1125014407 M * Bertl Aiken: ah, welcome! 1125014431 J * Doener ~doener@p54876F86.dip.t-dialin.net 1125014448 M * Bertl Aiken: I built your kernel with debug info, to analyze your oops, but no luck there, the alpha kernel seems to change a lot with debug info enabled 1125014457 M * fobi Bertl: you're doing great job here.. you should start paid support ;) 1125014487 M * Aiken I now have debug enabled so addr2line finally works 1125014503 M * Aiken have not had an oops since 1125014514 M * Aiken but getting many 'too many links' errors 1125014523 M * Bertl fobi: be my guest, here is the donations page: http://www.13thfloor.at/vserver/donate/ 1125014553 M * Bertl fobi: or let me know if you want to buy consulting time :) 1125014579 M * Bertl fobi: anyway, thanks for the comment! (don't take me serious) 1125014598 M * fobi hehhe 1125014660 M * Bertl Aiken: the too many links issue means that something went wrong when the link should have been copied/broken 1125014844 M * Aiken it will be nice when this is sorted, my master image is 260 meg, the disk cost is only 3 meg for a new vserver 1125014953 M * Bertl yes, maybe for a first test, only create a few files as cow links ... 1125014972 M * Bertl (but files which will be changed, like /var/log/* or so) 1125015007 M * Bertl then see what breaks, with debugging (echo 255 >/proc/sys/vserver/debug_misc) enabled 1125015063 M * Bertl Greek0: especially the cleanup is only half assed, we need to remove the temporary copies in some cases ... 1125015140 M * Greek0 Bertl: yep, already saw that 1125015213 M * Bertl also, I think we could avoid the pathlen check ... 1125015225 M * Bertl s/check/calculation/ 1125015301 M * Aiken the link breaking is not working 1125015349 M * Aiken did the above echo, no errors are being reported 1125015441 M * Aiken I am creating a file, created the link, change the test in the linked file 1125015447 M * Aiken text in both files changes 1125015453 M * Bertl expected .. sec 1125015455 M * Aiken ls -id both files shows the same inode 1125015538 M * Bertl http://vserver.13thfloor.at/Experimental/cow_test.sh 1125015546 M * Bertl have a look at that (updated) test 1125015566 M * Bertl setattr --iunlink $DIR/five.link 1125015571 J * keyser_soze ~cimarron@host241.201-252-63.telecom.net.ar 1125015579 M * Bertl this is what makes it a COW candidate ... 1125015581 M * Aiken it was the setattr I was forgetting in my test :( 1125015685 M * Aiken and back to Too many links 1125015721 M * Bertl okay, hope you get some logging from the system :) 1125015809 M * Aiken http://pastebin.com/346441 1125015839 M * Aiken I have done the echo twice to set the debug level 1125015844 M * Aiken no debugging messages 1125015859 M * Bertl then you did something wrong with the echo 1125015886 M * Aiken I copy paste from here 1125015888 M * Bertl it roughly produces about 20 lines of log messages on each COW attempt 1125015931 M * Bertl Aiken: this isn't NFS or so again? 1125015948 M * Aiken this is all local filesystem, ext3 1125015971 M * Aiken I took nfs and the network block devices out of the equation until this lot was sorted 1125015983 M * Aiken (root@pebbles) cat /proc/sys/vserver/debug_misc 1125015983 M * Aiken 255 1125016038 M * Bertl k, you checked with dmesg? (sorry for asking those stupid questions) 1125016074 M * Aiken sb: Init: Starting Probe... 1125016075 M * Aiken sb: Probing legacy card with io=220, irq=5, dma=1, dma16=5 1125016075 M * Aiken SB 4.13 detected OK (220) 1125016075 M * Aiken sb: Init: Done 1125016075 M * Aiken aoe: aoe_init: AoE v2.6-10 initialised. 1125016076 M * Aiken devfs_mk_dir: invalid argument.<6> etherd/e9.0: unknown partition table 1125016077 M * Aiken aoe: 0000e8d80d95 e9.0 v4000 has 2097152 sectors 1125016081 M * Aiken last six lines from dmesg 1125016093 M * Aiken uptime 21 min 1125016098 M * Bertl okay, let's try: 1125016119 M * Aiken just check and I do have vserver debug set 1125016121 M * Bertl echo 255 >/proc/sys/vserver/debug_switch 1125016137 M * Bertl and then 'chcontext --xid 1 true 1125016148 M * Aiken was 0 now 255 1125016163 M * Aiken vxD: vc: VCMD_00_0[0], 63,0000000000000000,0 1125016164 M * Aiken vxD: vc: VCMD_00_0[0] = 00020001(131073) 1125016164 M * Aiken vxD: vc: VCMD_11_1[0], 1,0000000000000000,0 1125016164 M * Aiken vxD: vc: VCMD_11_1[0] = 00000000(0) 1125016178 M * Bertl so debugging works ... but not cow debugging? 1125016210 M * Bertl vxdprintk(VXD_CBIT(misc, 2), "break link »%s«", pathname); 1125016230 M * Bertl this is the first debug message and it's unconditional once cow_break_link() is called 1125016231 M * Aiken the only thing I can think of that I changed between this kernel and yesterday kernel was I turned kernel debugging on for the next time you want the output of addr2line 1125016251 M * Bertl let me see what has to happen that it isn't called ... 1125016293 M * Bertl CONFIG_VSERVER_COWBL <-- is that enabled? (another silly question :) 1125016296 M * Greek0 Aiken: dumb question, but are you sure you compiled that kernel with COW support? 1125016316 M * Greek0 hehe, same idea ;) 1125016324 M * Bertl yeah, and I guess that's it 1125016333 M * Bertl but that is actually another bug ... 1125016346 M * Aiken I remember that option from yesterday, where should it be? 1125016347 M * Bertl as we should not return the -EMLINK in this case 1125016361 M * Bertl Aiken: in your .config ... 1125016384 M * Aiken where should it be is I do a make menuconfig? 1125016401 M * Greek0 Aiken: directly under the Linux Vserver menu 1125016405 M * Bertl press '/' followed by COWBL and enter 1125016407 M * Aiken I am positive I saw it yesterday and I did have the cow stuff working yesterday 1125016439 M * Aiken it is not showing up 1125016447 M * Bertl sec 1125016480 A * Greek0 curses gcc-4.0 1125016486 M * Bertl Aiken: wrong kernel? 1125016500 M * Greek0 not only that I don't want to use a kernel compiled with gcc-4.0, it doesn't even compile with 4.0 *grml* 1125016508 M * Aiken v2.6.13-rc7-vs2.1.0-pre4 1125016567 M * Bertl Symbol: VSERVER_COWBL [=y] 1125016570 M * Bertl Prompt: Enable COW Immutable Link Breaking 1125016573 M * Bertl Defined at kernel/vserver/Kconfig:56 1125016576 M * Bertl Depends on: EXPERIMENTAL 1125016578 M * Bertl Location: 1125016581 M * Bertl -> Linux VServer 1125016589 M * Bertl (the search result) 1125016593 M * Aiken http://members.optusnet.com.au/~boddingt/menuconfig.png 1125016629 M * Greek0 Aiken: do you have "Prompt for experimental drivers" enabled? 1125016635 M * Bertl check for EXPERIMENTAL, but if that is enabled too, then your kernel is broken ... 1125016657 M * Aiken thats it 1125016689 M * Bertl Greek0: probably answers your previous question regarding experimental :) 1125016708 M * Aiken that option got caught in the quest to make the kernel small enough it would boot 1125016768 M * Bertl I guess it will not increase the size per se ... so it should be fine 1125020517 M * Aiken after a full kernel rebuild the size was increasd by abt 600 bytes 1125020535 M * Aiken and my first one for the day :( http://pastebin.com/346469 1125020584 M * Aiken pc include/linux/dcache.h:294 1125020591 M * Aiken ra fs/namei.c:2492 1125020598 M * Bertl excellent! 1125020626 M * Bertl looks like the one we had yesterday ... 1125020647 M * Aiken I had one cow working, it was a case of so far so good then started a 2nd cow 1125020657 M * Aiken very shortly after it went splat 1125020680 M * Bertl okay, let me take a look at the relevant code (will take a few minutes) 1125020729 M * Aiken any other info before I reboot? 1125020743 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net 1125020764 M * Bertl Aiken: did you enable the debug logging? 1125020772 M * Bertl (the echo command) 1125020808 M * Aiken no :( 1125020829 M * Bertl np, please reboot ... 1125020852 M * Aiken added the echo to rc.local 1125020857 M * Bertl then try to reproduce it (maybe enable the debug before you start the second one) 1125020894 M * Bertl the first one will probably give a huge pile of useless debug info 1125020917 M * Aiken it is back to locking the terminal when I try ls the directory it died in 1125020957 M * Bertl hmm, sounds like an inode sem which is still held 1125020965 M * Aiken -rw-r--r-- 3 root root 742 2005-06-30 08:33 pkg-list 1125020965 M * Aiken -rw-r--r-- 1 root root 0 2005-08-26 11:40 pkg-list© 1125020969 M * Bertl could you dump the tasks? 1125020972 M * Aiken that is the file it died on 1125020975 M * Aiken ? 1125020989 M * Bertl did you already reboot? guess so, no? 1125020998 M * Aiken yes 1125021021 M * Bertl and simply doing ls in that dir does what? 1125021068 M * Aiken I use rxvt, the terminal just stops responding. I have to kill it and start a new terminal 1125021081 M * Aiken when I ls in the affected directory 1125021118 M * Bertl could you try pressing CTRL-Q when that happens? 1125021155 M * Aiken you think it could be causing the equiv of a ctrl-s? 1125021176 M * Bertl maybe ... do I know rxvt :) 1125021220 M * Aiken only need the 1 vserver running 1125021263 M * Aiken ctrl-q does nothing 1125021283 M * Greek0 Bertl: I guess there is no reverse mapping inode->[path1,path2,..], so COW could be implemented in terms of this, is there? 1125021283 M * Bertl k, it was worth a try ... 1125021297 M * Bertl Greek0: nope :) 1125021365 M * Aiken some light reading for you http://pastebin.com/346479 1125021393 M * Bertl ah, thanks! 1125021403 M * Aiken pc include/linux/dcache.h:294 1125021410 M * Aiken ra include/asm/semaphore.h:131 1125021445 M * Aiken top line of bt fs/namei.c:1574 1125021490 M * Bertl hmm ... 1125021683 M * Bertl /etc/egg/pkg-list <-- what is that ? 1125021707 M * Bertl a symlink or a file or a dir? 1125021710 M * Aiken text file 1125021809 M * Aiken can come up with a simple test case 1125021819 M * Bertl ah, cool! 1125022023 M * Aiken simple test case executed on the host http://pastebin.com/346483 1125022059 M * Aiken with an oops of http://pastebin.com/346485 1125022191 M * Bertl hum ... trying ... 1125022313 M * Bertl yeah, really nice .. I can reproduce it here ... 1125022404 M * Aiken dcache.d and semaphore.h again 1125022413 M * Aiken same lines even 1125022418 M * Bertl yep 1125022440 M * Bertl it is not a race or so, it is a definitely unhandled case, where dentry seems to be invalid 1125022700 Q * OliverA Ping timeout: 480 seconds 1125023199 J * OliverA ~kvirc@ti200710a080-15997.bb.online.no 1125023873 M * Greek0 Bertl: wouldn't something like this work: open(oldfile); unlink(oldfile); create(newfile); newfile>? 1125023895 M * Greek0 or are there other drawbacks (like having to copy the data manually, which could miss some optimisations)? 1125023898 M * Bertl well, what do we do when the copy fails? 1125023944 M * Bertl (we can't blame it on reiser, can we? *G*) 1125023949 M * Greek0 hehe 1125024000 M * Greek0 hmm. no way to keep the old inode around while the dentry(?) is deleted? so we could re-hardlink it. 1125024025 M * Bertl does your QEMU setup work yet? 1125024041 M * Greek0 I still don't understand the concepts behind linux vfs stuff :-/ 1125024048 M * Greek0 no, haven't set it up yet 1125024064 M * Greek0 what doesn't help either is that I'm kind of running out of disk space on my laptop 1125024079 M * Greek0 I probably need to purge the windows partition.. 1125024114 M * Bertl yeah, make it a nice kernel tree space ... 1125024166 M * Greek0 heh, 30 gigs for kernel trees.. 1125024176 M * Greek0 unbelievable how much space that damn partition takes 1125024208 M * Bertl /dev/vgd/src 28G 24G 2.8G 90% /src 1125024221 M * Bertl that's how much my 2.6 trees take up :) 1125024233 M * Greek0 hehe 1125024278 M * Greek0 another nice thing would be if I got my old pc working again, that would make a fine build host too (160gb hdd, 512mb ram :( and athlon 2000+ cpu) 1125024387 M * Bertl Aiken: it's almost working now .. just some missing put/up somewhere ... 1125024838 M * Aiken so what was different about copying a file to a link compared with directly editing the link? 1125024871 M * Bertl the way open_namei() is called 1125026542 P * stefani parting (is such sweet sorrow) 1125027350 Q * nokoya Ping timeout: 480 seconds 1125027474 M * Greek0 Bertl: , first in the line of my really trivial patches (TM) :) 1125027868 M * Aiken just had the alpha say something about the ext3 journal is inconsistant, making root ro 1125027874 M * Aiken I don't like it when that happens 1125027882 M * Aiken the fsck afterwards was not fun 1125027901 M * Bertl I don't think that was related ... 1125027920 M * Bertl I mean not directly ... of course, ext3 can screw up easily on any oops 1125027953 M * Aiken I was doing lots and lots of file copy at the time 1125027970 M * Aiken it gets me wary of ext3 1125028003 M * Aiken I have had it happen a few times even on machines that had normal use 1125028410 M * Bertl *argl* 1125028456 M * Greek0 hm? 1125028487 M * Bertl I hate it if I succeed to confuse myself :) 1125028637 M * Aiken the trick is not to do it in public 1125028656 M * Bertl I'm too honest for that ... 1125028711 M * Bertl but I think we are there now .. just one more test 1125028760 M * Bertl yep, works ... 1125029075 M * Bertl Aiken: http://vserver.13thfloor.at/Experimental/delta-cow-fixup.diff 1125029090 M * Bertl (those are just the COW fixes, not much to recompile, I hope) 1125029099 M * Bertl Greek0: still around? 1125029134 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.13-rc7-vs2.1.0-pre5.diff 1125029155 M * Bertl (this version of the cow_break_link is probably more readable) 1125029178 M * Aiken bad, very bad http://pastebin.com/346571 1125029210 M * Bertl try again with -l 1125029222 M * Bertl ah, we did a small patch pre4.1 1125029227 M * Bertl sec 1125029240 M * Bertl (changes are against pre4) 1125029330 M * Bertl Aiken: you get a new one, just a second 1125029397 J * aznboifaka ~brianlam9@dialup-4.248.7.23.Dial1.Honolulu1.Level3.net 1125029401 M * Aiken just rm -rf the tree, it as in a bad state 1125029403 M * Bertl Aiken: okay, please reload 1125029418 M * Bertl welcome aznboifaka! 1125029446 M * aznboifaka lol wow you remembered me 1125029453 M * aznboifaka i was just gonna say hi to you 1125029481 M * Aiken as I am starting a new tree I'll go pre5 1125029489 M * Bertl aznboifaka: you cn still do so, no? :) 1125029502 M * aznboifaka lol 1125029534 M * Bertl Aiken: okay .. will take a while, I guess ... 1125029566 M * Aiken the tree was in a bad way and some files got nuked while recovering from the filesystem errors 1125029694 M * Bertl aznboifaka: so you are back to ask some more quesions? 1125029701 M * Bertl *questions even ... 1125029766 M * Bertl aznboifaka: and how is your friend? (with the languages :) 1125029970 A * MooingLemur pretends to moo. 1125029982 M * Bertl moo moo! 1125030130 M * aznboifaka oh sorry 1125030159 M * aznboifaka i'm tryign out the idling thing so to see if i can learn anything from the convos like you said 1125030161 M * aznboifaka remember? 1125030172 M * Bertl yup 1125030177 M * Greek0 Bertl: yep 1125030206 M * Greek0 but I'll be off to work :-/ in a few minutes 1125030223 M * Bertl np, just wanted you to know ... 1125030240 M * Greek0 thx 1125030313 T * Bertl http://linux-vserver.org/ | latest stable 2.0, 1.2.10, devel 2.1.0-pre5 -- 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 ;) 1125030601 J * nokoya young@hi-230-82.tm.net.org.my 1125030609 Q * nokoya Remote host closed the connection 1125030990 Q * keyser_soze Quit: Abandonando 1125031121 A * aznboifaka thinks that idleing doesn't teach much 1125031127 M * Bertl Aiken: the /proc/mounts was actually a 'copy' of your procfs (in the template) no? 1125031166 M * Bertl aznboifaka: well, you have to be patient, but I agree, asking teaches a lot more .. especially if you get answers :) 1125031180 J * nokoya young@hi-230-82.tm.net.org.my 1125031190 M * Bertl wb nokoya! 1125031325 M * aznboifaka lol i see 1125031327 M * Aiken the master copy has a proc/mounts, probably from when the mtab symlink was created 1125031333 M * aznboifaka wow do you check up on this channel constantly? 1125031390 M * Bertl Aiken: okay, so you cloned that one for your guests, right? 1125031459 M * Aiken the master is a guest that I had just right, now I don't run it anymore but copy it every time I want a new guest 1125031505 M * Bertl okay, but proc is not mounted inside when you 'run' them ... 1125031556 M * Bertl (because I tried to recreate the mtab case in http://pastebin.com/346479) 1125031730 M * Aiken inside a guest proc looks mounted for me 1125031745 M * Aiken certainly many files I did not put there 1125031768 M * Bertl hmm .. k, let's see what pre5 does ... 1125031775 M * Aiken in the master mount is a file probably from when that image was created 1125031832 M * Aiken in a running guest /proc/mount is a sym link to /proc/self/mounts 1125031954 M * Aiken hoppy is a guest http://pastebin.com/346601 1125032893 M * ray6 Bertl: ah, back to a normal day-timing? :) 1125032943 M * Bertl ray6: hmm, well, yes ... ( means I'm off to bed pretty soon ... :) 1125032974 M * ray6 Bertl: ah, I'm completely off-shedule, think I'll work through today %-> 1125033027 M * ray6 Bertl: and I didn't find the time to try out the xen-3-vserver-kernel I built 2 weeks ago :/ just too busy... Going to New York next week, so many things to finish before... 1125033047 M * Bertl ah, vacation or business? 1125033095 M * ray6 more vacation... 1125033172 M * ray6 I've been to New York several times in the past years and I promised my parents to show them the city when they retire... and now they're coming back to this offer %-> 1125033198 M * Bertl nice ... 1125033237 M * ray6 at least I've got WLAN in the Hotel room and on the Flight :) 1125033253 Q * aznboifaka Ping timeout: 480 seconds 1125033256 M * Bertl Aiken: any estimation? is it worth waiting a few more minutest to get a first feedback? 1125033269 M * Bertl *minutes 1125033287 M * Aiken anything upto 30 1125033319 M * Bertl ray6: yeah I heard that wlan is now in many hotels/motels ... 1125033408 M * Aiken it has just hit drivers/net/ne2k-pci.o 1125034392 M * Hollow morning 1125034434 M * Bertl morning Hollow! 1125034972 M * Bertl Aiken: okay, I guess I give up ... let me know how it did ... 1125034985 M * ray6 g'night Bertl! 1125035009 M * Bertl night folks! back later ... 1125035022 N * Bertl Bertl_zZ 1125035066 M * Aiken it is still compiling 1125035091 M * ray6 your machine is too slow :) 1125035146 M * Aiken the cross compiler is building now so hopefully won't have to put up with this again 1125035320 M * Aiken ray6 if someone wants to donate me a very fast alpha I won't say no :) 1125035500 M * Aiken so much for pre5 1125035554 M * Aiken looks like I'll have a nice OOPS for Bertl_zZ tomorrow 1125036409 M * ray6 Aiken: ah, I'm sure my alpha is slower than yours :) 1125036423 M * ray6 Aiken: the only alpha slower than mine is the 166MHz Multia :) 1125036478 M * Aiken mine is 433MHz 1125036892 J * aznboifaka ~brianlam9@dialup-4.248.4.90.Dial1.Honolulu1.Level3.net 1125037185 A * aznboifaka slaps Bertl_zZ around a bit with a large trout 1125037239 M * sven111 hi everyone 1125037250 M * sven111 I am having trouble with localhost 1125037263 M * sven111 how do set it up properly for a guest? 1125037338 M * sven111 is it enough to just make a new dir under guests settings interfaces, and set dev as lo, ip as 127.0.0.1 and prefix as 32? 1125037376 M * sven111 I just trying to get lighttpd and php-fastcgi to work under guest 1125037485 M * sven111 using that method I described I am getting lo visible in guest and I can ping it and even connect point lynx browser to it and test page comes visible 1125037547 M * sven111 but fastcgi-php spawns don't start up 1125038922 M * Aiken not bad 12 1/2 min for a kernel compile compared to abt 74 1125040647 Q * aznboifaka Quit: 1125041514 N * dsoul_ dsoul 1125041913 J * ouSmia ~jojo@147.171.62.62.internet9t.9massy1-1-ro-bas-2.9tel.net 1125043710 Q * monrad Quit: Leaving 1125044307 M * sven111 I got it fixed already :) 1125044316 M * sven111 php works well now 1125044719 J * Neubix ~brian@p54B03CE9.dip.t-dialin.net 1125044725 Q * Neubix Quit: 1125044745 J * Neubix ~brian@p54B03CE9.dip.t-dialin.net 1125044929 J * meebey meebey@booster.qnetp.net 1125045945 M * Aiken svenlll what did you do? 1125046653 M * sven111 just rechecked the things 1125046666 M * sven111 I am using Arch linux 1125046677 M * sven111 and one guy has made nice guide for setting up lighttpd 1125046733 M * sven111 http://wiki.archlinux.org/index.php/Lighttpd_For_Both_SSL_And_Non-SSL 1125046770 M * sven111 I using apache as a proxy server 1125046776 M * sven111 on host 1125046793 M * sven111 and each guest will have its lighttpd instance 1125046868 M * sven111 I like lighttpd because it uses so little memory 1125046895 M * sven111 but also I want to use apache's security modules, like mod_security or mod_bwshare 1125046921 M * sven111 this way I can combine the best parts from both :) 1125047306 Q * Aiken Remote host closed the connection 1125047353 Q * Neubix Quit: Verlassend 1125047376 M * sven111 I've been also thinking of putting that apache proxy server to a vserver guest,too 1125047393 M * sven111 so that it would be secured in a better way 1125047498 J * Aiken ~james@tooax6-010.dialup.optusnet.com.au 1125047983 Q * Aiken Quit: Leaving 1125048364 M * fobi hi all 1125048910 Q * sven111 Ping timeout: 480 seconds 1125050450 Q * Hollow Read error: Connection reset by peer 1125050451 J * Hollow ~Hollow@home.xnull.de 1125051707 Q * Hollow Remote host closed the connection 1125052340 T * * http://linux-vserver.org/ | latest stable 2.0, 1.2.10, devel 2.1.0-pre5 -- 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 ;) 1125052340 T * Bertl - 1125052355 Q * BWare Max SendQ exceeded 1125052358 J * BWare ~bware@office.intouch.net 1125052996 J * eyck eyck@81.219.64.71 1125053183 Q * Hollow Remote host closed the connection 1125053331 J * Hollow ~Hollow@home.xnull.de 1125053758 J * Milf ~Miranda@ipsio131.ipsi.fraunhofer.de 1125053875 M * Milf Can anyone tell me which tools to use when upgrading from 2.6.8.1-vs1.9.2 i686/0.30/0.30 [E] to 2.6.12.4-vs2.0Vserver-2.0 1125053923 M * daniel_hozac 0.30.208 1125053961 M * Milf Won't that get me problems switching the config type from files to configdir? 1125054083 M * daniel_hozac why would it? 1125054108 M * Milf My knowledge is a bit old on the subject. 1125054124 M * Milf Right now, I have all the vservers configed in /etc/vservers 1125054155 M * Milf It used to be that the alpha tools wanted the vservers to be configured in configdirs somewhere else. 1125054170 M * Milf I wouldn't want to have to reconfigure everything 1125054171 M * daniel_hozac that's entirely dependent on how you configure it. 1125054189 M * daniel_hozac without --sysconfdir passed to ./configure, it'll be /usr/local/etc/vservers. 1125054236 M * Milf So I can configure the configstyle with .configure? 1125054255 M * daniel_hozac no, the new tools support the legacy configurations as well. 1125054259 M * daniel_hozac but you won't get any of the new features. 1125054280 M * Milf That can wait until I get the time to try them out (TM) 1125054312 M * Milf Thanks for the help. 1125054312 M * Milf Is there any document I can read to get an overview when I finally get around to checking out the new features? 1125054507 M * daniel_hozac http://linux-vserver.org/Release+FAQ 1125055080 M * fobi "Milf" what a nick ;) 1125055089 M * Milf Ok, thanks. 1125055124 M * Milf Why do you say so? I'm using this Nick for about eleven years. 1125055402 M * fobi when i saw it my first th 1125055409 M * fobi thought was m.i.l.f. 1125055421 M * fobi let call it shortcut 1125055482 M * Milf Well my right to that nick are older than any of the affiliated pr0n sites. 1125055482 M * Milf Let's call it a freudian misperception on your part :) 1125055488 M * fobi but maybe it should be like this (soorrry for my english anyway ;) 1125055553 M * fobi Milf: ;) 1125055560 M * Milf Hmmm, problem: i don't have vconfig. Can I disable a feature so I dont need it? 1125055730 M * Milf Hmmm, there was a binary on the page mentioned; I'll use that. 1125055734 J * monrad ~monrad@213083190134.sonofon.dk 1125056081 M * Milf Will Bertls old testme.h script work with the new tools 1125056168 M * daniel_hozac yes. 1125056180 M * daniel_hozac but there may be updated versions of the script. 1125056191 M * Milf Ok, here goes the build and make install 1125056233 M * Milf Good thing I had the stable tools installed from rpm; I can reinstall them if the new ones don't work with both kernels. 1125056974 M * Milf Yep, It worked. 1125057063 M * Milf Not everything does: on trying to enter a vserver from the host I get: chcontext: 1125057063 M * Milf vc_new_s_context(): Operation not permitted 1125059165 J * Neubix ~brian@p54B03CE9.dip.t-dialin.net 1125059196 M * Neubix Medivh: today you are here ? 1125059662 M * Milf here today, gone tomorrow? 1125060473 J * Aiken ~james@tooax6-010.dialup.optusnet.com.au 1125061013 Q * cryo Remote host closed the connection 1125061265 J * cryo ~say@212.86.243.154 1125062270 Q * dsoul Ping timeout: 480 seconds 1125064769 J * keyser_soze ~cimarron@host241.201-252-63.telecom.net.ar 1125065162 P * Milf 1125065985 Q * OliverA Ping timeout: 480 seconds 1125066543 J * OliverA ~kvirc@ti200710a080-12368.bb.online.no 1125066679 J * spd1snd ~psingh@68-232-133-13.chvlva.adelphia.net 1125066872 M * spd1snd when I restart my vservers, i get "RTNETLINK answers: File exists" ...i tried to delete ou the contents in /path/to/vserver/var/lib/init.d/* but that doesnt seem to fix it... the only way i get it to go away is by rebooting the host machine, is there a better way to handle this? 1125067049 Q * ouSmia Quit: BitchX: causing all sorts of havok! 1125067070 M * spd1snd anyone around? 1125067456 M * Greek0 re 1125067491 M * Greek0 spd1snd: what vserver version? 1125067513 M * spd1snd i followed the gentoo-vserver howto... one sec, let me find out 1125067532 M * spd1snd i believe it is 1.9.5 1125067610 Q * Aiken Ping timeout: 480 seconds 1125067643 M * Greek0 I'm not sure if this has something to do with your problem, but it'd probably be a good idea to try out linux-vserver 2.0 1125067672 M * Greek0 (and util-vserver 0.30.208) 1125067699 M * spd1snd alright, ill give that a shot, thanks 1125067802 M * wibble hmm 1125067808 M * wibble why dont my vservers start on reboot? 1125067817 M * wibble /etc/rc2.d/S90vservers-default is there! 1125067830 M * spd1snd wibble: did you mark the vserver to start on by default? 1125067836 M * wibble erm no! 1125067849 M * wibble cant i just get them to all start on boot? 1125067850 M * spd1snd echo "default" > /etc/vservers//apps/init/mark 1125067852 M * wibble some kind of default? 1125067864 M * spd1snd wibble: give that command a try 1125067872 M * wibble would prefer all just to start! 1125067891 M * spd1snd wibble: me too, hehe, although this was the only way i got it to work, i had to mark each of them 1125067897 M * wibble right, rebooting them now 1125067907 M * wibble ok, i will add that into my custom newvserver script 1125067917 M * wibble maybe we should recommend that option :) 1125067995 J * renihs ~renihs___@193.170.52.70 1125068006 M * Greek0 wibble: stop 1125068023 M * wibble ... 1125068028 A * wibble stops 1125068050 M * renihs what did i miss? :) 1125068051 M * Greek0 the vserver init script looks through /etc/vservers/*/apps/init/mark 1125068074 M * Greek0 if that file contains "default" (without quotes) for a given vserver, that vserver is started by default 1125068078 M * wibble yes, I understnad 1125068085 M * wibble so you can set "times" for each one to load 1125068091 M * Greek0 if it doesn't exist, or contains another string, it will not be started 1125068116 M * Greek0 so if you just want your vservers to get started on bott time just do what spd1snd said for every vserver 1125068127 M * Greek0 (donno if it works putting it into .default) 1125068143 M * Neubix someone know hot to use the ifconfig-wrapper to get the IP Address by eth0 ?? 1125068174 M * renihs ? 1125068284 M * Neubix yes, to register confixx, you need a IP address for eth0 .. Medivh found a solution with the ipconfig wrapper. but he is noct here and I try to fond a way since 5 days .. :( 1125068376 M * Neubix http://www.cpan.org/authors/id/T/TP/TPABA/Net-Ifconfig-Wrapper/ 1125070096 J * stefani ~stefani@superquan.apl.washington.edu 1125070420 P * spd1snd 1125070421 N * Bertl_zZ Bertl 1125070465 M * Bertl morning folks! 1125070593 M * Neubix morning Bertl ! 1125070702 M * Bertl Neubix: I'd say he wrote a script which gives similar output as ifconfig on the host .. you can either 'fake' that completely or use the information from 'ip addr ls' to make up a realistic one ... 1125070725 M * Bertl (and maybe takes similar arguments too, basically an ifconfig replacement) 1125070726 Q * ntrs Quit: Leaving 1125070770 M * Bertl Neubix: because IIRC, confixx actually 'uses' the ifconfig binary to get/set it's information 1125070910 M * Hollow heya Bertl! 1125070927 M * Neubix we have create a script.. it's running. I can set the path to the ifconfig but confixx don't use it .. 1125071019 M * Bertl hey Hollow! 1125071027 M * Bertl Neubix: so _what_ does confixx use? 1125071151 M * Neubix I don't know .. I set the path to my script called ifconfig. the output of this script is right .. in the database of confixx I see only the MAC is knowed. but not the IP .. I think, I can change the path to something .. confixx will use the normal ifconfig binary 1125071183 M * Bertl what about _replacing_ the ifconfig with your script? 1125071268 M * Neubix I try it to .. but the ifconfig-wrapper is using the normal ifconfig binary. The server get a load average mor then 20 .. 1125071290 M * micah morning everyone 1125071292 M * Neubix I must killall ifconfig 1125071323 M * menomc will major-distributions deprecate ifconfig some day? 1125071349 M * daniel_hozac hopefully. ;) 1125071432 M * micah I'm looking at vserver-stat output... is VSZ the virtual memory usage of each vserver, RSS is the actual memory use, userTIME the amount of CPU time in userspace and sysTIME the amount of time in system? 1125071493 M * micah I'm trying to figure out what is eating up my swap 1125071502 M * micah and chewing my CPU :) 1125071541 M * Bertl micah: yes, but check the /proc/virtual/* entries they might provide more clues 1125071564 M * stefani micah: which machine? 1125071580 M * micah stefani: black 1125071705 M * micah i think I need coffee to interpret these results 1125071963 M * stefani micah: i have seen a new guest take up memory and cpu, but after a restart, that goes away. i've not investigated that further. 1125072024 J * dsoul darksoul@vice.ii.uj.edu.pl 1125072055 M * micah stefani: these vservers have been up for almost 40 days 1125072072 M * micah stefani: I'm just getting warnings from nagios about load and swap 1125072174 M * micah I see the mysql vserver has a VSZ of 1.8G 1125073565 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com 1125073614 Q * DuckMaster Quit: Leaving 1125073773 Q * zobel Ping timeout: 480 seconds 1125073773 J * dsoul_ darksoul@vice.ii.uj.edu.pl 1125073805 M * Bertl okay, back later .... 1125073988 Q * renihs Remote host closed the connection 1125074028 Q * dsoul Ping timeout: 480 seconds 1125074058 J * zobel zobel@zobel.irc.ftbfs.de 1125074095 J * renihs ~renihs___@193.170.52.70 1125074164 J * dsoul darksoul@vice.ii.uj.edu.pl 1125074190 N * Bertl Bertl_oO 1125074396 M * Neubix bertl, I have get it !! :) .. copy /sbin/ifconfig to /sbin/ifconfig-org and replace in /usr/share/perl5/Net/Ifconfig/Wrapper.pm all "/sbin/ifconfig" to "/sbin/ifconfig-org" ... 1125074437 M * Hollow Bertl_oO: i reimplemented many of the tools today: nidof vcontext vdlimit vflags vinfo vkill vnamespace vsched vuname vwait xidof 1125074531 M * Greek0 wow Hollow 1125074564 M * Hollow Greek0: http://dev.croup.de/proj/vserver-utils/browser/ ;) 1125074570 Q * dsoul_ Ping timeout: 480 seconds 1125074625 M * Hollow but i'm quite sure there are still bugs ;) 1125074634 M * Hollow (i already know some ;) 1125074714 M * Greek0 aeh, what's up in vinfo? 1125074729 M * Greek0 (the first file I looked into and it was pretty empty afaics) 1125074758 M * Hollow it only shows the interface version currently 1125074776 M * Greek0 ah, ic 1125074784 Q * ntrs Ping timeout: 480 seconds 1125074785 M * Greek0 makes sense ;) 1125074787 M * Hollow ;) 1125074797 M * Hollow look at vcontext or vflags... there's the real stuff ;) 1125074827 J * brian_ ~brian@p54B0401B.dip.t-dialin.net 1125074925 M * Greek0 mhm 1125075017 M * Hollow well, they're kinda plain but doing their job quite well ;) 1125075260 Q * Neubix Ping timeout: 480 seconds 1125075355 M * Hollow dinner time, back later 1125075865 J * mnemoc ~amery@200.75.27.69 1125076219 Q * menomc Ping timeout: 480 seconds 1125078694 M * Greek0 re 1125079344 Q * brian_ Quit: Verlassend 1125079580 M * Hollow leavin' for a birthday party, back tonight.. 1125080002 J * ntrs ~ntrs@m433a36d0.tmodns.net 1125080586 Q * keyser_soze Quit: Abandonando 1125082129 Q * ntrs Ping timeout: 480 seconds 1125082225 J * jim ask@208.53.159.170 1125082347 Q * mlong168 Quit: Leaving 1125082627 Q * jim Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1125082883 J * Nicoli ask@208.53.159.170 1125082897 N * Bertl_oO Bertl 1125082915 M * Bertl evening folks! 1125083060 Q * renihs Remote host closed the connection 1125083270 M * Nicoli evnin 1125083393 M * Bertl evening Nicoli! 1125083641 M * Greek0 Bertl: I'm currently trying to restore my athlon 2k, I hope to be able to use that as a build host then 1125083691 M * Greek0 and I was asking myself if I should look a bit into oom_kill.c, and try to come up with some vserver patch for it, that favors guests instead of hosts.. 1125083691 M * Bertl excellent! 1125083718 M * Bertl hmm .. so you want to kill off customers first :) 1125083751 M * Greek0 well, I thought this was what your "soft limit" was running out on.. 1125083762 M * Greek0 i.e. something you wanted to do sooner or later anyway 1125083775 M * Bertl seriously, I think it would make perfect sense to look for the guests which are over soft limit, so that's fine, yes 1125083818 M * Bertl of course, will require to get the soft limits working too ... 1125083831 M * SiD3WiNDR hm 1125083835 M * Greek0 I think the oom stuff is reasonably self-contained, otoh badness (which calculates oom score) is seems pretty delicate 1125083837 M * SiD3WiNDR my membership to the list was disabled due to bounces 1125083847 A * SiD3WiNDR wonders what bounced where with what reason 1125083848 M * Bertl Greek0: btw, did you test the 'tivial' -2 patch in a real world setup? 1125083862 M * SiD3WiNDR I'm getting enough spam and other stuff 1125083871 M * Greek0 Bertl: not really :-/ 1125083873 M * Bertl SiD3WiNDR: cross check with martin! 1125083880 M * Greek0 got to set up some test stuff first 1125083897 M * Greek0 but I'm pretty sure it's at the very least less broken then your version :) 1125083910 M * SiD3WiNDR Bertl: is he on irc? 1125083940 M * Bertl sometimes (marlow), but not now, send him email ... 1125083975 M * SiD3WiNDR k :) 1125083980 M * SiD3WiNDR I just resubscribed 1125084032 M * Greek0 I've grepped the source vor uid_t.*-2, there was only one occurence, which tested for uid == (uid_t)-2 || uid == (uid_t)(65535 - 2) 1125084095 M * Greek0 so the only major problem seems to be the 24/24 tagging, which would result in 2**24-2 being returned, which probably isn't what we want 1125084189 M * Greek0 of course that's rationalizing. as soon as the machine is up again I'll setup qemu on it, and start to test things there. 1125084202 M * Greek0 It's probably a good idea to wait till then 1125085330 Q * OliverA Ping timeout: 480 seconds 1125085890 J * OliverA ~kvirc@ti200710a080-5793.bb.online.no 1125086088 M * Bertl wb OliverA! 1125088227 M * Bertl okay, feeling a little tired, so I guess I'm off to bed now .. maybe back later though ... don't know yet ... 1125088252 N * Bertl Bertl_zZ 1125088475 Q * Doener Quit: Leaving 1125088509 J * Doener ~doener@p54876F86.dip.t-dialin.net 1125088580 P * stefani I'm Parting (the water) 1125089680 Q * Doener Quit: Leaving 1125089691 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com 1125089703 J * Doener ~doener@p54876F86.dip.t-dialin.net 1125092856 Q * Nicoli Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1125094086 J * Nicoli ask@208.53.159.170 1125094626 J * Aiken ~james@tooax8-144.dialup.optusnet.com.au 1125098941 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1125099028 Q * lilo Ping timeout: 480 seconds 1125100740 Q * Doener Read error: Operation timed out 1125100750 J * Doener ~doener@p548776E0.dip.t-dialin.net