1168041629 M * daniel_hozac rpm -q --provides iproute2 1168041677 M * nayco # rpm -q --provides iproute2 1168041677 M * nayco iproute2 = 2.6.16-3mdk 1168041757 M * daniel_hozac so i guess that's expected, the packages are named differently. 1168041769 M * daniel_hozac you'll have to fix the BuildRequires in the specfile. 1168041787 M * nayco ahem... How can I do that ? 1168041810 M * nayco tar x + edit + tar c ? 1168041822 M * daniel_hozac easier to just put it in the right place. 1168041831 M * nayco ? 1168041869 M * daniel_hozac put the spec in /usr/src/*/SPECS, and the tarball in /usr/src/*/SOURCES, and then run rpmbuild -ba /usr/src/*/SPECS/util-vserver.spec 1168041900 M * nayco and before that, I untar to get the spec file, right ? 1168041904 M * daniel_hozac right. 1168041993 M * nayco Mmmm, and could the "official" tarball be fixed anyway for other mandriva users, or is it iimpossible to ship a Mandriva specfile inside it (and still be compatible with other distros) ?? 1168042130 M * daniel_hozac is there a reason there is no mandriva repository? 1168042143 M * daniel_hozac or mandriva maintainer, or whatever. 1168042151 A * daniel_hozac has no idea how mandriva works. 1168042251 M * nayco Mmmm... I don't even know who maintains the util-vserver package I found in mandriva"s contribs 1168042306 J * gagaaa ~aa@213.91.167.179 1168042332 M * daniel_hozac no changelog? 1168042345 M * nayco Yess !!! It builds \o/ => My first specfile editing ;-) 1168042352 M * nayco daniel_hozac: Wait... 1168042354 P * gagaaa 1168042486 M * nayco daniel_hozac: * Sat Feb 19 2005 Erwan Velu - 0.30.204-1mdk 1168042494 M * nayco This is old... 1168042503 M * daniel_hozac well yes, 0.30.204 is ancient. 1168042625 M * nayco If I understant it all, the package I'm building now will generate a SRPM I can submit to mandriva contribs, huh ? I mean, will this SRPM be usable by other (and mandriva) to build the RPMS ? 1168042707 Q * dna Quit: Verlassend 1168042962 M * daniel_hozac sure. 1168043193 M * nayco Mmmm, I gonna ask some cooker contributors. Oh, and I just installed the utils, it went fine. Now, I just need to reboot on my new kernel ;-) Thanx a lot, daniel, I learnt tonight :) ! 1168043209 M * daniel_hozac you're welcome! 1168043918 M * nayco ok, reboot 1168043920 Q * nayco Quit: Bonne nuit ! 1168044239 M * Borg- [root@jumpbox-adm] iptables -A INPUT -i lo -m state --state RELATED -j ACCEPT 1168044240 M * Borg- iptables: Unknown error 4294967295 1168044243 M * Borg- guys.. any clues? 1168044251 M * Borg- misconfigured kernel or what? ;) 1168044273 M * daniel_hozac 64-bit anything? 1168044292 M * Borg- nah.. simple 32bit PIV box 1168044300 M * Borg- Linux 2.6.17.13-vs2.0.2.1 1168044342 M * daniel_hozac did you compile iptables yourself? 1168044349 M * Borg- yes 1168044358 M * Borg- after installing kernel 1168044417 M * daniel_hozac very strange. 1168044420 M * daniel_hozac that's -1. 1168044477 M * Borg- yep.. 1168044515 M * Borg- probably some misconfiguration I guess ;) 2.6 different in configuring netfilter from 2.4 a bit.. 1168044521 M * Borg- Ill check that in moday. 1168044523 M * Borg- monday 1168044547 M * yarihm uh guys, i screwed up with routing here ... are there examples somewhere to help me orientate? my server has two interfaces in different networks (on the same switch though). one network is for the vservers and an alias of the server (for in-kernel-nfs) and the other network is for management of the host. I do something like this on each vservers startup: 1168044549 M * yarihm ip rule add from 10.0.5.18 table vservers 1168044549 M * yarihm ip route flush cache 1168044577 M * yarihm where the routing-table vservers contains something like: ip route add default 10.0.5.1 dev eth0 table vservers 1168044624 M * yarihm ok, now the problem is this: if i connect to the vserver on ip 10.0.5.18, it will send its answer to 10.0.5.1 instead of say 10.0.5.20 (my client) 1168044716 M * yarihm it is as if i would have to define something like this: ip rule add from 10.0.5.18 to 10.0.5.0/24 "send this directly, stupid" because there is no routing needed to reach 10.0.5.20 1168044773 Q * gerrit Ping timeout: 480 seconds 1168044805 M * Borg- yarihm: are you sure you want to use ip rules? this is for policy routing.. 1168044846 M * yarihm Borg-: well, it is the way i was told to do it here ... worked fine for quite some time, but now i seem to have borked something 1168044870 M * yarihm Borg-: i mean you can do all sorts of routing with ip rules, no? 1168044885 M * Borg- yeah. but in rule you specifi routing table to use 1168044888 M * Borg- not routes itself 1168044949 M * Borg- ip route add 10.0.5.0/24 via eth1 table 50 1168044961 M * Borg- ip rule add from 10.0.5.18 lookup 50 1168044964 M * Borg- but its ugly hack. 1168044983 M * yarihm Borg-: what would be less ugly? 1168044996 M * Borg- everything should work fine w/o hacks unless you have 2 NICs in same networks.. 1168045077 M * Borg- dunno what setup you have and what you want to achieve 1168045098 M * Borg- I have bouch of vservers w/ multiple NICs.. everything works fine w/o policy routing 1168045105 M * yarihm Borg-: are you quite sure? i remember seeing things like seeing ssh packets that are supposed to leave on the management-nic and be routed over a gateway being sent over the other nic that directly connects to the client's network without going over the gateway (and by this resulting in asymetric routing somehow) 1168045163 M * yarihm (which is kind of obvious as the kernel just takes the shortest path available) 1168045195 M * Borg- yarihm: absolutly.. using 2.6.17.13-v2.0.2.1 1168045224 M * yarihm 2.6.19.1-vs2.2.0-rc6 here ... so that should work ... 1168045231 M * Borg- so im quite suprised. 1168045256 M * yarihm well, i'm probably scratching a non-itching itch here ... i'll take out all special routing and have a look on what happens 1168045275 M * Borg- good idea.. 1168045281 M * Borg- okey.. good luck.. time to take a rest for me 1168045303 M * Borg- nite 1168045399 J * gerrit ~gerrit@bi01p1.co.us.ibm.com 1168045409 M * Bertl wb gerrit 1168045444 M * gerrit rehi bertl. just bouncing from network to network here. ;) 1168045485 Q * bandsaw Quit: The computer fell asleep 1168045867 M * yarihm Borg-: ah, now i remember ... the itch was not quite that non-itching ... 1168046038 M * yarihm Borg-: somehow i have to tell the vservers what their gateway is, no? the host itself has a gateway via the management-network, but that one isn't accessible for the vservers, so they have to get a different default gw somehow 1168046041 Q * gerrit Ping timeout: 480 seconds 1168046067 M * Bertl yarihm: there is no per guest networking (yet) 1168046085 M * Bertl yarihm: so you setup _all_ the routes on the host 1168046104 M * yarihm Bertl: of course ... but that's where i get into trouble at the moment 1168046106 M * Bertl those could of course use different gateways, no problem with that on linux :) 1168046158 M * Bertl http://archives.linux-vserver.org/200311/0470.html 1168046175 M * yarihm yes, i know, i remember you told me how to do that (define a table in /etc/iproute2/rt_tables and then do something like ip route add vserver-gateway table vservers ; ip rule add from vserver-ip table vservers ... 1168046178 M * yarihm ) 1168046227 M * Bertl good, that plus an ip rule add ... will do the trick 1168046291 M * yarihm what ip rule add ... ? 1168046334 M * Bertl e.g. ip rule add from table vservers 1168046345 M * yarihm Bertl: i posted a description of my problem above (at around 1:50) ... if you happen to have any hints ... 1168046369 M * yarihm ah, yes, i wrote that above as you can see :) "ip rule add from vserver-ip table vservers ... should have written $vserver-ip i guess 1168046400 M * Bertl so, what happens if you now do: 1168046412 M * Bertl ping -I www.google.com ? 1168046417 M * yarihm but anyway, i guess my setup suffers from the fact that i have only one physical network at the moment, i gotta change that to get more clearness i guess 1168046433 M * Bertl check with tcpdump -vvnei icmp 1168046451 M * yarihm Bertl: i have to reverse the setup above again as i removed all ip route-stuff as suggested by borg- :) one moment please, that will take some time 1168046872 M * yarihm Bertl: well, the ping works, should i still use tcpdump to have a closer look? i guess there won't be any surprises 1168046885 M * Bertl sure, you want to check the gateway 1168047122 M * yarihm Bertl: well, with the given options (and some additional filters to get out the noise) i can't see directly what gateway is being used, but telling from the mac it seems as if it was the correct one ... 1168047188 M * Bertl hum, what noise? 1168047191 M * yarihm however, what is strange is this: i can enter the vserver from the host via vserver foo enter and then ping a client in the vservers network 1168047221 M * Bertl you have that much icmp traffic? 1168047243 M * yarihm Bertl: well, there was an Azureus client running on that machine before, now i stopped azureus but there are lots of other bittorrent-users that try to access the azureus client's port which will result in ICMP port closed replies at the moment 1168047259 M * yarihm (azureus was running from within the vserver i mean) 1168047261 M * Bertl ah, ic .. 1168047275 M * yarihm however ... what i see when i ping the client in the same subnet: 1168047276 M * yarihm PING whitehead2.yarihm.unixwerkz.net (10.0.5.20) 56(84) bytes of data. 1168047276 M * yarihm From gateway.yarihm.unixwerkz.net (10.0.5.1): icmp_seq=1 Redirect Host(New nexthop: whitehead2.yarihm.unixwerkz.net (10.0.5.20)) 1168047276 M * yarihm 64 bytes from whitehead2.yarihm.unixwerkz.net (10.0.5.20): icmp_seq=1 ttl=64 time=3.74 ms 1168047276 M * yarihm From gateway.yarihm.unixwerkz.net (10.0.5.1): icmp_seq=2 Redirect Host(New nexthop: whitehead2.yarihm.unixwerkz.net (10.0.5.20)) 1168047283 M * Bertl the guest has the _first_ ip on the network? 1168047295 M * yarihm nono ... 1168047299 M * yarihm the guest is 10.0.5.18 1168047307 M * yarihm the host is 10.1.5.4 1168047315 M * yarihm (both networks /24) 1168047333 M * Bertl yeah, but the first guest assigned ip is the 10.0.5.18, yes? 1168047351 M * yarihm but as you can see, the guest (10.0.5.18) tries to get to 10.0.5.20 via 10.0.5.1 instead of getting there directly 1168047376 M * yarihm Bertl: all guests only have one IP, but there are other guests on the same subnet (e.g. 10.0.5.14 and 10.0.5.12) 1168047387 M * Bertl that's fine 1168047401 Q * haxier Remote host closed the connection 1168047420 M * yarihm i shall hope so :) but still, the question is: why is 10.0.5.20 being accessed via 10.0.5.1? 1168047446 M * Bertl I assume your routing table says so? 1168047448 J * dmax ~semaj@81.193.58.87 1168047470 M * Bertl what does 'ip route ls table vservers' give? 1168047495 M * yarihm the router then redirects as one can see from ping's output ... tcpdump shows the same: i do ssh root@10.0.5.18 from 10.0.5.20 then the reply is being sent to 10.0.5.1 1168047523 M * yarihm Bertl: one surely suspects that the routing table must have something to do with it :D i do so, too by the way 1168047541 M * yarihm kant:~# ip route ls table vservers 1168047541 M * yarihm default via 10.0.5.1 dev eth0 1168047553 J * s0undt3ch ~s0undt3ch@bl4-58-87.dsl.telepac.pt 1168047556 M * Bertl so no other rule there, what do you explect? 1168047563 M * Bertl do 1168047597 M * Bertl ip rule add 10.0.5.0/24 dev eth0 table vservers 1168047610 M * yarihm Bertl: i would have expected that the machine was smart enough to see that if the NIC is configured with a netmask of /24 ... however, in fact i didn't think about it at all i guess, i just assumed it worked 1168047618 M * yarihm ok 1168047633 M * Bertl well, routing works by using routing tables :) 1168047641 M * Bertl i.e. no entry there, no routing :) 1168047657 M * Bertl would be silly if it would assume various things ... 1168047711 M * yarihm there is no implicit entry there if I do "ifconfig eth0:foobar 10.0.1.18 netmask 255.255.255.0"? normally you don't have to tell the host that this means that 10.0.1.0/24 is accessible via eth0:foobar :) yet that assumption must be false when using routing tables i see 1168047738 M * daniel_hozac the route is added when you do that. 1168047740 M * daniel_hozac see ip route ls 1168047741 M * Bertl you have that entry _on the host_ routing table 1168047766 M * Bertl (to be precise, the main routing table 1168047770 M * Bertl ) 1168047799 M * yarihm yes, and as it seems _only_ in the main routing table, not in the vservers-table i defined later on ... makes sense somehow I must agree 1168047830 M * yarihm Bertl: 1168047831 M * yarihm kant:~# ip rule add 10.0.5.0/24 dev eth0 table vservers 1168047831 M * yarihm Error: argument "10.0.5.0/24" is wrong: Failed to parse rule type 1168047845 M * yarihm should i add "to 10.0.5.0/24" maybe? 1168047845 M * daniel_hozac s/rule/route/ 1168047848 M * yarihm oi 1168047864 M * Bertl indeed, sorry 1168047879 M * yarihm should have seen that myself ... 1168047891 M * yarihm ok, now. new try 1168047915 Q * duckx Remote host closed the connection 1168047923 M * yarihm har! it works 1168047943 M * yarihm thanks a lot! 1168047966 M * Bertl you're welcome! 1168047990 M * Bertl and I suggest to read the advanced routing guid and the ip man/info pages in the evening :) 1168048000 M * Bertl s/guid/howto/ 1168048102 M * yarihm you wouldn't believe, but i've already done so! but for some reason i didn't catch that this implicit rule that is calculated from the netmask and created in the main-routing table when assigning an address to an interface isn't created in other routing-tables ... 1168048116 M * yarihm maybe i should read again there ... 1168048383 J * nayco ~nayco@lns-bzn-59-82-252-154-246.adsl.proxad.net 1168048399 M * Bertl wb nayco! 1168048417 M * nayco Hello, Bertl ;-) 1168048668 M * nayco Bertl: You now have an X86รจ64 plateform ? 1168048677 M * nayco X86_64 1168048688 M * Bertl I have one to test on, yes ... 1168048699 M * Bertl (it's not mine, but available :) 1168048738 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net 1168048746 M * Bertl wb marcfiu! 1168048747 M * marcfiu hello vsWorld 1168048766 M * marcfiu I'm running into all sorts of badness with FC5 1.2260 based on 2.6.18.6. 1168048779 M * marcfiu The bugs appear to be either in mainline or FC5. 1168048796 M * marcfiu A while back I heard that vs2.0.3 could not be used with 2.6.19. 1168048796 M * Bertl hmm, why not simply replace it with mainline 2.6.19.1? 1168048805 M * yarihm Bertl: somewhat off-topic, but on what distro do you develop your stuff anyway? 1168048813 M * nayco Bertl: I saw that you built Mdv X86_64 packages... As you may have read upper, daniel_hozac helped me modify the util-vserver tarball for Mandriva 2007.0 1168048833 M * nayco I'm now submitting it to some Cooker developper. 1168048844 M * Bertl marcfiu: 2.6.19.1 + vs2.2.x 1168048848 M * daniel_hozac marcfiu: what's the problem? 1168048863 M * Bertl yarihm: mostly a self modified mandrake/iva 1168048875 M * yarihm Bertl: i see 1168048876 M * Bertl nayco: yep, I did recently 1168049061 M * marcfiu daniel_hozac: seeing slab corruption, badness in madvise, and what appears to be netfilter related Oops, spin lock bugs. 1168049115 M * daniel_hozac that sounds really bad. 1168049117 M * marcfiu daniel_hozac: it could well be that this is due to the additional stuff that we've integrated, but that stuff has worked quite solidly with <2.6.18 based kernel.s. 1168049144 M * marcfiu my colleague is working on figuring out if the madvise badness is due to FC or actually in mainline. 1168049144 M * Bertl well, a lot changed between 2.6.17 and 2.6.18 1168049182 M * marcfiu Bertl: my mistake for upgrading to it. 1168049224 M * marcfiu All I need to do is come up with a stable FC-based kernel release using vs2.0.x, and then I can move forward with vs2.2.x with >=2.6.19. 1168049235 M * marcfiu So maybe the right thing to do is back out to 2.6.17. 1168049268 M * Bertl maybe, you probably just have to backport the fixes we did between the last 2.6.17 and the recent 2.6.18 release 1168049269 M * nayco Hey, guys, who maintains the flower page (http://www.nongnu.org/util-vserver/doc/conf/configuration.html) ? Enrico ? 1168049286 M * daniel_hozac i've been updating it recently. 1168049286 M * Bertl nayco: ensc and daniel_hozac, AFAIK :) 1168049338 M * nayco Bertl: Ok, 'cause I had the nice surprise today to see that my CSS is now the default one ;-) 1168049396 M * daniel_hozac yeah, the blinking thing got to me :) 1168049403 M * daniel_hozac i hope you don't mind 1168049406 M * nayco Bertl: and by the way, i'd like to make a very small change to it : Change the licence to CC-By-SA (or GPL if you need it) 1168049427 M * nayco daniel_hozac: I don't mind, and i'm proud ;-p 1168049474 M * nayco I'd just like to remove "and all that ;-)", and change the licence to something more free. 1168049548 M * daniel_hozac send me an updated one and i'll commit it. 1168049625 M * nayco daniel_hozac: er... Dunno if I still got the original file. Ok, I gonna download yours ;-) => Which licence do you prefer ? 1168049659 M * daniel_hozac doesn't matter much since they're not shipped in the tarball, but GPL would make that possible... 1168049760 M * nayco ok, GPL, so. Dunno if CC-By-SA and GPL are compatible, anyway, but, hell... 1168049823 M * marcfiu Bertl: backport vs2.0.3-rc1 to 2.6.17? Seems doable. 1168049882 M * marcfiu Bertl: involves rolling forward from vs2.0.2.1 to vs2.0.3-rc1. 1168049946 M * Bertl yeah, get the deltas between that, analyze the changes and add them (as far as possible) to the last known good state 1168050165 M * marcfiu yup... I'll do that. Will be less work than figuring out what badness slipped into the 2.6.18 based FC5 releases. 1168050197 M * Bertl if you have specific questions regarding the changes, feel free to bother me ... 1168050220 M * daniel_hozac most of changes are 2.6.18 specific, IIRC. 1168050307 M * yarihm n8 everyone 1168050324 M * yarihm (those in europe at least) 1168050327 Q * yarihm Quit: Leaving 1168050706 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1168050782 M * nayco daniel_hozac: is this right ? "Stylesheet and images (c)2005 by Nayco. They are Licenced under the terms of the GNU/GPL version 2." 1168050810 M * nayco I don't even know if GPL apllies to image.. 1168050813 M * nayco +s 1168050815 M * daniel_hozac GNU GPL, IMHO. 1168050821 M * nayco ok 1168050838 M * daniel_hozac well, feel free to use some CC license. 1168050856 M * daniel_hozac as i said, it doesn't really matter. they're not in the tarball anyway. 1168051054 M * nayco Oh, I gonna keep the GPL now ;-) 1168051092 M * nayco so, can I send it to you ? 1168051170 M * daniel_hozac sure. 1168051179 M * marcfiu Bertl: bother you... oh, I wouldn't do that. :) 1168051181 M * daniel_hozac pastebin, email, http, whichever you prefer. 1168051220 M * nayco go ! 1168051376 M * daniel_hozac there we go. 1168051471 M * nayco ok, great ;-) 1168051542 M * nayco utils 212 are at Mandriva, they are looking at the diffs with their previous version (204). God, if they have patches to apply... 1168051770 M * Bertl not more than 500k of changes, I'd guess :) 1168051827 M * daniel_hozac i'd expect more, actually. 1168051883 M * daniel_hozac the tarball is over 100k larger. 1168051918 M * daniel_hozac 4.3 MiB. 1168051994 M * daniel_hozac i guess quite a bit of that is just autofoo noise though. 1168052038 M * Bertl nayco: you are somehow related to mandriva? 1168052160 M * Bertl s/related to/affiliated with/ 1168052361 M * nayco Bertl: Nope. I use it, I sometimes make bug reports, and try to follow cooker developpement on my old test machine. I too subscribed to the mailing-list, and... I am french ;-). Oh, and once a year, I try to go to Paris to attend to "solution Linux", where the developpers and users meat in a restaurant... 1168052416 M * Bertl sounds nice ... so you have a good contact to the developers? 1168052549 M * nayco Well, good... It cannot be bad, because it's tiny ! Yes, I once or twice met some of them, some even belong to the LUG near my city, and I am currently chatting with them on IRC ;-) But I cannot say I have much contact with them (by any means that be) 1168052553 M * Hollow daniel_hozac: http://paste.linux-vserver.org/855 ;) 1168052599 M * daniel_hozac Hollow: hmm? 1168052604 M * Bertl nayco: hmm, there is a channel? maybe even an english one? 1168052610 M * nayco Bertl: You want to contact them or something ? 1168052616 M * nayco Bertl: Yes :D 1168052622 M * Hollow just playing around with sloccount.. neat tool i found lately ;) 1168052633 M * Bertl nayco: let's hear the details ... 1168052646 M * daniel_hozac Hollow: what is it? lines of (ansic|sh|perl) in each directory? 1168052656 M * Hollow yeah... SLOC = source lines of code 1168052662 M * nayco #mandriva-cooker , irc.freenode.net , english channel ! 1168052673 M * daniel_hozac ok. 1168052684 M * nayco Well, I now must stop to say bad thing about vserver there :p 1168053548 M * Bertl nayco: hehe 1168054129 M * marcfiu We don't do networking on PlanetLab the "usual vserver" way and so have little experience w/ chbind etc.. 1168054129 M * marcfiu Is it possible to have, say, 2 wireless eth devices and expose wlan0 only in vserver-guest0 and wlan1 in vserver-guest1 and let each of them be able to invoke iwconfig etc. on those just the wlanX devices that are exposed in their respective guests? 1168054172 M * daniel_hozac no. 1168054186 M * Bertl daniel_hozac: do you have a minute for mandriva? 1168054190 M * nayco :D 1168054192 M * daniel_hozac iwconfig would require CAP_NET_ADMIN. 1168054195 M * daniel_hozac Bertl: sure. 1168054284 M * marcfiu daniel_hozac: ok 1168054291 M * marcfiu thx 1168054303 M * nayco . 1168054323 M * marcfiu suppose one gave just those guests CAP_NET_ADMIN, then they presumably could hose up the network settings for the whole system, right? 1168054393 J * blino olivier@blino.org 1168054395 M * blino hello 1168054404 M * Bertl welcome blino! 1168054421 M * Bertl daniel_hozac: meet blino, blino meet daniel_hozac :) 1168054460 M * daniel_hozac marcfiu: right 1168054464 M * daniel_hozac hello blino 1168054495 M * Bertl daniel_hozac: could you give blino a short overview of the debian debootstrap integration in util-vserver? 1168054504 M * Bertl (maybe with pointers to the scripts?) 1168054509 M * blino hi daniel_hozac 1168054525 M * Bertl blino: is interested in a similar process (if possible) for urpmi 1168054525 Q * Fire_Egl Quit: ... 1168054534 M * daniel_hozac ah, cool. 1168054558 M * daniel_hozac is debootstrap the method most similar to urpmi though? 1168054581 M * daniel_hozac IMHO yum or rpm would be more interesting. 1168054582 M * Bertl nah, but probably the most automated one 1168054595 M * Bertl i.e. which works on _any_ distro 1168054611 M * daniel_hozac indeed, but is urpmi a simple script using common utilities? 1168054648 M * blino it requires a rpm library and perl-URPM bindings 1168054661 M * blino how do you setup a Debian guest from a running RedHat OS for example, 1168054662 M * blino ? 1168054663 M * daniel_hozac right, so it needs actual software to be installed. 1168054693 M * daniel_hozac i think the reverse is more interesting, i.e. setting up a Fedora/CentOS guest on Debian. 1168054710 M * daniel_hozac since debootstrap makes building Debian guests very easy. 1168054729 M * Bertl yes, but that's the point 1168054747 M * daniel_hozac but RPM is a more complex format. 1168054748 M * Bertl would be nice to get urpmi working on non rpm based systems too 1168054760 M * daniel_hozac well, you'd just have to install it and the requirements there, no? 1168054778 M * daniel_hozac like recent Debian's and Gentoo have yum in their repos, IIRC. 1168054791 M * Bertl but not urpmi, for example :) 1168054802 M * daniel_hozac right, so all it'd need is someone packaging it up. 1168054806 M * blino we also have apt 1168054812 M * daniel_hozac apt-rpm? 1168054847 M * blino yep 1168054849 M * Bertl blino: mandrake can be installed via apt-rpm? 1168054857 M * daniel_hozac but that's not a good idea, as it'll be next to impossible to use on Debian-based systems. 1168054880 M * daniel_hozac i.e. where a regular apt is already used. 1168054888 M * blino to install a debian guest, debootstrap should be present on the host, right? 1168054947 M * daniel_hozac no, the scripts will download it. 1168054980 M * Bertl that's the nice part, it doesn't require anything ... 1168054991 M * daniel_hozac but that's because debootstrap itself doesn't require anything. 1168054992 M * blino ok, that's nice 1168055015 M * daniel_hozac having the scripts download and install e.g. rpm automatically would not be a good idea. 1168055112 M * nayco I can confirm I already installed an ubuntu vserver on a mandriva host... Debootstrap is cool. 1168055155 M * blino rpm can also be extracted easily, using the perl version of rpm2cpio 1168055160 M * blino it does not require a rpm library 1168055182 M * daniel_hozac well, rpm2cpio won't be doing the right thing. 1168055197 M * daniel_hozac i.e. add it to the rpm database, run the scripts in the right order, check deps, etc. 1168055233 M * daniel_hozac reimplementing rpm in bash doesn't strike me as the best idea either. 1168055253 M * daniel_hozac IMHO it's quite fine to just require the software be present on the system already. 1168055259 M * blino it could be enough to have an rpm lib and perl-URPM in a chroot 1168055276 M * blino or just in a local directory 1168055284 M * nayco could perl-URPM and urpmi run on e.g. debian ? 1168055302 M * blino I think so 1168055314 M * daniel_hozac that's what we do for the rpm, yum, fai, apt-rpm, template and rsync build methods, i.e. all of them _except_ debootstrap. 1168055342 M * blino how do you proceed to install Fedora on Debian for example? 1168055366 M * daniel_hozac first you'd have to install yum, then vserver ... build -m yum ... -- -d fcX. 1168055395 M * blino and probobably rpm and python-rpm before yum 1168055403 M * daniel_hozac apt resolves deps ;) 1168055439 M * blino yes, but it means that it requires both rpm, python-rpm and yum to be available for the host distro 1168055453 M * daniel_hozac it requires yum be available, yes. 1168055471 M * blino AFAIK, our perl-RPM bindings and urpmi are on the CPAN, so they should be easy to package 1168055479 M * Radiance any one can point me to a doc or so where i can see for example that the host is running distro X but the vserver is installed with distro Y ? 1168055484 M * Radiance hi btw:) 1168055489 M * daniel_hozac Radiance: hmm? 1168055532 M * Radiance well, i mean a document which shows how to install a distro in a vserver which is different from the host itself. 1168055550 M * daniel_hozac you install the guests the same way, regardless of the host distro. 1168055577 M * daniel_hozac there may be some preparations required before you _can_ install it, but the actual building is the same. 1168055639 M * blino sorry, I should really go to sleep now, I'll see that tomorrow 1168055642 M * Radiance ok cool :) 1168056084 Q * gerrit Ping timeout: 480 seconds 1168056117 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1168056677 M * nayco same for me guys, good night ;-) 1168056695 Q * nayco Quit: Bonne nuit ! 1168058070 Q * MrX Quit: X 1168058971 J * FireEgl Proteus@adsl-147-90-220.bhm.bellsouth.net 1168059365 Q * nox Ping timeout: 480 seconds 1168059770 J * nox ~nox@static.88-198-17-175.clients.your-server.de 1168061074 P * marcfiu 1168061461 M * pflanze Hello. 1168061466 M * daniel_hozac hi 1168061491 M * pflanze a day ago or two I've asked about whether the vserver kernel code is available in a git repository. 1168061507 M * pflanze (insert " ago" after "two") 1168061525 M * pflanze What's the answer? 1168061528 M * daniel_hozac no. 1168061535 M * pflanze doh. 1168061568 M * pflanze I'm not sure (not being used to git enough) but I think that would make updating patches to newer kernel releases easier. 1168061620 M * daniel_hozac IIRC doener tested that theory and it didn't turn out to be much easier than the manual way. 1168061625 M * pflanze One could "just" merge the current upstream kernel. The main point being that the history of both projects is present. 1168061652 M * pflanze Well probably it would mainly be easier for those who don't know the history already. 1168061691 M * pflanze (Well maybe it wouldn't safe time per se, but it would probably tend to avoid dangerous guesswork.) 1168061703 M * pflanze s/safe/save/ 1168062630 M * Bertl pflanze: just make your own git repository, no? 1168062639 M * Bertl check in the vserver patches, and voila 1168062647 M * pflanze But that would not contain your history! 1168062675 M * Bertl that's correct, but if you check the patches carefully, you'll see what they do 1168062701 M * pflanze hm it was that way some time ago, yes 1168062725 M * pflanze but I wanted to patch a recent kernel with 2.0 and I gave up when I saw ~90 reject files. 1168062742 M * Bertl recent means what exactly? 1168062749 M * daniel_hozac a recent kernel meaning 2.6.18.x? because 2.0 isn't going to be ported to 2.6.19 at all. 1168062778 M * pflanze newest vanilla 2.6 stable, and patch-2.6.18.5-vs2.0.3-rc1.diff 1168062790 M * pflanze why that? 1168062804 M * daniel_hozac because 2.6.19 introduced the uts and IPC namespaces. 1168062809 M * Bertl because a lot of the vserver (related) code got into mainline 1168062810 M * daniel_hozac i.e. mainline virtualization 1168062818 M * pflanze yep, I wondered about those "IPC Namespaces". 1168062827 M * pflanze ah 1168062869 M * pflanze what do you suggest I do (for a production server)? 1168062880 M * Bertl use the upcoming stable 2.2.x release? 1168062886 M * pflanze ok 1168064992 M * Bertl okay, off to bed now ... have a good one everyone! cya! 1168064998 N * Bertl Bertl_zZ 1168071183 Q * Loki|muh Read error: Connection reset by peer 1168071304 J * Loki|muh loki@satanix.de 1168074492 J * Aldo ~sebastian@pD957DE8C.dip.t-dialin.net 1168074505 M * Aldo hi 1168074563 Q * FireEgl Remote host closed the connection 1168075428 Q * Aldo Quit: leaving 1168075851 Q * softi42 Ping timeout: 480 seconds 1168076458 J * softi42 ~softi@p549D7B56.dip.t-dialin.net 1168078690 J * _mcp ~hightower@wolk-project.de 1168078690 Q * mcp Read error: Connection reset by peer 1168080234 J * dna ~naucki@p54BCD18D.dip.t-dialin.net 1168080352 J * bandsaw me@host217-45-242-181.in-addr.btopenworld.com 1168080485 J * Syscalo ~BlackArt@213-140-21-224.fastres.net 1168081520 J * dna_ ~naucki@p54BCD18D.dip.t-dialin.net 1168081785 Q * dna Ping timeout: 480 seconds 1168082056 Q * Aiken Ping timeout: 480 seconds 1168082479 Q * bandsaw Read error: Connection reset by peer 1168085116 Q * ensc Ping timeout: 480 seconds 1168085163 J * dna ~naucki@p54BCD18D.dip.t-dialin.net 1168085221 J * dna___ ~naucki@p54BCD18D.dip.t-dialin.net 1168085540 Q * dna_ Ping timeout: 480 seconds 1168085645 Q * dna Ping timeout: 480 seconds 1168085816 J * dna_ ~naucki@p54BCD18D.dip.t-dialin.net 1168086240 Q * dna___ Ping timeout: 480 seconds 1168086284 J * dna ~naucki@p54BCD18D.dip.t-dialin.net 1168086526 J * dna___ ~naucki@p54BCD18D.dip.t-dialin.net 1168086539 J * bonbons ~bonbons@83.222.37.103 1168086587 J * Piet hiddenserv@tor.noreply.org 1168086665 Q * dna_ Ping timeout: 480 seconds 1168086885 Q * dna Ping timeout: 480 seconds 1168087135 J * Belu Belu@88.134.218.136 1168087145 M * Belu hello 1168087192 J * dna ~naucki@p54BCD18D.dip.t-dialin.net 1168087249 J * shedi ~siggi@inferno.lhi.is 1168087314 J * BeLu_zZZz B.Lukas@mail.openvcp.org 1168087349 J * s0undt3ch_ ~s0undt3ch@80.69.34.154 1168087553 Q * Belu 1168087605 Q * dna___ Ping timeout: 480 seconds 1168087678 N * BeLu_zZZz Belu 1168087751 J * dna_ ~naucki@p54BCD18D.dip.t-dialin.net 1168087890 A * Belu is away (bye bye) 1168087890 N * Belu BeLu_zZZz 1168087891 N * BeLu_zZZz Belu 1168087898 A * Belu is away (bye bye) 1168087902 N * Belu BeLu_zZZz 1168087902 N * BeLu_zZZz Belu 1168087915 J * dna___ ~naucki@p54BCD18D.dip.t-dialin.net 1168087915 A * Belu is away (bye bye) 1168087915 N * Belu BeLu_zZZz 1168087944 N * BeLu_zZZz Belu 1168087972 A * Belu is away (bye bye) 1168087972 N * Belu BeLu_zZZz 1168087979 N * BeLu_zZZz Belu 1168087986 M * daniel_hozac Belu: please stop. 1168088013 M * Belu oh sorry i am testing :D 1168088160 Q * dna Ping timeout: 480 seconds 1168088275 Q * dna_ Ping timeout: 480 seconds 1168088287 N * _mcp mcp 1168088373 A * m4z facefucks Belu 1168088380 M * m4z oh sorry i am testing d: 1168088638 M * Belu .... 1168088817 J * cryptronic crypt@mail.openvcp.org 1168089164 J * duckx ~Duck@tox.dyndns.org 1168089620 J * dna_ ~naucki@p54BCD18D.dip.t-dialin.net 1168089874 J * dna ~naucki@p54BCD18D.dip.t-dialin.net 1168090030 Q * dna___ Ping timeout: 480 seconds 1168090194 J * dna___ ~naucki@p54BCD18D.dip.t-dialin.net 1168090220 Q * dna_ Ping timeout: 480 seconds 1168090512 J * dna_ ~naucki@p54BCD18D.dip.t-dialin.net 1168090600 Q * dna Ping timeout: 480 seconds 1168090731 M * m4z Belu: /me is away is _really_ bad. please please use /away instead 1168090885 Q * dna___ Ping timeout: 480 seconds 1168091101 M * waldi m4z: in several chanels he will be killed for both, public away and nick change 1168091128 M * m4z and thats a good thing! 1168091308 J * ensc ~irc-ensc@p54B4D3AF.dip.t-dialin.net 1168092265 J * yarihm ~yarihm@84-75-123-221.dclient.hispeed.ch 1168092894 Q * Syscalo Quit: KVIrc 1168094472 J * Aldo ~sebastian@pD957DE8C.dip.t-dialin.net 1168094477 M * Aldo hello 1168095010 J * _Medivh ck@paradise.by.the.dashboardlight.de 1168095011 Q * Medivh Read error: Connection reset by peer 1168095419 J * TrueBrain truelight@openttd.org 1168095456 M * TrueBrain Hi all :) Question: the data vserver-stat shows doesn't match the data proc/virtual/xid/limit shows (not for proc, not for rss) 1168095462 M * TrueBrain is this normal? Do they show different things? 1168095560 J * waldi_ ~waldi@bblank.thinkmo.de 1168095820 Q * waldi Ping timeout: 480 seconds 1168095987 N * waldi_ waldi 1168096493 M * Belu www.openvcp.org searching for php / frameworks dev anyboy intrests? query me 1168096510 M * blino hello 1168096541 M * Belu hello 1168097009 M * bonbons TrueBrain: vserver-stat show just a basic addition of values, it's not the best reference and values shown are not correct/absolute 1168097047 M * TrueBrain not correct seems to be an understatement... 1168097116 M * Aldo hi 1168097119 M * bonbons I don't remember the exact details, you would have to ask daniel_hozac for that... 1168097129 Q * Aldo Quit: Lost terminal 1168097250 M * TrueBrain some other thing I fail to understand: inside a VPS, the free swapspace shows always 0 bytes 1168097261 M * TrueBrain while the memory has more total then available 1168097344 M * TrueBrain hmm, no, the total memory is correct, just the swap shows odd numbers 1168097611 M * bonbons I see no difference for those (running top on the host or guest shows exactly the same), but I don't set memory limits for my guest 1168097670 M * TrueBrain I have no limits eithers 1168097678 M * TrueBrain but the swap space always shows 0 bytes free in any guest 1168097684 M * TrueBrain while the host really shows something different 1168097690 M * TrueBrain oh, I enabled virt_ stuff 1168097691 M * TrueBrain all 4 1168097715 Q * yarihm Quit: Leaving 1168102234 M * nebuchadnezzar hi everybody 1168102248 J * ntrs ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com 1168102328 J * Syscalo ~BlackArt@213-140-21-224.fastres.net 1168102355 Q * ntrs 1168103574 J * _mcp ~hightower@wolk-project.de 1168103629 Q * mcp Read error: Connection reset by peer 1168103633 N * _mcp mcp 1168103668 J * Medivh ck@paradise.by.the.dashboardlight.de 1168103671 Q * _Medivh Ping timeout: 480 seconds 1168103921 Q * Syscalo Ping timeout: 480 seconds 1168105255 M * Radiance hi :) 1168105492 Q * m`m`h Remote host closed the connection 1168107002 Q * thunder1 Remote host closed the connection 1168107017 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net 1168108280 J * Sebast1an ~sebastian@office.star-hosting.de 1168108365 Q * Sebastian|aw Ping timeout: 480 seconds 1168110180 J * nayco ~nayco@lns-bzn-31-82-252-249-229.adsl.proxad.net 1168110204 M * nayco Hi all ! 1168110211 M * daniel_hozac hello 1168110265 M * blino hello nayco 1168110279 M * blino nayco: I've uploaded a new util-vserver in Mandriva 1168110296 M * blino actually, we have a patch to support urpmi, I don't know why it has not been forwarded upstream 1168110337 M * nebuchadnezzar hi daniel_hozac 1168110356 M * daniel_hozac hey nebuchadnezzar 1168110408 M * nayco blino: *urpmi.update -a* 1168110432 M * blino nayco: I'm not sure it has been built and mirrored yet 1168110447 M * blino hmm, it's built at least 1168110447 M * nayco blino: I need to update anyway :-D 1168110457 M * nayco Thanks a lot. 1168110472 M * nayco I gonna have a look at this kernel-linus stuff... 1168110488 M * blino nayco: I'll try to build a kernel-vserver tonight 1168110506 M * daniel_hozac blino: URL? 1168110517 M * blino daniel_hozac: for the patch ? 1168110526 M * daniel_hozac yes. 1168110534 J * FireEgl Proteus@adsl-147-90-220.bhm.bellsouth.net 1168110576 M * nayco blino: Ok for the kernel. Anyway, I want to have a look by myself to learn... Well, to try to ;-) 1168110583 M * blino daniel_hozac: I've not completely reworked it, I yet have to update it for the Mandrake/Mandriva switch, but you can have a quick look here: http://svn.mandriva.com/cgi-bin/viewvc.cgi/packages/cooker/util-vserver/current/SOURCES/vserver-mandrake.patch?view=markup 1168110811 M * daniel_hozac it was submitted, but it seems the concerns raised in the review haven't been addressed: http://www.paul.sladen.org/vserver/archives/200502/0090.html 1168111127 M * blino daniel_hozac: thanks, I'll have a look 1168111773 M * nebuchadnezzar daniel_hozac: no new idea about the sparc prblem ? 1168111828 M * daniel_hozac not really, but gcc is definitely doing something wrong. 1168112176 M * nebuchadnezzar daniel_hozac: maybe I can try to install gcc-4.2 from experimental to see if it works ? 1168112230 M * nebuchadnezzar I'll build a new vserver to make the compilation, experimental packages can cause troubles :-) 1168112263 M * daniel_hozac hehe, sounds like a plan. 1168112657 J * bandsaw ~bandsaw@host217-45-242-181.in-addr.btopenworld.com 1168114307 J * lilalinux ~plasma@dslb-084-058-204-040.pools.arcor-ip.net 1168115998 M * TrueBrain then you want to open openvps.org, then it doesn't work! :) 1168116930 Q * Johnnie Ping timeout: 480 seconds 1168116933 J * Aiken ~james@tooax6-087.dialup.optusnet.com.au 1168119351 N * Bertl_zZ Bertl 1168119355 M * Bertl morning folks! 1168119392 M * nayco Hello Bertl ! 1168119407 M * nayco Hmmm, isn't this a little late for a morning ;-) ? 1168119425 M * Bertl a good late morging then :) 1168119430 M * Bertl *morning 1168119929 J * Piet_ hiddenserv@tor.noreply.org 1168120221 J * Aldo ~sebastian@pD957DE8C.dip.t-dialin.net 1168120226 M * Aldo hi 1168120335 Q * Piet Ping timeout: 480 seconds 1168120553 M * Bertl welcome Aldo! 1168120576 M * Aldo hi 1168120589 M * Aldo are you a supporter auf linux-vserver? 1168120723 M * Bertl most of us 'support' it :) 1168120744 M * Bertl do you have any questions or issues regarding Linux-VServer? 1168120753 M * Bertl if so, please go ahead and ask ... 1168120801 M * Aldo First I'm 15 and came from Germany, I hope my English is not sooo bad :D 1168120828 M * Bertl seems we can understand eachother, so it's sufficient ... 1168120869 M * Aldo I'm installed the Linux vServer Kernel 1168120881 M * Aldo then the Linux vServer Util 1168120886 M * Bertl okay, mainline with patches? 1168120895 M * Bertl or some distro kernel? 1168120912 M * Aldo Kubuto 2.6.19.1 Kernel 1168120918 M * Aldo patched 1168120942 M * Bertl okay 1168120966 M * Aldo the Problem is when I built a vServer I get that: 1168120970 M * Aldo segmentation fault 1168120989 M * Aldo I have no idea whats wrong 1168121001 M * Bertl that's not good, looks more like a compiler issue to me (at first glance) 1168121018 M * Bertl you did compile the utils yourself? 1168121023 M * Aldo yes 1168121032 M * Bertl with dietlibc? 1168121082 M * Bertl let's start with 'vserver-info - SYSINFO' 1168121083 M * Aldo http://www.debianhowto.de/doku.php/de:howtos:sarge:linux-vserver 1168121093 M * Bertl and upload the output to paste.linux-vserver.org 1168121110 M * Aldo the System is at the moment not booted with the vServer Kernel 1168121148 M * Aldo vserver-info 1168121157 M * Aldo sebastian@Linux:~$ vserver-info 1168121162 M * Aldo Segmentation fault 1168121165 M * Aldo mh :( 1168121186 M * Bertl so I'd say you have a problem with your toolchain (compiler + binutils) 1168121201 M * Aldo ok T look tomorrow 1168121214 M * Bertl okay, no problem ... 1168121249 M * Aldo am @home the linux-vServer shouldn't run on this system 1168121279 M * Aldo I wait of my root system, i will get it next week 1168121307 M * Bertl some hints: 1168121328 M * Bertl - do not patch a distro specific kernel with mainline patches 1168121345 M * Bertl - compile the tools with dietlibc (you get a warning if not) 1168121366 M * Bertl - test the kernel with the testme.sh script (basic functionality) 1168121393 M * Bertl - try to install a debian guest first, it is very simple 1168121394 M * Aldo Sorry, but what is a "mainline" patch? 1168121422 M * Bertl patches you get from vserver.13thfloor.at are against 'vanilla' or 'mainline' kernels 1168121430 M * Bertl i.e. kernels from kernel.org 1168121461 J * Johnnie ~jdlewis@jdlewis.org 1168121461 M * Aldo What do you mean i make the kernel? 1168121467 M * Aldo hello 1168121482 M * Bertl well, if you 'build' the kernel yourself 1168121493 M * Bertl you should take one from kernel.org 1168121498 M * Aldo also make make_modules e.g.? 1168121513 M * Bertl yep, kernel build process is like this: 1168121518 M * Aldo ok, that I do always :) 1168121525 M * Bertl - get the mainline kernel (from kernel.org) and unpack it 1168121534 M * Aldo yes I know 1168121539 M * Bertl - apply the vserver patches (matching kernel version) 1168121549 M * Aldo and what is "dietlibc" ? 1168121559 M * Bertl that is a library replacing the glibc 1168121569 M * Bertl well, to be precise, it's not a library 1168121580 M * Bertl it is more a modification at compile time 1168121589 M * Aldo how can I usw the glibc? 1168121602 M * Bertl glibc is not suited for guest installation 1168121614 M * Bertl the problem there is, glibc is modular 1168121628 M * Bertl and the modules (e.g. resolver) are loaded at runtime 1168121634 M * Aldo look here: 1168121638 M * Aldo http://www.debianhowto.de/doku.php/de:howtos:sarge:linux-vserver 1168121646 M * Aldo is that Howto o.k.? 1168121652 M * Bertl when the tools switch between host and guest context, the glibc will fail horribly 1168121665 M * Bertl sec, checking 1168121756 M * Bertl yes, looks fine, check the output for this line: 1168121761 M * Bertl Use dietlibc: yes 1168121775 M * Aldo yes is goot and no is bad? 1168121779 M * Bertl if you get there something like: no (you've been warned) 1168121780 M * Aldo *good 1168121792 M * Bertl then you are compiling against glibc, which will fail 1168121824 M * Aldo one last questition 1168121828 M * Aldo *question 1168121864 M * Aldo Have you a ICQ Account? 1168121870 M * Aldo or msn? 1168122341 Q * phreak`` Quit: leaving 1168122365 J * phreak`` ~phreak``@styx.xnull.de 1168122390 M * Bertl Aldo: nope, no microsoft here ... 1168122414 M * Bertl and I'm not very fond of ICQ either, but IRC is fine :) 1168122468 Q * Aldo Ping timeout: 480 seconds 1168123244 Q * bonbons Quit: Leaving 1168125384 Q * TrueBrain Ping timeout: 480 seconds 1168125630 J * TrueBrain truelight@openttd.org