1141603260 M * marl ok thanks, one other thing, is there any docs on how to assign an ip to a vserver via dhcp? i can see stuff on running a dhcp server in the faq (i take its meening running the actual server to assign out addys to other machines) 1141603292 M * marl but i cant see anything on just telling a vserver to get its ip via dhcp from another machine 1141603306 M * daniel_hozac you'd need to run the client on the host. 1141603307 J * Aiken ~james@tooax6-135.dialup.optusnet.com.au 1141603312 M * marl preferable setup via the inital verser build command :) 1141603324 M * daniel_hozac and the guest wouldn't be bound to only that IP address. 1141603335 M * marl ? 1141603342 M * daniel_hozac so really, using DHCP for guests is all cost and no benefits. 1141603365 M * daniel_hozac better to use private IP addresses and use NAT. 1141603429 M * marl yup the vservers are all run internally at the moment, but its starting to get a real headache tring to work out what ips have been assigned and what havnt :( as not all the vservers run all the time :( 1141603485 M * daniel_hozac grep -H . /etc/vservers/*/interfaces/*/ip 1141603515 M * marl lol, thanks, ive been looking for something like that for ages :) thanks again 1141603563 M * FaUl marl: what about documentation? I personally documentate our ip-adress-usage in a wiki :-) 1141603581 J * samuel_ ~samuel@levinux.UQAR.UQUEBEC.CA 1141603600 A * samuel_ is away: (Auto-Away after 10 mins) [BX-MsgLog On] 1141603630 M * marl this is a test box, and as most of my work is done in the early hours of the morning (1am - 5am) writing docs is the last thing on my mind LOL!!!! 1141603641 M * marl you are write thow its the best way :) 1141603745 M * FaUl marl: ok, seems to make sense 1141603769 M * marl im also dyslecsic aswell :( 1141604207 M * doener hm, how does one manage to join a channel and have an "auto-away after 10 minutes" only a minute later? 1141604240 M * FaUl doener: join/part etc. does not affect idle-counter 1141605232 J * Smutje_ ~Smutje@xdsl-87-78-1-250.netcologne.de 1141605349 Q * Smutje Ping timeout: 480 seconds 1141605349 N * Smutje_ Smutje 1141608002 Q * Dr4g Read error: Connection reset by peer 1141613103 Q * doener Quit: leaving 1141613392 Q * michal` Ping timeout: 480 seconds 1141613865 J * michal` ~michal@www.rsbac.org 1141618108 M * marl any one have apt-proxy running inside a vserver? im getting 'couldnt listen on ipaddy' errors when trying to start it, i cant see anything that would indicate it needs extra caps, so wantedt o ask if anyone had it running at the mo 1141618265 M * Bertl maybe the host is already using that port? 1141618361 M * Bertl anyway, off to bed now .. back tomorrow .. have a nice one everyone! 1141618369 N * Bertl Bertl_zZ 1141618423 M * marl nite Bertl_zZ 1141618438 M * marl ok host isnt using that port, anyone any other sujestions? 1141618971 M * marl tis ok, got it going (i think!!!!!!) 1141624919 Q * lonewolff Read error: Operation timed out 1141625000 J * lonewolff lonewolff@adleman.lonewolff.info 1141626812 J * coocoon ~coocoon@p54A06170.dip.t-dialin.net 1141626819 M * coocoon morning 1141628570 Q * FireEgl unununium.oftc.net plasma.oftc.net 1141630379 Q * Aiken Ping timeout: 480 seconds 1141630882 J * teukka ~tmatilai@backport.ri.fi 1141631242 J * ||Cobra|| ~cob@146.50.22.204 1141631406 A * ||Cobra|| yop 1141632986 J * Viper0482 ~Viper0482@p549763C5.dip.t-dialin.net 1141633817 Q * shedi Quit: Leaving 1141636148 J * wally ~homebase@62.116.83.242 1141636162 J * pagano ~pagano@lappagano.cnaf.infn.it 1141636168 M * pagano hi people :) 1141637680 Q * coocoon Ping timeout: 480 seconds 1141637903 J * shedi ~siggi@tolvudeild-204.lhi.is 1141638260 J * coocoon ~coocoon@p54A053FE.dip.t-dialin.net 1141639376 Q * rs Quit: rs 1141639922 J * dothebart ~willi@xdsl-213-196-255-169.netcologne.de 1141640379 Q * tudenbart Ping timeout: 480 seconds 1141640868 J * FireEgl Atlantica@Atlantica.Tcldrop.Org 1141641153 Q * wally Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1141641594 Q * pagano Read error: Connection reset by peer 1141641728 J * rs ~rs@office.dailymotion.com 1141641751 Q * rs Quit: 1141641752 J * pagano ~pagano@lappagano.cnaf.infn.it 1141641766 M * pagano re-hi 1141641830 J * rs ~rs@office.dailymotion.com 1141641857 Q * rs Quit: 1141642645 J * rs ~rs@office.dailymotion.com 1141642820 Q * FireEgl unununium.oftc.net plasma.oftc.net 1141643139 Q * teukka Read error: Connection reset by peer 1141643204 J * teukka ~tmatilai@backport.ri.fi 1141643513 J * lilalinux ~plasma@80.69.35.186 1141643578 M * phreak`` Bertl_zZ: might you take a look at http://dev.gentoo.org/~phreak/delta-vs2.0.2-rc10-ldebug-feat01.patch after you finished your nap ? :) 1141643832 M * pagano do u know 1141643848 M * pagano if can i connect 2 console port? 1141643895 M * pagano because i don't have serial port 1141644009 M * pagano for a control channel with heartbeat 1141644677 T * * 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}-rc10 | 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 ;) 1141644677 T * Bertl - 1141644958 J * fwl ~fwl@83-215-237-2.seek.stat.salzburg-online.at 1141649386 Q * fwl Quit: Leaving 1141649726 J * Smutje_ ~Smutje@xdsl-87-78-85-80.netcologne.de 1141649834 Q * Smutje Ping timeout: 480 seconds 1141649834 N * Smutje_ Smutje 1141649846 J * brc bruce@20151181056.user.veloxzone.com.br 1141650515 J * Hollow_ ~hollow@home.xnull.de 1141650516 Q * Hollow Read error: Connection reset by peer 1141650721 J * restill ~restill@c-24-11-171-10.hsd1.mi.comcast.net 1141651414 Q * dothebart Quit: brb. 1141652236 J * jrc ~jrc@81.7.189.11 1141652241 N * jrc Wenix 1141652358 Q * Wenix Quit: 1141652743 J * jrc ~jrc@81.7.189.11 1141652745 N * jrc Wenix 1141653904 Q * matta Ping timeout: 480 seconds 1141654357 J * GeDaST ~mircc4478@85.104.120.196 1141654871 Q * GeDaST Quit: mIRCTurk 1141654875 J * h01ger ~holger@bone.digitalis.org 1141654935 M * h01ger hi... i get those "setrlimit limit #11 to soft=-1, hard=-1 failed:"-errors, but neither "man limits.conf" nor "man setrlimit" tell me which nr means which limit. (i also get #12 and #6..) 1141655015 M * blizz can i have 32bit guest systems if i have a 64bit host? 1141655157 J * liquid3649_ ~Viper0482@p54975A84.dip.t-dialin.net 1141655201 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net 1141655580 Q * Viper0482 Ping timeout: 480 seconds 1141656874 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1141657558 N * Bertl_zZ Bertl 1141657562 M * Bertl morning folks! 1141657584 M * h01ger morning Bertl :) 1141657613 M * h01ger i've read http://linux-vserver.org/Resource+Limits but still dont get the answer for my question three lines up in scrollback... 1141657641 A * Bertl reading up ... 1141657675 M * Bertl ah, those are issues with the guest trying to raise 'ulimits' over the given host ulimits 1141657676 M * matta morning bertl! 1141657682 M * Bertl hey matta! 1141657683 M * matta but how is my morning like your morning? :) 1141657704 M * Bertl h01ger: so, check your 'ulimit -a' on the host 1141657713 M * h01ger Bertl, yes. but which limit ? limits.conf lists quite a few as well as http://linux-vserver.org/Resource+Limits :) but the error say #12 and not $LIMITNAME 1141657756 M * h01ger and then the item from the "ulimit -a" output ? lol :) 1141657763 M * Bertl RLIMIT_SIGPENDING11 1141657783 M * matta what is the pending signals limit actually limiting? 1141657784 M * Bertl if it is really limit #11 1141657791 M * matta and don't say pending signals! 1141657798 M * Bertl hum, yeah :) 1141657810 M * matta like, are we talking signals sent to processes via kill? 1141657814 M * matta IPC? 1141657821 M * Bertl yup, kill 1141657827 M * matta oh 1141657829 M * matta I have 2112 1141657836 M * matta that should be way more than sufficient 1141657842 M * matta don't even have 2000 procs :) 1141657843 M * Bertl usually 1141657873 M * matta right, I assume you want it to equal maxprocs 1141657883 M * h01ger so i count in the list at http://linux-vserver.org/Resource+Limits and start with 0 ?! thanks! :-) 1141657894 M * Bertl matta: not necessary but a limit is better than none 1141657955 M * Bertl h01ger: actually it's better to consult your resource.h file 1141657964 M * Bertl as the numbers vary from arch to arch 1141658015 M * h01ger aaah. thanx again :) 1141658081 M * Bertl h01ger: you're welcome! 1141658171 J * Val ~val@194.214.126.5 1141658175 M * Val Hi :) 1141658183 M * Bertl hey Val! LTNS! 1141658195 M * Val LTNS ? :) 1141658205 M * Bertl Long Time No See 1141658211 M * Val yes :~( 1141658246 M * Val i did a test of util-vserver 30.210 /etc/vservers//dlimits support : it works well ! 1141658253 M * Val great work 1141658269 M * Bertl good to hear! 1141658286 M * Val but i now got a problem : 'Could not find a place for the hashified files at '/etc/vservers/vskel/apps/vunify/hash'.' 1141658288 M * Val ... 1141658303 M * Val when hashifying a vserver 1141658322 M * Val i don't remember if i had to do something before hashifying 1141658337 M * Val (make a directory or somewhat else) 1141658369 M * Val note i use the "exclude" vunify file 1141658433 M * Bertl sec 1141658445 M * Val all the time you need ;-) 1141658502 M * Bertl http://linux-vserver.org/alpha+util-vserver 1141658519 M * Bertl check out the part 'Directory/vserver unification' 1141658562 M * Val (for those who are interested to test util-vserver 30.210 dlimit support under debian sarge take a look at http://zbla.net/debian/util-vserver-0.30.210-0sarge0/) 1141658570 M * Val Bertl : o, thanks 1141658571 M * Val ok 1141658855 M * Val hum... well, for hashify problem it seems that a directory /etc/vservers//apps/vunify/hash i created fail hash process... 1141658864 M * Val that's my fault :-/ 1141659598 M * daniel_hozac Val: i think i tested most configurations before sending it to Enrico. 1141659617 M * daniel_hozac (all i could think of at the time, pretty much) 1141661062 M * coocoon hello bertl 1141661132 M * Bertl hey coocoon! 1141661144 M * coocoon bertl: question to disk limiting via tagxid the third, I have seen that all my tarballed vservers with different images sizes, have all the same space 232MB 1141661158 M * coocoon Bertl: could this be, I don't think so 1141661200 M * coocoon Bertl: her is my DiskLimit.sh 1141661265 M * Bertl it's at least unusual :) 1141661286 M * coocoon Bertl: /usr/sbin/vdlimit --xid 100 --set space_total=1000000 --set space_used=236984 --set inodes_total=1000000 --set inodes_used=8950 --set reserved=5 1141661311 M * coocoon Bertl: /vservers/sin00/ 1141661319 M * coocoon Bertl: the last line 1141661332 M * Bertl well, after that all will have 232MB, that is right :) 1141661346 M * Bertl (as you set it to 232MB :) 1141661371 M * coocoon Bertl: space_used=236984 this is the entry right 1141661399 M * Bertl yup 1141661431 M * coocoon Bertl: how can I set it that it doesn't affect the vservers space 1141661456 M * coocoon Bertl: or is it not so important 1141661459 M * Bertl that's a strange question, as you set it to actually used space 1141661481 M * Bertl i.e., you figure the space a guest uses on disk, then set this value accordingly 1141661493 M * Bertl the kernel will then continue to track the space for you 1141661494 M * coocoon Bertl: must this setting be 1141661504 M * coocoon Bertl: oh aha 1141661511 M * coocoon I have no probs with that 1141661562 P * djudko Leaving 1141661733 M * coocoon Bertl: another question is using plain or gentoo for style settings for a gentoo vserver-guest which one is better 1141661807 M * Bertl according to Hollow, the gentoo style is not maintained 1141661818 M * Bertl and the plain init style is sugegsted 1141661836 M * Bertl but, IIRC, there was some efford to 'revive' the gentoo style 1141661859 M * Bertl as it would be a comparable solution to the sysv one 1141662038 M * coocoon Bertl: but plain init-style makes error when I am starting vserver with plain init-style maintenance tool starts, if I'm startng with gentoo init-style, everything starts fine without error messages, but after stopping getting /dev/initctl fehler 1141662060 M * coocoon error 1141662073 M * Bertl hmm, plain gives you an error? 1141662086 M * SiD3WiNDR is sam vilain on IRC somewhere? 1141662119 M * coocoon yes und util-vserver they doesn't start anymore, so I am using 209 1141662126 M * Bertl SiD3WiNDR: yep, here, goes by the nick mugwump 1141662127 M * coocoon util-vserver210 1141662133 M * SiD3WiNDR ah, cool. 1141662136 M * SiD3WiNDR thanks Bertl 1141662151 M * SiD3WiNDR a /who #vserver could have told me that ;/ 1141662189 M * Bertl coocoon: could you upload the output of: "vserver-info - SYSINFO" and "testme.sh" to some pastebin service (and your error with 0.30.210) 1141662218 M * coocoon Bertl: yes I installed it again 1141662226 M * coocoon Bertl: takes a little bit 1141662228 M * coocoon ;-) 1141662236 M * Bertl coocoon: take your time 1141662318 Q * shedi Quit: Leaving 1141662323 M * Val is vunification (hashify) compatible with dlimits xid tag ? 1141662388 M * Val got lots of lookup hidden ressources when trying to restart hashified vservers with xid tags 1141662589 Q * rs Ping timeout: 480 seconds 1141662608 M * Bertl Val: then something is misconfigured 1141662619 M * Val ok 1141662629 M * Bertl unified guests consist of two file taggigns 1141662633 M * Bertl *tagging 1141662635 M * Bertl +s 1141662647 M * Bertl tags with xid=0 -> shared 1141662657 M * Val ok, that's it 1141662660 M * Bertl and tags with xid=N (context id) not shared 1141662671 M * Val ok ok 1141662684 M * Bertl if your guest has files tagged with a different xid than it has 1141662696 M * Val ...well, the first hashified vserver must have xid=0 ? 1141662697 M * Bertl then something is wrong :) 1141662704 Q * pagano Ping timeout: 480 seconds 1141662712 M * Bertl the vhashify process should take care of that 1141662755 M * Bertl it moves the hashified files into the hash store 1141662768 M * coocoon Bertl: http://pastebin.com/587138 1141662771 M * Val well 1141662784 M * Val ...i need to change xid tag for hash directory 1141662811 M * Bertl hmm, no, that should be all xid=0 tagged 1141662826 M * Val yes it is 1141662948 M * Val when i do "vserver start", i got now : "chroot-shopen("/etc/mtab"): Permission denied" 1141662975 M * Bertl and how is your mtab tagged? 1141663001 M * coocoon Bertl: starting gentoo vserver -> error with util-vserver-0.30.209 init-style plain, under serial console: http://pastebin.com/587147 1141663015 J * stefani ~stefani@superquan.apl.washington.edu 1141663017 M * Val Bertl: with the good tag 1141663033 M * Bertl Val: try to configure the tools with --apis=NOLEGACY 1141663040 M * Val hum 1141663042 M * Val ok 1141663046 M * Bertl Val: (or update to a newer kernel :) 1141663109 M * phreak`` Bertl: had a chance to take a peak into the patch ? 1141663114 M * Bertl coocoon: looks like you are missing the proper gentoo layout for vserver guests 1141663123 M * Bertl phreak``: which one? 1141663145 M * Bertl ah, jsut found it 1141663165 M * coocoon coocoon: hm ok u mean I use x86 and should use it for PIV 1141663178 M * Bertl phreak``: looks good, let's compare it to mine, shall we :) 1141663185 M * phreak`` Bertl: it should be nearlly the same as http://vserver.13thfloor.at/Experimental/delta-ldebug-feat01.diff but for 2.6.15 :) 1141663198 M * phreak`` Bertl: sure :) thats what I've hoped ;) 1141663231 M * coocoon Bertl: stages u mean with the base layout 1141663289 M * Bertl http://vserver.13thfloor.at/Experimental/delta-ldebug-feat01b.diff 1141663304 M * phreak`` coocoon: thats probably happening becaue you're using baselayout instead baselayout-vserver :) 1141663305 M * Val unrecognized option: --apis=NOLEGACY 1141663322 M * Bertl coocoon: I'm no gentoo guy, so I just heard about the layout and stuff :) 1141663329 M * Val it is --enable-apis 1141663338 M * Bertl Val: congrats! :) 1141663380 M * Val grep power :p 1141663399 M * coocoon phreak: oh yes right, phreak that was another question which I wanted to aks u, so I should use vserver layout, but with init-style gentoo, it looks better 1141663408 M * phreak`` Bertl: thanks a lot :) (yay! my first own, backported patch :) 1141663466 M * coocoon phreak: and it works, but maybe there could happen some errors or appear some problems, right?! 1141663612 M * phreak`` coocoon: well but there is a problem with the gentoo init-style ... (as its used by the utils) .. the call /sbin/rc directly vs. plain starting init 1141663678 M * Bertl coocoon: if it works for you, it's fine 1141663679 M * phreak`` Bertl: got another question :) where would the HANDLE_CASE(QUOTA, quota, value); (quota-feat01.diff) belong to in 2.6.15 ? I don't seem to find anything similar 1141663725 M * phreak`` *fix01.diff even :) 1141663759 M * Val Bertl : It doesn't work :~( 1141663769 M * coocoon Bertl: ok but which errors could appear in the future 1141663798 M * Val Bertl : got the same error, chroot-shopen("/etc/mtab"): Permission denied 1141663929 M * coocoon phreak: which errors could appear if I am using baselayout instead of vserver-layout 1141663945 M * Val Mar 6 17:51:48 localhost kernel: vxW: xid=3 did lookup hidden de1cf8e0[#42,539628] »/etc/init.d/rc«. 1141663978 M * Val 42 is the context number of the first hashified vserver 1141664027 M * phreak`` coocoon: it could happen that your vserver gets unuseable :) despite the ugly error messages .. 1141664033 M * Val founded 1141664056 M * coocoon phreak: ok thanx so I will install vserver-layout 1141664057 M * Val in hashified directory there is xid tags that are not set to 0 1141664075 M * coocoon Bertl && phreak: thanx a lot 1141664080 M * coocoon ;-) 1141664087 M * Val should i change all of them to xid tag 0 ? 1141664220 M * coocoon phreak: one more question how to change the syslog-ng referring to /proc/kmsg 1141664236 M * Bertl Val: yep 1141664494 M * restill Device lo does not seem to be present? 1141664571 M * Bertl restill: where, when, how please? 1141664677 M * restill oh, I thought that was a common one. Ok............. 1141664761 M * restill I get it when I start a VS. "Bringing up loopback interface: Device lo does not seem to be present, delaying initialization. [FAILED}" 1141664785 M * daniel_hozac your guest hasn't been properly cleaned. 1141664788 M * restill it seems to run fine. 1141664798 M * restill define cleaned. 1141664804 M * daniel_hozac for some reason, the network initscript is set to run. 1141664823 M * daniel_hozac it serves no purpose other than generating bogus error messages within a guest. 1141664892 M * restill that is in rc3.d ? 1141664906 M * daniel_hozac if 3 is the default runlevel, yes. 1141664913 M * daniel_hozac i'd suggest you use chkconfig though. 1141665015 M * daniel_hozac if you build your guests with the utils instead, they will take care of disabling all services, etc. 1141665033 M * restill very good. It is set "on" for rc3.d 1141665036 J * bonbons ~bonbons@83.222.39.180 1141665073 M * Bertl welcome bonbons! 1141665093 M * bonbons hallo Bertl 1141665118 M * restill ok set it to off, and restarting it for fun 1141665146 M * restill beautimus 1141665266 M * Val Bertl : hum... hash dir is all now under xid 0, but hashified vserver directories too :( 1141665319 M * daniel_hozac Val: directories? they shouldn't be unified. 1141665348 M * daniel_hozac Val: the hashified files in the guests will be xid 0 though. 1141665396 M * Val well... 1141665408 M * Val how to separe xid tags ? 1141665414 M * daniel_hozac what do you mean? 1141665432 M * Val because now i have : hashified directory files all in xid 0 1141665447 M * Bertl Val: that is fine, because they are _shared_ 1141665455 M * daniel_hozac they must be xid 0. 1141665463 M * Bertl Val: if you really mean the _files_ not the _dirs_ 1141665474 M * Val hum 1141665483 M * Val i did a trick : 1141665498 M * Val 1) chxid -c 0 -R -x 1141665520 M * Val 2) chxid -c -R -x (for all vserver) 1141665525 M * Val then it works... :) 1141665536 M * daniel_hozac really? 1141665544 M * Val realy 1141665550 M * daniel_hozac it shouldn't. 1141665558 M * Val i just have to verify if dlimit works 1141665567 M * daniel_hozac only the last vserver would get access to the shared binaries. 1141665570 M * Val humm 1141665575 M * Val i doesn't work 1141665616 M * Val shit 1141665623 M * Val all is broken now :( 1141665630 M * daniel_hozac you need to do 1 last. 1141665640 M * daniel_hozac to give all vservers access to the shared files. 1141665645 J * shedi ~siggi@inferno.lhi.is 1141665651 M * Val daniel_hozac : ok 1141665662 M * Val daniel_hozac : i retry 2) then 1) 1141665846 M * Val Starting vservers of type 'default'...chroot-shopen("/etc/mtab"): Permission denied 1141665869 M * Val well... 1141665869 M * daniel_hozac are you unifying /etc? 1141665876 M * Val daniel_hozac : yes 1141665888 M * daniel_hozac so i assume you're running devel with CoW? 1141665900 M * Val CoW ? 1141665906 M * daniel_hozac Copy on Write. 1141665926 M * daniel_hozac to break links when you try to write to the files. 1141665955 M * Val nope, i don't do anything 1141665958 M * Val ... 1141666004 M * daniel_hozac you realize that unifying /etc without CoW will lead to that sort of error, right? 1141666012 M * Val right 1141666016 M * Val but just for now :) 1141666018 M * Val ... 1141666019 M * Val so 1141666030 M * Val i'll copy (cp -a) all my vserver dirs 1141666040 M * Val an then break unification 1141666046 M * Val then restore all of them 1141666053 M * daniel_hozac that will automatically break the links. 1141666060 M * Val yup 1141666062 M * Val then 1141666074 M * Val for the next step, while doing unification 1141666102 M * Val ...i have to exclude all files that are not the same ? 1141666119 Q * gerrit Ping timeout: 480 seconds 1141666129 Q * ||Cobra|| Remote host closed the connection 1141666137 M * daniel_hozac files that are not the same will not be unified. 1141666165 M * Val but that was not, why ? 1141666170 M * daniel_hozac without CoW (i.e. vserver 2.1.*), you won't want to unify /etc, /var, or any other place that has writable files. 1141666187 M * Val huh 1141666188 M * Val ok ok 1141666202 M * Val so, how to configure it ? 1141666212 M * daniel_hozac configure what? the exclude file? 1141666214 M * Val add "/var" and "/etc" in exclude file ? 1141666224 M * daniel_hozac yes. 1141666231 M * Val will this be recursive ? 1141666239 M * daniel_hozac yes. 1141666242 M * Val ok fine 1141666257 M * Val many thanks for all 1141666312 M * Val but... the last question :) 1141666326 M * Val when i upgrade a binary package on a guest 1141666349 M * Val ... same file with same name will be overwriten, yeah ? 1141666359 M * Bertl Val: yep 1141666373 M * Val will it make problem like the one we talk ? 1141666508 J * webbist ~blowme@cpe-024-211-255-023.nc.res.rr.com 1141666605 M * daniel_hozac no. 1141666627 M * daniel_hozac most package managers will work with unification, by first unlinking the file and then rewriting it. 1141666661 M * Val ok 1141667261 M * Bertl okay, off now .. back later ... 1141667265 N * Bertl Bertl_oO 1141667307 M * webbist Anybody here use OpenVPS? 1141667676 M * daniel_hozac no, why? 1141667692 M * webbist Just curious 1141668724 J * dearaujo ~dan@pixpat.austin.ibm.com 1141668871 Q * Val Quit: bye bye all :) 1141671403 P * dearaujo 1141671671 N * Bertl_oO Bertl 1141671675 M * Bertl back now ... 1141671711 M * michal` hey Bertl 1141671764 M * michal` you are lucky supporting only 2.6 kernels i'm telling you ;) 1141671784 M * Bertl huh? 1141671788 M * michal` (porting anything bigger to 2.4 is tragic ;) 1141671800 M * Bertl we still have stable releases for 2.4 1141671810 M * michal` but not development ? 1141671822 M * Bertl nope, no devel branch on 2.4 1141671878 M * coocoon thanx for all byebye 1141671884 A * Bertl .o( at this point usually eyck pops up and requests a new stable relase for 2.4 ... waiting for that to happen now :) 1141671901 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1141671954 M * michal` somehow kernel folks have decided it is good to shift right 16 bits (>> 16) flags that are beeing passed as arguments to clonse in kernel thread 1141671962 J * fwl ~fwl@83-215-237-1.seek.stat.salzburg-online.at 1141671962 M * michal` in 2.6 1141671983 M * michal` i cannot imagine why (2.4 just sets them and does nothing more) 1141672188 M * blizz Bertl, got my new rootserver on the net with lots of hdd/ram/traffic, about to install vserver there :-P 1141672196 M * blizz thatll be fun ;) 1141672274 M * blizz btw, newest development release patched nicely regardless of some hunks with 2.6.15.6 1141672301 J * doener ~doener@i5387C20F.versanet.de 1141672383 M * doener evening! 1141672392 M * daniel_hozac evening doener! 1141672415 M * blizz hey doener 1141672437 M * blizz (i dont even like doener ! :P) 1141672582 M * doener http://linuxgazette.net/124/smith.html -- just sent a mail to the author, pointing out Linux-VServer as an add-on for the system (for example process isolation + token bucket scheduler) 1141673191 J * rs ~rs@office.dailymotion.com 1141673223 M * Cru but where to buy PCI gracards with recent chipsets? 1141673288 M * eyck hmm, did Bertl said anything about new stable release for 2.4? 1141673303 M * eyck yeah, go Bertl! 1141673322 M * doener Cru: they used mx400 cards, and a dumb kiosk system won't need a ati x1000 or whatever ;) 1141673408 M * Cru I mean a Geforce at all - I have not seen any PCI gracards with nVidia chipset at all for years now, only SiS, Via, etc. 1141673443 M * Bertl eyck: lol! 1141673459 M * Cru heh 1141673526 J * Aiken ~james@tooax8-173.dialup.optusnet.com.au 1141673534 M * Cru at least not at the common street stores 1141673660 J * doener_ ~doener@i5387C24B.versanet.de 1141673753 M * Cru doener: I mean a Geforce at all - I have not seen any PCI gracards with nVidia chipset at all for years now, only SiS, Via, etc. - at least not at the common street stores 1141673970 M * doener_ they stopped doing pci after the fx series IIRC 1141674075 Q * doener Ping timeout: 480 seconds 1141674644 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1141675635 M * doener_ that's possible already 1141675638 M * doener_ oops 1141678345 Q * gerrit Ping timeout: 480 seconds 1141680376 Q * liquid3649_ Remote host closed the connection 1141680631 J * yoh ~yoh@ravana.rutgers.edu 1141680665 M * yoh Hi to Vserver peple! 1141680706 M * yoh from a Debian ... human 1141680774 M * yoh I decided to give it a try: installed 0.30.209-2 and it kinda works (running apache2) but it is waay to slow -- takes up to 5 secods on react on the simplest HTTP request: where could I look for the source of the problem? 1141680784 J * gerrit ~gerrit@129.33.1.37 1141680859 M * daniel_hozac does DNS work inside the guest? 1141680899 M * daniel_hozac does equivalent access to the host avoid the problem? 1141680962 M * yoh DNS seems to work... the weird thing is that interface (ifconfig) doesn't have any address assigned although I believe it had to be 192.168.100.2... 1141680969 Q * click Ping timeout: 480 seconds 1141681007 M * yoh and I am not sure what is "equivalen access" means in the context 1141681034 M * doener_ does an apache server on the host (i.e. not in a vserver) show the same behaviour? 1141681052 M * doener_ the ip address issue is because the ip address has no name assigned 1141681063 M * yoh doener_, o - nope -- original apache (although 1.3) is working fine 1141681070 M * doener_ ifconfig only knows about named ip addresses, because it uses the old ioctl interfaces. 1141681091 M * daniel_hozac yoh: that's not equivalent. 1141681094 M * doener_ util-vserver configures ip addresses using "ip" from iproute, which uses the netlink interface and can handle unnamed ip addresses 1141681094 M * yoh doener_, relating ifconfig: thank you for the answer 1141681165 M * doener_ you can add a file called "name" (IIRC) in /etc/vservers//ip/0/ and put a name in there, then ifconfig will also show it 1141681184 Q * bonbons Quit: Leaving 1141681194 M * yoh BTW - for installation I followed http://www.unixshell.com/wiki/index.php/Creating_and_using_vserver_virtual_servers (networking option B) 1141681200 M * doener_ name like "eth0:0", but it can be almost anything up to 15 characters IIRC 1141681207 M * daniel_hozac no, name like 0. 1141681223 M * daniel_hozac eth0:0 would make it eth0:eth0:0... 1141681235 M * daniel_hozac (assuming that's allowed) 1141681254 M * daniel_hozac hmm, seems everyone is putting up their own documentation. 1141681267 M * daniel_hozac is linux-vserver.org really that hard to work with? 1141681284 M * doener_ hm, does the newvserver stuff handle the new configuration format now? 1141681326 M * daniel_hozac i don't think so. 1141681356 J * Aiken_ ~james@tooax6-032.dialup.optusnet.com.au 1141681412 M * yoh I have /etc/vservers/www/interfaces/0 with dev and ip files, and ifconfig does show eth0 but with no ip assigned, should I create ip/0 and then which file to dump the interface name? 1141681435 M * yoh in any case that is probably not the cause of the problem with slow reaction 1141681435 M * doener_ yoh: sorry, interfaces is right, not ip 1141681447 M * daniel_hozac lol, i missed that. 1141681454 Q * Aiken Ping timeout: 480 seconds 1141681467 M * daniel_hozac it most certainly is not the cause. 1141681472 M * daniel_hozac it's just a cosmetic issue. 1141681478 M * doener_ echo "foo" > interfaces/0/name should do the trick 1141681490 M * daniel_hozac (until you find other dumb programs that depend on the ioctls...) 1141681513 M * doener_ anyway, now to the real problem... 1141681590 M * doener_ does a ssh connection to the vserver show the same problem? 1141681727 M * yoh doener_, I didn't install ssh inside the vserver... should I? 1141681744 J * click click@ti511110a080-3151.bb.online.no 1141681752 M * yoh doener_, I just do vserver www enter to get inside 1141681809 M * doener_ ssh is the preferred way to get into the vserver 1141681815 M * Bertl yoh: did you stop the guest, add the name entry, then restart the guest? 1141681941 M * yoh Bertl, I tried echo ethX > interfaces/0/name, stop, start, had no effect.... BTW after start or during stop I have "RTNETLINK answers: Cannot assign requested address" which might be the relevant message :-) 1141681959 Q * webbist Ping timeout: 480 seconds 1141682027 M * yoh BTW is that a general evil practice to place vdirbase inside /etc?? or it was a debian maintainer decision (then I need to kick him with a bugreport) 1141682053 M * yoh I mean I have /etc/vservers/.defaults/vdirbase/www/ as a dirbase for my "www" vserver 1141682154 M * daniel_hozac vdirbase should be a symlink. 1141682159 M * daniel_hozac probably to /var/lib/vservers. 1141682234 M * yoh daniel_hozac, SHAME ON ME!!!! I didn't check it out after I saw www/vdir -> /etc/vservers/.defaults/vdirbase/www 1141682240 M * yoh daniel_hozac, ;-) 1141682339 Q * restill Quit: Leaving 1141682432 M * Bertl yoh: that's exactly what I meant 1141682449 M * Bertl yoh: you did it the wrong way round, and messed up the ips 1141682475 M * Bertl yoh: also ethX is not waht you want there 1141682525 M * Bertl yoh: ethX in name will give you (assumed that dev contains eth0) eth0:ethX 1141682545 M * Bertl and it doesn't work right now, because the IP is _already_ configured on the host 1141682567 M * Bertl because you didn't stop the guest first, and change the config then 1141682582 M * yoh Bertl, thank you! Trying to fix it 1141682587 M * Bertl yoh: use 'ip addr ls' on the host to show the ips 1141682597 M * Bertl and ip addr del to remove the offending one 1141682623 M * Bertl then use something like 'karl' or '1' in name to get eth0:karl or eth0:1 1141682623 M * yoh that works :-) 1141682650 M * Bertl of course, you can simply install ip (from iproute2) inside the guest and ignore the old aliases 1141682650 M * yoh now I got nasty eth0:100.2 ;-))) 1141682679 M * yoh that seems have no impact on the slow connection problem 1141682689 M * Bertl didn't say that :) 1141682735 M * Bertl the problem of your delays is probably a bad resolver config 1141682763 M * Bertl I _assume_ that the guest is not able to do name resolving properly (missing/bad /etc/resolv.conf) 1141682787 M * Bertl this will take a few seconds every time a resolver call is done until the relover times out 1141682810 M * yoh well -- whenever I ping any name within vserver it takes almost no time to resolv the name... 1141682901 M * Bertl from your previous statements I conclude that you are not really inside the guest, but jsut entered it 1141682935 M * Bertl depending on your tool config ('vserver-info - SYSINFO' you might upload that to pastebin.com or so) 1141682950 M * Bertl it might be that you are still using the host ip/resolver 1141683027 M * yoh YIKES! damn me -I better go RTFM - I had no idea that "entering" is not the same as to be within the guest :-) 1141683053 M * Bertl well, it's the management backdoor :) 1141683111 M * yoh http://pastebin.com/587793 1141683143 M * yoh ok - installing ssh and getting into ;-) 1141683157 M * Bertl looks good ... 1141683191 M * Bertl how do you access your http? 1141683221 M * Bertl +d 1141683441 M * yoh sshed into ... nslookup works fine 1141683498 M * yoh http from mozilla on a nearby box (so different IP). I am forwarding port 8888 to the vserver 1141683506 M * yoh +d? 1141683519 M * Bertl http+d :) 1141683556 M * Bertl okay, so resolver works, now let's try to reach the box via ssh 1141683566 M * Bertl (from inside the guest) 1141683570 M * yoh yeap - I am in via ssh -- everything smooth... 1141683585 M * yoh although I am in ssh from the same host... 1141683598 M * Bertl okay, now the other way round, from the box to the vserver (via ssh) 1141683654 M * yoh I meant from the same box into vserver... ok - will try to check from outside box into vserver box to port 2222 which must be forwarded to vserver 1141683723 M * yoh ssh works perfectly from the box I am trying to browse httpd from 1141683734 M * Bertl forwarded? 1141683741 M * Bertl you mean DNAT, right? 1141683741 M * yoh yeap 1141683757 M * yoh yeap 1141683776 M * yoh may be it has to do with apache2? let me quickly try 1.3? 1141683802 M * Bertl well, it works quite fine here (with apache2) 1141683810 M * Bertl check the logs for unusual messages 1141684010 J * dearaujo ~dan@cpe-66-25-189-193.austin.res.rr.com 1141684014 M * yoh the same with 1.3 apache :-/ couldn't see anything abnormal in logs besides initial attempts to bind to port 80 which is used by the mother-OS ;-) then I Listen to 192.168.100.2:8888 1141684018 M * Skram Yey, another austinite. 1141684029 M * yoh Bertl, and it works but very slowly 1141684033 N * bubulak_ bubulak 1141684066 M * yoh Bertl, actually it speeded up on the last inqueries... let me place some doc in there :-) 1141684108 M * yoh nah - probably was from the cache :-/ 1141684289 M * djudko_ been looking at the quota work don't see anything for 2.6 kernel...how current is that work? 1141684301 M * Bertl yoh: okay, let's get the output from 'testme.sh' on pastebin.com 1141684329 Q * mugwump Quit: Reconnecting 1141685072 Q * gerrit Ping timeout: 480 seconds 1141686189 M * SiD3WiNDR argh 1141686197 M * SiD3WiNDR I was going to /query mugwump with some questions 1141686202 M * SiD3WiNDR and he hasn't come back yet :p 1141686351 M * yoh Bertl, should testme.sh be ran within vserver? 1141686389 M * Bertl nope, on the host as root 1141686401 M * Bertl will test various vserver functionality 1141686417 M * dearaujo hello - can anyone explain quota hashes? high-level of course... 1141686453 M * Bertl sure, basically it is 'more than one' quota per superblock 1141686478 M * yoh Bertl, said succeeded for all of them 1141686486 M * yoh do you want it on pastebin? 1141686523 M * yoh http://pastebin.com/587916 appended to the previous one :-) 1141686560 M * Bertl excellent 1141686669 M * Bertl dearaujo: what would you like to know about it ... 1141686710 M * dearaujo i guess my confusion lies in xid tagging, and quotas 1141686720 M * dearaujo whether they are hand in hand 1141686725 M * dearaujo or seperate entieies 1141686729 M * dearaujo *entities 1141686756 M * dearaujo am I even making sense :) 1141686759 M * Bertl well, it is probably even more complicated as you expect right now 1141686780 M * Bertl I'll try to explain the various aspects, okay? 1141686784 M * dearaujo sounds good to me 1141686805 M * yoh Bertl, ok - it seems that the issue in the "repsponse" packets since vserver log of apache states that GET was very same second as it was on the box... 1141686829 M * yoh could it have something to do with postroting? 1141686867 M * Bertl yoh: postrotin, nice typo :) 1141686887 M * Bertl guess that might be the case, what about using tpdump on the host? 1141686978 Q * rs Quit: rs 1141687046 M * Bertl dearaujo: okay, first thing is the filesystem tagging 1141687062 M * dearaujo ok 1141687079 M * Bertl the tagging is currently tied to the xid, but we are in the process of making that independant 1141687102 M * Bertl i.e. the context will provide a 'tag' for the filesystem to use 1141687125 M * Bertl individual filesystems might override that tag by special filesystem (vfsmount) tags 1141687179 M * dearaujo ok 1141687219 Q * ntrs Ping timeout: 480 seconds 1141687238 M * yoh Bertl, heh heh -- I am not a knowledgable person of tcpdump... so don't kick hard :-) please see what I got in the simplest ever tcpdump http://pastebin.com/587916 1141687253 M * yoh Bertl, let me just run ethereal :-) 1141687283 M * Bertl dearaujo: the second aspect is disk limits 1141687308 M * dearaujo before you start - the xid tagging is useful shared disk space right 1141687309 M * Bertl disk limits have to be per context and per filesystem 1141687330 M * Bertl dearaujo: yes, for permission purposes and the disk limits (or quota later) 1141687337 M * dearaujo ok 1141687515 M * yoh Bertl, Thank you for all your help so far - I need to rush home to pick up my baby 1141687544 M * Bertl okay, do that, we'll look at it later 1141687572 M * Bertl dearaujo: okay, so disk limits are basically something with (sb,xid) characteristics 1141687587 M * dearaujo sb? 1141687587 Q * michal` Ping timeout: 480 seconds 1141687593 M * Bertl superblock 1141687599 M * dearaujo oh ok sorry 1141687601 M * dearaujo thx 1141687605 M * Bertl np 1141687624 M * Bertl and this is also true for quota (user/group) 1141687663 M * Bertl so (a long time ago) I came to the conclusion that there are two options to make quota per guest (on a shared partition) possible 1141687701 M * Bertl one would be to have a separate filesystem layer for each guest (this was actually the first approach I took/tested) 1141687733 M * Bertl the other, to make the quota and limits (sb,xid) instead of sb only 1141687779 M * Bertl for the disk limits, we have special structures storing the current and total size (this is static info) per (sb,xid) 1141687808 M * Bertl quotas are a little more complicated, as they do not have static structures/entries 1141687815 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1141687827 M * dearaujo oh? 1141687829 M * Bertl i.e. they are dynamic, and inmainline only based on (sb,qid) 1141687846 M * Bertl qid being the quota type user or group 1141687887 M * Bertl so I decided to make it a little more flexible, and have basically arbitrary quota structures in a hash 1141687943 J * michal` ~michal@www.rsbac.org 1141688040 M * Bertl dearaujo: the 'quota hashes' are the result of this endavor, and the form the foundation for per guest quota on a shared partition 1141688060 M * Bertl (which results in several quota hashes per superblock) 1141688114 M * dearaujo so to me it sounds like the xid tagging provides a space for sharing and the quota/limits limit each guest to a particualr size... 1141688136 M * Bertl precisely 1141688142 M * dearaujo and this can be done via limits or quotas 1141688150 M * Bertl or both :) 1141688158 M * dearaujo heh :) 1141688174 M * dearaujo interesting 1141688193 M * dearaujo and this is available in the 2.01 release? 1141688212 M * Bertl no, only the disk limits is, and the quota hashes 1141688222 M * Bertl the per guest quota is not included yet 1141688235 M * dearaujo i see 1141688237 M * Bertl (but it is available for the 2.4 kernel) 1141688245 M * dearaujo i assume there is plan for it? 1141688251 M * dearaujo for 2.6? 1141688257 M * Bertl it is planned, but the interest is very low 1141688281 M * Bertl (and when I do it, I want to rewrite a bunch of things in the current quota system) 1141688304 M * dearaujo so for 2.6 I can merely use limits 1141688310 M * dearaujo per guest 1141688516 M * dearaujo Bertl: thanks for the help 1141688623 M * Bertl yes, limits are working fine, quota is to 80% here but not usable 1141688746 P * stefani I'm Parting (the water) 1141688910 J * Thorsten ~Thorsten@dslb-084-058-167-040.pools.arcor-ip.net 1141689343 Q * lilalinux Remote host closed the connection 1141689352 P * dearaujo