1156550413 M * nayco quota-3.13-1mdk, precompiled binary 1156550505 M * nayco \o/ 1156550513 M * nayco Quota work in vserver too ! 1156550532 M * nayco (provided that one does not use mountpoints as parameters ! 1156550543 M * Bertl of course ... 1156550586 M * nayco is it helping you ? 1156550699 M * nayco (___(__________(#~ 1156550722 M * nayco going for 7 centimeters... 1156550732 M * Bertl k, getting the quota package :) 1156551845 M * Hecticee vcontext: vc_set_cflags(): Operation not permitted 1156551845 M * Hecticee An error occured while executing the vserver startup sequence; when 1156551845 M * Hecticee there are no other messages, it is very likely that the init-script 1156551845 M * Hecticee (/sbin/init) failed. 1156551849 M * Hecticee whats wrong? 1156551895 M * Bertl could you run testme.sh and upload the output to paste.linux-vserver.org please? 1156551921 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh 1156551933 M * Hecticee [000]# succeeded. 1156551933 M * Hecticee [001]# succeeded. 1156551933 M * Hecticee [011]# succeeded. 1156551933 M * Hecticee [031]# succeeded. 1156551933 M * Hecticee [101]# succeeded. 1156551935 M * Hecticee [102]# succeeded. 1156551935 M * Hecticee [201]# succeeded. 1156551937 M * Hecticee [202]# succeeded. 1156551969 M * Bertl well, nice, but a) we are missing the _important_ info (first few lines) and b) please use the pastebin 1156552001 M * Hecticee oki 1156552007 M * Bertl tx 1156552014 M * Hecticee so whats about a point? 1156552077 M * Hecticee Bertl? 1156552143 M * Bertl testme.sh -> paste.linux-vserver.org then we probably know more 1156552201 M * Hecticee http://paste.linux-vserver.org/298 1156552205 M * Bertl tx 1156552296 M * Bertl ah, you are using the grsec enhanced version 1156552316 M * Hecticee ye 1156552318 M * Bertl looks like this might apply here: http://www.paul.sladen.org/vserver/archives/200512/0260.html 1156552554 Q * Hecticee 1156552606 M * Bertl nayco: ping? 1156552950 M * nayco pong § 1156552987 M * Bertl could you do an strace -fFv (note the 'v') of the failing repquota (with a single mountpoint) and upload that please? 1156553393 M * ntrs will there be a rc30? 1156553409 M * Bertl for stable or devel? 1156553422 M * ntrs both. 1156553436 M * Bertl stable is already there, devel will probably skip to rc31 1156553459 M * Bertl I know, topic is not updated yet :) 1156553490 M * ntrs anything critical happened between devel rc29 and devel rc31 stability wise or security wise? 1156553518 M * Bertl no, you are already on 2.6.17.11, IIRC, so security should be fine 1156553719 M * ntrs ok, thanks. 1156553727 M * nayco Bertl: http://paste.linux-vserver.org/299 1156553810 M * Bertl tx 1156553944 J * yarihm ~yarihm@84-74-17-70.dclient.hispeed.ch 1156553972 M * nayco Bertl: I thnik I gotta go to bed... Is it a problem now ? Do you need more "pastes" ;-) ? 1156554071 M * Bertl nah, that seems fine, tx 1156554082 M * Bertl have a good night then, cya tomorrow? 1156554186 M * nayco Yes :-) ! Thanx for all ! 1156554192 M * nayco 'night 1156554345 Q * nayco Quit: My sleep quota is low ! 1156554476 M * derjohn2 Bertl, what about rc31 ? Are those 'release c.'s' ? 1156554508 M * Bertl well, it's kind of confusing, I know, but we had to do some fixes 1156554541 M * Bertl so we will have *rc31 for both pretty soon 1156554542 M * derjohn2 it was there as file but the topic here didnt change. what is the refernce URL for RCs ? 1156554554 M * derjohn2 wel'll skip rc30 ? 1156554555 M * Bertl nah, I just forgot to update the topic 1156554573 M * Bertl but yes, we can skip rc30 now (almost) 1156554602 M * derjohn2 Bertl, so I wait for rc31 before my next compile run? I assume rc31 stable wont be a release= 1156554604 M * derjohn2 ? 1156554633 M * Bertl I 'hope' that rc31 will be released 1156554650 M * derjohn2 stable. or devel, too? 1156554664 M * Bertl stable, devel will take a little longer 1156554679 M * Bertl cleanups and so are still missing there 1156554772 M * derjohn2 ok, thx! abiut the changelog; what to write for rc30? 1156554822 M * Bertl nothing, we just list rc31 with _all_ the changes, no? 1156554831 M * derjohn2 (devel rc30 wasnt released at all, so i write 'no changes' for devel?) 1156554869 M * derjohn2 Bertl, OK, I'll skip that release silently :) (he, no work ;)) 1156554952 M * derjohn2 n8 ! 1156554953 M * ntrs derjohn, is there an official changelog somewhere? 1156554969 M * Bertl derjohn2: good night! and tx! 1156554982 M * ntrs derjohn2, is there an official changelog somewhere? 1156554992 M * ntrs derjohn2, good night. 1156555002 M * derjohn2 ntrs, http://linux-vserver.org/ChangeLogExperimental 1156555033 M * derjohn2 at least for the -rcs. IN real releases there will be a changelon somewhere with the release file I hope. 1156555046 J * nachoguy ~boster@216.146.94.55 1156555056 M * Bertl yes, definitely, we have that for all previous releases too 1156555063 M * Bertl welcome nachoguy! 1156555063 M * ntrs Oh, that's a beautiful changelog. 1156555076 M * nachoguy hello 1156555083 M * Bertl derjohn2: see, folks like it >:) 1156555120 M * derjohn2 Bertl, c00l, now I sleep even better *tired* :) I'll keep it clean ;) 1156555156 M * nachoguy so, I think I've done the undoable 1156555177 M * Bertl sounds interesting ... :) 1156555203 M * nachoguy http://sial.org/pbot/19269 1156555312 M * Bertl hmm, you meant the routeing table list? or what? 1156555328 M * ntrs Bertl, from the latest changelog entry for devel "Guests can now use private loop-mounts". What does that mean in practice? 1156555344 M * nachoguy I mean I have a routing table, without a network interface 1156555354 M * nachoguy I didn't realize it could be broken this way 1156555358 M * Bertl ah, okay, which kernel version is that? 1156555369 M * Bertl (best upload the output of testme.sh) 1156555390 M * Bertl ntrs: it means, that guests could have privacy from the 'host' admin (to some degree) 1156555534 M * nachoguy I'm running a 2.6.17.8 kernel (latest stable (IIRC)) but the problem is because I attempted to bind it to br0 (a bridge I set up) 1156555570 M * Bertl interesting, how did you 'bind' it to a bridge? 1156555601 M * Bertl you mean, the bridge is carrying the ip address, right? 1156555613 M * ntrs Bertl, with devel, could I give the guests a separate /tmp loop mount so that I can mount it as noexec but at the same time one that will not utilize any ram as it is the case with the stable patch? 1156555627 M * nachoguy when I built the image, I used --interface br0:192.168.0.3/24 1156555653 M * nachoguy when i use the interface eth0 it's fine and works as expected 1156555654 M * Bertl ntrs: yes, but you can do the same with stable 1156555674 M * ntrs Bertl, hmm, how? 1156555698 M * ntrs If I put it in the fstab config file it would use the memory right? 1156555701 M * Bertl nachoguy: ah, interesting ... so the bridge seems to change the checks (probably answering to 'more' ips ... 1156555732 M * nachoguy I'm just trying to make it so the physical host has an ethernet port and another device that the virtual devices connect to 1156555746 M * Bertl ntrs: not if it is a loopback device, but I would not suggest to do so, because the 'memory' it uses is faster than the loop, and it's swap backed anyway :) 1156555792 M * Bertl nachoguy: hmm, in Linux-VServer, there is ip isolation, so no need to 'connect' any 'virtual' devices 1156555812 M * Bertl nachoguy: nevertheless the routing table looks like a bug 1156555819 M * ntrs Yes, but some guests need large /tmp dirs and it is not practical to give them a large partition that fully resides in memory. 1156555841 M * Bertl ntrs: well, then make it an lvm volume (or other partition) 1156555875 M * Bertl or 'just' remove the tmp completely 1156555891 M * ntrs But the guests need /tmp, how could I just remove it? 1156555904 M * nachoguy but they all are more or less bridged to eth0 though? Even with ip isolation? doesn't taht mean that the main physical machine can't filter the virtual machines' traffic using iptables 1156555906 M * ntrs Bertl, i do not use lvm and the whole disk is already prepartitioned. 1156555925 M * Bertl well, make it a 'normal' dir or a link to /var/tmp (which is, btw, the place where larger files should go, by default) 1156555986 M * Bertl nachoguy: no, they are not bridged or anything else (unless you configure that :) the networking happens as on every other linux box 1156556009 M * Bertl nachoguy: so if you have three ips 'assigned' to eth0, you can use one (or more) of them for a guest 1156556033 M * Bertl iptables, routing and tc will work as normal 1156556045 M * ntrs Bertl, yes, I could do that but how do you make that noexec? 1156556051 M * Bertl nachoguy: (and all the networking will happen on the host) 1156556206 M * Bertl ntrs: with BME? 1156556228 M * Bertl mount --bind -o noexec /tmp /tmp (for example) 1156556265 M * ntrs inside the guest? 1156556276 M * ntrs Can it be done in the fstab vserver config? 1156556283 M * Bertl sure, why not? 1156556299 M * nachoguy hrm, also, you should not build a vserver and use --interface dummy0 1156556303 M * nachoguy bad things happen 1156556326 M * Bertl just 'dummy0' ? 1156556344 M * nachoguy no dummy0:/24 1156556346 M * Bertl that's not a proper interface specification, should give an error 1156556357 M * Bertl ah, that's fine 1156556363 M * nachoguy apparently, it's not 1156556365 M * nachoguy heh 1156556370 M * Bertl what 'bad' things do you observe? 1156556420 M * nachoguy actually, bah, I misspoke 1156556436 M * nachoguy I typoed. what I have is not nearly as interesting 1156556449 M * ntrs Bertl, what do I need to change the following line to? 1156556450 M * ntrs none /tmp tmpfs noexec,size=16m,mode=1777 0 0 1156556487 M * nachoguy I thought I had started that vserver, I infact started another. so when trying to enter a vserver that isn't running, one should facepalm 1156556518 M * Bertl okay :) 1156556578 M * Bertl ntrs: try something like: 1156556600 M * Bertl /path/to/guest/tmp /tmp none bind 0 0 1156556613 M * Bertl ah, bind,noexec :) 1156556672 M * ntrs '/vserver//tmp /tmp none bind,noexec 0 0' 1156556693 M * ntrs or '/tmp /tpm none bind,noexec 0 0' 1156556695 M * ntrs ? 1156556695 M * Bertl well, probably /vserver_s_, but yes 1156556735 M * ntrs ok 1156556863 T * Bertl http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc31, stable+grsec 2.0.2-rc28 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1156556880 M * Bertl *rc31 1156557060 J * mire_ ~mire@212-167-222-85.COOL.ADSL.VLine.verat.net 1156557090 M * harry rc31 allready??? 1156557101 M * Bertl yeah, we skipped rc30 (basically) 1156557106 A * harry will make a interdiff for rc28-rc31 soom 1156557111 M * harry soon 1156558064 M * Bertl okay, guess I'm off to bed for tonight ... have a good one everyone! cya later! 1156558070 N * Bertl Bertl_zZ 1156558554 M * nachoguy hrm, quick question, if anyone knows off hand, do the vservers share a routing table with the physical server? 1156558790 M * nachoguy quick answer, is yes 1156559185 Q * DreamerC Quit: leaving 1156559460 J * DreamerC ~dreamerc@59-112-20-59.dynamic.hinet.net 1156560400 Q * yarihm Quit: This computer has gone to sleep 1156560803 J * MrX ~urk@219.95.234.6 1156561021 N * lolilol olilo 1156561105 Q * FireEgl Ping timeout: 480 seconds 1156561321 J * FireEgl FireEgl@Sebastian.Atlantica.US 1156565462 J * alchemus ~kvirc@c-68-45-119-139.hsd1.nj.comcast.net 1156565538 Q * alchemus 1156565545 J * alchemus ~kvirc@c-68-45-119-139.hsd1.nj.comcast.net 1156566747 Q * ekc Ping timeout: 480 seconds 1156567290 Q * alchemus Quit: KVIrc 3.2.0 'Realia' 1156568078 Q * derjohn2 Ping timeout: 480 seconds 1156568461 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1156569831 J * s0undt3ch_ ~s0undt3ch@bl7-240-6.dsl.telepac.pt 1156569929 J * kir_home tis-751532@213.152.157.70 1156569966 Q * s0undt3ch Ping timeout: 480 seconds 1156569973 N * s0undt3ch_ s0undt3ch 1156571398 Q * kir_home Ping timeout: 480 seconds 1156573144 M * phreak`` daniel_hozac: hrm, is your template.patch obsoleting the clone function ? 1156575175 J * dna ~naucki@114-223-dsl.kielnet.net 1156575675 M * matti Hm... 1156575700 M * matti Where -rc31 is? :) I cannot find any usable download address :) 1156576063 J * Administrator__ ~Administr@82.114.185.60 1156576148 Q * Administrator__ 1156578107 J * WhataDUCK ~id@p50812021.dip0.t-ipconnect.de 1156578114 N * WhataDUCK id23 1156578121 M * id23 greetz #vserver 1156578277 J * Viper0482 ~Viper0482@p54977DD8.dip.t-dialin.net 1156578654 M * Hollow phreak``: around? 1156578966 M * phreak`` Hollow: yup 1156578989 M * phreak`` id23: http://vserver.13thfloor.at/Experimental/?C=M&O=D 1156579026 M * Hollow phreak``: i'm adding rc31 atm 1156579057 M * phreak`` Hollow: ok, I'm currently working on util-vserver-0.30.210-r18 1156579114 M * Hollow i guess it's time to give the util-vserver patches numbers as prefix ;) 1156579156 M * phreak`` think so :) 1156579167 M * phreak`` 000 -> bugfixes 1156579178 M * phreak`` 010 -> feature additions 1156579179 M * phreak`` ? 1156579188 M * Hollow sth like that.. 1156579381 J * DreamerC_ ~dreamerc@59-112-19-152.dynamic.hinet.net 1156579776 Q * DreamerC_ Quit: leaving 1156579778 Q * DreamerC Ping timeout: 480 seconds 1156579829 J * DreamerC ~dreamerc@59-112-19-152.dynamic.hinet.net 1156579864 J * kir_home tis-81bc64@213.152.157.70 1156580277 J * coocoon ~coocoon@p54A056B2.dip.t-dialin.net 1156581067 J * derjohn2 ~aj@dslb-084-058-204-200.pools.arcor-ip.net 1156581508 J * bonbons ~bonbons@83.222.36.236 1156581946 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156582848 Q * Hollow Quit: Konversation terminated! 1156583000 J * Hollow ~hollow@2001:a60:f026::1 1156583163 Q * kir_home Ping timeout: 480 seconds 1156583260 J * yarihm ~yarihm@84-74-17-70.dclient.hispeed.ch 1156584202 J * pisco hiddenserv@tor.noreply.org 1156585549 J * nayco ~nayco@lns-bzn-53-82-65-0-115.adsl.proxad.net 1156585894 M * nayco Hello, all ! 1156588827 M * id23 -rc31 nice 1156589164 J * serving ~serving@86.108.87.70 1156589177 M * serving hi all 1156589198 M * serving I have a problem shutting down vservers 1156589215 M * serving vserver-stat hangs teh shh connection 1156589224 M * serving ssh 1156589264 M * serving how do I find a list of my veservers and force shut them down 1156589266 M * serving ? 1156589287 M * serving :) please help I think I am under some sort of an attack 1156589319 M * serving 06:51:27 up 16:02, 3 users, load average: 251.99, 251.55, 249.43 1156589334 M * serving and climbing 1156589371 Q * mnemoc Ping timeout: 480 seconds 1156589542 M * daniel_hozac phreak``: no, it's for building guests from tarballs, like a stagew. 1156589548 M * daniel_hozac s/stagew/stage3/ 1156589744 J * mnemoc ~amery@kilo105.server4you.de 1156590293 M * daniel_hozac serving: what does dmesg on the host say? 1156590329 M * daniel_hozac serving: probably you had an oops or similar while under an important lock, so things will start hanging, waiting for the lock to be released... 1156590340 M * daniel_hozac (this happened to me just a few days ago :)) 1156590682 M * harry serving: vserver-stat 1156591234 M * serving Daniel: I posted dmesg outout here 1156591235 M * serving http://paste.lisp.org/display/24773 1156591261 M * serving vserver-stat hangs the ssh connection 1156591271 M * daniel_hozac ls /proc/virtual 1156591315 M * daniel_hozac serving: so no oops or anything? 1156591322 M * serving ls: /proc/virtual: No such file or directory 1156591332 M * serving i am using 2.4 and old tools 1156591349 M * daniel_hozac oh, hehe. 1156591360 M * daniel_hozac i have no idea then. i've never used vserver on 2.4. 1156591364 M * serving :( it wasn't broken ;) 1156591398 M * serving do you know how to force reboot the server without stopping processes 1156591401 M * serving ? 1156591403 M * daniel_hozac reboot -f 1156591408 M * phreak`` daniel_hozac: if you have some time, could you take a look at http://dev.croup.de/proj/gentoo-vps/browser/util-vserver/patches/0.30.210-r18/000_README (and tell me if I missed someone/thing) ? 1156591499 M * daniel_hozac debootstrap-script is from Debian, so probably micah should be listed as the author on that one. 1156591525 M * daniel_hozac fstab doesn't implement fstab.local/fstab.remote, it fixes the unmounting to do the opposite of the mounting ;) 1156591585 M * daniel_hozac the testsuite-fix is your patch, i don't even have it (testsuite WORKSFORME). 1156591610 M * daniel_hozac hmm, what is the yum-2.6 patch? 1156591626 M * phreak`` daniel_hozac: iirc from your repo 1156591645 M * phreak`` *shrug* 1156591646 M * daniel_hozac cpuset should probably list Jan Rekorajski as the author. 1156591660 M * daniel_hozac (http://www.paul.sladen.org/vserver/archives/200511/0245.html) 1156591712 M * daniel_hozac Bastian Blank wrote the namespace-cleanup patch, i just fixed some bugs with it. 1156591828 M * phreak`` daniel_hozac: http://daniel.hozac.com/vserver/util-vserver-0.30.210-yum26.patch <-- yum-2.6.patch 1156592009 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156592009 Q * ntrs Read error: Connection reset by peer 1156592204 M * daniel_hozac phreak``: hmm, that's included in the more recent fc5 patches. 1156592257 M * daniel_hozac generally, the patches that are just in /vserver/ are q'n'd hack patches, or works in progress. 1156592275 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156592277 Q * ntrs_ Read error: Connection reset by peer 1156592325 M * phreak`` daniel_hozac: ah :) 1156592551 M * serving oh well, rebooting does your body good :D 1156592587 M * serving Thanx daniel_hozac for your help 1156592645 M * serving Guys, do you know a reliable outfit I can move my vservers to while I work on my host hardware ? :) 1156592931 M * daniel_hozac phreak``: the rest looks fine. 1156593125 M * phreak`` daniel_hozac: ok, thanks a lot then (and yeah, thats why I asked you, you're probably the one person knowing which patch is from whom) 1156593197 M * daniel_hozac hehe. 1156593221 M * derjohn with devel rc31 (and rc29?) I made the observation that the load avg of the host is higher than expected. even if the machine has nearly 0 disk I/O and cpu (dual core ath64) is 95% idle, teh load avg is between 0.6 and 1. if I enable virtual load avg, the guests load avg seems to be calc'ed fine. I am using sched_hard. 1156593241 M * Hollow phreak``: the prefix patch is kind of useless for our ebuild... econf sets prefix anyway 1156593381 M * phreak`` dropped. 1156593402 M * Hollow did you add bash completion use flag to the ebuild? 1156593417 M * phreak`` Hollow: not yet :P 1156593428 M * Hollow ok.. just saw the patch.. nice thing 1156593446 A * phreak`` 's curently looking after something other (kde app to rip a audio-cd) 1156593468 M * daniel_hozac what do you guys think about a 0.30.210.1 with all the patches applied? 1156593503 M * Hollow daniel_hozac: heh, i thought about sth like that too yesterday... did you ask enrico if you can have commit access and/or make releases? 1156593522 M * phreak`` daniel_hozac: yeah, that would be probably be good (or even a .211) as tracking all the changes / fixing the patches sucks :) 1156593523 M * daniel_hozac not yet... i'm still hoping he'll get around to it. 1156593538 M * daniel_hozac phreak``: yeah, i'm starting to get conflicts between them now. 1156593555 M * daniel_hozac (well, conflicts have been around for a long time, but now it's rejects that can't be fixed by reordering them) 1156593612 M * daniel_hozac i guess i could ping him again and ask what happened, he told me a couple of weeks ago he'd look into a 0.30.211 with at least some of the patches. 1156593692 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156593692 Q * ntrs__ Read error: Connection reset by peer 1156593716 M * daniel_hozac derjohn: have you tried a vanilla kernel? 1156593752 M * daniel_hozac (i.e. are you sure it's vserver related? there haven't been any significant changes that would affect something like that, AFAICT) 1156593930 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156593930 Q * ntrs_ Read error: Connection reset by peer 1156593992 M * derjohn daniel_hozac, I build .deb always from vanilla (die the fw restirctions). I think devel got a new scheduler some timwe ago? 1156594004 M * derjohn huh ... 1400h ... sry ... /me away ..... 1156594033 M * daniel_hozac derjohn: i meant a really vanilla kernel, like without vserver ;) 1156594066 M * daniel_hozac derjohn: but the new scheduler has been around for a really long time, like since the early 2.1.1-rc days. 1156594106 M * daniel_hozac it did get a fix in... -rc29 though. 1156594139 M * daniel_hozac but IIRC that was an actual bug that would cause the context to hang under certain circumstances. 1156594289 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156594290 Q * ntrs__ Read error: Connection reset by peer 1156594840 N * Bertl_zZ Bertl 1156594844 M * Bertl morning folks! 1156594877 M * daniel_hozac morning! 1156595005 M * phreak`` morning Bertl! 1156595030 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156595031 Q * ntrs Read error: Connection reset by peer 1156595053 M * Bertl hey daniel_hozac! phreak``! how's going? 1156595125 M * phreak`` Bertl: fine :) had *a* look after gentoo's util-vserver (and finally added a README to track the patches) 1156595246 M * michal_ hey Bertl :] 1156595277 M * michal_ mode: vserver is great 1156595289 M * michal_ whatever machine i'm trying it on it always works 1156595292 M * michal_ mode: off 1156595297 M * nayco Hi Bertl :-)! 1156595306 M * michal_ really - only vanilla and vserver can boot this xseries 1156595326 M * michal_ xen hangs, even rsbac lockups, grsec lockups 1156595331 M * phreak`` michal_: what xSeries ? 1156595337 M * michal_ 220 1156595339 M * michal_ older one 1156595345 M * michal_ my workstation ;p 1156595376 M * phreak`` michal_: I'm having some 336/366 at work they work like a charm (at least with vServer/grsecurity) 1156595380 M * phreak`` ah, okee *g* 1156595385 M * Bertl nayco: I think I have a solution :) 1156595392 M * michal_ this 220 also works like a charm 1156595525 J * e-devil ~e-devil@cpe.atm2-0-7594.0x535a0562.boanxx18.customer.tele.dk 1156595541 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156595918 J * mkhl mkhl@200-148-41-158.dsl.telesp.net.br 1156595952 Q * ntrs_ Ping timeout: 480 seconds 1156596195 J * coocoon ~coocoon@p54A0598E.dip.t-dialin.net 1156596197 M * daniel_hozac Bertl, phreak``, Hollow: what bugs are left in util-vserver? the ones i can think of are: 1) context-mount should be a secure-mount option, 2) vshelper doesn't work on sync (i.e. without legacy) 1156596271 M * Hollow regarding the context mount.. bertls patch does mounting always inside, right? isn't secure_moun flag needef ro this? 1156596290 M * Bertl daniel_hozac: IIRC, a lot of problems arise with a guest specific lvm/root mount 1156596307 M * daniel_hozac i think Bertl's patch works in an odd way, because it tags devices before the context really exists... 1156596310 M * Bertl daniel_hozac: also the 'missing' init environment 1156596345 M * Bertl daniel_hozac: yeah,we might want to clean that up on kernel layer 1156596368 M * Bertl i.e. provide a syscall to do a context tagged mount or to 'tag' a device/mount 1156596411 M * Bertl but mounting right after context creation (with all caps) would suffice (IMHO) 1156596444 M * Bertl a few things I do not consider bugs, but general issues are: 1156596462 M * Bertl - novlandev should be the default IMHO 1156596514 M * Bertl - scheduler should be updated to handle the new values 1156596542 M * Bertl (maybe break it down into separate files like rlimits?) 1156596561 M * daniel_hozac yeah, the scheduler configuration needs to be fixed. 1156596576 M * Bertl - ulimit handling could be improved, i.e. lift hardlimits of the host 1156596578 M * daniel_hozac but handling the new values would require 2.1.1 headers. 1156596603 M * Bertl well, you have them, no? :) 1156596634 M * daniel_hozac right, but util-vserver currently uses the stable headers. 1156596651 M * Bertl folks like ntrs are currently using my vdlimit and vsched utils to configure this, not quite nice :) 1156596651 M * daniel_hozac i guess enrico had some idea behind that choice. 1156596685 M * daniel_hozac vdlimit? 1156596691 M * Bertl the headers should be backwards compatible, except for some very old legacy calls 1156596711 M * Bertl but as we did not replace anything, those could be added by hand 1156596882 M * daniel_hozac i'm starting to think util-vserver should carry the kernel headers for all the releases it supports. 1156596898 M * daniel_hozac especially when 2.2 is released, removing all the legacy stuff. 1156596907 M * Bertl I do that in the vcmd (except that I do not use it yet :) 1156596931 M * daniel_hozac hehe. 1156596942 M * Bertl but yes, it should be fine to build libs with 'different' headers (quite fine) 1156596962 M * Bertl and have _all_ calls available only taking those from the proper vci/version 1156596974 M * daniel_hozac i also think enrico's lovely API choosing macros need to be rewritten to take an upper version. 1156597015 M * Bertl maybe have a talk with enrico regarding his plans? he seems unreachable for me atm :/ 1156597039 M * Bertl i.e. he is not replying to my emails ... 1156597049 M * daniel_hozac well, he's on freenode. 1156597068 M * Bertl indeed? which channel? 1156597075 M * daniel_hozac #fedora-devel, #fedora-extras 1156597077 M * daniel_hozac at least. 1156597087 M * Bertl interesting ... 1156597090 M * daniel_hozac i had to ping him there to get him to read my util-vserver email. 1156597113 M * Bertl already says a lot 1156597163 M * Hollow imo we should start releasing the next util-vservers on our own.. someone should bug enrico for cvs access 1156597437 M * Bertl daniel_hozac: just to clarify, would you be willing to 'officially' take over util-vserver? 1156597525 M * daniel_hozac do we want CVS access? 1156597541 M * Hollow svn? ;) 1156597544 M * daniel_hozac i guess i'm already doing most of the work... 1156597549 M * daniel_hozac yeah, that's what i was thinking. 1156597573 M * Hollow sure, you can get a repo beside vcd & friends 1156597587 M * Hollow i'd prefer it as well 1156597628 M * Bertl okay, so let me state that again, you would take over maintainership completely (if possible), right? 1156597680 M * daniel_hozac sure, if Enrico doesn't have the time/interest anymore. 1156597697 M * Bertl okay, that's what we are going to fi nd out 1156598507 Q * cehteh Ping timeout: 480 seconds 1156598585 M * e-devil is there an easy way to set up a default gateway for vserver? 1156598596 M * waldi -v? 1156598668 M * Bertl http://archives.linux-vserver.org/200311/0470.html 1156598701 M * e-devil thx 1156598721 M * Bertl np 1156598772 M * tokkee Can I use fuse inside a guest? 1156598789 M * nayco Bertl: Which solution ? 1156598812 M * Bertl tokkee: probably, but it might not be secure, unless you mount it from outside 1156598846 M * waldi Bertl: hmm, it should be rather easy to extend fuse to also take the xid into account 1156598873 M * Bertl yes, but 'userspace' filesystems can always crash the kernel 1156598918 M * Bertl (by injecting wrong/evil data) 1156598948 M * tokkee Hmmm... I keep getting "No such device or address", guess I created the wrong device :-/ 1156598951 M * waldi fuse per default allows a user to use its own executables 1156598963 M * waldi so it can't be that broken 1156598988 M * Bertl I have no problem with making/adding that as a cap 1156599000 M * waldi you need to modify fuse themself 1156599002 M * Bertl so that folks can decide how far they want to go 1156599045 M * Bertl the fuse interface is kernel side by now, IIRC, they are not using the nfs interfaces anylonger, or? 1156599075 M * waldi fuse was never nfs based 1156599082 M * waldi this was cfs 1156599107 M * waldi or so 1156599118 M * Bertl hmm, I remember early fuse stuff (by erez) using nfs, but maybe I'm wrong 1156599149 M * Bertl nayco: 'very' experimental patch, but it might just work :) 1156599190 M * Bertl http://vserver.13thfloor.at/Experimental/delta-vroot-feat01.diff 1156599366 J * e-dvl ~e-devil@cpe.atm2-0-7594.0x535a0562.boanxx18.customer.tele.dk 1156599366 Q * e-devil Read error: Connection reset by peer 1156599372 N * e-dvl e-devil 1156599563 M * nayco Bertl: Ok, give me a little hour, i'm busy with family ;-) 1156599604 M * Bertl np, whenever you get around 1156600121 J * |gerrit| ~kvirc@dslb-084-060-199-237.pools.arcor-ip.net 1156600430 N * ntrs__ ntrs 1156600489 M * ntrs OK, Bertl, do we now know for sure if daniel_hozac will take over util-vserver from enrico? Did you talk to enrico yet? When can we expect a new updated version with vdlimit and vsched? 1156600508 M * daniel_hozac what do you need a special vdlimit for? 1156600520 M * Bertl hehe, not yet, we haven't reached enrico yet 1156600582 M * ntrs Let me try to get him over here from the fedora channel. 1156600608 M * ntrs daniel_hozac, what do you mean by "special"? 1156600621 M * daniel_hozac ntrs: why doesn't vdlimit from util-vserver work for you? 1156600628 M * Bertl well, I'm there, so once he comes back, we should be able to clarify that 1156600636 M * daniel_hozac or does vdlimit do something other than disk limits? 1156600638 M * ntrs Ok, great Belu_zZz 1156600646 M * ntrs I meant Bertk 1156600651 M * ntrs ahhh, Bertl 1156600657 M * Bertl :) 1156600658 M * ntrs early morning here, sorry 1156600699 M * ntrs I think it was the formatting of the parameters or something. 1156600720 M * ntrs I don't remember exactly. 1156600727 M * daniel_hozac hmm? so devel changed something in this regard? 1156600730 M * Bertl might already be superceeded by the work daniel did so far 1156600742 M * daniel_hozac because i'm using disk limits on stable just fine. 1156600777 M * Bertl daniel_hozac: IIRC, it (the tool) allows to dump the current values in a way it can read them back next time (was relevant before your changes) 1156600804 M * daniel_hozac vdlimit? yes. 1156600827 M * Bertl yes, vdlimit 1156600869 M * ntrs daniel_hozac, yes that is correct. 1156600903 M * ntrs The output of the vdlimit was such that if you simply put it in a script, you can execute the script next time you reboot and all your limits will be set. 1156600934 M * daniel_hozac how is that a problem? 1156600948 M * Bertl which should not be required anymore with recent 0.30.210 + patches 1156600954 M * ntrs That is not the problem, but it only worked that way with Bertl's version of vdlimit 1156600967 M * daniel_hozac 0.30.210 supports disk limits in the configuration. 1156600981 M * ntrs I run .210 and it does not support. 1156600985 M * daniel_hozac you just set dlimits/0/{inodes_total,space_total,reserved,directory}. 1156601023 M * ntrs Are the current number of used inodes/space persistent now between reboots? 1156601042 M * daniel_hozac no, it saves the current values when the guest is shut down. 1156601052 M * daniel_hozac then it restores those when it's started. 1156601060 M * Bertl (means yes :) 1156601076 M * ntrs Ok, what if the kernel panics or the server is shut down without shutting down the guests first? 1156601087 M * daniel_hozac if the guest wasn't stopped cleanly (i.e. cache isn't available), it will run vdu again to recalculate the inodes and space. 1156601104 M * ntrs automatically? 1156601107 M * daniel_hozac yes. 1156601113 M * Bertl (which is better than using old values :) 1156601128 M * ntrs Well, you must be talking about some other version because I am 100% certain .210 cannot do that. 1156601134 M * daniel_hozac it can. 1156601136 M * daniel_hozac i got it in there. 1156601140 M * daniel_hozac see THANKS and NEWS. 1156601152 M * ntrs Ok, so that is not 210 anymore, that would be 211 or something. 1156601166 M * daniel_hozac no, it is in vanilla 0.30.210. 1156601182 M * Bertl (i.e. enciro accepted and included it) 1156601184 M * daniel_hozac it's even listed on the flower page, http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1156601197 M * ntrs you got to be kidding me. 1156601215 M * daniel_hozac http://linux-vserver.org/Disk+Limits was also updated once it was released. 1156601216 M * ntrs Bertl, is this as safe to use as your vdlimit app? 1156601351 M * Bertl well, I guess so ... 1156601380 M * Bertl but best is to try it out and provide _feedback_ ... 1156601673 M * nayco Bertl: Ready, I can compile something if you want :-D 1156601800 M * ntrs Bertl, the only problem I see with this new approach is after a non-graceful restart it would take a lot of time to actually start the guests because all the values would have to be calculated again. My current setup writes all the values once per hour in a script. It would be nice if util-vserver could write down the cache with the current values every X number of minutes for example. That way we could choose to run vdu again or not to 1156601800 M * ntrs run it if we need the guests started ASAP even though the numbers may not be 100% accurate. 1156601930 M * Bertl well, you have to make such 'feature requests' to daniel_hozac or enrico, or maybe Hollow for the vserver-utils, that's userspace stuff :) 1156601933 M * ntrs daniel_hozac, is it possible to force the cache update at a given time interval? 1156601988 M * Hollow Bertl: to stop confusion vserver-utils is now vcd, vstatd and vwrappers (vps, vtop and friends) 1156602098 Q * e-devil Ping timeout: 480 seconds 1156602216 M * Bertl Hollow: hmm, and how is that going to stop confusion? :) 1156602216 M * ntrs daniel_hozac, is it possible to force the cache update at a given time interval not only on shutdown? So that it does not have to be recreated after a reboot? 1156602241 M * Hollow Bertl: these are completely different names 1156602247 M * Hollow not util-vserver vs. vserver-utils 1156602250 M * Hollow everyone confused it 1156602261 M * Bertl yeah but we have vps vtop and such too? 1156602278 M * nayco Find another name, not related, there will be no confusion. 1156602281 M * Hollow yep, they are in the vwrappers package... with many many other wrappers 1156602310 M * Bertl okay, so how to refer to the 'sum' of tools now? h-utils? 1156602331 M * Hollow refer to it as vcd 1156602336 M * Hollow as a whole 1156602358 M * Bertl okay, tx, will do so ... 1156602358 M * nayco Something like "sheep", "aeroplane", "banana"... 1156602402 M * Hollow Bertl: no all tools in these packages are directly related to vcd, but "we" commonly refer to all of it as vcd 1156602404 M * nayco ;-) 1156602423 M * ntrs daniel_hozac, it now takes about a minute to do a vdu on a single 2.5 GB guest which means that in order to start them all 40 on the server it will now take 40 more minutes to start them all than it took before. 1156602440 M * daniel_hozac ntrs: only on a non-graceful stop. 1156602452 M * ntrs daniel_hozac, yes, but those happen as well. 1156602454 M * daniel_hozac ntrs: the same would be true for your old solution. 1156602459 M * ntrs No 1156602466 M * daniel_hozac how so? 1156602490 M * ntrs the old solution just reads the latest values and starts the servers even though the numbers may be slightly inaccurate as they are X minutes old. 1156602538 M * daniel_hozac so if you really want a cron-script doing it, write one. 1156602559 M * ntrs It would be nice to be able to force the cache update every X minutes and then not require the vdu update after a non-graceful stop 1156602560 M * Bertl hint: it might be a nice feature to have the vdu do it's update lazy 1156602561 M * daniel_hozac should only be a few lines... 1156602572 M * daniel_hozac Bertl: hmm? 1156602574 M * ntrs what is the format of the cache then? 1156602600 M * daniel_hozac ntrs: source /usr/lib/util-vserver/f:qunctions 1156602611 M * Bertl i.e. start with the last 'saved' value and calculate the correct values, once done, update them on the running guest 1156602611 M * daniel_hozac uh, -:q + vserver. 1156602617 M * ntrs daniel_hozac, yes, but how do I make the utils not to recalculate the numbers again after a non-graceful reboot? 1156602624 M * daniel_hozac then just saveDiskLimits /etc/vservers/. 1156602638 M * daniel_hozac ntrs: if the cache exists, they won't recalculate. 1156602646 M * ntrs daniel_hozac, ah, good. 1156602704 M * daniel_hozac Bertl: that would work. but if we have saved values, why wouldn't we just use those? 1156602734 M * Bertl IMHO it would be best to make that two level 1156602756 M * Bertl i.e. have values which are 'known good' and values which 'might be good' 1156602772 M * Bertl and a way to differentiate between them (e.g. a flag)\ 1156602778 M * Bertl thing is this: 1156602792 M * Bertl - admin shuts down guest xy 1156602804 M * Bertl - then copies over some files or removes some of them 1156602811 M * Bertl - then restarts the guest 1156602827 M * Bertl similar case for the ungraceful shutdown of a guest 1156602852 M * daniel_hozac he would have to copy over the files and chxid them, no? 1156602864 M * daniel_hozac but yeah, the removal is a problem. 1156602865 M * Bertl having a set of values, plus a flag which states, those are correct valus, just use them, no recalc required 1156602881 M * daniel_hozac how would we determine if the guest has been tampered with though? 1156602885 M * ntrs can these functions be used from a bash script? 1156602886 M * daniel_hozac i mean, how would the flag get unset? 1156602887 M * Bertl would suffice to give an admin a tool to handle that 1156602889 M * daniel_hozac ntrs: yes. 1156602917 M * Bertl having a second option for a lazy vdu would help to speed up things 1156602917 M * daniel_hozac ntrs: as i said, source /usr/lib*/util-vserver/vserver.functions; saveDiskLimits /etc/vservers/ should save all the disk limits for that guest. 1156602934 M * Bertl I do similar for quotas 1156602948 M * daniel_hozac Bertl: if there's a tool for it, they might as well just rm -f the cache file. 1156602957 M * Bertl i.e. use the recorded values, start a background recalc, then update on the fly 1156602958 M * daniel_hozac as is noted on the disk limits how-to. 1156602978 M * derjohn daniel_hozac, the only guest: load average: 0.34, 0.27, 0.28. the host (with nothing running than the only guest):load average: 0,92, 0,66, 0,65 1156602978 M * ntrs daniel_hozac, it seems that saveDiskLimits does not take any parameters. 1156602983 M * daniel_hozac ntrs: it does. 1156602987 M * daniel_hozac first line... 1156602991 M * daniel_hozac local vdir=$1 1156602999 M * nayco So, Bertl , what is your solution for XFS quota ? 1156603019 M * daniel_hozac nayco: he posted the patch, http://vserver.13thfloor.at/Experimental/delta-vroot-feat01.diff 1156603019 M * Bertl the patch I pasted (the url) above 1156603037 Q * sladen Ping timeout: 480 seconds 1156603038 M * nayco Oh, sorry ! Eeweego ;-) ! 1156603043 M * Bertl nayco: http://vserver.13thfloor.at/Experimental/delta-vroot-feat01.diff 1156603068 M * Bertl it might be an interesting solution to the quota issues in general 1156603074 M * daniel_hozac derjohn: very odd. 1156603079 M * derjohn Bertl, maybe you have an idea about that, too. with -rc29 ? 1156603092 M * daniel_hozac derjohn: any process on the host whose CPU time is increasing? 1156603099 M * daniel_hozac like in ps output. 1156603108 M * Bertl derjohn: well, I assume something is running on the host, no? 1156603142 M * Bertl host shows whole system load, guest only virtualized one 1156603160 J * sladen paul@starsky.19inch.net 1156603164 M * Bertl so for example a raid thread will add to that too 1156603169 M * Bertl welcome sladen! 1156603254 M * nayco *compiling* 1156603260 M * nayco small patch, indeed... 1156603327 M * derjohn daniel_hozac, Bertl: 89.9%idle on host 1156603469 M * nayco *reboot* 1156603471 Q * nayco Quit: Bonne nuit ! 1156603623 M * Bertl derjohn: so? 1156603636 J * nayco ~nayco@lns-bzn-53-82-65-0-115.adsl.proxad.net 1156603746 M * ntrs How do I deteremine if a guest is running from a script? 1156603917 Q * mire_ Quit: Leaving 1156603988 J * kir_home tis-cf7a04@213.152.157.70 1156604105 M * Bertl wb kir_home! 1156604254 Q * kir_home 1156604376 M * nayco Quick as a flash ;-) 1156604579 M * nayco Bertl: Present ! http://paste.linux-vserver.org/304 1156604618 M * nayco I forgot : # uname -a 1156604619 M * nayco Linux samba4.basscorp.prive 2.6.17.8-vs2.1.1-rc29 #2 Sat Aug 26 16:41:16 CEST 2006 i686 AMD Athlon(tm) XP 1700+ unknown GNU/Linux 1156604681 M * daniel_hozac ntrs: vserver guest status &> /dev/null 1156604696 Q * pisco Ping timeout: 480 seconds 1156604849 M * phreak`` hrm, --enable-apis=NOLEGACY is confusing me a bit .. that would enable or disable legacy support for util-vserver ? 1156604867 M * daniel_hozac yes :) 1156604882 M * daniel_hozac it enables all APIs, except for the truly ancient one. 1156604898 M * daniel_hozac (which i can't even get to compile...) 1156604985 M * s0undt3ch Hollow: arround? 1156605043 M * Bertl nayco: so that is sufficient to 'cheat' the tools :) 1156605183 Q * nayco Ping timeout: 480 seconds 1156605200 J * nayco ~nayco@lns-bzn-53-82-65-0-115.adsl.proxad.net 1156605313 M * nayco damn mdv 2006... All my gtk apps (like xchat) freeze when I open a open/browse/save dialog >:-( 1156605331 M * derjohn Bertl, what I wanted to point out is that the machine cant have load of > 0.7. I assume the claculation of the loadavg is wrong for some reason.- 1156605333 M * nayco Bertl: Great ! 1156605363 M * Bertl derjohn: huh? 1156605380 M * derjohn daniel_hozac, the only guest: load average: 0.34, 0.27, 0.28. the host (with nothing running than the only guest):load average: 0,92, 0,66, 0,65 1156605385 M * derjohn daniel_hozac, sry 1156605394 M * derjohn Bertl, isn't that odd ? 1156605399 M * nayco Bertl: So, I'm now going to try samba/quota/vserver. If that succeeds, my students' file servers will run inside vservers this year :D 1156605408 M * derjohn I use virtualized load avg in the guest. 1156605450 M * derjohn how can the host be at nearly 1 while 89% idle and no disk I/O= 1156605479 M * Bertl could be a kernel thread 1156605546 M * derjohn How can I make such a thread 'visible' ? 1156605574 M * derjohn the host has a RAID 10 , maybe that created load, but that much? (the RAID is in sync ...) 1156605633 Q * |gerrit| Remote host closed the connection 1156605734 M * daniel_hozac kernel threads should be visible in the ps output of the host. 1156606131 M * Bertl upload ps auxwww 1156606143 M * waldi Linux zee 2.6.17-2-vserver-sparc64 #1 SMP Sat Aug 26 12:28:58 UTC 2006 sparc64 GNU/Linux 1156606151 M * waldi okay, machine works like a charm :) 1156606295 M * waldi Bertl: oh, oh, testme have a bad race condition in it 1156606320 M * waldi or such 1156606524 Q * id23 Ping timeout: 480 seconds 1156606531 M * derjohn waldi, ah, a success story with sparc64? did you have to patch custom stuff? 1156606537 M * waldi nope 1156606549 M * waldi works fine 1156606593 M * waldi and it eats kernels for breakfast, with ccache I can build debian kernels in something like 2 and half minute per image 1156606603 M * FaUl derjohn: sparc64 works fine beside the fact that it panics unregulary about once a week 1156606605 M * waldi without it needs 8 1156606625 M * derjohn nice to hear! waldi can you tell me whats the status of OpenVZ in Debian? Do you kernel folks already provide binary kernel? As a flavor ? 1156606645 M * FaUl derjohn: i guess this has nothing to do with the vserver-patches, anyway 1156606646 M * derjohn FaUl, sounds promising ;( 1156606659 M * waldi derjohn: the kernel team have no knowledge 1156606664 M * waldi about openvz 1156606676 M * FaUl derjohn: thats why we replaced our dual-ultrasparc with something x86 1156606688 M * derjohn waldi, there was an annocement on prolinux that ovz goes into debian now ?!?! 1156606695 M * waldi derjohn: only the patch 1156606709 M * waldi nothing installable 1156606765 M * derjohn waldi, ah. ok ;) Suse made similar annoucements for SLES 10 Update 2 which a tech of them not having heard about ... 1156606807 Q * Viper0482 Ping timeout: 480 seconds 1156606810 M * waldi derjohn: the press announcement was from swsoft AFAIK, not from debian 1156606890 M * derjohn waldi, ahhhh ;) /me must use openeyes ;) 1156606899 M * derjohn Bertl, daniel_hozac ps auxwww |awk '{print $3}' | sort |uniq -> 0.0 1156606922 M * derjohn w -> load average: 0,40, 0,61, 0,64 1156606936 M * Bertl did I say funny one liners, or upload? :) 1156606960 M * derjohn ok, hold a sec ! 1156607210 M * waldi derjohn: do you know someone called bill baker? we is the contact listed in that announcement 1156607215 M * derjohn Bertl, http://paste.linux-vserver.org/305 1156607246 M * derjohn waldi, bill baker sounds like an artist name of a pornstar. never heard .... 1156607277 M * derjohn but we could ask kir on the channel here, he is from ovz 1156607356 M * derjohn google tells: "Bill Baker Baker Communications ". maybe a PR pro ? 1156607393 M * Bertl I assume they use a network of PR agents 1156607481 J * Viper0482 ~Viper0482@p54976EF1.dip.t-dialin.net 1156607539 M * Bertl \ 1156607674 M * daniel_hozac waldi: vserver enter doesn't work though, right? i thought dietlibc was broken on sparc. 1156607674 M * derjohn Bertl, I might add the I set all kings of limits on the guest, i.e. AS, RSS, CPU sched ... 1156607798 M * phreak`` daniel_hozac: hrm, I don't see a difference between ./configure w/ --enable-apis=NOLEGACY and a ./configure with --enable-apis omitted 1156607814 M * daniel_hozac phreak``: hmm? in the output or what? 1156607858 M * Bertl daniel_hozac: you are using irssi too? 1156607878 M * phreak`` daniel_hozac: after the script is finishing the configure run, you'll get a list of all available API's (which shows this on both: "available APIs: v13,net") 1156607882 M * daniel_hozac Bertl: yep. 1156607929 M * Bertl I'd like to highlight urls somehow (i.e. in a different color) do you know/have/want something like that? 1156607984 M * daniel_hozac that would be useful, but i don't have anything like that. my irssi setup is pretty much vanilla. 1156608016 M * waldi daniel_hozac: have to check that 1156608086 M * Bertl derjohn: well, I see enough threads xfs* pdf* raid*, to explain a load of 0.4-0.9 1156608216 M * derjohn Bertl, hm, i just checked with iostat that there are really between 100-250 tps constanty. what much for sata.... 1156608472 A * derjohn off now folks! thanks for the help ! 1156608517 M * Bertl np, have fun! 1156608549 M * derjohn Bertl: if I use virtualized load avg on the guest and the XFS stuff runs on the host, we guest 'see' when there is much load-trouble, even if the guest itself is responsible for the load? 1156608579 M * daniel_hozac no. the process is on the host. 1156608618 M * daniel_hozac phreak``: can't reproduce... i get Available APIs: compat,v11,fscompat,v13,net,oldproc,olduts when configuring with --enable-apis=NOLEGACY 1156608636 M * derjohn hm, I think i'll go back to non-virtualized loadavg. it 'says' more about the sytems state then the virtualized one. 1156608666 M * Bertl well, it is an option :) 1156608680 M * derjohn IMVHO the load the hosts itself has, should be added to each guests virtualized load 1156608699 M * daniel_hozac why? 1156608703 M * derjohn of, at least a fraction of it 1156608742 M * daniel_hozac the point of the virtualized load averages is that they are only for that guest... 1156608753 M * derjohn well, because the guest has load of 0.1, while it created massive i/o and gets slow. Then host might be loaded > 1 , but the guest doesnt see it from inside. 1156608800 M * derjohn daniel_hozac, well, if the guest copies 1 GB file in a loop he has a (virt) load of 0.1, but the hist might have > 1.. 1156608826 M * derjohn even is the guest itself is the cause if the load (kernelthreads, xfs blabla ...) 1156608855 M * derjohn but i really have to leave!!! sorry !!! byte ~ 1156608872 M * daniel_hozac cya! 1156608952 M * phreak`` daniel_hozac: ah, sorry :) (sudo resetted it - don't try a USE=legacy sudo emerge util-vserver and expect it to work) 1156608962 M * daniel_hozac hehe. 1156609992 M * Bertl daniel_hozac: /hilight -word -color %B -regexp http://[^ ]* 1156610447 M * daniel_hozac Bertl: thanks! 1156610671 M * Bertl surprisingly simple with irssi, as usual :) 1156610865 M * matti You can also make url logging script - or download one from scripts repository ;p 1156610907 Q * shedi Ping timeout: 480 seconds 1156610975 J * shedi ~siggi@inferno.lhi.is 1156611063 Q * shedi 1156611334 M * daniel_hozac ctrl+a n grep -Hr http:// ~/irclogs works for that too ;) 1156611359 M * matti daniel_hozac: This is the easy way ;p 1156611360 M * matti ;pp 1156611368 A * FaUl has no irclogs 1156611420 M * daniel_hozac so... how do you remember things? :) 1156611442 M * Hollow irc.13thfloor.at 1156611443 M * Hollow ;) 1156611473 A * FaUl just does /last -file if he needs a particular log - anything else is just waste of space 1156611473 M * matti daniel_hozac: He probably tap some written thing on his fridge ;p 1156611525 M * FaUl and - if gouvernment pwnes my harddisk they don't have logs of me - i don't want to make things to easy to them;-) 1156611544 M * daniel_hozac lol 1156611547 M * mnemoc luks is your friend :) 1156611550 M * daniel_hozac indeed. 1156611551 M * matti LOL 1156611558 M * Bertl FaUl: that's what encryption is for 1156611568 M * matti Bertl: Paranoia also ;p 1156611572 M * matti :> 1156611578 M * FaUl Bertl: ;-) 1156611583 A * daniel_hozac can't even imagine what a life without paranoia would be like :) 1156611593 M * matti daniel_hozac: Boring ;p 1156611797 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156612044 J * coocoon ~coocoon@p54A0797A.dip.t-dialin.net 1156612331 Q * matti Quit: 8-X 1156613906 J * |gerrit| ~kvirc@dslb-084-060-199-237.pools.arcor-ip.net 1156616007 J * mire ~mire@212-167-222-85.COOL.ADSL.VLine.Verat.NET 1156616474 J * shedi ~siggi@inferno.lhi.is 1156616647 M * nayco Bertl: quota work too with samba 3.0.20 :) ! 1156617186 J * id23 ~id@p50812021.dip0.t-ipconnect.de 1156617218 M * nayco Bertl: When do you thinks this will make it to the devel release ? 1156617223 M * nayco -s 1156618376 M * daniel_hozac nayco: probably the next one ;) 1156619044 M * Bertl daniel_hozac: did you look at the patch? 1156619053 M * daniel_hozac yeah. 1156619077 M * Bertl kind of hacky but an interesting idea, what do you think? 1156619115 M * daniel_hozac a bit, but i guess it sort of makes sense. 1156619431 M * nayco daniel_hozac: you mean 2.1.1-rc32 ;-) ? 1156620714 M * s0undt3ch hello ppl 1156620730 M * daniel_hozac hi 1156620744 M * s0undt3ch how does one secure it's services inside a guest? since there no access to iptables and the guest only has a public ip 1156621274 M * s0undt3ch anyone arround, or you haven't got yourself in the same sictuation? 1156621309 M * daniel_hozac heh, sorry, busy upgrading my kernels... 1156621316 M * daniel_hozac you can use iptables on the host. 1156621363 M * daniel_hozac if you have customers inside of guests, you could setup an iptables daemon that your guest's iptables connects to for setting up the rules. 1156621367 M * daniel_hozac or some sort of web interface. 1156621432 M * s0undt3ch we'll I'm actually on the guest, so this is a sictuation where everything I can do is on the guest 1156621460 M * daniel_hozac ask your provider to it, then ;) 1156621478 M * s0undt3ch yeah, I couldm and I'll have to :) 1156621500 M * daniel_hozac Bertl: http://daniel.hozac.com/vserver/delta-headers-fix83.diff for devel. 1156621524 M * s0undt3ch so, the way to do it is, have an iptables daemon running on host, and have iptables inside guest connect to that host iptables? 1156621535 M * s0undt3ch never though that was even possible 1156621539 M * daniel_hozac basically. 1156621574 M * daniel_hozac i don't know if anyone has released their solution yet. 1156621759 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156621759 Q * ntrs Read error: Connection reset by peer 1156622093 M * waldi daniel_hozac: vserver enter have some problems with the pty, it looses input 1156622112 M * daniel_hozac waldi: right. 1156622120 M * daniel_hozac http://daniel.hozac.com/vserver/dietlibc-0.29-sparc-termios.patch 1156622148 M * waldi hmm 1156622150 M * waldi okay 1156622167 M * daniel_hozac the easy fix is to disable vlogin. 1156622176 Q * |gerrit| Read error: Connection reset by peer 1156622199 J * |gerrit| ~kvirc@dslb-084-060-211-186.pools.arcor-ip.net 1156622224 Q * ntrs_ Read error: Connection reset by peer 1156622229 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156622293 M * waldi daniel_hozac: hmm, can you report the bug to the debian bts? thats a rather working way to get such patches upstream 1156622425 M * daniel_hozac i guess, or we could just bug Hollow or phreak`` (IIRC they're maintaining diet in Gentoo). 1156622439 M * daniel_hozac FTR, i did report it upstream when we found it. 1156622496 M * waldi okay 1156622524 M * daniel_hozac i guess it lost somewhere, as it's not in 0.30. 1156622533 M * daniel_hozac +got 1156622622 M * waldi hmm, it is in the debian version 1156622629 M * daniel_hozac really? 1156622672 M * waldi yes 1156622730 M * daniel_hozac what version? 1156622733 M * waldi 0.30-1 1156622750 Q * ntrs_ Read error: Connection reset by peer 1156622752 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156622854 M * waldi so i need to rebuild util-vserver 1156623160 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156623160 Q * ntrs Read error: Connection reset by peer 1156623271 M * waldi daniel_hozac: much better 1156623395 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156623396 Q * ntrs_ Read error: Connection reset by peer 1156623462 M * daniel_hozac waldi: great! 1156623519 M * waldi and I asked for a rebuild of the package 1156623553 M * waldi now I want to test the t2000 a little bit ... 1156623612 Q * ntrs__ Read error: Connection reset by peer 1156623622 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156623727 M * waldi and I think I need to use ipvs 1156623754 M * Bertl okay, off for now .. probably back later 1156623756 Q * Viper0482 Remote host closed the connection 1156623762 N * Bertl Bertl_oO 1156623868 M * daniel_hozac hmm, you're right. it seems to be fixed in 0.30. 1156623879 M * daniel_hozac i could've sworn i checked 0.30 and it wasn't there. 1156623940 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156623940 Q * ntrs__ Read error: Connection reset by peer 1156625134 J * cehteh ~ct@cehteh.homeunix.org 1156625873 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156625873 Q * ntrs_ Read error: Connection reset by peer 1156626281 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156626282 Q * ntrs Read error: Connection reset by peer 1156626939 Q * nayco Quit: Bonne nuit ! 1156627144 Q * mire Quit: Leaving 1156627628 J * GNUcifer ~ct@cehteh.homeunix.org 1156627723 Q * cehteh Ping timeout: 480 seconds 1156627821 Q * serving 1156628007 J * cehteh ~ct@cehteh.homeunix.org 1156628128 Q * GNUcifer Ping timeout: 480 seconds 1156629151 J * DanishKingPin ~coolman_p@0x50c46cb0.arcnxx10.adsl-dhcp.tele.dk 1156629164 M * DanishKingPin Would someone please help me set up a proxy for use with mirc ? 1156629341 M * daniel_hozac how does that relate to linux-vserver? are you going to set it up in a guest? 1156629343 Q * dna Quit: Verlassend 1156629372 M * weasel daniel_hozac: not at all. the kingping just spammed its question all over the net. 1156629398 M * daniel_hozac ah. 1156629418 M * daniel_hozac as i expected ;) 1156629442 M * lylix anyone using vcd? 1156629459 M * mnemoc Hollow :) 1156629477 M * lylix k, thought id prolly have to root him up... 1156629499 M * lylix jsut wondering if there is a way/command to import existing vservers into the vxdb 1156629518 M * lylix everything is setup/working execpt for an empty db :( 1156629975 Q * balbir Ping timeout: 480 seconds 1156630458 P * DanishKingPin 1156630499 M * daniel_hozac lylix: can't you create an empty one and move the existing one to its place? 1156630925 M * lylix i supposed that would work 1156631007 M * lylix be ok for a couple, but to implement on a system w/ many, an importing command would be more feasible 1156631015 M * daniel_hozac indeed. 1156631043 M * daniel_hozac i guess patches are appreciated ;) 1156631070 M * lylix i figured perhaps when vcd is started it would do some system checks to grab existing setups, but this may be somewhat obtrusive 1156631621 J * wiesel_ weasel@rev86-59-21-36-reserved.sil.noreply.org 1156631751 Q * mnemoc Quit: leaving 1156631766 M * daniel_hozac Bertl_oO: hey, won't the delta-vroot-feat01 oops if vroot isn't loaded when you stat a vroot device? 1156631840 Q * |gerrit| Quit: KVIrc 3.2.0 'Realia' 1156631999 Q * wiesel_ Quit: wiesel_ 1156633499 J * mnemoc ~amery@kilo105.server4you.de 1156633854 M * daniel_hozac damn, still connection refused on list.linux-vserver.org. Martin's email is still the tuxbox one, right? (or has someone else already emailed him?) 1156634053 J * GNUcifer ~ct@cehteh.homeunix.org 1156634058 Q * cehteh Ping timeout: 480 seconds 1156635620 M * ntrs_ derjohn2, could you upgrade the devel changelog with the changes in rc31? 1156635627 M * ntrs_ upgrade=update 1156635928 M * GNUcifer ct 5791 0.0 0.0 0 0 ? Zs 23:12 0:00 [keepalive] 1156635948 M * GNUcifer .. is there still a known bug which keeps zombies around in vservers? 1156636053 M * GNUcifer mhm there was a cron, not collecting it 1156636492 Q * mkhl