1113178615 N * Doener_zZz Doener 1113178620 M * Doener evening! 1113178625 M * Bertl hey Doener! 1113178632 M * Doener can't sleep :/ 1113178647 M * Bertl so what do we do now? 1113178665 M * Doener step a: make coffee, step b: yet to decide ;) 1113178682 M * Bertl sounds like a .. hmm, plan? 1113178863 M * Doener i guess i'll finally do the -L test with debugging enabled :) 1113178923 A * mikegrb dashes Doener's L 1113178924 M * Bertl when you're done with that, we can discuss what we need for a 2.0 release ;) 1113178930 A * mikegrb runs off laughing hystericaly 1113178942 M * Doener o.O 1113178968 M * Bertl see, he desperately needs some Ls ... 1113179256 M * Doener eek! 255 was too much for debug_limits 1113179339 M * Doener Bertl: got a sane value for me? 1113179429 M * Bertl no need to set any 1113179450 M * Bertl the kernel option will activate the exit check 1113179474 M * Doener ok... i thought i had to... i don't get any messages with -L 1113179486 M * Doener so i guess i did that part right :) 1113179500 M * Bertl excellent! 1113180208 Q * tchan Ping timeout: 480 seconds 1113180518 M * johnny excellent! 1113180529 M * johnny you are excellent Bertl :) 1113180539 M * Bertl am I? 1113180543 M * johnny sure am 1113180559 M * Bertl and why in particular, if I may ask? 1113180566 A * johnny was reading blog posting about free software peeps getting bugged and pestered to much.. and yet not many thanks 1113180576 M * johnny so.. i've been trying to fix that personally :) 1113180591 M * Bertl ah, well, thanks then! 1113180624 M * johnny i happened to run into greg kh yesterday.. and said the same to him 1113180628 M * johnny for the lovely udev and hotplug :) 1113180640 M * mikegrb Bertl: I baked you some cookies, made with real L 1113180644 M * johnny he happened to be on the monotone channel 1113180645 M * Bertl johnny: really? an no comment on the book? 1113180651 M * johnny book? 1113180658 M * Bertl ldd3 1113180664 M * johnny never read it 1113180669 M * johnny i don't write device drivers :) 1113180683 M * Bertl well, okay, more leading a monotone life ;) 1113180700 M * johnny i have enough on my plate with php coding 1113180711 M * johnny and general activism 1113180720 M * Bertl hey that's cool! 1113180737 M * johnny so i'm sure you've heard about this whole bk mess ... :) 1113180741 M * Bertl nobody will force you to do device drivers ... 1113180753 M * johnny i read greg kh's article in some linux mag 1113180755 M * Bertl johnny: well, yes, I'm glad that linus decided to part 1113180757 M * johnny in how to write a usb device driver 1113180759 M * johnny that was neat 1113180766 M * johnny have you checked out the alternatives? 1113180770 M * johnny discounting cvs of course 1113180789 M * Bertl they are very sad, but this is a good chance for them to improve ... 1113180799 M * johnny i wouldn't call them sad ... 1113180805 M * Bertl and I'm pretty sure they will ... 1113180818 M * johnny happen to be partial to a particular one? 1113180838 M * Bertl I know (because I used) cvs and svn ... 1113180853 M * Bertl heard and read a lot about arch, but nothing else 1113180933 M * Bertl but I'm pretty sure kernel deveopment will come up with a bunch of new, free 'cvs' hybrids ... 1113180977 M * johnny there are already enough! 1113180981 M * johnny we don't need more 1113180988 M * johnny wanna see something? 1113181001 M * johnny http://www.xaraya.com/~marcel/scm_compare.html/ 1113181003 M * Bertl if it's interesting, sure! 1113181172 M * Doener http://better-scm.berlios.de/comparison/ 1113181233 M * johnny that thing is too old Doener 1113181240 M * johnny or was.. 1113181312 M * Doener 2005 april 6th... looks pretty recent ;) 1113181343 M * johnny oh wow.. it got updated 1113181344 M * johnny wow 1113181348 A * johnny is amazed.. 1113181358 M * johnny i bet it is the kernel thing that made it happen 1113181371 M * johnny cuz it was out of date.. last year 1113181373 M * johnny by a year 1113181452 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1113181537 M * Bertl welcome tchan! 1113181619 M * Bertl johnny: so if I get that right, you would use monotone for kernel development? 1113181635 M * Bertl s/use/suggest/ 1113182342 M * johnny that's not mine 1113182352 M * johnny that's marcel.. the one who chose bitkeeper 1113182354 M * johnny for us 1113182361 M * johnny for xaraya 1113182363 M * johnny not kernel dev 1113182368 M * johnny in fact.. 1113182373 M * johnny we are the #3 project 1113182376 M * johnny on openlogging.org 1113182392 M * johnny http://openlogging.org/csets.html 1113182406 M * johnny right after the kernel and mysql.. 1113182460 M * johnny i'm leaning towards monotone ... cuz it has an awesome design 1113182463 M * johnny personally 1113182508 M * johnny kinda weird to work on hashes as verson #'s toh 1113182510 M * johnny tho* 1113182523 M * Bertl hmm ... 1113182621 M * johnny i'd say monotone needs a bit more speedup to work with kernel sources tho 1113182630 M * johnny they are workin on it tho 1113183593 M * Doener Bertl: ok, let's go (2.0 release talk...) 1113183721 M * albeiro Bertl: ever been wondering about roboust channel traffic logging ? mayby you should see http://loglibrary.com ? 1113183794 M * Doener brainstorm: ngnet -> postpone to 2.1, we've seen that folks can live without it for now and inclusion into a 2.1 dev branch can happen soon after 2.0... 1113183830 M * Bertl precisely, no ngnet for 2.0 1113183850 M * Doener cpu virtualization would be good for 2.0 i'd say... vc_wait_context() sync command... hmm... need to ask enrico why exactly we need that ;) 1113183863 M * Bertl also no varhz, and no bme for 2.0 IMHO 1113183904 M * Doener varhz: agreed, bme might be nice, but we could just have that as an addon patch, should be as good 1113183925 M * Bertl I'm inclined to put the split into 2.0 (still too many x86 around) 1113183935 M * Doener ah, btw, acct_update_integrals seems to have been removed in 2.6.12-rc2 IIRC 1113183950 M * Bertl sounds good ... 1113184001 M * Doener adding more token buckets is definately for 2.1... 1113184019 M * Bertl yep, the one and only is stable now (not optimal, but stable) 1113184065 M * Doener hm, what x86_64 about 32bit compatibility? guess we can do without for now, maybe have the tools complain about 32bit with a 64bit kernel? 1113184085 M * Doener s/x86_64 about/about x86_64/ 1113184120 M * Bertl no, I would like to add that, but with a trick ;) 1113184132 M * Doener ok, cast the runes ;) 1113184150 M * Bertl I did some thinking about this, and we have a single case which breaks the out of the box compatibility 1113184184 M * Bertl vcmd_ctx_iattr_v1 1113184191 M * Bertl with the char * (for the name) 1113184210 M * Bertl and I see two easy solutions for that ... where I prefer the second one 1113184235 M * Bertl first solution: just make a 32bit version of that and fix it up (like ioctls do) 1113184255 M * Bertl second solution: make it a char array ... all done, problem gone 1113184326 M * Doener ok, i have no idea what you're talking about :) (and no real idea what to care about when doing this 32bit<->64bit stuff...) 1113184351 M * Doener i.e. i don't even know what's the problem ;) 1113184362 M * Bertl it's simple, I really really really tried to avoid this from the beginning 1113184381 M * Bertl so I decided to use c99 types for _all_ structures in the API 1113184402 M * Bertl unfortunately I fogot about pointers :/ 1113184418 M * Bertl pointers are 32bit on x86 but 64bit on x86_64 ... 1113184451 M * Doener ah, and we get a 32bit pointer from userspace that is unusable with the 64bit syscalls, right? 1113184468 M * Bertl yes, especially is messes up the struct ... 1113184478 M * Bertl (because the 32bit version is smaller) 1113184831 M * Bertl so it would be the best IMHO to make a new API version vcmd_ctx_iattr_v2 which is immune to that ... 1113184848 M * Doener agreed 1113184851 M * Bertl next tool release could use just this interface and we are done 1113185183 M * Bertl okay, so what is missing? 1113185221 M * Bertl ah, a 'stable' tool release for 2.0, of course 1113185256 M * Bertl and probably the mask capabilities ... 1113185264 M * Doener hm, guess that could be hard to get... (the stable tools release) 1113185275 M * Bertl although I think most stuff can be fixed up in the stable branch 1113185309 M * Doener back to "vserver child reaper (initpid)" (the last discussion died somewhere), why do we need it, and what exactly does it do? 1113185350 M * Bertl the idea was, when init is being killed, to prevent zombies by reaping its children ... 1113185358 M * Doener i had the idea that it was all about those unremovable zombies we've seen from time to time, but in the last discussion it felt like something else 1113185371 M * Bertl i.e. reparenting the children to the real init once they are dead 1113185382 M * Bertl no, that's exactly it ... 1113185392 M * Doener hm, isn't that done atm? 1113185407 M * Bertl well, not sure ... didn't check recently ;) 1113185420 M * Bertl as the (re-)parenting was changed twice in the kernel 1113185424 M * Doener looking at the code, init's children won't find any suitable relatives and thus get reparented to child_reaper 1113185432 M * Bertl and signal handler changed too ... 1113185769 M * Doener if it wanted to test that, i'd start an init-based vserver, kill init and then check what happens to its childs, right? 1113185846 M * Bertl correct .. that should be sufficient ... 1113186037 M * Doener ok, inside the vserver, everything stays the same (i.e. PPID is still 1), what i expected... 1113186042 M * Doener uh? 1113186047 M * Doener doener ~ # vps -Af 1113186047 M * Doener UID PID CONTEXT PPID C STIME TTY TIME CMD 1113186047 M * Doener root 13600 100 naucki 1 0 04:17 pts/6 00:00:00 init [2] 1113186081 M * Doener that's interesting... the host's init got kind of 'overlayed' 1113186094 M * Bertl yeah, that's the virtualization 1113186106 M * Doener was a temporary effect though... after a cat /proc/1/status it was gone... 1113186164 M * Doener that line is what should contain the host's init (which didn't show up at all in the output), but the vserver's init was shown twice 1113186209 M * Bertl can happen, ps just 'lists' the pids, no guarantee that they will be unique ... 1113186243 M * Doener but why didn't it listen pid 1? 1113186267 M * Bertl don't ask me ... 1113186295 M * Bertl it's like: hey why did I get the queen of spades? 1113186320 M * Bertl after shuffling a deck of cards and dealing you five ;) 1113186466 M * Doener doener ~ # chcontext --xid 1 cat /proc/1/status 1113186470 M * Doener s_context: 100 1113186527 M * Doener after accessing /proc/1 from ctx 0 it's back to normal 1113186549 M * Doener let's see if that's just a bug in my port... 1113186718 M * Doener no, happens with 2.6.11.5-1.9.5.5, too 1113186760 M * Bertl probably the inode/dentry isn't destroyed ... 1113186768 M * Bertl (well or lazy destroyed) 1113186817 M * Doener yep, guess so... the output is updated only in xid!=1... 1113187023 M * Bertl okay, we can live with that ... 1113187090 M * Bertl and the children are reaped properly? 1113187308 M * Doener after vkill-ing the init process there's no problem with killing the other process, they disappear fine 1113187463 M * Doener they also seem to get reparented correctly... can't tell for sure, as i don't trust xid 1 proc output any longer ;) 1113187465 M * Doener root 503 20 500 0 04:42 pts/0 00:00:00 bash 1113187465 M * Doener root 504 20 503 0 04:42 pts/0 00:00:00 bash 1113187474 M * Doener and after vkilling the parent bash: 1113187481 M * Doener root 504 20 1 0 04:42 pts/0 00:00:00 bash 1113187499 M * Bertl okay, so we consider that working/done 1113187584 M * Bertl btw, doing some code review ... I stumbled over a 'possible' issue with the reboot vshelper 1113187741 M * Doener damn, firefox is so damn broken when it comes to handling multiple windows... 1113187775 A * Bertl still uses galeon 1.2.x :/ 1113187798 M * Bertl (but there's nothing broken with multiple windows ;) 1113187808 M * Doener more often than not, it doesn't handle input in the right window (f.e. window A is focussed and when hitting ctrl+t i get a tab in window B) 1113187848 M * mugwump yes, you need to de-focus then re-focus it. I've noticed that too 1113187851 M * mugwump with Galeon too 1113187851 M * Doener and just a moment ago it even opened a tab in window B when i middle-clicked on a bookmark in window A 1113187883 M * mugwump I find I can't type into the location bar, etc, without re-focusing sometimes 1113187894 M * Bertl cool! 1113187906 M * Bertl morning mugwump, btw! 1113187920 M * mugwump hi there! It's not morning here, but it is to my body clock :) 1113187936 M * mugwump just come back from YAPC::Taipei 1113187940 M * Bertl that's what's important ... 1113187942 M * Doener mugwump: alt+tab focus switching helps most of the time in fvwm here, mouse focus doesn't help at all, and in gnome often the best i could do was restarting firefox 1113187971 M * Bertl mugwump: do you have a few minutes to join our 2.0 discussion? 1113187977 M * mugwump Doener: that sounds ... subtly different to what I'm experiencing, yet I'm fvwm too :) 1113187993 M * mugwump Oh, sure... I'll just catch up on a little backlog 1113188057 M * Doener btw, did you experience firefox displaying pages in separate windows sometimes? if you close them, firefox crashes, you got to reload the page till the window disappears and is display in the browser window again... 1113188114 M * Doener anyway, let's get back to topic... 1113188222 M * Bertl yes .. stuff required for 2.0 1113188237 M * Bertl (for those who just tuned in ;) 1113188297 M * Bertl yes, I would like to make the reboot (vshelper) call synchronous ... 1113188319 M * Doener what's the issue you've discovered? 1113188324 M * Bertl yes 1113188333 M * Bertl vs_reboot() 1113188344 M * Bertl I investigated because of the big stack usage 1113188359 M * Bertl and discovered that we allocate the arguments on the stack 1113188383 M * Bertl but then pass them on to call_usermodehelper() 1113188407 M * Bertl which hands them over to an asynchronous thread as far as I can tell 1113188434 M * Bertl no idea if and when the arguments are copied 1113188438 M * Doener yep, wait is set to 0 1113188463 M * Bertl might have interesting implications regarding reboot behaviour if we set it to one, tough 1113188467 M * Bertl *though 1113188529 M * Doener yes, we're in the context of a process inside the vserver... 1113188576 M * mugwump Has anyone worked on a good vs1.2.x -> vs1.9.x migration guide? There are a lot of combinations of legacy vs new versions of things, I don't know how well those have been tested 1113188614 M * Bertl migration seems not to be any issue ... 1113188620 M * mugwump in particular, when I first tried a new kernel with (new|old) tools and old config, that didn't work 1113188637 M * Bertl did you try debian tools? ;) 1113188673 M * Bertl all joking aside, the alpha tools can handle the legacy config quite fine ... 1113188688 M * Bertl there is also a 'conversion' script (not official though) 1113188723 M * mugwump The flower page is very synoptic, but there aren't very clear examples (or weren't when I looked) 1113188745 M * Doener mugwump: is fakeinit in use? that might fail... didn't check for some time if we fixed that... 1113188747 M * Bertl that's correct, I guess we have to make some example config dirs to browse 1113188767 M * mugwump exactly, it's easy enough to grok instantly when you see an example. 1113188786 M * Bertl yeah, guess we can do that easily ... 1113188833 M * Bertl I'm more worried about the tools and missing stuff (i.e. stuff I added to the Todo list which Enrico tends to ignore ;) 1113188840 M * Doener (fakeinit with legacy config and new tools that is IIRC) 1113188841 M * mugwump I can't remember the exact setup, Doener, I didn't try to set up fakeinit though 1113188869 M * Doener does Enrico know about that Todo list? ;) 1113188910 M * Bertl well, I mention it in every second personal email, so unless we assume that he misses every second one ... then yes 1113188919 M * Doener ok :) 1113188954 M * Bertl but it can't hurt to post it to the ml .. just for the record (tried to avoid that until now) 1113189073 M * Bertl well, I guess we have to fix one or the other thing in the tools ourselves and wait for enrico to continue on the larger parts ... 1113189110 A * Doener hardly tried to avoid that, as he doesn't understand all that bash stuff too well... 1113189131 M * Doener s/hardly/really/ 1113189132 M * mugwump You can make binaries as small as 188k with Haskell, that run as fast as C and are as strongly typed as ML FWIW 1113189142 M * mugwump (that's an a.out size) 1113189171 M * Doener ML? 1113189181 M * Bertl what about a full featured conversion tool in haskell then? 1113189203 M * Doener eeek! 1113189205 A * Doener hides 1113189277 M * Bertl okay, back to the topic, 2.0 release 1113189319 M * Bertl the reboot help should be easy to test, maybe rs will help us there ... 1113189328 M * Bertl *helper 1113189341 M * mugwump sorry, I lied about the size. 1113189350 M * mugwump 28k seems to be the smallest 1113189378 M * mugwump All I'm saying is that you don't need to write tools in C to make them suitable for putting in /sbin 1113189387 M * mugwump but that is perhaps offtopic for an impending stable release 1113189389 M * Bertl hmm, I had an hello world with 50 bytes IIRC ;) 1113189410 M * mugwump sure, I could do it in about 15 with a .COM ;-) 1113189424 M * mugwump was that an a.out? 1113189433 M * Bertl yep 1113189439 M * mugwump cool 1113189476 M * Bertl well, it was always padded to 4k, but hey ... 1113189556 M * Bertl okay, mugwump, do you see anything terribly missing for a 2.0? 1113189578 M * Bertl something we just have to get done before we release it? 1113189600 M * mugwump I've been happily on a 1.9.3 for ages... and yeah, my only real moan is about the examples 1113189675 M * Bertl okay, I understand that this is important, we are working on a new wiki design/layout too, and we will focus on documentation for a little while 1113189694 M * Bertl (well, at least I will ;) 1113189861 M * Bertl sidenote: hmm, seems I managed to hack/compile recent binutils for the following archs: alpha arm cris hppa hppa64 i386 ia64 m68k mips mips64 ppc ppc64 s390 s390x sh sh3 sh4 sh64 sparc sparc64 v850 x86_64 1113189962 M * Doener phew.. quite a list... 1113190063 M * Bertl yeah, now checking for gcc 3.3.5 ... 1113190076 M * mugwump you have access to a s390? 1113190093 M * Bertl well, everybody has (there is a good simulator ;) 1113190116 M * Bertl mugwump: cross compiling (on x86) 1113190217 M * Bertl btw, hercules does s390 1113190297 M * mugwump is that a simulated ia64 or a real one? 1113190317 M * mugwump :) 1113191061 M * Bertl as I said, all cross compiling ... 1113191114 M * Doener hm, can't come up with anything i'd desperately need... 1113191135 M * Bertl and we have zero known stability issues ... 1113191155 M * Bertl (last one fixed was the hard scheduler issue) 1113191171 M * Bertl doesn't mean that there are none ... 1113191201 M * Bertl mugwump: btw, if your interested http://linuxvm.org/penguinvm/hercules/ 1113191252 M * micah Bertl: just finished a test compile of 2.4.27-9-vs1.2.10 + the quota patch, all good 1113191261 M * Bertl excellent! 1113191409 M * mugwump The only reason I asked, was because like the man on the moon the existence of IA-64 is something I hold true not based on physical evidence or even second hand rumours ;) 1113191440 M * Bertl hehe, well, nevertheless we can compile kernels against that arch ;) 1113191571 M * Doener so, as we have no problems (and no urgent ideas, either) atm, we should just sort out what we may have to drop from current development versions to make a 2.0 release... and then focus on tools, documentation and whatever we come up with to get a broader audience for Linux-VServer... 1113191620 M * Bertl that's the idea ... well, some code cleanup and FIXME removal should be done as well, but that's more consmetical 1113191630 M * Doener of course... 1113191728 M * Doener what about the legacy stuff? do we still want that to be a kernel option in 2.0? or do we require recent tools? 1113191769 M * Doener hmm... would force migration to the new configuration i guess, the legacy support is probably relying on the legacy api... 1113191793 M * Doener so we should probably keep that... 1113191814 M * Bertl I'd say the larger part of the legacy stuff can go by now 1113191833 M * Bertl just keep 0.30 compatibility 1113191852 M * Bertl and of course the network legacy *sigh* 1113193091 M * Doener i got to go now... university awaits me in a few hours and i got stuff to prepare... back in the evening (probably) 1113193105 M * Doener have a good whatever everyone! 1113193109 N * Doener Doener|gone 1113193254 M * Bertl okay, cya 1113194441 M * Bertl okay, folks, I'm off to bed ... have a nice whatever ... 1113194450 N * Bertl Bertl_zZ 1113197916 Q * DaPhreak Ping timeout: 480 seconds 1113199222 Q * Bertl_zZ Ping timeout: 480 seconds 1113200335 J * Bertl_zZ ~herbert@janus.mc.tuwien.ac.at 1113200652 Q * Bertl_zZ Read error: Operation timed out 1113201239 J * Bertl_zZ ~herbert@janus.mc.tuwien.ac.at 1113201673 Q * Bertl_zZ Read error: Operation timed out 1113202523 J * BWare ~bware@212.26.196.195 1113203396 Q * DaCa Ping timeout: 480 seconds 1113203515 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1113203949 Q * erwan_ho Remote host closed the connection 1113205770 J * prae ~prae@134.106-14-84.ripe.coltfrance.com 1113206756 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1113207067 N * DaPhreak DaPhreak|work 1113207772 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1113207780 M * IceTi good morning 1113207804 M * IceTi ONE Question: Where can I get some useful vServer grahics ???, 1113208470 J * erwan_taf ~erwan@81.80.43.68 1113208500 M * IceTi Where can I get some useful vServer grahics ???, 1113208520 M * rs hi 1113209721 Q * erwan_taf Ping timeout: 480 seconds 1113210129 J * DaCa ~danny@mail.limehouse.org 1113210929 J * alexx ~alexx@82.225.136.176 1113210953 M * alexx hello all 1113211176 J * erwan_taf ~erwan@81.80.43.77 1113211199 M * cereal hi folks 1113211218 M * cereal anyone who wants coffee? 1113211279 J * jsambrook ~jsambrook@aelfric.plus.com 1113211283 P * jsambrook 1113211372 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net :: 1113211477 M * erwan_taf cereal: I prefer cereals :) 1113211673 A * cereal give a cup of coffee to erwan_taf 1113211843 M * prae erwan_taf: my lover !! 1113211844 M * prae erwan_taf: :) 1113212169 M * SiD3WiNDR cereal port? :) 1113212240 P * virtuoso 1113212257 J * virtuoso ~s0t0na@80.253.205.251 1113212318 M * cereal port? 1113212470 M * SiD3WiNDR yea, next to the parallel port 1113214094 Q * Pazzo Quit: .. 1113214431 M * erwan_taf prae: oh darling ! 1113214438 M * prae :) 1113221630 Q * berni Ping timeout: 480 seconds 1113222110 J * berni ~berni@svr01.mucip.net 1113222989 N * Doener|gone Doener 1113223005 M * Doener cereal: coffee! 1113226933 J * wurd ~kvlt@modemcable143.235-201-24.mc.videotron.ca 1113227722 M * wurd what is this and do i need it? : [devel util-vserver >= 0.30.x] 1113229562 J * ydupont ~dupont-y@tomintoul.cri.univ-nantes.fr 1113229573 J * Bertl ~herbert@janus.mc.tuwien.ac.at 1113229724 M * ydupont bertl: small question about 2.6 WITHOUT legacy 1113229737 M * ydupont bertl: & hello, BTW 1113229821 M * ydupont Is a compat compiled alpha utils can work without problems in a 2.6 kernel with latest vservers compiled WITHOUT legacy support 1113229851 M * ydupont HighlandPark:~# vserver-stat 1113229851 M * ydupont vc_create_context(): Invalid argument 1113229851 M * ydupont HighlandPark:~# vps 1113229851 M * ydupont vc_create_context(): Invalid argument 1113229858 M * Bertl hmm, sec just trying to get my irssi right ... 1113229867 M * ydupont BUT vservers ARE working ok... 1113229868 M * ydupont ok 1113229872 M * Bertl (somehow lost connection this morning, completely) 1113230156 M * Bertl okay, ready now ... 1113230185 M * Bertl you say latestkernel, means 1.9.5.12 or so? 1113230291 M * ydupont yes 1113230322 M * ydupont I don't want to turn this in a flamewar and I hope this is NOT a debian specific thing (once again) 1113230344 M * ydupont HighlandPark:/vservers# vserver-info 1113230344 M * ydupont Versions: 1113230344 M * ydupont Kernel: 2.6.11.7-vs1.9.5.12TTEC-SAN- 1113230344 M * ydupont VS-API: 0x00010025 1113230344 M * ydupont util-vserver: 0.30.204; Mar 30 2005, 07:45:12 1113230346 M * ydupont Features: 1113230347 M * ydupont CC: gcc, gcc (GCC) 3.3.5 (Debian 1:3.3.5-5) 1113230349 M * ydupont CXX: g++, g++ (GCC) 3.3.5 (Debian 1:3.3.5-5) 1113230349 M * Bertl okay, we 'change' the interfaces there a little, so some tools break 1113230351 M * ydupont CPPFLAGS: '' 1113230353 M * ydupont CFLAGS: '-Wall -g -O2 -std=c99 -Wall -pedantic -W' 1113230355 M * ydupont CXXFLAGS: '-g -O2 -ansi -Wall -pedantic -W -fmessage-length=0' 1113230359 M * ydupont build/host: i386-pc-linux-gnu/i386-pc-linux-gnu 1113230361 M * ydupont Use dietlibc: yes 1113230363 M * ydupont Build C++ programs: yes 1113230365 M * ydupont Build C99 programs: yes 1113230367 M * ydupont Available APIs: compat,v11,v13,fscompat,net,oldproc,olduts 1113230369 M * ydupont ext2fs Source: e2fsprogs 1113230371 M * ydupont syscall(2) invocation: fast 1113230373 M * ydupont vserver(2) syscall#: 273/glibc 1113230375 M * ydupont Paths: 1113230377 M * ydupont prefix: /usr 1113230379 M * ydupont sysconf-Directory: /etc 1113230381 M * ydupont cfg-Directory: /etc/vservers 1113230383 M * ydupont initrd-Directory: $(sysconfdir)/init.d 1113230385 M * ydupont pkgstate-Directory: /var/run/vservers 1113230391 M * ydupont Kernelheaders: /lib/modules/2.6.10-1-686-smp/build/include 1113230393 M * ydupont vserver-Rootdir: /var/lib/vservers 1113230395 M * ydupont ok 1113230397 M * ydupont so I just need a recompilation 1113230408 M * Bertl you need a patch to the tools actually ... 1113230413 M * Bertl (or a cvs version IIRC) 1113230416 M * ydupont ok not a big problem. 1113230430 M * ydupont so workiong with no legacy in the kernel is safe ? 1113230434 M * Bertl sec, I have a patched version somewhere 1113230461 M * Bertl yes, it is safe unless you have legacy vservers or tools ;) 1113230467 M * Bertl (this is not true for the legacy networking ;) 1113230484 M * Bertl http://vserver.13thfloor.at/Experimental/MDK/util-vserver-0.30.205-3mdk.src.rpm 1113230502 M * Bertl alienate/unpack it, will contain 2 patches 1113230509 M * ydupont I 'm migrating my legacy vservers so it's just a good trick to be sure I have totally migrated all my vservers ;-) 1113230520 M * prae Mister Dupont :) 1113230530 J * Pazzo ~thomas@host130-250.pool8172.interbusiness.it 1113230538 M * ydupont bertl: okay, thanks 1113230583 M * Bertl you're welcome! 1113230592 M * Bertl hey prae! Pazzo! 1113230602 M * prae Missa Bertl :) 1113230667 M * Pazzo hi Bertl! 1113230895 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.11 1113231843 Q * erwan_taf Remote host closed the connection 1113231873 J * erwan_taf ~erwan@81.80.43.77 1113232143 J * BWare ~bware@212.26.196.195 1113232153 Q * ydupont Quit: Leaving 1113232432 M * rs re 1113232456 M * rs morning bertl 1113232622 M * Bertl hey rs! 1113232648 M * Bertl yesterday we were talking about you ;) 1113232708 M * cereal hi Bertl 1113232722 M * Bertl greetings cereal! 1113232723 M * cereal and what´s going on with Doener´s -L test? 1113232740 M * Bertl quite fine, so everything was working there ;) 1113233179 N * kevinp|gone kevinp 1113233193 M * kevinp good morning all! 1113233241 M * kevinp Question on running this command - 'yum --installroot=/vservers/test update' 1113233243 M * Bertl morning kevinp! 1113233253 M * rs Bertl: yeah I saw that :) 1113233261 M * kevinp should the vserver be stopped or does it matter? 1113233288 M * Bertl guess it doesn't matter as long as you can do the same 'yum' on a running system 1113233364 M * kevinp I'm just wondering because it seems like yum restarts some services depending on what it installs 1113233375 M * kevinp So how would it no to restart the right process? 1113233401 Q * monrad Quit: Leaving 1113233412 M * Bertl probably it would chroot there ... sorry don't know the yum details ... 1113233424 M * Bertl but of course it's definitely better to use vyum 1113233509 M * kevinp I saw the command at http://linux-vserver.org/FedoraHowTo 1113233533 M * kevinp at the end of the doc 1113233533 M * Bertl it should do all the proper things ... 1113233859 M * kevinp hmm, this can get tricky if you have different distributions ... 1113233897 M * Bertl of course, that's where all the fun is ... 1113233929 M * kevinp :) 1113233929 M * daniel_hozac that's why yum installed in all the vservers and vserver X exec yum is easier ;) 1113233959 M * kevinp yeah, I would agree, that would make more sense 1113234337 M * wurd i have kernelversion 2.6.11.6 and would like to install vserver, which file should i download from http://www.13thfloor.at/vserver/project/ ??? 1113234391 M * Bertl want to get soemthing more stable or help with testing latest releases? 1113234407 M * wurd me, Bertl ? 1113234411 M * wurd i want something stable 1113234465 M * Bertl just a second then ... 1113234495 M * wurd ok 1113234649 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.11.5-vs1.9.5.1.diff.bz2 1113234659 M * Bertl should apply fine to 2.6.11.6 too 1113234686 M * Bertl you might also add the following patches 1113234797 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-initvx-clean01.diff 1113234802 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-initvx-fix01.diff 1113234811 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-kconfig-feat01.diff 1113234812 M * Pazzo bertl: what about 2.6.11.7? ;-p 1113234819 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-namei-fix01.diff 1113234831 M * wurd so, i download these files, then i follow steps 3 to 10 of http://linux-vserver.org/Step-by-Step+Guide ? 1113234851 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-x25-fix02.diff 1113234868 M * Bertl wurd: let me take a look at that ... 1113234921 M * Bertl yup looks good ... 1113234925 M * wurd ok thanks 1113234944 M * Bertl Pazzo: should work for 2.6.11.7 too 1113234963 M * Bertl there will be a new release shortly ... 1113235173 M * kevinp I applied 1.9.5 to 2.6.11.7 last night and it worked fine 1113235316 M * Bertl yes, should work too, just there are some things which got fixed since 1.9.5 so it would be bad _not_ to include them ;) 1113235562 M * kevinp Where do you define what is installed when you create a vserver with apt-rpm? 1113235606 M * Bertl there is a list of packages (don't ask me where, just list your tool package) 1113235805 M * kevinp looks like /usr/lib/util-vserver/distributions/fc3/pkgs 1113236672 M * Pazzo Bertl, kevinp: thnx! 1113236788 Q * yarihm Quit: Leaving 1113236853 M * Pazzo Bertl: it's really time for upgrading some kernels now - uptime 12 mins, probably because of a freeze of 2.6.9-vs1.9.3 :-( 1113236910 M * Bertl yeah, time to upgrade ;) 1113236950 M * Pazzo no idea why people have nothing better to do than call me, just because some little vservers aren't running - they should be happy and take their free time (no stupid emails, no spam...) to go out of their office to have some fun ;-) 1113237183 M * Bertl people are weird ;) 1113237320 M * Pazzo bertl: 1.9.5 or 1.9.5.12 ? 1113237390 M * Bertl see my list above ;) 1113238566 J * kalou ~kalou@ASte-Genev-Bois-109-1-1-164.w80-13.abo.wanadoo.fr 1113238570 M * kalou Hi 1113238803 M * Bertl hey kalou! 1113238886 Q * Pazzo Ping timeout: 480 seconds 1113239057 M * kalou :) 1113239279 J * Pazzo ~thomas@host130-250.pool8172.interbusiness.it 1113239285 M * Pazzo bertl: thnx for the list! so you consider 1.9.5 + .1 + some-delta-patch to be more stable than 1.9.5.12, right? 1113239343 M * Pazzo (was back for some minute, nautilus ate all of my memory - but then it was automagically killed :o) 1113239350 M * Pazzo s/back/away/ 1113239550 M * wurd Bertl should i apply the patches in any particular order? 1113239571 M * Bertl okay, folks, give me a few minutes, I'll upload a patch ... 1113239598 M * Bertl Pazzo: to answer your question, 1.9.5.12 contains additional features (basically untested) 1113239938 Q * prae Quit: Client exiting 1113240234 J * mst ~matthewt@81.29.65.220 1113240241 M * Bertl welcome mst! 1113240314 M * Bertl wurd, Pazzo: okay we should have a patch in about 4 minutes 1113240316 M * wurd Bertl ? 1113240342 M * wurd a patch for what? 1113240356 M * wurd i was asking you in which order i should apply all the patches you told me 1113240380 M * wurd (for example, those two:) http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-initvx-clean01.diff 1113240380 M * wurd http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-initvx-fix01.diff 1113240393 M * wurd (and all the others) 1113240393 J * Art512 ~chatzilla@10.228.dial-up.xter.net 1113240400 M * Bertl < Bertl> okay, folks, give me a few minutes, I'll upload a patch ... 1113240416 M * Bertl (tht was my answer ;) 1113240438 M * Bertl welcome Art512! 1113240495 M * Art512 Bertl: Hi. Thanks 1113240594 M * wurd so the patch youre uploading will replace all the 5 patches you previously told me to apply ? (or what?) 1113240837 M * Bertl yes, it will be a _single_ patch against 2.6.11.7 1113240853 M * wurd oh, great, thanks a lot 1113240875 M * Bertl no problem, just testing it right now ;) 1113241170 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.11.7-vs1.9.5.x.diff.bz2 1113241175 M * Bertl here it is ... 1113241259 M * Pazzo bertl you are great! 1113241273 M * Bertl thanks! 1113241324 M * Pazzo patch applies cleanly 1113241351 M * Bertl good, means that I didn't patch against the wrong kernel this time ;) 1113241677 M * Pazzo BLK_DEV_VROOT? 1113241751 M * Pazzo cool - is it "stable"? 1113241770 M * Pazzo docs? 1113241788 M * DaCa evening folks 1113241801 M * wurd so.. i can apply this patch even if my kernel is 2.6.11.6 ? 1113241829 M * Pazzo wurd: think so - try it! 1113241834 M * Bertl probably, but why not use 2.6.11.7 (which is the latest stable) 1113241855 M * Bertl anyway ... have to leave now ... cya later ... 1113241859 N * Bertl Bertl_oO 1113241861 M * Pazzo cya bertl! 1113241866 M * wurd you mean get a patch for 2.6.11.7 from kernel.org ? 1113241880 M * wurd im VERY new to compiling and stuff 1113241900 M * Pazzo good luck ;-) 1113241918 M * wurd well, what should i do 1113241933 M * DaCa wurd: yes, you can get a patch there to upgrade from 2.6.11.6 to 2.6.11.7, and then apply this new patch from Bertl_oO 1113241933 M * wurd try the patch, or upgrade my kernel to 2.6.11.7 and then try the patch 1113241934 M * Pazzo if you have a look at kernel.org you can see that there is already 2.6.11.7 1113241985 M * Pazzo on kernel.org there is a "F" link (full source) and a 2.6.11.7 (patch only) 1113242008 M * wurd which one do iget 1113242016 M * wurd F or patch only 1113242049 M * kevinp Does anyone know what the difference is between what is listed in the pkgs folder versus the rpmlist.d? 1113242058 M * DaCa wurd: your choice, if you already have 2.6.11.6 its a smaller download for the patch 1113242062 M * kevinp in /usr/lib/util-vserver/distributions/ 1113242068 M * Pazzo DaCa: isn't patch-2.6.11.7 a patch against 2.6.11 and not against 2.6.11.6? 1113242102 M * DaCa Pazzo: hmm, I honnestly dont know, I am still in the 2.4 era :) 1113242125 M * Pazzo kevinp: sorry, no idea - it's a long time I haven't been playing with rpm... apt rulz ;-) 1113242209 M * kevinp When creating a new vserver it references one of the two (I think) to install packages. What does apt use? 1113242225 M * Pazzo DaCa, wurd: patch-2.6.11.7.bz2 is just 11k - if you download and open it you can see in the first lines of the diff that it is against 2.6.11 1113242233 M * Pazzo -EXTRAVERSION = 1113242235 M * Pazzo +EXTRAVERSION = .7 1113242272 M * Pazzo wurd: so if you have the full source for 2.6.11.6 you can either: 1113242288 M * Pazzo remove the .6 patch and add the .7 patch 1113242322 M * Pazzo or, easier (if your downlink is fast enough :-) download the full 2.6.11.7 one 1113242359 M * wurd where do i click for the full 2.6.11.7 ? 1113242377 M * Pazzo "F"# 1113242381 M * wurd ok 1113242387 M * wurd what will i do with this file, then ? 1113242427 M * Pazzo tar xfj linux-2.6.11.7.tar.bz2 1113242434 M * Pazzo cd linux-2.6.11.7 1113242451 M * Pazzo patch -p1 < ../patch-2.6.11.7-vs1.9.5.x.diff 1113242460 M * Pazzo (first bunzip patch-2.6.11.7-vs1.9.5.x.diff.bz2) 1113242474 M * Pazzo has anyone a kernel baking howto link lying around? 1113242481 M * Pazzo wurd: btw, what distro? 1113242493 M * wurd fedoracore3 1113242500 M * wurd will i have to rebuild the kernel ? 1113242512 J * jsambrook ~jsambrook@aelfric.plus.com 1113242530 M * Pazzo wurd: sure! what do you think are we talking about? 1113242536 M * wurd patching... 1113242582 M * wurd isnt there any way i can NOT-have to rebuild the kernel? its a very long process 1113242592 M * wurd (sorry for my bad english) 1113242609 M * Pazzo wurd: a) would you like to use linux-vserver? b) why? 1113242625 M * Art512 wurd: where are you from??? 1113242632 M * wurd a)yes b)for running multiple servers on one machine 1113242636 M * wurd Art512 canada 1113242653 M * Art512 ok 10x 1113242654 M * Pazzo wurd: if a): yes, you have to install a new kernel - you either get someone created by somebody else... 1113242662 M * Pazzo ...or you create your own kernel 1113242665 P * jsambrook 1113242676 M * wurd Pazzo i have rebuild the kernel before, and the only reason i did it is for vserver. i dont really want to go through that again 1113242685 M * mst in order to provide context support vserver *has* to go against the main kernel tree 1113242701 M * mst the capabilities it provides can't currently be done as a loadable module, I don't think 1113242766 M * wurd before i was on kernel 2.6.9, but i installed a new kernel, 2.6.11.6 1113242775 M * wurd especially for being able to install vserver 1113242796 M * wurd (i mean that i did this ONLY to be able to install vserver. not because i wanted to or whatever) 1113242854 M * wurd am i making any sense? 1113242880 M * Pazzo wurd: linux-vserver is a kernel patch and you cannot install patches without recompiling your kernel or getting a already-built patched kernel 1113242910 M * wurd so i did this for absolutely nothing? 1113242920 M * wurd (rebuilding the kernel from 2.6.9 to 2.6.11.6) 1113243036 M * Art512 wutd: U can use patched kernel after compiling ONLY, Why you do not want compiling kernel? 1113243036 M * kevinp the answer is pkgs not rpmlist.d - put all the rpms you want installed in 01 1113243091 M * wurd i dont really understand your question Art512... if you speak french you can ask me in french 1113243146 M * Pazzo wurd: you didn't do it for absolutely nothing - you will learn much more from errors that hurt a little bit ;-) 1113243192 M * wurd well Pazzo what do i need to do now (the most simple and faster solution) ? 1113243194 M * Art512 wurd: no I cannot. why you need Linux-Vserver??? 1113243239 M * wurd Art512 i need it for running many servers on one maching 1113243250 M * Pazzo wurd: ok, slowly: YES, you did it for absolutely nothing. what you need to do now is to get kernel sources, add vserver patch, compile, get vserver utils, install / create vserver images... 1113243270 M * Pazzo ...but I would really ask you again: are you sure you need linux-vserver? 1113243282 M * wurd yes i am sure 1113243286 M * Pazzo :-) 1113243289 M * wurd why do you ask ? 1113243334 M * Pazzo I ask you because people who need vservers normally run servers on it - and people running servers (should) know how to build kernels... 1113243390 M * Art512 wurd: http://www.digitalhermit.com/linux/Kernel-Build-HOWTO.html 1113243422 M * wurd thanks Art512 but i know this website already, its the one i followed for rebuilding my kernel 1113243474 M * wurd i'm quite discouraged to realize i have to go through this once more 1113243499 M * mst well ... if you'd read the docs correctly you'd have known in advance and only had to do a kernel compile once ... 1113243534 M * wurd thanks mst but i dont need this kind of "advice" right now, nor do i need to justify myself 1113243540 M * Pazzo wurd: you have a long, long way in front of you - don't let you discourage! 1113243564 M * mst wurd: fair enough, but stop asking the devs to justify the technical realities 1113243577 M * wurd am i asking them that ? 1113243615 M * mst yes, the way linux-vserver has to work it was bound to require a kernel patch + recompile 1113243629 M * wurd i'm not accusing anyone here... 1113243639 M * wurd i'm not asking for justifications... 1113243671 M * mst well, basically, untar the thing 1113243684 M * mst use the patch commandline mentioned above with the appropriate vserver patch 1113243692 M * mst then follow the Kernel Build HOWTO from there on in 1113243748 M * Pazzo wurd: http://linux-vserver.org/ProjectOverview 1113243751 M * Pazzo wurd: http://linux-vserver.org/FedoraCore3_HowTo 1113243753 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1113243825 M * Pazzo wurd: there are many many other useful docs you can find at linux-vserver.org! read them first and feel free to ask if you run into trouble doing so... 1113243871 M * wurd like they say in my language (and probably in yours too), too much is like not enough 1113243879 M * wurd but yeah, i'll do that 1113243889 M * Pazzo wurd: people may get angry if being asked the same questions over and over again by people not willing to read the available docs ;-) 1113244117 M * kevinp Who has access to let me write a new doc on the linux-vserver site? 1113244135 M * Pazzo wurd: btw, I'm using Debian only, but I had a short look at http://linux-vserver.org/FedoraCore3_HowTo - it's an easy to understand step-by-step howto for your distro - and for sure not "too much" 1113244158 M * kevinp I want to put one up that outlines how to get CentOS going w/ vservers 1113244180 M * Pazzo kevinp: everyone, just write! 1113244211 M * kevinp Pazzo, the only problem with that doc is it skips the rebuilding the kernel part - which wurd needs 1113244239 M * wurd if i can manage to install vserver only by following the fedoracore3 HowTo thats fine with me 1113244257 M * wurd i thought you were asking me to real all documentation about vserver 1113244321 M * kevinp Pazzo, let me show my ignorance on wiki - how do I create a new doc? 1113244380 M * Pazzo kevinp: just replace http://linux-vserver.org/whatever with http://linux-vserver.org/Your_Title :) 1113244442 M * kevinp cool 1113244451 M * Pazzo wurd: all docs would really be a little bit too heavy at the beginning - but's also not wrong to read some of them to get an idea what linux-vserver is actually doing to / with your kernel / server 1113244508 Q * berni Remote host closed the connection 1113244590 M * wurd Pazzo yeah sure, but i would definitely fall asleep if i had to read all docs BEFORE starting to install vserver 1113244630 M * wurd :) 1113244647 M * wurd now lets just hope this kernel rebuilding wont go wrong 1113244657 M * Pazzo wurd: there are also projects asking you to read the source to find out how they should be used :-) 1113244683 M * wurd heh 1113244699 M * Pazzo wurd: I hope you are using your current config as a base and don't start it from scratch! 1113244709 M * wurd whats 'scratch ? 1113244727 M * Pazzo from nothing - starting at zero 1113244731 M * wurd oh 1113244748 M * wurd well yeah, good idea 1113244850 M * Pazzo wurd: ...and I hope you are using linux-vserver first for a long time at home to test / learn things and don't use it on some root-server to put a bunch of new servers in the wild - you should really get some experience first before doing so! 1113244938 M * wurd you mean, "dont install vserver on your 'real' server right away, test it before" ? 1113245129 M * Pazzo no, I mean I don't think it would be a good idea if you where administrating a real-world hosting-server - with or without vserver 1113245144 M * wurd lol 1113245190 M * Pazzo don't laugh - take me serious! 1113245197 M * wurd well, why do you say so? 1113245208 M * wurd cause i cant rebuild a kernel? 1113245231 M * wurd cause i dont enjoy reading docs too much? 1113245245 M * Art512 :) 1113245258 A * Doener yawns 1113245264 M * Doener evening 1113245282 M * DaCa hi Doener 1113245311 J * berni ~berni@svr01.mucip.net 1113245311 M * Pazzo hurd: maybe ;-) 1113245314 M * Pazzo hi Doener! 1113245401 M * wurd a 1113245492 M * wurd damn this text-only irc client, i lost all your above explainations due to display bugs 1113245542 M * wurd oh well. i'll just follow the HowTo and we'll see 1113245568 M * DaCa wurd: there is a link to live irclogs at the wiki 1113245838 N * Bertl_oO Bertl 1113245854 M * Bertl back now ... 1113245944 M * wurd do i absolutely need a seperate partition for vservers? 1113246006 M * Pazzo wb bertl! 1113246013 M * Pazzo wurd: absolutely no 1113246050 M * Art512 wurd: It depend on for a what do you need Linux_VServer!!!! 1113246061 M * Bertl wurd: no, unless you want to use xid tagging ... 1113246085 M * wurd is there some tweaking to do if i _dont_ want to use a separate partition for vservers? 1113246162 M * wurd Art512 for now its just a test 1113246181 M * wurd i just want to see first if i can make it work 1113246191 M * wurd then i might install it for real and have real uses for it 1113246243 M * Art512 If you do not want separate partition, u can crate vserver's root directory (/vservers) at your root partition, 1113246265 M * wurd ok good 1113246329 M * Art512 but for more effective testing pocedure, use separate partition:) 1113246353 M * wurd why is it more effective that way % 1113246355 M * wurd ? 1113246416 M * Art512 i thing you must testing storage limits too!!! 1113246416 J * yarihm ~yarihm@80-218-3-32.dclient.hispeed.ch 1113246423 M * Art512 I'm true? 1113246460 M * wurd well cant i test it even if it's on the same partition? 1113246499 M * Bertl thing is, you could but you really do not want to ;) 1113246531 M * Bertl the tagxid stores context information with the files (either in the uid/gid or on disk) 1113246561 M * Bertl if you want to use that you have to mount the partition with tagxid 1113246578 Q * berni Remote host closed the connection 1113246586 M * Bertl doing so for the root partition is not easy, and with good reason 1113246643 M * Bertl some files are accesses while a vserver is started from halfway into the vserver, those would then have this tags on your root fs 1113246924 J * berni ~berni@svr01.mucip.net 1113246948 M * Bertl wb berni! 1113247056 Q * kalou Read error: Operation timed out 1113247096 M * wurd quote from the fedoracore3 howto: "itialize the repository database and upgrade of the system: apt-get update && apt-get upgrade" 1113247101 M * wurd wow, did i really need this? 1113247140 M * wurd its long :) 1113247163 M * Bertl well, depends on what you plan to do ... 1113247216 M * wurd does it take like an hour, or more like 15 minutes ? 1113247325 M * Art512 bye guys... 1113247325 Q * Art512 Quit: ChatZilla 0.8.23 [Mozilla rv:1.3.1/20030425] 1113247518 M * Bertl wurd: no idea, really depends on your connection and setup 1113247529 M * wurd yeah i guess 1113247562 M * wurd well can i skip this step ? 1113247633 M * Pazzo wurd: probably depends on how up2date your fc3 is - just try it. if it fails - update and try again 1113247641 M * Bertl wurd: sure, I don#t know what the implications are though ... 1113247666 M * wurd Pazzo , at what point will it fail ? (if it does fail) 1113247689 M * Pazzo wurd: I don't think it will fail... 1113247697 M * wurd ok 1113247706 M * Pazzo ...last time I was using fedora it was still called "RedHat 6.2" :-) 1113247744 J * matta ~matta@69.93.28.254 1113247751 M * Bertl evening matta! 1113247754 M * matta hello! 1113247759 M * matta good afternoon! 1113247781 M * Bertl heh, thought that just a second after I wrote it ;) 1113247782 M * Pazzo hi matta 1113247940 M * matta added vci config information 1113247942 M * matta what is this? 1113248066 M * Bertl if you run the testme.sh or check the /proc/virtual/info 1113248071 M * Bertl you see something like: 1113248074 M * Bertl VCI: 0001:0025 273 03110076 1113248095 M * matta oh 1113248112 M * matta how is stability looking these days? 1113248124 M * Bertl excellent ... 1113248132 J * Jdogg Jdogg@c-24-1-240-54.hsd1.tx.comcast.net 1113248135 M * matta 1.9.x can run for weeks on end with a few thousand processes and loads of 10+ ? :) 1113248135 M * wurd Pazzo , the patch that you updated earlier.. is it a kernel patch, or a vserver patch ? 1113248139 M * Bertl welcome Jdogg! 1113248146 M * Jdogg hey! 1113248148 M * Jdogg whats up 1113248151 M * Bertl matta: definitely! 1113248152 M * wurd s/updated/uploaded 1113248234 M * Bertl matta: 2.6.11.6-vs1.9.5+ up 13 days on test machine 4way 1113248235 M * Pazzo wurd: vserver patch is a kernel patch 1113248260 M * Bertl Jdogg: linux-vserver development of course ;) 1113248266 M * wurd yeah. but by "kernel patch" i meant, a patch that's meant for updating your kernelversion 1113248274 M * Jdogg si 1113248277 M * wurd which is not for vserver. 1113248313 A * Pazzo is getting a headache 1113248335 M * Pazzo wurd: what patch did I update?? 1113248342 M * wurd s/update/upload 1113248363 M * Pazzo wurd: just use 2.6.11.7 (full) and use the patch posted by Bertl before 1113248372 M * wurd you uploaded a patch earlier, in order to make my life simpler (not having to download like 6 patches) 1113248385 M * Bertl was probably me ;) 1113248387 M * Pazzo wurd: that wasn't me, that was bertl 1113248392 M * wurd oh, sorry, i mixed up bert and pazzo ;/ 1113248416 M * Bertl np, my patch is the vserver kernel patch, against a vanilla kernel 1113248423 M * wurd well, could you paste the url for the patch again ? 1113248435 M * Bertl sure ... http://vserver.13thfloor.at/Experimental/patch-2.6.11.7-vs1.9.5.x.diff.bz2 1113248436 M * Pazzo wurd: no problem - if this happens in this channel I'm proud of it ;-) 1113248451 M * wurd thanks 1113248457 M * Bertl you're welcome! 1113248482 M * wurd what do you mean by "my patch is the vserver kernel patch against a vanilla kernel" ? sorry, english isnt my primary language, i sometimes have difficulties to understand it well 1113248508 M * Pazzo vanilla kernel = kernels from kernel.org 1113248540 M * Pazzo not vanilla kernel = not virgin kernel, patched somehow 1113248571 M * Pazzo vserver kernel patch against a vanilla kernel = patch that applies to a certain kernel from kernel.org 1113248582 M * Bertl yeah, the 'kernel' slang callse the unmodified kernels from kernel.org 'vanilla' 1113248652 M * wurd ok, so your patch is made to be applied on a vanilla kernel; that's all i need to know? :) 1113248672 M * Pazzo yep 1113248676 M * wurd good :) 1113248696 M * Pazzo and you should also know the right kernel version the patch was made for 1113248700 M * Bertl yeah, means, get the mentioned version (or one close to that) in this case 2.6.11.7 and apply it (with patch) ontop of that 1113248701 M * Pazzo 2.6.7.11 in this case 1113248713 M * Pazzo *gg* 1113249010 M * wurd the patch doesnt work (most probably because i did something wrong) 1113249026 M * wurd i extracted the vanilla kernel 2.6.11.7 into /usr/src, then went inside this dir 1113249032 M * wurd ran this command : 1113249034 M * wurd patch -p0 < patch-2.6.??-vs1.9.5.diff 1113249048 M * Bertl try again with patch -p1 1113249053 M * wurd ok 1113249082 M * Pazzo F*#K... it's not my day today :-( I'll touch no more vserver host today... 1113249093 M * wurd yep its working, thanks Bertl 1113249104 M * wurd whats the difference between p0 and p1 ? 1113249180 Q * berni Remote host closed the connection 1113249185 M * Bertl it is the number of path components removed from the aptch 1113249191 M * Pazzo -p strips one "folder" level for every number => try "patch --help" or "man patch" 1113249197 M * Bertl (see man patch for details) 1113249401 M * kevinp Bertl: check out http://linux-vserver.org/CentOS_HowTo 1113249439 M * kevinp I'm working on a doc that is CentOS centric, how do I make better formatting? 1113249541 M * Bertl you should not bother too much about formatting ... it's the information which is important ... 1113249575 M * Bertl but there are several pages which might help you with a better formating (like the FAQ or problematic programs or the paper) 1113249575 M * Pazzo kevinp: just look at http://linux-vserver.org/Linux-VServer?action=edit for the most important things 1113249639 M * Bertl kevinp: and check out the preview feature ... 1113249781 M * wurd i did "make bzImage" but i am getting asked questions about the configuration of the kernel, even though i moved my old .config file in the new kernel's directory... 1113249799 M * Bertl try doing make oldconfig first 1113249811 M * Bertl it will also ask some questions (for new options) 1113249828 M * Bertl for vserver related stuff you can safely accept the defaults 1113249969 M * wurd ok thanks 1113249994 M * wurd youre right about the 'make oldconfig' step, i forgot about that 1113250039 M * Pazzo yeah, it sometimes happens that Bertl is right ;-) 1113250047 M * wurd ;p 1113250419 J * Nik ~Nik@cable-153-130.online.bg 1113250434 M * Bertl evening Nik! 1113250447 J * berni ~berni@svr01.mucip.net 1113250459 M * DaCa literal spelling ability 1113250459 M * DaCa oops, wrong window 1113250460 M * Nik hi all, Bertl 1113250646 M * Pazzo http://os.newsforge.com/article.pl?sid=05/04/11/118211&from=rss - Linus: 1113250649 M * Pazzo (Ironically, many users and distributions are likely to actually not mind slightly slower development for a while. One of the most common worries for users is just the fact that 2.6.x has continued to be developed at a very high rate thanks to just how smoothly it's been working, so I bet some people are both upset and gratified by this all. ;) 1113250750 Q * berni Remote host closed the connection 1113250769 J * berni ~berni@svr01.mucip.net 1113250831 Q * erwan_taf Read error: Operation timed out 1113250982 M * wurd i was told that on a 2.6 kernel i didnt need to do the "make modules" 1113250992 M * wurd but the website you gave me says to do so 1113251001 M * wurd i was told that only a "make" was required 1113251006 M * wurd is it true ? 1113251068 M * Bertl make modules is not required, they are built with 'make' 1113251077 M * Bertl you normally do not do make bzImage with 2.6 1113251084 M * wurd hmm... 1113251105 M * wurd i just did make bzImage .. so what should i do next (the website says to do make modules.. and then make modules_install) 1113251110 M * Bertl but if you do make bzImage you also need make modules ;) 1113251128 M * DaCa doesnt make also automagically do oldconfig when necessary? 1113251130 M * wurd ok 1113251154 M * Bertl DaCa: yes, make ensures that everything is done properly 1113251911 Q * Nik Ping timeout: 480 seconds 1113252292 Q * berni Remote host closed the connection 1113252389 J * berni ~berni@svr01.mucip.net 1113252536 J * Nik ~Nik@cable-153-130.online.bg 1113252840 J * ComplexHo ~ComplexHo@funk.gotadsl.co.uk 1113253021 Q * Nik Ping timeout: 480 seconds 1113253075 J * Nik ~Nik@cable-153-130.online.bg 1113253184 M * Pazzo is there in the meantime a secure way to execute commands inside a vserver - without fearing that the vserver owner knowing what I'm gonna go to do would be able to hack the whole host? 1113253240 M * Pazzo or, in other words: is "vserver exec" secure? 1113253368 M * Nik Pazzo: afaik these commands do not appear in the logs, but the processes are visible within the context/vserver 1113253414 M * Pazzo Nik: that's not a problem - but what if I'm doing a "vserver xy exec somecommand"... 1113253415 Q * ComplexHo Remote host closed the connection 1113253444 M * Pazzo ...and the vserver owner has replaced "somecommand" with some evil code 1113253458 M * Pazzo will this code be able to exit it's context? 1113253479 M * Pazzo like "exit" does after a "vserver enter" 1113253480 M * Nik Pazzo: if the command itself is e.g a daemon and wrutes a log..... forget it, it's not possible to switch context from within a context 1113253512 M * Nik s/wrutes/writes 1113253515 M * Pazzo from within a context not - but I'm executing the command from the host 1113253535 M * daniel_hozac but you're entering the context prior to running the command. 1113253546 M * Nik Pazzo: i assume noone but you has access to the host 1113253577 M * Nik Pazzo: and vserver exec changes into the context and executes the command transparent to the vserver 1113253593 M * Pazzo if I do "vserver xy enter" and then run "ls" they guy owning the vserver may replace it's ls with some binary doing exit and something evil to the HOST 1113253616 M * daniel_hozac really? that shoud not be possible. 1113253619 M * Nik Pazzo: think of it as starting a bew shell, e.g. bash within the context and then executing the commant 1113253640 M * Pazzo daniel_hozac: would this also be possible with "vserver xy exec"? 1113253663 M * daniel_hozac Pazzo: so you're saying that it's possible to break out of a vserver? 1113253664 M * Nik Pazzo: no way. you execute everything within the chrooted, secure-context-ed environement 1113253673 M * daniel_hozac Pazzo: that would be a serious issue. 1113253680 M * Pazzo daniel_hozac: if you are able to do so after a "vserver enter" why shouldn't a binary (executed by you) can do this? 1113253697 M * daniel_hozac Pazzo: but you're not. 1113253724 M * Pazzo I can remember that some time ago it has been discussed here, that "vserver xy enter" is evil and should only be done on trusted vservers 1113253745 M * Pazzo daniel_hozac: If I did a "vserver xy enter" I am 1113253775 M * daniel_hozac Pazzo: no, you are however capable of exiting the shell, thus returning to the parent process. 1113253827 M * Nik Pazzo: vserver xy enter does one thing: it starts a shell in the security context of the vserver. When you "exit" the vserver, you just destroy the shell.. Think of it as su to a user executed as root 1113253841 M * Pazzo hmmm... absolutely sure? 1113253896 M * Nik Pazzo: well.. Bertl is the one who might say 99.9%, but yes, I am positive such things cannot happen without a security bug 1113254046 M * Bertl hmm .. reading up now ... 1113254170 M * Bertl Pazzo: the interesting point is how would you execute code (your code) after doing exit in the parent process? 1113254187 M * Pazzo Bertl: I can remember that this was discussed time ago - but no idea, if it is still up2date... but I can remember the statement (no idea by whom), that "vserver enter" was dangerous on untrusted vservers 1113254189 M * Bertl if you can solve this, you probably can hack any system 1113254210 M * Pazzo if I rethink it you all you are right 1113254235 M * Pazzo no idea why I was thinking so or when I've read this... 1113254258 M * Pazzo ...but for sure I'll sleep better now and save some ssh daemons :o) 1113254299 M * Bertl good then ;) 1113254325 M * Pazzo sorry for this terrible sentence -> Pazzo: if I rethink it you all you are right 1113254343 M * Pazzo thnx berl, thnx Nik & daniel_hozac 1113254421 M * Pazzo now I can move a lot of administrative stuff away from vservers to let it be handled by the host :-p 1113254432 M * Pazzo thanks all, have a good night! 1113254481 M * Nik np :-) 1113254512 Q * Pazzo Quit: ... 1113255111 M * Doener Bertl: is schedprio available without CONFIG_VSERVER_HARDCPU? (working on a 2.6 step by step guide, currently writing explanations for some config options) 1113255226 Q * Nik Ping timeout: 480 seconds 1113255234 J * Nik ~Nik@cable-153-130.online.bg 1113255251 M * Bertl Doener: yep, it is 1113255260 M * Bertl and evening! 1113255270 M * Doener evening Bertl! 1113256610 Q * erwan_ho Quit: Leaving 1113257012 J * ComplexHo ~ComplexHo@funk.gotadsl.co.uk 1113257055 M * Bertl welcome ComplexHo! 1113257105 M * ComplexHo hi Bertl :) 1113257249 M * ComplexHo [hopefully] a quick question for you Bertl... Do I need to compile the kernel with specific support for namespaces to enable them 1113257288 M * Bertl no, namespace support is default 1113257299 M * Bertl (with 2.6/1.9.x that is) 1113257307 M * Doener there we go: http://linux-vserver.org/Step-by-Step+Guide+2.6 1113257318 M * ComplexHo Hmm strange - I am using 2.6.11-5/1.9.5 1113257327 M * ComplexHo ok will check that out now thanks :) 1113257332 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1113257338 M * Doener ComplexHo: that's not meant for you ;) 1113257350 M * Doener i've just written that as a step-by-step guide for new users ;) 1113257351 M * ComplexHo ok :) thanks for telling me :) 1113257356 Q * erwan_ho Remote host closed the connection 1113257362 M * ComplexHo I will check that out too then ;) 1113257364 M * Doener just wanted to tell folks about it ;) 1113257386 M * Doener no problem with that, it's just not telling about namespaces at all ;) 1113257464 M * ComplexHo ok.. my problem is with vyum - it doesn't work unless I configure 'nonamespace'.. It bombs out saying insufficient repository configuration - I have an strace if it helps... 1113257489 M * Doener http://vserver.13thfloor.at/Experimental/FOR-1.9.6/delta-namei-fix01.diff 1113257508 M * ComplexHo ahh bbiab ;-) 1113257508 M * Bertl Doener: suggestion: make another dir for the kernel compile, or compile the kernel in a separate dir ... 1113257544 M * Doener Bertl: good idea... i actually do that on my box... no idea why i didn't do it in the guide 1113257574 M * Doener Bertl: that patch was the one fixing the vyum issue, right? 1113257577 M * Bertl and maybe split the code sections into smaller parts, i.e. unpack kernel, apply patches, build kernel 1113257579 A * Doener is getting tired 1113257595 M * Doener ok 1113257629 M * Doener didn't care about layout yet, as i knew that it won't even get completed today ;) 1113257641 M * Bertl np 1113257654 M * Bertl and no idea what we fixed with the namei fix ;) 1113257669 M * Doener should be the .. issue 1113257681 M * Doener just not sure whether that applied to vyum 1113257797 M * Doener well, guess ComplexHo will tell us soon 1113257830 M * ComplexHo Doener: Nice tip about the hard-linked src tree ;-) Really like that... 1113257869 M * Doener you're welcome, but i just copied idea that from Bertl 1113257876 M * Doener s/idea that/that idea/ 1113257878 M * Doener ;) 1113257926 M * ComplexHo Another question - whilst I am recompiling the kernel, I am very tempted to try out the ngnet patch to get an lo device in the vservers - how experimental are these patches? 1113257956 M * Bertl highly experimental ... but they seem to work ;) 1113257977 M * Bertl not for the faint of heart though, ebcause you need to change a lot 1113257999 M * ComplexHo do they apply cleanly on 2.6.11-6 and 1.9.5? what's the best order - ngnet second? 1113258024 M * ComplexHo It might be worth the effort :) 1113258042 M * DaCa hardlinked src tree? 1113258095 M * Bertl ComplexHo: try them ontop of 1.9.5, if they do not apply cleanly, please let me know ... you get an update then ... 1113258115 M * Bertl DaCa: how do you think that I keep my 10k kernel trees? 1113258213 M * Doener i just wonder how you navigate between them ;) 1113258290 M * Bertl cd? ;) 1113258309 M * Doener :P 1113258311 M * Nik Bertl Enhanced Routing Through Linux ;-) 1113258327 M * Doener more like: how do you know which kernels you got and where they are 1113258341 M * DaCa Bertl: you mean you cp -al the tree before patching in order to safe space? 1113258359 M * Bertl Doener: are, much better question ... well a script based on slocate helps me there ... 1113258371 M * Bertl s/are/ah, a/ 1113258397 M * Bertl DaCa: yep, and because a diff between two such trees takes just a few seconds at most 1113258412 M * Doener hm, guess i have to extend my kernel dev script base then... only got some to patch/diff them 1113258458 M * Doener them = the kernels 1113258501 M * DaCa has diff some optimization so it doesnt compare the files if the inodes are equal maybe? 1113258641 M * Bertl well, diff doesn't compare hard linked files (they are per definition identical) 1113258681 M * Bertl Doener: looks good so far 1113258689 M * DaCa thats indeed a nice tip 1113258734 M * Bertl but you should be careful, it needs some discipline and a well configured editor (which makes backups) to work in such a hard linked tree (without causing disasters ;) 1113258870 M * Doener i'm off to bed... have a good whatever everyone! 1113258874 N * Doener Doener|zZz 1113258877 M * DaCa I guess you also need to use the immutable_unlink bit like in unification in order to avoid changing both sides when patching 1113258906 M * Bertl no, patch handles that quite fine 1113258926 M * DaCa without any special options? 1113258941 M * Doener|zZz yep, patch deletes the old file and replaces it with the patched one 1113258948 M * DaCa ic 1113259038 M * DaCa so you just have to avoid that your editor saves changes inplace 1113259061 M * Bertl and that yiou modify any files yourself 1113259206 Q * Nik Quit: Hybernating my brain 1113259231 M * Bertl btw I lied, just 2303 kernel trees here ;) 1113259265 M * DaCa I assumed the 10k was exaggerated :) 1113259316 M * Bertl yeah well, I cleaned up recently ... 1113259373 M * DaCa for vi(m) :set backup should do it I suppose? 1113259417 M * Bertl yep 1113259505 M * kevinp Ok, I cleaned up my doc: http://linux-vserver.org/CentOS_HowTo 1113259541 M * kevinp Can someone take a look and tell me if it's right or if it's a bunch of garbage? 1113259612 M * Bertl sure ... 1113259642 M * kevinp thanks 1113259897 M * Bertl " patch -p0 " not the best idea in general ... 1113259922 M * Bertl maybe you cold refer to Doener's excellent howto (in progress) there? 1113259930 M * kevinp hmm, that part I really just copied from the fedora howto 1113259942 M * kevinp Yeah, I have just been looking at that 1113259971 M * Bertl you might mention the gpg check or maybe Doener includes it as optional? 1113259983 M * kevinp Maybe we should just have distro specific hints pages or something that come off of what Doener wrote? 1113260015 M * Bertl well, yeah, I would suggest to link to such stuff, because it keeps things in sync 1113260038 M * Bertl if you have 10 pages to fixup once something changes it's not really optimal 1113260047 M * kevinp A lot of the docs will say things like ....your distro may vary.... at certain points, those points should be to a distro pecific index page or something 1113260068 M * kevinp s/pecific/specifc/ 1113260078 M * Bertl yup, agreed 1113260102 M * kevinp For example on the fedora howto it leaves off the rest of the instructions on compiling the kernel 1113260104 Q * yarihm Ping timeout: 480 seconds 1113260155 M * kevinp I'm going to modify mine, so that it is just the CentOS specific stuff 1113260167 M * kevinp to bad Doener already left 1113260176 M * kevinp s/to/too 1113260187 M * Bertl well, he'll get back tomorrow, I guess ... 1113260476 J * erwan_taf ~erwan@81.80.43.77 1113260652 M * kevinp Bertl, the only problem I have with that min-centos4 vserver I created is that when I stop it I get the error: vkill: vc_ctx_kill(): No such process 1113260773 M * Bertl should be fixed by recent kernel and tools 1113260788 M * Bertl (well, top recent actually ;) 1113260989 M * kevinp ok, I was planning on upgrading the kernel with you 7 patch tomorrow, I 1113260993 M * kevinp 'll try it out 1113261449 M * kevinp ok, I've modified it to be just CentOS Specific Hints now, should I rename the doc as well? 1113261469 M * Bertl no need for now ... 1113261501 M * kevinp Should I add a link to it anywhere? 1113261512 M * Bertl sure .. let me see where that would be ... 1113261565 M * Bertl http://linux-vserver.org/Documentation under howtos 1113261594 M * kevinp It amazes me how fast these vservers start now. I'm used to watching each process start individually 1113261625 M * Bertl compared to what? 1113261646 M * kevinp the old 1.2.x vservers 1113261663 M * Bertl really? such an improvement? 1113261690 M * kevinp yeah, I get a prompt right away with no output at all with the new tools 1113261735 M * Bertl well, it is now init based, no? 1113261815 M * kevinp http://deadbeefbabe.org/paste/253 1113261823 M * kevinp This is what I'm seeing 1113261883 M * Bertl yeah, I'd say it starts it in the background 1113261916 M * Bertl so it's probably not that much faster ... it just feels a lot faster ;) 1113261939 M * kevinp So is there a way to see that output? 1113261960 M * kevinp btw, just read the ml and I'm excited to here about the 2.0 branch! 1113262000 M * Bertl it seems to me that your 'old' server is not init based but sysv based ... 1113262033 M * Bertl probably changing to this would give you back the 'old' behaviour ... 1113262084 M * kevinp hmm, not really needed, just comfortable watching everything say OK :) 1113262115 M * Bertl yeah, it is also one process less running inside the vserver (the init) 1113262124 M * Bertl so if you don't need it ... 1113262186 M * kevinp I'm going to have to read up on it and see what the differences are. 1113263299 M * kevinp well, I'm off for now 1113263304 N * kevinp kevinp|gone