1258935745 Q * geb Quit: / 1258937748 J * blues_ ~blues@afc22.neoplus.adsl.tpnet.pl 1258937867 Q * blues Ping timeout: 480 seconds 1258938538 Q * urbee Ping timeout: 480 seconds 1258940020 Q * AmokPaule Remote host closed the connection 1258941657 Q * bzed Ping timeout: 480 seconds 1258941715 N * Bertl_zZ Bertl 1258941718 M * Bertl back now ... 1258942278 Q * Piet Ping timeout: 480 seconds 1258942542 J * Piet ~Piet__@659AADS59.tor-irc.dnsbl.oftc.net 1258943163 M * Bertl daniel_hozac: did you get around uploading any deltas for your kernel fixes/cleanups yet? 1258943369 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/ 1258943392 M * Bertl ah, great! tx! 1258943393 M * daniel_hozac the ones with names identical to yours were just applied without any problems. 1258943400 M * daniel_hozac (some minor offsets) 1258943462 M * Bertl how did you find the header fixes? 1258943468 M * daniel_hozac just reading through the patches. 1258943527 M * daniel_hozac shouldn't be too hard to write an awk script for that though... 1258943588 M * Bertl well, it has proven to be a little tricky to handle the various config dependancies properly, especially for unusual archs 1258943616 M * Bertl but of course, it's always good to identify obsolete headers 1258943741 M * Bertl ah, so you included the dlimit feat02 already there, i.e. I assume it worked out fine? 1258943804 M * daniel_hozac no, i am testing it soon. 1258943823 M * Bertl okay, I was wondering if we shouldn't add a flag or bump the revision anyway, or how would you detect that the feature is present? 1258943837 M * daniel_hozac ah... yes, we need to do that. 1258943848 M * Bertl good, any preference there? 1258943893 M * daniel_hozac a VCI_VERSION bump sounds good. 1258943992 M * Bertl okay, will upload a pre for the current .32-rc soon, with the new hard cfs patches (v4) 1258944048 M * Bertl I'll pump the version for that and include the patch (it seems to be fine for older tools) 1258944495 M * daniel_hozac http://people.linux-vserver.org/~dhozac/t/headers.awk when run against the patch will output the files whose only modification is adding an include. 1258944710 M * Bertl okay, which still might be required, if one of the other includes contains unresolved references ... 1258944744 M * daniel_hozac sure. 1258944757 M * daniel_hozac i'm not saying they're all problems. 1258944816 M * Bertl okay, I plan to update my cross compiling env anyway, so we should be able to test/check that soon :) 1258945240 J * derjohn_foo ~aj@e180192034.adsl.alicedsl.de 1258945331 M * Bertl daniel_hozac: the namespace-feat01, that is a restricted version of the nscap-feat01 delta, correct? 1258945357 M * daniel_hozac ah, there it is. i couldn't find the delta for it. 1258945365 M * daniel_hozac that's correct. 1258945404 M * Bertl I'm wondering where the first hunk has gone? isn't it needed/wanted? 1258945419 M * daniel_hozac hmm. i missed that one. 1258945449 M * Bertl was it tested? 1258945472 M * daniel_hozac for pam_namespace? 1258945494 M * Bertl yeah, I presume that's what the original change was for 1258945517 M * daniel_hozac no. 1258945539 M * daniel_hozac i was just getting hunks from the 2.6.31 patch. 1258945572 M * Bertl okay, why do we want to restrict the VXC_NAMESPACE? 1258945648 M * daniel_hozac they're not all hierarchial. 1258945670 M * daniel_hozac e.g. NEWIPC could be used to hide resources from a casual inspector. 1258945671 Q * derjohn_mob Ping timeout: 480 seconds 1258945681 M * daniel_hozac and consume more resources than they have been allotted. 1258945689 M * Bertl as the NEWNS could, no? 1258945766 M * daniel_hozac i suppose. 1258945786 M * daniel_hozac but not quite the same. 1258945839 M * Bertl okay, I was thinking of adding some kind of clone/space mask instead, like with the bcapabilities 1258945870 M * Bertl so that you can configure what unshare flags are fine 1258945881 M * daniel_hozac sounds good. 1258945904 M * Bertl the question is, how should we go about the userspace API for that 1258945998 M * daniel_hozac similar to the bcapabilities? 1258946024 M * Bertl okay, so with a mask? 1258946051 M * daniel_hozac makes for a prettier interface. 1258946107 M * Bertl should we still keep the VXC_NAMESPACE as enabler, or drop that in favor of a restrictive default mask? 1258946207 M * daniel_hozac an initial mask of 0 seems good to me. 1258946258 M * Bertl okay, I'll prepare something then .. testing shouldn't be too hard with vcmd ... 1258946923 Q * Piet Quit: Piet 1258947098 J * aj__ ~aj@e180193139.adsl.alicedsl.de 1258947401 M * Bertl so I guess, we leave the VXC_NAMESPACE for the recursive changes, and remove it from the unshare checks, yes? 1258947471 Q * derjohn_foo Ping timeout: 480 seconds 1258947500 M * daniel_hozac i suppose... 1258948573 J * bzed ~bzed@devel.recluse.de 1258948612 J * saulus_ ~saulus@c207065.adsl.hansenet.de 1258949021 Q * SauLus Ping timeout: 480 seconds 1258949028 N * saulus_ SauLus 1258949213 Q * aj__ Ping timeout: 480 seconds 1258950314 J * Piet ~Piet__@659AADS70.tor-irc.dnsbl.oftc.net 1258954402 Q * nenolod Quit: Leaving 1258954432 J * nenolod ~nenolod@petrie.dereferenced.org 1258954838 Q * nenolod Quit: Leaving 1258954849 J * nenolod ~nenolod@petrie.dereferenced.org 1258956586 M * Bertl okay, uploaded two umask deltas and a patch for 2.6.32-rc8 (considered experimental) 1258956606 M * Bertl the umask stuff is untested, so please double check :) 1258956625 M * Bertl off to bed now ... have a good one everyone! 1258956631 N * Bertl Bertl_zZ 1258958617 J * aj__ ~aj@tmo-108-124.customers.d1-online.com 1258962624 J * davidkarban ~david@80.250.18.198 1258964264 Q * davidkarban Quit: Ex-Chat 1258964312 J * davidkarban ~david@80.250.18.198 1258965786 Q * aj__ Ping timeout: 480 seconds 1258965794 J * barismetin ~barismeti@zanzibar.inria.fr 1258967161 J * aj__ ~aj@c189231.adsl.hansenet.de 1258967826 Q * aj__ Ping timeout: 480 seconds 1258968381 J * friendly ~friendly@ppp118-209-248-164.lns20.mel6.internode.on.net 1258968470 J * scientes ~scientes@174-21-223-62.tukw.qwest.net 1258969772 J * kir ~kir@swsoft-msk-nat.sw.ru 1258971473 J * nas ~chatzilla@opengw.lga.net.sg 1258971552 M * nas guys i added echo eth0:0 > /etc/vservers/myserver/interfaces/1/dev; and echo 10.10.10.10 > 1258971554 M * nas /etc/vservers/myserver/interfaces/1/dev 1258971571 M * nas after vserver restart; does not register IP on ifconfig 1258971578 M * nas anything i miss out? 1258971617 M * blathijs nas: Perhaps you shouldn't do :0, but just use eth0 ? 1258971621 M * bobnormal hrrm, i never tried >1 interface but interested to see if you can get it working as i might try it later :) 1258971632 M * nas i mean echo 10.10.10.10 > /etc/vservers/myserver/interfaces/1/ip 1258971642 M * nas i have eth0 1258971651 M * nas but i want to add 2 more IPs for the vserver 1258971657 M * nas so total of 3 IPs 1258971687 M * nas so i have ../dev/0; ../dev/1; ../dev/2 1258971729 M * nas and eth0; eth0:0; and eth0:1 resp 1258971754 M * blathijs nas: Yeah, but you might be able to do that without using :0 and :1, just put multiple addresses on the eth0 interface 1258971806 M * blathijs I think the Linux kernel has 2 ways of doing multiple addresses. One is the :0 virtual interface approach (used by ifconfig IIRC), the other is just putting multiple addresses on a single interface (which is what "ip addr add" & friends do) 1258971831 J * AmokPaule ~amokpaule@brsg-4dbb1e65.pool.mediaWays.net 1258971836 M * blathijs The latter is also what happens when you have multiple vservers on the same interface 1258971855 M * blathijs So unless you really want those virtual interfaces separately, I'd try that approach 1258971860 M * bobnormal whats the status with the whole vserver networking anyway, i read something somewhere that there was a vserver context-id equivalent that can be used as a vnetwork id 1258971881 M * bobnormal is the vnetwork id possible to match on using netfilter/iptables? 1258971903 M * bobnormal it was sort of unclear from the documentation i read, which made it look like an upcoming feature rather than something established 1258971911 M * bobnormal anyone tried that method? 1258971915 M * blathijs Dunno about that 1258971928 M * nas if place multiple IPs /lines inside ../interfaces/ip becomes error 1258971964 M * nas if /etc/vservers/myserver/interfaces/0/ip contains multiple lines error 1258971994 M * blathijs nas: I meant having multiple interfaces with dev eth0 1258971999 M * bobnormal nas: i think the 'best practice' approach is to not alter /etc/vservers/* manually but instead to use the vserver-build script with multiple arguments 1258972020 M * blathijs e.g., put eth0 in interfaces/1/dev and 10.something in interfaces/1/ip 1258972026 M * nas nas: I meant having multiple interfaces with dev eth0 --> how? 1258972059 M * blathijs /e.g./i.e./ 1258972067 Q * fback Remote host closed the connection 1258972142 J * fback ~fback@red.fback.net 1258972160 M * bobnormal nas, cant you just add multiple arguments to vserver-build ? 1258972178 M * nas oh so /etc/vservers/mail01/interfaces/0/dev; /etc/vservers/mail01/interfaces/1/dev; /etc/vservers/mail01/interfaces/2/dev is all eth0? 1258972180 M * nas and /etc/vservers/mail01/interfaces/0/ip up to 2/ip have the IP values resp? 1258972189 M * blathijs nas: Yes 1258972207 M * blathijs nas: It's a guess, though, but worthwile to try :-) 1258972241 M * blathijs bobnormal: That's probably a fine idea as well, though it doesn't work for existing vservers. Modifying /etc/vservers manually is not a bad thing, it's pretty well documented on the flower page 1258972264 M * blathijs bobnormal: Still, creating a dummy vserver with multiple ip's and seeing what that looks like might not be bad either :-) 1258972295 J * geb ~geb@earth.gebura.eu.org 1258972309 M * nas hmmm 1258972354 M * nas i will try in my test server 1258972362 M * bobnormal blathijs, ok 1258972538 M * geb hi 1258973022 Q * nenolod Ping timeout: 480 seconds 1258973152 Q * fback Quit: leaving 1258973292 Q * nas Quit: ChatZilla 0.9.85 [Firefox 3.0.11/2009060309] 1258973298 J * fback ~fback@red.fback.net 1258973327 J * nenolod ~nenolod@petrie.dereferenced.org 1258973401 Q * nenolod 1258973414 J * nenolod ~nenolod@petrie.dereferenced.org 1258974795 J * ghislain ~AQUEOS@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr 1258976276 M * bobnormal /etc/vservers/.defaults/fstab doesnt seem to auto-deploy to new vms 1258976283 M * bobnormal err vservers even 1258976328 Q * kir Remote host closed the connection 1258976336 Q * friendly Quit: Leaving. 1258976408 J * derjohn_mob ~aj@213.238.45.2 1258977632 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net 1258977793 J * kir ~kir@swsoft-msk-nat.sw.ru 1258979405 M * ghislain bobnormal: the default are not a skeleton they are a default that is applied at all vserver start if the setting is not overrided in the vserver guest settings 1258979450 M * ghislain bobnormal: so if you have no fstab it could be used but i never tested the default part myself 1258979460 M * bobnormal ghislain, so how can i make normal calls to 'vserver newinstance build ....' use a specified fstab? 1258979504 M * ghislain bobnormal: i do not think the build method allow you that, you will have to wrap a little script that copy your fstab just after the build 1258979532 M * ghislain bobnormal: but here again perhaps there is a better way, not sure here. seems simple enough to be used anyway :) 1258979542 M * bobnormal ghislain, that's what im doing, but 'http://www.nongnu.org/util-vserver/doc/conf/configuration.html' says /etc/vservers/.defaults/fstab = The default fstab file to put in newly built guests. it just doesn't work. 1258979608 M * ghislain bobnormal: i see, perhaps daniel will pass by and answer on this one i never used this 1258979624 M * bobnormal ghislain, ok thanks 1258980147 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk 1258980676 J * evilhackerdude ~stephan@78.46.203.42 1258980686 M * evilhackerdude hey 1258980730 M * evilhackerdude i want to build a squeeze guest from a lenny host system - newvserver of vserver debiantools tells me it "doesn't know how to build squeeze" 1258980758 M * evilhackerdude anyway, it works - but is there anything i have to do manually after the install, or should i create the vserver differently all along? 1258980864 M * bobnormal evilhackerdude, i think this question depends more on your distribution and the instantiation mechanism you use with 'vserver build' / your particular usage scenario rather than vserver itself 1258980972 M * evilhackerdude i'm running debian 5.0.3 on the host and i want to use the newvserver tool to instantiate the vserver 1258980996 M * evilhackerdude i didn't completely get what you said :D 1258981012 M * bobnormal evilhackerdude, it seems newvserver is a debian tool rather than a linux-vserver project tool 1258981025 M * bobnormal evilhackerdude, so any bugs/queries around that should be resolved with debian 1258981025 J * gat ~chip@c-76-103-219-154.hsd1.ca.comcast.net 1258981037 M * evilhackerdude i c 1258981049 M * bobnormal evilhackerdude, basically that will be a frontend for the vserver tool, ie: 'vserver build ' 1258981080 M * bobnormal evilhackerdude, maybe try bypassing debian tools and using vserver directly, this probably gives you more power/insight in to the operation of the vserver system anyway. 1258981289 M * evilhackerdude i will, thanks 1258981421 M * blathijs evilhackerdude: Look for the "debootstrap" build method of vserver build, it will should create squeeze guests just fine 1258982495 M * evilhackerdude that worked much better than using `newvserver` - thanks alot 1258982674 M * blathijs :-) 1258982943 Q * jrdnyquist Quit: Leaving 1258983809 J * jrdnyquist ~jrdnyquis@slayer.caro.net 1258985092 M * evilhackerdude brr, locale problems 1258985386 J * ghislain1 ~AQUEOS@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr 1258985688 M * BenG hey all 1258985707 M * BenG anyone tell me the current state of cgroups support, and how I implement it 1258985710 M * BenG ? 1258985746 Q * ghislain Ping timeout: 480 seconds 1258985772 M * BenG http://linux-vserver.org/util-vserver:Cgroups seems to be the only doc, marked draft 1258986014 J * ghislain ~AQUEOS@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr 1258986065 M * BenG okay, it does look like all of this is a work in progress 1258986131 M * BenG nobody has added any real world examples yet, shame 1258986139 M * BenG I could really do with clueing up 1258986261 Q * ghislain Read error: Connection reset by peer 1258986411 Q * ghislain1 Ping timeout: 480 seconds 1258986437 M * bobnormal beng: maybe try to contact some of the larger-scale users listed on the users page? 1258987110 M * BenG I understand pretty much all of it, but the cpusets values I don't understand at all 1258987254 J * raceme ~tof@ombos.raceme.org 1258989029 M * bobnormal BenG, i haven't used it but basically it looks like it's just a percentage-based division of all available cycles 1258989047 M * BenG the soft scheduling yes 1258989055 M * BenG I'm having ago now 1258989104 M * bobnormal good luck :) it looks like it works judging by that page. it was written by ghislain who was here a few minutes ago so if you have problems maybe ask him 1258989933 Q * arachnist Remote host closed the connection 1258989941 M * blathijs BenG: Just stick around here a bit, there's probably devs who can tell you about it (like Bertl_zZ or daniel_hozac) 1258989996 M * BenG sure blathijs, I'll be calling out if I get into difficultly 1258990383 N * Bertl_zZ Bertl 1258990387 M * Bertl morning folks! 1258990401 M * bobnormal anyone know how to solve problems with gentoo as vserver guest when using 'emerge'? - /usr/lib64/portage/bin/ebuild.sh: /usr/lib64/portage/bin/filter-bash-environment.py: /usr/bin/env: bad interpreter: No such file or directory 1258990404 M * bobnormal morning bertl :) 1258990427 M * Bertl get a recent util-vserver :) 1258990441 M * bobnormal bertl: that's all? 1258990452 M * Bertl very likely 1258990458 M * fback morning Bertl! 1258990463 M * bobnormal Bertl, i am on 0.30.215-r3 1258990500 M * fback is 2.6.31.6 + vs2.3.0.36.24 right combo to test ipv6 functionality? 1258990510 M * Bertl bobnormal: which is almost two years old 1258990528 M * Bertl fback: yep, should be fine 1258990556 M * bobnormal Bertl, great, thanks! that sounds likely then... i will upgrade to newest not-necessarily-stable version provided by gentoo, and if that fails i'll go to svn or something :) 1258990585 M * Bertl make that, and let us know how it goes ... 1258990590 M * fback Bertl: single ip special case is for ipv4 only? 1258990610 M * Bertl nope, affects both, v4 and v6 1258990618 M * fback (so i can safely use it where I don't need loopback and want to use ipv6?) 1258990625 M * bobnormal Bertl, is moving to 0.30.216_pre2855 enough? 1258990680 M * Bertl should be, although 2855 has a minor bug in vdlimit output 1258990693 M * bobnormal Bertl, im not planning to use resource limits yet 1258990738 M * Bertl it works fine, it just confuses the hell out of the testfs script :) 1258990775 M * bobnormal bertl, installed, restarted vserver-utils service + vprocunhide script, started vm, same issue. 1258990787 M * bobnormal Bertl, i guess i need to go newer? 1258990815 M * Bertl how did you build your guest? 1258990823 M * fback Bertl: so I should turn off single ip special case for every guest that has just single ipv4 AND single ipv6 interface, or not? :) 1258990833 M * bobnormal bertl: using template from a gentoo stage3 archive originally, this one is a rsync'd version of that. 1258990966 M * bobnormal Bertl, strange, seems a few binaries are missing. i did ln -s /bin/env /usr/bin/env and got further, but now 'C compiler cannot create executables' ... ld and asm are missing, possibly amongst other things 1258990988 M * Bertl how did you rsync it? 1258991009 M * Bertl fback: one ipv4 and one ipv6 are _two_ IPs :) 1258991016 Q * scientes Ping timeout: 480 seconds 1258991044 M * bobnormal Bertl, vserver newname build --context 1234 --hostname blah --interface eth0:1.2.3.4/24 --initstyle plain -m rsync -- -d gentoo --source=/vservers/reference -o -r -o -l -o --safe-links -o --devices -o --specials -o --delete 1258991050 M * fback ok... so I should probably drop special casing completly... 1258991094 M * fback most guests have ipv4 and ipv6 address, and some other needs virtualized loopback :) 1258991097 M * Bertl if you don't have single IPs and don't want the performance gain, yes 1258991099 M * PowerKe bobnormal: if your host is also gentoo, you can emerge python or portage in the guest by using a chroot 1258991128 M * bobnormal PowerKe, surely stage3 contains that? 1258991146 M * fback Bertl: the only guest without ipv6 requires virtualized loopback, so... 1258991151 M * PowerKe bobnormal: It should 1258991157 M * Bertl bobnormal: and a recent baselayout (openrc?) works without cleanups 1258991188 M * bobnormal Bertl, #define recent ? im using this month or last month at the latest 1258991219 M * Bertl should be fine, regarding rsync, I guess your options to rsync might have done some of the damage 1258991247 M * Bertl i.e. the rsync the build method uses should already contain everything you need 1258991257 M * bobnormal Bertl, im already retrying with those removed ... hopefully it works :) 1258991273 M * Bertl the reference guest works as expected? 1258991284 M * Bertl (and contains all the binaries?) 1258991337 M * bobnormal Bertl, verifying that now... 1258991357 M * bobnormal Bertl, think this might be the cause, since the rsync now takes FOREVER :) 1258991414 M * PowerKe bobnormal: does the reference also include a portage tree? 1258991419 Q * Mr_Smoke Read error: Connection reset by peer 1258991424 M * Bertl a good option, if you have the reference guest locally, and use unification, is the 'clone' method 1258991441 M * bobnormal PowerKe, all VMs mount a 'bind,ro' version of the tree from the host system 1258991454 M * Bertl with properly unified guests it takes just a few seconds to 'clone' one 1258991467 M * bobnormal Bertl, is it very manageable? 1258991471 J * Mr_Smoke ~smokey@layla.lecoyote.org 1258991474 J * dowdle ~dowdle@scott.coe.montana.edu 1258991489 M * bobnormal Bertl, reference works 100% 1258991513 M * Bertl on gentoo, you can improve with the portage bind mount, but yes, works fine for all known distros 1258991542 M * bobnormal Bertl, cool i will read up on it then. thanks for your help. still hasn't finished rsync'ing with the default options! 1258991739 M * fback Bertl: leaving automatic loopback should be safe for ipv4 / ipv6 guests? 1258991922 M * Bertl every selectable kernel option (Linux-VServer related) is safe :) 1258991943 M * Bertl BenG: what's the problem? 1258991988 M * BenG Bertl, I'm looking at scheduling,, but actually getting stuck at just building a guest on Debian Sid 1258991998 M * BenG :( 1258992043 M * Bertl http://linux-vserver.org/Building_Guest_Systems 1258992079 M * BenG I've followed: "Building guests using the debootstrap build method" 1258992090 M * Bertl okay, and? 1258992113 M * BenG appears to build, then deletes everything 1258992121 M * BenG last line of output is 1258992127 M * Bertl util-vserver version? 1258992129 M * BenG /bin/rm: invalid argument: `' 1258992146 M * BenG 0.30.216~r2855 1258992173 M * Bertl let's see, I have that installed, please paste your build line somewhere 1258992186 M * Bertl (pre2855, not the debian version of course :) 1258992240 M * BenG http://pastebin.org/56385 1258992269 M * BenG I used a local domain though, not example.net 1258992290 M * Bertl your interface is called intel0 .. interesting :) 1258992296 M * BenG but true! 1258992304 M * BenG changed using udev 1258992319 M * BenG for reasons of convienience 1258992333 J * SubZero ~SubZero@chello089076140236.chello.pl 1258992391 M * Bertl okay, that version tells me that debian messed again with the debootrap binary, did you fix the path on your side or is debootstrap installed? 1258992429 M * BenG deboostrap is installed 1258992432 M * Bertl daniel_hozac: debootstrap path changed again, JFYI 1258992436 M * bzed are you using lenny or testing/sid? 1258992444 M * BenG testing/sid 1258992460 M * Bertl neither nor 1258992462 M * fback Bertl: finally reconfigured home lan, so ipv6 network is available by default. I'll report any problems, althou my setup is really simple 1258992468 M * bzed debootstrap fails for another reason there also probably 1258992628 M * BenG " did you fix the path on your side" Bertl, what do mean by that exactly 1258992788 M * Bertl yes, I fixed it, but it seems the recent debootstrap is broken 1258992803 M * Bertl trying with the lenny one now 1258992908 M * Bertl doesn't work either, yeah I guess we have to wait for daniel_hozac here 1258992997 M * bobnormal Bertl, takes me 50 seconds now to rsync a new vserver without hashify. i tried to use hashify but it says 'Can not determine packagemanagement style / failed to determine configfiles' when i run vserver reference hashify 1258993029 M * Bertl BenG: ah, no, sec, I'm getting the -arch/--arch wrong again 1258993083 M * Bertl bobnormal: on the reference guest or the new one? 1258993092 M * bobnormal Bertl, reference 1258993176 M * Bertl try to configure it as 'gentoo' guest then 1258993200 M * bobnormal Bertl, do i need to re-build or can i edit something in /etc/vservers/reference/* 1258993228 M * Bertl editing is fine, compar it to the rsync copy, the -d gentoo sets it 1258993232 M * Bertl *compare 1258993285 M * Bertl BenG: building with the lenny debootstrap right now ... when completed, I'll test with the latest one 1258993299 M * BenG cheers 1258993390 M * Bertl 91 3 44.6M 1.2M 0m00s20 0m00s10 0m05s93 twotime 1258993395 M * Bertl works fine ... 1258993402 M * Bertl now to the recent debootstrap 1258993611 M * Bertl 91 3 44.6M 1.2M 0m00s00 0m00s20 0m04s17 twotime 1258993648 M * Bertl http://paste.linux-vserver.org/13898 1258993692 M * Bertl works fine here, I'd suggest you uninstall the debootstrap on the host, and fix the debootstrap url in /etc/vservers/.defaults/apps/debootstrap/uri 1258993725 M * Bertl then start again, the --arch amd64 is just because debian calls x86_64 amd64 (i.e. not required in your case) 1258993907 M * BenG /etc/vservers/.defaults/apps/debootstrap/uri is blank for me 1258993911 M * BenG what should it be? 1258993932 M * Bertl http://ftp.debian.org/debian/pool/main/d/debootstrap/debootstrap_1.0.20_all.deb 1258993938 M * bobnormal Bertl, can't find any gentoo-related stuff in the -d gentoo clone's /etc/vserver/x dir 1258993977 M * Bertl check apps/pkgmgmt/internal 1258994032 M * Bertl or simpler, try to unify the copy instead of the reference first 1258994052 M * Bertl if that works fine, you know it is the config you need to fix :) 1258994170 M * bobnormal Bertl, doesn't work. 1258994224 M * Bertl fails with? 1258994238 M * bobnormal Bertl, same error as above ... cannot determin pkg management style 1258994256 M * bobnormal Bertl, does the guest need to be running? :) maybe i forgot that ... 1258994280 M * bobnormal Bertl, no diff. 1258994479 M * Bertl hmm, I see that error on a posting from 2006, where the question was if unification makes senso on gentoo (which I think it does :) but it looks to me like there is no handler for gentoo yet to determine the config files (which shouldn't be a big deal anyway) but maybe gentoo folks never tested that 1258994503 M * bobnormal ok no worries, it's small enough anyway 1258994548 M * Bertl nah, try to disable the pkgmgm 1258994556 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1258994568 M * Bertl search for vunify 1258994596 M * Bertl add a manual 'exclude' list containing e.g. /etc 1258994613 M * Bertl and touch the pkgmgmt-ignore 1258994630 M * bobnormal /etc/vservers/.defaults/apps/vunify/pkgmgmt-ignore # i think 1258994668 M * Bertl well, I'd make it guest specific, but if you just have gentoo guests, yeah, why not 1258994707 M * bobnormal yep im doing reference specific .. hashify is running .. unsure what the outcome will be :) 1258994723 M * bobnormal previously 49 seconds per clone, i'll see how many i can save :) 1258994753 M * Bertl you'll be surprised :) 1258994792 M * bobnormal Bertl, thanks a lot for your help. really need vservers for our applications as we need io throughput for video processing ... hypervisors are way too slow 1258994816 Q * fback Quit: kernel upgrade 1258994818 M * bobnormal Bertl, previously we used freebsd jails, but package management under freebsd = masochism 1258994825 M * Bertl you're welcome! 1258994838 M * bobnormal :) 1258994857 J * arachnist arachnist@insomniac.pl 1258994961 Q * davidkarban Quit: Ex-Chat 1258995040 M * bobnormal bertl, so i need to do -m clone as the vserver x build ... argument to take advantage of hashify, right? how can i determine if the source vserver is hashified or not? 1258995069 M * BenG Bertl, seems to work now, not sure what was going on there 1258995084 M * bobnormal wow, from 49 to 2.8 seconds 1258995090 M * bobnormal hashify is godlike :) 1258995119 M * Bertl bobnormal: that's a tough one, each guest can be completely hashified or not hashified at all (or anything inbetween) 1258995158 M * Bertl so, probably the best way is to make sure to hashify on a regular basis (after changing a guest or updating) 1258995184 M * Bertl besides that, clone will wor in any case, just faster on unified guests :) 1258995222 M * Bertl and it isn't just a build advantage or reduction in disk space, you also get a big memory advantage from unified guests 1258995234 M * bobnormal Bertl, so i guess every time a potentially cloned vserver is shut down i could just re-hashify? 1258995272 M * Bertl you can rehashify anytime ... 1258995279 M * bobnormal Bertl, great :) 1258995291 A * bobnormal loves linux-vserver 1258995336 M * Bertl ah, right moment for: http://linux-vserver.org/Donations I guess :) 1258995361 M * bobnormal bertl, if i had any money i would :) but will suggest to my company 1258995403 M * Bertl not just money we can use (see page :) 1258995452 M * bobnormal noted! 1258995628 M * bobnormal Bertl, unsure how this hard-linking operates, but i have /vservers/.hash 82M, /vservers/reference 42M, /vservers/clone0 541M ... ? 1258995637 M * bobnormal Bertl, that's du -h -s output 1258995657 M * Bertl try to run it separately :) 1258995700 M * bobnormal aha! wow. hard links are like black magic, affecting otherwise trustworthy tools in evil evil ways ... 1258995724 M * bobnormal .hash 581M, each clone 541M. nice. 1258995755 Q * BenG Quit: I Leave 1258995854 M * Bertl it's basic unix hard links, with some sugar to make it more guest friendly (including CoW link breaking) 1258995930 M * Bertl but the benefit is huge, especially if you have lots of almost identical guests ... 1258995948 M * bobnormal its great :) no further comment necessary 1258996406 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1258996551 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1258996885 J * fback fback@red.fback.net 1258997157 M * fback Bertl: just one issue noticed, oidentd binds to 0.0.0.0 or ::, but not to both, but this may be identd problem 1258997177 M * fback otoh it works on non-vserver dual stack 1258997495 Q * fback Quit: Reconnecting 1258997496 J * fback fback@red.fback.net 1258997513 M * daniel_hozac which kernel? 1258997540 M * fback daniel_hozac: 2.6.31.6 + vs2.3.0.36.24 1258997573 M * fback forcing it to :: did the trick 1258997585 M * daniel_hozac hmm. 1258997590 M * daniel_hozac i thought that was supposed to be fixed. 1258997616 Q * BenG Quit: I Leave 1258997625 M * fback daniel_hozac: sshd automagically listens on both 0.0.0.0 and :: 1258997654 M * fback so this may be a problem with oidentd 1258998308 Q * derjohn_mob Ping timeout: 480 seconds 1258998650 Q * ruskie Read error: Connection reset by peer 1258999576 J * derjohn_mob ~aj@c193169.adsl.hansenet.de 1258999645 Q * barismetin Quit: Leaving... 1259000046 J * dna ~dna@p54BCA39A.dip0.t-ipconnect.de 1259000355 Q * gnuk Quit: NoFeature 1259000664 Q * kir Quit: Leaving. 1259001054 M * _Shiva_ fback: it really seems to depend on oidentd's version.. there are various Bugreports (on non-vserver systems) that are all resolved by: >> just add "-a ::" to make oidentd listen to v6 again << 1259001069 A * _Shiva_ ran into this just yesterday 1259001104 M * Bertl sounds plausible .... vs2.3.0.36.24 has the proper ipv6 fixes 1259001161 Q * SubZero Read error: Connection reset by peer 1259001233 M * _Shiva_ other possible cause may be that one here: http://www.mail-archive.com/debian-bugs-dist@lists.debian.org/msg686678.html 1259001260 M * _Shiva_ (which also lies w/i oidentd's code) 1259001439 J * ruskie ruskie@ruskie.user.oftc.net 1259001585 M * _Shiva_ . o 0 ( hmm .. and it may also be just when running on a bridged interface.. ) 1259001738 M * fback _Shiva_: it's 2.0.8 on both machines 1259001747 M * _Shiva_ fback: distro? 1259001794 Q * ensc|w Remote host closed the connection 1259001823 M * fback OpenWRT (works) and debian with custom kernel (vanilla 2.6.31.6 + vs2.3.0.36.24) 1259001957 M * _Shiva_ fback: there are bug reports (i.e. #520134 @ bugs.debian.org) that state it might also be glibc related.. that would fit into OpenWRT as it uses micro-libc 1259002012 M * _Shiva_ fback: i ran into this issue on Gentoo running net-misc/oidentd-2.0.8-r2 1259002029 M * _Shiva_ also on 2.6.31.6-vs2.3.0.36.24 1259002060 M * _Shiva_ fback: have you tired the sysctl setting mentioned? 1259002068 M * fback _Shiva_: possible, I reported similar issues with something on OpenWRT (thus using uClibc) 1259002104 M * _Shiva_ fback: strace -f -e trace=bind oidentd -d 1259002112 M * fback they provided some patches against uClibc that solved it, I have no idea if upstream included them 1259002129 M * fback _Shiva_: and I don't mind adding -a :: to options :) 1259002288 M * _Shiva_ fback: but then it's listening on v6 only.. ;-) (ok - v4 too, but using ::v4.ad.dr.ess) 1259003046 Q * opuk Ping timeout: 480 seconds 1259003085 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk 1259003208 M * BenG right then, I've got a set up to test scheduling, but I don't know how to test if it's working, any ideas? 1259003229 M * BenG soft scheduling that is 1259003616 M * BenG I'm guessing I should see the priorities change in top 1259004880 J * opuk ~kupo@pipe.intertubez.net 1259004983 Q * sardyno Quit: leaving 1259005028 J * sardyno ~me@pool-173-75-5-88.pitbpa.fios.verizon.net 1259009026 J * imcsk8 ~ichavero@nat.ti.uach.mx 1259009243 Q * BenG Quit: I Leave 1259009677 Q * AmokPaule Remote host closed the connection 1259010385 J * AmokPaule ~amokpaule@brsg-4dbb1e65.pool.mediaWays.net 1259012727 Q * bonbons Quit: Leaving 1259013938 Q * dna Quit: Verlassend 1259016613 M * micah Bertl: whats the situation with the testfs CoW attribute flags situation? 1259017074 M * Bertl I presume it is fixed in svn, but no release from daniel_hozac yet 1259017327 M * Bertl OTOH, there is a kernel patch for 2.6.32-rc8 you folks could test/package 1259018760 Q * geb Quit: / 1259018913 J * ptiphus ~guillaume@APuteaux-554-1-66-134.w82-124.abo.wanadoo.fr 1259019136 M * Bertl off to bed now ... have a good one everyone! 1259019140 N * Bertl Bertl_zZ 1259019817 Q * gat Quit: gat 1259020109 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1259020133 Q * BenG 1259020495 Q * imcsk8 Quit: Leaving 1259020557 J * aland_ ~aland@81.95.52.68 1259020773 Q * dowdle Remote host closed the connection