1138406584 J * michal_ ~michal@www.rsbac.org
1138411818 Q * Johnnie Remote host closed the connection
1138411853 J * Johnnie ~jdlewis@dynamic-acs-24-154-53-16.zoominternet.net
1138412469 Q * cohan Ping timeout: 480 seconds
1138413971 Q * anarcat Quit: bonne nuitte!
1138414154 N * Bertl_zZ Bertl
1138414173 M * Bertl back now ... had a good rest ...
1138414360 J * cohan ~martin@p54AD32F9.dip0.t-ipconnect.de
1138414376 N * cohan cohan_zzz
1138414404 M * Bertl good night cohan_zzz!
1138414473 M * cohan_zzz thx - just thinking if we/you should provide a patched version for 2.6.15(.1), so ppl trying development with 2.6.15 get a working version?
1138414497 M * Bertl 2.1.0.8 is out, and should fix that, no?
1138414523 M * cohan_zzz (because the wiki mainpage shows 2.6.15 with 2.1.0.5 - should this just be updated to 2.1.0.8?)
1138414569 M * cohan_zzz ah, okay, got confused - i'll try it one more "clean" time with 2.6.15.1, patched with most current 2.1.0.8
1138414712 M * Bertl yes the wiki probably needs updating
1138414743 M * cohan_zzz if you are busy, i'll fix this the next minutes (as good as i can)
1138414745 M * Bertl (will do that with the release of 2.1.0.9
1138416266 M * cohan_zzz Bertl, how should a minimalistic changed patch of 2.1.0.5 which will cleanly apply to 2.6.15.1 be called ? 2.1.0.5p1 ?
1138416294 J * jkl eric@c-67-173-248-142.hsd1.co.comcast.net
1138416327 M * Bertl cohan_zzz: for example ...
1138416332 M * Bertl welcome jkl!
1138416339 M * jkl hey bertl
1138416361 M * jkl i had to take my box down
1138416377 M * Bertl why that?
1138416387 M * jkl for some reason the partition that holds my vservers thought it was mounted ro
1138416396 M * jkl even tho it said it was mounted rw
1138416398 M * Bertl ah, filesystem errors
1138416412 M * Bertl they lead to a fielsystem remount ro
1138416422 M * jkl oh really
1138416436 M * jkl well that is what mustve happened
1138416437 M * Bertl usually .. i.e. if the disk gets corrupted and such
1138416452 M * Bertl then the 'better' filesystems switch to ro mode
1138416454 M * jkl there was no way for me to unmount it, even after i shut all the vservers down
1138416476 M * jkl so i had to reboot it
1138416491 M * jkl and i was wicked drunk when this happened
1138416495 M * Bertl you should definitely do a filesystem check there
1138416500 M * jkl i was in no condition to be fooling around
1138416501 M * jkl haha
1138416521 M * Bertl I've heard that happens sometimes :)
1138416526 M * jkl yeah i think this box needs some major maintenance
1138416529 M * jkl hehe
1138416599 M * Bertl cohan_zzz: do you think there is demand for patches against 2.6.15.1?
1138416702 M * Bertl jkl: but I guess after the reboot, everything seems/seemed fine again
1138416764 M * cohan_zzz Bertl: well, do you think everybody uses rc1 already? i don't think there's a big demand, but right now, there is no (but 2.1.0.4?) version you could "savely" suggest to use with 2.6.15
1138416768 M * jkl yeah, i would hope that after a reboot the disk got checked for errors
1138416783 M * jkl but it might not have, so on sunday ill take it down for awhile and check all the disks
1138416792 M * jkl maybe even tonight
1138416803 M * jkl provided i dont get wasted again :/
1138416805 Q * Skram Ping timeout: 480 seconds
1138416869 M * Bertl cohan_zzz: well, hmm, let's do a 2.1.0.5.1 then
1138416911 M * cohan_zzz ookay, ready in a second
1138417149 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15.1-vs2.1.0.5.1.diff
1138417163 M * Bertl (contains the cleanups and the fix)
1138417337 M * cohan_zzz hm, you are faster then me, i just uploaded "my" version to https://servy.dynip.yawsp.de/~install/patch-2.6.15.1-vs2.1.0.5.1.diff
1138417401 M * Bertl well, I have more practice I guess
1138417511 M * cohan_zzz ;) almost identical, but i added an Makefile~, what a shame
1138417576 N * ebiederm_zZ ebiederm
1138417592 M * Bertl morning ebiederm!
1138417618 M * ebiederm Morning.
1138417638 M * ebiederm Have I mentioned that switch_exec_pids is evil?
1138417698 M * Bertl not that I remember, but I can imagine ...
1138417832 M * ebiederm I have just spen several hours building and testing a couple of simple obviuosly correct patches so it won't be in my way.
1138417899 M * Bertl how do the cross compilers work?
1138417921 M * ebiederm Unfortunatley I haven't quite gotten that far.
1138418072 M * ebiederm I am in the final stages of getting a task_reference implementation that won't pin gobs of memory, and works for process groups.
1138418093 M * Bertl sounds good
1138418349 M * ebiederm The final piece of the puzzle has been smacking switch_exec_pids around so it won't get in my way.
1138418406 M * ebiederm Because even if I am exposing a task_struct I need to update the pointers when entries are added and removed from the pid hash table.
1138418509 M * ebiederm It looks like you had a long productive debugging session.
1138418534 M * Bertl yeah, well, stupid off by one issue
1138418556 M * Bertl that's the problem if you can't count to six ...
1138418582 M * cohan_zzz and i thought i am just to dumb to get a vserver running properly... ;)
1138418608 M * Bertl cohan_zzz: and now you're famous :)
1138418631 M * cohan_zzz i cannot believe that yet ;) i am feeling honoured
1138419002 M * Bertl cohan_zzz: btw, what does the zzz mean in your nick, are you dreamwalking right now?
1138419059 M * cohan_zzz you are right - i am not nocturnal yet - i wantet to go to sleep an hour ago, will do so now, thx
1138419091 M * Bertl good night then, and thanks again!
1138419169 Q * gerrit Ping timeout: 480 seconds
1138419347 J * gerrit ~gerrit@c-24-22-19-162.hsd1.or.comcast.net
1138421727 N * ebiederm ebiedermOo
1138422871 J * shuri ~shuri@64.235.209.226
1138422886 M * Bertl welcome shuri! LTNS
1138422950 M * shuri hello Bertl Yes LTNS
1138422959 M * shuri very busy :)
1138422972 M * Bertl what's up?
1138422976 M * shuri but always a pleasure to talk to you
1138422984 M * Bertl the pleasure is mine ..
1138423018 M * shuri work,music,chicks :)
1138423052 M * shuri i read about a bug in dev branch
1138423055 M * Bertl no computers?
1138423063 M * shuri work=computer
1138423064 M * shuri :)
1138423073 M * Bertl yeah, well, stupid bug, but it's already fixed
1138423101 M * shuri great
1138423109 M * shuri will update then
1138423288 M * shuri got only one box with dev anyway
1138423342 M * Bertl and, how is it doing?
1138423383 M * shuri well nothing in it 
1138423410 M * shuri i try to stay up2date but 
1138423435 M * shuri i am not upgrading system that work fine and are stable for now
1138423448 M * Bertl makes sense
1138423467 M * shuri 2.4.30-vs1.2.10 #1 SMP Thu Apr 28 19:25:20 AST 2005 i686 i686 i386 GNU/Linux
1138423480 M * shuri stable as rock!
1138423490 M * Bertl that's how it is supposed to be!
1138423501 M * shuri 20 vservers
1138423521 M * shuri Tasks: 400 total,   1 running, 398 sleeping,   0 stopped,   1 zombie
1138423523 M * shuri lol
1138423732 M * shuri and you what's up?
1138423789 M * Bertl well, a lot of new features (in devel), eric is working on gettings parts in mainline (virtualization)
1138423872 M * shuri any news of Jack G.?
1138423887 M * Bertl not really, I saw he released newer tools some time ago
1138423960 M * shuri  Virtual private servers 	0.40	2005-08-10 15:24:2
1138423983 M * Bertl yes, but he didn't even announce them on ML or wiki
1138424014 M * shuri IPROOT=dhcp
1138424047 M * shuri humm is this working with utils-vservers?
1138424070 M * Bertl well, it's not implemented, but it should be quite easy
1138424080 M * Bertl you need to use a 'guest' identifier
1138424102 M * Bertl (as the MAC will be the same as the host)
1138424128 M * shuri why a guest?
1138424141 M * shuri switchs should do this...
1138424159 M * shuri humm
1138424163 M * shuri MAC will be the same as the host
1138424167 M * Bertl as all guests share the network device, it will be the same MAC
1138424169 M * shuri thats right...
1138424215 M * Bertl so unless you put in a card per guest, you'll have to specify the guest when requesting an ip via dhcp
1138424265 M * Bertl kind of defeats the purpose, but that's just MHO
1138424364 M * shuri maybe is time for network virtualisation
1138424378 M * Bertl we are working on it :)
1138424407 M * shuri gnet?
1138424420 M * Bertl yes ngnet
1138424466 M * shuri it work?
1138424488 M * Bertl it will work .. for now just proof of concept and brainstorming
1138424525 M * shuri k
1138424597 M * shuri this is weird Jacques releases new tools  
1138424608 M * shuri i don't understand
1138424610 J * Smutje ~Smutje@xdsl-87-78-41-206.netcologne.de
1138424617 M * shuri hi Smutje 
1138424620 M * Bertl shuri: me neither ...
1138424719 Q * Smutje_ Ping timeout: 480 seconds
1138424756 M * shuri Bye Smutje 
1138424909 M * Bertl we have a new (much better) SMP scheduler now ...
1138424947 M * shuri yes i know
1138424961 M * Bertl ah, you are up to date .. good ... :)
1138425068 M * shuri member of the maillist :)
1138425193 J * bwana ~bwana@71-82-125-40.dhcp.roch.mn.charter.com
1138425200 M * bwana hello people
1138425206 M * shuri hi bwana 
1138425211 M * Bertl welcome bwana !
1138425220 M * bwana thanks Bertl
1138425250 M * bwana question about vserver, what is the 'default' dir that the vservers will be in on the system
1138425263 M * bwana $VROOTDIR
1138425277 M * Bertl depends on your installation, but vserver-info - SYSINFO will tell you
1138425281 M * bwana what is default and has it changed in preavious implementations
1138425314 M * shuri defaults should be /vservers
1138425325 M * shuri with ./configure
1138425335 M * bwana is /vservers is what is reccomended/best practice
1138425342 M * Bertl well, some distros think different
1138425350 M * bwana i see
1138425353 M * Bertl but the canonical path is /vservers
1138425407 M * bwana thanks
1138425423 M * Bertl you're welcome!
1138426144 Q * Johnnie Quit: G'bye!
1138426670 M * bwana so second question, at what version will you say,  'this is really really good and will be the new stable release' for the kernel patch and the tools?
1138426715 M * Bertl tool are up to enrico, kernel side, well, I guess a new stable release will happen in a few weeks
1138427648 M * jkl hmpf
1138427668 M * jkl Bertl: any good advice on diagnosing trouble entering a vserver?
1138427706 M * jkl i know it's running but it doesnt show up in vserver-stat
1138427819 M * Bertl jkl: sure about that?
1138427825 M * jkl hehe nope
1138427835 M * jkl the IP is up
1138427838 M * jkl and i can ping it
1138427846 M * Bertl well, that's nothing unusual
1138427855 M * jkl yeah
1138427859 M * Bertl add an ip to your network card, and the host will reply
1138427891 M * Bertl in general, if vserver-stat doesn't show it, it's not there
1138427908 M * jkl ok, what can i do to try and bring it back to life
1138427922 M * Bertl well, usually start it ...
1138427930 M * jkl heh, tried taht
1138427936 M * Bertl and? what is the result?
1138427953 M * jkl well, it looks like it's starting as normal
1138427962 M * jkl but then it doesnt show up in vserver-stat
1138427993 M * Bertl maybe nothing is started in the guest?
1138428009 M * Bertl i.e. nothing that outlives the startup process
1138428032 M * jkl ohh
1138428049 M * jkl so if nothing is getting started it doesn't continue to run
1138428064 M * Bertl unless you use persistant contexts
1138428071 M * jkl which is above me
1138428071 M * Bertl (so by default, no)
1138428077 M * jkl ok i fixed it
1138428087 M * jkl thanks, i didn't realize it followed that type of behavior
1138428092 M * jkl (by default)
1138428121 M * Bertl also depends on the installation (i.e. init style)
1138428129 M * jkl yeah, it's gentoo
1138428146 M * jkl very frustrating - when a vserver dies and isn't stopped
1138428150 M * Bertl hmm, IIRC, that should bring it's own init
1138428165 M * jkl you try to start it again, but gentoo thinks everything is already started
1138428185 M * jkl so i have to chroot in, and zap all the startup services and reset them manually to stopped
1138428207 M * Bertl hmm, IIRC, Hollow's layout/stuff should do that too ...
1138428213 M * jkl really?
1138428214 M * jkl hmm
1138428228 M * Bertl anyway, you now figured how to fix it, so it should not be a problem
1138428236 M * jkl yeah
1138428255 M * jkl maybe if i'm ever here when he's here ill ask him about it
1138429025 M * Bertl should be here this evening, I'd say
1138431208 M * Hollow jkl: still around? ;)
1138431236 M * Bertl good morning Hollow!
1138431246 M * Hollow morning Bertl
1138431255 M * Hollow oi, 2.1.0.8
1138431287 M * Bertl yeah, seems we need a devel release soon :)
1138431343 M * jkl ahhh
1138431345 M * jkl hello
1138431350 M * Hollow heya ;)
1138431367 M * jkl so i had udev in my world file
1138431371 M * jkl in the guest
1138431380 M * Hollow yeah, strange enough anyway ;)
1138431388 M * jkl i just removed that and things seem to be working in that aspect
1138431392 M * jkl i have a general question tho
1138431415 M * jkl when a vserver goes down, and I haven't explicitly stoped it
1138431426 M * jkl the next time i start it, none of the services start
1138431441 M * jkl i have to chroot in and zap everything to reset them to stopped manually
1138431444 M * Hollow jkl: which init style? which baselayout-vserver version?
1138431445 M * jkl is there anyway to get around that
1138431450 M * jkl gentoo
1138431478 M * jkl and the baselayout ver is ... 
1138431510 M * jkl hm, on this one it's 1.0 
1138431516 M * jkl i see 1.12 is available
1138431519 M * Hollow 1.0?
1138431530 M * jkl sys-apps/baselayout-vserver
1138431530 M * jkl       Latest version available: 1.12.0_pre8-r2
1138431530 M * jkl       Latest version installed: 1.0
1138431533 M * Hollow never saw a baselayout with this version
1138431535 M * Hollow wth
1138431556 M * Hollow hm
1138431557 M * jkl hmpf
1138431558 M * jkl hehe
1138431559 A * Hollow goes looking
1138431575 Q * gerrit Ping timeout: 480 seconds
1138431577 M * Hollow but you should update to 1.11.4 in the stable case, and 1.12.* in the devel case ;)
1138431585 M * Hollow and set the init style to plain
1138431592 M * Hollow and all your init problems should go away
1138431593 M * jkl plain eh?
1138431609 M * Hollow yeah, the gentoo init style is obsolete
1138431616 M * jkl ohh
1138431620 M * jkl where do i change that?
1138431632 M * Hollow  /etc/vserver/<name>/apps/init/style
1138431650 M * jkl and that's all i have to do, it shouldnt screw anything else?
1138431690 M * Hollow yeah, just update baselayout-vserver in your guest(s) and use plain style.. probably stop the guests before chaning the init style
1138431749 M * jkl when did gentoo become obsolete
1138431768 M * Hollow hm.. about half a year ago?
1138431777 M * Hollow or maybe even longer
1138431785 M * Bertl btw, why was the gentoo init style dropped?
1138431799 M * Bertl I mean, it would have been the pendant to sysv, no?
1138431808 M * jkl hm, i dunno if i like plain - it doesnt show all the services starting like gentoo does
1138431825 M * Hollow well, yeah, but it causs many headaches du to the init script dependancy cache
1138431835 M * jkl vkill: vc_ctx_kill(): No such process
1138431836 M * jkl em
1138431846 M * Bertl Hollow: is that unfixable or just too much work?
1138431848 M * jkl thats not pretty :(
1138431909 M * jkl i get that error when i try to stop a vserver with initstyle plain
1138431922 M * jkl looks like it stopped though
1138431928 M * Hollow Bertl: hm, it may be fixable, but needs heavy testing i guess, we once supported it in baselayout-vserver but it always revealed new bugs
1138431954 M * Hollow jkl: did you change the init style before stopping?
1138431968 M * Bertl guess the folks who want lightweight guests (like jkl) with startup/shutdown output would do the testing, no?
1138431989 J * gerrit ~gerrit@c-24-22-19-162.hsd1.or.comcast.net
1138431994 M * Bertl wb gerrit!
1138432016 M * jkl Hollow: well, i changed it before i stopped it, but then changed it back, stopped it, changed it and then started it
1138432016 M * jkl hehe
1138432031 M * Hollow did you update baselayout-vserver too?
1138432053 M * jkl well, on the vserver that im testing on, it already has the updated baselayout
1138432062 M * jkl some of the others have the older one
1138432068 M * Hollow ic
1138432086 M * jkl what does that error mean?
1138432101 M * Hollow the context was already gone when util-vserver tried to kill it
1138432175 M * jkl hm, same message on another after changing the init style as well
1138432278 M * Hollow jkl: willing to try patches for gentoo init style?
1138432296 M * jkl sure,
1138432311 M * jkl will it affect all my vservers or just some of them
1138432312 M * Hollow ok, just a minute
1138432323 M * Hollow just the one you're testing with ;)
1138432328 M * jkl cool
1138432339 M * Hollow do you use 1.11.* or 1.12.*?
1138432387 M * jkl 1.12.*
1138432603 A * jkl will be back in 5
1138432835 Q * gerrit Ping timeout: 480 seconds
1138432985 M * jkl ah that was fun
1138433179 M * Hollow jkl: guess it'll take a bit still
1138433193 M * jkl k
1138433207 M * Bertl jkl: go for a second round, but come back :)
1138433221 M * jkl heh, i drove my buddy to the bar
1138433225 M * jkl he forgot his wallet
1138433269 A * jkl is going skiing early in the AM 
1138433276 M * jkl no drinking for me tonight
1138433440 M * Bertl as in alcohol, you mean?
1138433451 M * jkl heh, yeah
1138433532 J * gerrit ~gerrit@c-24-22-19-162.hsd1.or.comcast.net
1138433549 M * Bertl wb gerrit, again!
1138433660 M * gerrit hi Bertl - I'm playing with my home network.  ;)
1138433671 M * gerrit bertl:  it isn't *supposed* to be that obvious.  ;)
1138433682 M * Bertl heh :)
1138433698 M * Bertl gerrit: well, as long as we know that you are just fooling around, it's fine :)
1138433710 M * gerrit bertl:  all I do is fool around some days.  ;)
1138433738 M * gerrit bertl:  I'm flying Luftansa from the US to Germany this weekend - I'll be playing using the wireless on the airplane if all goes well, too.
1138433739 M * Bertl btw, you do not know some jfs folks by any chance?
1138433748 M * gerrit bertl:  I know them all.  ;)
1138433760 M * Bertl ah, excellent, could you make some kind of contact?
1138433778 M * gerrit I *think* they still have an official mailing list.
1138433781 M * jkl wireless and airplanes
1138433786 M * jkl fun combination :)
1138433795 M * gerrit Steve Best is currently in china/beijing but I think he's still maintaining
1138433797 M * Bertl because I'd like to get support for immutable and such into jfs ...
1138433815 M * gerrit if not, Shaggy (David Kleikamp) in Austin is also active still
1138433819 M * Bertl as almost all 'sane' filesystems have that now ...
1138433853 M * gerrit bertl: if you have patches, shaggy@austin.ibm.com will look at them.  I'll have to see if the mailing list is at sourceforge not
1138433906 M * Bertl well, it's not so much about patches, it's more the on-disk inode layout which would require adjustments
1138433918 M * Bertl to actually store those attributes
1138433932 M * gerrit oh - they wouldn't be quite as keen about that - jfs is primarily in maintenance mode.
1138433959 M * gerrit and since it isn't as widely used, writing conversion scripts and support for those would be more effort than they'd be willing to take on, I'd bet
1138433982 M * gerrit http://jfs.sourceforge.net/jfslllst.html
1138433994 M * Bertl let me put it this way, maybe the jfs folks _know_ where they left some space :)
1138434041 M * gerrit heh heh - good point.  Ask shaggy@austin.ibm.com or ask on the jfs mailing list above
1138434047 M * gerrit shaggy is pretty responsive
1138434056 M * Bertl okay, will do so, thanks for the pointers
1138434082 M * gerrit and now to go make sure my two tivo's can transfer programs over my newly configured network.  ;)
1138434147 M * Bertl linux based?
1138434184 M * Hollow hm, gentoo init style is quite easy with vserver-utils.. but i doubt there is an elegant solution for util-vserver
1138434215 M * jkl hm, which one am i using again?
1138434215 M * jkl hehe
1138434222 M * Hollow util-vserver
1138434225 M * Hollow i guess
1138434230 M * Bertl lol
1138434233 M * jkl hehe, what is the difference?
1138434260 M * Hollow u-v is bloated and buggy, v-u is lightweight and (still) buggy
1138434288 M * jkl how difficult is it to make the switch between them
1138434290 M * daniel_hozac hmm, where is u-v buggy?
1138434304 M * Hollow v-u is a new implementation, because i was not willing to hack util-vserver
1138434325 M * jkl hm, no ebuilds for vserver-utils
1138434326 M * Hollow daniel_hozac: just take start/stop sequences..
1138434333 M * Hollow jkl: oh yay
1138434340 M * Hollow when did you emerge sync the last time?
1138434347 M * Hollow in your previous life? ;)
1138434348 M * jkl today
1138434351 M * Hollow hm
1138434353 M * jkl hehe
1138434355 M * Hollow then it should be there
1138434364 M * daniel_hozac Hollow: where are they buggy?
1138434365 M * Hollow sys-cluster/vserver-utils
1138434372 M * jkl oh yah
1138434378 M * jkl man, im retarded sometimes
1138434381 M * jkl it's masked
1138434394 M * bwana whats that tool similar to wget many people are using
1138434408 M * daniel_hozac curl?
1138434426 M * Hollow well, dunno, it's been a long time since i last used util-vserver ;)
1138434448 M * jkl would you recommend switching to v-u ?
1138434456 M * Hollow but even from the dev perspective alone, it's horrible, and i don't like it
1138434458 M * Hollow ;)
1138434506 M * bwana that might be it
1138434509 M * bwana curl
1138434516 M * bwana thanks daniel_hozac
1138434562 M * daniel_hozac Hollow: you get used to it pretty quickly, like if you hurt yourself. it will still hurt like hell, but you won't notice it all the time.
1138434632 M * Hollow don't hink so.. it will always hurt me.. i don't get used to pain
1138434680 M * Hollow every time i approach util-vserver code this little voice in my head tells me "reqwrite it, jee, rewrite it! please!"
1138434711 M * Hollow and now i'm off to breakfast ;)
1138434715 M * Hollow back later
1138434750 M * daniel_hozac haha.
1138434772 M * daniel_hozac i guess i'm just a lot lazier than you, then :)
1138435049 M * Bertl well, lazy is not bad (in this particular case) it seems ...
1138435125 Q * gerrit Ping timeout: 480 seconds
1138435201 M * Hollow hm, i don't want to criticize util-vserver too much, it's just.. i'd never buy intel, i just like amd better, and util-vserver is just like intel, i don't like it that much ;)
1138435711 M * Bertl that's your right ...
1138435778 M * bwana anyone know what the deal is with perl and my local problems
1138435785 M * bwana like a quck fix
1138435797 M * Bertl what is your local problem?
1138435800 M * bwana well
1138435806 M * bwana although it is not crippeling 
1138435811 M * bwana it gernerates and error
1138435819 M * bwana when trying to run perl scripts n whatnot
1138435822 M * bwana it says
1138435854 M * Bertl yes?
1138435859 M * bwana perl: warning: Setting locale failed.
1138435859 M * bwana perl: warning: Please check that your locale settings:
1138435859 M * bwana         LANGUAGE = "en",
1138435870 M * bwana         LC_ALL = (unset),
1138435870 M * bwana         LANG = "en_US.UTF-8"
1138435870 M * bwana     are supported and installed on your system.
1138435879 M * bwana perl: warning: Falling back to the standard locale ("C").
1138435879 M * bwana locale: Cannot set LC_CTYPE to default locale: No such file or directory
1138435879 M * bwana locale: Cannot set LC_MESSAGES to default locale: No such file or directory
1138435884 M * Bertl ah, well, that's just that you have some locale set, in the environment, but it's not installed
1138435893 M * bwana hmm
1138435896 M * Bertl you can do two things there
1138435899 M * bwana ok
1138435911 M * Bertl a) use export LC_ALL=C LANG=C
1138435922 M * Bertl b) install the missing en/utf locale
1138435932 M * bwana so how does one do letter b
1138435945 M * bwana without a and c finding out
1138435959 M * Bertl that depends on your distro .. here the packages are called locale and locale-en
1138436002 M * bwana i will check
1138436028 M * bwana any idea what they are for ubuntu
1138436093 M * Bertl on debian it seems to be: locales and liblocale-gettext-perl
1138436114 M * Bertl not sure how the -en thingy is called (if there is one)
1138436124 M * bwana excellent thanks
1138436276 M * bwana damn
1138436282 M * bwana apparently they are already installed
1138436558 M * Bertl well, you host is ubuntu too?
1138436777 M * Bertl *your
1138436923 M * Hollow jkl: still around?
1138437263 M * Hollow Bertl: do you have a guest image with sysv init style for testing?
1138437294 M * Bertl sure .. tgz?
1138437298 M * Hollow yup
1138437932 M * Bertl http://vserver.13thfloor.at/Stuff/QEMU/XXXX.tar.bz2
1138437953 M * Bertl (it's a mandrake 9.1 guest with apache)
1138438008 M * Bertl actually with apache, bind, sshd and postfix
1138438125 M * Bertl will need hostname, ip and resolv.conf
1138438187 M * Hollow ok, thx
1138438498 M * Bertl okay, off to bed now .. back in the evening ...
1138438503 M * daniel_hozac good night!
1138438507 N * Bertl Bertl_zZ
1138438691 N * cohan_zzz cohan
1138438711 M * cohan *yawn* morning! and good "night", Bertl_zZ ;)
1138438749 M * daniel_hozac morning cohan!
1138438950 M * cohan moin daniel_hozac ;) *reading up*
1138439079 M * daniel_hozac Hollow: so what is needed to make the gentoo init style relevant again?
1138439202 M * Hollow daniel_hozac: well, you have to call /sbin/rc three times, once with sysinit, boot and default
1138439266 M * daniel_hozac hmm, what would sysinit and boot do in a vserver?
1138439288 M * Hollow sanity checks and depency cache
1138439299 M * Hollow all the mount etc stuff was removed
1138439458 J * Viper0482 ~Viper0482@p549771CA.dip.t-dialin.net
1138439800 M * cohan btw, all this locales-stuff and issues (who _really_ needs is anyway?) also led me to crux. for the kiss-concept, they ./configure every packet with --disable-nls etc ;) </adve
1138439813 M * cohan rtise>
1138439845 M * daniel_hozac hmm, so i guess the gentoo init style is doomed then.
1138439867 M * Hollow daniel_hozac: as i said, we tried to wrokaround it, but it always resulted in a corrupted cache
1138439872 M * Hollow it's quite fragile
1138439912 M * daniel_hozac i guess you really do want to run these things inside the vserver too, right?
1138439918 M * Hollow i'll take a nother look if we can do it nevertheless
1138439949 M * cohan plain is just signalling the guests init... gentoo and sysv are trying to emulate the rc-behaviour of the guest distro without init, right?
1138439958 M * Hollow cohan: right
1138439988 M * cohan what are the drawbacks of the plain / guests own init-version that there is so much effort in emulating it?
1138440002 M * daniel_hozac you run another process.
1138440006 M * Hollow daniel_hozac: well, we could just remove sysinit and boot, and put the cache routines in default, but the more we converge from default baselayout the more work to keep up to date
1138440018 M * daniel_hozac Hollow: yeah, i understnad.
1138440020 M * daniel_hozac +spelling
1138440060 M * daniel_hozac i guess the benefits are far outweighed by the cost.
1138440122 M * Hollow we'd just need a way to know if /sbin/rc default was called for boot purposes or for unlevel change
1138440152 M * cohan hm, i never realized that init is such a ressource issue... besides, it's really just one process - a vserver runs 10 on averay (syslog, sshd, cron, inetd and/or mail daemons, apache or webserver...)
1138440166 M * daniel_hozac it's not.
1138440170 M * Hollow cohan: well, many people like to see their guests booting
1138440179 M * Hollow you loose this feature with init
1138440205 M * daniel_hozac does init set its controlling tty?
1138440221 M * cohan ah, ookay - thats a point, still it would be nicer to see it WITH init ;)
1138440222 M * Hollow iirc yes
1138440223 M * daniel_hozac or would something like putting /dev/tty in /vservers/*/dev/console work?
1138440238 M * cohan iirc thats what we do on our old layout... *check*
1138440241 M * Hollow try to put /dev/console in it.. and you'll notice the keyboard to die on the host ;)
1138440252 M * daniel_hozac not the actual /dev/console.
1138440254 M * daniel_hozac /dev/tty
1138440289 M * Hollow well, that would print it inside the guest nevertheless, no?
1138440301 M * daniel_hozac would it?
1138440305 M * Hollow guess so
1138440318 M * daniel_hozac wouldn't /dev/tty point to the current tty, from which the (re)start was executed?
1138440325 M * cohan hm, our old (kernel 2.4.X-based) feature a /dev/console, at least they do not block.. not sure about seeing them booting, though
1138440327 M * daniel_hozac i guess that has other issues later though.
1138440330 M * Hollow we could try to do it like in vlogin, proxy a pseudo terminal
1138440348 M * daniel_hozac that would require vtd though.
1138440359 M * Hollow vtd never made it in svn.. ;)
1138440364 M * Hollow it's vlogin now
1138440367 M * daniel_hozac so it keeps on running and is available for attachment later.
1138440386 M * cohan something like screen?
1138440390 M * daniel_hozac pretty much.
1138440416 M * Hollow the client-server design was bit of an overkill for the login purpose, but we could do something with inti ehre i guess
1138440427 M * daniel_hozac yeah.
1138440462 M * Hollow vlogin is pretty straight, it should be usable for util-vserver too
1138440465 M * daniel_hozac i'll have to go now (pick up another box), i'll bbl.
1138440470 M * Hollow ok, cu
1138440470 M * daniel_hozac yeah, i've been meaning to port it.
1138440478 M * cohan that could almost replace my special sshd-instance which allows vserver owners to enter their vserver via vserver enter from outside almost without beeing present on the host
1138440519 M * Hollow cohan: if you start sshd in the guests they can even login without ever being on the host ;)
1138440554 M * cohan yes, but this sshd service is for those stopping their sshd or misconfiguring it ;)
1138440561 M * Hollow ic :P
1138440567 M * cohan also, they can enter if startup fails, whatever
1138440613 M * Hollow hm.. great idea btw..
1138440615 M * cohan its quite convenient, not discovered a security impact yet, too - but what will need some review before "releasing" the upgraded infrastructure
1138440628 M * Hollow we could include some rescue shell in vserver-utils which only allows vserver rescue commands to customers
1138440683 M * cohan don't get that now... where would that shell run? on the host? 
1138440729 M * Hollow yeah, you set /bin/vrescue as default shell, and it gives the customer a prompt, probably a help text, and some simple commands to start/stop clean their vservers
1138441187 M * cohan ah, okay, this is about what i have now, there is status, boot, kill, enter
1138441195 Q * Viper0482 Remote host closed the connection
1138441268 M * cohan could need adjustment, e.g. there is no reboot yet, need to enter, cleanup, if sth. is left, kill on the "rescue console", and type boot again)
1138441493 J * Viper0482 ~Viper0482@p549771CA.dip.t-dialin.net
1138441509 M * cohan btw, which is the nowadays-recommended way to start vservers after bootup with u-v ? init.d/vservers-default?
1138441655 M * cohan (is there a featured way to log this activity?)
1138441890 Q * AllenJB Remote host closed the connection
1138448436 Q * zobel Ping timeout: 480 seconds
1138449041 J * zobel zobel@zobel.irc.ftbfs.de
1138452124 Q * matti Remote host closed the connection
1138453616 J * bteoman ~bteoman@81.215.172.69
1138453645 M * bteoman hello!
1138453804 Q * bteoman Quit: 
1138454225 J * mdx ~perl@201-34-129-38.jvece701.dsl.brasiltelecom.net.br
1138454274 M * derjohn Q: when I use vserver build ..foo.. -- -d sarge ... what is "-d" for? I assumed -d was an debootstrap option, but I think my assumtion was wrong. I want to build an ubuntu breezy guest, so I need different args for debootstrap. Or shoud I create a /etc/vservers/.distributions/breezy/apt/sources.list ?
1138455490 M * derjohn hm, what does the .distribution dir do? is it only to create files _inside_ the guest?
1138456982 J * matti matti@linux.gentoo.pl
1138458148 A * mdx is away: comer
1138458679 J * meandtheshell ~markus@85-125-230-195.dynamic.xdsl-line.inode.at
1138459513 M * daniel_hozac the .distributions is for building, mostly.
1138459851 M * derjohn for the logs, this works: vserver vubuntu build --force -m debootstrap --hostname vubuntu.myvservers.net --netdev eth0 --interface 192.168.0.2/24 --context 42 -- -d breezy -m http://de.archive.ubuntu.com/ubuntu
1138459898 M * derjohn daniel_hozac, thx ... but vserver build didnt respect my ./apt/sources.list in the .distributions dir.
1138460191 A * mdx is back (gone 00:34:02)
1138460241 J * fhh ~plutark@gam75-4-82-235-223-10.fbx.proxad.net
1138460498 J * liquid3649_ ~Viper0482@p549771CA.dip.t-dialin.net
1138460749 Q * Viper0482 Ping timeout: 480 seconds
1138461175 J * monrad_ ~mikkel@213083190131.sonofon.dk
1138461187 Q * monrad Read error: Connection reset by peer
1138461801 P * mdx Leaving
1138462138 M * derjohn What does (should) getAllVservers names do? how do i call it correctly?
1138462171 J * liquid3649__ ~Viper0482@p54975EC0.dip.t-dialin.net
1138462215 N * liquid3649__ Viper0482
1138462216 M * derjohn I mean, how to call it from shell
1138462227 M * daniel_hozac what is getAllVservers?
1138462366 M * daniel_hozac ah, i see. i guess it puts all the vservers matching the criteria in the array named by the first argument.
1138462604 Q * liquid3649_ Ping timeout: 480 seconds
1138462837 M * derjohn daniel_hozac, I just ran over http://linux-vserver.org/Vserver+Completion . This would be really a p0rn feature. But doesnt work for me by now, I dunno  why. But I cannot find 'getAllVservers', which is references in the file. is it from legacy utils? I could replace it by some $(ls /etc/vservers) ...
1138462876 M * daniel_hozac it's in .../lib/util-vserver/functions
1138463181 M * derjohn h, not for me: locate util-vserver |grep getA -> NULL ;(
1138463191 M * derjohn daniel_hozac, you run .210 ?
1138463236 M * daniel_hozac yes.
1138463392 M * daniel_hozac it's a core file. vserver includes it.
1138463402 M * daniel_hozac (i.e. if you don't have it, your utils are screwed)
1138463703 M * derjohn ah, now I got it: /usr/lib/util-vserver/functions is a FILE, not dir ... so .. hm ... so how do I call it from (ba)sh?
1138463722 M * daniel_hozac like the completion script does it.
1138463904 M * derjohn oh, the thing should work (i.e. you are using it?). hm. I have bash3, but I assume this is backward compatile to 2.05 ?
1138463913 M * derjohn *g* # vserver
1138463913 M * derjohn Usage: /usr/sbin/vserver {start|stop|suexec|restart|condrestart|exec|enter|chkconfig|running|status}
1138463951 M * derjohn well, the usage hint is wrong. There must be a vserver <guestname> {start|stop|suexec|restart|condrestart|exec|enter|chkconfig|running|status} , or ?
1138463976 M * daniel_hozac yep.
1138463993 M * daniel_hozac and i don't see why it shouldn't work (but i've never done completion in bash)
1138464445 M * derjohn so, filed a bug about the usage on savannah.
1138464640 M * cohan ah, bash completion sounds interesting... i'll also give it a try
1138465233 J * prae ~benjamin@sherpadown.net
1138465269 Q * prae Quit: 
1138465882 M * michal_ "pay 122usd or 100EUR to unpaid X developer of your choice first."
1138465893 M * michal_ that's the answer x developers gave me to my bug reports
1138465904 M * cohan weird
1138465922 M * cohan hm, i just realize that bash_completion _really_ messes up one's environment ;)
1138466042 M * cohan derjohn: what was wrong with the vserver-bashcompletion? seems to work quite good, just the method of getting the vserver names could be rethought
1138466187 M * derjohn cohan, http://linux-vserver.org/Vserver+Completion -> bottom
1138466206 M * derjohn cohan, and the # # # lines.
1138466309 M * cohan i just installed it, feels good - what exactly on the bottom of this file? (i seem to be blind right now)
1138466441 M * derjohn in the wiki .... on  two lines there is missing some " stuff if you put it directly into the bashrc 
1138466473 M * derjohn (this is the way i tested it .. and ran into trouble with syntax errrors)
1138466474 M * daniel_hozac michal_: link?
1138466518 M * derjohn michal_, real or fraud ?
1138466561 M * cohan derjohn: it worked perfectly in my case (bash 3.00.16(1)), i did the following:
1138466589 M * cohan cut'n'paste the code from http://linux-vserver.org/Vserver+Completion into /usr/lib/bash-completion/vserver_completion
1138466606 M * derjohn cohan, if start bash completion with ". /etc/bash_completion" then it works
1138466637 M * derjohn cohan, but not if you put the code into .bashrc
1138466642 M * cohan yep, thats what i did - in my case with . /usr/share/bash-completion/init_bc
1138466673 M * derjohn cohan, it works for me now :) IMVHO this should be interagted into the debian package.
1138466679 M * cohan . /usr/share/bash-completion/init_bc
1138466685 M * cohan ups, sorry, wrong paste ;)
1138466691 M * derjohn micah, look at the last few lines .. would be nice :)
1138466702 M * cohan yes, it should - i'll also make a crux port for it
1138466737 M * cohan hm, i am curios now, i'll try it directly in .bashrc now as well..
1138466740 M * derjohn cohan, dont forget to provide your 100% upcybered crux-guest image :)
1138466769 M * derjohn cohan, yes! ( I have bash3 )
1138466773 M * cohan yes, still working on it ;) 
1138466804 M * cohan but i fear it will take another week until i get _inside_ the guest, i am right now preparing our next "vserver framework" for our housing environment
1138466976 M * michal_ daniel_hozac: no link, long talk on irc
1138466982 M * michal_ derjohn: real
1138467021 M * derjohn michal_, hm, paid support ;) oh well, is it xorg or xfree86 or ... ?
1138467093 M * michal_ xorg
1138467101 M * michal_ nvm
1138467663 N * Bertl_zZ Bertl
1138467669 M * Bertl morning folks!
1138467696 P * meandtheshell 
1138467720 M * cohan moinsen
1138467805 Q * shedi Ping timeout: 480 seconds
1138467845 M * derjohn Tachchen Bertl!
1138468082 J * meandtheshell ~markus@85-125-230-195.dynamic.xdsl-line.inode.at
1138468325 Q * meandtheshell Quit: a brave heart left ... 
1138468352 A * Bertl .o( hmm, wrong channel? :)
1138468357 J * shedi ~siggi@inferno.lhi.is
1138468440 M * Bertl wb shedi!
1138468948 M * Bertl derjohn: regarding savannah bug report "bonding" .. I'm the reporter, and no it's not sarge :)
1138469023 M * mnemoc it's mandrake!
1138469028 M * derjohn Bertl, *lol* 
1138469088 M * derjohn Bertl, ok... t I think ensc needs more info on that :)
1138469158 M * Bertl hmm, interesting, because I have a report that it was fixed :)
1138469201 M * Bertl 0.30.210+ has now a 'novlandev' flag
1138469219 M * Bertl he is just asking if that works for me ...
1138469239 M * Bertl and as I did not test it yet (never change a running system) I don't know yet ...
1138469267 J * Doener doener@i5387F88A.versanet.de
1138469277 M * Bertl welcome Doener!
1138469282 M * Doener hi Bertl 
1138469594 T * Bertl http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.0.9, 2.0.1.2 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)
1138469601 M * Bertl 2.1.0.9
1138469741 M * Doener just the vserver name fix
1138469741 M * Doener ?
1138469782 M * daniel_hozac i thought that was 2.1.0.8?
1138469845 M * Bertl yup, 2.1.0.9 has restructured resource limits
1138469859 M * Bertl will upload the delta(s) shortly
1138469877 M * Doener hm, ok, i lag behind ;)
1138469900 M * Bertl we missed a few things in the tag patches
1138469913 M * Bertl and bme missed a few entries too
1138469916 J * Smutje_ ~Smutje@xdsl-84-44-242-19.netcologne.de
1138469924 M * Bertl wb Smutje_!
1138470042 M * Bertl hmm, OVZ wants stories ...
1138470044 Q * Smutje Ping timeout: 480 seconds
1138470099 M * cohan looks good @ Bertl - i'll have some movie evening now with some friends *really looking forward to*
1138470147 M * Bertl cohan: enjoy!
1138470205 M * daniel_hozac stories of what?
1138470223 M * Bertl no idea, probably OVZ success stories ...
1138470255 M * daniel_hozac ah, that sort of thing.
1138470402 Q * fhh Quit: Leaving
1138470486 M * cohan and off ;) cul8r - i'll add that channel to my irssi next time i join
1138470491 Q * cohan Quit: ((42))
1138471055 J * AllenJB ~Allen@stuEAED.kent.ac.uk
1138471281 M * Bertl welcome AllenJB!
1138471312 M * Bertl Doener: okay, all deltas should be in PAT-2.1.1 now
1138471371 M * Bertl guess it's really time for a new devel release ... what do you folks think?
1138471798 M * Bertl ah, okay, guess you have to think about it ... will ask later again :)
1138471808 M * Bertl off for now .. back later ...
1138471815 N * Bertl Bertl_oO
1138472175 M * derjohn Bertl, without substantially know what the differences are, I would vote for 'yes'.
1138472198 M * daniel_hozac derjohn: PAT-2.1.1/* ;)
1138472277 M * derjohn daniel_hozac, I don't speak C :) .. or is PAT the new name for changelog ... But i'll have a look on the function named that are patched :)
1138472530 M * daniel_hozac derjohn: well, the filenames usually tell you something.
1138472544 J * meandtheshell ~markus@85-125-230-195.dynamic.xdsl-line.inode.at
1138472805 Q * zobel Ping timeout: 480 seconds
1138473339 J * zobel zobel@zobel.irc.ftbfs.de
1138473348 Q * emp Quit: Leaving
1138474232 J * bonbons ~bonbons@83.222.38.150
1138474232 Q * monrad_ Read error: Connection reset by peer
1138474264 M * bonbons Hollow: ping
1138474531 J * liquid3649_ ~Viper0482@p54975EC0.dip.t-dialin.net
1138474608 J * emp ~emp@70.57.239.35
1138474659 M * Hollow bonbons: pong
1138474762 M * bonbons Hollow: I've fixed a small issue with vattr today, wasn't working too well on folders inrecursive mode, the "root" was skipped!
1138474784 M * Hollow ah great :)
1138474786 M * Hollow patch?
1138474805 M * bonbons thus vserver procfs unhide and hide did update /proc/net/*, but not /proc/net itself (so guests still had no /proc/net)
1138474820 M * bonbons jup, on my vserver-test-box :)
1138474825 Q * Viper0482 Ping timeout: 480 seconds
1138474840 M * bonbons and checking the week's updates right now
1138474863 M * Hollow i added the first try for libvconfig... and off to dinner now, back shortly
1138474996 M * bonbons Hollow: good appetite!
1138476985 M * Hollow back now
1138477056 M * bonbons ok, I'm reading the updates, src/libvconfig/*.c
1138477072 M * Hollow yeah, still very rough
1138477238 M * bonbons yeah
1138477327 M * bonbons I'm wondering how the 'xid_t vconfig_get_xid(char*name)' and reverse should work, they are just querying the kernel, aren't they?
1138477680 M * Hollow xid->name looks in the uts data, name->xid looks in the config
1138477796 M * bonbons name->xid first checks the config, but afterwards asks kernel and fails if kernel does not know anything about xid read from config
1138477876 M * Hollow bonbons: it just verifies that if the xid is running the context field in uts matches the one from config
1138477881 M * Hollow at least it should do
1138478019 M * bonbons what does the kernel return to syscall when given xid does not exist?
1138478039 Q * liquid3649_ Quit: bin raus, 
1138478041 M * Hollow hm, -1 ;)
1138478045 M * Hollow you're right, it's a bug
1138478053 M * matti Not feature?
1138478054 M * matti ;]
1138478082 M * bonbons matti: on kernel a feature, on libvconfig a bug :)
1138478194 M * bonbons Hollow: for the functions like vconfig_isXXX(char*key) it woud be more logical to return 1 if it is, zero if it's not, and eventually -1 on error (unknown key), would be more intuitive as functions looks "boolean"
1138478222 A * Hollow nods
1138478266 J * Viper0482 ~Viper0482@p54975EC0.dip.t-dialin.net
1138478301 M * Hollow also the is* functions all do the same.. we could combine them with an additional argument for the type
1138478363 M * bonbons that's true, the type is currently hardcoded in them, so just make it a param (enum)
1138478466 M * bonbons in addition, there should be a vconfig_init(...) function that loads configuration information, e.g. to a config-handle which would then be used afterwards (avoids reloading at each query)
1138478533 M * daniel_hozac vconfig_open sounds better :)
1138478648 M * bonbons that's fine as well, and have a corresponding vconfig_close. How about saving, immediate or (optional) lazy on close?
1138478658 M * Hollow well, since you have a different file for each config option, that wouldn't be useful i guess
1138478694 M * daniel_hozac hmm, i thought you were going for multiple backends?
1138478709 M * daniel_hozac (database, single file, current, etc.)
1138478713 M * Hollow yeah, but only one is implemented
1138478718 M * bonbons I'm not too much for many files, I prefer to have a limited amount of files, but that would depend on backend!
1138478743 M * daniel_hozac the API would have to be the same though, no?
1138478779 M * Hollow yup
1138478808 M * daniel_hozac and for the others, _open and _close certainly make sense.
1138478822 M * bonbons so at least having both options would be really useful. sure, same API for all, but if we have a "config-handle" such backend-specific info could be hidden properly from outside
1138478858 M * Hollow yeah, depends completely on the backend
1138478960 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1138479007 M * bonbons vconfig_init/open/close would use structures with a common part an a backend-specific part, all other functions whould be really identical from outside
1138479077 M * bonbons best thing would be to be able to switch backends and even copy from one backend to the other (state-less API, all state hidden in our handle)
1138479260 M * Hollow yeah, another question would be if xid/name options to programs should compleltey transparent
1138479329 M * daniel_hozac being able to use either?
1138479338 M * bonbons that would be cool, and should not be too difficult while operating on running contexts (we can query kernel), using libvconfig in the other cases
1138479368 M * daniel_hozac util-vserver does it, so it'd be a regression not to do it ;)
1138479372 M * Hollow yeah, i.e. vnflags -x vs1 ...
1138479402 M * bonbons daniel_hozac: currently it's a "regression"
1138479614 M * bonbons but there are quite a few features of util-vserver that are still missing, so regression is not really the appropriate word
1138479684 Q * nox Read error: Connection reset by peer
1138479724 J * nox ~nox@noxlux.de
1138479857 M * Hollow but it hink we're (slowly) getting somewhere
1138479963 M * bonbons yeah, with some more time vserver-utils will be more complete (especially with libvconfig for makeing config operations easy)
1138480279 M * bonbons Hollow: do you know what can cause sshd in guests to not get any tty? I get sshd user@notty processes in the guests, and ssh client just waits for shell. opening /dev/ptmx inside the guest works fine (/dev/pts/x appears as expected)
1138480318 M * Hollow hm.. no idea
1138480375 M * bonbons looks like sshd misses something (works using util-vserver, but not with vserver-utils). The vattr bug cannot be considered responsible, corrected vattr has handled procfs unhide
1138480378 J * cyfer ~cyfer@ALICE-WHACKER.MIT.EDU
1138480399 M * bonbons daniel_hozac: do you know if sshd needs some extra files from /proc?
1138480477 M * daniel_hozac it shouldn't.
1138480506 M * daniel_hozac tried with debugging enabled?
1138480540 M * bonbons for getting a virtual console, does it need something special except the /dev/ptmx? Not yet, but on the way to look deeper at it
1138480756 Q * shedi Quit: Leaving
1138480821 J * duplex ~simplex@72.242.34.141
1138480892 M * duplex I'm curious, I've been looking around for a long time, and can't seem to find any documentation on running a gentoo guest on a debian host.  Does anyone know where a good howto is on this?  I dont know anything about vserver at all, really.
1138480980 M * bonbons duplex: for vserver guests, the host distro is not important at all, reading gentoo install doc and vserver-doc should get you started with gentoo inside vserver
1138481041 M * duplex bonbons:  Oh okay, thanks.
1138481078 M * daniel_hozac but creating the guests does depend on the host distribution.
1138481109 M * duplex That's the problem, I dont know how to actually create the guests.
1138481187 M * Doener *lol* i just wondered why some "grep foo *" took ages, then i noticed that the 2.0.1.1 patch creates a file called -
1138481188 M * daniel_hozac IIRC Hollow provides stage3 tarballs.
1138481203 M * duplex daniel_hozac:  Hollow?
1138481211 M * Hollow yup, on the gentoo mirrors
1138481215 M * daniel_hozac Doener: heh, yeah, i did that too...
1138481233 M * duplex Hollow:  Is there any documentation on how to install them with vserver?
1138481250 M * Hollow build skeleton, unpack to vserver dir, done
1138481255 M * Hollow use plain init style
1138481256 J * gerrit ~gerrit@c-24-22-19-162.hsd1.or.comcast.net
1138481279 M * duplex no vserver whatever build?
1138481288 M * Hollow sure, with skeleton build method
1138481294 M * duplex oh okay
1138481336 M * duplex Hollow:  Do you have a link to those tarballs?
1138481358 M * Hollow http://distfiles.gentoo.org/experimental/x86/vserver
1138481370 M * duplex Thanks =D
1138481412 M * daniel_hozac Hollow: re gentoo initstyle, would adding a simple script to baselayout-vserver that does /sbin/rc {sysinit,boot,default} (in the right order) work?
1138481419 M * Doener bonbons: any messages in the log about the context messing with devpts?
1138481424 M * Hollow daniel_hozac: probably
1138481479 M * bonbons Not yet done with the advanced test, hard to do 3 things at same time :)
1138481772 M * bonbons Doener: seems to be a permission issue Allocating pty, openpty: permission denied
1138481800 M * bonbons no neet to find out why, sshd running as root, loging in as root!
1138481806 J * monrad ~mikkel@213083190131.sonofon.dk
1138481812 M * bonbons s/no /now/
1138481897 M * Doener bonbons: that means no messages about messing with the devpts?
1138481924 M * Doener (in syslog, kern.log or whatever=
1138481928 M * Doener s/=/)/
1138482011 M * bonbons Hmm, log not available yet on that guest ;), dmesg has not message
1138482102 M * Doener would appear on the host
1138482355 J * Johnnie ~jdlewis@dynamic-acs-24-154-53-16.zoominternet.net
1138482373 Q * Johnnie Quit: 
1138482937 M * bonbons Doener: no, nothing to kernel.log, and sshd's output is not too detailled (even with 3 times -d)
1138483024 Q * gerrit Ping timeout: 480 seconds
1138483354 J * storm ~sil@134.Red-88-6-252.staticIP.rima-tde.net
1138483362 P * storm 
1138483375 J * milex_ ~milex@62.162.237.146
1138483440 Q * milex_ Quit: 
1138483474 M * bonbons Hmm, looks like it's a capabilities problem, default capabilities are not the same (looking at /proc/self/vinfo on util-vsefver guets and vserver-utils guest)
1138483564 M * bonbons fffffffffffffeff on vserver-utils, 00000000344c04ff on util-vserver!
1138483690 M * Doener hm, that's more caps... not sure the kernel likes the upper 32bit being set
1138483867 M * bonbons it's was BCaps, CCaps differs as well, 0 for vserver-utils, 0...0101 for util-vserver, cflags have some differences as well
1138484202 Q * aba Ping timeout: 480 seconds
1138484229 Q * Hollow Ping timeout: 480 seconds
1138484453 Q * Viper0482 Remote host closed the connection
1138484645 J * gerrit ~gerrit@c-24-22-19-162.hsd1.or.comcast.net
1138485455 J * aba ~aba@eos.turmzimmer.net
1138485473 P * cyfer 
1138485856 Q * Blissex Remote host closed the connection
1138486503 J * Hollow ~hollow@home.xnull.de
1138486676 Q * meandtheshell Quit: a brave heart left ... 
1138487365 M * bonbons Hollow: are you sure that vflags works correctly? The contents of /proc/self/vinfo in my vserver-utils guests look incorrect for BCaps: fffff...feff and vflags on a util-vserver guest with totally different bcaps produces the same result
1138487578 M * Hollow well, fo r the initial bcaps setting this is correct
1138487711 M * bonbons what explains the nearly identical output of vflags in both cases? (LEASE and F:* are the only differences)
1138487774 M * Hollow sorry, can't follow.. if you start a vserver the bcaps value is not altered by default
1138487890 M * bonbons On one box I start a guest with default caps which whos as BCaps: 00000000344c04ff, CCAPS: 0
1138487964 M * bonbons CCAPS: 0...0101 (util-vserver), on another box a guest started by vserver-utils has BCAPS: FFFF...FEFF, CCAPS 0...0
1138488006 M * Hollow iirc util-vserver alters bcaps
1138488037 M * bonbons in both cases default setting, on both hosts vflags outputs the same list with exception of "B: LEASE" and one or two "F: ..." items
1138488080 M * bonbons vflags -G -x <xid>
1138488094 M * Hollow but one was started with util-vserver and one with vserver-utils right?
1138488129 M * bonbons exactly, the one with mostly ffff in bcaps with vserver-utils, the other one with util-vserver
1138488147 M * Hollow as i said, util-vserver alters bcaps
1138488151 M * Hollow iirc
1138488159 M * bonbons I'm trying to get out the difference(s) that cause sshd not to get it's tty on the one started by vserver-utils
1138488180 M * Hollow hm, the one with vserver-utils has more bcaps i guess..
1138488193 M * Hollow all except one
1138488195 M * Hollow the e
1138488217 M * bonbons the 'e'?
1138488226 M * Hollow in fffff..ffeff
1138488258 M * Hollow all bcaps are set except 0x100
1138488297 M * bonbons ah ok, the one of util-vserver does not seem to have that either: 0x4ff does not include 0x100
1138488318 M * Hollow guess that's cap_mknod
1138488375 M * bonbons Should be possible. And for the CCAPs? vserver-utils has 0, util-vserver has 101 What do those include?
1138488386 Q * shuri Quit: Quitte
1138488440 M * Hollow raw_icmp and set_utsname
1138488456 J * ScoobyD00 ~foo@80-195-186-201.cable.ubr08.newm.blueyonder.co.uk
1138488488 M * bonbons that's extra vserver-capabilities, isn't it?
1138488514 M * ScoobyD00 hello - can anyone explain how to reverse the "setattr --barrier" command? :o)
1138488528 M * Hollow yup
1138488541 M * Hollow ScoobyD00: settattr --~barrier
1138488591 M * ScoobyD00 thanks!
1138488674 M * ScoobyD00 the reason i asked that is because when I did "setattr --barrier /etc/vservers/.defaults/vdirbase/test1" I could not start the vserver any more (chroot permission denied). Is that supposed to happen?
1138488697 M * Hollow yeah, you should set the barrier on the vserver dir
1138488711 M * Hollow i.e. /vservers or whereever your vserver dir is
1138488748 M * ScoobyD00 i have "/var/lib/vservers/" which contains the individual vservers
1138488760 M * Hollow ok, than that's your vserver dir ;)
1138488862 M * ScoobyD00 so, does the barrier stop a vserver from chdir'ing to /var/lib/vservers/ or /var/lib/ ?
1138488917 M * Hollow it stops from exploit the chroot with a simple cd ../..
1138489076 M * ScoobyD00 ok, thanks, just want to make sure they cannot get to other vserver directories :o)
1138489091 M * Hollow set it on /var/lib/vservers and everything should be fine ;)
1138489153 M * bonbons Hollow: from strace of sshd, get permission error when opening /dev/pts/0
1138489313 M * Hollow bonbons: ic, if devpts is mounted we have to investigate further, but i'm off to bed now, too tired
1138489394 M * bonbons Hollow: yep, it's mounted, an if I "cat /dev/ptmx &" I can see a new entry in /dev/pts/
1138489411 M * ScoobyD00 another newbie question :o) is there any other way than using tagxid to apply a disk quota to a vserver?
1138489447 M * bonbons really need to investigate, it's anoying to not being able to ssh to a guest!
1138489567 M * ScoobyD00 because i only have 1 big root partition and cannot get tagxid to work (it will not mount as rw)
1138489728 Q * bonbons Quit: I'm to bed too
1138489903 J * menomc ~amery@200.75.27.27
1138490011 Q * mnemoc Ping timeout: 480 seconds
1138490011 N * menomc mnemoc
1138490048 J * shedi ~siggi@inferno.lhi.is
1138490929 N * Bertl_oO Bertl
1138490934 M * Bertl back now ...
1138491022 M * ScoobyD00 can anybody point me at a working reboot command for a vserver? :o)
1138491033 M * daniel_hozac reboot -f
1138491034 M * Bertl reboot?
1138491043 M * ScoobyD00 test1:/# reboot
1138491043 M * ScoobyD00 shutdown: /dev/initctl: No such file or directory
1138491043 M * ScoobyD00 init: /dev/initctl: No such file or directory
1138491049 M * ScoobyD00 ^^ thats the output i get
1138491060 M * Bertl because you ahve no init running inside
1138491068 M * ScoobyD00 oh :o)
1138491069 M * Bertl in this case reboot -f will work
1138491089 M * Bertl because reboot contacts init, which in turn does the reboot
1138491100 M * Bertl (unless used with the -f option)
1138491111 M * ScoobyD00 ummmm, "reboot -f" just did something odd
1138491134 M * ScoobyD00 it killed my ssh connection to the main host - should it do that?
1138491190 M * Bertl well, really depends on _many_ things .. first, what's your kernel and patches?
1138491202 M * ScoobyD00 debian stable
1138491212 M * ScoobyD00 2.6.8
1138491217 M * Bertl (best would be to run testme.sh and upload the output somewhere, e.g. pastebin)
1138491249 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh-0.15
1138491260 M * blizz lol, i see similarities between my host system and his :P
1138491279 M * blizz i should setup a cronjob rebooting the network connection every once in a while :)
1138491286 M * blizz s/booting/starting/
1138491290 M * Bertl why?
1138491309 M * Loki|muh hehe, still connection 'problems'? *g*
1138491352 M * blizz Loki|muh, yup.. but didnt work on it lately.. had no time :/
1138491379 M * blizz Bertl, i still have the same problem scooby has, guest can influence the host's network
1138491417 M * Bertl well if configured properly that _does not_ happen ... but we can track it down if you folks like ...
1138491434 M * ScoobyD00 do i run testme.sh inside the vserver, or outside?
1138491445 M * Bertl ScoobyD00: on the host, as root
1138491447 M * ScoobyD00 k
1138491481 M * blizz Bertl, my problem is, i only have hollow's stuff, so it could be possible that i forgot something inside the guest
1138491486 M * ScoobyD00 sent output as pm
1138491491 M * ScoobyD00 looks good
1138491513 M * blizz but i will get the utils-vserver to run testme.sh from cohan soon..
1138491606 M * Doener blizz: bonbons noticed that caps are wrong with vserver-tools (is that the name?), so I guess that reboot is not safe with them
1138491658 M * blizz really? oh.. thats nasty
1138491684 M * Hollow you can't reboot with vserver-utils from inside anyway
1138491701 M * Hollow you can just halt it
1138491765 M * blizz hmmkay
1138491891 M * Bertl ScoobyD00: yes, kernel looks fine so far
1138491944 M * Hollow always keep in mind that the vserver command is just a proof-of-concept of the low-level tools
1138491955 M * Hollow it's not supüposed to be sucure, complete or usable in any way..
1138491962 M * ScoobyD00 ok, it doesnt seem to disconnect a host ssh session if i ssh into the vserver (rather than use "vserver test1 enter)
1138491975 M * ScoobyD00 so thats ok :o)
1138491995 M * Bertl ah, well, that's expected :)
1138492017 M * Bertl the 'enter' is for maintainance only, you are still on the host there
1138492028 M * blizz Hollow, yes, i should keep that in mind.. but anyway, great work so far
1138492042 M * Hollow thanks :)
1138492055 M * ScoobyD00 thanks also :o)
1138492115 M * ScoobyD00 is there any other way to get vserver disk limits than using the tagxid thing?
1138492130 M * Hollow Bertl: hm, with the vlogin command of vserver-utils.. is it possible to do secure migration probably?
1138492144 M * ScoobyD00 i tried that but could not get my root filesystem (which everything is on) to mount as rw
1138492178 M * Bertl Hollow: you can do a simple test, the chroot escape
1138492193 M * Bertl Hollow: if that works, you can assume that you're secure
1138492227 M * Hollow ok, is there a test program?
1138492231 M * Bertl ScoobyD00: well, yes, you can use separate partitions per guest
1138492261 M * ScoobyD00 ummm, i am limited to this 1 big partition on my live machine (in a datacenter where i cant play with it easily)
1138492293 M * ScoobyD00 i have got it working with a loopback filesystem image thing, but it is about 33% slower for disk accesses
1138492298 M * ScoobyD00 :o(
1138492306 M * Doener ScoobyD00: you got a swap partition?
1138492316 M * Bertl Hollow: http://vserver.13thfloor.at/Stuff/rootesc.c
1138492335 M * ScoobyD00 doener: yep
1138492356 M * Doener http://linux-vserver.org/Remote+Server+Repartitioning
1138492401 M * Bertl Hollow: there is another one too, give me a minute to find it
1138492426 M * ScoobyD00 whoa - thats serious messing around with the disks :o)
1138492435 M * Hollow ok, breakout works
1138492488 M * Doener ScoobyD00: well, it worked for me some years ago ;)
1138492509 M * ScoobyD00 i did some similar stuff to image remote machines with FreeBSD a few years ago :o)
1138492517 M * Hollow but with barrier set it does not work
1138492560 M * Bertl Hollow: hmm, at least something ...
1138492648 M * Bertl blizz: what does 'chcontext --xid 42 --secure -- grep CapE /proc/self/status' give you?