if they fail, they clock it down until it pass [17:27] at least for my experience [17:27] newer p4 system's have a fsb800 (200mhz quad-pumped) [17:27] sim0n_: correct [17:27] sim0n_: but sometimes it is needed [17:28] ok [17:28] sim0n_: i'm sorry that i suggested you didnt know anything [17:28] i leave the current setings and as soon as i receive the new ram modules i hope these problems are gone [17:28] JonB: np :-D [17:28] sim0n_: ecc ? [17:29] of course :-) [17:29] JonB: i know, just an explanation for sim0n_ as he seemed to be (and is) using modules that are specified for fsb266 which have been used in earlier p4 boxes... [17:29] Doener: okay [17:44] Nick change: talon_zz -> talon [18:12] kramer (~kramer@80.86.103.47) left irc: Quit: Leaving [18:12] hmm thats odd. running into strange NFS problems on one of my vserver patched boxes. trying to rm a file causes a kernel ooops and locks the nfs mount. [18:13] trying to isolate which patchsets cause that to pop up or if its just a 2.4.25 vanilla bug. [18:13] kramer (~kramer@80.86.103.47) joined #vserver. [18:28] Khahan (~Filbert@D577451D.kabel.telenet.be) joined #vserver. [18:50] sim0n_ (~simon@cable-212.76.255.43.coditel.net) left irc: Quit: Client exiting [19:11] kramer (~kramer@80.86.103.47) left irc: Quit: Leaving [19:18] Doener_zZz (~doener@pD9588ADB.dip.t-dialin.net) joined #vserver. [19:22] Doener (~doener@pD9E12951.dip.t-dialin.net) left irc: Ping timeout: 501 seconds [20:16] _shur1 (~shushushu@vserver.electronicbox.net) left irc: Read error: Connection reset by peer [20:17] kramer (~kramer@80.86.103.47) joined #vserver. [20:24] _shur1 (~shushushu@vserver.electronicbox.net) joined #vserver. [20:46] Nick change: Bertl_zZ -> Bertl [20:46] hi everyone! [20:47] Bertl: WB [20:47] Warner Brothers? [20:48] Welcome Bertl [20:48] ah? ;) [20:48] how are you jon? [20:49] i'm relaxed, ready for the next week [20:50] Linux dualp200mmx 2.4.25-rc2-vs1.3.7 #8 SMP Sat Feb 14 23:30:17 CET 2004 i586 unknown [20:50] nice [20:51] so we 'just' have to start some testing, once the console is up ... [20:51] allright, i'll install cu [20:52] that was a statement not a public incitement ... [20:55] hehe [20:55] if i dont do it now, i wont get it done [20:57] remember how long time it toke me to get the server done ? [20:59] only a few moments compared to the age of the universe 8-) [20:59] yeah, well in my universe i am the center, and it took too long time [21:12] Bertl: i might have another bug for you in a minute. [21:12] hey great, a vserver or quota bug? [21:13] vserver realted i think. [21:14] gotta try a vanilla rc5 kenrel firts to be sure. [21:14] but basicly runnign rm on a file thats on nfs causes a kernel opps [21:14] and locks up that nfs fs. [21:14] hmm, interesting ... [21:24] hmm ok cant reproduce it on a clean kernel. [21:24] will start from the bme patch and then on to vs and quota. [21:24] Bertl: damn, cu is only in testing and unstable :/ [21:31] Bertl: do you think this one is useable ? http://packages.debian.org/stable/net/modemu [21:32] hmm, well yes, but looks to me like the other way around ... [21:34] hmm [21:37] ahh, it is a part of uucp [21:37] hmm, what cu, yes, didn't I mention that? [21:37] if you did i wasnt looking [21:39] so, got it installed [21:42] Bertl: ~! command [21:42] Run command in a shell. If command is empty, starts up a shell. [21:42] damn [21:43] which means? [21:43] that potentialy anyone getting access to cu can get access to the host running it [21:44] hogwash, use a chroot environment ... with no shell ;) [21:44] how safe was it chroot was recently ? [21:45] well, if there is no shell, it can't be started, so no exploit ;) [21:45] hmm [21:45] special user, no rights except for serial line ... [21:47] or just patch the cu, to not allow that ... [21:47] i'm looking at netcat [21:47] netcat is something different ... [21:48] not useable for bidirectional communications [21:49] netcat < /dev/ttyS0 > /dev/ttyS0 ? [21:51] ser2net - Allows network connections to serial ports [21:51] yes, that sounds good ... [21:52] i was thinking of just giving you a port to telnet to [21:54] The program does not do UUCP file locking and it won't play nice with getty. [21:57] what shall I say, I would create a restricted user, which can execute minicom or the parport tool, maybe with different ssh commands, and that's it ... [21:57] but it's your site, and your decision ;) [21:58] it looks like what i need, except the getty part [21:58] Bertl: is plain telnet access good enough, or should it be more secured ? [21:58] i mean, how often are you going to use the anyway ? [21:58] different ssh commands ? [21:59] Bertl: yes, but if my security is too high, you wont use it [21:59] well, very often, every time the kernel gives a panic for example ... [21:59] hmm [21:59] you can configure ssh accounts to execute a specific command [21:59] anyway, i gotta go now, i'll be back later on [21:59] okay, cu ... [22:00] haha, [22:00] cu [22:00] haha [22:45] ok so its not the bme patch. [22:45] Action: talon tries teh vs patch. [22:46] we modify the nfs somewhat, so it might cause some troubles ... [22:47] mhepp (~mhepp@r72s22p13.home.nbox.cz) joined #vserver. [22:47] hi mhepp! [22:47] hi Bertl [22:48] I am looking to the topic and there is a new experimental version... [22:48] Bertl: could you try moutning an nfs export and creating adn then runnign rm on a file? [22:48] its fine! [22:50] talon: hmm, please elaborate ... which side is the nfs? [22:51] I mean which side is the server, patched or unpatched or both patched? [23:11] JonB (~NoSuchUse@kg203.kollegiegaarden.dk) left irc: Ping timeout: 501 seconds [23:36] bertl: unpatched server. [23:36] patched client. [23:36] and the client fails, right? [23:37] yeah. well sort of the rm succeds in removing the file on teh server, but it generates an opps and causes rm to segfault. [23:37] then teh nfs mount locks on the client. [23:37] could you ran that oops through ksymoops? [23:37] s/ran/run/ [23:39] yeah. [23:39] hopefully it sheod up in the messages file. [23:39] probably better to wait tin i have the new kernel built and try it again. sincei dont remember which kernel i had booted last. [23:40] dmesg will show it, and compiling the kernel with -g would be great! [00:00] --- Mon Feb 16 2004