1104969803 M * Seraph mmh? 1104969809 M * Seraph for apt-get update? 1104969818 M * Loki|muh vapt-get 1104969915 M * pusling in which debianpackage is vapt-get ? 1104969976 M * Seraph util-vserver? 1104969983 M * pusling or where to get it ? 1104969983 M * Seraph at least the alpha package has it 1104970024 M * Loki|muh yeah the alpha-tools rock :) 1104970047 M * Loki|muh http://www.linux-vserver.org/alpha+util-vserver 1104970079 M * Seraph pusling: there's debs for that too in case you wanna try.. 1104970092 M * pusling Seraph: is it in unstable? 1104970094 M * Seraph ndim: is the one on my webspace still in sync with svn or do i need to make a new one? 1104970107 M * Seraph pusling: not yet.. the one in SID is stable tools.. 1104970123 M * Seraph pusling: that's why we're trying to get the alpha tools into packaging shape =) 1104970295 M * pusling Seraph: is the vapt-get in some i386-version? And how do I get it ? 1104970344 M * Seraph the util-vserver alpha debs are currently i386, yes.. 1104970363 M * Seraph they are shipped without guarantee and/or warranty.. 1104970372 M * Seraph but we'd love to have feedback on 'em *g* 1104970373 M * Loki|muh url? 1104970385 M * Seraph http://backend.verfaction.de/~kk/util-vserver/ 1104970396 M * Loki|muh build with dietlibc? 1104970401 M * Loki|muh thx 1104970410 M * Seraph yep it does 1104970438 M * pusling so I shouldn't try it if I only have ssh-acces to the server or ? 1104970458 M * Seraph mmmh, i'd say it's safe.. 1104970474 M * Seraph but as said.. i won't promise 1104970477 M * Loki|muh if the kernel is right and with capabilities 1104970488 M * Loki|muh there is not much that can go wring 1104970494 M * Loki|muh wrong* 1104970504 M * pusling I will give it a go 1104970638 M * pusling hmm.. is there a way to do a command on all vservers? like vapt-get allservers update or ? (no manpage for vapt-get) 1104970658 M * Loki|muh Usage: /usr/sbin/vapt-get * [--all] -- + 1104970678 Q * flock Remote host closed the connection 1104970890 M * pusling hmm... the 'vapt-get --all -- update' doesn't seem to do anything ? 1104970966 M * pusling (is it just me who is stupid?) 1104970992 M * chrish01 is there a way to do iptables within contexts 1104970993 M * Seraph you have setup the new vserver style config files? 1104971001 M * Loki|muh chrish01: no, not yet :( 1104971010 M * chrish01 oohh we need that lol 1104971015 M * Seraph pusling: if you migrate from old vserver (stable) to new alpha tools you need to upgrade the config too 1104971020 M * Loki|muh chrish01: with the new network code there will be hopefully 1104971032 M * Seraph pusling: or it'll not find any vserver 1104971034 M * Loki|muh chrish01: yeah, we wait all for stuff like this ;) 1104971048 M * Seraph chrish01: you can try NGN patches and report back *g* 1104971049 M * chrish01 is there a roadmap? 1104971060 M * Seraph chrish01: NGN also has IPv6 in case you need that ;) 1104971071 M * Loki|muh an context-specific routing 1104971073 M * Loki|muh +d 1104971079 M * chrish01 Seraph, i would love to test that, but unfortunately there isnt any docs on configuration differents for it 1104971087 M * chrish01 omg i would love context based routing 1104971094 M * Loki|muh chrish01: best would be to wait for Bertl_zZ 1104971114 M * pusling Seraph: is there like a howto or something? 1104971115 M * chrish01 hmm...austria...so what time does he get on? 1104971116 M * Seraph yep, he's all the docs you need =) 1104971140 M * Loki|muh chrish01: i think in about 7 hours 1104971154 M * pusling (I just downloaded the package and did dpkg --install ) 1104971158 M * Loki|muh but thats heard to estimate for him *g* 1104971162 M * Seraph pusling: for util-vserver? mmh, you can try http://www-user.tu-chemnitz.de/~ensc/util-vserver/doc/conf/configuration.html if your webbrowser can ignore CSS *ggg* 1104971180 M * Loki|muh Seraph, pusling: its possible to switch styles 1104971182 M * Loki|muh ;) 1104971192 M * Loki|muh i like the alternate grass-css :) 1104971200 M * Seraph pusling: yep converting with legacy tools is planned if i got ndim right =) 1104971252 M * pusling think I will try the webpage in lynx ;) 1104971262 M * Loki|muh hehe 1104971333 A * Seraph likes the boring better 1104971418 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1104971441 M * pusling hmm.. it looks like the way my vservers are configured ? Or am I too tired to look at this ? 1104971478 M * Loki|muh if you have a /vserver/vserverxy/ directory style config 1104971482 M * pusling or is it like /etc/vservers/vservername-directory? 1104971486 M * Loki|muh then it is the new config style 1104971490 M * Loki|muh ups 1104971492 M * Loki|muh yeah 1104971496 M * Loki|muh yours is right 1104971501 M * Loki|muh I forgot a /etc 1104971516 M * pusling well didn't read the guide as directorys. I do have the old way 1104971557 M * pusling should I downgrade the tools until I feel able to change the config-format, or does the new tools work with old format = 1104971561 M * pusling ? 1104971593 M * Loki|muh i would get used to the new config-style and change to it 1104971628 M * Loki|muh since the alpha-tools are better than the legacy tools imho 1104971640 M * Loki|muh but there is no vserver-copy in the alpha-tools yet 1104971658 M * Loki|muh but rsync does the job ;) 1104971747 M * Seraph and you cannot downgrade new config style 1104971754 M * Seraph thus once new config, always new config 1104971767 M * Seraph so you might want to keep a copy for backup restore *g* 1104971859 M * pusling I do not have the time right now to change config-formats - and I think I need to sleep soon ;) But that might be my next project ;) 1104971866 M * Seraph hehe 1104971868 M * Seraph have fun 1104971879 M * Seraph and please report any success/problem back =) 1104971889 M * pusling But until then, should I keep the new tools or should I downgrade? 1104971915 M * Seraph if your vservers are productive, probably downgrade is better 1104971998 M * pusling they are semiproductive... Not end of the world if they stop function, but it is nice to have ;) 1104972031 M * pusling downgraded 1104972041 M * chrish01 anyone got the link to the experimental patches we were talking about earlier 1104972051 M * Loki|muh http://vserver.13thfloor.at/Experimental/NGNET/ 1104972055 M * chrish01 thanks 1104972185 M * Loki|muh chrish01: http://sonderschueler.de/vserver-irclog.txt <-- here is the irclog of the last months, Bertl explained a few time how it is installed and used 1104972191 M * Loki|muh times* 1104972235 M * Loki|muh I think grep NGNET would give you the right starting points 1104972300 M * chrish01 yeah, im trying to make a virtual router like cisco, foundry, alcatel for much less $$ 1104972328 M * Loki|muh ah, nice 1104972379 M * chrish01 but it needs to be able to have its own routing tables, firewall, quagga, etc... 1104972404 M * chrish01 usermodelinux does this great, but its *sooo* slow 1104972668 M * albeiro ? 1104972740 J * tchan_ ~tchan@c-24-13-81-164.client.comcast.net 1104972776 M * chrish01 does bertl take bribes for custom development work? (or anyone else in hear for that matter) 1104972855 Q * tchan Ping timeout: 480 seconds 1104972875 M * Seraph chrish01: ever tried XEN? 1104972899 M * chrish01 not too much. it has similar speed to UML 1104972911 M * chrish01 im looking to start about 100 of these vservers in a box :) 1104972917 M * Seraph it has? their graphs read a bit different 1104972919 M * chrish01 uml and xen dont really do that so well 1104972986 M * Loki|muh i think he does 1104972997 M * Loki|muh lycos is sponsoring some development, too 1104973035 M * chrish01 cause id like to through some bounties up on some specific stuff id like to use 1104973047 M * chrish01 im pretty generous monetary-wise too 1104973190 M * Loki|muh so, time for bed now, here its already 2am 1104973195 M * Loki|muh gn8 1104973206 M * chrish01 night dude 1104974133 Q * chrish01 Quit: Leaving 1104974406 Q * Alecsandro Ping timeout: 480 seconds 1104974426 J * Alecsandro alecsandro@redelocal.sampa.zl.a.la 1104974720 M * ndim Seraph: I havent modified anything in SVN yet. 1104974838 M * ndim Seraph: What am I supposed to have said about "converting with legacy tools is planned"? 1104974894 M * Loki|muh i think converting from legacy config style to alpha-utils-config-style 1104974980 M * ndim I don't think there is a conversion script yet. 1104975191 M * Loki|muh yeah, thats the problem we talked about 1104975596 N * Bertl_zZ Bertl 1104975620 M * Bertl morning folks! 1104975677 M * Loki|muh morning 1104975704 M * Loki|muh you missed some guys which were really interested in NGNET stuff 1104975745 M * Loki|muh I guess you have a backlog ;) 1104975782 M * Bertl yeah, thanks, I (and we all) have ;) 1104975856 M * Bertl the virtual router idea is new ... at least ... 1104976064 M * Bertl Loki|muh: so how is linux-vserver working for you? 1104976064 Q * infowolfe Read error: Connection reset by peer 1104976088 M * Loki|muh i'm very happy with it 1104976118 M * Bertl any 'important' missing features? (besides maybe ipv6) 1104976129 M * Loki|muh I think I will set up a test machine here as I wanna play a little with the ngnet stuff 1104976173 M * Bertl yes, I guess I will continue to work on that the next few week ... 1104976217 M * Loki|muh uhm, the only thing I heard, is that setfacl is not working, but i haven't took a look at it yet 1104976278 M * Bertl that's access control list stuff, right? 1104976304 M * Loki|muh yeah 1104976326 M * Loki|muh are there special capabilities needed or should it work? 1104976330 M * Bertl do you have that installed/active now? 1104976365 M * Bertl yes, I guess some caps are required ... but I never checked ... so an strace of a failing setfacl would be a good start/hint ;) 1104976437 M * Loki|muh one moment 1104976452 J * infowolfe ~infowolfe@66-230-110-131-cdsl-rb1.nwc.acsalaska.net 1104976472 M * Bertl wb infowolfe! 1104976487 M * infowolfe thx 1104976493 A * infowolfe mutters about the hal daemon 1104976713 N * tchan_ tchan 1104976742 M * Loki|muh Bertl: i think these lines are of interest for you: 1104976743 M * Loki|muh |>stat("test", {st_mode=S_IFREG|0644, st_size=0, ...}) = 0 1104976744 M * Loki|muh |>getxattr("test", "system.posix_acl_access", 0x7fbffff830, 132) = -1 EOPNOTSUPP (Operation not supported) 1104976747 M * Loki|muh |>setxattr("test", "system.posix_acl_access", 0x50b420, 44, ) = -1 EOPNOTSUPP (Operation not supported) 1104976762 M * Bertl hmm ... sec 1104976897 M * Bertl you sure that is compiled into the kernel or the module is loaded? 1104976953 M * Loki|muh CONFIG_EXT3_FS_POSIX_ACL=y 1104976974 M * Bertl (because I see no check except for a search for an xattr handler/name ... which doesn't match anything ... which results in EOPNOTSUPP 1104977026 M * Bertl could it be that the setfacl tool isn't recent enough for that kernel? and are we speaking of 2.6.10 at all? ;) 1104977087 M * Loki|muh no 2.6.9 1104977092 M * Loki|muh hmmm 1104977113 M * Bertl well, I guess 2.6.9 should be similar to 2.6.10 (which I'm looking at) 1104977135 Q * infowolfe Read error: Connection reset by peer 1104977146 J * infowolfe ~infowolfe@66-230-110-131-cdsl-rb1.nwc.acsalaska.net 1104977203 M * Loki|muh installed setfacl is 2.2.23 1104977237 M * Bertl * Mount With ACL Capabilities 1104977238 M * Loki|muh just searching for a homepage or a changelog or something like that 1104977244 M * Bertl does that apply? 1104977293 M * Bertl example: mount -o loop,encryption=aes,acl ./disk-aes /fs 1104977326 M * Bertl another example for setfacl: setfacl -R -m d:u:donkey:rwx,d:u:chirico:rwx,d:u:bozo2:rwx /fs 1104977352 M * Loki|muh hmmm, could be a possibility 1104977383 M * Bertl note: never played with acls, so I am just guessing ... 1104977478 M * Loki|muh uh, you are right 1104977500 M * Loki|muh thanks :) 1104977505 M * Bertl does it work now? 1104977523 M * Loki|muh yes, it does 1104977525 M * Bertl would you care to give me an example how to use it? 1104977576 M * Loki|muh thats easy (but i don't know very much about it yet :( ) 1104977593 M * Loki|muh 1.install package acl (in debian) 1104977599 M * Loki|muh 2. mount with option acl 1104977606 M * Bertl what, for example, does the d:u:donkey:rwx do? 1104977618 M * Bertl define the property? 1104977632 M * ndim Hmm. Is "lsxid" supposed to do something different from printing !!ERR!! all the time? 1104977658 M * Loki|muh d <- default : u<-user : donkey <- username : rwx 1104977660 M * Bertl ndim: yep, try it on an xid tagged filesystem 1104977702 M * ndim How may I get an xid tagged FS? 1104977727 M * Bertl by mounting a filesystem with tagxid 1104977736 M * Bertl (tagxid is the option) 1104977753 M * Bertl but to see that you _can_ get something different, try 1104977758 M * Bertl lsxid /dev/pts 1104977779 M * Bertl (pts is xid tagged by default) 1104977813 M * Bertl Loki|muh: hmm, the default is something like a namespace? 1104977814 M * ndim Ah good. I was fearing I had b0rked something. 1104977830 M * Bertl Loki|muh: or scope? 1104977849 M * ndim IIRC, the default was the permissions newly created stuff will get. 1104977855 M * Loki|muh default only applies to directories 1104977859 M * Loki|muh yeah 1104977862 M * Loki|muh thats it 1104977889 M * Bertl aha, okay, and d and u are just shortcuts, right? 1104977900 M * Loki|muh jep 1104977920 M * Bertl so I could also use default:user:donkey:rwx, correct? 1104977956 M * Bertl is there something like a 'group' too? 1104978082 M * Loki|muh yes 1104978084 M * Loki|muh and mask 1104978089 M * Loki|muh and other 1104978098 M * Bertl ah, what's that used for? 1104978194 M * Loki|muh to set the permissions for a group and others ;) 1104978211 M * Loki|muh the thing with the masks I didn't get yet 1104978241 M * Bertl okay ... 1104978356 M * Loki|muh I hope I will have some time for that the next days 1104978387 M * Bertl found a nice document: http://www.suse.de/~agruen/acl/chapter/fs_acl-en.pdf 1104978484 M * Loki|muh http://www.suse.de/~agruen/acl/chapter/fs_acl-de.pdf ^^ 1104978598 M * Loki|muh looks fine, easy stuff for reading in my bed *g* 1104978610 M * Loki|muh nighty night finally ;) 1104978618 M * Bertl night, and thanks! 1104978697 M * Loki|muh i have to thank ;) 1104978884 M * Bertl you're welcome! 1104979019 Q * infowolfe Quit: Leaving 1104979506 Q * Alecsandro Ping timeout: 480 seconds 1104985415 Q * we2by Ping timeout: 480 seconds 1104985456 J * we2by ~we2by@dc5146d009.adsl.wanadoo.nl 1104986716 J * nox- ~nox@213.39.207.82 1104987046 Q * nox Ping timeout: 480 seconds 1104987075 N * nox- nox 1104989184 J * infowolfe ~infowolfe@209-112-197-222-cdsl-rb1.nwc.acsalaska.net 1104989256 J * infowolfe_ ~infowolfe@209-112-197-222-cdsl-rb1.nwc.acsalaska.net 1104989257 Q * infowolfe Quit: 1104993945 J * chrish01 ~chrish01@69.90.131.10 1104993967 M * Bertl welcome chrish01! 1104993982 M * chrish01 hey bertl...i came just to see you :) 1104993989 M * chrish01 i had some questions about your NGN stuff 1104993998 M * Bertl yeah? let's hear! 1104994020 M * chrish01 well, you know how alcatel, cisco, foundry have virtual routers? 1104994035 M * chrish01 i would like to do the same thing with vserver (cause freevps is very unstable IMO). 1104994048 M * Bertl hmm, is it? 1104994050 M * eyck hmm, what about this allegro thingie? 1104994063 M * chrish01 i am currently doing this with UML, but that is very slow 1104994076 M * chrish01 eyck: whats that? 1104994086 M * eyck virtual router 1104994102 M * Bertl so you basically want/need iptables/rt/if virtualization 1104994111 M * chrish01 yes 1104994124 M * Bertl eyck: any links? 1104994150 M * chrish01 i need private routing tables and private iptables per context. it would also be great if i could have over-lapping ip-space in different contexts 1104994172 M * chrish01 at least overlapping ip space in routing tables if not in the machine itself 1104994201 M * Bertl hmm, you already know what ngnet does and how it works (atm)? 1104994203 M * chrish01 oh, and im willing to pay $$ too ;) 1104994212 M * chrish01 not really 1104994233 M * mugwump sssh Bertl, he'll give you a cheque soon 1104994250 M * Bertl yeah, I should wait for the cheque, yes? 1104994251 M * chrish01 i would probably support vserver monetarily either way 1104994257 M * chrish01 hehe 1104994267 M * mugwump that sounds like a really good idea chrish01, maybe Bertl could be convinced to put that stuff in ngn ;-) 1104994269 M * chrish01 i paid a few uml developers to do some stuff for me as well 1104994277 M * Bertl just kidding ;) 1104994309 M * Bertl okay, let me explain what the ngnet does, or if you have a machine at hand, you can check it out yourself ... 1104994316 M * chrish01 what would be rad is if we could get 100 - 1024 contexts per box (since they only do routing). then we could sell a competing product to foundry for like 100th of hte price 1104994328 M * chrish01 bertl, i have vmware :) 1104994338 M * Bertl okay, will do I guess ;) 1104994383 M * Bertl get 2.6.10, vs1.9.3.17 (to be uploaded in a few minutes) 1104994417 M * chrish01 ok...i got a box at work with 2.6.9 and util-vserver debian package from mailinglist 1104994423 M * chrish01 ill just update it with a new kernel 1104994439 M * Bertl the tools are meaningless for this purpose atm ... 1104994444 M * chrish01 ah ok 1104994459 M * Bertl just chcontext is required ... 1104994469 M * chrish01 Bertl: you can count on me to do quite a bit of documentation if this works out well 1104994479 M * Bertl excellent! 1104994521 M * chrish01 ive been doing the same for the mono project as well. although i contribute code to them. 1104994560 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.10-vs1.9.3.17.diff 1104994591 M * chrish01 grrr no debian source package for 2.6.10 yet 1104994596 M * chrish01 your patch for vanilla? 1104994604 M * Bertl yep 1104994609 M * mugwump mono? as in m0n0wall or GNU mono? 1104994624 M * chrish01 as in ximian mono 1104994629 M * chrish01 i work a bit with m0n0wall too 1104994668 M * mugwump stay away from debian kernel sources, is the experience of this channel... 1104994702 M * mugwump for latest version stuff anyway 1104994724 M * chrish01 i like it cause it has the cramfs patch for initrd 1104994751 M * chrish01 Bertl: check out www.mosaix.net for what my company does 1104994786 M * mugwump People actually bother with initrd? :) 1104994804 M * mugwump I mean, for dists its useful of course 1104994834 M * chrish01 well, we have half our servers running legacy SCSI, and have running SATA 1104994845 M * chrish01 its nice so we dont over complicate kernels 1104994852 M * chrish01 and we can still deb package stuff 1104994912 M * chrish01 Bertl: ok, i have source patched 1104995083 M * Bertl k, there's another patch, sec, I'm test compiling it right now (needed some adaptations to .17) 1104995106 M * Bertl but there is something you can do in the meantime ... 1104995114 M * chrish01 separate patch, or reverse other and then this one 1104995124 M * Bertl no, it's ontop of the .17 1104995135 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/kernel-ng8.config 1104995161 M * Bertl try to match your kernel config (regarding networking) with this config 1104995184 M * chrish01 ok 1104995206 M * Bertl because some things like MULTIPLE_TABLES and such are not implemented yet 1104995295 M * chrish01 can i just modify that config for what else i need? 1104995367 M * Bertl yes, that would be fine ... 1104995380 M * Bertl (or just copy over the network section to your config) 1104996058 J * Raj_Online ~amit@203.124.158.219 1104996242 M * chrish01 Bertl: configured 1104996274 M * Bertl k, my test compile finished too ... 1104996290 M * chrish01 k, ill make-kpkg 1104996299 M * Bertl sec, you are missing a patch ;) 1104996306 M * chrish01 oh ok 1104996441 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-vs1.9.3.17-ng8.7.diff 1104996461 M * Bertl okay, this will have a few new config options, please enable them all 1104996534 M * chrish01 ok 1104996578 M * Bertl especially 1104996579 M * Bertl CONFIG_IP_NF_MATCH_VNET=y 1104996585 Q * infowolfe_ Ping timeout: 480 seconds 1104996587 M * Bertl CONFIG_IP_NF_VNET=y 1104996587 M * Bertl CONFIG_IP_NF_TARGET_VNET=y 1104996604 M * Bertl CONFIG_VNET=y 1104996621 P * oberonc Leaving 1104996624 M * chrish01 mmm...make oldconfig works great for this :) 1104996648 M * chrish01 k, patched and configured 1104996750 M * Bertl okay, you need two other things inside your image ... 1104996764 M * Bertl a patched iptables version 1104996794 M * eyck aah, I meant argante not allegro: http://agt.rogue-squirrel.org/AgtDoc.txt 1104996809 M * eyck it looks like it's abandoned though... 1104996839 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/iptables-1.2.9-vnet.diff or the one for 1.2.11 1104996852 M * Bertl and the vnet tools: http://vserver.13thfloor.at/Experimental/NGNET/vnet-0.02.tar.bz2 1104996855 M * chrish01 is that a kernel patch or userspace app patch 1104996864 M * Bertl userspace, iptables 1104996875 M * chrish01 ah, so i can build my kernel right now then 1104996880 M * Bertl yep 1104997388 M * chrish01 Bertl: was that patch against vanilla iptables (i ask cause i thought you used debian) 1104997419 M * Bertl yes, it's against vanilla, I avoid debian if possible ... 1104997440 M * Bertl (too many things not done yet, but it will get better this year, I hope) 1104997454 M * chrish01 ah 1104997470 M * Bertl at least we managed to get in contact with the maintainers ... 1104997482 M * chrish01 ill make deb's for most of this stuff anyway so i can deploy easier 1104997488 M * chrish01 oh nice 1104997768 M * chrish01 what do i need to patch this? -Np1 and Np0 dont work 1104997773 M * chrish01 iptables that is 1104997776 M * chrish01 1.2.11 1104997798 M * chrish01 nm...Np2 works :) 1104998268 J * infowolfe ~infowolfe@209-112-197-222-cdsl-rb1.nwc.acsalaska.net 1104998275 Q * infowolfe Quit: 1104998284 J * infowolfe ~infowolfe@209-112-197-222-cdsl-rb1.nwc.acsalaska.net 1104998628 M * Bertl chrish01: let me know when you are ready ... 1104998689 M * _are_ moin 1104998714 M * Bertl morning _are_! 1104998731 M * _are_ I remember something about me saying the opterons are less tested than Xeons. 1104998765 M * _are_ well, the Opterons at least reboot when they are told to, the Xeon hangs with a dead keybord and writes on screen it is about to reboot. 1104998834 M * Bertl try specifying reboot=b on the kernel command line 1104998851 M * chrish01 Bertl: make-kpkg is still running...ill let you know when its ready 1104998862 M * _are_ will try, the boix hangs atm and is 10km apart, will pay it a visit later today 1104998937 M * Bertl chrish01: hope this isn't messing with the .config ;) 1104998969 M * chrish01 shouldnt be 1104999670 M * chrish01 Bertl: you got an assumption of about how many vserver contexts i could get running in a single box with 2.8ghz p4? (just running iptables and quagga basically) 1104999731 M * Bertl no idea what quagga is ... but we reached about 2500 processes on a dual PIII/2GHz 1104999745 M * chrish01 quagga is the newer zebra 1104999750 M * chrish01 zebra is a routing daemon 1104999762 M * chrish01 Rip/OSPF/BGP and such 1104999774 M * Bertl ah yes, I know zebra ... and the context limit is theoretically at 2k 1104999790 M * chrish01 nice 1104999792 M * Bertl (not much overhead in the context itself) 1104999806 M * chrish01 thats what made me want to check out vserver :) 1104999816 M * Bertl might be that the iptables/rt/whatever virtualization will eat up some more memory ... 1104999848 M * chrish01 its so little though...like only 350 bytes per connection for conn_track if i remember right 1104999927 J * tchan_ ~tchan@24.13.81.164 1104999981 Q * tchan_ Quit: 1104999993 J * tchan_ ~tchan@c-24-13-81-164.client.comcast.net 1105000016 J * jsambrook ~jsambrook@host-62-69-64-93.bsve.net 1105000028 M * Bertl welcome jsambrook! 1105000114 Q * tchan Killed (NickServ command used by tchan_) 1105000119 N * tchan_ tchan 1105000129 M * Bertl welcome tchan! 1105000305 M * chrish01 Bertl: could you give me a little info on how vnet prog works? 1105000324 M * chrish01 i would like to start documenting a bit 1105000378 M * Bertl yes, the vnet tool itself just calls a new syscall command 1105000403 M * Bertl which in turn creates a vnet device ... similar to a vlan device 1105000445 M * chrish01 can i still use vlans? cause our network uses vlans quite heavily 1105000469 M * Bertl yes, it should not affect vlans ... 1105001801 M * chrish01 time to reboot and hope for the best :) 1105001849 M * Bertl ;) 1105002237 M * chrish01 well damn, it didnt come back up. ill have to setup a local machine i guess 1105002317 M * Bertl hmm, thought you where using vmware? 1105002335 M * chrish01 i was going to, but then i decided to just use my remote vserver devel box at work 1105002393 M * Bertl do you have a machine in the same physical lan? 1105002487 M * chrish01 not really, just my ppc laptop and my main machine 1105002509 M * chrish01 im at home...its around 1am so im just trying to get something quickly working 1105002542 M * Bertl well, it's 10am here, and I'm off to bed soon ... ;) 1105002562 M * chrish01 hahaha 1105002910 Q * infowolfe Ping timeout: 480 seconds 1105003034 M * sannes :) 1105003155 M * sannes Bertl : tried to upgrade my production server last night (for me) .. it totally froze after 15-20 minutes .. a real bummer 1105003166 M * Bertl what kernel patch? 1105003212 M * sannes 1.9.3.16.3 ? 1105003242 M * Bertl hmm ... shouldn't freeze ... but maybe we are still missing something ... 1105003242 M * sannes but I havn't tested it with a vanilla kernel or anything like that.. so doesn't have to be vserver related at all.. 1105003266 M * Bertl with preemption enabled? 1105003276 M * sannes no, with it off actually.. 1105003297 M * Bertl good, magic-sysrq enabled? 1105003308 M * sannes yip, didn't respond to that either.. 1105003337 M * Bertl hmm, that means it was hanging with irqs off ... 1105003344 M * sannes hm.. just going to double check the sysrq was enabled.. 1105003470 M * Bertl chrish01: okay, if you get it running (either in vmware or on a local machine) try the http://vserver.13thfloor.at/Experimental/NGNET/vnet3_setup.sh and after this script has been run, enter context 100 with 'chcontext --ctx 100 /bin/bash' and contact (e.g. ping) something in 192.168.0.x (which should be connected to eth0) 1105003505 M * sannes ugh, it wasn't .. figures.. I'll try again to night.. without SMP and see what happens.. 1105003507 M * Bertl the two contexts will assume ips 192.168.0.2 and 192.168.0.3 you can modify that in the script if it doesn't work for you ... 1105003541 M * chrish01 Bertl: thanks a million 1105003556 M * Bertl you're welcome! 1105003584 M * Bertl more to come ... for now ... have a good whatever everyone! 1105003600 M * Bertl and thanks for testing linux-vserver stuff ... 1105003609 N * Bertl Bertl_zZ 1105003614 M * chrish01 later bertl 1105003786 M * sannes night Bertl :) 1105004484 J * rs rs@ice.aspic.com 1105004489 M * rs hi there 1105004576 M * chrish01 night guys, thanks for all the help 1105004674 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105004693 Q * chrish01 Quit: chrish01 1105006357 N * Doener|gone Doener 1105006363 M * Doener morning! 1105006553 M * Loki|muh moin 1105007353 J * mboman ~michael@cm48.sigma230.maxonline.com.sg 1105008953 J * psychops psychops@d221250.upc-d.chello.nl 1105009384 M * psychops is there some documentation on how to get an xserver running in a vserver ? 1105009399 M * psychops on 1.9.3 latest dev tools kernel 2.6.9 (gentoo) 1105009516 M * Loki|muh hmmm, the server itself is not the problem, I can open a xterm on a vserver over X-Forwarding 1105009532 M * Loki|muh but for a complete x-window-system... hmmm... never heard of that 1105009663 M * psychops i now get Failed to ser IOPL for I/O when i do startx 1105010854 M * psychops what is the correct syntax in bcapabilities for adding a CAP ? 1105010860 M * psychops CAP_SYS_RAWIO ? 1105011055 M * Doener have a look at the source referenced on the flower page... IIRC it's just SYS_RAWIO 1105011100 M * _are_ I have seen some post/page that claimed it to be possible to run an X-Server within a vserver 1105011165 M * psychops is it possible to give a vserver all caps ? 1105011166 M * _are_ http://www.linux-vserver.org/Linux-Vserver+FAQ check topic G 5 1105011175 M * psychops yeah i was reading that 1105011183 M * _are_ psychops: yes, it is, but no use in a vserver then :-> 1105011237 M * psychops How, just to test if its a caps problem 1105011244 M * _are_ sample from my config for capabilities: 1105011245 M * _are_ # less /etc/vservers/misc/bcapabilities 1105011245 M * _are_ NET_RAW 1105011354 M * _are_ well, all capabilities -> add al keywords to the bcapabilities file. 1105011386 M * psychops :D 1105011401 M * _are_ next question is where you get all keywords? ;) 1105011570 M * psychops na i have those from flower page 1105011657 M * _are_ :-) 1105011878 M * psychops ./dev/mem error now 1105012053 N * Doener Doener|gone 1105012110 M * psychops kewl works now :) 1105012162 M * Loki|muh could you write a howto in the wiki? ;) 1105012176 M * sannes but it is insecure iirc .. 1105012188 M * psychops well lemme test more first, cuz atm i have just added all caps and /dev entrys 1105012197 M * Loki|muh hehe ;) 1105012201 M * psychops also /dev/mem and i am guessing thats not good :) 1105012209 M * psychops also gonna upgrade to latest kernel also 1105012230 M * psychops but ill write a wiki when i have that done :D 1105012239 M * Loki|muh :) 1105013081 Q * psychops Read error: Connection reset by peer 1105016106 Q * sannes Read error: Connection reset by peer 1105016330 N * Seraph dominance 1105017344 Q * flock Ping timeout: 480 seconds 1105018541 J * Alecsandro alecsandro@redelocal.sampa.zl.a.la 1105019306 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1105019415 M * ndim Does anybody know what "vconfig" may be? alpha util-vserver checks for its presence at configure time, and I can't see it being used, or available from anywhere. 1105021699 Q * _are_ Ping timeout: 480 seconds 1105021779 J * _are_ ~are@84.56.141.60 1105023161 M * ndim ensc: Are you here? I have a set of patches for alpha util-vserver at http://vserver.lauft.net/util-vserver/patches--hun/ :) 1105023306 J * sannes ~ace@home.skarby.no 1105023342 J * monrad ~monrad@213083190130.sonofon.dk 1105024133 Q * infowolfe Remote host closed the connection 1105024158 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105024178 Q * albeiro Remote host closed the connection 1105024211 Q * infowolfe Remote host closed the connection 1105024268 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105025020 Q * Raj_Online Quit: using sirc version 2.211+KSIRC/1.2.4 1105025991 J * pusling_ ~pusling@195.215.29.124 1105026004 Q * pusling_ Quit: 1105027728 J * DuckMaster ~Duck@dyn-83-157-130-205.ppp.tiscali.fr 1105028155 Q * duckx Ping timeout: 480 seconds 1105028586 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1105028628 N * Bertl_zZ Bertl 1105028646 M * Bertl ndim: send an email to the ml, cc to enrico ... 1105028650 M * Bertl morning folks! 1105028731 M * daniel_hozac ndim: vconfig is for VLANs. 1105028948 M * we2by Bertl, I have the ld SV spawning to fast problem 1105029028 M * Bertl hmm, and? what's in your inittab(s)? 1105029074 M * we2by SV:123456:respawn:/command/svscanboot 1105029096 M * Bertl excellent, comment it out ... 1105029110 M * we2by if I do so, then would my vservers startup at boot? 1105029136 M * Bertl is this on the host? 1105029144 M * we2by yea 1105029163 M * Bertl svscanboot is daemontools, right? 1105029173 M * we2by ye 1105029177 M * we2by it is version 0.30 1105029180 M * we2by no 1105029188 M * we2by thet is the vserver utilities 1105029200 M * Bertl hope so ;) 1105029207 M * we2by :\ 1105029236 M * Bertl well, basically your 'error' just says that your /command/svscanboot is broken (not working as intended?) ... 1105029256 M * Bertl if that happens on vserver start/stop, it's not the one on the host reporting that 1105029275 M * Bertl it's the entry _inside_ the vserver instead. 1105029294 M * Bertl (probably has a similar entry in each vserver) 1105029377 M * we2by I don;t think so 1105029415 Q * matti Ping timeout: 480 seconds 1105029465 J * matti matti@linux.gentoo.pl 1105029701 M * sannes :) 1105029726 M * Bertl we2by: well, then you have to ask D.J.B what's wrong ;) 1105029743 M * sannes What has changed between 1.9.3.16.3 -> 1.9.3.17 ? 1105029769 M * Bertl the version number ... 1105029775 M * we2by :P 1105029790 M * sannes heh.. ah, figures :) 1105029814 M * Bertl we2by: when do you get that error? 1105029823 M * we2by Bertl, a long time ago 1105029830 M * Bertl (or message) 1105029837 M * we2by I think after a started with creating vservers and reboot 1105029845 M * we2by I have all my vserver to start at boot 1105029854 M * sannes hm, got an odd error message for one of the vservers chroot: Permission denied. 1105029861 M * sannes ugh 1105029876 M * sannes think I know what it is now that I think about it.. 1105029887 M * sannes but why would it still work with 2.4 kernel .. hmz 1105029913 M * Bertl we2by: what I mean is: when did it happen? 1105029923 M * Bertl s/did/does/ 1105029931 M * we2by after startup 1105029931 M * sannes With 1.9.3 should /vservers still be chmod 0 and chattr +I /vservers (or what it was again) 1105029948 M * Bertl nope 1105029950 M * we2by why chattr +i 1105029962 M * we2by what would make the dir not-writeable for all users including root 1105030011 M * we2by ha, that;s a good trick to make n00bs work around to make change in ur system files 1105030014 M * we2by :D 1105030041 M * sannes ugh, can't remeber what it was again, there was a flag that were a barrier or something.. 1105030060 M * sannes but that isn't needed for 1.9.3 anymore because of namespaces so all is well :) 1105030087 M * Bertl yep, the so called 'barrier' flag 1105030131 M * sannes I just love 1.9.3 so much.. if 2.6 was just a little bit less harsh to all the hardware I have.. 1105030170 M * Bertl we2by, so you get the message about the svscanboot respawning too fast on the host after you boot? 1105030170 M * sannes one of the computers I were to test 2.6 on just kills the interfaces, think it is a SMP + ISA-bus thing .. heh 1105030172 Q * Doener|gone Quit: changing servers 1105030179 M * we2by Bertl, yea 1105030194 M * sannes and the production server just freezes so I don't know what to think about that.. 1105030205 M * Bertl we2by: and did you check the logs? 1105030207 M * sannes guess 2.6.10 is still a bit early.. heh 1105030210 J * Doener|gone doener@193.24.208.125 1105030228 M * Bertl sannes: to me it looks like you are doing something wrong ... 1105030240 M * sannes Bertl : when everything freezes? 1105030254 M * sannes Bertl : randomly between 10 minutes and 20 minutes ? 1105030259 M * Bertl yes 1105030283 M * Bertl for example ACPI and powersaving/console blanking comes to mind ... 1105030331 M * sannes hm, .. but I was typing at the screen.. was updating some configuration.. 1105030392 M * Bertl k, what about testing an unpatched kernel on that machine? 1105030424 M * sannes hm.. ACPI is enabled on that computer.. not that it is ever needed.. 1105030443 M * Bertl it is probably needed to get the irq/boot stuff right ... 1105030462 M * ndim daniel_hozac: Thx, I've just found out about vconfig myself. Now I'm going to revert my patch to remove it :) 1105030468 M * Bertl sannes: could you upload the dmesg output after a bootup somewhere? 1105030477 M * sannes sure thing 1105030499 M * sannes but, Power Management isn't really needed is it.. on a server? 1105030519 M * sannes Bertl : the one from 2.4 or the one from 2.6 ? 1105030761 M * Bertl the 2.6 one 1105030910 M * sannes http://www.sannes.org/panics/dmesg.v26 1105030948 M * Bertl hmm, this is with linux-vserver? 1105031031 M * sannes yes 1105031047 M * Bertl then you should switch to a 2/2 split and disable highmem 1105031058 M * Bertl (2GB with highmem is suicide) 1105031087 M * sannes how will userspace react to that? 1105031101 M * Bertl it will thank you many times! ;) 1105031125 M * Bertl also avoid the eepro100, use e100 instead 1105031145 M * sannes well, actually I have troubles with e100 and e1000 .. 1105031155 M * Bertl hmm, in what way? 1105031173 M * sannes well, just doesn't work (tm) 1105031190 M * Bertl ahem, the e100 doesn't work with your e100 card? 1105031193 M * sannes eepro100 doesn't work very reliably either.. have to unplug and the plug the server again for it to work.. 1105031211 M * Bertl well, sounds like broken hardware to me?! 1105031225 M * Bertl you are sure you're not beating a dead horse? 1105031230 M * sannes well, it works excellent aslong as I don't enable SMP 1105031252 M * Bertl another hint that the hardware is broken ... 1105031272 M * Bertl (or at least the bios is doing bad things) 1105031278 M * sannes I think it gets confused on which of the cards it should be using.. 1105031299 M * Bertl what did you configure in the bios? regarding MPS and ACPI? 1105031304 M * sannes hm, there is a BIOS update for that BIOS, but it only says it should be applied if win2k bluescreens.. heh 1105031341 M * sannes I don't really have two processors (altough it is supported by the mobo) is hyperthreading.. 1105031375 M * Bertl update the bios to the latest version (maybe even beta version) configure MPS to 1.4, enable pnp aware OS 1105031384 M * sannes could you explain to me what the 2G/2G split is? 1105031392 M * Bertl yes sure ... 1105031404 M * Bertl there are 4GB address space on x86 1105031410 M * sannes yip 1105031422 M * Bertl and usually the kernel gets 1GB and userspace 3GB 1105031438 M * sannes yip 1105031442 M * Bertl this means that you can have slighly below 1GB lowmemory 1105031458 M * Bertl eerything above that has to be 'mapped' into that range 1105031463 M * sannes paging :) 1105031466 M * sannes yip 1105031481 M * Bertl well, not paging but highmem 1105031499 M * Bertl the kernel does not get paged (at least not the linux one) 1105031531 M * sannes the mapping I was thinking of, is paging.. atleast that is what we called it .. 1105031551 A * sannes is the proud programmer of his own OS.. heh 1105031569 M * Bertl k, the 2/2 split makes this 2GB kernel and 2GB userspace 1105031596 M * Bertl which basically obsoletes the overhead of highmem 1105031629 M * Bertl sorry, but I have to leave now .. will be back later ... 1105031636 N * Bertl Bertl_oO 1105031687 M * sannes ok, so the split changes where the userspace starts in the addressspace .. but then all processes must be changed right? 1105031725 M * sannes unless ELF actually supports relocation.... hehe 1105031851 M * sannes so to shorten this a bit down.. I should upgrade BIOS, change to 2G/2G split, test to see if e100 and e1000 works with those changes, and if it doesn't try to turn of power management... ? 1105032868 J * psychops psychops@d221250.upc-d.chello.nl 1105032882 M * psychops is it possible to get alsa working within a vserver ? 1105033621 M * sladen psychops: it doesn't really work like that 1105033625 M * psychops ? 1105033646 M * psychops if i get it to work on root server should it also work in a vserver then ? 1105033647 M * sladen psychops: but you can copy the /dev/dsp etc devices to the vserver and it will be able to access the sound device 1105033664 M * psychops hmm okay :D 1105033716 M * sladen I'll rephrase. It's not a question of ''getting it to work'', it a question of which devices you want a vserver to be able to access. If you want it have access to, say a serial port, or a sound card; then you need to create (copy) the devices into that vserver. 1105033738 M * psychops okay get it 1105034482 M * ndim Hmm. A root user inside a vserver cannot call mknod himself? 1105034553 M * daniel_hozac ndim: not without CAP_MKNOD. 1105034802 J * chrish01 ~chrish01@69.90.131.10 1105034848 Q * rs Quit: leaving 1105035397 M * _are_ why use dietlibc, btw? what are the benefits for util-vserver? 1105036083 M * chrish01 isnt dietlibc much smaller in size and footprint? which makes it ideal for lightweight setups 1105036307 M * _are_ i don't care about the lightweight, I have enough ram and diskspace 1105037058 N * Doener|gone Doener 1105037213 M * Doener chrish01: that's only valid if you use shared libs, and IIRC util-vserver doesn't do so and i've heard it is not stable for dietlibc on x86_64 1105037235 M * chrish01 ah...im off to work 1105037239 M * chrish01 cya ya guys later 1105037243 M * Doener later chrish01 1105037258 M * _are_ ah, yes, now I know there had been a reason not to use dietlibc on the opterons. 1105037258 M * chrish01 Bertl_oO: if you are here later, ill go over some of that NGN stuff with you 1105037273 Q * chrish01 Quit: off to work 1105037273 M * _are_ but this time it is on Xeons, so I guess I can use it safely? 1105037344 M * psychops Doener, prolly a stupid question but what us ng8.7 ? 1105037344 M * Doener _are_: it's only unstable with shared libs, the reason not to use it on x86_64 with util-vserver 0.30.196 is that it misses a syscall (sys_nice IIRC) 1105037394 M * Doener psychops: next generation networking. the new networking code for linux-vserver. but it's still under development and not suitable for production use 1105037432 M * ndim _are_: glibc has issues working in chroot wwhich dietlibc doesnt have. 1105037478 M * _are_ so far it seems to work in my vservers. 1105037491 M * Doener "glibc's NSS functions are very unreliable in chroot() environments, so you have to expect problems there." 1105037492 M * _are_ and users come mixed from ldap and /etc/passwd 1105037506 M * Doener from the alpha tools' configure.ac 1105037533 M * _are_ yes, seen that 1105037536 M * psychops what are the differences with "old" code ? 1105037695 M * Doener psychops: a lot, the whole concept behind the network virtualization has changed. also provides virtual routing tables and ipv6 AFAIK, but i'm not really up2date with ng development 1105037743 M * psychops ill try it out then :D 1105037827 M * Doener you'll probably need someone to tell you how it works ;) the tools don't support it yet 1105037858 M * psychops hmm, okay :D 1105038195 M * ndim Argh. I am so stupid. 1105038888 J * chrish01 ~chrish01@69.90.131.10 1105038980 Q * sannes Read error: Connection reset by peer 1105039439 J * shuri ~shuri@dsl.speedline209.226.electronicbox.net 1105039555 M * we2by is there any php script to do the admin job? 1105039580 M * _are_ which admin job? of ngn or vserver in general? 1105039610 M * we2by I want to control my vservers via the my webbrowser 1105039613 M * we2by make a new vserver 1105039617 M * we2by stop start 1105039628 M * we2by or even open up a shell using java if posible 1105039643 M * we2by give me the status 1105039649 M * shuri useless 1105039720 M * shuri some project was born for that 1105039726 M * shuri but never come up 1105039746 M * _are_ not that i know of 1105040150 Q * shuri Read error: Connection reset by peer 1105040592 Q * infowolfe Remote host closed the connection 1105040612 M * chrish01 Bertl_oO, vc_add_vndev: No Such Process 1105040619 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105040620 M * chrish01 Bertl_oO, that occurs when i try to use vnet 1105040639 Q * infowolfe Quit: 1105040687 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105040725 Q * infowolfe Quit: 1105040734 J * infowolfe ~infowolfe@209-112-211-167-cdsl-rb1.nwc.acsalaska.net 1105042341 N * chrish01 chrish|lunch 1105043359 Q * infowolfe Remote host closed the connection 1105043490 J * Count ~count@pD9E7D319.dip0.t-ipconnect.de 1105043495 M * Count hiho :) 1105043532 M * Count uhm - I'm having problems creating directories as root (i.e. via dpkg etc) as root in a vserver using 2.4.27-v1.29 ? 1105043533 N * Count Guest30 1105043542 N * Guest30 elCount 1105044147 M * elCount any idea what the problem is, and/or how to work around it? :) 1105045102 N * chrish|lunch chrish01 1105045529 Q * serving Ping timeout: 480 seconds 1105045919 M * _are_ elCount: mkdir works? 1105046096 J * sannes ~ace@home.skarby.no 1105046334 J * axu_ ~Axu@81.223.125.177 1105046343 M * axu_ good evening folks 1105046349 M * _are_ Hi axu_ 1105046354 M * axu_ hi are 1105046406 M * axu_ i am about to try patching 2.6.10 with the patch-2.6.9-vs1.9.3.diff patch... is there a patch for 2.6.10 ? 1105046436 M * _are_ yes, as you run devel versions anyway... 1105046449 M * axu_ are: yes 1105046462 M * axu_ are: hmm, any hint where to find ? 1105046473 M * _are_ yes, just waiting for the page to load... 1105046482 M * _are_ http://vserver.13thfloor.at/Experimental/ 1105046490 M * axu_ thanks are :) 1105046516 M * _are_ I currently run 2.6.10-vs1.9.3.16 1105046571 M * axu_ are: nice, i run only older vserver versions but want to try the improved cpu and memory limitationoptions. also quotas. 1105046596 M * _are_ well, I only started with vservers 3 weeks ago and always used devel version. 1105046618 M * _are_ have 1 vserver with amd64 architecture, no use in older kernels there 1105046623 M * elCount _are_ yupp 1105046630 M * axu_ are: yeah not really :) 1105046631 M * elCount _are_: yupp, mkdir works 1105046655 M * _are_ elCount: then what directories dpkg can't create? debian (sarge) works fine here 1105046685 M * _are_ (even the 'oh-no-we-are-not-even-officially-unstable' amd64 version ;) 1105046693 M * elCount _are_: e.g. /usr/lib/perl/5.8.4/auto .. 1105046703 M * ndim _are_: you're using Debian, right? Did you compile the alpha util-vserver yourself or did you use one of the debs various people are currently building? 1105046733 M * _are_ ndim: compiled myself with the configure line I found in some old package 1105046775 M * _are_ elCount: I have no real idea about the old tools and ...-1.29 sounds like old tools/patch, I think I have read something about a 'dpkg fix' for old versions somewhere 1105046798 M * elCount *gooooogle* ;) 1105047202 N * Bertl_oO Bertl 1105047215 M * Bertl evening folks! 1105047229 Q * sannes Read error: Connection reset by peer 1105047258 M * _are_ Hi Bertl 1105047370 M * elCount hmm. was http://www.paul.sladen.org/vserver/archives/200402/0052.html ever solved? dpkg seems to do mkdir("blah/blah.dpkg-new", 0); mkdir("blah/blah.dpkg-new", 0);, the latter fails ... 1105047502 M * Bertl hmm ... no idea what the issue actually was (although it seems that I did comprehend it back then) 1105047608 M * _are_ :-> 1105047633 M * chrish01 Bertl, hey dude 1105047661 M * chrish01 Bertl, "vc_add_vndev: No Such Process" // When i try to run `vnet` 1105047692 M * sladen elCount: I think that with create a directory with permissions 000 which then cannot be accessed outside the host-server 1105047712 M * elCount sladen: sounds plausible 1105047756 M * sladen elCount: what dpkg should actually do is mkdir(..., 0755); chmod (... 0755); 1105047789 M * sladen elCount: it's dpkg which is in the wrong here; ...however, how to work around it? 1105047855 M * elCount I agree. 1105047868 A * elCount has a tremendous headache and will get some aspirine 1105048767 M * Bertl hey chrish01! 1105048838 M * Bertl try a modified version of the vnet3_setup script 1105048853 M * Bertl (or the script itself, if the ips are fine for your system) 1105048864 M * chrish01 ah shit, i knew i forgot something :) 1105048868 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/vnet3_setup.sh 1105048886 M * chrish01 thanks, ill try that in a few...pushing an svn patch for some stuff at work 1105048894 M * Bertl np 1105049244 M * elCount hm. 1105049260 M * elCount mkdir foo ; chmod 0000 foo; chmod 0700 foo fails, too. 1105049263 Q * Alecsandro Quit: brb 1105049356 M * Bertl elCount: please describe your system (kernel/patches/tools) 1105049419 M * elCount Bertl: kernel 2.4.27-vs1.29; Debian unstable; util-vserver 0.29-5 on host 1105049449 M * Bertl debian tools .. hmm ... probably that is the issue here ... 1105049457 M * Bertl sec 1105049635 M * Bertl you should have a tool called showattr, please try it on the host with the directory in question as argument 1105049743 M * elCount Bertl: hm. I don't seem to have showattr 1105049785 M * Bertl hrm, okay, get the util-vserver 0.30 tools, configure and compile them from source ... 1105049791 M * Bertl (not the debian version) 1105049831 M * elCount ah, there. /var/lib/vservers/shell/home/count/foo 00000000 1105049852 M * Bertl that is the directory you try to change? 1105049853 M * elCount it is there, it's just not in the path and doesn't show up for me using locate 1105049858 M * elCount Bertl: yupp. 1105049879 M * elCount d--------- 2 1000 users 4096 Jan 6 23:07 /var/lib/vservers/shell/home/count/foo 1105049882 M * Bertl okay, then what does ls -lda /var/lib/vservers/shell/home/count /var/lib/vservers/shell/home/count/foo 1105049913 M * Bertl and lsattr -d /var/lib/vservers/shell/home/count /var/lib/vservers/shell/home/count/foo 1105049916 M * Bertl return 1105049943 M * axu_ hi bertl :) 1105049954 M * Bertl hey axu_! 1105049993 M * elCount shall I flood the channel with the 6 lines? 1105050006 M * Bertl yep 1105050010 M * axu_ bertl: i am about to compile patch-2.6.10-vs1.9.3.17.diff :) 1105050011 M * elCount slop:~# ls -lda /var/lib/vservers/shell/home/count /var/lib/vservers/shell/home/count/foo 1105050012 M * elCount drwx-----x 7 1000 users 4096 Jan 6 23:07 /var/lib/vservers/shell/home/count 1105050012 M * elCount d--------- 2 1000 users 4096 Jan 6 23:07 /var/lib/vservers/shell/home/count/foo 1105050012 M * elCount slop:~# lsattr -d /var/lib/vservers/shell/home/count /var/lib/vservers/shell/home/count/foo 1105050012 M * elCount ----------------- /var/lib/vservers/shell/home/count 1105050012 M * elCount ----------------- /var/lib/vservers/shell/home/count/foo 1105050023 M * axu_ on intel 1105050030 M * Bertl excellent! 1105050051 M * axu_ ;) 1105050108 M * Bertl elCount: and chmod 0700 /var/lib/vservers/shell/home/count/foo says? 1105050145 M * elCount slop:~# chmod 0700 /var/lib/vservers/shell/home/count/foo 1105050145 M * elCount slop:~# echo $? 1105050145 M * elCount 0 1105050192 M * Bertl hmm, so where was the problem again? 1105050718 M * elCount Bertl: can't do the same in a vserver. 1105050734 M * elCount Bertl: at least not on vs1.29 1105050747 M * Bertl hmm ... okay, I understand, sec checking ... 1105050912 M * elCount tia :) 1105051202 M * axu_ bertls: my compile fails with this errors: http://www.revolto.at/error.txt any ideas ? 1105051244 M * Bertl 2.6.10 issue, disable the gamma_drm or add a fix ... 1105051246 M * Doener elCount: FYI the issue with 1.25 was that the notail flag (+t for chattr) was inherited. That flag is used by linux-vserver (stable branch) to protect the chroot barrier. 1105051261 M * Doener that was fixed in 1.26, i.e. +t is no longer inherited 1105051274 M * Bertl Doener: I can reproduce his issue with 2.4.27-vs1.29 :/ 1105051277 M * Doener evening btw ;) 1105051293 M * axu_ hi doener :) 1105051298 M * axu_ bertl: ic, ok :) 1105051316 M * Doener Bertl: hm, got instructions at hand? didn't follow to closely 1105051320 M * Doener s/to/too/ 1105051328 M * Bertl simple, enter a vserver and do: 1105051357 M * Bertl cd /tmp; mkdir foo; mkdir foo/bar; chmod 0000 foo; chmod 0000 bar 1105051371 M * Bertl oops, last should be chmod 0000 foo/bar 1105051381 M * Bertl and the last one will fail ... 1105051474 M * Doener yep... 1105051497 M * Doener but you can still chmod 700 foo... 1105051563 M * Bertl yes, but I don't understand why that happens at all? 1105051598 M * Bertl stat64("foo/bar", 0x8ffffa20) = -1 EACCES (Permission denied) 1105051628 M * Bertl elCount: I consider this a bug and I will fix this ... 1105051652 M * elCount Bertl: great :) truly appreciated and will help a lot ;) especially for the debian users 1105051672 M * elCount Bertl: asking for a timeframe wouldn't be nice, would it? ;) 1105051686 M * Bertl np, I guess it will at least take 20minutes ;) 1105051846 M * elCount Bertl: cool, I'll check back later ;) 1105051852 M * Bertl k 1105051923 M * chrish01 Bertl, sorry im not gonna be able to test this until tonight...i got a meeting with some investors (yeah!) 1105051936 M * Bertl np, take your time ... 1105052310 J * serving ~serving@213.186.171.54 1105052483 M * Bertl welcome serving! 1105052542 Q * chrish01 Quit: investors (yeah!) 1105052685 Q * jsambrook Ping timeout: 480 seconds 1105053087 M * Bertl Doener: bingo! 1105053102 M * Bertl fs/namei.c, vfs_permission() doesn't check the barrier correctly 1105053104 M * Doener 5 in a row? 1105053106 M * Doener ;) 1105053134 M * Bertl what if we do an IS_BARRIER() macro for that? 1105053163 Q * psychops Quit: 1105053174 M * Bertl and much more important, how do we call the next 1.2 release? 1105053182 M * Doener didn't we have such a macro once? maybe my memories are playing tricks on me... macro sounds good anyway 1105053201 M * axu_ time for reboot :) 1105053227 Q * axu_ Quit: Leaving 1105053260 M * Doener well... 1.30 is 'free' but probably confusing. 1.210 is even more confusing, so 1.2.10 seems right, although we would change the naming scheme within a branch :/ 1105053313 M * Bertl hmm, and starting a 1.4 release is probably against the 'let it die in peace' strategy ;) 1105053349 M * Doener indeed, and we have no real point for calling it 1.4.x ... 1105053392 M * Bertl right ... what about 1.2A then? *G* 1105053511 M * Doener maybe -final? 1105053522 M * Bertl yes, final that's it! 1105053540 M * Bertl no, seriously, we'll ask the community, they have to live with it ... 1105054268 J * sannes ~ace@home.skarby.no 1105054614 M * Bertl welcome sannes! 1105054635 M * we2by I;m still trying to fix the respawn too fast problem 1105054718 M * we2by Bertl, how do I share a dir? ? 1105054881 M * Bertl hum, with whom? 1105055204 M * we2by with my vservers 1105055215 M * we2by I want to share a dir on the host OS with my vservers 1105055243 M * Bertl well, just use it, the host can access it anyways? 1105055856 M * we2by Bertl, why not 1105055861 M * we2by it is just a dir on the host os 1105055871 M * we2by and I like 2 or more vservers to access the dir 1105055917 M * Bertl ah, okay, so you want to share a dir among several vservers? 1105055932 M * Bertl read only or read write?