1120089901 M * Bertl good, a lot of work, but good ... 1120090717 M * Hollow heh, i'm still learning C + vserver.. ;) 1120090881 M * Hollow Bertl: i read 2.0 will be out soon... any new features, or just bug fixes? and what about the sync problem? 1120090895 M * Bertl sync problem is userspace ... 1120090901 M * Hollow ok 1120090933 M * Bertl I'm not happy, but we will have to live with the current tools until new tools are created/released 1120090968 M * Hollow Bertl: maybe you want to look at http://home.xnull.de/work/libvserver/ ... 1120090977 M * Hollow just the beginning 1120090998 M * Hollow but if things go well, it could be a clean 2.0 api implementation of the basic tools 1120091033 M * Bertl excellent! maybe you should advertize it on the ML? 1120091051 M * Bertl others might want to join?! 1120091063 M * daniel_hozac hmm, what's wrong with util-vserver's libvserver? 1120091065 M * Hollow yeah, just give me a few days to complete some more functionality so it can do something usefull 1120091097 M * Hollow vcontext vnamespace and vsignal are currently implemented 1120091131 M * Bertl daniel_hozac: I guess the main issue is that nobody can read enricos code well enough to fix/change it? 1120091140 A * Hollow nods 1120091155 M * daniel_hozac ah, that. 1120091318 M * Hollow vcontext example: http://phpfi.com/67646 1120094767 Q * Aiken Quit: Leaving 1120095948 Q * mugwump Remote host closed the connection 1120096364 Q * Johnsie Quit: 1120096444 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1120096744 M * Bertl okay, I'm off to bed now .. night folks! 1120096750 N * Bertl Bertl_zZ 1120097305 Q * DaCa Ping timeout: 480 seconds 1120097835 J * DaCa ~danny@mail.limehouse.org 1120097848 Q * mugwump Remote host closed the connection 1120097850 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1120098050 M * mugwump ? 1120098054 M * mugwump strange 1120100667 J * eXplasm explasm@p549F7B95.dip.t-dialin.net 1120103588 Q * eXplasm Remote host closed the connection 1120110877 M * jkl anyone know how to make a vserver of type default on gentoo? 1120110896 M * jkl or define it as default rather.. 1120111159 M * maharaja jkl: i too fail to do that :) 1120111162 M * maharaja on debian 1120111378 M * maharaja ah, found my error 1120111403 M * maharaja # cat /etc/vservers/raoul/apps/init/mark 1120111403 M * maharaja default 1120111412 M * maharaja the mark file has been in the wrong dir :) 1120111538 M * maharaja /etc/vservers/raoul/apps/init/mark is the correct way 1120113589 J * alexx ~alexx@82.225.136.176 1120115350 J * sukria ~sukria@zenon.apartia.fr 1120115815 J * Doener` ~doener@p5487596D.dip.t-dialin.net 1120115830 Q * atsab Ping timeout: 480 seconds 1120116262 Q * Doener Ping timeout: 480 seconds 1120116376 J * atsab ~as@lotes.vtu.lt 1120116860 Q * atsab Ping timeout: 480 seconds 1120120053 J * erwan_taf ~erwan@81.80.43.67 1120122013 Q * sukria Remote host closed the connection 1120122261 T * services.oftc.net : http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc4, ng9.5 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1120122565 P * erwan_taf Leaving 1120122714 T * services.oftc.net : http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc4, ng9.5 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1120122920 J * _mcp ~hightower@wolk-project.de 1120122920 Q * mcp Read error: Connection reset by peer 1120122954 M * Vudumen Hollow: did you wrote libvserver? 1120122960 M * Vudumen did you write :) 1120123011 J * sukria ~sukria@zenon.apartia.fr 1120123397 T * services.oftc.net : http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc4, ng9.5 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1120123633 Q * _mcp Read error: Connection reset by peer 1120123633 J * __mcp ~hightower@wolk-project.de 1120124132 Q * __mcp Ping timeout: 480 seconds 1120124151 M * DaPhreak Vudumen: yeah he did ;) he is actually ... 1120124397 J * _mcp ~hightower@wolk-project.de 1120125385 M * Vudumen fine. i'm started debianizing it :) 1120125461 M * Vudumen ahh shit. i cannot speak english today :( 1120125600 J * rs_ ~rs@ATuileries-153-1-5-47.w82-123.abo.wanadoo.fr 1120126577 Q * sukria Quit: see you 1120127433 M * Doener` Nick /doener 1120127438 M * Doener` ehr 1120127444 N * Doener` Doener 1120127454 M * Doener morning! 1120127467 M * rt hi Doener 1120127550 M * Doener wow, .13-rc1 is out... that was a fast one... 1120127999 Q * rs_ Quit: rs_ 1120128013 M * Doener *sigh* and quite a number of rejects with the vserver patch... 1120128864 Q * stephenM Ping timeout: 480 seconds 1120128906 M * Doener micah: i can now confirm that screen works when you enter a vserver using my experimental daemon+client 1120130263 J * _are_ ~are@gateway-dsl.lihas.de 1120130265 M * _are_ hi 1120130274 M * Doener hey _are_ 1120130359 J * sukria ~sukria@213.223.184.194 1120130699 M * pusling how do I mount something from outside into a vserver without restarting the vserver ? 1120130863 M * Doener http://linux-vserver.org/Namespaces 1120131074 M * pusling thanx! 1120131081 M * Doener np 1120132093 N * _mcp mcp 1120132375 J * rs ~rs@ATuileries-153-1-5-47.w82-123.abo.wanadoo.fr 1120132556 M * DaPhreak hmm Doener, whats that about your "experimental daemon+client" ? 1120132575 M * DaPhreak lo btw ;) 1120132585 M * Doener if you don't mind reading ugly code, I'll upload it ;) 1120132599 M * DaPhreak yeah .. 1120132607 A * DaPhreak doesn't care on ugly code ;) 1120132660 M * Hollow Vudumen: which libvserver? ;) 1120132671 M * Hollow the one in util-vserver or the one at home.xnull.de ? ;) 1120132719 M * DaPhreak heh, wonder he's awake ;P 1120132752 M * Hollow he'll read it in the backlog then ;) 1120132758 M * DaPhreak yeah *g* 1120132767 M * DaPhreak nah .. still fscked *grml* 1120132782 M * Hollow DaPhreak: i finished vsched this morning 1120132785 M * Doener http://www.13thfloor.at/~doener/vserver/tools/vservd-0.02.tar.bz2 1120132789 M * DaPhreak great to hear :) 1120132799 M * Doener untar, make, start ./vservd as root, start ./client as root 1120132814 M * Doener currently it supports a unbelievable number of 3 commands! 1120132817 M * Doener ;) 1120132827 M * Doener quit = stop the daemon 1120132853 M * Doener create 1234 = create context 1234, without any setup, keep it for 10 seconds 1120132882 M * Hollow Doener: remote api calls? 1120132888 M * Doener enter 1234 = enter context 1234, its namespace and chroot, then start /bin/login 1120132902 M * Doener the client then displays the output of /bin/login 1120132907 M * Hollow ah 1120132942 M * Hollow maybe this could be integrated to my libvserver as well 1120132955 M * Hollow i'll set up a svn repo soon 1120132958 M * Doener client takes the user interaction and communicates with the server via a unix domain socket, which does all the work 1120133035 M * Doener hm, wouldn't it make more sense to have the daemon take advantage of your libvserver? not the other way round... 1120133069 M * Hollow yeah, that's what i meant ;) 1120133090 M * Hollow libvserver i.e. lib + basic tools 1120133096 M * DaPhreak depends on what that daemon provides ;) 1120133105 M * Hollow yep ;) 1120133139 M * Doener the daemon may become a complete vserver management thingy... 1120133148 M * DaPhreak hmm 1120133153 M * Doener somewhat like the stuff i've seen from solaris zones admin 1120133160 M * DaPhreak something like vserver #name enter ? 1120133173 Q * ruuth Quit: Nettalk6 der Freeware IRC-Client 1120133175 A * DaPhreak never looked at solaris 1120133192 M * Doener i didn't use it, but i've read some articles lately 1120133213 M * DaPhreak well we got some (~20-30) solaris machines here ... also some HP-UX and AIX but thats not my territory (sadly) 1120133230 M * Hollow Doener: what about using RPC or SOAP? 1120133259 M * Vudumen Hollow: liveserver from your server :) (xnull) 1120133278 M * Hollow Vudumen: yeah, i wrote it ;) 1120133678 A * Vudumen hates debianizing... :) 1120134013 J * sukria_ ~sukria@213.223.184.201 1120134032 Q * sukria_ Quit: 1120134345 M * Doener Hollow: hmm... 1120134351 M * Hollow Doener: hm? 1120134405 M * Doener any specific reasons why i should do that? 1120134433 M * Hollow you can't use unix sockets remotely, can you? 1120134473 M * Doener no, but i don't want to reinvent ssh either... 1120134517 M * Hollow but e.g. a webapp could easily connect to that daemon using rpc or soap 1120134639 M * Doener yep... maybe some intermediate process could handle translations 1120134693 M * Doener the basic daemon should provide almost nothing but the stuff the tools currently do 1120134775 M * Doener some very basic programs could then act in place of vcontext and friends (just connect to the socket, send a packet through the unix socket and be done) 1120134812 M * Doener for soap, there could also be some program, that accepts SOAP stuff and talks to vservd 1120134822 M * DaPhreak Doener: did you guys (you and Bertl) hear anything from Enrico ? 1120135189 M * Doener my conversation with enrico only happend on irc (till he left) and the ml... in early may i sent him some examples for the wait-for-context stuff, don't remember if he replied (too lazy to check ;), but savannah cvs shows that he started working on that... 1120135210 M * Doener last thing I saw was his mail on the ml on May 30th 1120135728 M * DaPhreak hmmm .. yeah that was also last sign here :) since Bertl said, he will only release 2.0 (AFIAK, or better as far as i _remember_) if the tools get some *serious* more testing (and becomes more bug-free) 1120136036 M * Doener what about "IIRC"? ;) 1120136156 M * DaPhreak yeah ;) "If I recall correct" *g* was also not able to recall that 1120136647 J * stephenM ~stephen@user-2774.l6.c5.dsl.pol.co.uk 1120137350 M * Hollow so... here it is ;) http://dev.gentoo.org/~hollow/vserver/libvserver/ 1120137353 M * Hollow libvserver-0.1 1120137809 Q * rs Remote host closed the connection 1120137893 J * rs ~rs@ATuileries-153-1-5-47.w82-123.abo.wanadoo.fr 1120138019 M * Doener 15:26:58 ERROR 403: Forbidden. 1120138124 M * Doener Hollow: not yet sure about releasing it? ;) 1120138166 M * DaPhreak heh 1120138220 M * Hollow ehm 1120138222 M * Hollow i hate toucan 1120138231 M * Hollow so should work 1120138241 A * Hollow should change his umask on d.g.o 1120138276 M * _ag_ does anyone know how to do chattr +t /vservers on XFS? 1120138662 M * Doener Hollow: hmm... you combined all the headers? 1120138778 M * Hollow Doener: which headers? 1120138806 M * Doener I'm looking at lib/vserver.h 1120138852 M * Doener that one includes many of the other headers... if you now change some network related header, all tools will get recompiled, although they're not affected 1120138881 M * Hollow i was just to lazy to look which file needs which header ;) 1120138884 M * Hollow low prio 1120138908 M * Doener hm, i just canceled my 2.6.13-rc1 port, I'll try to clean that up... 1120138934 M * Hollow :) 1120138964 N * jonsmel_zZ jonsmel 1120139271 A * DaPhreak stabs QT 1120139274 M * DaPhreak *grml* 1120140689 M * Hollow back later... 1120140812 Q * stephenM Ping timeout: 480 seconds 1120141053 A * DaPhreak starts to like Thunderbird ;P yeah I know mutt is even better ;) 1120141120 J * AprilDL ~chatzilla@ip68-9-200-247.ri.ri.cox.net 1120141339 M * AprilDL my purpose is Debian hosting for customers - mainly php mysql apache - isolating problems they may create from harming other customers.... I have been studying vserver for several days and started reading about Xen today. For a hosting company - what makes vserver a better choice than Xen? I see that a large number of hosting companies use vserver. 1120141481 M * DaPhreak well vserver-guest or simply guests don't use their own (maybe different) kernel as xen does ... 1120141823 J * eXplasm explasm@p549F7B95.dip.t-dialin.net 1120141826 M * _are_ vserver = ~0.2% cpu overhead compared to the plain kernel 1120141840 M * _are_ xen. 1 extra kernel per instance 1120141876 M * eyck vserver is stable for such purposes, Xen is highly experimental, 1120141877 M * _are_ so if you want to run all the same kernel version anyway for some virtual hosting, I'd use vserver, if you want to run different kernels, I'd use Xen 1120141885 M * eyck Xen is the future though, 1120141907 M * TheSeer doubtly 1120141921 M * TheSeer Xen has a totally different purpose 1120141932 Q * eXplasm Remote host closed the connection 1120141939 M * eyck so? 1120141969 J * eXplasm explasm@p549F7B95.dip.t-dialin.net 1120141979 M * TheSeer vserver is more like bsd jails 1120141987 M * TheSeer while xen is not 1120142007 M * TheSeer xen is more like vmware 1120142023 M * TheSeer to compare both with other existing products 1120142024 M * eyck ...so? 1120142036 M * _ag_ eyck: so you can use both xen and vserver :) 1120142044 M * eyck ...and? 1120142051 M * TheSeer what's your point? 1120142069 M * eyck what is YOUR point, you're the one arguing 1120142087 M * TheSeer i didn't argue.. i just said doubtly about your "future" statement 1120142098 M * TheSeer since xen and vserver don't "fight" each other 1120142102 M * eyck what exactly doubtfull there is about that? 1120142105 M * TheSeer they have a different purpose 1120142110 M * eyck ..so? 1120142125 M * TheSeer so the future of xen has no effect on vserver 1120142135 M * eyck ..and? 1120142145 M * TheSeer no and 1120142147 M * _ag_ this is the muppet show :) 1120142149 M * eyck what's your point? 1120142196 M * TheSeer your statement - made in this channel and in the context you said it - is wrong - imho 1120142205 M * TheSeer since, again, vserver and xen don't conflict 1120142206 M * eyck which statement? 1120142221 M * TheSeer Xen is the future though, 1120142226 M * eyck like mainframe business and unix busines 1120142236 M * eyck and you think Xen has no future, 1120142241 M * TheSeer i didn't say that 1120142255 M * eyck because being different from vserver takes it's future away from it? 1120142263 M * TheSeer i didn't say that either 1120142272 M * eyck what exactly did you try to say then? 1120142280 M * TheSeer i took your statement in context to the future of vserver 1120142304 M * TheSeer whereas your statement that Xen is the future, translates to vserver has none 1120142321 M * eyck I didn't say that, 1120142352 M * eyck I do believe that Xen will eat some of vserver's "marketshare" , but I didn't say that either, 1120142365 M * eyck this discussion has no sense. 1120142368 M * _ag_ eyck: this is not the same market 1120142378 M * eyck let's postpone it untill we're all a bit more drunk. 1120142394 M * eyck _ag_: exactly like mainframe and unix markets in the past, 1120142497 M * _ag_ eyck: i get your point here :) 1120142890 M * AprilDL I'm not too smart. For my first creation of vserver, shall I stick with the 2.4 kernel and the 1.2.10 vserver? I have a nice fresh debianized dell 2450 server waiting for me to give vserver a try. It has 2.4 altho I could probably get one of our tech guys to upgrade it to 2.6. 1120142916 M * TheSeer i'd use 2.6 1120142932 M * TheSeer 2.4 is dead - or will be soon 1120143063 M * _are_ I would use 2.6, most recent patches and alpha utils. 1120143111 M * _are_ well, actually I do. 1120143126 M * eyck for fun I would go with 2.4, 1120143135 M * eyck 2.6 is only slowly starting to stabilisie now, 1120143162 M * eyck argh, for fun, 2.6 1120143166 M * eyck 2.4 for servers. 1120143249 M * _are_ eyck: always depends on what you want to do, the functions I use had no changes for quite some time 1120143456 M * AprilDL mysql - php - apache - sendmail type of hosting is the purpose. Keeping customers isolated from each other and hopefully measuring their resource usage. 1120143524 M * eyck uh, toys, go with 2.6 1120144338 Q * janra Ping timeout: 481 seconds 1120145500 N * Bertl_zZ Bertl 1120145553 M * Bertl morning folks! 1120145605 M * aba Bertl: since when are you in US time zone? 1120145607 M * AprilDL 2.6.11 ? 1120145623 M * DaPhreak lol aba ;) 1120145631 M * DaPhreak lo Bertl 1120145639 M * Loki|muh AprilDL: bertl has his own timezone ;) 1120145642 M * Loki|muh ups 1120145645 M * DaPhreak ;) 1120145649 M * Loki|muh that was for aba 1120145708 M * Bertl aba: BTC = Bertl Time Zone :) 1120145722 M * alexx :-D 1120145764 M * Bertl it often happens to be closer to US time than to CET .. 1120145805 M * DaPhreak so it's CET+10 ? 1120145830 M * Bertl sometimes :) 1120145846 M * DaPhreak heh 1120145963 M * Bertl ah, AprilDL is new? 1120145974 M * Bertl welcome AprilDL then! 1120146220 M * AprilDL thanks Bertl or Bert1 1120146292 M * Zoiah It's Bertl, not Bert1. ;) 1120146302 A * Bertl .o( hmm, some clients seems to use unreadable fonts :) 1120146401 M * Bertl *seem 1120146429 M * DaPhreak yeah .. something like cp1252 ;) 1120146452 M * Bertl :) 1120146604 J * janra janra@paradox.homeip.net 1120147104 Q * _are_ Quit: bbl 1120147279 M * AprilDL Ok I now have 2.6.11 kernel and I have completed apt-get install kernel-patch-vserver. So far its easier than the 200 different vserver websites make it sound, on Debian anyway. But then I guess I haven't really done anything significant. 1120147296 M * AprilDL yet 1120147312 M * aba AprilDL: you need to apply the patch and build the kernel 1120147368 M * Doener AprilDL: http://linux-vserver.org/Step-by-Step+Guide+2.6... guess you can leave out the part about getting the sources ;) 1120147399 M * Bertl aloha Doener! 1120147404 M * Bertl off for dinner now :) 1120147411 N * Bertl Bertl_oO 1120147415 M * Doener Bertl: 2.6.13-rc1 introduces configurable HZ (100, 250, 1000) thus i rather not try to port vserver to that on 1120147416 M * Doener ;) 1120147422 M * Doener s/on$/one/ 1120147670 M * DaPhreak hmmm the CONIG_HZ is already in <2.6.13-r1 .. or is that something different ? (or was it the vs-patchset that introduced that?) 1120147740 M * DaPhreak off for back treatment ;( 1120147743 M * DaPhreak cu later guys 1120148897 N * Bertl_oO Bertl 1120148923 M * AprilDL any idea where the vserver patch goes or what it is named? (installed with apt-get) ... i tried find, i tried guessing... so far i cannot find it 1120148944 M * Bertl Doener: well, not the first thing which was rejected by the lkml folks and sneaked in some time later :) 1120148956 M * Doener /usr/src/something... should be includable via make-kpkg... 1120148980 M * Doener yeah, but I don't know the least about it, thus i won't touch it *g* 1120149094 M * Bertl Doener: k, np 1120149126 M * Doener the other stuff seemed to be pretty obvious, so i guess you don't miss the few rejects i fixed 1120149153 M * Doener and the macintosh file that is missing was removed because of being obsoleted 1120149209 M * Doener dinner time, back later 1120151805 M * _ag_ AprilDL: /usr/src/*.tar.bz2, I suppose 1120151860 M * AprilDL I found it, but I can't outwit make-kpkg somehow it keeps insisting I'm not in the top level linux source dir 1120151896 M * AprilDL debian:/usr/src/kernel-patches/all/apply# ls 1120151898 M * AprilDL vserver vserver-quota 1120152001 M * _ag_ AprilDL: you have to compile a kernel (apt-get install kernel-source-), go in top level linux source dir, and then issue : make-kpkg --added-patches vserver 1120152008 M * Bertl AprilDL: hmm, what's the point in 'outwit'ing kpkg? why not jsut build it yourself, with your config? 1120152042 M * _ag_ Bertl: the sake of debianess? ;P 1120152045 M * AprilDL its nice to at least learn the debian way i guess 1120152082 M * Bertl ah, I see ... :) 1120152304 M * jonsmel but much easier to build it yourself 1120152309 M * jonsmel :) 1120152317 M * Bertl evening jonsmel! 1120152337 M * jonsmel I love debian but sometimes there are easier ways to do things 1120152341 M * jonsmel hey bert 1120152348 M * jonsmel how's it going today 1120152362 M * Bertl fine, thanks! and for you? 1120152370 M * jonsmel btw, it's still morning here, hehe :) 1120152379 M * jonsmel good tnx! 1120152388 M * Bertl so then a good morning to you :) 1120152400 M * jonsmel yep! :) 1120152417 M * jonsmel still working on breaking gfs to get it to do what I want 1120152438 M * jonsmel but linux-cluster say it can't be done, but I'm sure there is a way :) 1120152540 M * Bertl who/what is linux-cluster? 1120153017 M * DaPhreak Bertl: maybe openmosix ? 1120153116 M * Bertl maybe ... 1120154484 M * jonsmel sorry, gone for a min 1120154506 M * jonsmel linux-cluster are the ones who make gfs and all the pieses that go with it 1120154513 M * jonsmel pieces 1120154552 M * jonsmel they have a irc #linux-cluster on freenode 1120154604 M * jonsmel anyway, I am working with a couple of ppl that are trying to acomplish the same thing I am 1120154640 M * jonsmel and we might have the linux-cluster ppl talked into making a patch to allow for what we are wanting to do 1120154645 M * jonsmel so we'll see 1120154666 Q * sukria Quit: see you 1120154683 J * tomi ~tomi@pha-84-242-95-4.nat.karneval.cz 1120154687 M * tomi Hi 1120154708 M * Bertl welcome tomi! 1120154908 M * tomi where can I download grp secure patch ? on 2.6.12? 1120154992 M * Bertl even you mean a combo linux-vserver + grsec? 1120155000 M * Bertl s/even// 1120155001 M * tomi now without.. 1120155011 M * Bertl probably at some grsec site, no? 1120155050 M * tomi yes. 1120155060 M * tomi where are patchs?? 1120155090 M * Bertl no idea, I just know the linux-vserver + grsec patches ... after all this is the linux-vserver channel :) 1120155104 M * tomi I know.. 1120155114 M * tomi but I need other.. 1120155134 M * Bertl well, then you have to shoot your own .... 1120155863 A * Hollow waves 1120156181 M * Bertl hey Hollow! 1120156221 M * Hollow hey Bertl, did you read about libvserver on the ML 1120156229 M * Bertl yep 1120156249 M * Bertl looks good, hope folks will help/participate 1120156283 M * Hollow thx.. i hope so too, DaPhreak already told me he will help me in his vacation 1120156413 M * Hollow Bertl: maybe you could take a look at the perror thingy when using shared objects, i don't quite know what to look for and where 1120156649 M * Bertl have you decided which libc to use yet? 1120156679 M * Bertl personally I'd prefer sticking with something more lean than the default glibc ... 1120156684 M * Hollow tbh i'd like to use glibc, but i don't know what is the best one.. 1120156685 A * DaPhreak waves at Hollow ;P 1120156717 M * Bertl Hollow: okay, then you ahve to figure how glibc does the threaded/pic errno handling 1120156729 M * DaPhreak uclibc ;) 1120156749 M * Hollow hm.. uclibc requires an extra toolchain.. bleh 1120156777 M * DaPhreak yeah :) either diet, glibc or even another (which i don't know) 1120156806 M * Bertl the thing is, glibc is really bloated for our purpose ... 1120156820 M * Bertl and we already spent a lot of time fixing dietlibc 1120156828 M * DaPhreak yeah .. 1120156877 M * Hollow Bertl: so uclibc would be your choice too? 1120156905 M * Bertl probably not :) 1120156931 M * Hollow hm.. i only know glibc, diet and uclibc 1120156936 M * DaPhreak heh, rather dietlibc i guess ;) 1120156955 M * Hollow but if we need to fix dietlibc all the time? 1120156959 M * Bertl yes, especially as it seems to be pretty cross compatible 1120156972 M * Bertl i.e. dietlibc can be used with cross compilers quite fine 1120157022 M * Hollow hm 1120157029 M * DaPhreak Bertl: what says your testing currently on the other archs? i know about the amd64 stuff .. but not about alpha/sparc and friends 1120157058 M * Bertl we had to fix some alignment issues on various 64bit archs 1120157073 M * Hollow ok, dietlibc is fine 1120157077 M * Bertl but once detected, they were easy to fix 1120157090 M * Bertl the syscall for dietlibc is still broken on many archs 1120157097 M * Bertl but we do not use that anymore ... 1120157110 M * Hollow i.e.? 1120157114 M * DaPhreak yeah, we've got our own 1120157115 M * DaPhreak ;) 1120157121 M * Hollow yeah.. 1120157125 M * Hollow :P 1120157131 M * Bertl an Hollow, btw: the library should also add methods for related syscalls 1120157141 M * Hollow e.g.? 1120157143 M * Bertl especially the personality and cpuset stuff 1120157152 M * Bertl cpuset/cpu affinity 1120157161 M * Hollow ah personality btw: what is that? 1120157175 M * Bertl if you have amd64 (aka. x86_64) 1120157198 M * Bertl you can get a 32bit compatibility mode (aka linux32 personality) 1120157203 M * Hollow ah 1120157226 M * Bertl same is true for sparc/64 and mips/64? 1120157285 M * Hollow another question regarding flags and caps... i'm a bit lost with this mask and bit-shift stuff... do you know any doc about the idea? 1120157302 M * Bertl bit shift? 1120157314 M * Hollow a & b or a |= b etc pp 1120157330 M * Hollow bitwise operators 1120157332 M * Hollow ;) 1120157341 M * Bertl ah. ic ... 1120157362 M * Bertl well, you know how logic operations (AND, OR, XOR, NOT) work? 1120157368 M * Hollow i know what these operators are doing and how they work, but i don't get the idea how to set the caps and flags 1120157369 M * daniel_hozac boolean algebra ;) 1120157420 M * Bertl Hollow: okay, so you know how boolean algebra works? 1120157431 M * Hollow yup 1120157442 M * Bertl 1 ^ 1 = ? 1120157450 M * Bertl (^ being xor) 1120157460 M * Hollow 0 1120157470 M * Bertl good , just checking :) 1120157473 M * Hollow :P 1120157491 M * Bertl you know what a |= ~2 means ? 1120157521 M * Hollow i know what |= does... but ~2 dunno 1120157530 M * daniel_hozac negate 2 1120157532 M * Bertl ~ is the complement (inversion) 1120157564 M * Hollow hm.. what would ~2 be then? 1120157593 M * Bertl 000...010 = 2 1120157602 M * Bertl 111...101 = ~2 1120157606 M * Hollow ah 1120157623 M * Bertl okay, you ahve the kernel code at hand? 1120157628 M * Hollow yup 1120157655 M * Bertl vc_set_cflags 1120157680 M * Hollow in kernel/vserver/context.c? 1120157681 M * Bertl gets as argument a flagword and a mask both 64bit 1120157689 M * Bertl yep 1120157711 M * Hollow ok 1120157745 M * Bertl at the end of that procedure, you see the actual 'masking' done by vx_mask_flags() 1120157764 M * Bertl #define vx_mask_flags(v,f,m) (((v) & ~(m)) | ((f) & (m))) 1120157781 M * Bertl v = value, f = flags, m = mask 1120157789 M * Bertl now what happens here is: 1120157806 M * Bertl f & m .. cut out only flags mentioned in mask 1120157823 M * Hollow so i set mask to the flags i want to change? 1120157831 M * Bertl v & ~m .. remove flags mentioned in mask 1120157848 M * Bertl a | b (take flags from both) 1120157864 M * Bertl so basically we do remove all bits contained in the mask 1120157874 M * Bertl and replace them by the appropriate bits from the flags 1120157889 M * Hollow ah 1120157948 M * Bertl v f m q 1120157948 M * Bertl ------------- 1120157948 M * Bertl 0 X 0 0 1120157948 M * Bertl 1 X 0 1 1120157948 M * Bertl X 0 1 0 1120157950 M * Bertl X 1 1 1 1120157961 M * Bertl this would be the truth table for the vx_mask_flags() 1120157977 Q * rs Quit: rs 1120158000 M * Bertl so the mask works as 'selector' between old and new vlaues 1120158012 M * Bertl (does this make sense for you?) 1120158045 M * Hollow can you give an example what v,f and m woudl be if i want to set e.g. fakeinit 1120158080 M * Bertl you would 'mask' out fakeinit (m = fakeinit) and set the corresponding flag (f = fakeinit) 1120158103 M * Bertl to clear it, you would use m = fakeinit, and f = 0 1120158108 M * Hollow so the mask removes the bit regardless of its value 1120158121 M * Hollow and i can set it with the new value 1120158146 M * Bertl the mask decides where the flag is taken from, 0 emans from the cutrrent setting, one means from the 'new' flags 1120158158 M * Hollow ok 1120158179 M * Bertl we also define so called 'one time' flags 1120158193 M * Bertl they can be cleared once, that's it ... 1120158217 M * Bertl that's what the vx_mask_mask is used for 1120158243 M * Bertl #define vx_mask_mask(v,f,m) (((v) & ~(m)) | ((v) & (f) & (m))) 1120158311 M * Bertl v f m q 1120158311 M * Bertl ------------- 1120158311 M * Bertl 0 X 0 0 1120158311 M * Bertl 1 X 0 1 1120158311 M * Bertl ------------ 1120158314 M * Bertl 0 X 1 0 1120158316 M * Bertl 1 0 1 0 1120158319 M * Bertl 1 1 1 1 1120158345 M * Doener ARGH 1120158348 M * Bertl the difference here is, that the flag cannot be set, once it has been cleared 1120158358 M * Bertl argh? 1120158368 M * DaPhreak he typed rm -rf / ;) 1120158380 M * Hollow lol 1120158385 M * Doener just wanted to diff foo-0.1 and foo-0.1.1... 1120158405 M * Doener i had messed up foo-0.1 and wanted to replace it with the clean tarball contents 1120158415 M * Doener and yes, i did rm -r foo-0.1.1 1120158425 M * Bertl oops :) 1120158444 M * Hollow Bertl: still bit confusing, but i'll try to understand it and if i'll stuck i ping you ;) 1120158447 M * Doener yeah, now i need to figure out all the automake stuff again :( 1120158466 M * DaPhreak *uugh* 1120158480 M * DaPhreak yeah, "great" automake :) 1120158655 M * jkl does anyone run courier-imap on a gentoo vserver? I'm having trouble with the init script 1120158680 M * DaPhreak what exactly says the init-script ? 1120158687 M * jkl [!!] 1120158688 M * jkl hehe 1120158699 M * jkl if i run the rc script by hand courier starts ok 1120158703 M * Hollow Bertl: so i'd use v = vxi->vx_flags, f = 1 and m = VXF_STATE_INIT to set fakeinit? 1120158706 M * DaPhreak heh ;) something in syslog or even an courier-imap log ? 1120158737 M * Hollow jkl: set the ip of your alias interface in /etc/courier-imap/imapdrc 1120158744 M * DaPhreak ah :) 1120158745 M * daniel_hozac Hollow: f = VXF_STATE_INIT 1120158775 M * Hollow and m = ? 1120158776 M * jkl Hollow: should that file exist? says it's new.. 1120158791 M * Hollow jkl: look for similar name.. dunno 1120158798 M * Hollow the imap config file.. 1120158800 M * jkl hehe ok, looks like its just impad 1120158802 M * DaPhreak jkl: maybe in conf.d ? no idea *g* 1120158822 M * Hollow DaPhreak: nope.. 1120158827 M * jkl hrm, i have it set to the alias address already 1120158834 M * Hollow hm.. 1120158840 M * jkl ADDRESS=10.0.1.15 1120158843 M * Hollow any logs? 1120158848 M * jkl maybe i should specify the port too 1120158864 M * Hollow jkl: maybe on the host another process listens on 0.0.0.0 on that port? 1120158883 M * jkl no, unfortunatley not 1120158892 M * jkl i may be looking at the wrong config file though 1120158893 M * jkl hm 1120158932 M * jkl no, it looks like its the one the rc file is calling up 1120158990 M * jkl it's strange, if i call the rc file directly it starts up just fine 1120159082 M * Hollow jkl: which baselayout version? 1120159142 M * jkl hmmm 1120159149 M * jkl i have baselayout-vserver 1.0 1120159164 M * jkl looks like 1.11.12-r4 may be a good upgrade? 1120159178 J * Aiken ~james@tooax6-049.dialup.optusnet.com.au 1120159214 M * Hollow yup... though i don't know if it fixes the issue i think you have.. i remember a issue with the start-stop-daemon wrapper provided by rc-services.sh... but you should give it a try 1120159229 M * Hollow though... 1120159244 M * Hollow maybe first try to replace start-stop... lines in the init script with /sbin/start-stop... 1120159316 M * jkl so change start-stop-daemon to /sbin/start-stop? 1120159353 M * Hollow ./sbin/start-stop-daemon 1120159356 M * Hollow without . 1120159412 M * jkl ha 1120159424 M * jkl well that was incredibly easy 1120159432 M * jkl is that an issue with my current path or somethign? 1120159437 M * jkl *something 1120159461 M * jkl maybe this will fix the problems im having with all my other init scripts =) 1120159533 M * Hollow ok.. try to upgrade to the latest baselayout-vserver and see if it fixes the problem 1120159562 M * jkl i just did the same for mysql, and that fixed that problem too 1120159575 M * jkl so, ill be merging up to the new baselayout here shortly 1120159594 M * Hollow fine, i really need testers on that one to mark it stable asap 1120159612 M * DaPhreak heh, don't know how many already tested it *g* 1120159618 A * DaPhreak knows of 2 ;) 1120159620 M * Hollow dunno too 1120159632 M * Hollow reminds me to ping bonsaikitten 1120159796 M * jkl ah, so ill be a guinea pig 1120159877 M * jkl dude, that bonsaikitten stuff on the net is funny 1120159881 M * jkl but its kinda messed up too 1120159984 M * jkl are there any good docs on using the same portage tree for multiple vservers? 1120160060 M * Hollow jkl: sec 1120160104 M * Hollow jkl: http://phpfi.com/67841 1120160113 M * Hollow in /etc/vservers/name/fstab 1120160187 M * DaPhreak lol never saw a windows-app eating 407M Ram .. until know 1120160399 M * DaPhreak s/know/now/ 1120160653 M * jkl Hollow: so that mounts a portage dir from the host on the vserver 1120160694 M * Hollow jkl: yep, and if you have a BME enabled kernel (such as vserver-sources) you can use ro and rw in the bnid mount command... 1120160733 M * Hollow so you can share a read-only portage tree with a read-write distfiles inside 1120160877 M * jkl i'm using a custom kernel, what's BME? 1120160889 M * Bertl Bind Mount Extensions ... 1120160900 M * jkl thanks, bertl ill look around for that 1120161061 M * Bertl http://vserver.13thfloor.at/Experimental/BME/ 1120161344 M * maharaja Bertl: i told you about the server problems i encountered 1120161351 M * Bertl yup? 1120161364 M * maharaja you guessed that it was an hardware fault 1120161370 M * Bertl yup? 1120161373 M * maharaja i change the everything 1120161382 M * maharaja p4, 1.5ghz -> p4 3.2ghz 1120161389 M * maharaja everything except the hds is new 1120161399 M * maharaja i still encounter the same problems 1120161423 M * maharaja the hds are 2x brand new 200gb disks (never used before) - raid1 1120161435 M * Bertl could you remind me what the issues were? 1120161443 M * maharaja the server is unavailable 1120161450 M * maharaja it hangs 1120161452 M * maharaja freezes 1120161467 M * Bertl ah, yep, but no panic or so, right? 1120161469 M * maharaja no 1120161479 M * maharaja i enabled the serial line 1120161492 M * maharaja most of the time, there is no oops/panic/... 1120161507 M * Doener most of the time? doesn't sound that good ;) 1120161519 M * maharaja 2 times i encountered a kernel oops 1120161533 M * maharaja but i had no debugging options on so i wasn't able to get any information 1120161568 M * Bertl do you have a record of the oops? 1120161648 M * maharaja nope, i deleted the information after the hardware exchange and some basic testing 1120161665 M * maharaja it used to oops/freeze at io operations 1120161673 M * Bertl good decision, means less work for me :/ 1120161682 M * maharaja i did some 500mb dds from /dev/random and hoped for the best 1120161705 M * maharaja /dev/urandom of course 1120161802 M * maharaja well, tell me how to help 1120161815 M * Bertl collect good oops traces :) 1120161841 M * maharaja what kernel debug settings do i have to enable? 1120161842 M * Bertl if it doesn't oops, just hang, use the magic sysrq to dump the tasks 1120161872 M * maharaja mhm 1120161877 M * maharaja do you know ser2net? 1120161884 M * Bertl yes 1120161891 M * maharaja is it possible to do that via ser2net? 1120161904 M * Bertl no, but you can use the netconsole for that 1120161920 M * Bertl but sysrq via serial line is better ... 1120161921 M * maharaja never heard of netconsole 1120161942 M * maharaja i do that via serial console 1120161957 M * maharaja crashy server <-> monitoring server (running ser2net) 1120161973 M * maharaja so i can do stuff from the monitoring server 1120161990 M * maharaja my configuration: http://raoul.bhatia.at/~raoul/config-2.6.11.12-vs2.0-rc4-p4b 1120161994 M * maharaja do i have to enable anyting else? 1120161995 M * Bertl yeah, well, you can use minicom there, don#t see the point in network redirection 1120162090 M * maharaja ok, i got minicom up 1120162134 M * maharaja "main:~# echo "-- MARK" > /dev/ttyS0" is transmitted and i see the output via minicom 1120162158 M * Bertl CONFIG_DEBUG_KERNEL=y 1120162162 M * Bertl CONFIG_DEBUG_SPINLOCK=y 1120162165 M * Bertl CONFIG_DEBUG_SPINLOCK_SLEEP=y 1120162169 M * Bertl CONFIG_DEBUG_INFO=y 1120162173 M * Bertl CONFIG_VSERVER_DEBUG=y 1120162189 M * Bertl CONFIG_VSERVER_HISTORY=y 1120162194 M * Bertl CONFIG_VSERVER_HISTORY_SIZE=256 1120162213 M * Bertl should be enabled on the crash server 1120162257 M * Bertl then boot it with console=ttyS0,115200n8 1120162278 M * Bertl and watch the bootup via minicom (set to similar) 1120162280 M * maharaja ok, ill recompile the kernel 1120162286 M * Bertl then make the test machine hang/crash 1120162306 M * Bertl and either record the oops or dump the tasks via 1120162317 M * Bertl magic-sysrq (break in minicom) T 1120162377 M * maharaja can i test the magic sysrq stuff prior to a crash? 1120162379 M * maharaja so i know that it works 1120162416 M * Bertl yep, sure magic-sysrq H (just h no shift) will give you a help message 1120162430 M * maharaja ok, thank you 1120162437 M * Bertl yw 1120162449 M * maharaja should minicom be in "online" mode? 1120162473 M * Bertl depends on the kind of conenction 1120162487 M * maharaja because with the above config, it is in "offline" mode 1120162490 M * maharaja (kernel config) 1120162491 M * Bertl i.e. if the crossover cable is correctly wired 1120162514 M * maharaja it is as i used it before (6-12 monts ago when helping with smp debugging) 1120162529 M * Bertl it is supposed to go online, if something like mgetty connects 1120162540 M * Bertl (but we do not need mgetty here) 1120162596 M * maharaja mhm 1120162603 M * maharaja i try to "send break" in minicom 1120162611 M * Bertl CTRL-A Z F 1120162614 M * maharaja but the window does not disppear 1120162649 M * maharaja it writes "sending break" 1120162663 M * Bertl if it 'hangs' then either minicom is badly configured or the connection is bad 1120162683 M * maharaja /proc/sys/kernel/sysrq is 1 1120162690 M * maharaja mhm 1120162719 M * maharaja well, i know that i too used this cable 2 weeks ago for configuring my hp switch 1120162725 M * maharaja via serial line 1120162730 M * maharaja so it should work, shouldn't it? 1120162738 M * maharaja damn, i hate that... :) 1120162742 M * maharaja going on vacation tomorrow ... :-/ 1120162788 M * maharaja is there any minicom configfile you can provide? 1120162844 M * maharaja ok, now it worked 1120162851 M * maharaja "sending break" flashed 1120162869 M * maharaja but the second time it hangs 1120162959 M * Hollow off to bed.. cya 1120162962 M * maharaja mhm 1120162991 M * maharaja ok, i get some kind of feedback 1120163145 M * maharaja *compiling new kernel* 1120163320 M * Doener Bertl: hm, I'm trying to clean up Hollow's libvserver a little... what would be your suggestion regarding header files? i don't think the lib should come with the kernel header, should they? 1120163420 Q * stupidawy Ping timeout: 480 seconds 1120163506 M * Bertl Doener: hmm, you eman library headers, no? 1120163510 M * Bertl *mean 1120163550 M * Doener yeah, i mean that the lib headers (i.e. what users of the lib include) shouldn't include the kernel headers (those that are used in the hack tools) 1120163566 M * Doener i got some headache deciding about types atm... 1120163606 M * Doener i can either basically duplicate a few types and create extra code (might be useful later on though), or just move the types from the kernel headers to the lib headers... 1120163718 M * Bertl sec, phone 1120163799 M * daniel_hozac duplicate them. 1120163819 M * DaPhreak night guys :) 1120163823 M * Doener night DaPhreak 1120163968 M * Doener http://pastebin.ca/16552 -- that's what i mean by duplication... 1120164004 M * maharaja damn it... 1120164016 M * maharaja i can send the break cmd and get no feedback 1120164152 M * AprilDL vserver-quota is not for kernel 2.6.11 according to the message I get when I try that patch. What does vserver-quota do? 1120164173 M * AprilDL using 1.95 of vserver 1120164349 M * maharaja any1 who can help me getting minicom to work? 1120164389 M * Bertl maharaja: check without hardware handshake 1120164465 M * maharaja do you mean "hardware flow control"? 1120164494 A * Bertl is still on the phone 1120164498 M * maharaja k 1120164722 M * maharaja or maybe there is any1 else who can send me a minicom rc file 1120164940 Q * eXplasm Ping timeout: 480 seconds 1120165383 M * maharaja *still waiting* 1120165433 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1120165607 J * eXplasm explasm@p549F7B95.dip.t-dialin.net 1120165677 Q * stupidawy Quit: 1120165722 J * _mountie ~mountie@24.42.99.232 1120166028 M * maharaja Bertl: still abusing the phone? :) 1120166080 M * Bertl yep 1120166187 Q * mountie Ping timeout: 480 seconds 1120166303 M * maharaja damn 1120166435 M * maharaja right now i get the boot stuff in minicom 1120166440 M * maharaja but the magic sysrq still fails 1120166476 M * Bertl okay, so what do you try to get magic sysrq, and did you enable it in the kernel config? 1120166489 M * maharaja i did enable it 1120166545 M * maharaja cat /proc/sys/kernel/sysrq is 1 1120166576 M * maharaja i press: strg+a (inside screen) a f H 1120166603 M * Bertl CTRL-A Z F H IIRC 1120166635 M * maharaja i think that z is only used to display the help screen and is not neccesairy 1120166653 M * Bertl whatever, if it says sending break it's fine 1120166664 M * Bertl and you press 'h' not shift-h, okay? 1120166671 M * maharaja yes it does 1120166672 M * maharaja i once got the help stuff 1120166675 M * maharaja but only once 1120166680 M * Bertl well, then it works ... 1120166690 M * maharaja pressed some enters and did an echo "xxx" > /dev/ttyS0 on the crash-server 1120166695 M * maharaja then it magically appeared 1120166718 M * maharaja and somehow there seems to be a delay or so 1120166722 M * Bertl sounds like an unreliable serial line 1120166729 M * maharaja mhm 1120166741 M * maharaja should a "mount" be displayed in minicom? 1120166758 M * maharaja like in dmesg 1120166785 M * Bertl if you booted with the console=ttyS... option, then yes, everything in dmesg will be on the console too 1120166812 M * maharaja mhm 1120166814 M * maharaja it is not :) 1120166815 M * Bertl Doener: ad headers 1120166819 M * maharaja damn it 1120166844 M * Bertl Doener: the API (i.e. kernel headers in include/vservers are _clean_ 1120166859 M * maharaja Bertl: it worked with ser2net 1120166860 M * Bertl i.e. they are supposed to work inside the kernel as well as for userspace 1120166867 M * maharaja Bertl: in minicom, things seem broken 1120166877 M * Bertl maharaja: well, I have absolutely no problem with ser2net either 1120166890 M * Bertl maybe you should compile the minicom yourself? 1120166894 M * maharaja but it won't transmit the sysrq key, right? 1120166902 M * Bertl maybe you currently use some debian version *G* 1120166907 M * maharaja is VT102 the right choice? 1120166909 M * Doener yeah, and i see no problem using them with userspace apps, but it seems a little strange to me, to ship them with a lib... 1120166945 Q * eXplasm Remote host closed the connection 1120166996 M * Bertl Doener: yeah, well, no problem with 'redefining' some types, but they have to be ket in sync 1120167000 M * Bertl *kept 1120167028 M * Bertl IMHO, the best way would be to split it into 1120167048 M * Bertl vserver-kernelheaders which go into /usr/include/linux/vserver 1120167060 M * Bertl and library headers, which depend on those 1120167234 M * Bertl maharaja: I use ASCII/ANSI but that's not really important 1120167246 M * Bertl the handshake and xon/xoff is important 1120167409 M * maharaja i set flow control to off 1120167413 M * maharaja is this the correct setting? 1120167633 M * Bertl xon/xoff should be off, the hw flow control depends on your cable 1120167654 M * Bertl if it connects RTS/CTS then hw flow control should be on 1120167674 M * Bertl also check that you have only _one_ client attached to the serial port 1120167705 M * Bertl if you have for example two minicoms attached, they will receive partial output 1120167834 M * maharaja ok 1120167842 M * maharaja rebooting into the new kernel 1120167845 M * maharaja with the options you told me 1120167929 M * Bertl AprilDL: you are referring to the 2.4 quota patches? 1120167963 M * AprilDL well all i know is that i got the apt-get install vserver 1120167973 M * AprilDL and one of the patches was vserver-quota 1120167995 M * Bertl well, debian stuff is (very) broken (unfortunately) 1120168013 M * AprilDL my kernel is still building. dual 733 mhz p3's... i wouldn't have thought it would take so long. 1120168020 M * Bertl I have no idea what patches you are looking at, but the quota patches are labeled q0.14 or similar 1120168032 M * AprilDL bertL so you are from the source kind of guy ? 1120168057 M * Bertl well, more the 'working way' kind of guy 1120168068 M * maharaja Bertl: there is still no output when mounting /boot or something else 1120168071 M * AprilDL after spending the afternoon learning how to do the kernel build the debian way i expect you were right it would have been easier the non debian way 1120168106 M * Bertl maharaja: on bootup you should get the complete boot loog as in dmesg (if not something on your monitoring server is broken) 1120168130 M * maharaja Bertl: i get everything from the bootup 1120168161 M * maharaja the first thing i do not see is 1120168162 M * maharaja eth0: no IPv6 routers present 1120168169 M * maharaja kjournald starting. Commit interval 5 seconds 1120168170 M * Bertl AprilDL: well, I use mandrake (now called mandriva) which is rpm based, and I can use about 50% of their packages quite fine ... 1120168175 M * maharaja maybe its because of metalog? 1120168214 M * maharaja as i use metalog for logging 1120168228 M * Bertl AprilDL: nevertheless I have to (re)build a lot of packages, because they are not compiled the way I like it ... 1120168242 M * Bertl maharaja: what does metalog do with the serial port/console? 1120168257 M * maharaja Bertl: hopefully nothing 1120168324 M * maharaja but i don't know why its not working as expected 1120168337 M * maharaja echo "bla" > /dev/ttyS0 is always transmitted 1120168378 M * Bertl you do that where? 1120168425 M * maharaja on the crash-server 1120168426 M * maharaja and i see it on the monitoring server 1120168437 M * maharaja i started minicom with "minicom -o " 1120168439 M * maharaja is that correct? 1120168444 M * Bertl good, maybe the other direction doesn't work then? 1120168489 M * maharaja has the speed got something to do with the cable wires? 1120168504 M * maharaja like baud 115200 is not working with less wires 1120168528 M * Bertl the baudrate just depends on the 'quality' of the cable 1120168540 M * Bertl i.e. good cable -> high baud rate 1120168546 M * maharaja maybe i should try baud 9600 1120168554 M * maharaja = 1120168554 M * maharaja ? 1120168558 M * Bertl you can do that .. 1120168569 M * Bertl although 19200 should be supported by any cable 1120168598 J * complexho ~mark@funk.gotadsl.co.uk 1120168605 M * Bertl welcome complexho! 1120168611 M * complexho hi Bertl! 1120168689 M * complexho has anyone been discussing libvserver at all? 1120168700 M * complexho ...or is it a bit early ;) 1120168716 M * Bertl we are discussing it ... ;) 1120168720 M * complexho ah 1120168782 M * maharaja damn it 1120168784 M * maharaja i give up for today 1120168793 M * maharaja have a nice week, i'm off till 10. of july 1120168798 M * Bertl okay, cya 1120168916 M * complexho Bertl - if you remember I was in here quite a lot a couple of months back - I have been building a hosting cluster based on vserver since then and it's more or less built now 1120168939 M * complexho I've said it before and I'll say it again - vserver really rocks :) 1120168941 M * Bertl yup, I remember ... somewhat ... 1120168966 M * Bertl complexho: you added yourself to the happy vserver users? 1120168977 M * complexho not yet, but I should do :) 1120169065 M * Doener i'm off to bed, good night folks... 1120169132 M * Bertl night Doener! 1120169385 M * complexho Bertl: do you know of anyone who has managed to get 2.0 working against RH Enterprise (or centos et al) kernels? 1120169395 M * complexho if if anyone has ever tried? 1120169404 M * Bertl I guess nobody bothered 1120169410 M * complexho right 1120169451 M * daniel_hozac the RHEL kernels shouldn't be too different from Fedora's kernel. 1120169475 M * complexho I tried myself a while back and got about 75% of it in but became unstuck with my limited kernel compiling abilities ;) 1120169494 M * complexho yeah I use Fedora now, but always put in a vanilla kernel 1120169516 M * Bertl there is no real point in aptching distro specific kernels 1120169533 M * complexho but I have been asked about it by a client who wants to use the GFS Clustering underneath vservers, and it's messy unless you go with RHEL 1120169544 M * complexho or seems to be 1120169552 M * Bertl well, the GFS will need modifications anyway 1120169559 J * EverD ~EverD@81.222.149.146 1120169560 M * Bertl otherwise the vserver stuff will not work 1120169566 M * Bertl welcome EverD! 1120169575 M * EverD hi! 1120169594 M * EverD I came with my problem 1120169600 M * complexho what sort of modifications would be needed? 1120169624 M * EverD I don't understand why can i run fdisk inside the vserver 1120169650 M * Bertl EverD: well, I don't udnerstand why you would like to? :) 1120169690 M * Bertl complexho: flags, xid tagging, network restrictions 1120169699 M * EverD No, i don't like it :( . 1120169745 M * EverD I tryed to modify my partition table. I created new partition. 1120169755 M * Bertl ah, sorry misread your message 1120169772 M * Bertl you mean, you do not like that you _can_ run fdisk inside, no? 1120169773 M * EverD I want to deny such operations 1120169788 J * _are_ foobar@dsl-084-056-130-240.arcor-ip.net 1120169788 M * EverD Bertl, yah! 1120169800 M * Bertl yes, the question is, on which device do you run it? 1120169823 M * EverD fdisk /dev/sda 1120169831 M * Bertl where did /dev/sda come from? 1120169844 M * Bertl evening _are_! 1120169850 M * EverD just minute 1120169952 M * complexho Bertl: ok... never realised there were so many fs hooks going on... So am I correct in thinking this type of modification is required to run vserver on any new filesystem? 1120169983 M * EverD Bertl: my /dev/sda came from the gentoo's distro . When i unpucked the distro, my /dev/ was populated with many devices, including the /dev/sda 1120170027 M * Bertl okay, this is not how it should be ... 1120170109 M * EverD Bertl: What is the right way to populate /dev ? 1120170127 M * Bertl usually the tools do that, but you can 'fix' it if you use 1120170140 M * Bertl the skeleton build method, which creates an empty guest 1120170150 M * Bertl and copy the few devices from there into your gest 1120170153 M * Bertl *guest 1120170177 M * Bertl ls /vservers/gentoo/dev/ 1120170178 M * Bertl MAKEDEV@ full initctl| log= null ptmx pts/ random stderr urandom zero 1120170190 M * Bertl only those devices/files should be present 1120170199 M * EverD not many :) 1120170217 M * Bertl yes, you can add devices, if you know what you are doing 1120170226 M * Bertl for example to allow acces to serial ports or so 1120170239 M * EverD I see now... 1120170240 M * Bertl but I would not allow access to any block device 1120170268 M * EverD How to deny access to blockdevs? 1120170284 M * Bertl well, don#t copy any device node into the guest :) 1120170425 M * EverD mmm... Is it possible to create /dev/sda being inside the vserver ? And get access to block device ^) 1120170443 M * Bertl no, because you do not ahve the required cap to create device nodes :) 1120170484 M * EverD Aha! that's great ^) I love vservers! 1120170505 M * Bertl great! 1120170526 M * Bertl btw, we are considering some kind of 'device' protection to remove this restriction ... 1120170537 M * Bertl but it isn't top priority 1120170617 M * albeiro could you elaborate a bit more about it ? 1120170706 M * Doener back... 1120170735 M * Bertl albeiro: sure, the idea is to 'block' device access for certain major/minor combos 1120170756 M * Bertl this way we could allow for mknod operations inside a guest without reducing security 1120170785 M * Bertl and the requirement for a 'clean' guest without dangerous nodes would be removed 1120170804 M * albeiro sounds good to me 1120170852 M * Bertl yes, the only downside is that we code 'policy' into the kernel 1120170864 M * Bertl (i.e. the major/minors to allow) 1120170984 M * albeiro could not it be extended so choosing numbers from userspace would be possible ? 1120171121 M * albeiro "fallback" function in proper place in kernel checking permisions which would do it against some db kept in memory 1120171165 M * Doener or YAUSH? (yet another userspace helper ;) 1120171188 M * albeiro well, making dev nodes is not that common so why not ;) 1120171261 M * Doener hm, you cannot safely do that when the node is created... policy changes would not work, the old node would still be there 1120171284 M * Doener the check has to be done when the node is accessed and that may be quite common 1120171403 M * albeiro yeah. well, we can block device access with rsbac + vserver too ;p 1120171522 M * Bertl excellent, so with rsbac it's already safe .. no need to hurry :) 1120171561 M * EverD excuse me, where can i find documentation on configuring 1.9.x, and info about vshelper. (i've read the flower page :) ) 1120171586 M * Bertl EverD: sidenot, 2.0-rc4 is out, you might consider that over 1.9.x 1120171591 M * Bertl *sidenote 1120171618 M * EverD Where can i find it? 1120171650 M * Bertl http://www.13thfloor.at/~doener/vserver/patches/patch-2.6.12-vs2.0-rc4.diff 1120171659 M * Doener http://vserver.13thfloor.at/Experimental/patch-2.6.11.11-vs2.0-rc4.diff 1120171669 M * Doener http://linux-vserver.org/Step-by-Step+Guide+2.6 1120171703 M * albeiro what i really love is recent possibility of assigning attributes / permisions in "inheritance" major:minor (major is sufficient of course) manner - so keeping attibutes on devfs / udev managed /dev is easy :) 1120171706 M * Doener vshelper should work out of the box in most cases, otherwise the tools show some error message with instructions how to fix the problem 1120171908 M * EverD I followed instructions (the last one, where i used sysctl). But I don't quite understand, What does vshelper do? 1120171996 M * EverD I've an idea, that vshelper allows to reboot the guest but command, given inside the guest 1120172060 M * Bertl yep, this is one (the main) purpose of the vshelper 1120172083 M * EverD How can i use it? 1120172103 M * Bertl inside the guest, you invoke 'reboot' 1120172117 M * Bertl outside (on the host) the vshelper does the work 1120172165 M * EverD hm. I tryed. But reboot command said that the initctl doesn't respond 1120172207 M * Bertl depends if you use init based guests, then reboot is fine (or shutdown) because it will contact init 1120172222 M * Bertl if you have no init inside the guest, you better use reboot -f 1120172228 M * Bertl (which bypasses the init) 1120172284 M * EverD yes, my guest is not init based. :) 1120172317 M * Bertl reboot for example is usually just a link to init itself, and just tries to contact init 1120172639 M * complexho Bertl: I didn't realise that vserver required the filesystems themselves to be patched up. Which other fs does vserver work on? 1120172664 M * Bertl depends on what features you want to use ... 1120172689 M * complexho which patches should I be looking at to understand what is going on here? 1120172691 M * Bertl but currently ext2/ext3/reiser/xfs/(and partially jfs) are supported 1120172703 M * Bertl also nfs as network based fs 1120172745 M * complexho should I be looking in here? http://vserver.13thfloor.at/Experimental/FOR-2.0/ 1120172769 M * EverD Bertl, I was reading maillists, the thread about iptables in vservers. There was writen that there is no way to use iptables inside vservers but using CAP_NETADMIN wich allows all operations with network. But the discussion was a year ago. Is there any changes in that derection? 1120172779 M * Bertl complexho: no, better a broken out version of 1.9.x 1120172809 M * Bertl EverD: no, not really, iptables is done on the host, as all network ... 1120172813 M * Bertl *networking 1120172825 M * EverD okey 1120172833 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1120172850 M * complexho do you mean the main patch broken down by filename? 1120172923 M * Bertl http://www.13thfloor.at/vserver/d_rel26/v1.9.5/split-2.6.11.5-vs1.9.5.tar.bz2 1120172932 M * complexho thanks bbiab :) 1120173035 Q * brc Ping timeout: 480 seconds 1120173540 Q * stupidawy jupiter.oftc.net uranium.oftc.net 1120173540 Q * Aiken jupiter.oftc.net uranium.oftc.net 1120173540 Q * Doener jupiter.oftc.net uranium.oftc.net 1120173540 Q * mugwump jupiter.oftc.net uranium.oftc.net 1120173540 Q * micah jupiter.oftc.net uranium.oftc.net 1120173540 Q * lilo jupiter.oftc.net uranium.oftc.net 1120173540 Q * flock jupiter.oftc.net uranium.oftc.net 1120173540 Q * Hollow jupiter.oftc.net uranium.oftc.net 1120173540 Q * locksy jupiter.oftc.net uranium.oftc.net 1120173540 Q * pattieja jupiter.oftc.net uranium.oftc.net 1120173540 Q * anonymousc jupiter.oftc.net uranium.oftc.net 1120173540 Q * monrad jupiter.oftc.net uranium.oftc.net 1120173540 Q * case jupiter.oftc.net uranium.oftc.net 1120173540 Q * BWare jupiter.oftc.net uranium.oftc.net 1120173540 Q * eyck jupiter.oftc.net uranium.oftc.net 1120173540 Q * nox jupiter.oftc.net uranium.oftc.net 1120173540 Q * _are_ jupiter.oftc.net uranium.oftc.net 1120173540 Q * complexho jupiter.oftc.net uranium.oftc.net 1120173540 Q * tomi jupiter.oftc.net uranium.oftc.net 1120173540 Q * alexx jupiter.oftc.net uranium.oftc.net 1120173540 Q * DaCa jupiter.oftc.net uranium.oftc.net 1120173540 Q * _ag_ jupiter.oftc.net uranium.oftc.net 1120173540 Q * rt jupiter.oftc.net uranium.oftc.net 1120173540 Q * matti jupiter.oftc.net uranium.oftc.net 1120173540 Q * _mountie jupiter.oftc.net uranium.oftc.net 1120173540 Q * jkl jupiter.oftc.net uranium.oftc.net 1120173540 Q * Pazzo jupiter.oftc.net uranium.oftc.net 1120173540 Q * SNy jupiter.oftc.net uranium.oftc.net 1120173540 Q * meebey_ jupiter.oftc.net uranium.oftc.net 1120173540 Q * maharaja jupiter.oftc.net uranium.oftc.net 1120173540 Q * Vudumen jupiter.oftc.net uranium.oftc.net 1120173540 Q * jonsmel jupiter.oftc.net uranium.oftc.net 1120173540 Q * Loki|muh jupiter.oftc.net uranium.oftc.net 1120173540 Q * tchan jupiter.oftc.net uranium.oftc.net 1120173540 Q * albeiro jupiter.oftc.net uranium.oftc.net 1120173540 Q * sladen jupiter.oftc.net uranium.oftc.net 1120173540 Q * neofutur jupiter.oftc.net uranium.oftc.net 1120173540 Q * Hunger jupiter.oftc.net uranium.oftc.net 1120173540 Q * EverD jupiter.oftc.net uranium.oftc.net 1120173540 Q * janra jupiter.oftc.net uranium.oftc.net 1120173540 Q * AprilDL jupiter.oftc.net uranium.oftc.net 1120173540 Q * mcp jupiter.oftc.net uranium.oftc.net 1120173540 Q * zimbo jupiter.oftc.net uranium.oftc.net 1120173540 Q * DaPhreak jupiter.oftc.net uranium.oftc.net 1120173540 Q * FaUl jupiter.oftc.net uranium.oftc.net 1120173540 Q * virtuoso jupiter.oftc.net uranium.oftc.net 1120173540 Q * id jupiter.oftc.net uranium.oftc.net 1120173540 Q * SiD3WiNDR jupiter.oftc.net uranium.oftc.net 1120173540 Q * Zoiah jupiter.oftc.net uranium.oftc.net 1120173540 Q * dsoul jupiter.oftc.net uranium.oftc.net 1120173540 Q * aba jupiter.oftc.net uranium.oftc.net 1120173562 J * _mountie ~mountie@24.42.99.232 1120173562 J * jkl ~eric@ool-182cef46.dyn.optonline.net 1120173562 J * neofutur ~neofutur@neofutur.net 1120173562 J * Loki|muh loki@satanix.de 1120173562 J * jonsmel ~jscottorn@209.33.206.3 1120173562 J * Vudumen vudumen@perverz.hu 1120173562 J * maharaja maharaja@ipax.at 1120173562 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1120173562 J * albeiro ~albeiro@albeiro.usercloak.oftc.net 1120173562 J * meebey_ meebey@booster.qnetp.net 1120173562 J * sladen paul@starsky.19inch.net 1120173562 J * SNy ~mfr@bmx-chemnitz.de 1120173562 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it 1120173564 J * Hunger Hunger.hu@levnor.hu 1120173573 J * stupidawy foo@you.wish.you.were.pimp.olicio.us 1120173573 J * Aiken ~james@tooax6-049.dialup.optusnet.com.au 1120173573 J * Doener ~doener@p5487596D.dip.t-dialin.net 1120173573 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1120173573 J * micah micah@micha.hampshire.edu 1120173573 J * lilo ~lilo@lilo.usercloak.oftc.net 1120173573 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1120173573 J * Hollow ~Hollow@home.xnull.de 1120173573 J * locksy ~locksy@mrtg.sisgroup.com.au 1120173573 J * BWare ~bware@office.intouch.net 1120173573 J * case ~case@donpanic.faveve.uni-stuttgart.de 1120173573 J * eyck eyck@81.219.64.71 1120173573 J * monrad ~monrad@213083190130.sonofon.dk 1120173573 J * anonymousc ~anonymous@staff.internode.com.au 1120173573 J * pattieja ~pattieja@adsl-69-153-174-41.dsl.stlsmo.swbell.net 1120173575 M * jonsmel wb all! 1120173576 J * nox ~nox@noxlux.de 1120173587 J * _are_ foobar@dsl-084-056-130-240.arcor-ip.net 1120173587 J * complexho ~mark@funk.gotadsl.co.uk 1120173587 J * tomi ~tomi@pha-84-242-95-4.nat.karneval.cz 1120173587 J * alexx ~alexx@82.225.136.176 1120173587 J * DaCa ~danny@mail.limehouse.org 1120173587 J * _ag_ ag@caladan.roxor.cx 1120173587 J * rt ~rt@195.246.161.1 1120173587 J * matti matti@linux.gentoo.pl 1120173590 J * EverD ~EverD@81.222.149.146 1120173590 J * janra janra@paradox.homeip.net 1120173590 J * AprilDL ~chatzilla@ip68-9-200-247.ri.ri.cox.net 1120173590 J * mcp ~hightower@wolk-project.de 1120173590 J * zimbo ~zimbo@callisto.dom.bonis.de 1120173590 J * Zoiah Zoiah@matryoshka.zoiah.net 1120173590 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1120173590 J * dsoul darksoul@pingu.ii.uj.edu.pl 1120173590 J * id ~id@relax-media.softwarezentrum.de 1120173590 J * virtuoso ~s0t0na@80.253.205.251 1120173590 J * FaUl ~immo@ip88.164.1211G-CUD12K-01.ish.de 1120173590 J * DaPhreak ~phreak@styx.xnull.de 1120173597 J * aba ~aba@eos.turmzimmer.net 1120173651 Q * tomi Quit: Ukončuji 1120173904 J * are|lunch foobar@dsl-084-056-136-105.arcor-ip.net 1120173963 M * AprilDL is Ola Lundqvist here? 1120174054 M * Bertl no, his nick is ola 1120174202 M * complexho Bertl: ok, so from the looks of it, it would be a fairly significant task (but not impossible) to get vserver and GFS playing nicely together. I will speak with one of our coders and see if he's up for looking more closely 1120174253 M * complexho Do you have any other golden tips on the subject? ;) 1120174259 M * Bertl k, IIRC some folks from possibility forge are trying to get GFS or lustre or code to work with linux-vserver 1120174275 M * jonsmel hehe 1120174283 M * jonsmel with not too much luck i might add 1120174287 M * Bertl ah, jonsmel is still here ... 1120174289 M * complexho really? 1120174295 M * complexho Lustre or GFDS? 1120174297 M * Bertl so no need to relay ... 1120174303 M * complexho :) 1120174309 M * complexho *GFS 1120174316 Q * _are_ Ping timeout: 480 seconds 1120174320 M * complexho hi jonsmel! 1120174322 M * jonsmel We want to use lustre but it did not like vservers that well 1120174329 M * jonsmel hi 1120174337 M * complexho did you try GFS at all? 1120174340 M * jonsmel bert can explain more as to why lustre is not working 1120174358 M * jonsmel gfs will work but there are other issues dealing with the cluster stuff 1120174407 M * jonsmel basically we want gfs to serve out the gfs share and not worry about the cluster 1120174418 M * complexho I did spend some time on it a while back, building everything from scratch but couldn't devote much time to it back then... But now I am coming back to take a closer look 1120174423 M * jonsmel but I haven't had much luck getting it to do that 1120174438 M * complexho what distro are you using? 1120174450 M * jonsmel Debian :) 1120174461 M * jonsmel 2.6.12 kernel currently 1120174463 M * jonsmel vs 2.0 1120174467 M * complexho ah ok - iirc Debian actually have GFS packages don't they? 1120174477 M * jonsmel they do but I didn't use them 1120174485 M * complexho right... 1120174487 M * jonsmel cvs checkout has more up to date stuff 1120174506 M * complexho yeah I got much further with that too :) But didn't get it working... 1120174518 M * jonsmel i'm dealing with fencing issues currently 1120174530 M * complexho what hardware are you using? F/C? 1120174544 M * jonsmel but I'm hoping that once I get those worked out it will operate the way I want it to 1120174576 M * complexho but you have managed to mount GFS filesystems (if only from one host) and build a vserver on there yes? 1120174583 M * jonsmel We basically have a high end data storage server and cheap client nodes to access the data 1120174591 M * jonsmel yes 1120174599 M * complexho ah that is encouraging... 1120174605 M * jonsmel that is what we were running on prior to trying to find a better solution 1120174613 M * jonsmel 4 node cluster 1120174657 M * complexho what is the fileserver? Shared SCSI or F/C? 1120174662 M * jonsmel lustre is a much more promising fs cluster if it was only able to play nice with vserver none of this would be an issue 1120174670 M * jonsmel much easier to get working as well 1120174683 M * jonsmel gnbd 1120174687 M * complexho ok 1120174721 M * Bertl jonsmel: well, you have my estimation for required changes to lustre ... 1120174731 M * jonsmel you can use iSCSI which we were going to but gnbd workd the same 1120174751 M * jonsmel bertl: right, it is just a little out of reach for us right now 1120174787 M * Bertl thought so .. but a lot off stuff could be done by interested folks ... 1120174797 M * complexho Bertl: I would be interested in hearing about your thoughts on Lustre too... 1120174806 M * Bertl e.g. lustre preparations and testing ... 1120174807 M * jonsmel so we are trying gfs again to see if I can tweak it actually break it in linux-cluster ppls opinions 1120174808 M * jonsmel hehe 1120174810 M * complexho have not seriously considered it 1120174836 M * complexho (Lustre that is) 1120174843 M * Bertl the prereq for lustre is to get it into the kernel 1120174848 M * jonsmel Lustre works awesome as a network filesystem 1120174868 J * rs ~rs@imhotep.rhapsodyk.net 1120174877 M * Bertl currently lustre consists of a kernel patch to vfs/ext3 and a huge number of out of kernel modules 1120174881 M * Bertl evening rs! 1120174907 M * AprilDL can you suggest a pastebin for the output of my testme.sh 1120174912 M * AprilDL tho its about 6 lines 1120174920 M * Bertl pastebin.com? 1120174924 M * jonsmel I am going to keep checking with them to see if they will integrate/buid in to the kernel those modules 1120174975 M * complexho jonsmel: so did you have to do anything special kernel-wise to get GFS working with vserver? 1120174983 M * AprilDL http://pastebin.com/305691 1120175018 M * Bertl AprilDL: seems you missed the kernel patch :) 1120175021 M * jonsmel I will have to check with our other guy, he did all the development for that part. 1120175037 M * jonsmel I don't think there were major changes 1120175054 M * Bertl AprilDL: well, but it looks fine the debian way *G* 1120175056 M * complexho I would be interested in talking more about it, I have some dev resources available put put on the problem for a while 1120175080 M * jonsmel As soon as I get fenced to start for me again I will know more 1120175085 M * complexho :) 1120175093 M * AprilDL bertL if i put on debian colored glasses perhaps 1120175110 M * complexho I will stay in the channel for a few days, let me know how you get on 1120175123 M * AprilDL i don't think we can blame this on debian since it was my first try at building a kernel 1120175133 M * jonsmel ok, I basically stay in this channel all the time 1120175158 M * Bertl AprilDL: no, debian is not to blame, debian maintainers maybe ... 1120175195 M * complexho ok great! hopefully speak again soon then 1120175222 M * complexho btw where are you (timezone wise) 1120175223 M * Bertl sounds like the cluster front is moving again :) 1120175229 M * complexho yeah :) 1120175229 Q * rs Quit: rs 1120175229 M * jonsmel if you try gfs and want to discuss item pertaining with that join #linux-cluster on freenode 1120175235 Q * jkl Ping timeout: 480 seconds 1120175243 M * jonsmel good, I could use some more help 1120175245 M * jonsmel hehe 1120175251 M * complexho :) 1120175257 M * complexho me too 1120175263 M * jonsmel I am also in that channel all the time 1120175327 M * jonsmel complexho: btw I am MST timezone 1120175331 M * jonsmel u? 1120175337 M * complexho GMT (UK) 1120175348 M * jonsmel ah, you're over there with Bertl 1120175350 M * jonsmel :) 1120175353 M * complexho yeah 1120175362 M * complexho Brighton, England 1120175377 M * jonsmel cool 1120175378 J * rs ~rs@imhotep.rhapsodyk.net 1120175421 M * jonsmel maybe we can conqure the cluster struggles together 1120175427 M * complexho yeah hopefully 1120175427 M * jonsmel ;) 1120175445 M * complexho are you using EVMS or LVM2? 1120175446 M * jonsmel have you setup gfs yet 1120175449 M * jonsmel evms 1120175474 M * complexho ok I tried with LVM2 - and LVM seemed to be the problem. Perhaps this time I will try EVMS instead... 1120175504 M * complexho I managed to get GFS built but never finished cos I couldn't get it all working together and LVM wasn't building properly 1120175505 M * jonsmel evms is very good, we haven't had any issues so far 1120175529 Q * rs Quit: 1120175532 M * jonsmel basically what I have done from rebuilding now is: 1120175549 M * jonsmel take vanilla 2.6.12 kernel with vs-2.0-rc4 patched in 1120175559 M * jonsmel cvs checkout gfs stuff 1120175560 M * complexho but it was only a half-assed attempt back then. Now I have got the cluster up and running with dedicated partitions, but no cluster fs. I have saved a disk set on the SAN for GFS devel 1120175579 M * jonsmel cool 1120175593 M * jonsmel sounds similar to what I have here 1120175626 M * complexho HP MSA1000 - Qlogic 2/8 switch, 7xDL-360 head ends 1120175649 M * Bertl hmm, both lvm2 and evms are solely userspace apps utilizing the kernel dm layer ... 1120175656 M * jonsmel ok, not quite what I have :) 1120175675 M * complexho but it's all the same basically 1120175681 M * jonsmel yep 1120175684 M * complexho but that 2Gb FC is fast ;) 1120175691 M * jonsmel i bet 1120175704 M * jonsmel I don't have any FC equipment though 1120175709 M * complexho I get 450MB vs builds in about 12 secs 1120175723 M * complexho including the yum lag 1120175729 M * jonsmel awesome 1120175743 M * jonsmel eventually I have to get that 1120175751 M * complexho so about 6-8 secs of that is the build, all pkgs coming from a local respository 1120175811 M * complexho yeah it rocks, but without GFS/Lustre I don't get to my original vision ;) 1120175846 M * jonsmel if you are going just straight fs, lustre is fast and easy 1120175869 M * jonsmel with vservers, gfs is the only functional fs that will do it 1120175878 M * jonsmel currently that is :) 1120175900 M * complexho that's what I thought last time I looked. And I must say GFS looks like the puppy whereas Lustre seemd more complex to me 1120175935 M * jonsmel yeah 1120175943 M * complexho so then, shall we get a wiki page going for GFS then? 1120175969 M * jonsmel they have some but they are not kept up that well 1120175985 M * complexho I was thinking rather more vserver-specific 1120175989 M * jonsmel it is also pretty hard to get support from them 1120175997 M * Bertl feel free to add a vserver GFS devel/setup page on the wiki 1120175999 M * jonsmel can do that as well