1112486485 M * Bertl s/../?/ ;) 1112486642 M * Venomous ok guys for me is time to sleep 1112486650 M * Bertl night Venomous! 1112486681 M * Venomous for me is a sad day, the Pope is no longer with us 1112486690 M * Venomous Rest in Peace 1112486696 M * Venomous Bye Bertl 1112486702 M * Venomous See you tomorrow 1112486713 M * complexho v2.6.11 1112486716 M * Venomous ( i have resolved all the network problem ) 1112486720 M * complexho w/1.9.5 1112486734 M * alban complexho, I have in my bookmarks: http://nfsv4.bullopensource.org/doc/testing_tools.php 1112486746 M * alban complexho, you may find some testing tools 1112486751 Q * Venomous Quit: KVIrc 3.2.0 'Realia' 1112486780 M * complexho so what is a useful test to perform? I have not really used this before 1112486787 M * complexho Thanks alban 1112486799 M * complexho That looks useful! 1112486809 M * alban I have only tested bonnie++ 1112486830 M * alban I don't know... 1112486883 M * Bertl complexho: basically everything you do in parallel is a good test 1112486894 M * Bertl 10 vservers will stress a host _in parallel_ 1112487018 M * alban iozone has an option to launch itself on several hosts 1112487021 M * alban (not sure) 1112487038 M * alban (in order to synchronize the tests) 1112487208 M * alban complexho, what do you want to do with your cluster? 1112487219 M * alban web hosting, HPC? 1112487467 M * complexho it will be used for webhosting - but all internal sites - not sold on or anything 1112487493 M * complexho well we're not planning to anyway ;) 1112487517 M * alban ok ;) 1112487529 M * complexho I need to get some vservers up... I only got it installed tonight (vserver that is) 1112487557 M * Bertl you might update to 1.9.5.6 1112487591 M * Bertl it has some instrumentation to measure block io and swapping/page faults 1112487603 M * complexho The hardware has taken 2 weeks to configure and build and I have spent a couple of weeks getting to grips with the filesystem - opengfs. The plan is to be able to dynamically startup vservers on any fc-connected host 1112487631 M * Bertl ah, that works over opengfs? 1112487643 M * complexho And then it is being load balanced from the other side using LVS 1112487659 M * Bertl did you have any issues with the vserver patch and gfs? 1112487684 M * complexho Yeah as far as I can tell opengfs seems the most suited since it is a truly shared mount at block level (well through LVM anyway) 1112487696 M * alban LVS? 1112487701 M * complexho still working through that now - not so far 1112487712 M * complexho sorry yes LVS ;) 1112487746 M * alban I don't know LVS ... Linux Virtual Server? 1112487759 M * complexho we have an interesting setup on the network side too ;) A pair of Linux/Quagga boxes running BGP and LVS to the cluster - yes Linux Virtual Server 1112487776 M * complexho we are running the vservers in pairs and load balancing them from the network side 1112487789 M * complexho so spreading eggs really 1112487825 M * complexho Eventually I would like to see if we could adapt OpenMosix and move the VServers around but we're a way off that yet... ! 1112487846 M * complexho It's part serious, part dev/play 1112487866 M * complexho (for now) 1112487866 M * Bertl mosix has no real advantage ... 1112487883 M * Bertl as long as you are bound to the I/O node 1112487920 M * complexho Does that problemn not go away with fc-attached opengfs filesystems? 1112487940 M * complexho (Like I said this is probably some ways off yet!) 1112487955 M * Bertl yeah, agreed, it's interesting though ;) 1112487971 M * complexho I will keep you posted on progress! 1112487989 M * Bertl please do so, and if possible please over the ml too ;) 1112488069 M * complexho yeah sure - we will be putting up a site with our configs in due course, there has been a lot of work just to get this far... We are working on some pretty nattly sysyadmin/firewall tools as well which we are hoping to integrate with util-vserver 1112488096 M * complexho The first is a vserver locking mechanism so that multiple hosts don't try to start the same vserver 1112488133 M * Bertl you might want to talk to the lycos folks, especially rs, they did something similar ... 1112488134 N * Snow-Man_ Snow-Man 1112488188 M * complexho I also have some old perl modules lying around that we wrote for freevsd back in 2000 that provided a nice api for creating/destroying/configuring vservers - I'm hoping to dust some of that off too 1112488218 M * complexho really - it would be interesting to speak with them - do they come in IRC? 1112488270 M * Bertl yes, rs is usually around, well, not at the weekend though ... but sometimes even then ;) 1112488303 M * complexho ok now I have my net back I will linger a little more - if you see him perhaps you would mention me? 1112488312 M * alban If I understand your config, you store your vservers files on GFS but the vservers files are used by only one node at a time? 1112488335 M * complexho yes, here is my schema atm: 1112488353 M * complexho ooh hang on rebooting! 1112488380 M * complexho basically there are two partitions being mounted: /etc/vserver and /vservers. These are mounted on all head end servers 1112488419 M * alban ah... only one volume for all vservers? 1112488432 M * complexho both file systems are openGFS. The plan is to be able to quickly failover to new hosts simply by noticing a downed vs, shooting it and bringing it up on the new host 1112488472 M * complexho Meanwhile we have LVS on the front end running feedbackd to try and orchestrate from the front 1112488491 M * Bertl so no issues with linux-vserver and gfs so far? 1112488509 M * Bertl just curious, I expected at least some initial issues 1112488513 M * complexho It's still very early days - I've only just this evening moved onto the vserver installation 1112488522 M * complexho still compiling/installing etc 1112488582 M * complexho Do you expect problems getting opengfs to compile with vserver? 1112488692 M * alban I am wondering about the pro and con between your config, and another config: not using GFS but ext3+NBD+LVM (and one volume per vserver) 1112488717 M * alban perhaps there will be less locks to manage on the cluster 1112488785 M * Bertl ext3 on LVM on NBD ;) 1112488788 M * complexho It will be interesting to try both 1112488803 M * Bertl with a local journal and swap ;) 1112488822 M * complexho Do you think that is the best way? 1112488836 M * alban I don't know at all ;) 1112488886 M * complexho seems very primitive - have not looked at nbd before 1112488935 M * complexho But much simpler perhaps... 1112488939 M * alban which version of GFS do you use? Red Hat's GFS or OpenGFS? (both are Open Source now) 1112488949 M * complexho I have downloaded opengfs 1112488999 M * Bertl I think (open)gfs has some advantages over nbd or iscsi in some situations 1112489075 M * Bertl OTOH, iSCSI and/or NBD might provide better performance ... 1112489084 M * complexho Well I have time on my side at the moment - about 3 months to try a few different options 1112489127 M * Bertl if you can make the testing somewhat public ... maybe over mailing list/wiki 1112489139 M * complexho It is definitely a good opportunity to get some interesting benchmarks 1112489144 M * Bertl I guess some folks would contribute and/or suggest stuff ... 1112489166 M * Bertl and if some linux-vserver related issues arise, we can fix them easily .. I guess 1112489184 M * complexho like I said the plan is to put up a site because there are three of us working on the overall project - quite a lot of front-end and housekeeping stuff being coded 1112489214 M * complexho Hoping to pick up where I left off a few yrs ago :) 1112489229 M * complexho With freeVSD... 1112489256 M * alban will you be on IRC theses 3 months? 1112489406 M * complexho I guess so :) I am dedicated to the project for at least this long 1112489430 M * complexho Plus a couple of others and of course a big pile of hardware on a DC :D 1112489452 M * complexho s/on/in/ 1112489490 M * alban ;) 1112489660 M * complexho And if any of this work can benefit vserver over the course of the project then I/we will be happy to accomodate 1112489771 Q * alban Quit: Leaving 1112489844 M * Bertl sounds good, feel free to 'abuse' existing infrastructure like wiki or mailing list (if you want to) 1112489859 M * complexho :) 1112489895 M * complexho Perhaps I should paste some of my work so far into a 'vserver-cluster' page 1112489916 M * Bertl won't hurt I guess ... 1112489965 Q * flock Ping timeout: 480 seconds 1112490535 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1112492934 M * complexho ok well thanks for the feedback Bertl I'm gonna go to bed now :) 1112492941 A * complexho is AFK, on another planet —I-n-v-i-s-i-o-n— 1112493330 M * micah I wonder, does anyone know if a debian 2.6.8 kernel version of the 1.9.x patch was ever made? 1112493365 M * Bertl probably .. but does it work? probably not ;) 1112493394 M * micah hehe 1112493405 M * micah that kernel for debian is the most stable/supported/secure 1112493439 M * micah I think I attempted a patch once, but it failed 1112493481 A * micah searches the arvhies 1112493483 M * micah archives 1112493567 M * micah ah, I think I got the patch-2.6.8.1-vs1.9.2.diff to apply 1112493579 M * micah but that was a while ago, and so much has been fixed 1112494295 M * Bertl yeah, it's probably almost a 2.6.11.6 kernel by now ;) 1112494728 M * micah seriously 1112494736 M * micah it looks like patch-2.6.11.5-vs1.9.5.diff doesn't apply to a vanilla 2.6.8 1112494746 M * micah which is not surpising, given its name 1112494851 M * micah I was thinking of trying it against the vanilla, and then applying each debian patch, one by one to see what I can get 1112495047 M * Bertl bad idea ... 1112495069 M * Bertl if you want to make it work somewhat .. you best do it the following way: 1112495082 M * Bertl get the splitted (broken out) 1.9.5 1112495139 M * Bertl and the previous version ... 1112495177 M * Bertl and try each split against the debian 1112495303 M * Bertl then select the one which applies, but make sure not to lose any patches 1112495309 M * Bertl s/lose/miss/ 1112495330 M * Bertl after that, upgrade the missing changes/features to recent 1112495339 M * Bertl then verify the result extensively ... 1112495487 M * micah this may be beyond my patching abilities, but I shall see how far I can get 1112495608 M * micah I've got the splitted 1.9.5, and will try each patch individually 1112495626 M * micah if one of them fails, you suggest that I try the previous version split patch? 1112495673 M * micah the first one failed pretty bad :P 1112495902 M * micah as does split-2.6.11-rc3-vs1.9.4/01_2.6.11-rc3_include.diff 1112496065 M * micah should I keep going versions back until I can find one that applies? 1112496517 M * Bertl try the 1.9.4 split for .10 and .9 (if there was one) 1112496723 M * micah there wasn't one... however split-2.6.8.1-vs1.9.2/01_2.6.8.1_base.diff applies with one Hunk #1 succeeded at 1346 (offset 36 lines), and one failure on the Makefile (the Makefile is a simple one on the EXTRAVERSION) 1112497294 M * Bertl well, it's a long way from 1.9.2 to 1.9.5 :/ 1112497338 A * micah looks off the ships bow and sees only ocean 1112497384 M * Bertl which debian kernel is now current? and can you make me a single patch against a vanilla kernel? 1112497494 M * micah the two kernels that will be shipped with Sarge, which is to be frozen any day now... are the heavily patched 2.6.8 kernel, and potentially a 2.6.11.6, that has not been determined yet 1112497522 M * Bertl yes, but it's 2.6.8-XX, right? 1112497523 M * micah I can easily make you a patch against a vanilla kernel, also, all the patches that are applied to 2.6.8 are split out into individual patches if you would find that more convenient 1112497537 M * Bertl no, I prefer a single one 1112497567 M * micah yes, it is 2.6.8-15 1112497647 M * micah do you prefer a diff -unR /usr/src/vanilla-linux-2.6.8 /usr/src/debian-2.6.8 > onebig.diff 1112497650 M * micah ? 1112498016 M * Bertl -NurpP --minimal , but yes 1112498030 M * micah mm, I meant -uNr... but I will use your options 1112498326 M * micah Bertl: https://munin.riseup.net/vanilla-vs-debian-2.6.8 1112498361 M * mikegrb munin! 1112498370 M * micah heh 1112498400 M * micah yes, munin is a great monitoring tool, I've submitted many patches and contrib plugins upstream 1112498434 M * mikegrb :D 1112498443 M * mikegrb I've got munin monitoring 44 servers 1112498455 M * mikegrb it's the greatest tool ever 1112498487 M * micah only 15 here 1112498488 M * mikegrb unfortunatly I can't give the url 1112498496 M * micah any custom plugins? 1112498511 M * mikegrb not yet 1112498516 M * micah they are really easy to write 1112498518 M * mikegrb just deployed it about a week ago 1112498520 M * mikegrb indeed 1112498529 M * mikegrb we just haven't needed any yet 1112498537 M * micah i've contemplated tying it into nagios, but nagios is such a behemoth 1112498568 M * Bertl micah: tx 1112498823 M * micah Bertl: no... thank you! 1112499415 M * Bertl you're welcome! 1112499701 M * Bertl okay, off for now .. maybe back later ... 1112499707 N * Bertl Bertl_oO 1112502689 Q * Vudumen_ Read error: No route to host 1112507179 N * Bertl_oO Bertl_zZ 1112514478 J * complexmi ~mark@funk.gotadsl.co.uk 1112514755 Q * complexho Ping timeout: 480 seconds 1112516275 Q * sannes Read error: Connection reset by peer 1112516603 J * aA-01 ~a@245.228.dial-up.xter.net 1112517370 Q * aA-01 Quit: 1112518244 J * fp- ~sean@user-1120804.dsl.mindspring.com 1112518251 M * fp- does anybody here run a hosting business? 1112518470 Q * fp- Quit: 1112518810 N * complexmi complexho 1112519528 J * Hollow ~Hollow@home.xnull.de 1112521541 J * Vudumen vudumen@perverz.hu 1112523300 J * Venomous ~kvirc@host143-154.pool8255.interbusiness.it 1112523307 M * Venomous Hi all 1112523313 N * Doener|gone Doener 1112523323 M * Doener morning folks 1112523404 M * Venomous hi Doener 1112523525 M * Venomous guys i'm running a 2.4 kernel with vserver patch, there are problem instally the grsecurity patch and the preemptive patch? 1112523587 M * Doener no idea regarding the preemptive patch, but grsec+vserver patches have been done by some folks... 1112523622 M * Doener links should be in the tools and patches section of the wiki 1112523848 M * Venomous i want try to speed up the scheduler of linux to see if i can get better performance of virtual servers 1112523914 M * Doener hm... with 2.6 that's already been done for you... 1112524165 M * Venomous yes i know 1112524169 M * Venomous i don't know 1112524182 M * Venomous probally after i will try with the low latency patch 1112524303 M * Doener AFAIK that won't give you better performance, just better interactivity... actually it might even lower performance i guess... 1112524365 M * Doener for high numbers of processes 2.4 won't beat 2.6 no matter what... O(n) sched vs. O(1) sched ;) 1112524440 M * Doener and preemption and stuff is just interesting for desktop environments 1112524526 M * Venomous well but a low latency patch speed up the scheduler, giving you a real time scheduling 1112524536 M * Venomous Never tried anyway... 1112525854 M * Venomous actually in my vmachine.conf i have IPROOT=a_public_ip how i can add more ip? 1112525937 M * Doener just did a quick read of http://www.linuxdevices.com/articles/AT8906594941.html ... the low-lat patch, as the preempt patch, allows syscalls to be preempted, it just uses pre-defined points to do so. thus both allow lower latencies when reacting to events (like your video grabbing card telling that it has data to be read). that improves interactivity, but the kernel does need extra checks and probably more context switches, thus lowering overall performanc 1112525938 M * Doener e (i guess ;)... other folks can tell better than me... 1112525948 M * Doener http://vserver.13thfloor.at/Stuff/VServer-IP-Setup-0.1.txt 1112526644 M * Venomous it says to separate it by space and use quotes 1112526649 M * Venomous doesn't work 1112526651 M * Venomous :\ 1112527013 M * Venomous "first_ip Second_IP" <-- now works :) 1112529273 N * DaPhreak DaPhreak|hates_school 1112529325 N * DaPhreak|hates_school hates_school 1112534046 J * alban ~alban@gar31-1-82-234-50-167.fbx.proxad.net 1112536303 M * alban Hello, 1112536343 M * alban "vservser vserver02 status" says my vserver is stopped, but /proc/virtual/102 is not empty 1112536348 M * alban Is it normal? 1112536493 M * Doener which vserver version? 1112536497 M * Doener and which tools? 1112536532 M * alban linux 2.6.11.6 vs1.9.5 with util-vserver-0.30.205 1112536560 M * Doener cat /proc/virtual/102/cvirt | grep threads 1112536568 M * Doener and let me know the output please 1112536569 M * alban My problem is that I can't start my vserver: http://www.pastebin.com/266492 1112536578 M * alban nr_threads: 2 1112536598 M * Doener chcontext --ctx 102 ps aux 1112536627 M * alban /usr/sbin/atd and /usr/sbin/cron 1112536654 M * Doener hm... seems they didn't get killed somehow... 1112536673 M * Doener you can kill them from the host as usual, i.e. kill 1112536740 M * alban /proc/virtual/102 does not exist any more 1112536742 M * alban thanks 1112536748 M * Doener you're welcome 1112536885 M * Doener btw, if it ever happens again, you can also use vkill to kill the whole context (i.e. all processes in that context) at once 1112536893 M * alban My problem is reproductible 1112536897 M * Doener like: vkill --xid 102 1112536908 M * alban when I run vserver vserver02 start 1112536950 M * Doener so: start -> stop -> start, BANG! ? 1112536976 M * alban so vkill --xid 102 -> start -> BANG 1112537056 M * Doener hm? question was: how do you reproduce the problem? 1112537073 M * Doener i don't think you always use vkill to stop your vserver, do you? 1112537092 M * alban I used the legacy config file and I am trying to use the new config file scheme 1112537105 M * Doener you did the config all yourself? 1112537144 M * Doener check if you got the following symlinks: 1112537148 M * alban yes, but the config seems to be incorrect 1112537198 M * Doener /etc/vservers//run -> /var/run/vservers/ (the latter path may differ, depending on how you configured the tools, IIRC default is /usr/local/var/run/vservers/ ) 1112537220 M * Doener /etc/vservers//run.rev -> /etc/vservers/.defaults/run.rev 1112537241 M * Doener /etc/vservers//vdir -> /etc/vservers/.defaults/vdirbase/ 1112537269 M * Doener you probably got the last one, but miss one or both of the first two... 1112537280 M * alban yes 1112537314 M * Doener those first two symlinks are needed by the tools to keep track of running vservers 1112537769 M * alban http://deadbeefbabe.org/paste/182 1112537788 M * alban it seems that I forget another thing 1112538787 Q * virtuoso Ping timeout: 480 seconds 1112538879 J * virtuoso ~s0t0na@80.253.205.251 1112539154 M * alban my legacy config still work, but my new config don't work yet. 1112539165 J * monrad ~monrad@213083190130.sonofon.dk 1112539502 M * Doener alban: hm, mkdir /usr/local/var/run/vservers/vserver02 ;) 1112539540 M * Doener err... s/mkdir/touch/ 1112539603 M * Doener hm, no... that file should be there by now... 1112539634 M * alban I just did a touch (the file wasn't there) 1112539644 M * Doener ok 1112539645 M * alban and stop && vkill && start 1112539652 M * alban but the problem is still there 1112539680 M * Doener which one exactly? the paste doesn't contain a 'status' after the last start 1112539761 M * alban the status says "Number of processes: 2" 1112539780 M * Doener and 'running' i guess? 1112539789 M * alban yes 1112539808 M * alban But I have still errors like 1112539809 M * Doener ok, so you should be able to stop it, without vkill, right? 1112539812 M * alban "Starting system log daemon: syslogdstart-stop-daemon: nothing in /proc - not mounted?" 1112539837 M * alban yes 1112539845 Q * virtuoso Ping timeout: 480 seconds 1112539845 M * alban "vserver vserver02 stop" works 1112539899 M * Doener yep, i saw those... probably because you have your vservers in /var/lib/vservers but didn't configure the tools to use that location... 1112539901 M * Doener vserver-info 2>&1 | grep Root 1112539923 M * alban vserver-Rootdir: /var/lib/vservers 1112539939 M * Doener hm, no that's fine... 1112539959 M * Doener ls -l /usr/local/etc/vservers/.defaults/vdirbase 1112539963 M * alban I play with util-vserver's ./configure in order to have this ;) 1112539985 M * alban gar31-1-82-234-50-167:/usr/local/etc/vservers/vserver02# ls -l /usr/local/etc/vservers/.defaults/vdirbase 1112539985 M * alban lrwxrwxrwx 1 root root 17 Apr 3 00:32 /usr/local/etc/vservers/.defaults/vdirbase -> /var/lib/vservers 1112539985 M * alban gar31-1-82-234-50-167:/usr/local/etc/vservers/vserver02# ls -l /usr/local/etc/vservers/.defaults/vdirbase/ 1112539985 M * alban total 8 1112539985 M * alban drwxr-xr-x 14 root root 4096 Apr 1 23:26 vserver01 1112539986 M * alban drwxr-xr-x 20 root root 4096 Apr 2 17:24 vserver02 1112540024 M * alban seems to be ok... 1112540029 M * Doener i just wanted the one without the slash ;) but it's fine anyway.. 1112540046 M * Doener chcontext --ctx 102 cat /proc/mounts 1112540116 M * alban http://deadbeefbabe.org/paste/183 1112540180 M * alban my vserver is stopped when I did cat /proc/mounts 1112540210 M * alban is it ok to have the same /proc/mounts in the vserver and in the host? 1112540229 M * Doener ah ok, please start it first... and use vnamespace -e 102 cat /proc/mounts (my fault ;) 1112540256 M * Doener yep, it is only virtualized if you're chroot'ed 1112540278 M * alban http://deadbeefbabe.org/paste/184 1112540282 M * alban seems to be the same 1112540311 M * alban oh .. sorry 1112540315 M * Doener no problem ;) 1112540329 M * alban I don't read your command correctly :; 1112540357 M * Doener well, on first try, i gave you the wrong one 1112540417 M * alban anyways, seems to be the same: http://deadbeefbabe.org/paste/185 1112540452 M * Doener so proc is really not mounted... 1112540466 M * Doener ah, you got no fstab there... 1112540511 M * Doener let's stop fiddling with the config and do it the 'usual' way... 1112540522 A * Doener .oO( could've thought of that earlier... ) 1112540530 M * alban what is the usual way? 1112540574 M * Doener move your hand-made config somewhere else, rename /var/lib/vservers/vserver02 to something else 1112540614 M * Doener see http://linux-vserver.org/alpha+util-vserver for examples on how to build a vserver 1112540636 M * Doener there's also a skeleton method, that just creates the config and an (almost) empty vserver directory 1112540702 M * Doener so do something like: vserver vserver02 build -m skeleton --context 102 --hostname whatever 1112540730 M * Doener you can delete the created vserver directory then and move your old vserver back in place 1112540756 M * Doener (keep the config, just the /var/lib/vservers/vserver02 thing ;) 1112540782 A * Doener .oO( hmm... guess the last wasn't more clear then the previous... ) 1112540858 M * alban ok, I am reading your page... thanks for your help ;) 1112540888 M * Doener you're welcome 1112541655 Q * shuri Read error: Connection reset by peer 1112542721 M * alban I did this: 1112542723 M * alban vserver vserver04 build -m debootstrap --hostname vserver04 --netdev eth0 --interface 10.0.0.104/24 --context 104 -- -d sarge -m ftp://ftp2.fr.debian.org/debian 1112542760 M * alban "vserver vserver04 start" works 1112542806 M * alban "vserver vserver04 enter" works :) 1112542970 M * Doener yup, that builds a new server... i used the skeleton method above, because we were aiming at creating a configuration for an existing vserver... 1112543018 M * alban ok. But I just begin, so I don't have an existing vserver which works ;) 1112543036 M * Doener ok 1112543054 M * Doener have fun with your new, first vserver then :) 1112543058 M * alban I don't have the network yet 1112543064 M * alban :D 1112543088 M * Doener guess some folks should be around that can help you with that, i've got to go now... back later (probably) 1112543092 N * Doener Doener|gone 1112544560 M * Venomous what is the skeleton method? 1112544608 M * alban Venomous, it is: "vserver build -m skeleton " 1112544664 M * alban Venomous, it create an empty vserver: 1112544686 M * alban # ls /var/lib/vservers/vserver03/ 1112544686 M * alban dev etc proc 1112544690 A * complexho is back from on another planet. I was gone for hrs 21mins —I-n-v-i-s-i-o-n— 1112544714 M * alban Hello complexho ! 1112544738 M * Venomous thnx 1112544750 M * Venomous witch problem you have with network? 1112544828 M * alban Well, my vserver has no network yet: http://deadbeefbabe.org/paste/186 1112544875 M * complexho hi alban :) 1112544919 M * alban Venomous, I try to setup it though: http://deadbeefbabe.org/paste/187 1112544923 M * alban complexho, hi ;) 1112544956 M * Venomous i have used a script 1112544970 M * Venomous i you want take a lokk 1112545004 M * alban yes 1112545032 M * Venomous http://www.marlow.dk/site.php/tech/vserver 1112545055 M * Venomous use the script deploy-vserver.sh 1112545065 M * Venomous it willsetup a vserver in few minutes 1112545176 M * alban ah, this is the script I tried yesterday ;) 1112545455 M * Venomous youdon't like it? 1112545499 M * alban I don't know... I tried another today: vserver vserver04 build -m debootstrap --hostname vserver04 --netdev eth0 --interface 10.0.0.104/24 --context 104 -- -d sarge -m ftp://ftp2.fr.debian.org/debian 1112545585 M * alban But both don't give me the network yet... I am reading the doc because I don't understand how it should be done 1112545672 N * Bertl_zZ Bertl 1112545680 M * Bertl morning folks! 1112545812 M * alban Hello Bertl ! 1112545843 M * Bertl hey alban: what do you mean by 'don#t give you the network?' 1112545887 M * alban Well, ping google.com don't work ;) 1112545893 M * alban from the vserver 1112545912 M * complexho Hi Bertl 1112545924 M * Bertl alban: kernel version? 1112545939 M * alban linux 2.6.11.6 vs1.9.5 with util-vserver-0.30.205 1112545977 M * Bertl okay, does 'ping -c 1 212.16.62.50' work? 1112546029 M * alban from my host, yes ; from my vserver no 1112546049 M * alban (1 packets transmitted, 0 received, 100% packet loss, time 0ms) 1112546051 M * cereal Hi Bertl 1112546073 M * Bertl alban: ah, okay, so your network setup on the host isn't finished? 1112546078 M * Bertl hey cereal! complexho! 1112546108 M * Bertl alban: try ping -c 1 -I 10.0.0.104 212.16.62.50 (on the host) 1112546123 M * Bertl (it will fail like from the vserver ;) 1112546141 M * alban yes, it fails 1112546162 M * Bertl do you have a public ip on the host? 1112546174 M * alban yes 1112546188 M * alban My public IP is bound to eth0 1112546193 M * alban on the host 1112546200 M * Bertl so you want to use something like: 1112546232 M * Bertl iptables -t nat -A POSTROUTING -s 10.0.0.104 -j SNAT --to-source 1112546255 M * alban I want to use NAT for my vserver + some port redirection for my sshd in the vserver 1112546384 M * alban I have this in my vserver: http://deadbeefbabe.org/paste/186 ; so I don't have 10.0.0.104 bound to any iface yet 1112546444 M * Bertl yes you ahve, just try 'ip addr ls' (inside) 1112546493 M * Bertl if you want alias (i.e. eth0:name) like ip addresses you have to specify that in the config or at creation time 1112546508 M * alban bash: ip: command not found (inside) 1112546525 M * Bertl outdated distro .. debian? yes of course! ;) 1112546547 M * alban outside (on the host), ip works and I have lo, eth0, and sit0 1112546563 M * Bertl no problem if you install iproute inside it will work too 1112546605 M * Bertl you can also change the creation part: 1112546606 M * Bertl --netdev eth0 --interface 10.0.0.104/24 1112546608 M * Bertl to 1112546621 M * Bertl --interface hansi=eth0:10.0.0.104/24 1112546636 M * Bertl then you'll get eth0:hansi (which will be seen by ifconfig too) 1112546638 M * alban (ip show my public ip and 10.0.0.104 on my host) 1112546664 M * Bertl as I said, (and demonstrated) your network is working fine ... 1112546670 M * alban ok ;) 1112546705 M * Bertl if you want to reach the outside, you have to SNAT outgoing packets to your public ip ... that's not different with linux-vserver from any other linux system 1112546818 M * alban I don't understand why my hosts have the IP 10.0.0.104. This is the IP I set for my vservern not for my host 1112546827 M * alban s/hosts/host/ 1112546863 M * Bertl linux-vserver (without ngnet) does not use virtualized networking, it uses shared networking with some restrictions 1112546882 M * Bertl so _everything_ you configure (regarding networking) for your vservers happens on the host 1112546903 M * Bertl the only vserver related part is a restriction to the selected IPs 1112546925 M * Bertl for example your guest can only send packets with src=10.0.0.104. 1112546966 M * alban So, if I have several vservers with 10.0.0.101, ..., 10.0.0.109,, my hosts will have 9 IP adress bound to eth0 1112546975 M * Bertl precisely 1112547023 M * Bertl of course you can bind them to any interface, even dummy0 for example 1112547177 M * alban Do I have to echo 1 > /proc/sys/net/ipv4/ip_forward ? 1112547335 M * Bertl nope, no forwarding involved 1112547343 M * Bertl it all happens _on_ the host ... 1112548089 J * Scytale ~Scytale@cmung291.cmu.carnet.hr 1112548102 M * Bertl welcome Scytale! 1112548129 M * Scytale hi 1112548270 P * Scytale Leaving 1112548677 M * alban I've named my iface. Now, ifconfig in the vserver shows eth0 and eth0:vs4 1112548689 M * Bertl exactly 1112548840 M * alban And I tried iptables -t nat -A POSTROUTING -s 10.0.0.104 -j SNAT --to-source $MY_PUBLIC_IP 1112549027 M * alban But my rule is not created: http://deadbeefbabe.org/paste/190 1112549065 M * Bertl try iptables -t nat -L -n -v 1112549109 M * alban oh ;) I have a lot of rules from my previous tests ;) 1112549126 M * Bertl iptables -t nat -F POSTROUTING ;) 1112549308 M * alban yeah !! 1112549310 M * alban it works ;) 1112549324 M * alban thank you 1112549329 M * Bertl of course! ... you're welcome! 1112550507 M * Bertl back later ... 1112550516 N * Bertl Bertl_oO 1112551464 Q * hates_school Quit: leaving 1112552033 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1112553446 M * alban My vserver works, have internet connection, and the internet can ssh on it ;) 1112553538 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1112554181 J * RedFox ~RedFox@ABayonne-152-1-69-140.w83-200.abo.wanadoo.fr 1112554185 M * RedFox hi 1112554273 M * RedFox Debian Sarge now have util-vserver 0.30.205. Is there a good documentation about the new configuration scheme? 1112554290 M * RedFox Something other than http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1112554542 M * alban RedFox, I don't know 1112554561 M * alban RedFox, but the conf can be set up automatically with: 1112554572 M * alban vserver vserver04 build -m debootstrap --hostname vserver04 --netdev eth0 --interface 10.0.0.104/24 --context 104 -- -d sarge -m ftp://ftp2.fr.debian.org/debian 1112554576 M * alban (for example) 1112554585 M * alban thanks you for the info on sarge 1112554607 M * alban since when does sarge have util-vserver 0.30.205 ? 1112554628 N * Bertl_oO Bertl 1112554634 M * RedFox I already have vserver, but configured in the old way 1112554647 M * Bertl RedFox: what kind of documentation would you expect? 1112554652 M * alban RedFox, my command will erase your vserver :( 1112554657 M * RedFox I've found this conversion script on the list http://list.linux-vserver.org/archive/vserver/msg09274.html 1112554812 M * RedFox I look for a doc that explain which settings are mandatory, optional... which settings are symlinks, directory, file. how the file are formatted...some examples... 1112554818 M * alban RedFox, my sarge mirror have util-vserver 0.30.204-4 1112554834 J * keyser_soze ~keyser@host201.201-252-39.telecom.net.ar 1112554849 M * erwan_ho alban: \o/ 1112554870 M * erwan_ho hey Bertl 1112554877 M * Bertl evening erwan! 1112554885 M * RedFox alban, heuu right :) 0.30.204-4 1112554900 M * alban erwan_ho, hello, are you THE erwan ? 1112554905 M * erwan_ho alban: ;) 1112554944 M * erwan_ho alban: but I mean you better go to bed, tomorrow GFS/Lustre will kick your mind :) 1112554958 A * erwan_ho flies to its bed 1112554963 M * erwan_ho see you guys 1112554969 M * Bertl RedFox: I do not know of such docu ... 1112554972 M * erwan_ho have a nice evenning 1112554988 M * Bertl RedFox: but you can answer some questions easily by yourself ... 1112555027 M * alban erwan_ho, do you know newz2000 ? He have a nice cluster with GFS + vservers 1112555042 M * Bertl RedFox: for example if you build a vserver with 'vserver mandatory build -m debootstrap -- -d sarge 1112555054 M * erwan_ho alban: newsz2000 ? 1112555069 M * alban erwan_ho, newz2000, here, on this channel 1112555090 M * alban But he is away I guess 1112555092 M * erwan_ho alban: oh I don't ... 1112555112 M * Bertl RedFox: this will give you and example and a somewhat minimal list of entries ... maybe not all mandatory ... 1112555130 M * RedFox ok 1112555131 M * alban erwan_ho, I'm sure you will be interested by his setup 1112555150 M * erwan_ho alban: thx 1112555167 M * Bertl RedFox: OTOH, a build with all the options mentioned on the alpha util-vserver example page, will give you examples for the files 1112555180 A * erwan_ho is listenning the "bed song" .... "coooooooome tooooooooo sleeeeeeeep" :) 1112555192 Q * Snow-Man Read error: Operation timed out 1112555313 M * Bertl RedFox: http://linux-vserver.org/alpha+util-vserver, just add '--flags nproc,lock --initstyle minit --interface a=dummy0:10.0.0.1 --interface b=dummy0:10.0.0.2/24 --hostname zaphod --context 42' 1112555342 M * Bertl that will give you a good example of configuration ... 1112555390 M * Bertl maybe --interface dummy0:10.0.0.1 --interface a=dummy0:10.0.0.2 --interface 10.0.0.3 are good options too ... 1112555409 M * Bertl (for testing various configurations) 1112556271 M * erwan_ho nenuit alban 1112556595 Q * keyser_soze Quit: Abandonando 1112556770 Q * erwan_ho Remote host closed the connection 1112556857 Q * alban Remote host closed the connection 1112558028 M * complexho evening all (back to work then) 1112559057 J * Snow-Man ~sfrost@snowman.net 1112560143 J * Tbery ~tb@pha-84-242-127-4.nat.karneval.cz 1112560228 M * Bertl evening Snow-Man! Tbery! 1112560307 M * Tbery HI 1112560317 M * Tbery Bertl Hi 1112560326 M * Tbery Bertl Im make new vservers.. 1112560345 M * Tbery Bertl but I have some trobles with set up ip.. 1112560351 M * Bertl hmm yeah? 1112560370 M * Tbery I set it but doesnt works... 1112560384 M * Bertl how do you set it? 1112560396 M * Tbery Yes.. 1112560423 M * Tbery I was inhabit on old version.. 1112560431 M * Tbery but with new is more harder.. 1112560448 M * Bertl well, how do you set it? ;) 1112560475 M * Tbery before with old packages.. 1112560478 M * Tbery S_HOSTNAME="klobouk2" 1112560478 M * Tbery IPROOT="192.168.0.100/255.255.255.0 lo:127.0.0.1/255.0.0.0" 1112560478 M * Tbery IPROOTDEV="eth0:2" 1112560478 M * Tbery ONBOOT="no" 1112560478 M * Tbery S_NICE="" 1112560479 M * Tbery S_FLAGS="lock nproc" 1112560485 M * Tbery ULIMIT="-H -n 1024" 1112560487 M * Tbery S_CAPS="CAP_NET_RAW" 1112560488 M * Tbery # *NOT* DNS domain name, for NIS only 1112560490 M * Tbery S_DOMAINNAME="" 1112560492 M * Tbery like this 1112560494 M * Tbery how to make same with new version?? 1112560500 M * Tbery all was in one file 1112560505 M * Bertl looks pretty broken, and this did work somehow? 1112560510 M * Tbery why now is in 6 files.. 1112560521 M * Bertl IPROOTDEV="eth0:2" <-- wrong 1112560539 M * Tbery but this worked.. 1112560549 M * Tbery well.. 1112560552 M * Bertl yeah, well, probably you were lucky ;) 1112560603 M * Bertl --interface 2=eth0:192.168.0.100/24 --interface lo:127.0.0.1/8 would be the equivalent 1112560623 M * Bertl but that would probably be wrong too, as you do not want to create 127.0.0.1 1112560636 M * Bertl so an appropriate setup might look like this: 1112560656 M * Bertl --interface eth0:192.168.0.100/24 1112560667 M * Tbery with new packaged how to reconfihute?? 1112560683 N * DaPhreak DaPhreak|school 1112560691 M * Bertl that's a missing feature for now (reconfiguration) 1112560694 M * Tbery vserver name bult -- 1112560698 M * Tbery atc.. 1112560700 M * Tbery ect.. 1112560842 M * Bertl but you can create a new one with -m skeleton and copy the old vserver to the new dir 1112560848 N * Doener|gone Doener 1112560849 M * Tbery in /etc/vservers/name/interfaces/ 1112560855 M * Tbery setup it?? 1112560860 M * Bertl yep, there you can edit it too 1112560861 M * Doener evening! 1112560865 M * Bertl evening Doener! 1112560889 M * Tbery in dev I have only eth0:namevserver 1112560895 M * Doener hm, wasn't there a old->new conversion thingy on the ml lately? 1112560900 M * Tbery if I want lo 1112560904 M * Bertl Tbery: which is wrong ... 1112560923 M * Bertl for the setup I mentioned: --interface eth0:192.168.0.100/24 1112560927 M * Bertl you want to have: 1112560932 M * Tbery if I run vserver ip does not set up.. 1112560944 M * Bertl /etc/vservers/name/interfaces/0/dev with eth0 1112560956 M * Bertl /etc/vservers/name/interfaces/0/ip with 192.168.0.100 1112560963 M * Bertl /etc/vservers/name/interfaces/0/prefix with 24 1112560988 M * Tbery dev I have in /interfaces/dev 1112560996 M * Bertl if you want a named alias like namevserver 1112560998 M * Tbery not interfaces/0/dev/ 1112561002 M * Bertl then you also have to make 1112561014 M * Bertl /etc/vservers/name/interfaces/0/name with namevserver 1112561070 M * RedFox Is there a way to bind the vserver to ppp0 without specifying the ip ? 1112561088 M * Bertl RedFox: no, because linux-vserver is solely ip based 1112561093 M * RedFox :(( 1112561110 M * Bertl but no need to do so actually ... 1112561144 M * RedFox My vservers at home are behind an ADSL connection with dynamic IP 1112561175 M * Bertl ADSL via ppp? 1112561186 M * RedFox pppoe 1112561199 M * RedFox PPP over Ethernet 1112561199 M * Bertl well, anyway, you assign an ip like 10.0.0.1 to your vserver 1112561228 M * Tbery WARNING: '/etc/vservers/klobouk/interfaces/0/name' does not end on newline 1112561234 M * Bertl RedFox: and whenever the ppp comes up (new ip7etc) you do add a NAT entry to iptables 1112561259 M * Bertl Tbery: line without a newline? 1112561265 M * Tbery after run vserver in ifconfgi missing virtualip.. 1112561285 M * Bertl don't use ifconfig unless you have a named alias, use 'ip addr ls' 1112561289 M * RedFox Bertl: ho yes, good suggestion :) 1112561331 M * RedFox Like a virtual ZDM :) 1112561352 M * Bertl you're from france or just clumsy fingers like me? 1112561363 M * Tbery is threre.. 1112561441 J * kalou ~kalou@193.253.182.140 1112561444 M * kalou Hello 1112561447 M * Bertl evening kalou! 1112561451 M * Tbery name [----] 0 L:[ 1+ 0 1/ 1] *(0 / 7b)= k 107 0x6B 1112561452 M * Tbery klobouk 1112561466 M * kalou Everything fine out there ? 1112561474 M * Tbery that is /interface/0/name 1112561490 M * RedFox kalou, salut 1112561497 M * Bertl Tbery: which means? 1112561509 M * Tbery klobouk is name of vserver 1112561545 M * Bertl try xxd .../interface/0/name 1112561677 J * redLED redled@d54C2CD58.access.telenet.be 1112561689 M * redLED hello :) 1112561690 M * Bertl welcome redLED! 1112561704 M * redLED how are you all hangin' on? 1112561733 M * Bertl kalou, redLED: so far so good ... 1112561747 M * redLED very well :) 1112561748 M * redLED hmm 1112561761 M * redLED i get some strangeness trying to build the latest alpha-util 1112561764 M * redLED src/rpm-fake.c: At top level: 1112561765 M * redLED src/rpm-fake.c:628: error: syntax error before '{' token 1112561765 M * redLED src/rpm-fake.c:585: warning: 'removeNamespaceMounts' defined but not used 1112561788 M * redLED get this both from 214 and 213 1112561806 M * daniel_hozac 214? 213? 1112561825 M * Bertl yeah, it's getting interesting ;) 1112561825 M * redLED 0.29.214 1112561829 M * redLED and 0.29.213 1112561838 M * Bertl latest, eh? 1112561844 M * Bertl 0.30.205 :)) 1112561856 M * redLED ... 1112561859 M * redLED *checks own head* 1112561868 M * redLED right. 1112561871 M * Tbery Bertl, try to help me?? where can I set cap-s ?? 1112561887 M * Bertl which caps do you want to set? 1112561898 A * redLED takes mental note not to pick out a tarball by the highest minor version 1112561995 M * Tbery reboot on setup time broadcast 1112562035 M * Bertl we do not have such cap(s)?! ... 1112562051 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1112562070 M * Bertl and 1112562071 M * Bertl http://savannah.nongnu.org/cgi-bin/viewcvs/util-vserver/util-vserver/lib/ccaps-v13.c?rev=HEAD 1112562082 M * Bertl (those are the ccaps) 1112562089 M * redLED yeah 1112562094 M * redLED was trying to build wrong version 1112562095 M * redLED mmm 1112562097 M * redLED sorry for that :) 1112562102 M * Bertl np 1112562107 M * Bertl (see topic ;) 1112562399 M * Tbery I will use older packages.. 1112562403 M * Tbery :-// 1112562412 M * Tbery are easiers... 1112562416 M * Bertl well, if you prefer ... 1112562741 M * Tbery I hope that will continue If I make now vservr with oldconf will able to use with new packages,... in future.. 1112562755 M * Tbery like with apache or samba ect... 1112562882 M * Tbery with one conf go on.. 1112562906 M * Tbery Bertl, would you like to come on Linux expo?? 1112562954 M * Bertl well, when is Linux expo? 1112562966 M * Bertl (and more important where) 1112562996 M * Tbery http://linuxexpo.cz/?lang=EN 1112563019 M * Tbery from vien You will here in flew hours.. 1112563102 M * Bertl what means 'dubna' ? 1112563199 M * Tbery April.. 1112563206 M * Tbery 11.4.2005 1112563213 M * Bertl hmm, quite soon ... 1112563269 M * Tbery yes.. 1112563274 M * Tbery realy/.. 1112563310 M * Bertl looks somewhat focused on commercial linux use ... 1112563328 M * Tbery Like this.. 1112563473 M * Doener Bertl: i just (finally ;) realized why the if(!vxi) check in init_vx_info was so damn stupid ;) 1112563490 M * Bertl ah, sorry, I could have explained it earlier ;) 1112563510 M * Bertl thought you just missed it at the first patch ... 1112563567 M * Doener no, i actually thought it was necessary back then... though it's soooo obvious... 1112563589 M * Bertl hey, read the qvm86 mail? *G* 1112563618 J * RedFox_ ~RedFox@ABayonne-152-1-54-17.w83-193.abo.wanadoo.fr 1112563620 M * Doener guess my weird thoughts on the (non existent) race conditions broke my brain ;) 1112563621 M * Doener yep 1112563636 M * Bertl excellent news, eh? 1112563641 M * Doener yep 1112563649 M * Bertl (well, of course could have been better, though) 1112563673 A * Doener .oO( i need to say something else than 'yep' this time... ) 1112563679 M * Doener hmm... of course! 1112563720 M * Bertl ah, well, I can rephrase the questions to have alternating expectations (you can then go yep, nope, yep, nope ... ;) 1112563724 M * Doener twice the work... 1112563738 M * Doener ... regarding kqemu/qvm86 1112563743 M * Doener not the yep/nope :D 1112563744 M * Bertl yep 1112563753 M * Doener *lol* 1112563790 M * Bertl so did you look at the getsid() issue yet? (nope expected ;) 1112563824 M * Doener just a lookup to confirm my memories on what getsid() does 1112563901 M * Bertl btw, we should make the clr_vx_info and init_vx_info behave similar regarding logging/history 1112563979 M * Bertl not sure if we want the 0 assignments recorded or not ... 1112564003 M * Doener 0 assigments? 1112564011 M * Doener s/g/gn/ 1112564020 Q * RedFox Ping timeout: 480 seconds 1112564050 M * Bertl init_vx_info now logs/records vxi=NULL events 1112564059 M * Bertl clr_vx_info does not do so ... 1112564092 M * Doener ah, right... 1112564196 M * Doener guess we don't want them... preparing patch... 1112564230 M * Bertl yeah, guess that's the best, just move the print/history into the if() check 1112564303 M * Doener hm... no 1112564315 M * Doener actually the current __init_vx_info is broken 1112564339 M * Doener because we don't set *vxp to zero... 1112564340 M * Bertl in what way? 1112564376 M * Bertl vxlprintk(VXD_CBIT(xid, 3), "init_vx_info(%p[#%d.%d])", 1112564376 M * Bertl vxi, vxi?vxi->vx_id:0, 1112564376 M * Bertl vxi?atomic_read(&vxi->vx_usecnt):0, 1112564376 M * Bertl _file, _line); 1112564376 M * Bertl vxh_init_vx_info(vxi, vxp); 1112564378 M * Bertl if (vxi) 1112564381 M * Bertl atomic_inc(&vxi->vx_usecnt); 1112564383 M * Bertl *vxp = vxi; 1112564387 M * Doener argh... 1112564401 M * Bertl if vxi == NULL, we'll log, record and set to NULL 1112564403 M * Doener messed up indenting in my vim... 1112564429 M * Doener hm, what about: 1112564438 M * Doener if(!(*vxp = vxi)) 1112564440 M * Doener return; 1112564451 M * Doener at the top of the function... 1112564456 M * Bertl no good idea ;) 1112564471 M * Bertl vxi->vx_usecnt isn't incremented yet ;) 1112564487 M * Bertl working on your race conditions, eh? 1112564522 M * Doener erhm? if vxi is NULL we do not want to increase that... really ;) 1112564541 M * Bertl yeah, but if it isn't when do we increment the count? 1112564559 M * Bertl (choices: before or afterwards ;) 1112564564 M * Doener order is: the above, then log, then inc 1112564574 M * Bertl which is plain wrong 1112564583 M * Bertl we have to do the increment before ... 1112564628 M * Doener where's a wall? i got a head to bang against it... 1112564649 M * Bertl /usr/bin/wall *G* 1112564760 M * Bertl something like this should be fine: 1112564761 M * Bertl if (vxi) { 1112564761 M * Bertl vxlprintk(VXD_CBIT(xid, 3), 1112564761 M * Bertl vxi, vxi?vxi->vx_id:0, 1112564763 M * Bertl vxi?atomic_read(&vxi->vx_usecnt):0, 1112564766 M * Bertl _file, _line); 1112564769 M * Bertl vxh_init_vx_info(vxi, vxp); 1112564771 M * Bertl atomic_inc(&vxi->vx_usecnt); 1112564774 M * Bertl } 1112564776 M * Bertl *vxp = vxi; 1112564787 M * Bertl ahem, sec (damn TAB in irssi ;) 1112564797 M * Bertl if (vxi) { 1112564797 M * Bertl vxlprintk(VXD_CBIT(xid, 3), 1112564797 M * Bertl "init_vx_info(%p[#%d.%d])", 1112564797 M * Bertl vxi, vxi?vxi->vx_id:0, 1112564797 M * Bertl vxi?atomic_read(&vxi->vx_usecnt):0, 1112564800 M * Bertl _file, _line); 1112564802 M * Bertl vxh_init_vx_info(vxi, vxp); 1112564805 M * Bertl atomic_inc(&vxi->vx_usecnt); 1112564807 M * Bertl } 1112564810 M * Bertl *vxp = vxi; 1112564884 M * Doener yep... 1112564984 M * Doener guess you have a patch ready then? 1112565016 M * Bertl hmm, nope ;) 1112565030 M * Bertl but I can do one, of course ;) 1112565159 Q * Venomous Ping timeout: 480 seconds 1112565732 M * Tbery Bertl, how to rebult vprounhide?? 1112565748 M * Tbery /etc/init.d/vprocunhide start 1112565748 M * Tbery Fixing /proc entries visibility.... 1112565797 M * Tbery is ok now.. 1112565800 M * Tbery soryy.. 1112565811 M * Bertl np 1112566237 M * Bertl k, leaving now .. back later ... 1112566250 N * Bertl Bertl_oO 1112567143 Q * kalou Quit: leaving 1112567278 Q * Tbery Quit: Ukončuji 1112567415 Q * monrad Ping timeout: 480 seconds 1112567640 P * RedFox_ Leaving 1112568342 J * ciphernaut ~a@61.88.18.130 1112569830 J * Slidey slidah@217-112-86-5.adsl.customer.flairnet.co.uk 1112569918 M * Slidey hi, can anyone tell me if i'd be going in the wrong direction by using the freevps.com vserver stuff? a customer has asked for that to be installed but having done a lil research ive seen logs of people saying its out of date (and the kernel they suggest is 2.4.21) 1112570016 M * Slidey i think it appealed to him as theres an rpm kernel for rhel 1112570052 Q * albeiro Remote host closed the connection 1112570169 M * Doener the freevps forked from linux-vserver a long time ago... all i actually know about it, is that it is (or was?) focussed on rhel... 1112570177 M * Slidey aha 1112570186 M * Slidey ta 1112570204 M * Doener and they still didn't make the move to 2.6? thought rhel had switched over to 2.6 some time ago 1112570210 M * Slidey nah they're 2.4 1112570228 M * Slidey the box is a centos 3.3 machine, which i believe is 2.4 1112570237 M * Slidey but an upgrade to 2.6 shouldnt be an issue anyway 1112570292 M * Doener yep, centos 4 is the first in the series to include an 2.6 kernel 1112570310 M * Slidey havent installed 4 yet 1112570323 M * Slidey anyway 1112570324 M * Slidey cheers for the info 1112570328 Q * Slidey Quit: Leaving 1112570894 N * Bertl_oO Bertl 1112570899 M * Doener wb Bertl 1112570903 M * Bertl thanks! 1112570938 P * redLED Leaving 1112571262 M * Bertl Doener: so time for the sid issue? 1112571291 M * Doener yep 1112571324 M * Bertl I was wondering iwhat we could/should do about session ids for the init process ... 1112571345 M * Bertl IMHO we have 3 options there ... 1112571368 M * Bertl a) just 'virtualize' the session id away (in sys_getsid()) for initpid processes 1112571396 M * Bertl b) remove/free the session at vserver start (for the initpid) and set the session to 0 1112571443 M * Bertl c) add some syscall command to start an init (or something else) _inisde_ a vserver, from outside (kernel) 1112571584 M * Doener hm, c) seems complicated... (on first look) 1112571658 M * Bertl yeah, but it 'might' allow some fancy stuff, like running the init over a 'virtual' console 1112571836 M * Bertl not saying that this is the way to go, tough ;) 1112571867 M * Bertl *though 1112571867 M * Doener hmm... about b)... any chance that this may cause trouble? 1112571896 M * Bertl yeah, maybe, no idea ... didn't try anything with sessions yet 1112571937 M * Bertl can't even say that I understand why we need 4 different pid types anyway ... 1112571941 M * Doener hm, can one list sids with ps? 1112571962 M * Doener 4? pid, sid, tgid and? 1112571970 M * Bertl after all we have pid, tid, gid and sid 1112572096 M * Bertl and to answer your question, there seems to be a --sid option to ps 1112572129 M * Doener yeah, that lists the processes belonging to the giving session 1112572374 M * Doener anyway, i'll try to find some information on sids...