[00:26] shuri (~ipv6@ipv6.electronicbox.net) got netsplit. [00:26] maharaja (maharaja@ipax.tk) got netsplit. [00:26] Bertl (~herbert@MAIL.13thfloor.at) got netsplit. [00:27] shuri (~ipv6@ipv6.electronicbox.net) returned to #vserver. [00:27] maharaja (maharaja@ipax.tk) returned to #vserver. [00:27] Bertl (~herbert@MAIL.13thfloor.at) returned to #vserver. [00:27] Topic changed on #vserver by !lepton.oftc.net: http://linux-vserver.org/ || latest stable 1.21, devel 1.3.0 "Revolutions". [00:31] maharaja (maharaja@ipax.tk) got netsplit. [00:31] Bertl (~herbert@MAIL.13thfloor.at) got netsplit. [00:32] click (click@gonnamakeyou.com) left irc: Remote host closed the connection [00:32] Bertl (~herbert@MAIL.13thfloor.at) returned to #vserver. [00:32] maharaja (maharaja@ipax.tk) returned to #vserver. [00:36] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [00:45] monako (~monako@ts1-a120.Perm.dial.rol.ru) joined #vserver. [00:45] hi monako! [01:08] monako (~monako@ts1-a120.Perm.dial.rol.ru) left #vserver (всем пока ... good bye all ...). [01:28] S3b4St14N (~email@p5081BCAA.dip.t-dialin.net) left #vserver. [02:09] serving (~serving@213.186.191.101) left irc: Ping timeout: 499 seconds [02:10] pflanze (~chris@cc-linux4.ethz.ch) joined #vserver. [02:10] hi chris! [02:10] hi! [02:11] Going to upgrade to 1.21 now. [02:13] SHA1(www.13thfloor.at/patch-2.4.23-vs1.21.diff.gz)= 13f6d5413ba1bc40132c2017c492b5bdad19763e [02:13] hmm, should this tell me something? [02:14] maybe that you don't trust signed md5sums? [02:14] I've learned from our last session ;M) [02:14] s/M// [02:14] ah, signed md5 sums? [02:14] yes, on each release page ;) [02:15] oh, overlooked them [02:15] sorry:) [02:16] np 8) [02:28] netrose (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [02:51] g'night all [02:51] Nick change: Doener -> doener_zZ [03:09] Hmm, vs1.21 is not as easy to merge with grsecurity anymore as 1.00 [03:09] hmm .. where are the problems? [03:09] With 1.00, there were 2 files with rejects, now it's eleven. [03:09] besides somebody has already done the merge, IIRC ... [03:10] Yes but it's 1.20 [03:10] well, the delta 1.20 -> 1.21 is about 4 lines?! [03:10] ah? [03:10] k [03:10] (well I'd like to patch in skas and freeswan as well.. going to see) [03:10] so it propably will apply without even noticing ... [03:11] well it's strange, the advertised patch didn't apply without rejects. [03:11] if you already have some patches for 1.00, you can also merge the delta 1.00 -> 1.20 [03:12] maybe [03:12] which advertised patch doesn't work? [03:12] url, please, I'll verify ... [03:13] http://www.linux-vserver.org/index.php?page=grsecurityHowto [03:13] i did procedure 1 [03:13] using http://www.firehead.org/~jeffrey/linux-vserver/vs1.20-diff-for-grsecurity-1.9.23-2.4.23.patch [03:14] do you ahve an url at hand for the grsec patch? [03:15] sorry, I'm confused 8-) [03:15] moment [03:15] wget http://www.grsecurity.org/grsecurity-1.9.13-2.4.23.patch [03:15] wget http://www.firehead.org/~jeffrey/linux-vserver/vs1.20-diff-for-grsecurity-1.9.23-2.4.23.patch [03:16] patch -p0 < vs1.20-diff-for-grsecurity-1.9.23-2.4.23.patch [03:16] cd linux-2.4.23 [03:16] patch -p1 < ../grsecurity-1.9.13-2.4.23.patch [03:16] gave rejects for me. [03:16] nope, I udnerstand it this way: [03:17] linux + vserver + (grsec+ grsec-fix) [03:17] so you apply the http://www.firehead.org/~jeffrey/linux-vserver/vs1.20-diff-for-grsecurity-1.9.23-2.4.23.patch [03:17] hm yes that makes sense [03:17] to the grsec, and this then to the vserver patched linux kernel ... [03:19] does this make sense and/or work for you? ;) [03:33] morning [03:33] morn! [03:34] hi herbert :) [03:34] how are you? [03:34] fine thanks, and you? [03:57] hmm .. seems you are still sleepy ... [03:57] Action: MrBawb yawns [03:58] hmm .. want something exciting, dqn? [03:58] s/dqn/dan/ [04:03] serving (~serving@213.186.191.134) joined #vserver. [04:03] anybody interested in testing something new? [04:06] what is it? [04:06] 2.6 stuff ... [04:06] ah, i can test it in qemu, but i imagine you've already done that :) [04:06] i haven't got a live 2.6 system, or one that i'm ready to test it on [04:07] how realistic is your qemu setup? [04:07] just boot, or vserver with services etc? [04:09] just boot :( [04:09] well, that's worse than mine then ... [04:09] hehe [04:10] you're putting lots of new things into the 2.6 code aren't you? consolidating around existing kernel services and so on... [04:10] am I? [04:11] like the vfs root thing... [04:11] vfs namespace or something like that? [04:11] hmm, that is userspace ... but yes ... [04:11] Action: kestrel can't remember the exact terminology used [04:11] and virtual interfaces [04:11] clone namespace and pivot root ... [04:12] yeah, virtual network, which reminds me on gandalf ... [04:12] lots of big changes.. [04:12] hmm .. probably ... is this bad? [04:12] are you planning on integrating any stuff with that other cpu, disk, etc. quota project? [04:13] no, it sounds good to me...using existing kernel infrastructure seems like the way to go to me [04:13] you mean ckrm, right? [04:13] especially as it sounds like they give you a cleaner implementation [04:13] yeah, that's the one [04:14] well, I'm currently trying the impossible ... (getting a common codebase for 2.4 and 2.6 ;) [04:14] ah :) [04:14] a worthy goal [04:14] sounds more like the holy grail ... but at least I'm trying .... [04:16] yep, absolutely [04:16] did you try the perl script vstop? [04:17] I have to admit, not yet ... [04:17] herbert! [04:18] :) [04:18] I had a little peek at the code but I forwarded it to enrico ... [04:18] np, it's just something useful [04:18] well it's userspace, so it wasn't that interesting ... :) [04:19] pfft [04:19] hehe [04:19] perhaps i should post it to the mailing list, that might be better [04:22] I would have done that in the first place, but I tested it now ... what does this output mean? [04:22] VSERVER COUNT %CPU [04:22] ALL_PROCS 2 4.7 [04:22] DUMA 27 1.9 [04:23] it's the total CPU usage of each vserver [04:23] COUNT is the number of processes [04:23] okay, and ALL_PROCS is the host context, right? [04:23] so ALL_PROCS is using 4.7% of the CPU, DUMA is using 1.9% [04:23] it's context 1 [04:23] MAIN is context 0 [04:23] i ripped the code from vps, so it uses the same terminology as that [04:24] hmm, good that I have no server named MAIN ... [04:24] isn't that a little odd? :) [04:24] okay, how do you explain that this doesn't sum up? [04:24] i have MAIN on my system [04:25] what doesn't? [04:25] shouldn't the sum of all servers be the % of ALL_PROCS, or is this something different? [04:26] ALL_PROCS is only processes with a ctx of 1, it's not actually ALL_PROCS [04:26] that terminology confused me too, but it's the same as vps [04:26] Action: kestrel shrugs [04:27] hmm, okay I add this to the 'have a word with jack' list ... [04:27] hehe [04:28] okay, looks good, post it to the list ... maybe some aggregated info would be nice too ... [04:29] and maybe a sort option, by name and/or by uptime 8-) [04:30] can you obtain the uptime from ctx 1, or do you have to change context to get that? [04:30] hmm, currently this is done via the /var/run/vserver/.ctx file ;) [04:31] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [04:31] uptime would be nice [04:31] but in near future this and other information will be available via proc ... [04:31] okay [04:31] what is your name for ctx 1? [04:32] if you want to grasp a glimpse, check out the 1.3.0 release ... [04:32] ctx 1 is VX_WATCH or the observer context ... [04:33] do you think it's okay for a vserver userspace utility to require non-default perl modules? [04:34] eg. Term::ReadKey [04:35] hmm, do not ask me that, as I consider it bad for a userspace utility to be written in perl at all ... [04:35] Debian has already a libterm-readkey-perl package. [04:36] So it wouldn't much hassle. [04:36] pflanze (~chris@cc-linux4.ethz.ch) left irc: Quit: [x]chat [04:36] hehe [04:36] pflanze throws in his two cents, then off he goes [04:37] pflanze (~chris@cc-linux4.ethz.ch) joined #vserver. [04:37] wb [04:37] oops (was wrong window) [04:37] herbert: what is your objection to perl utilities? [04:37] that they require perl ... ;) [04:38] personally I dislike perl for two reasons [04:38] ah, that was the expected answer [04:38] 1) if I need 1 hour to code something in perl, and then after 2-3 month, I have a look at the code, it takes me almost 1 hour to understand it ... [04:39] 2) if you permit perl on your machine, then sooner or later you end up as a cpan archive ... [04:40] but this does neither mean it's bad nor unsuited for such tasks, it's just _my_ opinion ... [04:40] well, both of those objections are not absolutes. perl can be just as easy to understand as C if you stick to some guidelines,and always use strict and use warnings [04:40] I've made worse experiences with C code than perl. [04:41] but then that was prolly really bad C code :) [04:41] the cpan thing can be difficult to avoid though, unless you have infrastructure for building packages on a build server [04:43] php for example, seems much more readable than perl, with almost equivalent regexp/array stuff ... [04:43] I forgot, to me! [04:44] system scripts in php would be much worse. [04:45] why do you think that? [04:45] Action: kestrel agrees with pflanze [04:45] php is much more verbose [04:45] after all, perl has it's home as system scripting langugae. [04:46] that is just because it's much older ... [04:46] eg. perl: while (<>) { if /someregex/ { print "lalal\n"; } } [04:46] nothing worse than a badly suited tool on a system where admins have to work that don't know it. [04:46] Bertl: yes, it's also much more mature. still. [04:48] php: while ($str = fgets("stdin://")) { if (preg_match("/someregex/", $str) { print("lalal\n"); } } [04:48] And don't forget that perl6 is around the corner. Then php will look moot. [04:48] still, as you said herbert, everybody has their own enthusiasms [04:48] hehe [04:48] what about perl6 is going to invalidate php? i haven't read that much about it [04:48] yeah, I used to program in postscript 8-) [04:49] that is sick [04:49] hmm, depends ... once I wrote a flower, which changed on every printout ... that was really nice ... [04:50] perl6 is even more mature, for example optional type safety, a clean vm for executing code, .. [04:50] exceptions everywhere [04:50] cleaner syntax. [04:50] revamped regexes. [04:50] i like exceptions [04:50] perl5 exceptions are a bit of a nasty hack [04:50] revamped how so? [04:50] okay, let me make one! *G* [04:51] kestrel: larry said he felt guilty for making regexes what they are now, so he took on responsibility and cleaned them up. [04:51] :) [04:51] and if you read his apocalypse about it, there are quite a few good ideas. [04:51] fair enough...any idea what he changed? url? [04:52] "apocalypse perl larry" should turn something up on lang.perl.org [04:52] it's more integrated with the perl language itself, for one. [04:53] cleaner usable for writing parsers. (like, no more \s+ and such stuff needed) [04:53] i tell you one thing i absolutely hate about perl is the variable handling, i much prefer php's way [04:53] what do you mean? [04:53] having to use %, @ and $ is .... redundant in my opinion, when the language should be able to figure it out itself, as it does in php [04:53] okay, i will [04:54] perl6 does change that too, [04:54] @ will always be an array, like @array[5] instead of $array[5] [04:55] I find php strange that it does't differentiate between hash and array. arrays are hashes there, right? [04:55] correct [04:55] not exactly ... [04:55] ordered hashes maybe [04:56] arrays can be associative but they are not required to map 1:1 ... [04:56] you can have more than one key for one value and the other way around too ... [04:57] well, since perl objects are references, you can have that too, there. [04:57] at least on the value side [04:58] not sure what you mean about the "same key different values" thing. [04:58] what shall I says, perl (turing compatible), php (turing compatible), so you can do the same with both languages ... [04:59] Lisp is probably cooler than any of those languages (but maybe perl6 is catching up :). [05:00] what about scheme? [05:00] But then I've never written lisp yet, so can't prove it. [05:00] perl6 is not going to be backwards compatible is it? [05:00] it is - it determines if the input is perl5 and behaves accordingly. [05:01] eek, ugh ... that will hurt ... [05:01] ? [05:01] ooh, users can add their own grammar rules to perl6 [05:01] that's kind of cool [05:02] That's how they make it both perl5 and perl6 compliant. [05:02] well consider you are writing perl6 code and want to integrate perl5 parts, or you do the other way around ... [05:02] and the interpreter suddenly decides that your perl6/5 is perl5/6 [05:03] You can do it, it decides on a per-file basis. [05:03] as far as I remember. [05:03] i.e. if your modules start with "package ...", it's perl5. if they start with (forgot the new syntax), it's perl6. [05:04] jesus, the apocalypse on regex is 30 pages [05:05] :) - Damian Conway has some kind of abstracts of them. [05:10] larry wall is a god amongst men [05:15] he's a (very) good designer - years before the implementation is there, he describes it in details. (If I compare that with my poor abilities to describe my project to my buddy..) [05:36] hehe [06:23] netrose (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [07:52] pflanze (~chris@cc-linux4.ethz.ch) got netsplit. [07:52] serving (~serving@213.186.191.134) got netsplit. [07:52] Medivh (ck@62.93.217.199) got netsplit. [07:52] apw (~apw@212.104.150.41) got netsplit. [07:52] ensc (~ircensc@ultra.csn.tu-chemnitz.de) got netsplit. [07:52] Alex^ (alex@alex.city17.org) got netsplit. [07:52] Zoiah (Zoiah@matryoshka.zoiah.net) got netsplit. [07:52] aka (~aka@h062040166017.gun.cm.kabsi.at) got netsplit. [07:52] lp (~lpressl@interner.SerNet.DE) got netsplit. [07:52] riel (~riel@riel.netop.oftc.net) got netsplit. [07:52] fbc (~fbc@ppp-104-136.26-151.libero.it) got netsplit. [07:52] doener_zZ (~doener@pD9E12124.dip.t-dialin.net) got netsplit. [07:52] kestrel (~athomas@dialup51.optus.net.au) got netsplit. [07:52] Nesh (~dmistry@ool-4352413d.dyn.optonline.net) got netsplit. [07:52] MrBawb (abob@swordfish.drown.org) got netsplit. [07:52] virtuoso (~shisha@ip114-115.adsl.wplus.ru) got netsplit. [07:52] MedivhWrk (ck@netops.multimedia-centrum.de) got netsplit. [07:52] netrose (~john877@CC3-24.171.21.47.charter-stl.com) got netsplit. [07:52] sannes (~ace@sannes.org) got netsplit. [07:52] CosmicRay (~jgoerzen@glockenspiel.complete.org) got netsplit. [07:53] Topic changed on #vserver by ChanServ!services@services.oftc.net: http://linux-vserver.org/ || latest stable 1.21, devel 1.3.0 "Revolutions". [07:53] netrose (~john877@CC3-24.171.21.47.charter-stl.com) returned to #vserver. [07:53] Topic changed on #vserver by !lepton.oftc.net: http://linux-vserver.org/ || latest stable 1.21, devel 1.3.0 "Revolutions". [07:54] pflanze (~chris@cc-linux4.ethz.ch) returned to #vserver. [07:54] serving (~serving@213.186.191.134) returned to #vserver. [07:54] Alex^ (alex@alex.city17.org) returned to #vserver. [07:54] Zoiah (Zoiah@matryoshka.zoiah.net) returned to #vserver. [07:54] aka (~aka@h062040166017.gun.cm.kabsi.at) returned to #vserver. [07:54] ensc (~ircensc@ultra.csn.tu-chemnitz.de) returned to #vserver. [07:54] riel (~riel@riel.netop.oftc.net) joined #vserver. [07:54] lp (~lpressl@interner.SerNet.DE) returned to #vserver. [07:54] apw (~apw@212.104.150.41) returned to #vserver. [07:54] Medivh (ck@62.93.217.199) returned to #vserver. [07:54] Topic changed on #vserver by !lepton.oftc.net: http://linux-vserver.org/ || latest stable 1.21, devel 1.3.0 "Revolutions". [07:54] fbc (~fbc@ppp-104-136.26-151.libero.it) returned to #vserver. [07:54] doener_zZ (~doener@pD9E12124.dip.t-dialin.net) returned to #vserver. [07:54] kestrel (~athomas@dialup51.optus.net.au) returned to #vserver. [07:54] Nesh (~dmistry@ool-4352413d.dyn.optonline.net) returned to #vserver. [07:54] MrBawb (abob@swordfish.drown.org) returned to #vserver. [07:54] MedivhWrk (ck@netops.multimedia-centrum.de) returned to #vserver. [07:54] virtuoso (~shisha@ip114-115.adsl.wplus.ru) returned to #vserver. [07:55] sannes (~ace@sannes.org) returned to #vserver. [07:55] CosmicRay (~jgoerzen@glockenspiel.complete.org) returned to #vserver. [09:04] Nick change: Bertl -> Bertl_zZ [09:26] ensc (~ircensc@ultra.csn.tu-chemnitz.de) left irc: Ping timeout: 499 seconds [10:30] pflanze (~chris@cc-linux4.ethz.ch) left irc: Quit: [x]chat [12:18] pflanze (~chris@ethlife-a.ethz.ch) joined #vserver. [12:18] hello [13:30] does the 000 barrier also protect if the subdirectory is mounted? [13:31] like: chmod 000 /var/lib/vservers; mkdir /var/lib/vservers/serv1; mount -o lo image /var/lib/vservers/serv1 [13:39] pflanze: I think you have to chmod 000 it after it is mounted [13:40] hm, but you have to chmod the parent dir, not the mount point itself, right? [13:40] if I run the vserver with the mount point as root. [13:40] yes [13:40] sorry, I misread the question [13:41] It's just that i'm suspicious about how the 000 hack is done. [13:41] I was wondering the same yesterday [13:42] But the people here passed me some documents which start to explaing this thing [13:43] Public ones? [13:43] however, if you don't mount the vserver in separate partitions but keep their directories in one partition, chmod 000 it after it is mounted [13:43] yes I see [13:43] let me find some urls [13:43] (which should only be needed once, since it's the perms of the root dir of the fs on the partition) [13:44] yes, of course [13:44] the problem is that sometimes people chmod 000 the mount point and then mount the partition [13:46] take a look at this: http://list.linux-vserver.org/archive/vserver/msg00729.html [13:47] I had an hard time yesterday by chmodding 000 by mistake the folder of a vserver [13:48] heh:) [13:50] Doener` (~doener@p5082D8AB.dip.t-dialin.net) joined #vserver. [13:51] thanks for the link. [13:53] Hm, I thought that grsecurity *does* have secure chroots. [13:54] Btw, maybe grsecurity isn't far from vserver's functionality. [13:54] Instead of security context numbers, it probably uses the actual root dir for differentiation. [13:58] doener_zZ (~doener@pD9E12124.dip.t-dialin.net) left irc: Ping timeout: 485 seconds [14:05] hmm, [14:05] vserver uml3 enter [14:05] Can't set the ipv4 root (Bad address) [14:05] tap2 Link encap:Ethernet HWaddr 00:FF:E0:CB:8D:C7 [14:05] inet addr:192.168.6.1 Bcast:192.168.6.255 Mask:255.255.255.0 [14:06] So doesn't vserver work with tap devices? or what's the problem? [14:07] ah, need to specify tap2:..ip.. in the conf [14:10] hmm, still doesn't work [14:16] straange, I can't use *any* interface anymore, not even eth0 [14:18] Always getting Can't set the ipv4 root (Bad address) [14:19] Is vs1.21 broken with userspace tools 0.26? [14:19] Or did I break it by compiling freeswan into the kernel (not running now). [15:12] fbc (~fbc@ppp-104-136.26-151.libero.it) left irc: Quit: I'm hungry [16:14] click (click@gonnamakeyou.com) joined #vserver. [16:15] hm, what's the /usr/bin/shellmod file? [16:15] bin or script? [16:18] huh, I don't have that. Is that OT? [16:19] seems like newvserver requires it [16:19] ah, dunno. [16:19] newvserver on debian doesn't seem to require it. [16:20] i see that, i installed newvserver from the vserver 0.29 package... might be some odd rewrite in the script I guess [16:20] *fetching the new 'newvserver' script* [16:21] Action: pflanze has 0.26 [16:21] well, 1.20 requires 0.29 it seems [16:22] ah? [16:22] I'm having problems using networking right now, could this be because of too old vserver usertools? [16:23] Action: pflanze compiling a new clean 2.4.23+vs1.21 kernel right now to try this first [16:24] vserver on a nat'ed connection? [16:24] or with real external ip's ? [16:24] even on the public ip of the machine, not nat. [16:24] Can't set the ipv4 root (Bad address) [16:25] Action: pflanze rebooting said machine now [16:25] yeah thats the tools [16:25] no need to reboot [16:25] oh. [16:25] just compile the 0.29 tools [16:26] sigh. [16:26] it worked fine after installing the newest tools here [16:26] s/worked/worts [16:26] works! [16:26] dammit [16:27] :) [16:27] warts n words [16:27] i know [16:27] glad I don't have warts [16:28] :) [16:29] grr, somehow on this machine the kernel's 'make install' always installs into / instead of /boot [16:29] haha, replace /vmlinuz with a link to /boot/vmlinuz [16:29] same with /System.map and /config [16:29] that will force it into /boot [16:30] Install will still be to /, which looks like a mess. [16:30] nah [16:30] What do I have a boot partition for?.. [16:30] it will always look for /vmlinuz, and it links to /boot, thus placing it there [16:30] that's what happens here at least [16:30] :) [16:30] after some tidying up that is [16:31] ah. [16:31] it checks /etc/lilo.conf [16:31] edit that one [16:31] :) [16:32] everything in lilo.conf points to /boot/.. [16:32] even the kernel image? [16:33] well, do this: [16:33] ln -s /boot/vmlinuz /vmlinuz [16:33] ln -s /boot/config /config [16:33] ln -s /boot/System.map /System.map [16:33] and make sure /boot/* symlinks are just mapped to the kernel you want to boot [16:34] or modify lilo to boot the correct /boot/vmlinuz-* kernel and rereun lilo [16:37] The links didn't help. [16:37] Lilo is configured to /boot/vmlinz and running it / booting works [16:38] Maybe I'd have to 'make dep' or something again first. [16:38] Just doing make install with only the above vmlinuz, config, System.map links in / doesn't help. [16:38] (Without *any* files and links it's the same thing.) [16:39] on a and you're on a woody debian? [16:39] -on a [16:39] yes [16:39] *shrug* works fine here :/ [16:39] It's the only machine with this strange behaviour. [16:39] Must have remembered that somehow. [16:40] (It's my oldest machine still in use, from potato times) [16:45] Action: pflanze has simply hacked /sbin/installkernel now to always use /boot as target [16:48] have to go now, cu later [16:49] pflanze (~chris@ethlife-a.ethz.ch) left irc: Quit: [x]chat [16:54] kestrel (~athomas@dialup51.optus.net.au) left irc: Quit: ZZZzzzzz.... [17:04] ah, there we go. [17:05] smart... [17:05] real smart :/ [17:12] Nick change: Bertl_zZ -> Bertl [17:12] hi all! [17:12] hi click! [17:23] bertl :) [17:23] g'mornin [17:23] (i guess) [17:23] morn, who or what is real smart? [17:27] Morning, guys! [17:27] hi virtuoso! [17:28] Bertl: Can you say something good or bad about BitKeeper? [17:29] hmm, sure ... Larry is fun, the BitKeeper License sucks! [17:30] Yes, I already managed to read it. [17:30] But I failed to find the price for 'single user' license. [17:30] well, depends on how you use it ... [17:31] Hm. So people prefer CVS? [17:32] people? I prefer anything without limitations and restrains ... [17:33] If this thing is really good why not even try it? [17:34] because you would have to agree to the license first ;) [17:36] I mean that the functionality has the higher priority than license to my opinion. [17:36] Ok, I got your point. [17:36] Thanks. [17:37] well, Linus uses bk so you are in good company ... [17:38] Is he a paying customer or not? :) [17:39] he fulfills the requirements to legally not-have-to-pay ... [17:39] Ok. This is called non-paying customer there. [17:40] hmm, it's early in the morning ;) [17:49] has anybody done some 2.6 testig yet? [17:50] qh only for now. [17:50] hmm, you read my posting (vs0.02)? [17:51] No, not yet. [17:52] you should be able to actually start a vps on 2.6 (no network) [17:55] Action: virtuoso rides to 13th floor. :) [17:57] you might want to comment out #define DEBUG_VX in include/linux/vcontext.h (to hide debug information) [17:59] bertl: jsut poking around with the damned newvserver-scripts [17:59] bertl: had a laugh when I founf out it was replaced with a debian-ized version [17:59] *found [18:00] yeah funny stuff that newvserver thingy ... [18:01] IIRC I filed a bugreport on every version ... [18:02] :) [18:02] haha :D [18:03] well, the vserver-shit was annoying me [18:03] I couldn't remember why it worked, and I checked twice, and didn't see that it was two different scripts [18:03] *shrug* need to get my glasses :D [18:03] onefor the debianized ver, and one for the damned vserver-0.29-package [18:07] There is a known memory leak when creating tens of thousands of [18:07] threads on RHEL AS3 (NPTL-0.60). It's fixed in 1.4.2_04, which [18:07] should get released early next year. [18:07] hmm, maybe this is the right bug fixing concept ... ;) [18:08] An exploitable leak? :) [18:08] yeah, RH can 'exploit' you by selling you an update ... [18:08] it's a DoS for local users only [18:09] and there are enough local user DoSes anyway [18:09] so no need for an emergency errata for this bug [18:09] hi rik! how are you? [18:09] I'm fine [18:09] a bit of a cold, but fine otherwise [18:10] It's wandering around, these days ... [18:10] mine wandered all the way from new hampshire to the netherlands and back [18:10] Action: riel notes that travelling with a cold sucks [18:11] I read your explanations on lkml regarding the different meminfo values ... [18:12] is there some updated rmap for 2.4.23 sheduled? [18:12] I'll keep explaining, in the hope that Mike Fedyk writes some documentation [18:12] it's planned, but not scheduled [18:12] well actually, I need to do that to get an updated kernel for nl.linux.org ;) [18:12] so I guess it is kind of scheduled [18:12] hehe [18:13] what do you think of Con's kernels, ever had a look at them? [18:14] nope, only at some of his individual patches [18:15] I'm thinking about using his kernel patches for vserver on 2.4 ... [18:15] ahh, rik, do you have another minute? [18:16] sure [18:17] there was a thread on the vserver ml, started with Message-ID: <20031211172321.GF11536@jsambrook> [18:17] from Jonathan Sambrook, on Thu, 11 Dec 2003 [18:18] he reports that replacing the uts_sem (rwsem) by a spinlock (in the vserver code) solved some issues for him ... [18:19] http://jonathan.dsvr.co.uk/kernel/patches/02-fix-sem_with_spinlock.diff (this is the patch) [18:19] interesting [18:19] no I do not understand how this could change anything, do you have any idea? [18:19] s/no/now/ [18:20] does he already have a spinlock locked while trying to lock the semaphore ? [18:20] if so, he could end up scheduling with a spinlock held [18:21] hmm, good point ... have to check that ... but it is possible ... [18:22] thank you for this hint, didn't think of that ... (limited view) [19:08] maharaja (maharaja@ipax.tk) got netsplit. [19:08] Bertl (~herbert@MAIL.13thfloor.at) got netsplit. [19:10] Bertl (~herbert@MAIL.13thfloor.at) returned to #vserver. [19:10] maharaja (maharaja@ipax.tk) returned to #vserver. [19:14] sannes (~ace@sannes.org) got netsplit. [19:14] CosmicRay (~jgoerzen@glockenspiel.complete.org) got netsplit. [19:14] Medivh (ck@62.93.217.199) got netsplit. [19:14] apw (~apw@212.104.150.41) got netsplit. [19:14] Alex^ (alex@alex.city17.org) got netsplit. [19:14] Zoiah (Zoiah@matryoshka.zoiah.net) got netsplit. [19:14] aka (~aka@h062040166017.gun.cm.kabsi.at) got netsplit. [19:14] serving (~serving@213.186.191.134) got netsplit. [19:14] lp (~lpressl@interner.SerNet.DE) got netsplit. [19:14] riel (~riel@riel.netop.oftc.net) got netsplit. [19:14] netrose (~john877@CC3-24.171.21.47.charter-stl.com) got netsplit. [19:15] CosmicRay (~jgoerzen@glockenspiel.complete.org) returned to #vserver. [19:15] sannes (~ace@sannes.org) returned to #vserver. [19:15] Medivh (ck@62.93.217.199) returned to #vserver. [19:15] apw (~apw@212.104.150.41) returned to #vserver. [19:15] lp (~lpressl@interner.SerNet.DE) returned to #vserver. [19:15] riel (~riel@riel.netop.oftc.net) returned to #vserver. [19:15] aka (~aka@h062040166017.gun.cm.kabsi.at) returned to #vserver. [19:15] Zoiah (Zoiah@matryoshka.zoiah.net) returned to #vserver. [19:15] Alex^ (alex@alex.city17.org) returned to #vserver. [19:15] serving (~serving@213.186.191.134) returned to #vserver. [19:15] netrose (~john877@CC3-24.171.21.47.charter-stl.com) returned to #vserver. [19:19] shuri (~ipv6@ipv6.electronicbox.net) got netsplit. [19:19] shuri (~ipv6@ipv6.electronicbox.net) returned to #vserver. [19:19] Nick change: Bertl -> Bertl_oO [19:36] hmm, this was odd... [19:36] any of you guys get a block when trying to erase an old vserver that has been shut down etc? [19:36] aka rm -rf /vserver/vserver-dir ? [19:47] maharaja (maharaja@ipax.tk) got netsplit. [19:47] Bertl_oO (~herbert@MAIL.13thfloor.at) got netsplit. [19:48] maharaja (maharaja@ipax.tk) returned to #vserver. [19:49] Bertl_oO (~herbert@MAIL.13thfloor.at) returned to #vserver. [19:50] Nick change: Bertl_oO -> Bertl [19:51] hmm, seems like the connection was lost ... [19:52] @click what 'block'? please elaborate! [19:53] couldn't remove the vservers /proc entries [19:53] wrong tool, eh? [19:53] yup [19:54] and, forgot to down it [19:54] :9 [19:54] thus, no wonder [19:54] whats the proper tool for removing a vserver btw? [19:54] except downing it and rm -rf ? [19:54] :D [19:54] never figured out that one [19:54] hmm, for removing the proc entries, the proper tool is umount ;) [19:55] ah, inside the vserver, or in the rootserver? [19:55] wherever you are allowed to do so ... [19:55] <-- enlightened [19:55] :) [19:55] thanks mate :) [19:55] my pleasure ... [19:55] same goes for /dev/pts by the way ... [19:56] well, no prob really, I just erased it, and added the same server again, plus downing it and moving it to the correct location, with a symllink to the new vservers locateion from /var/lib/vservers/ to the correct location [19:57] guess this issues will go away in a few releases ... [19:57] got a friend hooked on the conecpt as well... [19:57] he started installing it after trying a testbase I made for him [19:59] hm, btw... isn't there an option on the debianized ver to say where the vserver should be located? --rootpath or smth...? [19:59] *checking newvserver* [19:59] ah -vsroot [19:59] could be, I do not know the debian changes ... [19:59] makes it easier to point it to the location where I want the vserver [20:00] instead of creating it and then mv'ing it [20:04] okay, another vserver race/issue fixed (by Jonathan) [20:04] 1.22 ? [20:04] :) [20:04] will be there soon ... [20:05] security or stability-related? [20:05] stbaility related .. [20:05] anything in particular? [20:05] http://jonathan.dsvr.co.uk/kernel/patches/02-fix-sem_with_spinlock.diff [20:05] explanation is in my posting on the ml ... [20:06] ok, reading up [20:09] there... small fix on the debianized newvserver-script, making a symlink from /var/lib/vservers/ to where the vserver is installed to when using --vsroot [20:18] sannes (~ace@sannes.org) got netsplit. [20:18] CosmicRay (~jgoerzen@glockenspiel.complete.org) got netsplit. [20:18] Bertl (~herbert@MAIL.13thfloor.at) got netsplit. [20:18] serving (~serving@213.186.191.134) got netsplit. [20:18] Alex^ (alex@alex.city17.org) got netsplit. [20:18] Zoiah (Zoiah@matryoshka.zoiah.net) got netsplit. [20:18] Medivh (ck@62.93.217.199) got netsplit. [20:18] lp (~lpressl@interner.SerNet.DE) got netsplit. [20:18] aka (~aka@h062040166017.gun.cm.kabsi.at) got netsplit. [20:18] riel (~riel@riel.netop.oftc.net) got netsplit. [20:18] apw (~apw@212.104.150.41) got netsplit. [20:18] netrose (~john877@CC3-24.171.21.47.charter-stl.com) got netsplit. [20:19] Bertl (~herbert@MAIL.13thfloor.at) returned to #vserver. [20:19] CosmicRay (~jgoerzen@glockenspiel.complete.org) returned to #vserver. [20:19] sannes (~ace@sannes.org) returned to #vserver. [20:19] Medivh (ck@62.93.217.199) returned to #vserver. [20:19] apw (~apw@212.104.150.41) returned to #vserver. [20:19] lp (~lpressl@interner.SerNet.DE) returned to #vserver. [20:19] riel (~riel@riel.netop.oftc.net) returned to #vserver. [20:19] aka (~aka@h062040166017.gun.cm.kabsi.at) returned to #vserver. [20:19] Zoiah (Zoiah@matryoshka.zoiah.net) returned to #vserver. [20:19] Alex^ (alex@alex.city17.org) returned to #vserver. [20:19] serving (~serving@213.186.191.134) returned to #vserver. [20:19] netrose (~john877@CC3-24.171.21.47.charter-stl.com) returned to #vserver. [20:24] maharaja (maharaja@ipax.tk) got netsplit. [20:24] click (click@gonnamakeyou.com) got netsplit. [20:24] Doener` (~doener@p5082D8AB.dip.t-dialin.net) got netsplit. [20:24] virtuoso (~shisha@ip114-115.adsl.wplus.ru) got netsplit. [20:24] MedivhWrk (ck@netops.multimedia-centrum.de) got netsplit. [20:24] MrBawb (abob@swordfish.drown.org) got netsplit. [20:24] Nesh (~dmistry@ool-4352413d.dyn.optonline.net) got netsplit. [20:24] maharaja (maharaja@ipax.tk) returned to #vserver. [20:24] click (click@gonnamakeyou.com) returned to #vserver. [20:24] Doener` (~doener@p5082D8AB.dip.t-dialin.net) returned to #vserver. [20:24] Nesh (~dmistry@ool-4352413d.dyn.optonline.net) returned to #vserver. [20:24] MrBawb (abob@swordfish.drown.org) returned to #vserver. [20:24] MedivhWrk (ck@netops.multimedia-centrum.de) returned to #vserver. [20:24] virtuoso (~shisha@ip114-115.adsl.wplus.ru) returned to #vserver. [21:52] Nick change: Bertl -> Bertl_zZ [22:43] Alex^ (alex@alex.city17.org) left irc: Ping timeout: 499 seconds [22:46] serving- (~serving@213.186.191.134) joined #vserver. [22:47] serving (~serving@213.186.191.134) left irc: Read error: No route to host [22:49] Alex (alex@alex.city17.org) joined #vserver. [23:26] netrose_ (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [23:26] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Read error: Connection reset by peer [23:56] [S]ushi (Sushi@pD9E34EC8.dip.t-dialin.net) joined #vserver. [23:57] <[S]ushi> hi! [00:00] --- Sun Dec 14 2003