1193270404 M * daniel_hozac --rescue sets the initstyle to rescue, so if you want rescue mode, that's required. 1193270448 M * daniel_hozac --rescue-init makes whatever command you give run as pid 1 in the guest. 1193270486 M * daniel_hozac so it's something like [--rescue [--rescue-init] *]. 1193270569 M * micah i guess I dont really know what "rescue mode" is 1193270648 M * daniel_hozac it runs the specified command instead of whatever is configured. 1193270674 M * daniel_hozac so instead of running /etc/rc.d/rc (e.g.), it runs /bin/bash, or whatever. 1193270800 M * daniel_hozac FYI i'm adding a --rescue-init argument to stop too. 1193270817 M * daniel_hozac (to stop a guest started with it) 1193271052 M * micah exec isn't really limited to shell-commands is it? I guess thats semantics 1193271136 M * daniel_hozac right. 1193271710 M * micah i'm changing that to in the man page, but I got that from the vserver --help output 1193271742 M * daniel_hozac i figured, i'll be fixing it. 1193272445 M * micah daniel_hozac: can you tell me what --sync does? 1193272786 M * daniel_hozac it performs synchronization with vshelper to ensure that the guest has been stopped. 1193272814 M * daniel_hozac right now it's a noop, as it's forced automatically on stop. 1193272857 M * daniel_hozac (and stop is the only user) 1193273224 M * micah daniel_hozac: http://svn.debian.org/wsvn/pkg-vserver/util-vserver/trunk/debian/patches/20freshenvserverman.dpatch?op=file&rev=0&sc=0 1193273250 M * micah i'll do vserver-build next, but I have a meeting to attend to now 1193273312 M * daniel_hozac i don't know man-page syntax, but shouldn't --help have \fI too? 1193273477 M * daniel_hozac the enter text is inaccurate, enter also allocates a pty. the default is to run /bin/bash if it exists. 1193273578 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa 1193273604 M * micah daniel_hozac: yeah, it should be: [\fI--help\fR], changed that and the 'enter' text 1193273670 M * daniel_hozac thanks, will merge. 1193274129 J * Medivh ck@dolphin.serverbox.de 1193274235 M * daniel_hozac micah: you wouldn't happen to know how to get the raw file out of websvn? 1193274412 M * micah daniel_hozac: websvn sucks... maybe do: svn co svn+ssh://svn.debian.org/svn/pkg-vserver/util-vserver/trunk/debian/patches 1193274453 M * daniel_hozac no dav_svn either? 1193274473 M * micah I dont think so 1193274480 Q * mire Ping timeout: 480 seconds 1193274540 Q * _gh_ Ping timeout: 480 seconds 1193275425 Q * Johnnie Ping timeout: 480 seconds 1193275463 J * agamon ~derek@CPE-24-208-3-178.new.res.rr.com 1193281935 J * friendly12345 ~friendly@ppp59-167-134-217.lns3.mel6.internode.on.net 1193284460 J * _gh_ ~gerrit@c-67-169-199-103.hsd1.or.comcast.net 1193285097 Q * agamon Quit: Leaving 1193286980 J * zLinux ~zLinux@88.213.26.14 1193289233 J * balbir ~balbir@59.178.43.69 1193289309 J * misc__ ~misc@122.2.127.232 1193289314 Q * misc__ 1193289319 J * misc-- ~misc@122.2.127.232 1193289432 M * misc-- hello... to backup a vserver is it sufficient to just backup the /etc/vservers/ and /vservers/ directories? 1193289791 Q * balbir Ping timeout: 480 seconds 1193289913 M * hparker There was just a script posted to the mailing list yesterday to backup everything... But, afaik those two directories should do it 1193289980 M * misc-- ok cool thanks 1193290227 M * misc-- if I was to have my VMs on LVM partitions then I could do live backups. Hmmm I like the idea of that. I am setting up a fileserver so I'm thinking - have the main guest OS on an LVM, then have the actual user data on another LVM. Then when I do a backup, I just do a LVM snapshot which will be minimal, transfer that over the wire, then use rsync for the data 1193290514 J * balbir ~balbir@59.178.42.209 1193290845 Q * speedy Quit: BitchX: no additives or preservatives 1193291280 J * DavidS ~david@p5481172D.dip0.t-ipconnect.de 1193291485 J * David1 ~david@p54811834.dip0.t-ipconnect.de 1193291485 N * DavidS Guest494 1193291486 N * David1 DavidS 1193291897 Q * Guest494 Ping timeout: 480 seconds 1193292119 J * cirrus_ cirrus@duempel.org 1193292158 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193292186 M * misc-- hmm has anyone ever had problems doing an 'ls' inside a specific guest dir? It's just a plain dir and only to one that I know of. If I do ls then it just pauses until I push ctrl-c a few times. However if I use 'find' then I can list it fine. On the host, I can see it using ls without a problem 1193292210 M * misc-- there is also a message on the host in my dmesg, not sure if it's related, but it says vxW: !!! limit: f6758054[VM,9] = 60 on exit. 1193292272 M * misc-- ah nevermind. Was related to my nsswitch looking up an ldap server 1193292466 Q * balbir Ping timeout: 480 seconds 1193292944 J * virtuoso ~s0t0na@ppp91-122-59-121.pppoe.avangard-dsl.ru 1193292944 Q * virtuoso_ Read error: Connection reset by peer 1193293478 Q * hparker Quit: peer reset by connection 1193294012 J * balbir ~balbir@59.178.57.212 1193295042 N * ensc Guest495 1193295052 J * ensc ~irc-ensc@p54B4DEE6.dip.t-dialin.net 1193295160 Q * Guest495 Ping timeout: 480 seconds 1193295353 Q * FireEgl Read error: Connection reset by peer 1193296100 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa 1193296439 M * harry Bertl_zZ: stop sleeping! :) 1193296450 M * harry (hmm... that's rude of me...) 1193296453 M * harry Bertl_zZ: SLEEP! ;) 1193296829 J * dna ~dna@2-195-dsl.kielnet.net 1193297072 J * larsivi ~larsivi@85.221.53.194 1193297132 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1193297160 Q * larsivi Remote host closed the connection 1193297197 J * larsivi ~larsivi@85.221.53.194 1193298080 Q * balbir Ping timeout: 480 seconds 1193298092 M * Supaplex sudo killall -SIGSLEEP Bertl_zZ 1193298268 J * gebura ~gebura@77.192.186.197 1193298442 Q * JonB Quit: This computer has gone to sleep 1193298703 M * gebura hi 1193298798 M * harry ha 1193298843 J * balbir ~balbir@59.178.60.198 1193299127 M * bXi morning 1193299147 M * bXi can i mount an users home dir on 2 places? 1193299266 M * gebura mount -o bind /home/user1 /home/myusers/user1 ? 1193299412 M * bXi as in my home i want is in vserver A and i want it to be available in vserver B as well 1193299494 M * gebura hum 1193299509 M * gebura there is a doc in the wiki but i am unable to find it 1193299514 M * gebura i use that: 1193299526 M * gebura pluton:/etc/vservers/ssh# grep home fstab 1193299526 M * gebura /var/contenu/home /home auto rbind 0 0 1193299568 M * gebura /var/contenu/home is the path on the main host , and /home the path on vserver 1193299835 Q * balbir Read error: Operation timed out 1193300436 M * bXi crap i rebooted the wrong vserver 1193300608 J * balbir ~balbir@59.178.43.138 1193300631 M * dmnd heh 1193301415 J * mire ~mire@156-171-222-85.adsl.verat.net 1193301832 J * speedy ~speedy@194.126.63.47 1193303496 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193303521 M * alex__ does the dev version have the ability to creat localhosts for guests? 1193304481 M * speedy why do you need localhost? 1193304497 M * sid3windr localhost is just an entry in /etc/hosts.. 1193304561 M * alex__ sql etc 1193304564 M * alex__ for mysql eyv 1193304566 M * alex__ etc* 1193304574 M * alex__ i need a 127.0.0.1 device 1193304600 M * alex__ local loopback 1193304603 M * alex__ whatever you want to call it 1193304760 Q * balbir Ping timeout: 480 seconds 1193305129 M * speedy i can use localhost with no problems 1193305165 M * speedy my apps point to mysql localhost 1193305241 M * alex__ i get permission problems on my grants tables 1193305247 M * alex__ so i had to fix my grants 1193305254 M * alex__ said my hostname had no permission 1193305256 M * alex__ etc 1193305291 M * alex__ so anyway.... back to my question, does the dev version have localhost / 127.0.0.1 / local looback / whatever enabled? 1193305570 N * Bertl_zZ Bertl 1193305577 M * Bertl morning folks! 1193305585 M * Bertl alex__: yes 1193305589 M * matti Hi Bertl 1193305590 M * matti :) 1193305595 M * Bertl hey matti! 1193305648 M * matti How are you? 1193305650 M * Chr0nicles morning ;) 1193305829 J * balbir ~balbir@59.178.39.78 1193305852 M * JonB Bertl: who made the vhashify sccripts? 1193306095 M * Bertl JonB: I'd assume ensc and/or daniel_hozac 1193307241 M * JonB Bertl: okay, because i am trying to do something similar at my backup server, though with 7.5 million files and 3.6 TB data. It takes a month just to find all the files and make a single file pr. filesize with a text line, 1 pr. file with that size. After that have to check if they are identical. I think i do it hopelessly inefficient 1193307348 M * Bertl looks like, but it depends on many factors like number of identical files, filesystem used, etc 1193307385 M * JonB Bertl: the filesystem is ext3, and it's all on sata disks of at least 400GB size 1193307392 M * JonB Bertl: i dont even compare the identical files 1193307418 M * JonB i just find all the files which have the same size and make a 1 file for every filesize that exists 1193307425 M * Bertl crawling and checksumming about 2TB takes roughly a day here 1193307432 M * JonB okay 1193307454 M * JonB i use a month for 2x that size. I must be doing something wrong 1193307494 M * JonB and i only do crawling 1193307571 M * Bertl well, check how long: 'find /storage -type f | wc' takes 1193307602 M * Bertl that is roughly your lower bound 1193307618 M * Bertl matti: fine, thanks, and you? 1193307670 M * alex__ Bertl, how "stable" is dev? 1193307686 M * JonB Bertl: right, good idea 1193307709 M * Bertl alex__: it is 'devel stable' :) 1193307783 Q * meandtheshell Quit: Leaving. 1193308038 Q * JonB Quit: This computer has gone to sleep 1193308403 J * amax ~a@r5ct9.net.upc.cz 1193308422 M * amax hi all 1193308428 M * amax daniel_hozac hi, one q. 1193308464 M * amax whats the easy way to startup guests in the strong order ? 1193308498 M * amax for example if many guests depends on first, how to make _first_ to start really first ? 1193308533 M * amax at my point i have authorization server inside guest, and all other guests uses it too 1193308536 M * amax ldap/samba 1193308599 M * amax how can i specify a case "startup a guest in 2 min after starting 'myspecialguest' " ? 1193308796 M * Bertl the order is controlled by a dependancy in the guest config 1193308818 M * Bertl but for the timing, you have to put 'delays' into the scripts 1193308855 M * Bertl i.e. there is no concept of 'so and so long after' 1193308984 M * amax i see thanks 1193308998 M * amax dependancy which key of settings& 1193308999 M * amax ? 1193309013 M * amax maybe i have skipped something.. 1193309024 M * amax key for startup priority.. 1193309026 M * amax hm.. 1193309056 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1193309061 M * Bertl scroll down to 'depends' 1193309081 M * amax thanks 1193309087 M * Bertl you're welcome! 1193309128 M * amax Content of this file are vserver ids (one name per line). 1193309138 M * amax does it mean "context id" ? 1193309167 M * Bertl I think the name should be fine, but most likely xids will work too 1193309192 M * amax i see ) will try now it.. 1193309424 J * meandtheshell ~markus@85.127.115.87 1193309567 Q * pmenier Quit: Konversation terminated! 1193310213 M * matti Bertl: Not bad :) Thanks :) 1193310377 Q * DavidS Ping timeout: 480 seconds 1193310400 J * DavidS ~david@vpn.uni-ak.ac.at 1193310755 Q * derjohn Ping timeout: 480 seconds 1193310810 J * derjohn ~derjohn@80.69.41.3 1193311723 J * lilalinux ~plasma@dslb-084-059-002-135.pools.arcor-ip.net 1193311814 J * CWC ~CWC@89-215-37-177.2073053861.ddns-lan.pl.ekk.bg 1193312091 J * Yvo yvonne@vpn188.rz.uni-mannheim.de 1193312106 M * Bertl wb DavidS! CWC! lilalinux! Yvo! 1193312130 M * Yvo Hi :-) 1193312399 Q * CWC Quit: Client exiting 1193312496 Q * lilalinux Remote host closed the connection 1193313338 P * friendly12345 1193313795 M * Bertl daniel_hozac: how about that? http://paste.linux-vserver.org/7601 1193313822 M * daniel_hozac xid, nid, tag? 1193313826 M * daniel_hozac looks nice. 1193313832 M * Bertl yep 1193313860 M * daniel_hozac but could we use some character that displays correctly in all charsets? :) 1193313863 M * Bertl I'm not sure we can add the xid to the limits, checking that now 1193313874 M * Bertl daniel_hozac: suggestions? 1193313901 M * daniel_hozac ' or >< or [] or ()... :) 1193313917 M * Bertl problem there is, they are too common in filenames 1193314017 M * daniel_hozac well, is that really a problem? they're just separators. 1193314076 M * Bertl yes, but we want them to be really visible 1193314109 M * Bertl I think we could do with '' or "" but they have the disadvantage that you cannot tell end from start 1193314126 M * Bertl besides the fact, that they are present in filenames too 1193314311 J * Bytewurm ~bytewurm@bytewurm.rising-systems.de 1193314313 M * Bytewurm hi! 1193314359 M * Bertl welcome Bytewurm! 1193314375 Q * meandtheshell Quit: Leaving. 1193314400 M * Bytewurm hi bertl. would you have 1 minute to listen and let me know which part of the documentation I missed? 1193314468 M * Bertl sure ... 1193314520 M * Bytewurm i want to move a vserver (1.2 based) to another master. old system debian sarge 2.4 kernel, new one debian etch 2.6 kernel with 2.2.0 vserver 1193314547 M * Bytewurm when i try to launch the vserver on the new home (w/o patching itself) mysql refuses to start and points out a probelm with localhost-connection. 1193314562 M * Bytewurm I#ve tried to work around hat but failed .. another freshly-created vserver on the new box does fine. 1193314578 M * Bytewurm I'd expect some incosistencies between sarge/etch but cannot point it out to solve it ;-/ 1193314589 M * Bytewurm so I am watching for hints now :-) 1193314611 M * Bertl I would check the /etc/hosts file of both and the mysql config 1193314644 M * Bytewurm i did and got confused .. although 127.0.0.1 is bound to localhost on the freshly created vserver it launched mysql w/o pain (referring to 127.0.0.1) in my.cnf 1193314660 M * Bytewurm but the "moved" vserver does not ... 1193314694 M * Bytewurm I#ve tried to bind localhost to the ip of the vserver -> no success, rewriting my.cnf did not work as well. 1193314717 M * Bertl which probably means that the 'host access' configuration of your mysqls does differ 1193314719 M * daniel_hozac what's the problem you're encountering? 1193314749 M * Bytewurm bertl: ok, I'll double check that again ... 1193314889 M * Bertl daniel_hozac: I thought about showing link counts of unified files as '1' inside a context, what do you think about that? 1193314928 M * Bytewurm hmm, checked it .. bind-adress is 127.0.0.1 with the vserver that works and points to the primary ip of the vserver where it does not work (and 127.0.0.1) doesn't either. 1193314930 M * daniel_hozac to hide the fact that they're unified? 1193314948 M * daniel_hozac Bytewurm: so what does strace say? 1193314967 M * Bytewurm mompls 1193314969 M * Bertl daniel_hozac: yes, of course that could be a flag 1193315002 M * daniel_hozac Bertl: shouldn't we also hide the immutable bit in that case? 1193315007 M * Bertl daniel_hozac: looking at the test results, don't we have any barrier checks in testfs? 1193315023 M * daniel_hozac aren't they in testme? 1193315053 M * Bertl hmm? 1193315088 M * daniel_hozac nevermind, we do have them in testfs. 1193315103 M * daniel_hozac test 106 1193315135 M * Bertl ah, right, completely missed the output 1193315203 M * Bytewurm *gnarf* I've been on the wrong trail ... the socket seems to be in use and only the mysqladmin-connect to localhost lead me to the localhost/127.0.0.1 issue. 1193315228 M * Bytewurm hmmm, will check iptables setup and interfering with the other vserver running a mysql-database. 1193315247 M * Bytewurm thanks for you patience and simply letting me dump my problems to view them from outside ;-) 1193315274 M * Bytewurm and: thanks a lot for vserver!!! 1193315288 Q * Bytewurm Quit: *waves goodbye for today* 1193315508 Q * balbir Read error: Operation timed out 1193315764 M * Bertl daniel_hozac: btw, I think we should answer the questions: "how stable is XXX" with "XXX stable", what do you think? :) 1193315791 M * daniel_hozac hehe, makes sense. 1193316125 Q * pusling Ping timeout: 480 seconds 1193316408 J * balbir ~balbir@59.178.54.0 1193316449 M * bXi does the context id somehow affect the memory regions? 1193316487 M * Bertl no, it should not 1193316566 M * Bertl daniel_hozac: http://vserver.13thfloor.at/Experimental/delta-debug-feat02.diff 1193316616 J * pusling pusling@77.75.162.71 1193316662 M * daniel_hozac Bertl: nice, i guess it's hat eating time :) 1193316865 Q * Yvo Ping timeout: 480 seconds 1193317000 M * Bertl daniel_hozac: yep, looking forward to it :) 1193317040 Q * virtuoso Ping timeout: 480 seconds 1193317043 M * daniel_hozac amax: Bertl's done his part, now it's your turn :) 1193317151 Q * wibble Remote host closed the connection 1193317630 Q * amax 1193317634 M * daniel_hozac hehe. 1193317778 J * Piet ~piet@tor.noreply.org 1193317831 J * hparker ~hparker@linux.homershut.net 1193317880 Q * Aiken Remote host closed the connection 1193317887 M * Bertl daniel_hozac: is secure mount intended for normal usage? 1193317905 M * Bertl daniel_hozac: do we have plans for a vmount or so? 1193317919 J * wibble wibble@vortex.ukshells.co.uk 1193317926 M * daniel_hozac what would vmount do? enter the namespace and do the mount there? 1193317936 M * Bertl wb wibble! 1193317947 M * wibble thanks... someone pulled the plug! 1193317958 M * Bertl daniel_hozac: yes, maybe with the config fstab syntax or so? 1193317993 M * Bertl s/syntax/semantics/ 1193318044 Q * speedy Quit: [BX] Time wasted: 4h 30m 28s 1193318064 M * daniel_hozac should be a fairly simple wrapper script... 1193318168 M * Bertl I think it would help folks like the guy asking for the namespace mounts 1193318177 M * Bertl (on the ML) 1193318197 J * pmenier ~pmenier@ACaen-152-1-67-157.w83-115.abo.wanadoo.fr 1193318629 J * lilalinux ~plasma@dslb-084-059-002-135.pools.arcor-ip.net 1193318759 Q * lilalinux Remote host closed the connection 1193318937 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193319589 J * Yvo yvonne@vpn046.rz.uni-mannheim.de 1193319946 Q * balbir Read error: Operation timed out 1193320017 Q * larsivi Quit: Konversation terminated! 1193320163 J * virtuoso ~s0t0na@ppp91-122-103-25.pppoe.avangard-dsl.ru 1193320700 J * balbir ~balbir@59.178.47.85 1193321460 P * Yvo 1193322035 Q * nou Ping timeout: 480 seconds 1193322088 J * pmjdebruijn ~pascal@62.133.206.55 1193322105 M * Bertl welcome pmjdebruijn! 1193322181 M * pmjdebruijn hi 1193322248 J * nou Chaton@causse.larzac.fr.eu.org 1193322254 M * Bertl wb nou! 1193322290 Q * _gh_ Ping timeout: 480 seconds 1193322333 M * pmjdebruijn vserver VSERVERNAME exec 'ls | tee /var/log/ls.log' gives me issues 1193322350 M * pmjdebruijn I can reproduce it here, we just tested the latest vserver patch (2.6.22.10vs2.0.4) 1193322364 M * Bertl try vserver VSERVERNAME exec bash -c 'ls | tee /var/log/ls.log' 1193322392 M * Bertl pmjdebruijn: but vs2.0.4 is hardly the latest, do you mean 2.2.0.4? 1193322405 M * pmjdebruijn uh yes 1193322433 M * pmjdebruijn but why is that? 1193322440 M * Bertl what? 1193322446 M * pmjdebruijn why doesn't my command work? 1193322467 M * Bertl I'd say, because the shell doesn't know how to handle 'ls | ls' as command? 1193322482 M * Bertl $ 'ls | ls' 1193322482 M * Bertl -bash: ls | ls: command not found 1193322501 M * Bertl $ bash -c 'ls | ls' 1193322506 M * Bertl OTOH, works fine, no? 1193322522 M * pmjdebruijn oh 1193322525 M * pmjdebruijn right 1193322568 M * pmjdebruijn thankyou 1193322571 M * pmjdebruijn that's a big help 1193322585 M * Bertl you're welcome! feel free to hang around! 1193322682 M * pmjdebruijn heh 1193322689 M * pmjdebruijn it's slightly ennoying you're not on FreeNode 1193322719 M * Bertl well, _I_ am, but the channel is not, but I think any client nowadays can handle multiple servers 1193322741 M * Bertl (well, any IRC client worth using, that is :) 1193322760 M * harry Bertl: ever took time to look at the len>>PAGE_SHIFT patch i made? 1193322776 M * Bertl harry: yes, but I didn't include it yet 1193322789 M * harry is it any good? 1193322793 M * harry (just wondering) 1193322811 M * Bertl well, removing recalculations is always a good idea 1193322832 M * harry since it's done 8 times in a pretty extensively used function... ;) 1193322850 M * pmjdebruijn Bertl: that's not really the point, having to connect on 10 different network is very ennoying 1193322856 M * Bertl harry: care to provide the url once again, then I'll have a close look right now 1193322882 M * harry http://harry.enzoverder.be/charged.diff 1193322900 M * Bertl pmjdebruijn: options there are autojoin, irc-bouncer or screen+irssi 1193322934 M * pmjdebruijn Bertl: like I said the technical part isn't the problem... 1193322962 M * pmjdebruijn but thanks again 1193322967 Q * pmjdebruijn Quit: Ex-Chat 1193323109 J * dowdle ~dowdle@scott.coe.montana.edu 1193323125 Q * JonB Quit: This computer has gone to sleep 1193323244 M * Bertl wb dowdle! 1193323328 M * Bertl harry: well, the do_brk() change is a little gcc dependant 1193323350 M * Bertl i.e. you might as well end up with much more work in the hot path 1193323360 Q * balbir Ping timeout: 480 seconds 1193323384 M * Bertl but I think it should be fine with recent gcc 1193323389 J * onox ~onox@kalfjeslab.demon.nl 1193323452 M * dowdle Bertl: Morning/Evening. 1193323906 M * harry Bertl: i don't think that the patch will be a problem in older compilers either? 1193323937 M * Bertl well, you assign charged early, and it might be that it isn't used at all 1193323980 M * Bertl did you do some compile tests, to see how the code size behaves? 1193323984 M * harry hmm... 1193323990 M * harry not yet 1193323993 M * harry no time 1193323999 M * Bertl note: I think it should be fine 1193324021 M * harry hell, i don't even have a testmachine anymore 1193324105 J * balbir ~balbir@59.178.42.215 1193324220 J * dna_ ~dna@39-241-dsl.kielnet.net 1193324315 J * waldi ~waldi@bblank.thinkmo.de 1193324539 M * Bertl wb balbir! dna_! waldi! 1193324627 Q * dna Ping timeout: 480 seconds 1193324814 Q * balbir Read error: Operation timed out 1193324827 M * Bertl daniel_hozac: here a first try: http://vserver.13thfloor.at/Experimental/delta-cow-fix18.diff 1193324851 J * _gh_ ~gerrit@bi01p1.co.us.ibm.com 1193325343 M * daniel_hozac Bertl: hmm, out_unlock_new is below out_unlock, no? 1193325448 M * Bertl hmm, right ... should move that down then 1193325491 M * Bertl ah, won't work, messes up the ret code 1193325502 M * Bertl okay, needs some other solution I guess 1193325564 J * balbir ~balbir@59.178.61.97 1193326254 J * Johnnie ~jdlewis@c-67-163-142-234.hsd1.pa.comcast.net 1193326553 M * daniel_hozac Bertl: what about http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix19.diff? 1193326639 M * Bertl is the out_rel_both -> out_unlock_new a fix in the current code? 1193326649 M * daniel_hozac oh, that's an unrelated hunk. 1193326670 M * Bertl so the idea is to check only once after the lock 1193326672 M * dilinger Bertl: btw, i don't know if you saw this: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=ddbcc7e8e50aefe467c01cac3dec71f118cd8ac2 1193326692 M * dilinger framework for implementing containers 1193326722 M * daniel_hozac he's CCed in it ;) 1193326754 M * Bertl dilinger: yes, but a) it's quite new and not much tested 1193326771 M * Bertl dilinger: b) we can avoid the 'grouping' by using actual tags in many cases 1193326781 M * dilinger daniel_hozac: i get cc'd on plenty of stuff; doesn't necessarily mean that i pay attention ;) 1193326818 M * Bertl dilinger: so it currently is too heavy for Linux-VServer purpose 1193326925 Q * gebura Quit: Quitte 1193327024 M * Bertl daniel_hozac: so we want to drop the early exit completely? 1193327069 M * daniel_hozac i'm not sure... there is a point to dropping out before we've done the splice, but it will lead to a bit of code duplication. 1193327107 M * Bertl let's do that in two steps, I'll first clean up the ret/res mess 1193327166 M * daniel_hozac there's another problem too, in that we won't unlink the temporary file. 1193327195 M * daniel_hozac (my patch is missing a dput(new_dentry) too) 1193327820 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193328348 Q * DavidS Quit: Leaving. 1193328765 M * Bertl daniel_hozac: http://vserver.13thfloor.at/Experimental/delta-cow-fix19.diff 1193328780 M * Bertl this one is not supposed to change the semantics 1193328824 Q * balbir Read error: Operation timed out 1193328924 M * daniel_hozac ah, nice. 1193329459 Q * onox Quit: blaat 1193329487 M * Bertl and this one should take care of the redo 1193329488 M * Bertl http://vserver.13thfloor.at/Experimental/delta-cow-fix20.diff 1193329569 J * balbir ~balbir@59.178.51.215 1193329572 M * Bertl have to take a break now, let me know what you think of those changes ... 1193329577 M * Bertl will be back later ... 1193329581 N * Bertl Bertl_oO 1193329632 J * tanjix ~tanjix@dslb-084-058-028-253.pools.arcor-ip.net 1193329635 M * tanjix hello everyone 1193329642 M * daniel_hozac it looks good, will do some testing.. 1193329650 M * daniel_hozac hi tanjix 1193330150 Q * Johnnie Ping timeout: 480 seconds 1193330347 Q * FireEgl Quit: Bye... 1193330883 M * dowdle As seen on /. : http://kerneltrap.org/OpenBSD/Virtualization_Security 1193330932 J * tabjix2 ~tanjix@office.star-hosting.de 1193330938 N * tabjix2 tanjix2 1193331122 Q * tanjix Ping timeout: 480 seconds 1193331223 M * JonB dowdle: yeah i read it 1193331297 M * JonB dowdle: but vserver is not xen 1193331325 M * dowdle I think the security improvement is in isolating services so that if one service gets compromised, you can't get beyond that single service... so in that sense, it does improve security... but for those of us virtualizating everything and not using it to separate services, not so much. 1193331362 M * dowdle JonB: Same theory applies... if 1 thing isn't secure (the OS), 10 copies isn't secure. 1193331488 J * Yvo ~yvonne@91.64.217.106 1193331580 M * tanjix2 main80:~# setattr --~hide /proc/loadavg 1193331591 M * tanjix2 i get an error when doing that 1193331598 M * tanjix2 function not implemented 1193331626 M * daniel_hozac what kernel and how did you configure it? 1193331679 M * tanjix2 standard debian kernel 2.6.18.5-amd64... the problem is existent since last reboot... i cannt start any vservers any more because of an error regarding proc... i also do not have the cprocunhide script shich usually solves thatr 1193331681 M * JonB dowdle: i agree that most security is in 1 service pr machine making them small and manageable 1193331709 M * daniel_hozac how did you install the utils? 1193331720 M * tanjix2 apt-get install util-vserver ;) 1193331737 M * daniel_hozac did you run /etc/init.d/vservers start? 1193331744 M * tanjix2 yes 1193331750 M * tanjix2 the vservers cannot stat 1193331754 M * tanjix2 start 1193331761 M * JonB dowdle: but i do think that because vserver makes processes you get a much more well known security problem than with xen that emulates a computer 1193331783 M * tanjix2 hm but i see the problem... 1193331791 M * tanjix2 the reboot loaded a kernel without vserver support 1193331798 M * daniel_hozac well, that'll do it... 1193331893 M * bXi can you move vservers transparantly between hosts? 1193331947 M * daniel_hozac transparently meaning the same process with the same state will be running on the other host? 1193332016 M * tanjix2 http://paste.linux-vserver.org/7626 1193332028 M * tanjix2 anyone an idea what is wrong and how can i fix? 1193332102 M * dowdle JonB: Don't know. 1193332159 M * dowdle tanjix2: dpkg doesn't do dependenies (use apt-get)... and it bitching that you can't install that package without all of the packages it depends on... or at least that's how I'm reading it. 1193332187 M * tanjix2 how do i install my .deb package aith apt-get ? 1193332203 M * daniel_hozac doesn't look like deps at all to me. 1193332212 M * daniel_hozac more like something corrupted your dpkg database. 1193332238 M * tanjix2 any way to restore it ? 1193332291 M * daniel_hozac backups? i don't know how dpkg works. 1193332333 M * dowdle daniel_hozac: Same thing only different. :) 1193332508 M * tanjix2 hmm 1193332687 M * Yvo tanjix2: you wrote, that the reboot loaded a kernel without vserver support, is that your problem? 1193332740 M * tanjix2 the current problem is that i cannot install anything 1193332750 M * tanjix2 with apt 1193332892 M * Yvo and what did you install before? 1193332942 M * tanjix2 nothing 1193333014 M * pmenier tanjix2: did you try apt-get check and/or apt-get install -f 1193333055 M * tanjix2 pmenier: -f does not help, and apt-get check does not report errors 1193333098 M * pmenier and apt-get install --reinstall libmd5-perl didn't help any more ? 1193333131 M * tanjix2 no, i also tried that - rings me the same error for this package 1193333135 M * Yvo tanjix2: you did nothing and now it doesn't work anymore? 1193333145 M * tanjix2 Yvo: yes 1193333160 M * Yvo strange 1193333661 M * pmenier tanjix2: perhaps more info here : http://osdir.com/ml/debian.devel.dpkg.general/2003-06/msg00010.html 1193333666 J * meandtheshel1 ~markus@85.127.115.87 1193333672 M * pmenier same probleme as yours 1193333839 M * tanjix2 pmenier: thanks. therei is only soluted "apt-get -f install" - that did not help in my situation 1193333964 M * pmenier i see that, too : dpkg i --force-confmiss but not sure it may help you 1193334658 M * Yvo tanjix2: your paste-post says, that the file-list for some packages are missing, did you try to install especially these packages? 1193335834 Q * balbir Remote host closed the connection 1193336684 J * ema ~ema@rtfm.galliera.it 1193337903 J * DLange ~dlange@p57A30715.dip0.t-ipconnect.de 1193337918 Q * Vudumen Ping timeout: 480 seconds 1193337992 Q * sid3windr Ping timeout: 480 seconds 1193338118 J * sid3windr luser@bastard-operator.from-hell.be 1193338226 J * Vudumen ~vudumen@perverz.hu 1193341423 J * larsivi ~larsivi@101.84-48-201.nextgentel.com 1193342410 Q * JonB Quit: This computer has gone to sleep 1193343058 Q * brc Read error: Operation timed out 1193343884 Q * fatgoose Read error: Operation timed out 1193343890 J * raa ~raa@static-207-114-230-7.asisna.com 1193343915 J * fatgoose ~samuel@76-10-151-220.dsl.teksavvy.com 1193344403 Q * raa Quit: Leaving 1193344513 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa 1193345089 J * southtel ~southtel@68-114-23-151.dhcp.gwnt.ga.charter.com 1193345127 Q * djbclark Remote host closed the connection 1193345187 M * southtel Is there a more in-depth description of the /etc/vservers files other than the great flower page? 1193345198 M * daniel_hozac in-depth meaning what? 1193345211 M * southtel I'm trying to get more in-depth explanations of some of the files (specifically the interface files) 1193345246 M * daniel_hozac such as? 1193345283 M * southtel How do I use nodev, if it's even recommended. 1193345294 M * daniel_hozac you create it. 1193345306 M * daniel_hozac e.g. with touch. 1193345311 M * southtel Right. 1193345331 M * daniel_hozac nodev just means that the utils won't manage your IP addresses. 1193345342 M * southtel But, don't I then need to specify the "real" interface/device somewhere? 1193345361 M * daniel_hozac Linux-VServer doesn't care about interfaces. 1193345370 M * daniel_hozac it's entirely IP-based. 1193345397 M * southtel Let me rephrase... 1193345515 J * djbclark dclark@opensysadmin.com 1193345518 M * southtel When I'm using nodev to not have to specify the ip...do I still specify the device in dev? 1193345559 M * daniel_hozac uh, it's the other way around. 1193345567 M * daniel_hozac you always have to specify the IP address. 1193345572 M * daniel_hozac that's what it's all based on. 1193345587 M * daniel_hozac when nodev is present, dev doesn't matter. 1193345589 J * Aiken ~james@ppp121-45-206-11.lns1.bne1.internode.on.net 1193345628 M * southtel Okay, let's try this...when I specify nodev, what other files (if any) are used for configuration? 1193345670 M * daniel_hozac ip and prefix are always required. 1193345745 M * southtel Awesome, thanks. 1193346055 Q * djbclark Remote host closed the connection 1193346081 J * djbclark dclark@opensysadmin.com 1193346356 M * southtel And am I correct in also assuming that with nodev specified, that the ip is not setup/torn down upon start stop? 1193346371 M * daniel_hozac that is what nodev means... 1193346528 J * yarihm ~yarihm@84-75-130-73.dclient.hispeed.ch 1193346532 Q * djbclark Remote host closed the connection 1193346537 J * djbclark dclark@opensysadmin.com 1193346809 Q * DLange Quit: Bye, bye. Hasta luego. 1193346986 Q * djbclark Remote host closed the connection 1193347195 J * djbclark dclark@opensysadmin.com 1193347195 Q * southtel Read error: Connection reset by peer 1193347743 Q * djbclark Read error: Connection reset by peer 1193347798 Q * Piet Ping timeout: 480 seconds 1193347868 J * speedy ~speedy@home.speedy.org 1193348004 J * djbclark dclark@opensysadmin.com 1193348009 J * southtel ~southtel@68-114-23-151.dhcp.gwnt.ga.charter.com 1193348391 J * Piet ~piet@tor.noreply.org 1193348481 Q * larsivi Quit: Konversation terminated! 1193349023 Q * tanjix2 Ping timeout: 480 seconds 1193349222 P * Yvo 1193349222 Q * southtel Read error: Connection reset by peer 1193349732 Q * sid3windr Remote host closed the connection 1193349838 J * brc bruce@megarapido.cliquerapido.com.br 1193350076 J * southtel ~southtel@68-114-23-151.dhcp.gwnt.ga.charter.com 1193350302 J * tanjix tanjix@77.37.25.239 1193350319 J * tanjix2 ~tanjix@office.star-hosting.de 1193350377 J * sid3windr luser@bastard-operator.from-hell.be 1193350587 Q * nebuchad` Ping timeout: 480 seconds 1193350780 J * Johnnie ~jdlewis@c-67-163-142-234.hsd1.ct.comcast.net 1193350785 Q * tanjix Ping timeout: 480 seconds 1193350819 J * Beuc ~yo@82.238.35.175 1193350900 M * Beuc Hey, I get "RTNETLINK answers: Numerical result out of range" when starting a vserver. I configured interfaces/ with one new IP on dev=eth0,name=autobuilder,ip=192.168.30.11/32. I have a similar vserver (192.168.30.10/32) which runs just fine! Any clue? :) 1193350980 M * Beuc The main problem being that the eth0:autobuilder interface isn't created, so no network for the vserver. 1193350992 Q * fatgoose Quit: fatgoose 1193351270 Q * Piet Quit: Piet 1193351301 M * Beuc OK, the problem goes away when I rename "eth0:autobuilder" to "eth0:11" 1193351336 M * Beuc I guess there's a size limit (or a truncated ifconfig output parsing) somewhere :) 1193351776 Q * meandtheshel1 Quit: Leaving. 1193351777 Q * pmenier Ping timeout: 480 seconds 1193351818 J * pmenier ~pmenier@ACaen-152-1-111-155.w86-215.abo.wanadoo.fr 1193352208 M * bzed Beuc: I guess you can't use anythoign but numbers in a if name 1193352229 M * bzed Beuc: you don;t need the name at all btw 1193352229 M * Beuc bzed: it works manually 1193352257 M * Beuc bzed: true, though it shows in ifconfig with a name 1193352276 M * daniel_hozac a name cannot be longer than 16 characters. 1193352292 M * daniel_hozac name being "eth0:autobuilder" in this case, which is too long. 1193352429 M * Beuc daniel_hozac: thanks. Strange that ifconfig accepts it, maybe it's truncating it silently 1193352453 M * daniel_hozac ifconfig is using interfaces that were deprecated so many years ago it's not even funny. 1193352520 M * Beuc heh 1193352589 M * Beuc ok, apparently it truncates to eth0:autobuilde silently 1193353323 P * dowdle Konversation terminated! 1193354387 Q * ema Quit: leaving 1193354462 J * nebuchadnezzar ~nebu@zion.asgardr.info