1111018164 Q * eyck Ping timeout: 480 seconds 1111018216 M * Bertl douglas: the hard scheduler 'ensures' that no process is run without having tokens available, the soft one just adjusts the priorities 1111018227 M * Bertl (soft is smoother ;) 1111018345 M * matti Bertl: ;] 1111018357 M * douglas ahh 1111018371 M * douglas cool beans, reading the linux paper, does it explain that in the paper vertl? 1111018373 M * douglas errr 1111018375 M * douglas bertl 1111018376 M * Bertl matti: `:] 1111018383 M * matti ;) 1111018397 M * douglas sorry using a microshit natural keyboard. it was free and my old one broke, so it'll hav eto do until I get a new keyboard. 1111018410 M * Bertl that's no excuse!!! ;) 1111018435 M * DaCa Bertl: I am a bit worried about the clock on your webserver :) 1111018443 M * douglas you try typing on one of these POS's. 1111018463 M * Bertl well, the paper might not explain it in all detail, but I guess it should contain some basic concepts ... 1111018471 M * Bertl DaCa: hmm .. yeah? 1111018480 M * Bertl DaCa: which one? 1111018515 M * DaCa Bertl: the one hosting the wiki 1111018534 M * Bertl hmm, again, which one? (the vserver wiki?) 1111018579 M * DaCa yes, linux-vserver.org 1111018612 M * Bertl Wed Mar 16 19:13:11 EST 2005 hmm, should be fine, no? 1111018717 M * Bertl EST = UTC - 5 1111018750 M * DaCa is it hosted in the us? you should be in CET right? the RecentChanges doest show the timezone 1111018766 M * DaCa s/doest/doesn't/ 1111018788 M * Bertl hmm, good point, maybe we should add that or change it to UTC ... 1111018835 M * Bertl and yes, it's probably in canada somewhere (provided by Jacques) 1111019024 J * eyck eyck@81.219.64.71 1111019222 M * Bertl welcome eyck! 1111019410 M * douglas this paper if I'm reading it correctly mixes stable and devel stuff. 1111019424 M * douglas I mean it explains it, and explains stuff thats only available in the devel version. 1111019428 M * douglas thats kinda confusing. 1111019558 M * Bertl well, the 'development' version is what we are working on ... and the 'stable' version is a subset of that (well, functionality wise) 1111019595 M * douglas yea, but certain features that are offered in the devel are not offered in stable. 1111019609 M * Bertl that is what 'subset' means ;) 1111019619 M * douglas thats like grouping a brand new ford mustang with a mustange from 68. 1111019620 M * DaCa douglas: I found that too, thats why in my slides I separated the base infrastructure, and then the extras for stable and unstable 1111019627 M * Bertl (maybe I should have said, true subset ;) 1111019661 M * Bertl douglas: thing is, you can take a 2.4 kernel/vserver and upgrade to 2.6 instantly ... 1111019679 M * douglas bertl, I understand what your saying, but the paper doesn't say that. it just groups them together hoping the reader comes and ask questions or figured the person will try using some of these and find out they are not available and be bright enough to know that he didn't mess up, its just not offered. 1111019715 M * Bertl that's why we have http://linux-vserver.org/Release+FAQ 1111019727 M * DaCa there is a chart somewhere on the wiki that lists a matrix of what is supported in which version 1111019736 M * douglas yea I know, I'm doing that myself, its just it sucks, I would rather not upgrade to 2.6 unless I had to. and I'm one of many people who dont want to, yet I figured it out that I had to get the "good stuff" :) 1111019774 M * Bertl no, wrong, you can start backporting 2.6/1.9.x stuff to 2.4/1.2 or 1.3 ;) 1111019783 M * douglas haha 1111019791 M * douglas I'm no means a programmer. 1111019794 M * douglas that is WELL out of my scope. 1111019844 M * douglas I didn't mean for it to get this big of an argument, I was just stating that its confusing at times because of that. 1111019880 M * Bertl feel free to make it less confusing ;) 1111019917 M * DaCa :) 1111019945 M * douglas perhaps I will just to spite you! 1111019946 M * douglas :) 1111019978 M * douglas I'll give me self 1/3 odds on doing that. 1111019978 M * Bertl excellent idea! *G* 1111019997 M * douglas I'll be betting on I won't. but feel free to bet I will :) 1111021058 M * Bertl are some folks around using rsync to move vservers? 1111021179 M * douglas yes. 1111021228 M * douglas some. most easily is gzip and sftp or something. 1111021719 Q * mcp Ping timeout: 480 seconds 1111021966 M * douglas rsync through ssh tunnels. anyone using just plan rsync should be slaped with a security for dummies book :D 1111022276 J * chairuou ~chairuou@210.245.70.1 1111022304 M * Bertl evening chairuou! 1111022337 M * Bertl douglas: okay, can I tell rsync to copy over a root filesystem but not the mounted (sub) partitions? 1111022409 M * ciphernaut when doing == vserver $VSERVER exec dpkg -i /cached_configs/local_pkg_overides/*.deb == from a bash script on the host, the evalution of * fails 1111022412 M * DaCa Bertl: -x 1111022434 M * DaCa -x, --one-file-system don't cross filesystem boundaries 1111022453 M * ciphernaut do i need to dump the command into a seperate script and then execute it? 1111022454 M * Bertl ciphernaut: no it works as expected, but your expectations differ from that ;) 1111022545 Q * chairuou Read error: Connection reset by peer 1111022552 M * ciphernaut so the evalution happens before enetering the vserver, but the path does not exist on the host, and * expands to nothing... 1111022579 M * douglas I think there may be some exclude options, but I dunno 1111022580 M * Bertl precisely, as defined by your shell 1111022595 M * Bertl (that was for ciphernaut) 1111022597 M * ciphernaut bertl: is it possible to do this without resorting to generating yet another script? 1111022613 M * Bertl try with shell escapes and quoting ... 1111022701 J * chairuou ~chairuou@210.245.70.1 1111022710 M * Bertl DaCa: thanks, that was what I was hoping for ... 1111022952 M * ciphernaut vserver $VSERVER exec /bin/sh -c "dpkg -i /cached_configs/local_pkg_overides/*.deb" 1111025185 J * tad ~tad@h000f3d19edcc.ne.client2.attbi.com 1111025253 M * Bertl wb tad! 1111025276 M * tad Hi everyone 1111027844 Q * ciphernaut Ping timeout: 480 seconds 1111034672 M * Bertl okay, folks, I'm off for now ... cya all! 1111034677 M * Bertl night tad! 1111034704 N * Bertl Bertl_zZ 1111035163 Q * nox Ping timeout: 480 seconds 1111035192 J * nox ~nox@213.39.193.118 1111035251 Q * tad Remote host closed the connection 1111036949 Q * chairuou Ping timeout: 480 seconds 1111039531 J * DuckMaster ~Duck@dyn-83-157-201-31.ppp.tiscali.fr 1111039915 Q * duckx Ping timeout: 480 seconds 1111040567 J * chairuou ~chairuou@210.245.70.1 1111045865 J * anonymous-coward ~nwalsh@shaggy.internode.com.au 1111046378 Q * DuckMaster Quit: Client exiting 1111047034 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111048668 Q * anonymous-coward Quit: adios 1111048838 Q * erwan_ho Remote host closed the connection 1111048981 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111048998 Q * erwan_ho Remote host closed the connection 1111049657 J * prae ~prae@134.106-14-84.ripe.coltfrance.com 1111052135 Q * nox Quit: I shouldn't really be here - dircproxy 1.0.5 1111052163 J * nox ~nox@213.39.193.118 1111052819 J * mcp ~hightower@www.c-tera.de 1111054121 J * erwan_taf ~erwan@81.80.43.68 1111054602 Q * erwan_taf Remote host closed the connection 1111055073 Q * chairuou Quit: 1111055323 J * jsambrook ~jsambrook@aelfric.plus.com 1111055402 J * erwan_taf ~erwan@81.80.43.68 1111056129 Q * monrad Quit: Leaving 1111057570 M * prae http://www.linux.it/~carlo/somehacks/bup/README 1111057589 M * prae "Benjamin Gigon [11] (2005-02-28) sent me a patch he made for linux-vserver.org that "integrate hostname into log"" 1111057603 M * prae http://www.linux.it/~carlo/somehacks/bup/bash-2.05b-syslog_udp01.patch 1111057609 M * prae http://www.linux.it/~carlo/somehacks/bup/gigon-3.0.patch 1111057611 M * prae *happy* 1111057613 M * prae :) 1111057613 M * erwan_taf lo prae 1111057617 M * prae Mister Wan :) 1111057619 M * erwan_taf :b 1111059077 M * Doener morning... 1111059102 M * prae Hi Do' :) 1111059336 M * DaPhreak mornin you two ;P 1111059639 M * prae Mister DaPhreak :) 1111059660 M * DaPhreak yeah, thats my name :) 1111059711 M * DaPhreak whats it with this patch ? does it log every line typed in bash on a vserver to the vhost's syslog ? 1111060500 Q * _BWare_ Read error: Operation timed out 1111060942 M * prae yes 1111060962 M * prae original patch log every line typed in bash into syslog 1111061034 M * prae but, into vservers, le original patch can't differs vserver and main server 1111061038 M * prae s/le/the/ 1111061101 M * prae into main.syslog, it's a big bazaar 1111061138 M * prae with second patch, each line into main.syslog is added with vs_hostname 1111061145 M * prae easy to grep :) 1111061161 M * DaCa whats the point of this patch? there a tons of other ways to issue commands rather than typing them in your bash shell 1111061221 M * prae other ways ? ... grsec.Exec.log ? 1111061246 M * DaPhreak no prae, DaCa means there are other ways to _type_ in commands ;) 1111061284 M * DaCa other ways to execute commands, so this logging is worthless imho 1111061302 M * prae oh! ok 1111061305 M * prae yes of course 1111061331 M * DaCa using grsec to log exec's is indeed a far better auditing solution :) 1111061348 M * prae but it's additional layer for the security 1111061356 M * prae not the graal for the security :) 1111061370 M * prae yes of course, grsec.log.exec is better ;) 1111061434 M * prae it is a simply reach for logging :) 1111061471 M * prae (by example into vserver with minimal system _and_ without grsec :P 1111062455 Q * ola Read error: Operation timed out 1111062583 J * ola ~ola@c-adt-5.ataco.se 1111063883 M * pusling is it possible to get to use mount --bind inside a vserver ? 1111063958 M * alexx without good capabilities : "mount: permission denied" 1111063993 M * pusling what capabilities is needed? And what could be unwantet side-effects ? 1111064154 M * alexx with vs1.9.x ? 1111064295 M * alexx on vs1.9.x with new configuration system, put "mount" into /etc/vservers/xxx/ccapabilities 1111065132 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1111066733 M * pusling vs19 - that is for kernel 2.6, right ? (only using kernel 2.4) 1111066921 N * _maharaja maharaja 1111066966 M * Doener pusling: yes 1111067152 J * micah_ micah@micha.hampshire.edu 1111067505 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net :: 1111067574 Q * micah Ping timeout: 480 seconds 1111068412 J * tad ~truex@static-151-204-232-50.bos.east.verizon.net 1111068822 M * pusling ... I just had to learn how to set up apache instead of mount --bind stuf into /var/www ;) 1111068919 M * DaCa pusling: can't you mount --bind them from outside the vserver? (from the root server) 1111068944 M * pusling not without restarting the vserver, I think 1111069106 M * daniel_hozac sure you can. 1111069498 M * pusling okay... Instead I just learned a bit about the apache.conf ;) 1111069530 J * miller7 ~none@adsl49-static-gw1.access.acn.gr 1111069580 A * miller7 waves at everyone 1111070672 A * DaCa waves back 1111071130 J * witchdoc ~witchdoc@d463c3f1.datahighways.de 1111071137 M * witchdoc hi all 1111071417 M * witchdoc i am using util-vserver 0.30.203 with 2.6.11-vs1.9.5-rc1 with debian/sarge as host and vserver and i am confused with the some init-script errors on my vserver 1111071507 M * witchdoc i get Starting monitor daemon: monCan't exec "start-stop-daemon": No such file or directory at -e line 18. 1111071511 M * witchdoc Error running start-stop-daemon: No such file or directory 1111071545 M * witchdoc but if i start the init script from the vserver, it works. 1111073567 M * Doener pusling: http://linux-vserver.org/Namespaces ... should tell you why you didn't see your bind mounts ;) 1111073648 M * DaCa Doener: he said he's using 2.4 1111073682 M * Doener hm, right 1111073707 M * Doener but 2.4 also supports namespaces... 1111073737 M * Doener if he's using alpha tools + new style config with 2.4 (which is perfectly fine) he may be using namespaces 1111073765 M * pusling I have converted to alpha tools and new config style (took 3 hours) 1111073773 M * Doener DaCa: *g* 1111073788 M * DaCa Doener: with stable vserver? 1111073800 M * Doener namespaces are a vanilla feature 1111073806 N * BobR_oO BobR 1111073810 M * Doener were added somewhere around 2.4.19 or something 1111073822 Q * douglas Quit: 1111073878 M * DaCa Doener: then the matrix is incorrect or confusing 1111073898 M * DaCa Doener: I should look in the 1.26 column for stable right? 1111073928 M * DaCa there is no yes for syscall namespace commands 1111073938 M * Doener hm, good point, the additional vs namespace commands are missing on 2.4... 1111074023 M * Doener pusling: on the host, does /proc/mounts contain, for example, the proc mounts of the vservers? 1111075136 N * BobR BobR_afk 1111075189 N * DaPhreak DaPhreak|off 1111075672 J * Hollow ~Hollow@home.xnull.de 1111076540 M * Doener back later... 1111076543 N * Doener Doener|gone 1111076739 N * BobR_afk BobR 1111077297 N * BobR BobR_oO 1111077416 N * Doener|gone Doener 1111077655 N * Bertl_zZ Bertl 1111077663 M * Doener evening Bertl! 1111077678 M * Bertl evening Doener! Folks! 1111077795 M * Bertl received a mail telling me that my talk at 'Linuxwochen 2005' (Austria) has been accepted (so that's a chance to visit austria and meet interesting folks ;) 1111077881 M * Bertl Doener: I guess I fixed the ppc64 regressions ... 1111077915 M * Doener great 1111077922 M * Bertl but we are still lacking a syscall assignment for arm 1111077925 M * miller7 well done Bertl 1111077973 M * Bertl thanks, but it was just some compile time warnings ... 1111079102 M * rs hey Bertl! 1111079124 M * Bertl evening rs! LTNS! 1111079148 M * Doener LTNS? 1111079156 M * rs long time no see 1111079162 M * Doener ah 1111079167 M * rs yeah :/ 1111079171 M * rs I'm back :) 1111079180 M * Bertl for real? ;) 1111079189 M * rs I hope :) 1111079191 M * Bertl or just a short visit ... 1111079205 M * rs should be for real :) 1111079241 M * rs I'm porting the nfsall patch for 2.6.11.4-vs1.9.5-rc3 btw, maybe someone is interested 1111079247 M * rs I plan to test this kernel in production soon 1111079259 M * rs no bug discovered so far (in this release ?) 1111079285 M * Bertl hmm, wasn't some of that (NFSAll) included in mainline? 1111079310 M * Bertl rs: well, 1.9.5-rc3 seems stable and fine (at least on 2.6.11.3) 1111079345 M * rs the .4 is just a security fix in pppd IIRC 1111079350 M * rs shouldn't hurt 1111079373 M * albeiro pppd is b0rked as well as 2.4 is. does not matter ;p 1111079378 M * albeiro er.r.. 1111079382 M * albeiro i mean 2.6 of course 1111079637 M * Bertl rs: I got a request for disk I/O accounting from ntrs ... maybe that's something not just he is interested in ... in future maybe disk bandwith limitations? 1111079767 M * rs yeah I think we already talked about it? 1111079799 M * Bertl yes, so that's probably something for the next devel cycle ... 1111079819 M * Bertl we still have a lot of 'open' (to be completed) features ... 1111079877 M * Bertl - ipc limits are partially there, but mostly untested 1111079897 M * rs nice, si it in 1.9.5 ? 1111079900 M * rs is 1111079904 M * Bertl - the capability mask stuff needs to be completed 1111079926 M * Bertl rs: yes, what we added is in 1.9.5-rc3 ... 1111079971 M * rs k, it just accounted, so it shouldn't hurt to use it in production I guess? 1111079979 M * rs it's 1111080199 M * rs how does ngnet ? is there some new things ? 1111080200 Q * prae Quit: Client exiting 1111080316 M * Bertl rs: ngnet does work except for arp ... 1111080336 M * rs what's wrong with arp ? 1111080355 M * Bertl we managed to logon via ssh to the 'host' and switch over to the vserver, as well as http/bind and ssh connections directly to the guest 1111080366 M * Bertl the problem with arp is the arp-reply 1111080381 M * Bertl you do not know to which vserver it should be relayed ... 1111080399 M * Bertl a cruel hack would be to send it to each vserver 1111080422 M * rs yeah link if they are connected to an hub 1111080447 M * Bertl yes, precisely .. but we are aiming for a virtual switch, so it's not working yet ;) 1111080459 M * rs or assign a diff mac to each vservers, but would be a lot of work for the kernel I guess, and not that easy to generate some uniq mac 1111080520 M * Bertl well, yes, and actually that's something we didn't try yet ... 1111080558 M * Doener Bertl: brain just spits out something: what about a list of virtual interfaces attached to a neighbour and faking the receival of arp replies for the vservers? 1111080577 M * Doener hm, would break arp filtering... 1111080634 M * Doener ... probably that is... 1111080644 M * Bertl the idea isn't that bad, I guess we should do something like fake arp replies for the vservers/host 1111080694 M * Doener there's arp_resend() or something like that (for bridging IIRC) maybe we can make use of that stuff someho 1111080699 M * Doener somehow even 1111080737 M * Bertl and we also should make (or maybe modify) a list of 'waiting' interfaces for arp replies 1111080794 M * Bertl (that's how the current linux-implementation does it) 1111081287 M * Bertl we also need to think about some way to 'configure' the virtual switch 1111081300 M * Bertl (i.e. to create partitions and such in the future) 1111081405 M * rs what do you mean by "partitions"? 1111081457 M * Bertl well, it would be easy to say, vserver A,B and C see eachother, where D is connected over a vlan and doesn't see A-C ;) 1111081481 M * Bertl then A-C would for one partition, where D is another ... 1111081578 M * rs hmm I see 1111081605 M * rs we could do this with normal vlan btw 1111081620 M * rs do you think it's a good idea to do this at this level ? 1111081650 M * Bertl well, if you 'ping' A from D and the virtual switch does not know vlans, you will get the reply, right? 1111081676 M * Bertl (because it all happens on the same host) 1111081684 M * rs because it's on the same host ? 1111081686 M * rs yeah ok 1111081693 M * rs it's a good point 1111081721 M * Bertl but I guess those are details, just to be considered for the future ... 1111081733 M * Bertl nothing we have to implement right now ... 1111081745 M * rs I think so 1111081769 M * Bertl I also do not really consider the arp issue a big deal 1111081786 M * Bertl it requires to make a static arp entry, so what? ;) 1111081873 M * rs yeah indeed :) 1111082286 M * Doener back later.. 1111082289 N * Doener Doener|gone 1111083445 M * rs Bertl: perhaps you can put this in your experimental patch list: http://rs.rhapsodyk.net/vserver/linux-2.6.11.4-vs1.9.5-rc3-NFS_ALL.dif 1111083543 M * rs hum don't do it :) 1111083551 M * Bertl hmm? 1111083561 M * Bertl broken? 1111083566 M * rs seems yes :) 1111083636 M * Bertl anyway ... you should leave the ol' DOS times behind 1111083683 M * Bertl rs: do you use all that stuff from the NFS_ALL patch? 1111083686 M * rs the ol' DOS times ? 1111083701 M * Bertl nowadays it's .diff not .dif ;) 1111083718 M * rs yeah say that to trond :) 1111083724 M * rs http://linux-nfs.org/Linux-2.6.x/2.6.11/linux-2.6.11-NFS_ALL.dif 1111083767 M * Bertl hmm, that puts the NFS stuff into a totally new light ... 1111083785 M * rs hehe :) 1111083815 M * rs do you think that trond work with DOS ? :) 1111083883 M * Bertl no, never would assume that ... though ... 1111084131 M * albeiro Bertl: sorry about offtopic, but that is currently the best way of interacting kernel with userspace - /dev, procfs, sysfs, syscalls ? 1111084158 M * Bertl depends on who you ask ... 1111084216 M * Bertl there is only one thing all kernel folks (on lkml) seem to agree upon: that ioctl is evil ;) 1111084221 M * albeiro for sure not a device driver writer... 1111084275 M * Bertl I'd say, for read-only info /proc, for simple (one value per setting) read/write use sysfs 1111084284 M * albeiro i gues there is not reason to worry if i am using /proc now ? 1111084294 M * albeiro or will be in future ? 1111084308 M * Bertl for 'actions' or process specific stuff use a syscall 1111084322 M * Bertl and for large amounts of data to and from use netlink 1111084345 M * Bertl if you have a 'device' to handle, of course it's the device interface to use 1111084382 M * albeiro so i got a picture of using it all... 1111084420 M * albeiro are they going to rip out procfs ? or leave it only for process information ? 1111084611 M * Bertl the tendency is heading there ... but will take quite some time I guess 1111084726 M * albeiro thx ! 1111084758 M * albeiro hm, ontopic question (yes, unusual). was vserver tested on ppc ? 1111084916 M * rs Bertl: http://rs.rhapsodyk.net/vserver/linux-2.6.11.4-vs1.9.5-rc3-NFS_ALL.dif <= this one is ok 1111085052 M * Bertl k 1111085248 J * Vudumen vudumen@perverz.hu 1111085249 M * Vudumen hi 1111085265 M * Bertl welcome Vudumen! 1111085315 M * Vudumen well i compiled 2.6.11-vserver on my test-box without the legacy api. i updated the util-vserver package to 0.30-203. 1111085332 M * Vudumen newvserver creates .conf but not /etc/vservers//* 1111085353 M * Bertl newvserver is not part of util-vserver (the tools) 1111085363 M * Vudumen do you have any examples of this directory? i started creating one but it would be so much faster :) 1111085366 M * Vudumen Bertl: i know 1111085371 M * Vudumen it's in vserver-debiantools 1111085379 M * Vudumen creating the filesystem was easier with this :) 1111085394 M * Vudumen and it made much better fs than a simple debootstrap (/dev etc...) 1111085396 M * Bertl you can use 'vserver build ...' 1111085412 M * Vudumen Bertl: hmm let's give it a try :) 1111085417 M * Bertl which does all this _and_ creates a config ... 1111085434 M * Bertl check vserver - build --help first 1111085451 M * Bertl and have a look at http://linux-vserver.org/alpha+util-vserver 1111085457 M * Bertl (contains a few examples) 1111085492 M * Bertl you can also use the -m skeleton build method to create an empty filesystem plus config 1111085528 M * albeiro hm 1111085543 M * Bertl the * are typically things like hostname or ip address, flags or special caps 1111085551 M * albeiro http://www.solucorp.qc.ca/miscprj/s_context.hc <- flash with vserver is nice there, mayby it should go on main site of vserver ? ;p 1111085635 M * Bertl well, yes and no, personally I don't like flash stuff ... but if the community likes it, and Jacques agrees, I don't see a reason against it ... 1111085818 M * Vudumen hmm interesting 1111085875 M * Vudumen i have a separate partition mounted on /vservers/spray. okay i use --force to use this dir buit it wants to move to another location :/ 1111085945 M * albeiro Bertl: is it you having 33 vservers on your notbook ? 1111086004 J * witchdoc_ ~witchdoc@d463c3f1.datahighways.de 1111086083 M * Bertl albeiro: no currently no vservers on my notebook 1111086122 Q * witchdoc Ping timeout: 480 seconds 1111086136 M * Bertl Vudumen: hmm, where did you 'force' that? 1111086148 M * Bertl /vservers/spray looks very sane to me btw ;) 1111086194 M * albeiro anyway, somebody had. 1111086208 M * Bertl (sidenote, maybe you should try util-vserver 0.30.204 from the tar archive ;) 1111086256 M * Vudumen Bertl: i did 1111086259 M * Vudumen it wants to rename 1111086275 M * Vudumen but /dev/mapper/virtuals-spray is mounted on /virtuals/spray 1111086289 M * Vudumen well it was a streaming media broadcaster so it's called spray :) 1111086291 M * Bertl ah, okay, I understand what you mean ... 1111086304 M * Bertl you have something mounted on the vserver dir 1111086316 M * Bertl and you want the tools to install into that mount ... 1111086338 M * Vudumen exactly 1111086379 M * Bertl I heard of that giving some issues, and I guess Enrico (the genious doing the tools) fixed that in the cvs already ... 1111086399 M * Bertl but a simple and quikc workaround would be: 1111086420 M * Bertl mount the /dev/mapper/virtuals-spray on /vservers 1111086442 M * Bertl install the vservers and move the contents of the dir up one level 1111086451 M * Bertl *vserver 1111086461 M * Vudumen okay :) 1111086462 M * Bertl then unmount, and remount at the target location 1111086483 M * Bertl I know it's kind of hacky, but you could also adjust the vserver script 1111086535 M * Vudumen :) 1111086542 M * Vudumen now it's installing the base sytem 1111086544 M * Vudumen i go to smoke :) 1111086546 M * Vudumen will be back 1111086553 M * Vudumen thanks for the help :) 1111086559 M * Bertl you're welcome! 1111086963 M * Vudumen hmm 1111086970 M * Vudumen vserver spray start 1111086981 M * Vudumen and it sais the RTNETLINK answers: File exists 1111086992 M * Vudumen chbind: vc_set_ipv4root(): Function not implemented 1111087013 M * Vudumen i used eth0 as in /etc/vservers/spray/interfaces/0/dev 1111087015 M * Vudumen is it okay? 1111087037 M * Bertl depends, but the second mone indicates that the kernel is not patched (or not properly installed) 1111087047 M * Vudumen hmm 1111087055 M * Bertl s/mone/one 1111087056 M * Vudumen isn't it somethink in the legacy api? 1111087073 M * Vudumen s/ink/ing/ :) 1111087077 M * Bertl yep, which kernel did you compile/use? 1111087098 M * Vudumen 2.6.11 1111087111 M * Bertl patches? 1111087119 M * Vudumen + vserver 1.9.4 1111087131 M * Vudumen no 1111087134 M * Vudumen w8 a second 1111087151 M * Vudumen patch-2.6.11-rc3-vs1.9.4 1111087155 M * Vudumen this was the patch 1111087157 M * Bertl I'd suggest to use 2.6.11.4 and 1.9.5-rc3 (soon to be released as 1.9.5) 1111087170 M * Vudumen hmm can you giv eme an url to this patch? :) 1111087186 M * Bertl but it probably works with your kernel too, you have to enabled the legacy api (for networking) 1111087194 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.11.3-vs1.9.5-rc3.diff 1111087198 M * Vudumen hmm legacy api is required? 1111087219 M * Bertl (this has a separate option for legacy networking, with an appropriate help text) 1111087238 M * Vudumen fine :) thanks 1111087260 M * Bertl my pleasure! 1111087487 M * Vudumen sometime in one month we will thest virtuozzo (www.sw-soft.com) but i don't think we will like it too much 1111087495 M * Vudumen it's redhat/fedora or suse based... 1111087514 M * Vudumen do you know this product? 1111087564 M * Bertl if you are going to test it, it would be nice to get some feedback how it compares ... and feel free to ask about missing features and/or how to do something with linux-vserver 1111087666 M * Vudumen okay :) thanks for your support :) 1111087970 M * Bertl np 1111088057 M * Vudumen hmm interesting 1111088057 M * Vudumen virtual:~# vps 1111088057 M * Vudumen vc_create_context(): Invalid argument 1111088065 M * Vudumen with 2.6.11.4-vserver 1111088106 M * Bertl yes, that's aknown issue if you remove the legacy support for now, will be fixed in the next tool release, I guess 1111088152 M * Bertl you can use vcontext --migrate --xid 1 ps auxwww (instead) 1111088155 M * witchdoc_ last 1111088159 J * duckx ~Duck@dyn-83-157-201-31.ppp.tiscali.fr 1111088161 M * witchdoc_ re 1111088167 M * witchdoc_ :-) 1111088170 M * Bertl wb witchdoc_! 1111088176 N * witchdoc_ witchdoc 1111088187 M * witchdoc hi Bertl 1111088213 M * Vudumen so it's much better if i enable legacy support? :) 1111088264 M * Bertl probably for now, should be solved in the next release 1111088289 M * Vudumen it sais that sched is an unknown flag :( 1111088319 M * Bertl which is true for new config .. you want sched_prio or sched_hard there ... 1111088346 M * Bertl well, the sched flag is a very old legacy flag which has no meaning on 2.6 kernels ... 1111088546 M * Vudumen thanks :) 1111088581 M * Vudumen which one is that when i set the limits for the whole vps instead of the processes? 1111088585 M * Vudumen prio? 1111088605 M * Bertl the cpu 'limits' are now handled by the token bucket scheduler 1111088624 M * Vudumen so it equals which do i use 1111088665 M * Bertl probably sched_prio is the best match, but it's not identical in behaviour 1111088742 M * Bertl the token bucket gives you much more fine control over a vserver, including hard limits to the cpu usage ... 1111088755 M * Bertl (which was not available in 2.4) 1111088767 M * Vudumen i see 1111088844 M * Vudumen and when i have an eth0 on the host what interfaces should i use in VPSes? 1111088865 M * Vudumen i mean /etc/vservers/interfaces//dev 1111088891 M * Vudumen simply any valid name? (ethX, dummyX, etc.) 1111088903 M * Bertl as I mentioned before, depends, if you want to use an existing ip, then you do not specify the dev at all 1111088912 M * Bertl you then use 'nodev' instead 1111088929 M * Bertl if you want the tools to create an alias or add an ip, you have to specify the 'dev' 1111088943 M * Vudumen hmm. with nodev i don't see any interface 1111088976 M * Bertl again, this depends on your address setup .. here is an example: 1111088980 M * Vudumen well i have many IPs here. and i want to use one ip for each vhost 1111088992 M * Bertl ifconfig eth0:hansi 10.0.0.2 1111089005 M * Bertl now if you want to use that in your vserver, you specify 1111089015 M * Bertl nodev, ip=10.0.0.2 1111089037 M * Bertl if you want to create the tools this alias on vserver startup, you specify 1111089049 M * Bertl dev=eth0, name=hansi, ip=10.0.02 1111089109 M * Vudumen okay. now i have eth0:1 as 1.2.3.4/255. so i can use dev eth0:1 1111089139 M * Bertl yep, but if you created it by hand, you specify nodev 1111089139 M * Vudumen no. simply eth0 1111089157 M * Bertl (ando dev entry at all, just ip=1.2.3.4) 1111089167 M * Bertl *and no 1111089174 M * Vudumen okay so if the interface exists befor the strt of the vps use nodev. else use the dev where the ip will be used on the host? 1111089188 M * Bertl where it will be 'created' yes 1111089204 Q * sebd Remote host closed the connection 1111089216 M * Vudumen fine :) 1111089217 J * sebd ~sebd@lesdeveloppementsdurables.org 1111089441 M * witchdoc hmm i am confused why some vserver init script fails to load. i use 2.6.11-vs1.9.5-rc1 and 0.30.203 alpha+utils 1111089465 M * witchdoc Starting monitor daemon: monCan't exec "start-stop-daemon": No such file or directory at -e line 18. 1111089498 M * witchdoc but if i start the "mon" daemon in the vserver manual, it starts fine 1111089519 M * witchdoc any hints? 1111089583 M * Bertl hmm, what is the mon(itor) deamon? 1111089602 M * witchdoc an perl-script for monitoring, nice to use 1111089628 M * Bertl and it has a runlevel script inside one of your vservers, I presume? 1111089643 M * witchdoc yes thats right 1111089662 M * Bertl did you try to call that one with start and stop? 1111089678 M * Bertl (because that is probably what the tools do ;) 1111089715 M * witchdoc yes, and if i do so, it works. 1111089745 M * micah_ howdy folks... I just tried to do a package upgrade in my reference vserver, and I am getting an Operation not permitted 1111089779 N * micah_ micah 1111089786 M * micah failed to rmdir/unlink `//usr/bin/chage.dpkg-tmp': Operation not permitted 1111089793 M * Bertl witchdoc: could you upload the runlevel script somewhere please (maybe pastebin.com)? 1111089809 M * witchdoc moment ... 1111089839 M * micah is my /usr mounted readonly or something? mount does not show this 1111089850 M * Bertl micah: what does lsattr /path/to/vserver//usr/bin/chage.dpkg-tmp show? 1111089884 M * Bertl btw, what is chage? 1111089892 M * micah I didn't know either :) 1111089900 M * micah but its part of the login package 1111089906 M * micah "change user password expiry information" 1111089952 M * Bertl interesting ... 1111089955 M * micah Bertl: chage.dpkg-tmp doesn't exist, because it is created in the package upgrade when it does mv chage chage.dpkg-tmp 1111089963 M * micah but... cage has the immutible flag set in lsattr 1111089971 M * micah I did not set this 1111089980 M * Bertl ;) 1111089980 M * witchdoc Bertl: www.netzwech.de/karsten/debian/mon is the init.d script for it 1111089999 M * micah many things in /vserver/seedbank/usr/bin have +i :) 1111090021 M * Bertl well, that's fine as long as they have the iunlink flag too 1111090021 M * Vudumen hmm an interesting question. i have an ip in the VPS. can i administer iptables rules for this IP from the VPS? 1111090042 M * Bertl Vudumen: not at the moment and not with legacy networking ... 1111090052 M * micah Bertl: they are only ----i------------ /vservers/seedbank/usr/bin/chage 1111090056 M * Vudumen Bertl: will it be pssible with the new networking code? 1111090065 M * Bertl but you can add rules to the host's iptables 1111090075 M * Bertl yes ngnet will allow this (soon) 1111090112 M * Vudumen fine :) 1111090241 M * micah Bertl: it appears everything in /vservers/seedbank/usr/sbin is also the same attr (with +i) 1111090266 M * Bertl check with showattr, if the proper flags I and U are set 1111090292 M * micah Bertl: yes ----UI- /vservers/seedbank/usr/bin/chage 1111090302 M * Vudumen Bertl: thanks for all your help now i go home :) 1111090303 M * Vudumen bye 1111090311 M * Bertl micah: so that is fine, just directories should not have any I or U 1111090315 M * Bertl Vudumen: cya! 1111090341 M * micah Bertl: some files have i and some have I 1111090379 M * Bertl i means Immutable not set, u means IUnlink not set and I and U mean that the corresponding flags _are_ set 1111090405 M * micah hmm 1111090406 M * Bertl unification is based on IU 1111090414 M * Bertl non unified files should have iu 1111090431 M * micah ok, this is my reference server that I am unifying everything against 1111090473 J * DukeGangsta|aw ~uefgangst@p548697A0.dip.t-dialin.net 1111090494 M * Loki|muh sorry for beeing off-topic, but does anyone know if I must have enabled a option in the kernel for beeing able to communicate with some device over the serial port? screen says to me that /dev/ttyS0 is not a tty 1111090511 M * micah I am confused because unless I understand the unification process incorrectly, I should be able to update my packages on the reference server, and then re-run the unification scripts on the vservers that are being unified with the reference, right? 1111090513 M * Bertl micah: do you know _what_ command actually fails? (i.e. is it an mv or rm?) 1111090546 M * Bertl screen is the wrong tool to use ttyS0 1111090553 M * micah Bertl: I can look it fails to "rmdir/unlink" but I'll see what it specifically does 1111090563 M * Bertl Loki|muh: you want to use minicom or cu for that 1111090588 M * Loki|muh ah, will try that, thanks 1111090588 M * DukeGangsta|aw hi all 1111090594 M * Bertl micah: maybe try with strace -fF 1111090607 M * Bertl evening DukeGangsta|aw! 1111090680 M * micah Bertl: hrmmm, odd... if I run the package upgrade again, it works 1111090690 M * witchdoc hi DukeGangsta|aw 1111090772 M * DukeGangsta|aw is there a possibility to set a specific ulimit? i created a file named "ulimit" in the vserver-dir.. i filled it with "-HS -u 10000" , but the ulimit in that vserver is set to the ulimit of the host, not 10.000 like I wanted 1111090794 M * Bertl the limits are name based ... 1111090810 M * Bertl so if you want a limit for NPROC the file is called nproc 1111090827 M * Bertl but IIRC, the ulimits are not honored on 2.6 (new style config) 1111090842 M * Bertl you have to use the rlimits instead (which are per vserver) 1111090863 M * Bertl so that is similar to what -HS -u did with the nproc flag, for example 1111090918 M * Bertl DukeGangsta|aw: but apologize the dumb question, what do you plan with 10k processes? 1111090933 M * DukeGangsta|aw I'm only trying... 1111090946 M * DukeGangsta|aw testing all possibilities 1111091090 M * witchdoc if there newer alpha+utils as 0.30.203 1111091091 M * micah Bertl: since it apparantly worked the second time, I cannot tell if it was a mv or rm 1111091125 M * witchdoc ? 1111091128 M * Bertl hmm, okay ... if that issue reappears, let me know 1111091144 M * Bertl witchdoc: yes 0.30.204 is definitely out, maybe even a later version 1111091146 M * micah Bertl: heh, I'm looking through the dpkg source to see if I can figure out what it is doing 1111091150 M * micah the funny thing is the error: 1111091151 M * micah if (errno != ENOTEMPTY) /* Huh ? */ 1111091151 M * micah ohshite(_("failed to rmdir/unlink `%.255s'"),pathname); 1111091174 M * micah ohshite heh 1111091249 M * micah Bertl: it appears to do: a rmdir 1111091251 M * witchdoc thanks 1111091253 M * micah if (!rmdir(fnvb.buf) || errno == ENOENT || errno == ELOOP) continue; 1111091253 M * micah if (errno == ENOTEMPTY) { 1111091303 M * Bertl so aprerently some files were in whatever 'dir' it used, when it wanted to remove that dir ... 1111091307 M * micah "Do not report an error from rmdir () when 'errno' is ENOTEMPTY (Directory not empty) 1111091327 M * Bertl nothing which is really immutable related 1111091338 M * Bertl (especially not if it works the second time ;) 1111091364 M * micah it happened on another package 1111091371 M * micah Unpacking replacement apache ... 1111091371 M * micah dpkg: error processing /var/cache/apt/archives/apache_1.3.33-4_i386.deb (--unpack): 1111091374 M * micah failed to rmdir/unlink `//usr/lib/apache/suexec.disabled.dpkg-tmp': Operation not permitted 1111091413 Q * duckx Remote host closed the connection 1111091444 M * Bertl hmm, is that with tagxid in use? 1111091468 M * witchdoc btw - witch wiki is mean in the topic? all or a special vserver-wiki? 1111091483 M * micah Bertl: the vserver is started and I "entered" the vserver to do the package upgrade 1111091493 M * Bertl witchdoc: linux-vserver.org (in particular) 1111091503 M * micah Bertl: if thats what you mean by tagxid? 1111091528 M * Bertl no, do you use tagxid for the vserver partition (on mount)? 1111091555 M * micah Bertl: I dont believe I do, since I dont know what it is 1111091565 M * Bertl yeah, good argumentation ... 1111091589 M * micah :) 1111091606 M * micah that means, I did not set it up, so unless it is setup by default somehow, no 1111091608 Q * tchan Quit: leaving 1111091613 M * Bertl the EPERM suggests that you are missing some permnissions there ... 1111091626 M * Bertl that leaves us with: 1111091637 M * Bertl - context permissions (would require tagxid) 1111091664 M * Bertl - directory permissions (what does showattr -d tell about the directories above the files to be unlinked) 1111091666 J * tchan ~tchan@c-24-13-81-164.client.comcast.net 1111091681 M * Bertl - filesystem or host system permissions 1111091698 M * Bertl anything like selinux or grsec or whatever installed/activated? 1111091709 M * micah no, none of that 1111091729 M * Bertl okay, then let's focus on the directory permissions 1111091742 M * micah ok 1111091755 M * Bertl showattr -d /path/to/vserver//usr/lib/apache 1111091762 M * witchdoc Bertl: okay. i forgot this one 1111091799 M * DukeGangsta|aw ok, now I would like to set another ulimit -n (open files)... but if I set file ulimit "-HS -n 10000", the setting is still 1024 after vserver-restart.... in rlimit there is no setting for open files...any good tips? 1111091813 M * Bertl micah: ah, another idea, could you become a non root user inside your temple (e.g. with su - apache) and try to list one of those dirs? 1111091819 M * micah Bertl: ---bui- /vservers/seedbank//usr/lib/apache 1111091828 M * Bertl s/temple/template 1111091858 M * Bertl DukeGangsta|aw: nofile (is the rlimit for that) 1111091881 J * ntrs_ ntrs@Dardeene-68.188.50.87.charter-stl.com 1111091881 Q * ntrs Read error: Connection reset by peer 1111091896 M * micah Bertl: as a non-root user I can list /usr/lib/apache fine in that template 1111091913 M * Bertl okay, was just an idea ... 1111091940 M * Bertl what are the permissions of that dir and the dirs above? 1111091973 M * micah drwxr-xr-x 3 root root 4096 Mar 17 20:32 apache 1111091980 M * micah drwxr-xr-x 20 root root 4096 Mar 17 20:19 lib 1111091988 M * micah drwxr-xr-x 12 root root 4096 Feb 1 20:34 /usr 1111091997 M * micah drwxr-xr-x 20 root root 4096 Jan 31 17:25 / 1111092036 M * witchdoc hmmm, don't find the link for alpha+utils 0.30.204 1111092101 M * witchdoc someone have the url for me, pls 1111092144 J * duckx ~Duck@dyn-83-157-201-31.ppp.tiscali.fr 1111092281 M * micah witchdoc: it should be on 13th floor 1111092299 M * witchdoc sure ? ... hmmm 1111092322 M * Loki|muh http://www.tu-chemnitz.de/~ensc/util-vserver/alpha/ 1111092326 M * Loki|muh needed this? 1111092433 M * Bertl Loki|muh: hmm? don't understand the question ;) 1111092451 M * witchdoc yes Loki|muh thx 1111092456 M * Bertl ah, that was an answer, sorry ;) 1111092458 M * Loki|muh :) 1111092460 M * Loki|muh yeah :) 1111092505 M * Bertl micah: okay, no idea right now, maybe google a little for the errors, maybe it's a known debian issue 1111092522 M * DukeGangsta|aw hmmm.. I'm too stupid for this ;/.... what's wrong with this line: 1111092523 M * DukeGangsta|aw R245223:/usr/local/etc/vservers/gangsta/rlimits# cat /usr/local/etc/vservers/gangsta/rlimits/nofile 1111092523 M * DukeGangsta|aw 100000 80000 1111092523 M * DukeGangsta|aw R245223:/usr/local/etc/vservers/gangsta/rlimits# 1111092543 M * Bertl it contains two values? 1111092552 M * DukeGangsta|aw hard and soft in first line 1111092564 M * DukeGangsta|aw A file which contains the hard- and soft-limit of the given resource in the first line. 1111092580 M * micah Bertl: yeah, I will see if it happens in the future agani 1111092589 M * Bertl hard- and soft-limit (not limits) so one value 1111092620 M * Bertl (apllies to hard _and_ soft, otherwise use separate files) 1111092622 M * DukeGangsta|aw if it's the same value... I see.... not clear described ;) 1111092637 M * micah if I want to set eth1 to have 127.0.0.2 but keep eth0 the vserver default (so it can access the internet), I thought I would make /vserver/interfaces/1/ip with 127.0.0.2 and /vserver/interfaces/dev having eth0 and eth1 in there? 1111092656 M * Bertl okay, off for now, back later .... 1111092678 M * DukeGangsta|aw cu 1111092678 M * micah but if I do that, then eth0 *and* eth1 have 127.0.0.2 1111092678 M * Bertl micah: look in the archives, I explained the network config today (in some detail with examples) 1111092684 M * micah Bertl: ah, great 1111092690 N * Bertl Bertl_oO 1111092703 M * witchdoc good idea 1111092728 M * witchdoc enter sofa-mode on ;-) 1111092735 M * witchdoc bye folks 1111092753 Q * witchdoc Quit: n8 1111092755 M * DukeGangsta|aw anyone else here who changed the "ulimit -n " value with alpha-utils kernel 2.6 ? 1111092791 M * DukeGangsta|aw rlimits/nofile has no effekt on "ulimit -n" value 1111092948 M * micah hmm, I haven't seen Bertl_oO's message come through the list yet 1111093671 M * DukeGangsta|aw hmm. in /proc/virtual/49166/limit is the right value for open files.... hmm 1111093720 M * micah hmm, anyone know how to create two interfaces in vservers? 1111093725 M * micah I am trying every combination 1111093781 M * daniel_hozac micah: make another directory in /etc/vservers//interfaces 1111093843 M * micah daniel_hozac: the problem is I want it to have the networking that it has when there is no directory in /interfaces 1111093853 M * micah ie. the networking of the host server, plus this private network 1111093896 M * micah daniel_hozac: so if I create /interfaces/0 and put in /interfaces/0/ip 127.0.0.2 and /interfaces/0/dev has eth1 1111093906 M * micah then the vserver starts with only 127.0.0.2 networking 1111093960 M * micah if I have *nothing* in /interfaces/ then the vserver starts with the networking of the host server 1111093984 M * micah i want the combination of those 1111093986 M * daniel_hozac why do you need the 127.0.0.2 networking? 1111094015 M * micah daniel_hozac: because it is a mysql vserver and the other vservers connect to the mysql server via that network 1111094073 M * daniel_hozac do all your vservers have an empty interfaces directory? 1111094077 M * micah ie. it works better than sharing via --rbind the mysql socket 1111094100 M * daniel_hozac couldn't you just hardlink it? 1111094115 M * micah daniel_hozac: no, all my vservers do not have empty interfaces, some have specific static IPs assigned 1111094171 M * daniel_hozac couldn't you just setup 127.0.0.2 on the host? 1111094177 M * micah daniel_hozac: on what host? 1111094195 M * daniel_hozac xid 0. 1111094211 M * micah doesn't that defeat the purpose of solving the problem correctly? 1111094222 M * micah :) 1111094282 M * daniel_hozac solving the problem correctly would be to copy the host's network configuration to an interfaces directory. 1111094316 M * micah copy the host's network configuration to an interfaces directory? 1111094330 M * micah I do not see how that solves the problem 1111094365 M * daniel_hozac you'd be able to bind to the host's IP address. 1111094440 M * daniel_hozac that is the goal, no? 1111094499 M * micah yes, I want to be able to bind to the host's IP as well as create a 127.0.0.2 interface 1111094511 M * micah to test this, I have created an interfaces/0 and interfaces/1 1111094523 M * micah the 0 has the network config of the host (with a file dev having eth0) 1111094526 M * micah and the 1 has the 127.0.0.2 1111094542 M * micah with a dev of eth1 1111094549 M * micah daniel_hozac: this is what you mean, right? 1111094553 M * daniel_hozac you'd want nodev, or the vserver script will try to create the interface for you. 1111094563 M * daniel_hozac that's not what you want, as that is already set up. 1111094575 M * micah ah, in interfaces/0 I want a file called nodev? 1111094591 M * daniel_hozac yes, IIRC. 1111094608 M * micah ok, trying that 1111094648 M * micah ok, that seems to work for the host interface 1111094659 M * micah the ifconfig while in the vserver doesn't show a 127.0.0.2 address on eth1 1111094677 M * micah (it doesn't show an address on ether of eth0 or eth1) 1111094680 M * daniel_hozac how about ip addr ls outside? 1111094699 N * Doener|gone Doener 1111094763 M * micah daniel_hozac: outside, no, there is no 127.0.0.2 1111094786 M * micah it doesn't seem as if 127.0.0.2 is available 1111094919 M * micah maybe I should not have /vservers//interfaces/dev to detail the default network device? 1111094968 M * micah no, that does not seem to change anything 1111094991 M * daniel_hozac could you run vserver --debug start? 1111095034 M * micah ok 1111095079 M * micah it seems to do this: 1111095081 M * micah /usr/local/sbin/chbind --silent --ip 69.90.134.220/25 --ip 127.0.0.2 1111095104 M * micah /sbin/ip addr add 127.0.0.2 broadcast + dev eth1 1111095145 M * daniel_hozac tried adding a prefix file in interfaces/1 ? 1111095173 M * micah what prefix would I add for that? 1111095213 M * micah I guess /32? 1111095369 M * micah nothing different :( 1111095386 M * Doener micah: what exactly do you want to achieve? 1111095514 M * micah Doener: if I make an interface for my mysql vserver with the IP 127.0.0.2 then my other vservers can connect to it over the network. However, I want that vserver to also be able to access the internet. Without the 127.0.0.2 interface it uses the hosts network and can do this, but when I add 127.0.0.2 it cannot. 1111095533 M * micah does this make sense? 1111095609 Q * jsambrook Quit: Download Gaim: http://gaim.sourceforge.net/ 1111095688 M * micah I originally built the vserver as: vserver mysql build -m skeleton --hostname mysql --interface eth0:127.0.0.2 1111095719 M * micah so the shorter explanation is I want to add internet access to that 1111095755 M * Doener what's the setup for the other vservers accessing that one? 1111095939 M * micah Doener: some of them have their own IP configured, and some do not 1111095966 M * micah Doener: but all the vservers can communicate over 127.0.0 network, I assume 1111096008 M * Doener not 127/8 specific... you can always reach local addresses 1111096058 M * Doener what's the problem with adding the public ip address to the vserver? 1111096090 M * Vudumen micah: try using a private ip subnet like 10.0.0.0/8 or 192.168.0.0/16 1111096113 M * Doener i guess a simple solution (if adding the public address is a problem) is adding a private ip and doing nat 1111096113 M * Vudumen and create nat rules in iptables for these ips 1111096115 M * Doener heh :) 1111096122 M * Vudumen Doener :) 1111096183 M * micah well, I think I have something screwed up, because now when I start my mysql vserver, I do not see 127.0.0.2 1111096203 M * micah nor is it represented on the host in an ifconfig -a 1111096219 M * Doener what about "ip a"? 1111096235 M * Doener ifconfig only knows about named ip addresses 1111096321 M * micah ip a does show it... I thought it was represented in the vserver itself before, but I could be wrong 1111096341 M * Doener ip a should show it inside the vserver as well 1111096591 M * micah I dont have ip a inside the vserver 1111096721 M * micah Doener: it looks like now if I add the interfaces/1 and put the hosts networking config in there it works 1111096750 M * micah I must have been confused by the ips not showing up 1111096757 M * Doener that's how it should be :) 1111096770 M * Doener micah: you used "nodev" and not dev=eth0, right? 1111096781 M * micah yes, for the host's interface 1111096783 M * Doener otherwise a "vserver xxx stop" will kill your networking ;) 1111096787 M * Doener ok, great 1111096802 M * micah thanks to daniel_hozac who alerted me to that 1111097153 Q * tad Remote host closed the connection 1111097789 M * micah if I set up apache to do virtual hosting in a vserver, do I need to give that vserver a static IP address? 1111097860 M * micah because I set it up, and I see it listening, but I get a connection refused: 1111097860 M * micah tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 13824/apache 1111097898 M * Doener to what addresses does the vserver have access to? 1111097907 M * Doener s/what/which/ 1111097982 M * micah how can I tell? 1111098001 M * Doener well which ip addresses did you configure for that vserver? 1111098036 M * micah I configured none, so it is bound to the host's primary interface 1111098087 M * Doener it's bound to nothing then (i guess) 1111098116 M * Doener you actually see the ip addresses within the vserver, right? 1111098164 M * micah it might be easier to just assign an IP :)_ 1111102610 P * kevinp Leaving 1111102799 J * ciphernaut ~a@61.88.18.130