1236211589 Q * dowdle Remote host closed the connection
1236214369 Q * yarihm Quit: This computer has gone to sleep
1236215225 J * bliz42 ~kdsmith@c-68-53-128-240.hsd1.tn.comcast.net
1236215235 Q * bliz42 
1236215247 J * bliz42 ~kdsmith@c-68-53-128-240.hsd1.tn.comcast.net
1236215253 M * bliz42 hello hello
1236215832 M * bliz42 is there a significant performance cost to use a different guest distro than the container distro? or should that be avoided when possible?
1236215899 J * dowdle ~dowdle@71-32-17-196.blng.qwest.net
1236216208 M * daniel_hozac there's no performance cost at all.
1236216234 M * bliz42 perhaps more overhead because the libs will vary and can't be hardlinked/shared in memory?
1236216237 M * daniel_hozac the host and guest distro's aren't related in any way.
1236216248 M * daniel_hozac you can't unify with the host anyway.
1236216258 M * bliz42 oh ok.. tks :)
1236216463 M * bliz42 and context's need to be specified manually when creating a new vserver correct?   it looks like at one point these were optionally dynamic, but that stopped at some point?  i assume there's an advantage to this?
1236216563 M * daniel_hozac recent util-vserver versions will generate a static one for you.
1236216570 M * daniel_hozac if you don't specify it.
1236216609 M * bliz42 oh ok.. i thought i had created one without problems and without specifying a context id, but a co-worker told me i must be mistaken and didn't realize i told it what to use...
1236216750 J * saulus ~saulus@d026130.adsl.hansenet.de
1236220737 J * jpablo ~chatzilla@186.81.64.167
1236220752 Q * bliz42 Quit: bliz42
1236221921 J * yarihm ~yarihm@77-56-182-18.dclient.hispeed.ch
1236222056 J * derjohn_foo ~aj@e180207148.adsl.alicedsl.de
1236222221 Q * yarihm 
1236222228 J * sardyno_ ~me@pool-96-235-18-120.pitbpa.fios.verizon.net
1236222445 Q * sardyno Ping timeout: 480 seconds
1236222456 Q * aj__ Ping timeout: 480 seconds
1236222650 M * geb harry, i was just trying to use build a vserver-grsec for the last 2.6.27, it seems ok but thanks for your offer :)
1236222683 M * geb good night !
1236222687 Q * geb Quit: Quitte
1236222708 J * balbir_ ~balbir@122.172.52.181
1236223128 Q * dowdle Remote host closed the connection
1236223616 J * bono bono@118-160-164-52.dynamic.hinet.net
1236223670 Q * uva Ping timeout: 480 seconds
1236224978 J * doener_ ~doener@i577B9611.versanet.de
1236225082 Q * doener Ping timeout: 480 seconds
1236226253 Q * Piet Quit: Piet
1236226983 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1236229934 Q * balbir_ Read error: Connection reset by peer
1236230295 Q * kaner Remote host closed the connection
1236230299 J * kaner ~kaner@zzz.strace.org
1236231072 Q * DreamerC Quit: leaving
1236231092 J * DreamerC ~DreamerC@122-116-181-118.HINET-IP.hinet.net
1236233640 Q * _Shiva_ Remote host closed the connection
1236233650 J * _Shiva_ shiva@whatcha.looking.at
1236234010 Q * jpablo Quit: ChatZilla 0.9.84 [Iceweasel 3.0.6/2009020407]
1236234014 J * Slydder1 ~chuck@dslb-088-072-213-089.pools.arcor-ip.net
1236235005 Q * sladen Ping timeout: 480 seconds
1236235026 J * scientes ~scientes@75-165-95-28.tukw.qwest.net
1236236747 Q * Slydder1 Quit: Leaving.
1236237013 Q * kwowt Ping timeout: 480 seconds
1236237133 J * ncopa ~ncopa@149-182-13.oke2-bras2.adsl.tele2.no
1236237149 M * ncopa mornin
1236237544 J * Punkie ~Punkie@2a01:5f0:1:80:216:d4ff:fe05:5987
1236238288 M * hijacker_ morning
1236238471 J * kir ~kir@swsoft-msk-nat.sw.ru
1236239382 J * cluk ~cluk@p5B17F820.dip.t-dialin.net
1236239655 J * duckx ~Duck@81.57.39.234
1236239730 J * dna ~dna@53-215-103-86.dynamic.dsl.tng.de
1236240000 M * ghislainocfs2 daniel are you there ?
1236240122 M * ghislainocfs2 daniel: for info dpkg-shlibdeps: warning: dependency on libpthread.so.0 could be avoided if "debian/utilvserver/usr/lib/util-vserver/vhashify" were not uselessly linked against it (they use none of its symbols).
1236240214 J * balbir_ ~balbir@59.145.136.1
1236240244 Q * balbir_ Read error: Connection reset by peer
1236240270 J * balbir_ ~balbir@59.145.136.1
1236241244 J * cga ~weechat@62.196.2.6
1236241700 Q * balbir_ Ping timeout: 480 seconds
1236242427 J * harobed ~harobed@pda57-1-82-231-115-1.fbx.proxad.net
1236242676 Q * duckx Remote host closed the connection
1236243595 J * esa bip@62.123.8.77
1236243825 J * yarihm ~yarihm@whitehead2.nine.ch
1236244035 J * davidkarban ~david@88.86.104.103
1236244105 Q * cluk Quit: Ex-Chat
1236244174 J * ntrs ~ntrs@77.29.1.218
1236244180 Q * Punkie Remote host closed the connection
1236244232 J * Punkie ~Punkie@2a01:5f0:1:80:216:d4ff:fe05:5987
1236244813 J * ntrs_ ~ntrs@77.29.1.218
1236244814 Q * ntrs Read error: Connection reset by peer
1236245196 Q * derjohn_foo Ping timeout: 480 seconds
1236246729 Q * ntrs_ Ping timeout: 480 seconds
1236246939 J * mattzerah ~matt@121.50.217.178
1236246963 M * mattzerah hi guys - i have an issue with networking - hopeing someone can help me out with it
1236246994 M * mattzerah i want my vservers to be on a different ip subnet - but i also want to route traffic to the outside world (through nat is fine)
1236247019 M * mattzerah so i have an interface 192.168.1.200/255.255.0.0 default route to 192.168.1.254
1236247042 M * mattzerah and setup a vsever guest with 10.0.0.1/255.0.0.0
1236247056 M * mattzerah im' using centos 5.2
1236247092 M * mattzerah ive setup net.ipv4.ip_forward = 1
1236247102 M * mattzerah and now i'm at a bit of a loss to even know where to look
1236247107 M * mattzerah can anyone point me in the right direction ?
1236247270 J * derjohn_foo ~aj@51.42.69.80.in-addr.net-lab.net
1236247436 M * fb mattzerah: for a moment forget about guests
1236247479 M * fb mattzerah: set up your networking like there were no guest systems, then assign them desired ips
1236247522 M * mattzerah ok
1236247533 M * fb mattzerah: keep in mind, vserver doesn't do networking
1236247538 M * mattzerah yes, i understand that
1236247563 M * mattzerah i need to setup the routing rules (which is my guess) to allow traffic to route from 10.0.0.0 network through to the 192.168.0.0 network
1236247568 M * fb it only separates interfaces between guests
1236247682 M * mattzerah which i assume i can do through iptables, or not at all - if it cant be done then it cant be done - but i would assume something like iptables would be able to direct the traffic to where it needs to go ?
1236247734 M * fb mattzerah: you can assign multiple ips to one interface
1236247745 M * fb even on different segments
1236247779 M * mattzerah ok, so why does it not seem to work ?
1236247791 M * fb your host has also public ip to which it NATs your traffic?
1236247804 M * mattzerah no
1236247808 M * fb or you have separate router for this
1236247811 M * mattzerah my host is behind a router which does the nat
1236247826 M * mattzerah my host has an internal address of 192.168.1.201 
1236247841 M * mattzerah router = 192.168.1.254 (which is the default route and the nameserver)
1236247861 M * mattzerah setup the guest to have ip = 192.168.1.202 and it works like a charm
1236247861 M * fb and your router knows about the other, 10.0/8 network on the lan?
1236247869 M * mattzerah no it doesn't
1236247886 M * fb and there's your first problem :)
1236247929 M * mattzerah i did setup the host at one stage to do nat, but that didn't work so i have removed all that now
1236247933 M * mattzerah should i be concentrating on that ?
1236247967 M * fb mattzerah: you want to nat this 10.0/8 through 192.168?
1236247991 M * mattzerah i think i have a solution now anyway, thanx for your time fb - i'm just going to add a second interface to the guests that need to access the internet
1236247999 M * mattzerah that will be acceptable i blieve
1236248055 M * mattzerah at least the guests will be able to communicate using 10.x.x.x as they will all be on the same server
1236248079 M * fb they are on the same server :)
1236248113 M * mattzerah yea, only one server at this point in time - i'll worry about traffic between servers when i get big enough to need more than one server :)
1236248150 M * mattzerah although if they are on the same physical lan that shouldn't be an issue anyway should it ?
1236248165 M * mattzerah i'm thinking it must be late for me to miss that - and its only 8pm :/
1236248172 M * mattzerah thankyou for your time fb - much appreciated :)
1236248447 M * fb mattzerah: no, this shouldn't, unless you setup some strange filters or exotic routes :)
1236248467 M * mattzerah not here - i had an ipcop box installed - but that was way overkill for my setup
1236248513 J * cluk ~cluk@p5B17F820.dip.t-dialin.net
1236248530 M * mattzerah plus it slowed the traffic down too much - i know it might be useful to filter all the traffic and have a partitioned network with a safe green zone, and an unsafe orange zone, and a semi safe blue zone etc, but way overkill for my network
1236248555 M * mattzerah all i need is a couple of port forwards to a server or two - everything running over ssh or ssl
1236248572 M * mattzerah i dont want to get TOO paranoid :)
1236250246 N * pmenier_off pmenier
1236250432 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net
1236250972 J * esa` ~esa@ip-87-238-2-45.static.adsl.cheapnet.it
1236250983 Q * esa Ping timeout: 480 seconds
1236251959 N * Bertl_zZ Bertl
1236251964 M * Bertl morning folks!
1236252102 M * hijacker_ afternoon Bertl 
1236252120 M * Punkie hi Bertl, thanks a lot, the patch vs2.3.0.36.8 is working ;)
1236252174 M * Bertl excellent, thank you for spotting the issue!
1236253454 J * Piet ~piet@asteria.debian.or.at
1236253743 M * ard Heh...
1236253750 A * ard read the backlog
1236253753 M * fb morning Bertl  :)
1236253771 M * ard And yes, I wouldn't have spotted it myself (the bug punkie found)
1236253798 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it
1236253872 J * xdr ~xdr@gote2.244.cust.blixtvik.net
1236253941 M * ard Hmmm
1236253955 A * ard thinks his free statement isn't incorrect at all :-)
1236253970 M * Bertl do you use a 2.6..28 kernel?
1236253977 M * Bertl *2.6.28
1236254036 M * ard well, somebody ripped one of the .deb kernels I prepared without asking:
1236254045 M * ard 2.6.28.3-vs2.3.0.36.5-d64-core2
1236254066 M * ard (they usually ask which build I "approved" ;-) )
1236254085 M * ard But at home I might just have that specific kernel
1236254086 M * Bertl so that one _might_ show the issue, if caches goes above the assigned memory
1236254145 M * ard ah :-)
1236254173 M * _Shiva_ Bertl: can you give a short abstract what has changed from vs2.3.0.36.7 to .8? only the delta-fix?
1236254174 A * ard still has to do some assigning first
1236254207 M * ard (and it's needed because a server went down due to a lot of OOMs :-( )
1236254272 M * Bertl _Shiva_: http://vserver.13thfloor.at/Experimental/delta-cached-fix01.diff happened, and it fixes the broken free command (by not showing the cached value) inside a guest with memvirt enabled and limits set
1236254292 J * ktwilight_ ~ktwilight@91.178.146.168
1236254371 M * Bertl (the flag is actually called VIRT_MEM, just for the record)
1236254407 M * _Shiva_ Bertl: ok - so it's not really worth rebuilding/rebooting from .7 to .8 asap, right? :-)
1236254415 M * hijacker_ does it affect versions prior to 2.3.0.36.7?
1236254429 M * Bertl yes, but only on 2.6.28.x kernels
1236254466 M * Bertl and it is a purely cosmetic issue
1236254500 M * hijacker_ also on  2.3.0.36.4 the virt_uptime flag seems to have no effect?
1236254536 M * Bertl that was fixed with 
1236254538 M * _Shiva_ Bertl: apropos the 13th-floor webserver.. may you change the directory listing back to not displaying "folders first" when sorting by mtime? ;-)
1236254546 M * Bertl hijacker_: http://vserver.13thfloor.at/Experimental/delta-uptime-fix01.diff
1236254591 Q * ktwilight Ping timeout: 480 seconds
1236254601 M * Bertl _Shiva_: why? any reason for that?
1236254647 M * _Shiva_ Bertl: 'cause this would imply less scrolling needs to get to the filesbe displayed  :-)
1236254674 M * Bertl huh? the interesting files will always be at the end
1236254677 M * hijacker_ thanks Bertl, is that already applied to the latest devel patches?
1236254688 M * Bertl hijacker_: of course
1236254693 M * hijacker_ excellent
1236254711 M * _Shiva_ Bertl: not if you sort "recent files top-down": http://vserver.13thfloor.at/Experimental/?C=M&O=D
1236254781 M * _Shiva_ Bertl: it *used to be* "folders in-line" display - so one could see new patches right away
1236254792 M * hijacker_ ah _Shiva_ nice once
1236254797 M * hijacker_ *one
1236254813 M * hijacker_ never knew about those ?C=M&O=D stuff
1236254814 M * hijacker_ ;-)
1236254857 M * _Shiva_ hijacker_: note the column titles are clickable? :-)
1236254887 M * hijacker_ _Shiva_, what's a column title?
1236254898 M * hijacker_ you mean the file names?
1236254923 M * _Shiva_ hijacker_: first row: "Name" | "Last Modified" | "Size"
1236254944 M * _Shiva_ hijacker_: those provide sorting by this column
1236254947 M * hijacker_ yeah, spotted that one
1236254948 M * hijacker_ ;-)
1236254952 M * hijacker_ very nice
1236255521 M * _Shiva_ hmm... one my also give a search pattern to the url: http://vserver.13thfloor.at/Experimental/?C=M&O=D&P=*.diff 
1236255564 M * _Shiva_ that's ok with me ;-) Bertl -> CLOSE - resolution: WORKSFORME :-)
1236255592 M * cluk Hi
1236255783 M * cluk Bertl: I have build a vanilla 2.6.27.10 with unmodified vs2.3.0.36.4 applied and I get the same failure as with my ubuntu kernel with adapted patch:
1236255794 M * cluk some details: http://paste.linux-vserver.org/12767
1236256037 M * Punkie I would like to ask...Is here anybody, who use Debian Lenny 64bit (as a guest)? I have strange problem with apache2 and allocating memory. I don't have such problem in lenny 32bit or etch 32/64bit.
1236256249 M * ktwilight_ Punkie, what's the problem?
1236256320 M * Punkie After pure instalations of guest I try install apache2, and I always get (sometimes during instalation, sometimes after instalation) "Cannot allocate memory", "FATAL -> Failed to fork" 
1236256326 J * davidkarban_ ~david@193.85.217.71
1236256394 M * Punkie I never have that problem on 32/64bit etch or 32bit lenny, I tried that with the same settings of memory limit and all other settings.
1236256645 M * ktwilight_ you're running a 32bit lenny and want to have a 64bit lenny guest?
1236256649 M * ktwilight_ Punkie, ^
1236256737 M * Punkie no I have host machine - lenny 64bit, and I have problem with apache2 in lenny 64bit guest
1236256773 Q * davidkarban Ping timeout: 480 seconds
1236256817 Q * davidkarban_ Ping timeout: 480 seconds
1236257028 M * ktwilight_ Punkie, i vaguely remember i had this problem too, but can't remember how i solved it :/
1236257039 J * davidkarban_ ~david@88.86.104.103
1236257111 M * Punkie I solve it...to not use 64bit lenny guest ;) But it is not solution ;)
1236257169 M * ktwilight_ i use 64bit lenny guest too, so... ;)
1236257192 M * ktwilight_ Punkie, any caps?
1236257211 M * Bertl Punkie: what does /proc/virtual/<xid>/limits show for that guest?
1236257219 M * Bertl (after you get the error)
1236257345 M * Punkie http://paste.linux-vserver.org/12768 (VM hits)
1236257501 M * Bertl so, here you go
1236257513 M * Bertl raise the limit and you should be fine
1236257537 M * Bertl 64bit apps use much more VM than 32bit ones (usually)
1236257541 M * Punkie "as" limit?
1236257549 M * Bertl yep
1236257568 M * Punkie I tried to increase "as" number 2x...but it doesn't help
1236257622 M * Bertl remove the limit, check how much the guest actually uses
1236257641 M * Bertl then set the limit twice as high as the maximum
1236257692 M * Punkie ok, I will try that, thanks Bertl
1236257784 M * Bertl _Shiva_: try http://vserver.13thfloor.at/ExperimentalT/
1236257913 M * _Shiva_ Bertl: also ok :-)
1236257981 M * ktwilight_ isn't vm limits by default?
1236257993 M * Bertl there are no limits by default
1236257997 M * ktwilight_ or is this "as" limit a vserver specific?
1236258010 M * Bertl yes, that is a guest limit
1236258013 M * ktwilight_ ah
1236258115 Q * yarihm Quit: This computer has gone to sleep
1236258550 J * ghislainocfs21 ~Ghislain@adsl2.aqueos.com
1236258855 Q * ghislainocfs2 Ping timeout: 480 seconds
1236259215 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1236259619 M * Bertl off for a nap .. bbl
1236259624 N * Bertl Bertl_zZ
1236260852 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1236261609 Q * bono Quit: Leaving
1236261701 J * Inferior ~infer@tor-irc.dnsbl.oftc.net
1236261820 J * geb ~geb@157.4.82-79.rev.gaoland.net
1236262172 J * yarihm ~yarihm@guest-docking-nat-2-091.ethz.ch
1236263461 J * ntrs_ ~ntrs@77.29.1.122
1236263464 J * yarihm_ ~yarihm@guest-docking-nat-1-170.ethz.ch
1236263507 J * mauro__ ~yarihm@guest-docking-nat-1-106.ethz.ch
1236263652 Q * yarihm Ping timeout: 480 seconds
1236263678 J * zbyniu_ ~zbyniu@ip-62.181.188.13.static.crowley.pl
1236263724 Q * zbyniu Read error: Connection reset by peer
1236263801 J * yarihm ~yarihm@guest-docking-nat-2-057.ethz.ch
1236263948 Q * yarihm_ Ping timeout: 480 seconds
1236264008 Q * mauro__ Ping timeout: 480 seconds
1236265885 Q * Piet Remote host closed the connection
1236265955 J * Piet ~piet@asteria.debian.or.at
1236266281 Q * hparker Ping timeout: 480 seconds
1236266479 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1236267308 Q * xdr Read error: Connection reset by peer
1236267329 J * bono bono@118-160-175-240.dynamic.hinet.net
1236267653 Q * Punkie Quit: Punkie
1236268485 Q * mrfree Quit: Leaving
1236269818 Q * kir Quit: Leaving.
1236270000 N * Bertl_zZ Bertl
1236270321 M * jp- on a host machine, /tmp is it's own file system mounted noexec,nosuid, i want to use these same settings for the /tmp file system in a vserver. linux doesn't acknowledge these settings from within the vserver when i use mount /tmp /var/lib/vservers/<vserver>/tmp -o bind,noexec,nosuid
1236270382 M * jp- is there another way i can achieve this functionality?
1236270406 M * Bertl by default the guest gets a separate tmpfs mount
1236270444 M * Bertl at least with nodev, but you can add noexec and nosuid for that too in the guest config fstab
1236270471 M * Bertl inside the geust, you are not allowed to do bind mounts (by default)
1236270477 J * xdr ~xdr@gote2.244.cust.blixtvik.net
1236270493 M * jp- i'll check that out, i want to keep /tmp and /var/tmp out of memory allocated filesystems though, there are cases where we might need to move large sets of data to /tmp for work.
1236270496 M * Bertl and if you are using a newer kernel (since 2.6.25) then this syntax will not work on the host or guest
1236270535 M * Bertl /var/tmp makes sernse to have on a disk, but /tmp?
1236270564 M * jp- they are both bind mounted with the same options just for simplification
1236270567 M * jp- i'll reconsider though.
1236270575 M * Bertl what kernel do you use?
1236270591 M * jp- debian's 2.6.26-1-vserver-686
1236270613 M * Bertl then you need to get very recent util-vserver to do that
1236270705 M * jp- i'll check that out, might file a bug report since using bind doesn't seem to inherit mount options from the intial file system. i think this is a problem, dunno though.
1236270716 M * jp- bug report with the linux guys that is.
1236270720 M * Bertl well, that is intentional
1236270763 M * Bertl after 5 years of thinking how to integrate ro-bind mounts (before 2.6.24, it wasn't supported at all), mainline decided to integrate it this way
1236270794 M * Bertl i.e. you need to do a mount -o bind /path/to/src /path/to/dst
1236270811 M * Bertl and then after that, a mount -o remount,ro... /path/to/dst
1236270884 M * Bertl (options specified on the bind mount are silently ignored)
1236270964 M * jp- seems to pick them up when the bind mount is listed in fstab
1236270978 M * jp- ah well doesn't matter now, i need to rethink this stuff. :)
1236270998 J * jazzanova ~boris@dagny.hypervolume.com
1236270999 M * jazzanova hi
1236271007 M * jp- gonna go have a smoke and ponder on it, thanks for your time
1236271014 M * Bertl np
1236271014 M * jazzanova I got zombie processes in my vserver
1236271017 M * jazzanova parent is init
1236271023 M * jazzanova what do i do ?
1236271029 M * jazzanova I don't want to reboot the host
1236271033 M * Bertl init inside the guest?
1236271036 M * jazzanova yes
1236271057 M * Bertl well, then I'd replace the init process, as it is supposed to reap the zombies
1236271069 M * jazzanova how do I replace it ?
1236271105 M * Bertl that is a long term solution
1236271121 M * Bertl first let's see why you get zombies at all
1236271132 M * Bertl what kernel do you use, and what is in you dmesg
1236271136 M * Bertl (please use paste.linux-vserver.org for everything longer than 3 lines)
1236271162 M * jazzanova  2.6.22-3-vserver-686
1236271178 M * Bertl so _some_ debian kernel
1236271213 M * jazzanova my dmesg is empty on the guest
1236271222 M * Bertl on the host
1236271275 M * jazzanova what am I looking for in the dmesg
1236271291 M * Bertl for 'unusual' messages
1236271316 M * jazzanova vxW: [»fullpliant«,4623:#40004|40004] did lookup hidden dfdcfe6c[#0,4026531918] »/proc/acpi«.
1236271325 M * jazzanova fullpliant is my zombie process
1236271346 M * Bertl okay, but should be harmless
1236271360 M * Bertl no idea why it would check out /proc/acpi though
1236271369 M * jazzanova TCP: Treason uncloaked! Peer 38.108.180.128:39275/80 shrinks window 4191531453:4191534805. Repaired.
1236271383 M * Bertl any stack traces?
1236271406 M * jazzanova no
1236271422 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1236271466 M * jazzanova so fullpliant spawns a bunch of child processes
1236271481 M * jazzanova if the parent fullpliant process dies, the children become inherited zombies by init
1236271499 M * Bertl well, yes, they get reparented
1236271501 M * jazzanova if I restart the vserver it is not going to fix it right?
1236271543 M * Bertl depends, might help, if the guest init process (assumed there is one running) is actually the parent, then it might help to stop it, as the processes will ge treparented to the host init
1236271562 M * Bertl *get reparented
1236271576 M * jazzanova but that is worse
1236271592 M * Bertl init is supposed to reap those orphaned child processes
1236271625 M * jazzanova last time I remember I tried to do something like that, the restart of the vserver got locked up and I had to reboot.
1236271654 M * jazzanova can't I just kill init ?
1236271666 M * Bertl could as well happen, I do not even have a guess what kernel that might be you are using ...
1236271681 M * jazzanova from about a year ago
1236271688 M * Bertl you can try, vkill will allow you to send signals to any process
1236271703 M * jazzanova i shuold not use kill 1 from the guest ?
1236271717 M * Bertl init is protected against such signals
1236271734 M * Bertl you can try, but it won't work :)
1236271753 M * jazzanova ok, so my guest init is reported to hve PID 1
1236271758 M * jazzanova how do I use vkill to kill it ?
1236271774 M * Bertl vkill --help
1236271780 M * daniel_hozac use the real pid.
1236271795 M * Bertl 1 should work too, no?
1236271809 M * jazzanova don't I need to tell it which guest ?
1236271813 M * daniel_hozac when the context is specified, yes.
1236271826 M * Bertl jazzanova: that's the xid
1236271862 M * jazzanova boris@dagny:/var/log$ sudo vkill -c rearden -s 9 1
1236271864 M * jazzanova vkill: vc_ctx_kill(): No such process
1236271893 M * jazzanova boris@rearden:/var/log$ ps aux | grep init
1236271894 M * jazzanova root         1  0.0  0.0   1940   636 ?        Ss   Feb15   4:26 init [2]
1236271905 M * Bertl so, we are back at the question, does your guest have an init process of its own?
1236271911 M * jazzanova oh
1236271920 M * jazzanova guess not :)? how do I know ?
1236271944 M * Bertl if not, then it looks to me like a kernel bug, and is probably fixed since ages
1236271977 M * Bertl check the info about the context
1236272013 M * Bertl actually the status, i.e. cat /proc/virtual/<xid>/status
1236272056 M * jazzanova i'm not sure what is my xid
1236272066 M * jazzanova the name of the host is rearden
1236272072 M * jazzanova but i don't have that folder
1236272079 M * jazzanova i have to vguests on the box
1236272083 M * jazzanova two
1236272139 M * jazzanova ok, i found the context to be 40003
1236272142 M * Bertl the name is not relevant to the kernel
1236272144 M * jazzanova oris@dagny:/etc/vservers/rearden$ cat context
1236272144 M * jazzanova 40003
1236272153 M * Bertl yep, that's the xid
1236272177 M * jazzanova boris@dagny:/etc/vservers/rearden$ cat /proc/virtual/40003/status 
1236272178 M * jazzanova UseCnt: 32453
1236272178 M * jazzanova Tasks:  52
1236272178 M * jazzanova Flags:  0000001602020010
1236272178 M * jazzanova BCaps:  00000000344c05ff
1236272178 M * jazzanova CCaps:  0000000000000101
1236272178 M * jazzanova Spaces: 0c020200
1236272191 M * jazzanova what does that tell me ?
1236272194 M * Bertl daniel_hozac: would it be hard to get some commands to retrieve and parse those settings into vserver ... ?
1236272218 M * daniel_hozac vattribute --get --xid <guest> shows the flags, bcaps and ncaps.
1236272226 M * daniel_hozac s/ncaps/ccaps/
1236272235 M * Bertl one step ahead, as usual :)
1236272296 M * jazzanova     vattribute --set [--xid <xid>] [--bcap [~!]<cap>] [--ccap [~!]<cap>] [--flag [~!]<flag>] [--secure] -- [<program> <args>*]
1236272297 M * Bertl so, your flags tell us that you do not have an init
1236272299 M * jazzanova i don't have get
1236272307 M * jazzanova ok
1236272309 M * Bertl jazzanova: you have ancient utils
1236272323 M * jazzanova can I upgrade the utils without messing up the system ?
1236272342 M * Bertl I guess so, but no need right now
1236272362 M * jazzanova so how can I configure my guest to have its own init process ?
1236272378 M * Bertl by switching from the sysv init style to plain
1236272378 J * balbir_ ~balbir@122.172.62.109
1236272390 M * Bertl (given that there is a guest init installed)
1236272393 J * maio ~maio@217.196.213.98
1236272403 M * maio hi, how experimental is experimental branch? :)
1236272424 M * jazzanova bertl: so how would it know what to run ?
1236272430 M * Bertl maio: kind of stable experimental on the virge of becoming devel
1236272453 M * Bertl jazzanova: that depends on the config, of course there are defaults
1236272461 M * jazzanova do I need to change layout of rc.d stuff ?
1236272493 M * Bertl if your guest init uses different runlevels than you do now, then yes, otherwise it should be fine
1236272506 M * jazzanova no, it is 2
1236272515 M * Bertl maio: i.e. a lot of folks are running the experimental patches in production like environments
1236272533 M * maio Bertl: ok thank you
1236272537 M * jazzanova ok, so what do I do to sitch to plain init ?
1236272597 M * Bertl here is the config http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1236272628 M * Bertl change the 'style' to 'plain'
1236272644 M * Bertl (should be sysv now)
1236272660 M * daniel_hozac or the file doesn't exist.
1236272757 M * jazzanova directory: /etc/vservers/rearden/apps/init containes nothing
1236272770 M * jazzanova do I create a file in there ?
1236272787 M * Bertl yep
1236272835 M * jazzanova ok, but I stil lhave to resolve the problem I have now
1236272840 M * jazzanova do I have no option, but reboot ?
1236272857 M * Bertl probably not, except for changing the xid and ignoring the zombies for now
1236272864 M * jp- don't mean to interupt, can wait til later, but i read something about vserver being considered for inclusion in the default kernel tree, what's the status on that (still being considered)?
1236272886 M * jazzanova i am concerned that my zombies eat the RAM
1236272913 M * Bertl jazzanova: as I said, if that happens on a recent stable or devel kernel I consider it a bug, and it should be reported
1236272925 M * sid3windr being considered? by whom, jp- ?
1236272937 M * Bertl jp-: mainline is re-implementing Linux-VServer from scratch
1236272960 M * sid3windr in tiny bits :)
1236272962 M * Bertl jp-: should be mostly done in 2-3 years
1236272971 M * jazzanova well, I am not on a recent kernel
1236272976 M * jazzanova since I haven't upgradet in a year at least
1236273003 M * jp- ah, well i'm glad debian chooses to support it :p
1236273009 M * jp- 2 to 3 years is a long time.
1236273010 J * dowdle ~dowdle@scott.coe.montana.edu
1236273029 Q * yarihm Quit: This computer has gone to sleep
1236273045 M * jazzanova ok, i am going to reboot
1236273054 M * jazzanova bakc in 5
1236273180 Q * jazzanova Remote host closed the connection
1236273286 J * jazzanova ~boris@dagny.hypervolume.com
1236273289 M * jazzanova i restart the host
1236273297 M * jazzanova I started the vserver
1236273308 M * jazzanova i want to check that i have a separate init in the host
1236273311 M * jazzanova i mean in the guest
1236273313 M * jazzanova boris@dagny:~$ cat /proc/virtual/40003/status 
1236273313 M * jazzanova UseCnt: 99
1236273313 M * jazzanova Tasks:  48
1236273313 M * jazzanova Flags:  0000001402020010
1236273313 M * jazzanova BCaps:  00000000344c05ff
1236273315 M * jazzanova CCaps:  0000000000000101
1236273315 M * jazzanova Spaces: 0c020200
1236273320 M * jazzanova does that tell you anything ?
1236273331 M * jazzanova does it mean I have a separate init ?
1236273358 M * Bertl what does info show, and please us pastebin for longer pastes
1236273387 M * jazzanova info says: Init:   2505
1236273399 M * jazzanova that means the init is at process 2505 ?
1236273407 M * Bertl so, that is a good indication that you have an init running now
1236273422 M * jazzanova so I can do an experiment now
1236273595 M * jazzanova ok, I have a defunct process now
1236273603 M * jazzanova in the guest
1236273611 M * jazzanova do I kill the process 2505 ?
1236273699 M * jazzanova ok, that worked
1236273709 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1236273741 Q * harobed Ping timeout: 480 seconds
1236273745 M * jazzanova but now i tried this:
1236273747 M * jazzanova boris@dagny:~$ sudo vserver rearden restart
1236273748 M * jazzanova vkill: vc_ctx_kill(): No such process
1236273752 M * jazzanova and it gest stuck
1236273775 M * jazzanova boris@dagny:~$ A timeout occured while waiting for the vserver to finish and it will
1236273775 M * jazzanova be killed by sending a SIGKILL signal. The following process list
1236273775 M * jazzanova might be useful for finding out the reason of this behavior:
1236273779 M * jazzanova thats in dmesg
1236273793 M * jazzanova ok, the restart commad killed the vserver
1236273844 M * jazzanova well, looks like this is working, as a temp solution
1236273894 M * jazzanova thanks :)
1236273903 M * Bertl you're welcome! and don't forget to upgrade
1236273913 M * jazzanova i'm not sure how
1236273922 M * jazzanova what kerne do I want ?
1236273940 M * jazzanova I think my host system is too old of a version 
1236273990 M * Bertl either the latest stable branch, which uses 2.6.22.19, or one of the newer 2.6.27/2.6.28 versions
1236274419 Q * jazzanova Quit: jazzanova
1236274425 J * jazzanova ~boris@dagny.hypervolume.com
1236274433 M * jazzanova i remember i had a lot of trouble lasttime upgarding the hos.
1236274529 Q * cga Quit: WeeChat 0.2.6
1236274567 M * Bertl probably very much depends on what distro you use and how you prepare for the upgrade
1236275559 Q * esa` Quit: Coyote finally caught me
1236275591 Q * davidkarban_ Quit: Ex-Chat
1236275791 Q * pmenier Quit: Konversation terminated!
1236275941 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1236276132 J * harobed ~harobed@arl57-1-82-231-110-14.fbx.proxad.net
1236276779 P * pusling hest
1236277291 Q * derjohn_foo Ping timeout: 480 seconds
1236277976 J * duckx ~Duck@81.57.39.234
1236278024 J * ntrs__ ~ntrs@77.29.4.23
1236278460 Q * ntrs_ Ping timeout: 480 seconds
1236278884 M * cluk Bertl: seems as the problem i was seeing with the ubuntu kernel is just caused by broken utils.
1236278928 M * cluk it behaved the same with my ubuntu kernel, a vanilla 2.6.27.10 and vanilla 2.6.27.19
1236278948 M * cluk using a util-vserver package from hardy seems to work.
1236279279 J * Slydder ~chuck@dslb-088-072-213-089.pools.arcor-ip.net
1236279776 J * saulus_ ~saulus@d046163.adsl.hansenet.de
1236279801 M * ktwilight_ cluk, it's best to get .216+
1236279813 M * Bertl cluk: okay, so probably miscompiled or so
1236279831 M * Bertl cluk: do you have the 'vserver-info - SYSINFO' output from those utils?
1236280153 Q * saulus Ping timeout: 480 seconds
1236280492 M * cluk Bertl: http://paste.linux-vserver.org/12767
1236280564 M * cluk in the same pastebin you can see some messages from dmesg:
1236280567 M * cluk fork(): process `vcontext' used deprecated clone flags 0x2000000
1236280586 M * cluk using older working utils the message do not appear.
1236280646 M * Bertl might be interesting to recompile the tools, i.e. looks like old/broken kernel headers or so
1236280685 M * cluk ok, will try that and file a bug report. :)
1236280770 M * Bertl maybe daniel_hozac has more insight what went wrong there
1236281125 M * cluk recompiling did not help.
1236281139 M * Bertl same dmesg entry?
1236281289 M * cluk exactly the same.
1236281294 M * Bertl what kernel headers do you currently use?
1236281484 M * cluk the ones from the patched ubuntu kernel
1236281520 M * Bertl sure about that? I mean the headers in /usr/include/linux
1236281701 M * cluk ah ok. these files are owned by the ubuntu package linux-libc-dev
1236281722 M * Bertl okay, and from what kernel do they originate?
1236281805 M * cluk hm. any easy way to find out?
1236281825 M * Bertl changelogs? maintainer?
1236282005 Q * xdr Quit: Lost terminal
1236282088 M * cluk they are build from the unpatched ubuntu kernel source from which the original ubuntu kernel is build
1236282113 M * Bertl okay, so far so good, what kernel version would that be?
1236282155 M * cluk 2.6.27-11 as ubuntu version, which corresponds to a vanilla 2.6.27.10
1236282187 M * Bertl so that should have the very same clone flags defined as the kernel you ar eusing, right?
1236282193 M * Bertl *are using
1236282237 J * xdr ~xdr@gote2.244.cust.blixtvik.net
1236282239 M * cluk if i did not mess up those clone flags while adapting the patch, yes.
1236282240 M * Bertl daniel_hozac: any idea what goes wrong here? should we remove some clone flags (from the list)?
1236282337 M * daniel_hozac not exactly.
1236282346 M * cluk one thing I noticed: the exact ubuntu version of the intrepid util-vserver package is: 0.30.216~r2750-3ubuntu2
1236282348 M * daniel_hozac i'd have to look at it closer.
1236282393 M * daniel_hozac that's a known broken version.
1236282431 M * Bertl okay, let's try to update to the latest snapshot and see if that works
1236282449 M * Bertl (and produces the same dmesg line)
1236282516 M * cluk ok. i will try this tomorrow morning and report back. thanks for your excellent support. :)
1236282526 M * cluk bye
1236282528 M * Bertl you're welcome! 
1236282560 P * cluk Ex-Chat
1236282727 M * Bertl btw, the clone flag would be CLONE_STOPPED, which is definitely suspicious
1236283059 J * qozumeryj ~qozumeryj@193.43.249.169
1236283263 M * daniel_hozac yeah, that has no business being there.
1236283268 M * daniel_hozac i wonder where it's coming from.
1236283411 Q * Pazzo Quit: Ex-Chat
1236283435 M * maio OT question: how should I create duplicate of my (live) root partition so I can re-create it on another computer (so I can test upgrade process there)
1236283463 M * Bertl the rsync build method should work for that
1236283475 M * Bertl ah, you mean a real system?
1236283586 M * maio server which I want to clone is running (and using that root partition)
1236283622 M * Bertl okay, root partition here means a partition assgined/dedicated to that guest, yes?
1236283657 M * Bertl we are having a little nomenclature issue here, it seems :)
1236283674 M * maio no I'm talking about host :) (that's why it is Off Topic :)
1236283689 M * Bertl so you want to copy a running host to a new machine, yes?
1236283756 M * maio yes so I can test upgrade process. host is running kernel 2.8.6 + som old vserver patch, software raid, XFS, and some other strange things :)
1236283757 Q * jazzanova Quit: Leaving
1236283773 M * maio 2.6.8 :)
1236283782 M * Bertl okay, you are booted (on the new machine) from some CD?USB or so?
1236283788 M * Bertl *CD/USB
1236283836 P * neofutur 
1236283895 M * maio my thoughts were that I'll try it inside virtualbox/vmware (but I also have real machine)
1236283955 M * Bertl okay, basically the best approach is to recreate an identical partition scheme, then create the filesystems, and either rsync or dump/restore them
1236283982 M * Bertl you basically start with a 'crashed' system after that, which should recover, and you can then test the upgrade procedure
1236284047 M * maio ok I'll try rsync way. btw thanks for idea of booting from some CD (live/rescue) :)
1236284278 M * Bertl np
1236284414 Q * gnuk Quit: NoFeature
1236284871 J * yarihm ~yarihm@guest-docking-nat-1-017.ethz.ch
1236285462 Q * yarihm Quit: This computer has gone to sleep
1236285615 J * kwowt ~quote@pomoc.ircnet.com
1236285618 M * kwowt yo
1236285673 M * kwowt :>
1236285707 M * kwowt what option should i use with rsync when moving a guest vserver on a machine with slow cpu (high loads)?
1236285739 M * kwowt *from
1236286135 M * daniel_hozac don't use -z is probably the only thing you have to gain.
1236286439 Q * Slydder Remote host closed the connection
1236286671 Q * dowdle Remote host closed the connection
1236287292 J * Slydder ~chuck@dslb-088-072-213-089.pools.arcor-ip.net
1236287397 Q * qozumeryj Remote host closed the connection
1236287469 J * feditezyn ~feditezyn@193.43.249.169
1236287682 Q * Inferior Quit: Leaving
1236287764 J * dowdle ~dowdle@scott.coe.montana.edu
1236287922 Q * feditezyn Remote host closed the connection
1236287959 Q * bonbons Quit: Leaving
1236292411 J * dwery ~azummo@81-208-60-204.ip.fastwebnet.it
1236292442 M * dwery 'morning. do some kind of patch against 2.6.29-rc7 exists?
1236292464 M * Bertl not yet, feel free to do a port
1236292486 M * dwery :)
1236292497 M * kwowt :>
1236292513 M * mugwump any plans for a git repo for vserver Bertl ?
1236292526 M * dwery I believe this has something to to with the recent change sin the network subsystem
1236292568 M * Bertl mugwump: as soon as there are two other Linux-VServer kernel developers using git, I'll switch to git (at least for the commits)
1236292607 M * daniel_hozac lol.
1236292620 M * daniel_hozac i thought we only had one kernel developer? :-)
1236292644 M * sid3windr did you now!
1236292645 M * Bertl yeah, well, if they pop up somewhere, I'm all for reason :)
1236292648 M * mugwump it's not really about that, is it?  it's about getting work from the project moving into the mainstream kernel
1236292657 M * mugwump instead of it permanently staying a ghetto
1236292667 M * daniel_hozac git really doesn't help with that.
1236292675 M * Bertl mugwump: ah, that is a lost cause, but feel free to pursue it
1236292706 M * mugwump but it has to start at the source ... individual changes need the commenting
1236292708 M * Bertl mugwump: and keep in mind that the kernel might be git based, but the development and submission is patch based
1236292716 M * mugwump I know that
1236292726 M * mugwump but it's incremental, well described patches which are important
1236292739 M * mugwump ie, work that is reviewable
1236292741 M * daniel_hozac not really.
1236292755 M * daniel_hozac the deltas need to be described.
1236292757 M * Bertl not that git would help in any way there
1236292761 M * daniel_hozac i.e. the split set.
1236292776 M * daniel_hozac you don't want the split set + every single change made since then.
1236292792 M * daniel_hozac since, well, mainline submissions take forever and ever.
1236292797 M * mugwump you're using quilt right now are you?
1236292804 M * mugwump to keep those split sets
1236292808 M * mugwump and you manage the split sets?
1236292846 M * mugwump the tool does help achieve the process, but if the process isn't followed then no, the tool doesn't help
1236292971 M * mugwump daniel_hozac: the split set + every single change made since then would be better than one big set
1236292982 M * mugwump because it is maintainable.  *if* the descriptions were accurate
1236292993 M * mugwump that's a tree which could get merged
1236292999 M * mugwump a big patch, never
1236293008 M * daniel_hozac no.
1236293015 M * daniel_hozac you want changes recorded in the split set.
1236293023 M * daniel_hozac that's something you can merge.
1236293039 M * mugwump ok... so why not publish the split set using git?
1236293065 M * daniel_hozac i don't think Bertl would mind if you did.
1236293089 M * mugwump I can't come up with the descriptions for the changes
1236293093 M * dwery I've got some experienc in pushing things to the kernel. If there's the intention to do that, I might be of some help
1236293155 M * mugwump well, if the split out patch for a current version is published somewhere, importing that to git would be a good start
1236293332 J * yarihm ~yarihm@77-56-182-18.dclient.hispeed.ch
1236293388 M * mugwump Bertl: are you able to publish that?
1236293392 M * Bertl dwery: I'm not against it, but we did that twice
1236293404 M * dwery Bertl: I don't know the story, what happened?
1236293441 M * mugwump well, one time I actually submitted a few patches to add the API interface as a starting point
1236293452 M * mugwump but people wanted to solve the core problems cleanly first
1236293460 M * mugwump hence, PID namespaces, IPC namespaces, etc
1236293486 M * mugwump anyway my submission was considered to be not solving any of the interesting problems
1236293486 M * Bertl dwery: there was one efford back 4 years or so, together with rik van riel (when he 'loved' Linux-VServer :)
1236293503 M * dwery well the namespace thingy should be almost solved in mainstream, isn't it?
1236293514 M * Bertl dwery: the tag line was, a lot of stuff has to be cleaned up first
1236293528 M * Bertl (not in our code that is, in mainline)
1236293535 M * mugwump AFAIK it's core and Bertl has ported to use the new namespaces
1236293570 M * Bertl took mainline to clean up their stuff 3+ years, then they 'considered' OS level virtualization again
1236293575 M * dwery :D
1236293583 M * dwery you know.. we are all slow :)
1236293583 M * Bertl we did a code cleanup with OLPC, which was using Linux-VServer for their security
1236293607 M * mugwump I think what it will take is a shift in development approach though, to start using topgit or stgit so that reviewable patch series are easily extractable
1236293618 M * dwery mugwump: from my experience, it doesn't matter
1236293625 M * dwery I always used quilt
1236293626 M * Bertl eventually OLPC went down the river and dropped their security concept (and most of the other concepts :) 
1236293655 M * mugwump and to make sure that new changes are slotted into what remains as a large patch submission queue
1236293655 M * dwery andrew, in fact, hates git ;)
1236293675 M * mugwump right.  but *most* kernel developers and teams are using it
1236293682 M * dwery that's true
1236293696 M * mugwump and there are big benefits to that.  it's a standard way of exporting your development progress
1236293698 M * Bertl now, a 'task force' around Eric Biederman and a bunch of folks from IBM try to re-implement Linux-VServer and OVZ in the way the kernel should be
1236293721 M * mnemoc .oO
1236293762 M * Bertl but their focus is on snapshoting, not on lightweight virtualization
1236293796 M * dwery well, from what I can see the code is of high quality and appropriate for the kernel
1236293804 M * Bertl dwery: what we do is keep it light and simple, unintrusive and highly performant
1236293804 M * dwery you did an amazing wor Bertl
1236293828 M * mugwump the snapshotting stuff was more about making sure that the virtualization was done right
1236293848 M * dwery even checkpatches.pl agrees :)
1236293855 M * dwery 325 error on 25K lines of code
1236293886 M * dwery I once rewrote a driver from Garmin.. when I had finished it was 50% the code size with more functionality :)
1236293900 M * dwery and checkpatches had a lot of things to say about it
1236293903 M * dwery :)
1236293929 M * Bertl around the time when we did the OLPC cleanup, we were down to like 5 or so
1236293951 M * mugwump there's nothing which says that making sure every identifier you return to userspace can be virtualised is going to negatively impact performance
1236293955 M * mugwump at least, not tangibly
1236293967 M * mugwump that's what they were getting at
1236293983 M * mugwump but in any case, that's not important... that's what they're going for with namespaces and I agree with that goal
1236293991 M * Bertl the kernel has already gotten a lot slower by mainline virtualization
1236294000 Q * ntrs__ Ping timeout: 480 seconds
1236294005 M * mugwump the point is that vserver is continually rebasing on top of the mainline kernel
1236294006 M * dwery and vserver si an optional component
1236294026 M * mugwump and this rebasing is not being exported via a tree on git.kernel.org
1236294032 M * dwery If you guys want to try again I'd be happy to add my sign off.. it wouldn't weight too much but it still would be 0.02 grams more
1236294074 M * Bertl I'm fine with that, I'll even do the breakdown and/or integration of follow up patches, but I won't do the pushing
1236294099 M * mugwump topgit is very nice, it just doesn't support re-ordering of patches
1236294106 M * mugwump stgit is more quilt-like
1236294117 M * Bertl i.e. I do not want to spend endless hours discussing why we should use file based config instead of the syscall command switch or something like that
1236294130 M * dwery Bertl: tat would be my work :)
1236294263 M * dwery but I need to know if there's someone among the main kernel developers which is against vserver, either with a reason or without
1236294286 M * Bertl I think Andrew wasn't very fond of it, but I have no idea why ...
1236294316 M * dwery mmm that's not a good starting point :)
1236294349 M * dwery well, let me have a few chats here and there....
1236294426 M * Bertl maybe my impression was wrong ...
1236294479 M * Bertl towertech ... interesting
1236294490 Q * Slydder Quit: Leaving.
1236294497 M * dwery having Andrew agree on it would be very useful
1236294535 M * Bertl I think in any case you are up to a big lobby around OVZ, as they definitely want to push their patches into mainline (SWsoft that is)
1236294571 M * Bertl s/to/against/
1236294574 M * dwery Bertl: how far are they from that?
1236294611 M * Bertl till now, nothing got accepted, everything was reimplemented, although they have some copyrights in the networking and cgroup stuff
1236294648 M * mugwump well, in principle the two projects could be sharing patches
1236294649 M * Bertl not that the accepted namespaces would differ from what Linux-VServer was doing for 5 years or so
1236294658 M * mugwump and collaborating on them ... very similar projects etc
1236294667 M * Bertl mugwump: have you looked at the OVZ code? ever?
1236294698 M * mugwump not with the same fine eye for detail as yourself of course, but yes
1236294711 M * mugwump sure there are some big design differences
1236294731 M * Bertl then I'm really surprised that you even consider combining that bloat with Linux-VServer 
1236294808 M * Bertl http://vserver.13thfloor.at/Stuff/slide-page16.png
1236294868 M * mugwump look, I don't care about the extra stuff which they won't get merged
1236294892 M * mugwump but the stuff in common is slowing becoming common because namespaces are getting merged
1236294986 M * Bertl yes, and thus we pick up the namespaces wherever they make sense
1236295370 M * mugwump ok.  well, some more split out patches would help someone willing like dwery pick up the work, but the thing is that for new changes they will need to be documented too
1236295390 M * mugwump so that the submission maintainer can decide whether they get a new patch or are rolled into an existing one
1236295414 M * mugwump without that co-ordination, it will be difficult to achieve success..
1236295441 J * derjohn_foo ~aj@p5B23CD38.dip.t-dialin.net
1236295460 M * dwery agreed.
1236295492 Q * geb Quit: Quitte
1236295580 Q * yarihm Ping timeout: 480 seconds
1236295649 M * Bertl the split patches are part of the release/devel cycle, so they will happen when we approach a stable release for 2.6.27+
1236295667 M * Bertl (which shouldn't take too long IMHO)
1236295762 M * mugwump dwery, do you already have a git.kernel.org tree?
1236295796 A * dwery hates git :) I use git only to checkout the tree and then either stgit or quilt
1236295850 M * Bertl dwery: I'm using something very git like, basically hard linked kernel trees plus a bunch of tools/scripts I wrote a few years ago
1236295869 J * sebastiandeutsch ~sebastian@mue-88-130-113-187.dsl.tropolys.de
1236295885 M * dwery I made a quick count and a stripped down vserver patch, touching only the linux core would be around ~90 files (not counting the includes)
1236295912 M * mugwump it's not how you manage it, it's how it's published for review
1236295923 M * sebastiandeutsch hello, I created a vserver and entered it, when I want to ping the internet ping says me icmp open socket: Operation not permitted.... even if i ping 127.0.0.1
1236295935 M * mugwump assuming whatever you are doing is equivalent to or better than git, then you can easily export it to git
1236295964 M * daniel_hozac sebastiandeutsch: sounds like old utils on a recent kernel.
1236296010 M * sebastiandeutsch daniel_hozac: so I have to update utils and recreate the vserver again?
1236296067 M * dwery Bertl: so you can easily get out patches a-la quilt way?
1236296087 M * Bertl to some extend, see the split patches
1236296128 M * Bertl http://vserver.13thfloor.at/Experimental/split-2.6.18.2-vs2.1.1/
1236296132 M * Bertl (as an example)
1236296142 M * daniel_hozac sebastiandeutsch: just restarting it is enough.
1236296143 M * dwery Bertl: nice.
1236296236 M * mugwump those are mostly split by filename, right?
1236296365 M * Bertl mugwump: nope, they are split by subsystem or logical entity they apply to
1236296365 M * sebastiandeutsch daniel_hozac: noob question - how do I update util-vserver (ps my kernel is 2.6.26-bpo.1-vserver-686)
1236296395 M * mugwump ok, but they're not currently incremental
1236296400 M * mugwump ?
1236296532 M * sebastiandeutsch ok got latest version from backports (sorry no lenny around)
1236296588 M * Bertl mugwump: how could unrelated subsystems/logical entities be incremental?
1236296592 M * sebastiandeutsch daniel_hozac: thanks that was it
1236296613 M * mugwump Bertl: what I mean is the standard that no patch introduces a regression or depends on a subsequent patch
1236296640 M * mugwump and does one unit of reviewable change
1236296673 M * Bertl that's hard to do, as you would need to build dummies and stubs and similar
1236296733 M * mugwump right.  which is why it helps if such work is fed back into the upstream process as far as possible..
1236296748 M * Bertl go ahead :)
1236296805 M * mugwump I'll be happy to help ... but it needs someone really interested in the inner workings of the patch involved too
1236296980 M * maio any idea what should I do when I changed my raid devices from /dev/hdX to /dev/sdX? (I can't boot now because there is still /dev/hdX in initrd)
1236297115 M * Bertl change the initrd
1236297137 M * Bertl initrd is a misnomer nowadays, it actually is an initramfs (google for that)
1236297156 M * Bertl you can unpack it with cpio, do some changes and repack it into your new initramfs
1236297419 M * maio # cpio -i < /boot/initrd.img-2.6.8+vserver    throws some error at me :/
1236297445 M * Bertl try zcat /boot/initrd.img-2.6.8+vserver | cpio -i (in a dir)
1236297530 M * maio "not in gzip format"
1236297564 M * maio file /boot/initrd.img-2.6.8+vserver  -> Linux compressed ROM file