1156466081 J * ekc ~ekc@oxnard-cuda1-cable-bundle-68-69-253-79.ventca.adelphia.net 1156466501 M * Bertl welcome ekc! 1156467287 M * cehteh what was the trick with running bind in a guest without elevating privileges or recompiling? 1156467300 M * cehteh (was there one?) 1156467837 M * Skram brc_: ever work on the updating a vps-snapshot stuff? 1156468366 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net 1156470788 Q * Hecticee 1156471527 M * brc_ SKRAM 1156471535 M * brc_ I never reasd this channel, just switched in :) 1156471552 M * Skram :) right on 1156471565 M * brc_ Not much, i am having lot of problems lately (work related). 1156471571 M * brc_ hds dieing, crashes, etc :) 1156471574 M * Skram Icky 1156471580 M * Skram Sorry abou that-- never fun. 1156471652 M * brc_ everythning is almost 100% now 1156471660 M * brc_ hpe in some weeks i will be able to play agani 1156471679 M * brc_ Skram: Ever worked with vserver unification ? 1156473290 M * Skram im not in a position to teach anything about it 1156473294 M * Skram consider that a "no" 1156473369 Q * cehteh Remote host closed the connection 1156474375 Q * DreamerC Quit: leaving 1156474908 J * cehteh ~ct@cehteh.homeunix.org 1156476540 J * DreamerC ~dreamerc@59-112-1-201.dynamic.hinet.net 1156476883 Q * cehteh Read error: Connection reset by peer 1156477326 J * cehteh ~ct@cehteh.homeunix.org 1156477897 Q * cehteh Ping timeout: 480 seconds 1156477916 M * Bertl okay, off to bed now ... have a good one everyone! cya tomorrow! 1156477924 N * Bertl Bertl_zZ 1156479793 N * ekc ekc_away 1156480337 Q * comfrey Ping timeout: 480 seconds 1156480585 M * daniel_hozac derjohn2: installing it should just be a -m template --initstyle plain -- -t stage3... ;) 1156481684 Q * derjohn2 Ping timeout: 480 seconds 1156482095 J * derjohn2 ~aj@dslb-084-058-201-002.pools.arcor-ip.net 1156483317 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net 1156485208 J * Wonder_in ~wonder_in@83.102.213.20 1156485353 Q * Wonder_in 1156485465 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1156485744 J * dna ~naucki@31-243-dsl.kielnet.net 1156487195 Q * comfrey Remote host closed the connection 1156488798 Q * DreamerC Remote host closed the connection 1156489426 J * DreamerC ~dreamerc@59-112-1-201.dynamic.hinet.net 1156489932 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net 1156490072 Q * mire Ping timeout: 480 seconds 1156490931 J * pisco ~pampel@p5087948F.dip0.t-ipconnect.de 1156491091 P * pisco 1156491474 J * suka ~suka@bi01p1.co.us.ibm.com 1156491552 J * pagano_ ~pagano@131.154.5.20 1156491822 Q * pagano Ping timeout: 480 seconds 1156492179 Q * pagano_ Remote host closed the connection 1156494992 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156495298 J * coocoon ~coocoon@p54A065F2.dip.t-dialin.net 1156497114 N * Zaki_ Zaki 1156498735 J * pisco ~pampel@tor-irc.dnsbl.oftc.net 1156498796 M * nayco_work hello, all ! 1156500324 Q * michal_ Ping timeout: 480 seconds 1156500731 J * michal_ ~michal@www.rsbac.org 1156501786 J * shedi ~siggi@dsl-149-109-85.hive.is 1156502438 M * harry any ideas for 2.0.2 release? 1156502445 M * harry what extra patches are needed etc...?? 1156502559 J * mire ~mire@212-167-222-85.COOL.ADSL.VLine.verat.net 1156502664 M * derjohn harry, Bertl_zZ said he will consider a release with the last RC soon. more I dont know ;) 1156502791 M * Hollow hi derjohn 1156502792 M * nayco_work Is there any disk quota specialist, here, or someone who knos how to trace a program to see were a call fails ? 1156502801 M * Hollow did you look at consolidating faqs? ;) 1156502998 M * derjohn Hollow, what about makeing me a bad conscience ? 1156503005 M * nayco_work Hello, Hollow ! I used this : http://linux-vserver.org/Standard+non-shared+quota 1156503026 M * nayco_work But I just discovered http://www.5dollarwhitebox.org/wiki/index.php/Howtos_Linux-Vserver_With_LVM_And_Quotas 1156503039 M * derjohn Hollow, what is the status of the wiki-migration at all? 1156503049 M * derjohn read: how much time is left for me? 1156503085 M * derjohn nayco_work, I dunno much about quota, but with recent kernels and patches everything got simpler 1156503114 M * derjohn nayco_work, read: you simply set a quota for the guest and restart ist. 1156503122 Q * FireEgl Quit: Bye... 1156503140 M * derjohn nayco_work, or do you want (std fs ) quota within a guest? 1156503196 M * nayco_work derjohn: user/group quota inside the guest (setquota, repquota, quotaon...) 1156503305 J * lilalinux ~plasma@80.69.35.186 1156503768 Q * pisco Remote host closed the connection 1156503963 M * nayco_work quotactl(Q_XGETQSTAT|USRQUOTA, NULL, 0, NULL) = -1 EFAULT (Bad address) 1156503968 M * nayco_work sniffle ;-( 1156504173 M * derjohn nayco_work, user/group quota inside the guest ... that was some evil magick. you might check/search chatlogs. 1156504234 M * harry VServer Control Daemon <== wtf? 1156504246 M * harry where can one find that? is that the future? 1156504250 M * harry etc..etc... 1156504275 M * Hollow harry: svn.linux-vserver.org 1156504286 M * Hollow for info: http://home.xnull.de/work/vserver/manual/manual.pdf 1156504288 M * nayco_work derjohn: that's what i'm doing now ;-) 1156504302 M * harry Hollow: i printed that manual... 1156504311 M * harry which is quite useless since it's allmost completely empty :) 1156504317 M * Hollow i updaed it this morning 1156504323 M * harry bloody hell :) 1156504331 M * Hollow well, i can't write a manual in two days, i'm sorry ;) 1156504332 M * harry how much updated stuff is there? 1156504333 M * derjohn harry, there is also openvcp , which was a webfoo and a agent per host. 1156504339 M * Hollow 6 pages long xmlrpc intro 1156504348 M * harry Hollow: mkay , tnx 1156504470 M * harry how does it all work with networking etc??? 1156504526 A * harry reads more vserver stuff... 1156504532 M * Hollow which networking stuff? 1156504549 M * harry especially per vps quota etc... 1156504558 M * harry well... imagine i want 6 different servers 1156504564 M * harry on 6 different lan's 1156504572 M * harry routing is allways... wel... a problem 1156504584 M * Hollow well, i'm not a routing expert.. 1156504607 A * harry allways has to fix weird routing problems here... so it is a problem for me ;) 1156504618 M * harry welll... "problem" offcourse 1156504650 M * harry especially if you don't want to worry about every extra vps's routing 1156504864 M * derjohn nayco_work, there is http://vserver.13thfloor.at/Experimental/delta-quota-fix01.diff, which seems to be made at -rc10 times... But I do not even know if it's per-guest quota or per-vserver quota fix ;) 1156504875 A * derjohn now away for some mins. 1156505097 M * nayco_work derjohn: I gonna check... 1156505188 M * nayco_work small patch indeed... 1156505189 M * nayco_work - HANDLE_CASE(QUOTA, dlim, value); 1156505190 M * nayco_work + HANDLE_CASE(QUOTA, quota, value); 1156505653 M * derjohn nayco_work, I write the changelogs, but began with some -rc2x .. dunno what happend before. I would say: wait for Bertl_zZ or daniel_hozac ? 1156505817 Q * cdrx Ping timeout: 480 seconds 1156506097 M * Hollow harry: actually, now the pdf is updated.. i didn't export it... 1156506287 J * FireEgl FireEgl@Sebastian.Atlantica.US 1156506528 M * mnemoc harry: hi, are you the 'vserver+grsec' guy? .... there is a mail on grsec list about virtualization with grsec which i think you are right person to answer :) 1156508324 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156508338 N * Bertl_zZ Bertl 1156508343 M * Bertl morning folks! 1156508348 M * derjohn moin Bertl ! 1156508774 J * coocoon ~coocoon@p54A056B2.dip.t-dialin.net 1156508855 M * nayco_work 'llo, Bertl 1156508857 M * nayco_work ;-) 1156508911 M * nayco_work Bertl: # LC_ALL=C strace quotaon /home 1156508917 M * nayco_work quotactl(Q_XGETQSTAT|USRQUOTA, NULL, 0, NULL) = -1 EFAULT (Bad address) 1156508920 M * nayco_work :-( 1156509513 M * harry mnemoc: i'm the one :) 1156509523 M * harry Hollow: i'll check it out later :) 1156509543 J * cdrx ~legoater@242.32.96-84.rev.gaoland.net 1156509544 M * harry mnemoc: i answered it...b ut now that i think of it... i didn't reply to the list 1156509574 M * mnemoc doh 1156509589 M * mnemoc bad PR-guy :( 1156509601 M * harry there... fixed ;) 1156509649 M * harry ctrl-r or ctrl-shift-r... 1156509653 M * harry allways difficult :) 1156509691 M * mnemoc :) 1156509703 M * harry Your message to grsecurity awaits moderator approval 1156509717 M * harry bloody hell... i must have subscribed under another mailaddress... 1156510771 M * Bertl nayco_work: hmm, EFAULT sounds interesting ... 1156510776 M * Bertl nayco_work: which filesystem? 1156510789 N * Belu_zZz Belu 1156511155 Q * sladen Ping timeout: 480 seconds 1156511266 J * sladen paul@starsky.19inch.net 1156511340 M * nayco_work Bertl: Sorry, I was away. Well XFS, as usal ;-) 1156511349 M * nayco_work +u 1156511708 M * nayco_work Oh, I forgot : util-vserver 0.30.210 , 2.6.17.8-vs2.1.1-rc29 1156511738 M * nayco_work testme.sh-0.15 OK 1156511769 M * derjohn is there somethign like "vserver * stop" ? (I dont mean a for in $(vserver-stat | awk 'NR > 2 { print ($8) }') ....) 1156511782 M * nayco_work "halt" 1156511785 M * nayco_work :-p 1156511824 M * derjohn nayco_work, NR > 2 .. the root server should stay up :) 1156511907 M * Bertl nayco_work: did you a) enable xfs quota, and b) does it work with your tools on the host? 1156512036 M * derjohn Bertl, do we need xaatr for the feature? 1156512041 M * derjohn xattr ? 1156512053 M * Bertl for what? 1156512102 M * derjohn for the per 'guest quota' 1156512538 M * Bertl no, per guest quota requires xid tagging on shared partitions (nothing on separate ones) 1156512554 M * derjohn tagxid? 1156512566 M * Bertl for shared partitions, yes 1156512586 M * Bertl but shared quota is not implemented in 2.6 trees (yet) 1156512824 M * Bertl btw, http://plm.testing.osdl.org/patches/show/5228 1156513084 A * waldi needs to add a sparc64 vserver variant to the debian kernels :) 1156513113 M * Bertl should work out of the box, no? 1156513120 M * waldi should 1156513120 M * waldi cpu : UltraSparc T1 (Niagara) 1156513124 M * waldi ncpus probed : 32 1156513148 M * Bertl we got that running half a year ago :) 1156513162 M * Bertl ask derjohn, he was there ... 1156513192 M * waldi yeah, the machine is available since half a year but the location was not able to add a console so we can't install it ... 1156513229 A * waldi .o( yeah, i should've say: the machine comes to me ... ) 1156513235 M * Bertl ok :) 1156513597 Q * ekc_away Ping timeout: 480 seconds 1156513752 M * nayco_work Bertl: I read last our last year's discussion's log, so : 1156513767 M * nayco_work # zgrep _Q..T /proc/config.gz 1156513767 M * nayco_work CONFIG_XFS_QUOTA=y 1156513767 M * nayco_work CONFIG_QUOTA=y 1156513767 M * nayco_work CONFIG_QFMT_V1=m 1156513767 M * nayco_work CONFIG_QFMT_V2=y 1156513770 M * nayco_work CONFIG_QUOTACTL=y 1156513862 M * nayco_work But on the host : 1156513870 M * nayco_work (mtab) /dev/mapper/volume1-ic--srv--fichier--home /vservers/ic-srv-fichier/home xfs rw,quota 0 0 1156513880 M * nayco_work manually mounted, so nothing in fstab 1156513880 J * Viper0482 ~Viper0482@p54975601.dip.t-dialin.net 1156513882 M * nayco_work and : 1156513905 M * nayco_work # LC_ALL=C repquota /vservers/ic-srv-fichier/home 1156513905 M * nayco_work repquota: Mountpoint (or device) /vservers/ic-srv-fichier/home not found. 1156513906 M * nayco_work repquota: Not all specified mountpoints are using quota. 1156513914 M * nayco_work well... 1156514012 M * nayco_work Ok, I added it to fstab, then "mount -a", and same result... 1156514139 M * Bertl on the host? 1156514186 M * Bertl brb, getting something to eat ... 1156514329 M * nayco_work Bertl: Yes, on the host 1156514502 M * Bertl so it doesn't work on the host, right? 1156514521 M * nayco_work nope. 1156514547 M * Bertl so it is not really vserver related, but have you checked the xfs mount options? 1156514566 M * nayco_work # mount 1156514570 M * Bertl IIRC, xfs needs some weird options plus very rrecent quota tools 1156514571 M * nayco_work /dev/mapper/volume1-ic--srv--fichier--home on /vservers/ic-srv-fichier/home type xfs (rw,quota) 1156514591 M * Bertl try yo mount with usrquota instead 1156514597 M * nayco_work # cat /etc/fstab 1156514597 M * Bertl *to 1156514603 M * nayco_work /dev/mapper/volume1-ic--srv--fichier--home /vservers/ic-srv-fichier/home xfs defaults,quota 1 2 1156514616 M * nayco_work ok 1156514642 M * nayco_work is it mandatory to modify fstab, or can I manually mount ? 1156514654 M * Bertl might also be that you need to do something at filesystem creation time 1156514665 M * Bertl you can do it manually of course 1156514667 M * nayco_work I naver had to... 1156514674 M * nayco_work -a+e 1156514738 M * nayco_work mount -t xfs -o usrquota /dev/mapper/volume1-ic--srv--fichier--home /vservers/ic-srv-fichier/home 1156514745 M * nayco_work # mount 1156514749 M * nayco_work /dev/mapper/volume1-ic--srv--fichier--home on /vservers/ic-srv-fichier/home type xfs (rw,usrquota) 1156514754 M * nayco_work # cat /etc/mtab 1156514759 M * nayco_work /dev/mapper/volume1-ic--srv--fichier--home /vservers/ic-srv-fichier/home xfs rw,usrquota 0 0 1156514783 M * nayco_work # LC_ALL=C repquota /vservers/ic-srv-fichier/home 1156514783 M * nayco_work repquota: Mountpoint (or device) /vservers/ic-srv-fichier/home not found. 1156514785 M * nayco_work repquota: Not all specified mountpoints are using quota. 1156514826 M * nayco_work Oh, the trace is interesting : 1156514835 M * Hollow nayco_work: you know that there is a xfs_quota tool in the xfsprogs package? 1156514914 M * nayco_work how, what is the url of linux-vserver paste ? 1156514916 M * nayco_work pls 1156514974 M * Hollow http://paste.linux-vserver.org 1156514975 M * coocoon http://paste.linux-vserver.org/ 1156514982 M * coocoon ;-) 1156515050 M * nayco_work thwx 1156515068 M * nayco_work Hollow: no, I did not know... This is... interesting :D 1156515073 M * nayco_work thx 1156515109 M * Hollow i don't use quotas myself, but maybe worth a look ;) 1156515110 M * nayco_work http://paste.linux-vserver.org/286 1156515142 M * Hollow seems you're missing some functionality in your kernel 1156515155 M * nayco_work But... ? 1156515169 M * nayco_work I have the good option configured, I think :-( 1156515178 M * Hollow yeah, i just scrolled to your paste.. ;) 1156515235 M * nayco_work options 1156515253 M * nayco_work Hollow: and ? Did you see something ? 1156515260 M * derjohn is there anythign bas with setting sched_hard and sched_prio in flags? 1156515268 M * Hollow nayco_work: no, looks ok to me 1156515270 M * nayco_work because "quotactl(Q_XGETQSTAT|GRPQUOTA, "/dev/mapper/volume1-ic--srv--fic"..., 0, {0, 0, 0, 0, 0, 0, 0, 0}) = -1 ENOSYS (Function not implemented)" is... Er... 1156515271 M * derjohn which one is preferred? 1156515524 M * Bertl nayco_work: I'd suggest to configure/boot a vanilla 2.6.17.11 kernel and get xfs quota working there (if possible) then continue with a vserver kernel, I will check xfs quota in my test setup soon 1156515630 M * nayco_work Bertl: Ok, I gonna try 1156515929 M * nayco_work Ok, I'm compiling a new 2.6.17.8 kernel on an bi-Opteron test server. Wait and see 1156515932 M * nayco_work ... 1156515967 M * derjohn vdu: stat(): No such file or directory <---- what is that? 1156516008 M * Bertl that is vdu trying to stat a directory (or context?) which is gone now 1156516040 M * derjohn i gonna check my mounts (tried to remount tagxid) 1156516143 M * Bertl well, remounting tagxid does not work :) 1156516183 M * harry Bertl: any ideas on a release of 2.0.2 ? 1156516213 M * derjohn Bertl, yeah, sry for being not precise. I unmounted and mounted again :) 1156516241 M * Bertl harry: we got a bug report and we are currently seeing some ppc regressions in the compile 1156516256 M * Bertl harry: have to check what exactly happens there, maybe we want to fix that 1156516259 M * harry aha... fix that one first ;) 1156516267 M * harry need any help? ;) 1156516282 M * Bertl well, you can dig through the logs if you like :) 1156516291 M * Bertl http://plm.testing.osdl.org/patches/show/5228 1156516415 M * harry i'll ckeck it out a bit later... 1156516418 M * harry now... 0xc0ffee 1156516419 M * harry ;) 1156516436 M * derjohn Bertl, thx. the prob was /etc/vservers/${GUEST}/dlimits/0/directory pointing to the wrong path. 1156517070 M * nayco_work oh, 4 minutes to compile the kernel, not too bad. Now, the modules... Guys, is "make -j 2" enough for a dual-proc ? 1156517161 M * Bertl try -j 99 and see if that is faster :) 1156517208 M * coocoon for opteron dual core i use -j 4 1156517270 M * nayco_work Bertl: Hummm, welll... ;-) 1156517275 J * pisco ~pampel@tor-irc.dnsbl.oftc.net 1156517280 Q * pisco 1156517317 M * nayco_work coocoon: in fact, 2xnumber of cpu or cores ? 1156517341 M * coocoon core 1156517654 M * nayco_work coocoon: yes, I meant total number of cores, but my opterons are monocores ;-) 1156517868 M * Bertl nayco_work: is that with stable or devel branch, btw? 1156518114 M * nayco_work ok, rebooting... 1156518128 M * nayco_work Bertl: 2.6.17.8-vs2.1.1-rc29 1156518134 M * nayco_work This is devel branch, non ? 1156518136 M * nayco_work -n 1156518139 M * daniel_hozac yep. 1156518141 M * Bertl yep, okay, tx 1156518263 M * nayco_work Ok, here are the results : 1156518264 M * nayco_work Linux morpheus.pulry.univ-nantes.prive 2.6.17.8 #1 SMP Fri Aug 25 16:42:45 CEST 2006 x86_64 Dual Core AMD Opteron(tm) Processor 265 unknown GNU/Linux 1156518273 M * nayco_work # mount 1156518279 M * nayco_work /dev/mapper/volume1-home on /home type xfs (rw,usrquota,grpquota) 1156518297 M * Bertl ah, usr/grp quota options now? 1156518300 M * nayco_work # LC_ALL=C repquota /home 1156518300 M * nayco_work *** Report for user quotas on device /dev/mapper/volume1-home 1156518300 M * nayco_work Block grace time: 7days; Inode grace time: 7days 1156518300 M * nayco_work Block limits File limits 1156518300 M * nayco_work User used soft hard grace used soft hard grace 1156518302 M * nayco_work ---------------------------------------------------------------------- 1156518305 M * nayco_work root -- 40948 0 0 9 0 0 1156518307 M * nayco_work nayco -- 20 0 0 7 0 0 1156518328 M * nayco_work Bertl: Yes, I tried this way. I can revert to "quota", if you want 1156518342 M * Bertl please do so, but I guess I already know what causes that 1156518362 M * Bertl (seems the xfs superblock is missing the quota hashes) 1156518385 M * nayco_work # mount -o remount /home 1156518390 M * nayco_work # mount 1156518392 M * phedny afaik quota on xfs is working differently then on other os'es 1156518399 M * nayco_work # LC_ALL=C repquota /home 1156518399 M * nayco_work *** Report for user quotas on device /dev/mapper/volume1-home 1156518399 M * nayco_work Block grace time: 7days; Inode grace time: 7days 1156518399 M * nayco_work Block limits File limits 1156518399 M * nayco_work User used soft hard grace used soft hard grace 1156518402 M * nayco_work ---------------------------------------------------------------------- 1156518404 M * nayco_work root -- 40948 0 0 9 0 0 1156518407 M * nayco_work nayco -- 20 0 0 7 0 0 1156518468 Q * ag- Quit: BRB 1156518512 M * nayco_work Bertl: ;-) That remarmbers me of something last year... http://irc.13thfloor.at/LOG/2005-09/LOG_2005-09-12.txt 1156518527 M * nayco_work "1126485752 M * Bertl fs/xfs/linux-2.6/xfs_super.c line 771 1156518527 M * nayco_work 1126485759 M * Bertl sb->s_qcop = &linvfs_qops; 1156518527 M * nayco_work 1126485767 M * Bertl below that, add the following: 1156518529 M * nayco_work 1126485781 M * Bertl sb->s_dqh->dqh_qcop = &linvfs_qops;" 1156518546 J * ag- ag@caladan.roxor.cx 1156518578 M * Bertl *checking now* 1156518583 M * Bertl welcome ag-! 1156518696 M * Bertl nah, seems we do that (or similar), but it looks like xfs doesn't call alloc_super() 1156518730 M * nayco_work Last year, you had to make me add this for quotas to work _on-the_host_ ;-) 1156518750 M * nayco_work anyway, i've now just patched the above kernel, and am recompiling. 1156518787 Q * comfrey Ping timeout: 480 seconds 1156518922 M * daniel_hozac Bertl: hmmm, how do you get linux-2.6.17.11's plm id? 1156518953 M * Bertl it's linux-2.6.17.11 now 1156518957 M * Bertl http://plm.testing.osdl.org/patches/show/linux-2.6.17.11 1156518996 M * daniel_hozac oh, i was just using the wrong script... 1156519009 A * daniel_hozac was trying to run plmdiff_new 1156519061 M * daniel_hozac running the right script makes it all work nicely again, heh. 1156519075 M * Bertl good to hear :) 1156519126 M * daniel_hozac hmm, actually, no, i only get the sparse-2 results. weren't there other differences? 1156519134 M * Bertl btw, I can't make heads or tail of that diff (for ppc) 1156519163 M * Bertl I'll upload it, sec 1156519168 M * daniel_hozac ok, thanks. 1156519264 M * Bertl cross_5228.out, Experimental 1156519304 M * daniel_hozac thanks. 1156519431 J * stefani ~stefani@tsipoor.banerian.org 1156519496 M * daniel_hozac i love how it seems to mix all the lines together into really weird combinations... 1156519501 M * nayco_work *reboot* 1156519616 M * harry q: i want a specific gateway defined for my vps 1156519628 M * harry is there a simple way? 1156519632 M * Bertl multiple routing tables 1156519644 M * harry bwerk... advanced routing again :( 1156519681 M * Bertl cool, I found an elementary bug in our (or maybe mainline?) kernel checks 1156519691 M * Bertl kernel config checks that is 1156519709 M * Bertl CONFIG_XFS_QUOTA can be on, without CONFIG_QUOTA being enabled 1156519799 M * nayco_work # uname -a 1156519799 M * nayco_work Linux morpheus.pulry.univ-nantes.prive 2.6.17.8-vs2.0.2-rc29 #2 SMP Fri Aug 25 17:13:11 CEST 2006 x86_64 Dual Core AMD Opteron(tm) Processor 265 unknown GNU/Linux 1156519800 M * harry Bertl: when do i add the routing stuff? in pre-start? 1156519806 M * nayco_work # LC_ALL=C repquota -a 1156519806 M * nayco_work *** Report for user quotas on device /dev/mapper/volume1-home 1156519806 M * nayco_work Block grace time: 7days; Inode grace time: 7days 1156519806 M * nayco_work Block limits File limits 1156519806 M * nayco_work User used soft hard grace used soft hard grace 1156519808 M * nayco_work ---------------------------------------------------------------------- 1156519810 M * harry imho: not possible because the if won't be up 1156519811 M * nayco_work root -- 40948 0 0 9 0 0 1156519813 M * nayco_work nayco -- 20 0 0 7 0 0 1156519816 M * nayco_work Work in the host 1156519818 M * nayco_work +s 1156519833 M * Bertl okay, you are now with stable, there it should work fine 1156519835 M * harry post-start: impossible because some start-scripts need to be able to have the network running 1156519862 M * Bertl harry: any time, you can even add it permanently on the host 1156519885 M * harry no i can't... if i bring down the host, theere is no ip on that interface 1156519889 M * harry ==> no routing for that network 1156519930 M * Bertl routing tables should not care about that too much, IMHO 1156519941 M * harry : 17:31 lois ~ ;route add -net 134.58.243.0/24 gw 134.58.243.254 1156519941 M * harry SIOCADDRT: Network is unreachable 1156519959 M * Bertl you don't want route, use ip route and the proper table 1156519963 M * harry indeed they shouldn't... they do 1156519972 M * daniel_hozac that command doesn't make sense. 1156519984 M * harry daniel_hozac: sure it does 1156519987 M * daniel_hozac you can't tell it to route x.x.x.0 through x.x.x.y. 1156519997 M * harry ah, hehe... you're right :) 1156520047 M * harry Bertl: if i do a "vserver luditdb start", and i define 134.58.243.2/24 on eth0... 1156520107 M * nayco_work Bertl: I tried to build util-vserver-0.30.210-5mdk.src.rpm (rpm --rebuild...), and it fails. Do you want the output ? 1156520135 M * derjohn2 harry, routes with unreachable nexthops are removed by kernel 1156520143 M * harry derjohn2: no kiddin' 1156520166 M * harry the situation here is this: i want a host, which is reachable by a management network, nothing more 1156520178 M * harry the vps's on it, should have public ip's on other interfaces 1156520191 M * harry what's the sanest way to do that? 1156520212 M * Bertl nayco_work: do you have a mandrake system? 1156520222 M * derjohn2 harry, how does the host reach internet? some router on 192.xx ? 1156520224 M * harry on my production server, i have 1 "placeholder" ip per subnet, and advanced routing to get it all working "well" 1156520239 M * harry derjohn2: yes 1156520309 M * nayco_work Bertl: # cat /etc/mandriva-release 1156520309 M * nayco_work Mandriva Linux release 2006.0 (Official) for x86_64 1156520312 M * derjohn2 and the guests have public ip? 1156520312 M * harry it would be nice to be able to define a gateway for each vps 1156520316 M * harry derjohn2: yes 1156520324 M * derjohn2 well, you could do so 1156520330 M * Bertl nayco_work: yes please, upload it to paste.linux-vserver.org 1156520331 M * nayco_work rpm --rebuild util-vserver-0.30.210-4mdk.src.rpm works nice \o/ !!! 1156520334 M * harry an ip, an interface and a gateway for that ip 1156520336 M * daniel_hozac Bertl: given that vanilla is also broken, do we care? the errors seem pretty severe too, so i don't think they appear in real life. 1156520336 M * derjohn2 harry, or use 1:1 nat 1156520362 M * Bertl daniel_hozac: nah, I'm not worried much, just confused 1156520367 M * harry derjohn2: isn't that a large overhead??? 1156520372 J * comfrey ~comfrey@216-99-217-92.dsl.aracnet.com 1156520378 M * daniel_hozac yeah, i wonder how they managed to produce that. 1156520385 M * harry 1:1 NAT? 1156520386 M * derjohn2 harry, but adding a secondary IP to the host (public ip) and block access to the mgmt ip via -j DROP 1156520395 M * derjohn2 s/but/or 1156520408 M * harry derjohn2: that's what i do on our "production" server... 1156520417 Q * Viper0482 Ping timeout: 480 seconds 1156520417 M * harry but it s a waste of useless ip addresses 1156520428 M * harry useless waist of ip addresses :) 1156520435 M * harry waste lol 1156520447 M * daniel_hozac i wish i had a waist made of IP addresses... 1156520450 M * derjohn2 harry, my waist dosnt fire ip, yours ? :) 1156520473 M * harry that would mean, every real host in a subnet should have 1 fake ip... 1156520522 M * harry mkay, we have a B class range... but that doesn't mean we have that much in spare 1156520569 Q * stefani Ping timeout: 480 seconds 1156520571 M * harry so... the only possible thing to do is: set up advanced routing and use a useless fake address :S 1156520588 M * harry man, this sucks :( 1156520613 M * derjohn2 well, harry , you can setup a rt_table, which matches "from src bla" and give that ad ifferent default gw than the std rt_table 1156520636 M * harry derjohn2: 17:42 < harry> so... the only possible thing to do is: set up advanced routing and use a useless fake address :S 1156520641 M * derjohn2 harry, yes, but advanced routing is pretty ready 1156520643 M * derjohn2 *easy 1156520652 M * harry derjohn2: not really... 1156520666 M * harry right now, eth0 is down on the machine 1156520670 M * derjohn2 why? in your setup you should be fine with 5 lines 1156520678 M * harry i want to start the vps now... 1156520686 M * harry but then i won't have a route for that ip 1156520696 M * harry the vps should have a certain gateway 1156520697 M * nayco_work Bertl: http://paste.linux-vserver.org/287 1156520699 M * derjohn2 then ssh to the host ? 1156520712 M * derjohn2 and set up the rt:table ... 1156520735 M * harry derjohn2: how can i set up a route for a network, if there is no interface up on that network??? 1156520755 M * derjohn2 hey, the rt_tables are 'dumb' 1156520757 M * daniel_hozac nayco_work: what dietlibc version is that? 1156520763 M * derjohn2 harry, simply set them 1156520778 M * harry mkay... i'll follow your lead... tell me 1156520787 M * harry the ip of the vps will be: 134.58.243.2 1156520789 M * derjohn2 ok, give me 30 secs 1156520791 M * harry it s a /24 1156520793 M * nayco_work daniel_hozac: er..... 1156520796 M * harry gateway: 134.58.243.254 1156520820 M * nayco_work # rpm -qa | grep -i libc 1156520820 M * nayco_work glibc-2.3.5-5mdk 1156520820 M * nayco_work perl-Libconf-0.39.21-1mdk 1156520820 M * nayco_work glibc-devel-2.3.5-5mdk 1156520841 M * Bertl daniel_hozac: # 1156520841 M * Bertl vlogin.c:(.text+0x396): undefined reference to `openpty' 1156520854 M * Bertl (seems to happen on x86_64 :) 1156520864 M * harry ip route add 134.58.243.0/24 dev eth0 table 134.58.243-net 1156520864 M * harry ip route add default via 134.58.243.254 dev eth0 table 134.58.243-net 1156520864 M * harry ip rule add from 134.58.243.0/24 table 134.58.243-net pref 1000 1156520866 M * harry ? 1156520881 M * harry (if i add a 134.58.243-net 1156520908 M * daniel_hozac Bertl: probably a glibc thing though. :) 1156520928 M * daniel_hozac nayco_work: try with dietlibc. 1156520931 M * derjohn2 harry http://paste.linux-vserver.org/288 1156520950 M * Bertl daniel_hozac: ah, didn't assume he was using glibc 1156520964 J * Viper0482 ~Viper0482@p54977DD8.dip.t-dialin.net 1156520969 M * Bertl probably have to add that to the dependancies .. hmm, should be there IMHO 1156520982 M * daniel_hozac but that's a good point, i need to fix it so it compiles with glibc too. 1156521023 M * Bertl seems I removed the dietlibc dependancy 1156521050 M * harry derjohn2: http://paste.linux-vserver.org/289 1156521055 M * nayco_work Anyway, I installed 210-4mdk and it works. I'm building a test vserver 1156521117 M * derjohn2 harry, odd. try to add 'dummy' ? 1156521140 M * daniel_hozac is it just me or is http://list.linux-vserver.org down? 1156521184 M * harry derjohn2: where? 1156521197 M * derjohn2 harry 'modprobe dummy' 1156521229 M * harry boromir:~# ip route add default via 134.58.243.254 dev dummy0 table 134.58.243-net 1156521233 M * harry RTNETLINK answers: Network is unreachable 1156521237 M * derjohn2 harry, actually i didnt use dummy every, but IMO it thought for that kind of case. usually if you have routes with a ppp-dialup that is nit always there. 1156521238 M * harry it stays the same... 1156521248 M * harry you need an ip on that network to route to it 1156521258 M * derjohn2 harry, you have to add the ip to the dummy0 1156521268 M * harry what ip? the public one? 1156521290 M * derjohn2 well, I _think_ you use the dumm0 _instead_ the eth0 1156521310 M * harry let's check it out :) 1156521335 M * harry routing is ok now... 1156521336 M * derjohn2 i dunno howe to say that dummy0 is 'bound' to ethXY. googleing for dummy + linux +howto doesnt help much ;) 1156521339 M * harry but i can't use the ip 1156521347 M * derjohn2 ip link dumm0 up ? 1156521358 M * harry derjohn2: dummy0 is up 1156521371 M * Bertl derjohn2: you cannot 'bind' one interface to another? 1156521372 M * harry but the p needs to be up on eth0, not a dummy if 1156521397 M * Bertl dummy0 is an interface like eth0. period. 1156521431 M * derjohn2 harry, to Bertl words listen ;) 1156521443 M * Bertl the only difference is, while eth0 has some hardware attached which sends off the packet, and receives stuff, dummy0 just discards the packets :) 1156521479 M * harry yeah, i know that ): 1156521487 M * derjohn2 hm, so .... well ... that doesnt really help with the nexttop problem .. 1156521509 M * Bertl http://archives.linux-vserver.org/200311/0470.html 1156521512 M * derjohn2 harry, or you would have to build a bridge from dummy0 to eth0 1156521531 M * daniel_hozac Bertl: delta-vci-bump01 and delta-net-fix02 are the only changes to -rc30, right? 1156521539 M * derjohn2 harry,with 'brctl' 1156521553 M * derjohn2 then dummy0 should be always up and behave like eth0 ... 1156521555 M * harry that doesn't change a thing! 1156521564 M * derjohn2 no? 1156521572 M * harry i still need to set an ip on dummy0 then! 1156521589 M * daniel_hozac Bertl: or are the persistent and helper fixes included as well? 1156521594 M * harry i could aswell set the ip fixed on the host, and route it there 1156521599 M * harry and let the vps use the "dummy" 1156521620 M * derjohn2 harry, yes, you set the /30 to the dummy0 then it always up and you can fire up the route there 1156521640 M * derjohn2 harry, no matter if eth0 is up or down 1156521672 M * derjohn2 harry, but: all pure theory :) If you figure out, let us know ;) 1156521687 M * harry derjohn2: i can't set a route over eth0 when the network is defined on dummy0 1156521704 M * derjohn2 yes, you route over dummy0 1156521705 M * Bertl http://archives.linux-vserver.org/200311/0470.html 1156521709 M * harry all i need, is to give a vps a route , depending on the ip and interface 1156521718 M * derjohn2 and then build a bridge from dummy0 to eth0 1156521718 M * harry i kinda hoped that it could be done in vserver 1156521729 M * derjohn2 harry, or read Bertl's link ,) 1156521731 M * harry but you have to do it yourself by using "freaky" routing stuff 1156521743 M * derjohn2 harry, you setup is simply freaky ;) 1156521749 M * harry derjohn2: i konw that link... i'e done it a million times! 1156521752 M * harry derjohn2: not at all! 1156521765 M * derjohn2 harry, I would do the mgmt net with a VLAN dot1q anyway ;) 1156521773 M * harry what does everyone want: 1 host, running 20 vps'es 1156521798 M * harry the host has to be managed: so: host only reachable through management network 1156521810 M * harry all the vps'es are on different networks, using different ips etc... 1156521812 M * derjohn2 yup, via a vlan interfaces 1156521824 M * derjohn2 yes, no problem. 1156521826 M * harry mkay, whatever... suits me fine 1156521832 M * harry how is that done? 1156521852 M * derjohn2 set a static route like ip a add 192.168.0.0/16 and set the default to the public ip 1156521867 M * harry there is no public ip for the host 1156521871 M * harry why would you want that??? 1156521878 M * harry there is a public ip for the vps'es 1156521884 M * harry (1 for every vps) 1156521896 M * derjohn2 well, so make the 192er accessable from anywhere in your company? 1156521899 M * harry on its own interface/vlan/whatever... i don't care" 1156521905 M * harry it is 1156521911 M * harry i can reach it 1156521918 M * derjohn2 well, the 1st VPS could also share the Ip the host 1156521934 M * derjohn2 er, forget it :) 1156521940 M * harry ahm... uhu :) 1156521972 M * harry host: private network, not routed or routed for management stuff, but not public 1156521984 M * derjohn2 or? yes, 1st guest uses "publicip/24". thats also the ip of the host. additional guests use /32 ips from _the same net_ 1156521988 M * harry vps: public ip address on a different interface/vlan/... with it's route over that network 1156521990 M * nayco_work Strange : installing utils-vserver did not set the barrier attribute... What was the command, please ? 1156521992 M * harry is that so hard to ask? 1156521995 M * derjohn2 1st guest uses "nodev" 1156522000 M * harry so much to ask/so hard to get? 1156522028 M * derjohn2 well, the host simply get a listen 192.x.x.x in sshd_config 1156522032 M * harry derjohn2: uhu... first guest goes down... now what? 1156522047 M * derjohn2 harry, nodev .. you know that option? 1156522050 M * derjohn2 *option? 1156522056 M * harry yes 1156522072 M * harry ah... wait.. you mean... set that ip fixed on the host 1156522074 M * derjohn2 yes, then the 1st guest can go down and the host ip stays up 1156522080 M * derjohn2 yup 1156522084 M * harry and leave the ip online, even if the vps is down? 1156522094 M * harry so: server is down, but ip is up... 1156522100 M * harry weirdness++!!!!! 1156522120 M * harry so basicly: you say: share an ip between the host and 1 vps 1156522137 M * harry again: tricky business!!!!!! 1156522147 M * derjohn2 harry, well, you could also try to find out how to do a 'paket reinject/redirect' with IP tables 1156522161 M * derjohn2 harry, I use the shared IP way 1156522164 M * harry why not make it possible to set an ip, netmask AND ROUTE in the config of your vserver? 1156522187 M * derjohn2 harry, maybe that can be set in sysctl foo 1156522191 M * harry i mean... you can set ip, interface, netmask... but no route 1156522194 M * derjohn2 harry, i simply dont know 1156522220 M * nayco_work Oh, my mistake, the attribute is here, wrong command (lsattr instead of showattr 1156522223 M * nayco_work ) 1156522234 M * derjohn2 harry, the behavior can maybe be changed 1156522249 M * derjohn2 harry, did you also consider a hostroute? 1156522281 M * derjohn2 from the router a 'ip ro add publicvpsa/24 via 192.blaa' ? 1156522305 M * harry ? 1156522316 M * harry so use the management network to route all the traffic??? 1156522322 M * derjohn2 use /32 Ip for _every_ guest 1156522327 M * derjohn2 yes 1156522338 M * harry like... our 100Mbit management network should replace the 3 1Gbit lines ? 1156522340 M * derjohn2 but: the sending IP is always the one of the guest 1156522358 M * derjohn2 huh? 1156522373 M * harry eth0, eth2 and eth3 are Gbit lines 1156522378 M * harry eth1 (management) is 100Mbit 1156522380 M * derjohn2 you have differnt interfaces? 1156522397 M * derjohn2 ah, didnt mention that (to me) 1156522404 M * harry ahm... 1156522419 M * derjohn2 then you need some transfer net at all. 1156522424 M * harry that doesn't change anything 1156522450 M * derjohn2 you can use a 192.169.xx between host and router 1156522458 Q * shedi Quit: Leaving 1156522476 M * harry ??? 1156522482 Q * balbir Ping timeout: 480 seconds 1156522504 M * derjohn2 well, you have a dedicated interface foir the 'public traffic' ? 1156522512 M * harry yes 1156522520 M * harry eth1: management 1156522522 M * harry eth0: public 1156522527 M * harry eth2/3: not defined yet 1156522536 J * bonbons ~bonbons@83.222.36.236 1156522538 M * harry so: host: eth1: management 1156522544 M * derjohn2 and why is eth0 down? 1156522548 M * harry vps: eth0: public ip with route 1156522554 M * harry because there is no ip bound to it 1156522561 M * harry and: it isn't down 1156522578 M * derjohn2 and default route is to ... ? 1156522605 M * harry management off course 1156522614 M * harry there is no ip on eth0 unless the vps is started 1156522629 M * daniel_hozac harry: specifying routes in the config wouldn't make sense at all, since all the networking happens on the host. 1156522653 M * harry daniel_hozac: imho my setup isnt very "odd" 1156522661 M * harry management interface for the host 1156522668 M * harry vps'es have their ip's and routes 1156522678 M * harry now, they have ip's but no routes 1156522692 M * harry mkay, routing is done by the host now... but that sucks! 1156522722 M * harry there is no "good" sollution i'm afraid 1156522738 M * derjohn2 harry, you might ask on #quagga if there can be a route with an unreachable nexthop 1156522740 M * daniel_hozac you don't think it's odd to have a bunch of IP addresses on different networks? 1156522761 M * daniel_hozac i'm guessing most people assign larger chunks of their address space to each host. 1156522763 M * harry daniel_hozac: in a virtualisation environment? no 1156522787 M * harry you want a webserver in a vps (on a webserver dmz) 1156522794 M * harry a db server in the db-dmz 1156522802 M * harry a ftp server in a ftpserver-dmz 1156522812 M * harry and a management interface for the host in the management dmz 1156522824 M * harry all different interfaces on different vlans 1156522829 M * harry virtualized on 1 host 1156522864 M * daniel_hozac would make more sense to have all the FTP servers on one host, all the databases on another and all the webservers on a third, no? 1156522894 M * harry no, all on 1 host 1156522908 M * harry or... maybe you're right 1156522913 M * Bertl :) 1156522915 M * harry so you say: all the management on 1 host 1156522919 M * harry all the webservers on 1 host 1156522926 M * harry but... how will you manage the webservers??????? 1156522988 M * daniel_hozac you have management guests? 1156522995 M * harry no 1156522998 M * harry i want to manage servers 1156523004 M * harry that's why there is a management dmz 1156523027 M * harry i want to set up webservers on vps'es, running on servers 1156523035 M * harry they should be in a webserver dmz 1156523052 M * harry so... routing the webserver traffic over network ... which is all the vps traffic 1156523059 J * balbir ~balbir@59.144.51.222 1156523070 M * harry and route the management stuff over the management stuff... (the host/the server) 1156523086 M * harry pretty logic thinking imho 1156523113 M * derjohn2 harry, the question is : hey, if there a way to add a static route (on linux 2.6) to a ip that is unreachable? (i.e. add the route before the interface goes up) 1156523136 M * harry derjohn2: my question is different 1156523143 Q * ntrs Remote host closed the connection 1156523149 M * harry what does a server need to get networking done: 1156523155 M * harry interface, ip, netmask, gateway 1156523168 M * harry what can you set in vserver config: interface, ip, netmask 1156523182 M * harry if i have a vps... i want to give that one route too!!! 1156523185 M * daniel_hozac because gateway doesn't make sense. 1156523186 M * nayco_work Bertl: This is sooo strange... er.... Pffff... I started my new vserver, followed the procedure to add quotas to / (/dev/hdv1) inside the guest, 1156523198 M * harry daniel_hozac: tell me how a gateway doesn't make sense??? 1156523202 M * nayco_work and it half works.... 1156523204 M * derjohn2 harry, in which host would you set the gateway? 1156523204 M * harry my vps has 1 ip 1156523206 M * daniel_hozac harry: because networking happens on the host? 1156523210 M * harry on 1 network 1156523215 M * derjohn2 i mean: in which guest 1156523221 M * harry why doesn't it make sense to tell him what's the way out! 1156523221 M * Bertl nayco_work: what is the half which doesn't work? 1156523232 M * harry daniel_hozac: network doesn't happen on the host!!!! 1156523238 M * daniel_hozac ... yes it does. 1156523239 M * harry it happens in the vps 1156523243 M * daniel_hozac no, it doesn't. 1156523246 M * harry the vps gets an ip 1156523248 M * daniel_hozac vserver does IP isolation. 1156523252 M * harry the vps should get a route 1156523252 M * daniel_hozac nothing more. 1156523257 M * daniel_hozac the host handles all of the networking. 1156523264 M * harry okay 1156523267 M * daniel_hozac vserver 101. 1156523267 M * harry suits me! 1156523284 M * harry then give me a way to set the route for a vps 1156523296 M * harry i can't setup a route to a network we're not connected to 1156523302 M * harry if i start a vps, vserver sets the ip 1156523311 M * harry why not let vserver set a route for that ip? 1156523315 M * harry /network 1156523316 M * derjohn2 daniel_hozac, he has an extra ethX for the vservers. and wants to bring that up with the 1st VPS stared, as he doesnt want to wate one ip for that interface. 1156523342 M * derjohn2 daniel_hozac, the ssh is bound to ethY ( X != Y ) 1156523361 M * harry i agree networking is done on the host 1156523379 M * harry but vserver luditdb start SETS an ip on eth0 1156523382 M * daniel_hozac harry: because that's such a weird use-case it doesn't make sense to implement it. feel free though. 1156523385 M * harry it sets the netmask 1156523392 M * harry it should also set the route for that netwrok 1156523394 M * Bertl derjohn2: not releavant, just enable secondary propagation and leave it unconfigured 1156523394 M * daniel_hozac you could just use a pre-start script though. 1156523409 M * derjohn2 Bertl, and what about the routes ? 1156523410 M * harry daniel_hozac: how will that ever work? 1156523416 M * harry there is no ip defined then 1156523426 M * daniel_hozac harry: why wouldn't there be? 1156523432 M * Bertl derjohn2: as I already said several time, multiple routing tables with _rules_ 1156523434 M * harry why WOULD there be? 1156523437 M * daniel_hozac you seem to misunderstand pre*start. 1156523441 M * harry eth0 is down now... there is no ip bound to it 1156523448 M * daniel_hozac vserver adds it, no? 1156523452 M * harry yes 1156523457 M * daniel_hozac so how doesn't pre-start work? 1156523458 M * harry when does pre-start happen? 1156523464 M * harry AFTER the ip is bound to eth0? 1156523466 M * derjohn2 Bertl, doesnt work. we tried. it says as long as the interface isnt up "unreachabe nexthop" 1156523467 M * daniel_hozac right before starting, as the name suggests. 1156523471 M * harry or is it a PRE-start? 1156523475 M * nayco_work Bertl: I mean, repquota -a works, repquota / fails, repquota /dev/hdv1 works (interesting ;-)), 1156523478 M * derjohn2 Bertl, even with a oen rt_table 1156523492 M * harry if it's started AFTER the vps has started, it's too late, since some services require networking to be up 1156523499 M * nayco_work setquota / fails, setquota /dev/hdv1 is silent but does nothing.... 1156523506 M * Bertl nayco_work: what is /dev/hdv1 and what does /etc/mtab list inside your guest 1156523510 M * daniel_hozac harry: you might want to read the descriptions for the scriptlets. 1156523523 M * harry daniel_hozac: url? 1156523528 M * nayco_work and in fact, I given quotas to a user, and no limit is set. 1156523531 M * harry flowerpage/ 1156523533 M * daniel_hozac http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1156523557 M * nayco_work Bertl: Ok, I gonna paste ;-) 1156523577 M * harry * pre-start The scriptlet which will be executed after network-interfaces were enabled and the directories mounted, but before the vserver itself has been started. Before executing the script, the vserver root directory will be made the working directory. 1156523581 M * harry hahaaaaaaaaaaa 1156523584 M * harry superduper! 1156523590 M * derjohn2 harry, you might try the onlink flag: onlink pretend that the nexthop is directly attached to this link, even if it does not match any interface prefix. 1156523590 M * harry i can setup filtering there!!! 1156523605 M * harry advanced routing per server 1156523610 M * harry that's doable 1156523633 M * harry tough i still think that there should be an easy way to set a gateway for each vps... let vserver do the routing magic 1156523643 A * derjohn2 now off for some BBQ 1156523666 M * harry that way, you could set your own routing stuff in each vps for even more advanced stuff 1156523670 M * daniel_hozac harry: so what's stopping you? 1156523679 M * harry daniel_hozac: i haven't got a clue where to start 1156523679 M * harry ;) 1156523683 M * harry and... it's weekend 1156523693 M * harry and... i have to leave at 19u10 at home 1156523696 M * harry i'm still at work 1156523704 M * harry need to drive home, get lunch, eat it, take a shower 1156523710 M * harry so.... i'm gonna be late :) 1156523712 M * Bertl derjohn2, harry: with CONFIG_IP_ROUTE_MULTIPATH enabled? 1156523725 M * harry Bertl: ? 1156523734 M * Bertl the kernel config, is it on? 1156523735 M * harry Bertl: is that the advanced routing option? 1156523746 M * Bertl no 1156523785 M * harry why would i want route multipathing btw??? 1156523804 M * Bertl to allow for several routes 1156523814 M * Bertl is it off, just curious? 1156523826 M * harry it's not set 1156523830 M * harry i can't find it in the config 1156523834 M * Bertl okay, tx 1156523852 M * harry i don't know what it does 1156523855 M * harry could that help me? 1156523881 M * Bertl depends on what you plan to do ... 1156523890 M * harry i'll check out the docs, what it does 1156523896 M * harry but now... i REALLY have to run! 1156523896 M * harry :)( 1156523901 M * harry cya'll later! 1156523901 M * Bertl okay, cya 1156523906 M * harry and tnx for the info 1156523915 M * harry (and sry i'm such an asshole, it's just... i think i'm right ;)) 1156523932 M * harry there should be a way that vps'es can set their gateways 1156523951 M * Bertl with virtualized networking that will make sense :) 1156523959 M * daniel_hozac without it, not so much. 1156524241 M * Bertl we probably should add the persistant fixes (especially the unhash) to stable too 1156524264 M * Bertl not sure we want to mess with the helper return codes though 1156524291 M * nayco_work Bertl: Finished : http://paste.linux-vserver.org/290 1156524298 M * Bertl tx 1156524333 M * daniel_hozac yeah, i was thinking the same thing. 1156524383 M * Bertl nayco_work: try the following things (in this order and one by one): 1156524394 M * Bertl - change the mtab entry from ufs to xfs 1156524400 J * stefani ~stefani@tsipoor.banerian.org 1156524408 M * Bertl - add an fstab with an entry for / with ufs/xfs 1156524449 M * Bertl also I still would like to know what ls -la /dev/hdv* gives 1156524462 M * nayco_work Inside the guets ? 1156524466 M * Bertl yep 1156525122 A * Belu is away (iŽll be back later...) 1156525122 N * Belu Belu_zZz 1156525585 J * ekc_away ~ekc@oxnard-cuda1-cable-bundle-68-69-253-79.ventca.adelphia.net 1156525595 N * ekc_away ekc 1156525644 J * cehteh ~ct@cehteh.homeunix.org 1156526069 M * nayco_work Bertl: Is the paste mirrored, saved, archived, or crawled by search engines ? 1156526092 M * daniel_hozac it should be archived. 1156526093 M * Bertl no, yes, yes, probably? 1156526171 M * nayco_work Ok, because this may be interesting for knowledge... Maybe not ;-). Ok, I'm nearly happy :D : 1156526174 M * nayco_work http://paste.linux-vserver.org/294 1156526179 M * nayco_work \o/ 1156526237 M * mnemoc anyone knows a tool to 'apply differences' between two iptables-save ? 1156526248 M * Bertl the repquota / is probably fixed by an fstab entry 1156526261 M * cehteh mnemoc: diff3? 1156526287 M * cehteh well guess not ;) .. at least you need some fixup scripting afterwards 1156526329 M * mnemoc you need to calculate where to insert, which chains to create, which to delete, etc.. 1156526337 M * mnemoc it's not that trivial :p 1156526486 M * nayco_work Bertl: I'm trying fstab... 1156526548 M * Bertl daniel_hozac: do you know if c99 stuff is considered okay for the kernel by now? 1156526573 M * daniel_hozac we already have C99 initializers for the proc structs, no? 1156526586 M * Bertl yeah, I was more referring to things like: 1156526592 M * Bertl for (int cnt = 0; ... 1156526600 M * daniel_hozac oh, i have no idea. 1156526641 M * daniel_hozac but since we're already using other C99 constructs, i don't see why not. 1156526663 M * Bertl just scanning for an example : 1156526667 M * Bertl +) 1156526824 M * Bertl okay, we have at least 5 places where this is already done :) 1156526938 M * Bertl it's actually something I consider very appropriate, as it limits the scope and adds readability IMHO 1156526942 M * nayco_work Bertl: No way : http://paste.linux-vserver.org/295 1156527044 M * coocoon nayco_work: which guest are u using 1156527113 M * Bertl nayco_work: could you strace -fF the 'repquota /' ? 1156527122 Q * cdrx Ping timeout: 480 seconds 1156527393 M * mnemoc Bertl: i have a problem (using 2.0.2-rc29) which i have never seen before.... 'syslog' on the guest can't bind to /dev/log... because it's 'already in use' 1156527410 M * mnemoc (in fact it's socklog, but for the case is the same) 1156527451 M * nayco_work coocoon: Er... Which guest ? What do you mean ? The one in the vserver host I just installed (The one on my dual-opteron) 1156527456 M * nayco_work Bertl: yes 1156527509 M * Bertl okay, then let's do that :) 1156527535 M * nayco_work Bertl: http://paste.linux-vserver.org/296 1156527586 M * coocoon nayco_work: this guest i mean /etc/vservers/quota/ 1156527647 Q * stefani Quit: I Quit! 1156527647 J * xtz ~xtz@sc.xtz.ch 1156527657 P * xtz 1156527661 M * Bertl nayco_work: it seems like a quota tools internal decision, no idea what causes it to bail out 1156527734 M * Bertl maybe it compares devices or such, and figures that / is a different device than /dev/hdv1 1156527844 M * nayco_work Ok 1156527863 M * coocoon nayco_work: Bertl: sorry but maybe it is very simple and easy, because of fstab will be overwritten ????? 1156527902 M * phedny stat("/proc/fs/xfs/stat", 0x7fff9fe5c470) = -1 ENOENT (No such file or directory) 1156527906 M * nayco_work Now I'm trying to do the same with /home.... I don't know if I must bind /dev/vroot0 with /dev/hdv2 ? 1156527911 M * phedny maybe this file is required for xfs quota to work? 1156527926 M * nayco_work phedny: Oh, vprocunhide ? 1156527933 M * phedny just a wild guess 1156527935 M * Bertl phedny: doesn't seem so, it seems to work if you specify the device, no? 1156527950 M * phedny i've xfs quota working on a non-vserver machine, never tried in vserver 1156528153 M * phedny hmm, although repquota doesn't seem to work on it either 1156528198 M * phedny but that seems to be because I have group quota, not user quota 1156528235 M * Bertl nayco_work: this rather intrusive change makes xfs quota work with devel (on the host) for me: http://vserver.13thfloor.at/Experimental/delta-xfsquot-fix01.diff 1156528255 M * Bertl (with 'normal' quota disabled, that is, now testing with quota enabled) 1156528289 M * phedny when I strace it (repquota -g) this point is different: 1156528290 M * phedny stat64("/proc/fs/xfs/stat", {st_mode=S_IFREG|0444, st_size=0, ...}) = 0 1156528290 M * phedny stat64("/proc/sys/fs/quota", {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0 1156528612 M * Bertl daniel_hozac: hmm, seems that there are still checks left: fs/quota.c:75: error: `for' loop initial declaration used outside C99 mode 1156528616 M * nayco_work Bertl: ? er.... "makes xfs quota work with devel (on the host)" _On_the_host_ ? But it works here... On my opteron, not on the other. I mean, I thought it was me that forgot something... 1156528630 M * nayco_work You mean inside the guest ? 1156528632 M * Bertl nayco_work: you switched to stable 1156528667 M * nayco_work ??? 1156528678 M * nayco_work Ohhh :D !!! 1156528688 M * nayco_work I did not even see. It's late :D ! 1156528698 M * nayco_work 2.0.2 1156528715 M * nayco_work So, I go to the other host, apply your pathc, and try agian ? 1156528738 M * Bertl better wait a few minutes for an additional patch 1156528819 M * nayco_work Ok, I already applied this one. 1156528820 M * mnemoc ah, lovely.... if /dev is ro /dev/log doesn't work "as expected" :'( 1156528844 M * nayco_work Gotta smoke (___(__________#~ 1156528892 J * shedi ~siggi@inferno.lhi.is 1156528933 M * mnemoc Bertl: where is the mail archive now? 1156528968 M * Bertl according to http://linux-vserver.org/ 1156528975 M * Bertl - http://list.linux-vserver.org/archive/vserver/ 1156528981 M * Bertl - http://archives.linux-vserver.org/ 1156529013 M * Bertl the former seems to refuse connects right now 1156529062 M * nayco_work Bertl: How many time do I have before patch (estimating the length of my cigarette ;-)) 1156529065 M * nayco_work ? 1156529090 M * Bertl probably 30-40 cm, test compiling now :) 1156529096 M * nayco_work :D !!!! 1156529173 M * mnemoc Bertl: there is any security benefit if /vservers/ is nodev when the contexts are not mknod capables? 1156529228 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1156529229 M * Bertl well, it avoids cases where nodes are hidden/forgotten somewhere 1156529295 M * mnemoc that was the 'idea' pushing me to do that... but then i need a dev capable /dev on a different partition an mounted on each guest 1156529314 M * Bertl could be a shared tmpfs? 1156529328 M * mnemoc bind ? 1156529330 M * Bertl read only, but with dev 1156529348 M * mnemoc but if it's read ony /dev/log doesn't work :\ 1156529370 M * mnemoc one tmpfs (rw) per guest? uhm.... 1156529481 J * kuberkoos ~Miranda@wbs-41-208-193-179.wbs.co.za 1156529483 A * mnemoc puzzled 1156529742 J * stefani ~stefani@tsipoor.banerian.org 1156529817 A * mnemoc going for the per-guest tmpfs rw,dev capable populated on pre-start 1156529901 Q * kuberkoos Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1156529908 J * dna_ ~naucki@31-243-dsl.kielnet.net 1156529924 M * Hollow Bertl: just as a reminder.. we still need vc_rlimit_stat and extension of rlimit_mask for accounting values 1156529971 M * Bertl http://vserver.13thfloor.at/Experimental/delta-vstat-feat03.diff 1156529992 M * Hollow uhm.. i'm sorry 1156529993 M * Bertl the rlimit_mask extension is not there yet, but will be soon, tx 1156530000 M * Hollow seems like i missed it 1156530015 M * Bertl np 1156530044 M * Bertl .o( of all the things I lost ... I miss my mind the most :) 1156530076 M * Hollow heh 1156530086 M * nayco_work Bertl: back 1156530093 M * Bertl wb nayco_work! 1156530109 M * Hollow btw, i updated the manual a lot, the xmlrpc reference is now complete, and also the vcd architecture is described in detail... 1156530131 M * Hollow already ~40 written pages 1156530142 J * gerrit ~gerrit@bi01p1.co.us.ibm.com 1156530170 M * Bertl great! as you remind me of docu, how do we (plan?) to keep updates to the old wiki in sync with the old pages on the new wiki, if at all? 1156530178 M * Bertl wb gerrit! 1156530189 M * gerrit hiBertl! 1156530228 M * Hollow imo the best would be to allow write access to the old wiki for some time, until more pages have been migrated.. 1156530254 M * Bertl okay, so we do some kind of 'resync' and switch over? 1156530267 M * Hollow i'd say so 1156530283 M * Bertl okay, will look into the final color stuff this weekend 1156530302 Q * dna Ping timeout: 480 seconds 1156530318 M * Hollow yup, we should sync with each other regarding the announce etc then... 1156530331 M * Bertl nayco_work: http://vserver.13thfloor.at/Experimental/delta-xfsquot-fix02.diff 1156530337 M * Bertl (ontop of the fix01) 1156530400 M * Hollow btw, didn't you plan to rewrite (parts of) the paper? 1156530469 M * Bertl yes, it is still on my todo .. had some issues with moving my mail server and such 1156530562 M * Hollow sure.. i used parts of your paper and also extended it a bit in the manual (and also in the wiki to some degree) so maybe you take a look at it once you start rewriting it.. 1156530576 M * Hollow to keep the info in sync 1156530586 M * nayco_work Bertl: Ok, I can compile, now ? 1156530611 M * Bertl Hollow: excellent, tx for the info! btw, is it already linked somewhere? 1156530631 M * Bertl nayco_work: yes, let's give it a try! :) 1156530633 M * nayco_work eeweego ! 1156530640 M * Hollow Bertl: yep, http://home.xnull.de/work/vserver/manual/manual.pdf 1156530642 J * mkhl mkhl@200-153-181-145.dsl.telesp.net.br 1156530659 M * Bertl Hollow: yes, I meant from the wiki(s)? 1156530672 M * Hollow ah, no.. not yet... but will do so soon... 1156530680 M * Hollow it's just three days old 1156530688 M * Hollow some parts are still missing 1156530705 M * Bertl that's why I didn't find it with google (the missing links :) 1156530711 M * Hollow :) 1156530735 M * Hollow will do all the wiki stuff we discussed this weekend 1156530855 M * Bertl btw: http://www.linux.com/article.pl?sid=06/06/15/1444226 1156530913 M * matti Hollow: Nice manual. Thanks. 1156530915 M * Hollow nice.. this part e.g. is still missing :) 1156530922 M * Hollow matti: yw :) 1156530932 M * daniel_hozac Bertl: at least i have seen that before. 1156530938 M * Hollow please bug me if anything is unclear 1156531209 Q * cehteh Remote host closed the connection 1156531499 M * Bertl daniel_hozac: did you make/provide a patch to restructure the out_unhash logic as you suggested? 1156531519 M * daniel_hozac hmm, i think that was doener. 1156531530 M * Bertl ah, okay, sorry then ... 1156531554 J * cehteh ~ct@cehteh.homeunix.org 1156531777 M * Bertl could somebody bounce me the email titled 'Bug in vc_net_remove ..' from Cristopher? (only one please :) 1156531805 A * Hollow goes sending one to be sure.. 1156531807 M * Hollow ;) 1156531910 Q * comfrey Quit: Lost terminal 1156531977 M * Bertl Hollow: tx! 1156532009 M * nayco_work Bertl: The f***** kernel is compiling entirely.... It'll be long on this machine :-( 1156532024 A * Hollow pats kernel 1156532209 M * nayco_work :D 1156532239 M * matti Heh. 1156532288 A * Wonka compiles kernels in under 30min... 1156532382 M * matti On smp this can even take less than 10 min. 1156532498 M * nayco_work On my dual-opteron, 20 minutes from scratch, bzImage, modules and modules_install. But on this PIII 1133 MHz, it take more than an hour :-( 1156532526 M * Bertl nayco_work: you must have a lot of hardware in your machine :) 1156532590 M * nayco_work Bertl: ahemmm... *whistles* 1156532624 M * nayco_work Well, this is a Mandriva .config, and I do not have the knowledge to remove the drivers I do not need. 1156532657 M * nayco_work I tried once, and the machine did not work anymore... So, well ;-) 1156533348 J * shuri ~shuri@cpu183.adsl.qc.bellglobal.com 1156533373 M * shuri hello all 1156533394 M * Bertl welcome shuri! 1156533402 M * shuri :) 1156533408 M * shuri the master is here! 1156533428 M * shuri how are you6 1156533429 M * shuri ? 1156533454 M * Bertl I'm fine, tx, and you? 1156533495 M * shuri 50/50 1156533496 M * shuri hehe 1156533514 M * shuri i got a quesiton for you 1156533533 M * shuri is there a way to change the space allocated for the /tmp inside a vserver? 1156533540 M * shuri none 16384 3272 13112 20% /tmp 1156533542 M * Bertl yes 1156533562 M * shuri how? 1156533571 M * shuri conf? 1156533576 M * Bertl simply adjust the fstab in the config 1156533611 M * shuri :) 1156533639 M * shuri thx you! 1156533646 M * Bertl you're welcome! 1156533648 Q * lilalinux Remote host closed the connection 1156533956 M * nayco_work *reboot* 1156534236 Q * Viper0482 Remote host closed the connection 1156534955 M * nayco_work Bertl: It does not work :-( And I got to go home.... 1156535202 Q * cdrx Quit: Leaving 1156535274 M * Bertl nayco_work: okay, we'll continue another time 1156535556 Q * ptl_xid 1156536412 M * mnemoc Bertl: if /dev is tmpfs i can't mkdir pts between /dev is mounted and /dev/pts :\ 1156536427 M * daniel_hozac mnemoc: how come? 1156536457 M * mnemoc daniel_hozac: what do you suggest? 1156536472 M * daniel_hozac how are you mounting it? fstab? 1156536477 M * mnemoc yes 1156536490 M * daniel_hozac ah, i guess you are screwed then. 1156536516 M * mnemoc is it ok to mount /dev by hand at prepre-start ? 1156536535 M * daniel_hozac i think it should be. 1156536662 A * mnemoc wonders if having /vservers as nodev is really a good idea 1156537364 M * mnemoc ok, the 'elegant' solution was to mount /dev on fstab, populate on pre-start and mount /dev/pts on post-start :) 1156537970 M * daniel_hozac heh. 1156539848 M * Bertl okay, off for now ... back later! 1156539862 M * mnemoc gn8 Bertl 1156539868 N * Bertl Bertl_oO 1156540488 Q * mkhl 1156540559 J * Hecticee ~dman@host81-129-189-119.range81-129.btcentralplus.com 1156540867 J * nayco ~nayco@lns-bzn-59-82-252-179-122.adsl.proxad.net 1156540951 M * nayco hello again, all ! 1156540985 M * Hecticee hello, from where i can get patches to 2.6.17.11 sources 1156541004 M * coocoon http://vserver.13thfloor.at/Experimental/ 1156541029 M * derjohn2 Hecticee, you might need to patch wit fuzz 50" 1156541036 M * derjohn2 *with 1156541046 M * Hecticee fuzz 50? 1156541060 M * derjohn2 patch -p1 -F 50 -i pacthfile 1156541141 M * daniel_hozac there are already 2.6.17.11 patches. 1156541171 M * derjohn2 daniel_hozac, Bertl_oO : why is there a rc30 for stable? 1156541187 M * daniel_hozac PLM, mostly. 1156541192 M * daniel_hozac i think anyway. 1156541197 M * derjohn2 daniel_hozac, Bertl_oO : it's not in the topic 1156541203 M * derjohn2 PLM? 1156541209 M * daniel_hozac because 2.1.1-rc30 isn't released yet ;) 1156541225 M * daniel_hozac patch lifecycle manager(?) 1156541228 M * derjohn2 well,why is it then in the dir ? 1156541235 M * daniel_hozac http://plm.testing.osdl.org/patches/show/5228 1156541239 M * daniel_hozac hmm? 1156541239 M * derjohn2 there is patch-2.6.17.11-vs2.0.2-rc30.diff 1156541245 M * daniel_hozac yes, but not 2.1.1-rc30. 1156541274 M * derjohn2 daniel_hozac, yes, but it's also not on topic. to i'll wait until my next complile run ;) 1156541280 M * Hecticee so which patches i should use? 1156541350 M * daniel_hozac derjohn2: topic says 2.{0.2,1.1}-rc29. 2.1.1-rc30 would have to be released before the topic can be updated. 1156541356 M * daniel_hozac Hecticee: depends on if you want stable or devel. 1156541444 M * derjohn2 daniel_hozac, eh, as you say :) So the topic here is the refernce source for the current rc ? 1156541512 M * daniel_hozac derjohn2: no, the dir listing is the reference source for the current rc :) 1156541785 M * derjohn2 daniel_hozac, I would assume that, as the file is there and downloadable. if the file can change Bertl_oO should mark is as rc30-candicate1 :) 1156541793 M * derjohn2 *candidate 1156541810 M * daniel_hozac the files do change though :) 1156541817 M * derjohn2 daniel_hozac, really? 1156541818 M * daniel_hozac not too often, but it happens. 1156541832 M * daniel_hozac for the brown bag releases ;) 1156542829 M * nayco *compiling 2.6.11.8* 1156542966 M * daniel_hozac really? 1156542973 M * daniel_hozac that's _ancient_. :) 1156542986 M * mnemoc and it includes 2 CVEs :p 1156543087 J * lolilol hiddenserv@tor.noreply.org 1156543142 Q * olilo Remote host closed the connection 1156543774 Q * Hecticee Read error: Connection reset by peer 1156544226 J * Hecticee ~dman@host81-129-189-119.range81-129.btcentralplus.com 1156544519 Q * click oxygen.oftc.net strange.oftc.net 1156544519 Q * glut oxygen.oftc.net strange.oftc.net 1156544519 Q * Nam oxygen.oftc.net strange.oftc.net 1156544519 Q * Johnnie oxygen.oftc.net strange.oftc.net 1156544519 Q * Greek0 oxygen.oftc.net strange.oftc.net 1156544519 Q * mugwump oxygen.oftc.net strange.oftc.net 1156544519 Q * micah oxygen.oftc.net strange.oftc.net 1156544519 Q * tanjix oxygen.oftc.net strange.oftc.net 1156544635 Q * dna_ Quit: Verlassend 1156544653 N * Bertl_oO Bertl 1156544745 J * click click@ti511110a080-2980.bb.online.no 1156544745 J * glut glut@no.suid.pl 1156544745 J * Nam ~nam@S0106001195551ff0.va.shawcable.net 1156544745 J * Johnnie ~jdlewis@static-acs-24-154-32-33.zoominternet.net 1156544745 J * Greek0 ~greek0@85.255.145.201 1156544745 J * micah ~micah@micah.riseup.net 1156544745 J * tanjix ~tanjix@office.star-hosting.de 1156544745 J * mugwump ~samv@watts.utsl.gen.nz 1156545040 Q * Hecticee 1156545181 J * someoneonet ~someoneon@80.178.70.181.adsl.012.net.il 1156545193 M * Bertl welcome someoneonet! 1156545303 M * someoneonet Hi everyone, I've started using VServer recently and I must say it's really neat. 1156545319 M * Bertl glad to hear! 1156545404 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1156545474 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1156545487 Q * mire Quit: Leaving 1156545563 M * someoneonet I've notices something strange and I wonder whether it's my config mistake or just a VSServer limitation: when I satrt a service (lets say Exim) on a VServer,I can connect to it from another computer, but when I try to connect to another service (like SSH), i'm being connected to the host SSH Server 1156545608 M * Bertl that's not an unusual issue, your host is running an _unlimited_ sshd, which grabs all ips 1156545624 M * Bertl the guest sshd cannot bind and bails out 1156545652 M * Bertl just use some Listen directive to restrict the host's sshd and everything should be fine 1156545713 M * someoneonet Oh... I see, thanks :) Hmm... does VServer support dynamiccaly allocated IP (DHCP) for VServers? 1156545750 M * Bertl you can do dhcp requests (on the host) with some 'guest' identifier, and use that to (re)configure a guest, yes 1156545832 M * someoneonet Hmm.. great. thank you a lot :) 1156545847 M * Bertl you're welcome! feel free to hang around ... 1156545882 Q * gerrit Ping timeout: 480 seconds 1156545981 M * nayco *rebooting* ! 1156546058 Q * shuri Remote host closed the connection 1156546264 M * nayco *not rebooting* : Make install fails :O 1156546468 P * stefani I'm Parting (the water) 1156546540 M * nayco make modules_install is better before.... :-) 1156546751 M * nayco Ok, good, rebooting 1156546778 M * nayco Oh, Bertl , will you be available in a couple o minutes to continue quota tests ?test 1156546820 M * Bertl probably 1156547130 M * nayco Ok, brb 1156547136 M * nayco thx 1156547137 Q * nayco Quit: Bonne nuit ! 1156547396 Q * someoneonet 1156547810 J * nayco ~nayco@lns-bzn-59-82-252-179-122.adsl.proxad.net 1156548066 M * nayco Bertl: Quota seem to work on the host, with your patches ! 1156548072 M * nayco 2.6.11.8 1156548080 M * nayco 2.6.17.8 1156548085 M * Bertl excellent! 1156548320 J * Hecticee ~dman@host81-129-189-119.range81-129.btcentralplus.com 1156548376 M * nayco Bertl: quota does not work or root ? 1156548405 M * nayco *for root ? 1156548548 M * Bertl well, IIRC, quota does not apply to the root user 1156548563 M * Bertl so, it should account, but not limit you 1156548660 M * Bertl welcome Hecticee! 1156548688 M * nayco well it does not account. 1156548704 M * nayco # LC_ALL=C repquota /vservers/samba4/test/ 1156548704 M * nayco *** Report for user quotas on device /dev/mapper/volume_1-1 1156548704 M * nayco Block grace time: 7days; Inode grace time: 7days 1156548704 M * nayco Block limits File limits 1156548704 M * nayco User used soft hard grace used soft hard grace 1156548706 M * nayco ---------------------------------------------------------------------- 1156548707 M * nayco root -- 0 1000 1024 3 0 0 1156548729 M * Hecticee Bertl: im about to start new virtual server but im not sure how to do it 1156548743 M * nayco Before that, I just did "dd if=/dev/urandom of=/vservers/samba4/test/BigFile bs=1k count=2k" 1156548751 M * nayco Ok, now I gonna test inside a vs 1156548753 M * Hecticee interface must be my real ipv4 from ifconfi? 1156548794 M * nayco Hecticee: don't think so. 1156548800 M * Bertl Hecticee: there are a few documents how to get started, but the tools can do a lot of things for you, like assigning ips and so 1156548830 M * Hecticee * Missing required argument 1156548833 M * Bertl Hecticee: of course, you can also use an existing ip 1156548874 M * Bertl check out http://linux-vserver.org/Documentation 1156548899 M * Bertl nayco: but for 'other' users it does account? 1156548973 M * nayco Well, problem is I created the mountpoint under /vserver, so users can't go there :-( I'm trying again with another clever mountpoint :D 1156549058 M * Bertl maybe quota isn't active at all, for whatever reason 1156549209 M * nayco Bertl: No, it finally accounts, for root. There may be a delay, or a trigger 1156549218 M * Hecticee * Missing required argument 1156549222 M * Hecticee whats wrong? 1156549250 M * Bertl no idea what you are trying to do, so no idea what is actually wrong :/ 1156549283 M * nayco Hecticee: Are you trying to create a vserver ? 1156549292 M * Hecticee yes nayco 1156549329 M * nayco Bertl: Ok, quota in the host work. Now, Inside 1156549350 M * Hecticee nayco any ideas? 1156549351 M * nayco Hecticee: so what you want to know is how to give an IP to a vserver, right ? 1156549367 M * Hecticee right 1156549403 M * nayco so, vserver foo build --interface foo=eth0:192.168.0.10 1156549410 M * nayco "foo" is the vserver name 1156549416 Q * bonbons Quit: Leaving 1156549432 M * nayco 192.168.0.10 is the IP you can give to it, depending on your network 1156549472 M * nayco Well, it is one way, there are many others... but it's the one I use ;-) 1156549590 M * nayco Hecticee: Your error is not interface-related : give us the whole command, please ! 1156549644 M * Hecticee whats its arch? 1156549661 M * Hecticee stage3 /path/to/stage3--.tar.bz2 1156549670 M * nayco Paste your command here, please, we will se what's the problem 1156549673 M * Hecticee lol 1156549677 M * Hecticee im stupid 1156549701 M * Hecticee sorry 1156549702 M * Hecticee working 1156549703 M * nayco Wich distro is this ? 1156549707 M * nayco +h 1156549849 M * nayco Hecticee: yes, you've got to install ssh in the vserver, but you'll need to modify the host server's ssh config to bind only one one IP 1156549943 M * nayco Bertl: repquota /dev/hdv2 works, repquota -a too, and repquota /test fails 1156549952 M * nayco (as axpected ?) 1156549972 M * Bertl well, yes, kind of :) 1156549987 M * nayco setquota nayco 1000 1024 0 0 /test fails 1156549990 M * nayco ;-( 1156550007 M * Bertl I still don't know the details how the tools figure the 'name' for xfs 1156550026 M * Bertl it seems to work reasonably fine (not tested for some time) with ext2/3 1156550072 M * Bertl but digging through the quota tool code should give the required info 1156550089 M * nayco Well I treid today, but it's way above me... 1156550094 M * nayco tried 1156550142 M * nayco Bertl: Guess what I forgot ;-) ? 1156550174 M * nayco echo "quota_ctl" > /etc/vservers/samba4/ccapabilities 1156550218 M * Bertl ah, so that's why you weren't allowed quota ctls 1156550277 M * Bertl well, first, let's start with the tools you use, which version of the quota tools is it? 1156550286 M * Bertl did you compile it from scratch? 1156550301 M * Bertl Hecticee: do you still need something or is everything fine now?