1181865606 M * Bertl it basically configures the fair scheduling shares
1181865640 M * Bertl i.e. as long as work can be done, the hard cpu scheduler controls the distribution
1181865671 M * Bertl when all contexts have no tokens left, and the system would go idle, the idle time shares are used
1181865671 M * dallas I am using this command to change a live guest on the fly.  does that option not give me anything useful?
1181865691 M * dallas ah... so yeah I think that's exactly what I don't want
1181865703 M * Bertl well, it switches it from pure hard scheduling to fair scheduling
1181865732 M * Bertl simply leaving that out will switch the guest back
1181865744 M * dallas in this case the host itself is mostly unused as we're only just starting to test.  got it
1181865773 M * Bertl yep, as long as host/guest progress can be made, the idle time shares will not kick in
1181865850 M * dallas is fair scheduling the same as priority scheduling, or not at all?
1181865862 M * Bertl not at all ...
1181865863 M * dallas or is fair simply splitting the available cpu up?
1181865884 M * Bertl you probably figured the rate/interval for hard scheduling by now, yes?
1181865900 M * dallas yes, hard scheduling is working
1181865925 M * Bertl okay, now when all token buckets are empty, and the host is doing nothing, the following happens:
1181865962 M * Bertl for all guests which have idle time skipping enabled, the token bucket's rate/interval is switched to a second set
1181865973 M * Bertl (called rate2/interval2)
1181866000 M * Bertl and then time is artificially advanced till the first guest can be sheduled (instantly)
1181866032 M * Bertl thus, you can distribute the 'idle time' according to the rate2/interval2 shares among guests
1181866051 M * dallas aha.  I was also wondering what those rate2/interval2 were for.  makes sense.
1181866084 M * dallas thank you, I believe I'm on my way now.
1181866100 M * Bertl you're welcome!
1181868180 J * DoberMann_ ~james@AToulouse-156-1-86-253.w86-196.abo.wanadoo.fr
1181868184 Q * Piet Quit: Piet
1181868289 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds
1181870767 Q * harry Read error: Connection reset by peer
1181870769 J * harry ~harry@d54C2508C.access.telenet.be
1181870817 P * dallas 
1181873348 J * slacker403 ~root@cpe-71-74-73-121.insight.res.rr.com
1181873352 M * slacker403 yo yo
1181873362 M * slacker403 whats up Bertl daniel_hozac etc 
1181873505 M * Bertl working on Linux-VServer of course ...
1181873558 M * slacker403 yea
1181873581 M * slacker403 ive been getting very distracted doing dumb stuff 
1181873587 M * slacker403 spent 2 days compiling a kerrnel
1181873596 M * slacker403 finnaly did with vserver installed
1181873611 M * Bertl well, now you should know it inside out :)
1181873618 M * slacker403 well
1181873621 M * slacker403 its slackware 
1181873624 M * slacker403 and 
1181873641 M * slacker403 i need to get offf persoanl preferences and jus focus on money 
1181873647 M * slacker403 people dont want to use no slackware 
1181873661 M * slacker403 its old no package manager per say online anyways 
1181873703 M * Bertl shouldn't be problem, that's what Linux-VServer is for
1181873721 M * Bertl put the latest debian, ubuntu, fedora, etc on the guests
1181873742 M * slacker403 well yea i know i was creating a custom distro in slackware too
1181873777 M * slacker403 you rich ?
1181873788 M * slacker403 not ot be personal just wondering 
1181873807 M * Bertl unfortunately no :)
1181873823 M * slacker403 did you make vserver ?
1181873829 M * slacker403 who started it ?
1181873834 M * slacker403 you a big contributor ?
1181873884 M * Bertl no, but all current kernel code is from me, a canadian guy started the project (Jacques Gelinas) who handed it over to me about three years ago
1181873902 M * slacker403 so your in head of it now ?
1181873913 M * slacker403 people make millions off your software 
1181873927 M * coderanger Gotta love FOSS
1181873931 M * Bertl I'm currently maintaining the kernel (and mentoring the project)
1181873936 M * slacker403 well
1181873938 M * slacker403 man
1181873941 M * slacker403 i lost soooo much money
1181873957 M * slacker403 i had the full virtulization kind before had a few hundred customers 
1181873969 M * slacker403 decided to switch over to the native kind of thing vserver 
1181873975 M * slacker403 been about 2 months 
1181873985 M * slacker403 lost likje half the customers which was really wrong 
1181873994 M * slacker403 i dont blame them 
1181874031 M * slacker403 i refunded a bunch too ........but hopefully this time i will do it right .........better software this time 
1181874049 M * slacker403 i know people who use vserver and make 150 K a month 
1181874050 M * Bertl why didn't you make a smooth transition?
1181874056 M * slacker403 i was stupid
1181874063 M * Bertl okay :)
1181874069 M * slacker403 i had 175 support  tickets lol 
1181874086 M * slacker403 so i jus refunded 
1181874123 M * slacker403 i got about 25 loyal people i setup on a different server .......called them gave them it free until i get this up so thats good soo 25 x 60.00 for starting here but yea 
1181874135 M * slacker403 i need to finish this tonight
1181874136 M * Bertl please ask those folks (making 150K a month) to contribute back to the project :)
1181874157 M * Bertl (so that we can make it even better!)
1181874158 M * coderanger Bertl: You know they never will ;-)
1181874163 M * slacker403 well
1181874179 M * slacker403 i will ask them
1181874184 M * Bertl thanks!
1181874190 M * slacker403 i will contribute 
1181874213 M * slacker403 on a monthly basis :)
1181874216 M * doener they could e.g. clean up. like, ... amazon wishlist clean up? ;-)
1181874218 M * doener just kidding
1181874219 M * slacker403 but yea jus need ot get everything done 
1181874258 M * slacker403 this way i will fit 20 people on a server
1181874263 M * slacker403 otherwise it was about 5-7
1181874353 M * coderanger Bertl: It seems like ctx_create errors out with VC_NOCTX if you give it an already used xid, but it doesn't set errno or do anything else to tell you what the exact error way
1181874405 M * slacker403 anyways i formatted the server .......so back to square one ........will start with debian this time jus get a newer version 
1181874412 M * coderanger Nevermind, it does, ipython was just making other syscalls so errno got overwritten
1181874489 M * slacker403 Bertl: if i was you Bertli would have a vserver business going ........ my obstacle is seritng it up right ........you would be great at that part 
1181874526 M * doener Bertl: hm, just continued reviewing the whitespace changes. is "x..." an alternative to using VA_ARGS?
1181874539 M * Bertl doener: yep
1181874548 M * doener interesting, never seen that IIRC
1181874590 M * Bertl slacker403: if you want to buy consulting time, please contact me in private
1181874660 M * slacker403 no problem will do once i get it up .......maybe you can help make it better 
1181874689 M * slacker403 you got a job ?
1181874801 M * Bertl I'm working as consultant, yes
1181874835 M * slacker403 so vserver is your life ?
1181875045 M * Bertl not just Linux-VServer, but it is definitely a big part of it :)
1181875052 M * slacker403 i didnt mean that
1181875056 M * slacker403 but its kinda your career 
1181875094 M * slacker403 you live in germany right ?
1181875101 M * coderanger I don't think nerd have careers anymore, we just like to hack and people seem to think it is worth something :)
1181875113 M * slacker403 lol
1181875120 M * slacker403 i hate computers ...........
1181875141 M * slacker403 but i need them 
1181875153 M * slacker403 and makes our lives way easy 
1181875203 M * slacker403 Bertl: i got love ofr you making the software and all but i dont see how you make this software for other to make millions and you dont yourself 
1181875260 M * doener it's not all about the money...
1181875284 M * slacker403 nope 
1181875293 M * slacker403 but i never really met a happy poor person 
1181875343 M * Bertl well, I love doing what I do ... and researching what I'm interested in ... and maybe I'm just a philanthropist :) 
1181875348 M * doener I could probably be less "poor", but that would probably also make me less happy, because atm it would mean working more on things I do not like
1181875361 M * slacker403 lol 
1181875368 M * slacker403 i enjoy how i make my money dont get me wrong
1181875376 M * slacker403 but when you have money it opens possibilities 
1181875390 M * coderanger You can make 40k a week as a spammer in China, but I can't say I would really want to try
1181875412 M * slacker403 coderanger: you can make more with VPS
1181875439 M * slacker403 again i dont like computers or money .........but they are both needed 
1181875448 M * slacker403 and the more you got the more you can do what you want to do
1181875448 M * Bertl slacker403: look at Bill Gates, he has all the money one could possible have, and still can't get a working version of Windows :)
1181875455 M * coderanger slacker403: I think I would be hard pressed to find a more awesome job than i have now 
1181875470 M * slacker403 whats your job
1181875479 M * slacker403 Bertl: he set a good steppiung stone 
1181875505 M * coderanger slacker403: Programmer at One Laptop Per Child
1181875515 M * slacker403 maybe its jus i had it hard growing up and would never want to be that way 
1181875517 M * doener Bertl: I think he _could_, but then he probably couldn't enjoy the chair throwing actions of Steve anymore ;)
1181875562 M * slacker403 poor is no good 
1181875570 M * slacker403 bill set a nice stepping stone 
1181875614 M * coderanger daniel_hozac: What do you think of this as an API so far: http://paste.lisp.org/display/42799
1181875678 M * slacker403 Bertl: your from austria correct ?
1181875692 M * Bertl yep
1181875698 M * slacker403 you like it there ?
1181875706 M * Bertl yep
1181875712 M * slacker403 ever plan to move ?
1181875728 M * Bertl I'm moving quite a lot :)
1181875739 M * slacker403 i meant
1181875743 M * slacker403 move countires / places ?
1181875758 M * slacker403 isnt austria pretty cold ?
1181875795 M * Bertl depends, last few times I visited the US, it was colder there than in Austria :)
1181875809 M * slacker403 lol
1181875812 M * slacker403 where idd you go ?
1181875817 M * slacker403 i live in miami florida 
1181875827 M * slacker403 its not been cold here since 1940
1181875841 M * slacker403 half tghe us is cold but the bigger half is warm to hot 
1181875843 M * coderanger slacker403: It was plenty cold when I lived there in 1997
1181875856 M * slacker403 its a huge place 
1181875863 M * slacker403 depends on where you go 
1181875877 M * slacker403 thats like me saying i visited europe and it was cold ;)
1181875892 M * slacker403 theres 50 different independent nations ;)
1181875893 M * coderanger Miami is slightly smaller than Europe ...
1181875900 M * slacker403 miami is a city
1181875911 M * slacker403 i have lived in charleston south carolina 
1181875913 M * coderanger But to be precise, miami beach, about halfway down the island
1181875914 M * slacker403 miami flroida 
1181875928 M * slacker403 charleston was still hot
1181875931 M * slacker403 palm trees etc 
1181875954 M * slacker403 california texas arizone new mexivo texas all verrrry hot places 
1181875971 M * slacker403 GA
1181875974 M * slacker403 south carolina 
1181875980 M * slacker403 louisianna 
1181875983 M * slacker403 all hot ;)
1181875996 M * slacker403 but yea new york and maine get cold 
1181875998 M * doener Bertl: in proc_pid_nx_info there's "%d.%d.%d.%d/%d.%d.%d.%d", in a few other places you've changed that to NIPQUAD_FMT "/" NIPQUAD_FMT
1181876030 M * slacker403 coderanger: i think miami is bigger then austria ;)
1181876094 M * slacker403 anyways yea ......... when i was in the UK it was super cold ......almost all year, they thought 80 F was like death temperature 
1181876118 M * coderanger Miami: 143.15 sq km   Austria: 83,870 sq km
1181876145 M * slacker403 it was a joke 
1181876149 M * slacker403 miami is small
1181876165 M * slacker403 flroida is bigger im sure ;)
1181876169 M * slacker403 which is the actual state 
1181876192 M * Bertl doener: ah, missed that, can you draft a patch for me?
1181876231 M * doener yeah, will take a bit though. going through the rest of the patch and then need to get the right kernel first
1181876260 M * Bertl shouldn't be too kernel specific I guess ... but TIA
1181876308 M * doener hm, I doubt that 2.6.22-rc4 will do, will it? ;) I only got bleeding edge vanilla source trees lying around atm
1181876406 M * Bertl actually it will, a patch will probably be there when you are through :)
1181876425 M * Bertl just have to double check a few more things ...
1181876471 M * slacker403 coderanger: http://www.alpinehomeair.com/ewebeditpro4/upload/USA_-_Average_Temperature_-2.jpg
1181876471 M * slacker403 yellow down in palm trees ;)
1181876472 M * slacker403 all that blue in the west is mountians also
1181876474 M * slacker403 coderanger: you ever been to miami ?
1181876482 M * doener I'm through :)
1181876485 M * coderanger slacker403: Yes, lived there for a year
1181876497 M * slacker403 good
1181876499 M * slacker403 student ?
1181876507 M * slacker403 you should see the downtown now 
1181876516 M * Bertl doener: okay, then get the older kernel :)
1181876524 M * coderanger slacker403: No, I was 9 at the time
1181876549 M * slacker403 the downtown is super sweet now 
1181876555 M * slacker403 all re did and i mean pretty much all
1181876672 M * slacker403 coderanger: where you live now ?
1181876791 M * coderanger Outside Boston
1181876800 M * slacker403 cold as shit 
1181877100 M * doener http://people.linux-vserver.org/~doener/delta-clean-nipquad_fmt.diff
1181877301 M * Bertl tx
1181877357 M * Bertl doener: btw, I uploaded an updated version of the mainline-clean (02), you might want to diff that and check it too ...
1181877410 M * doener later... whitespace review isn't exactly exciting ;-)
1181877468 M * Bertl hehe, yeah, I know ...
1181877637 M * Bertl doener: thanks for your time!
1181877647 M * doener np
1181877665 M * doener a welcome change from looking at the nmi watchdog for days ;-)
1181877682 M * Bertl ah? let's hear!
1181877701 M * slacker403 Bertl: how much you charge ?
1181877800 M * doener I've starting looking at various regressions and bugzilla entries for the kernel and started submitting debugging reports and/or patches. The latest area was the nmi watchdog, which saw a cleanup lately, but that unfortunately introduced a set of bugs.
1181877944 M * doener now, 3 patches fix bigger bugs (incl. oopses) and are pending at akpm, one fixes an older minor bug (no reaction so far), and another one fixes a regression when oprofile is used without the nmi watchdog (no reaction here either).
1181878015 M * doener given the fact that I had no clue about the stuff, and that I also tried to hunt down another regression that Ingo Molnar reported, you can probably tell that I spent ages on it ;)
1181878183 M * Bertl but you finally found and fixed it, no?
1181878266 M * doener 5 bugs, but not the regression that Ingo reported. I can't reproduce it either, and till now, Ingo didn't answer my request for his .config, so I've mostly given up on that. Reading the code and patches up and down 5 times is enough ;)
1181878275 M * Bertl doener: btw, do you have any clue why #include <linux/sched.h> keep popping up in several files in 2.6.22*
1181878325 M * doener as new #include's you mean? IIRC sched.h was removed from module.h
1181878353 M * Bertl ah, thanks, that makes sense
1181878581 M * slacker403 is thier a place with the newst debian packages ? i think someone posted it a bit ago 
1181878587 M * slacker403 cant find it in logs 
1181878692 M * Bertl for Linux-VServer, or in general?
1181878711 M * slacker403 vserver obviosuly ;) 
1181878760 M * Bertl waldi probably knows that ...
1181878776 M * Bertl micah would know too, but I'm not sure who's around atm
1181878780 M * slacker403 i was looking through logs 
1181878799 M * slacker403 i was trying to get the newest version i could get 
1181878838 M * slacker403 also you know where the config is to make bind listen on the host ? so i can run them on the guest as well
1181878923 M * Bertl check out: http://packages.debian.org/unstable/devel/
1181878938 M * doener wow, I just checked how big the current GPLv2/3 flamewar on lkml is... 400 messages in 6 days...
1181879105 M * Bertl will rs remove the GNU from Linux? :)
1181879231 M * doener He's probably more likely to resurrect Lignux (v3?) and complete the Hurd (after he killed the final boss in Duke Nukem Forever, of course)
1181879255 M * Bertl right ...
1181879282 M * Bertl what the hell is DM_DELAY for?
1181879377 M * doener It's description says "Useful for testing." :)
1181879383 M * doener s/It's/Its/
1181879403 M * doener A target that delays reads and/or writes and can send them to different devices.  Useful for testing.
1181879479 M * Bertl that is now the proof that the kernel folks keep making Linux slower ... and they do it on purpose! :)
1181880119 Q * coderanger Ping timeout: 480 seconds
1181882224 Q * Vudumen Ping timeout: 480 seconds
1181886610 M * Bertl okay, off to bed now ... have a good one everyone! cya!
1181886615 N * Bertl Bertl_zZ
1181887036 N * DoberMann_ DoberMann
1181889112 N * DoberMann DoberMann[PullA]
1181889869 J * dna ~naucki@115-247-dsl.kielnet.net
1181890720 J * dlezcano1 ~dlezcano1@AToulouse-252-1-89-132.w86-201.abo.wanadoo.fr
1181891829 J * phreak`` ~phreak``@deimos.barfoo.org
1181891831 Q * phreak`` 
1181891901 J * phreak`` ~phreak``@deimos.barfoo.org
1181893104 Q * ntrs__ Read error: Connection reset by peer
1181893143 J * ntrs ntrs@68-188-55-120.dhcp.stls.mo.charter.com
1181894717 Q * nkukard Quit: Leaving
1181896198 M * blizz do the VIRT_* cflags add significant overhead?
1181896473 M * blizz well, *enabling them
1181897205 M * arachnist http://images.teamsugar.com/files/users/1/13254/21_2007/german_0.preview.jpg 
1181897959 J * dna_ ~naucki@115-247-dsl.kielnet.net
1181898364 Q * dna Ping timeout: 480 seconds
1181899257 J * dreamind ~dreamind@p54A7A1A6.dip0.t-ipconnect.de
1181899289 N * dreamind Guest1953
1181899784 J * chand ~chand@212.99.51.254
1181900122 M * daniel_hozac blizz: some do, some don't. ;)
1181900180 M * daniel_hozac though i wouldn't call it significant...
1181902180 J * lilalinux ~plasma@dslb-084-058-213-243.pools.arcor-ip.net
1181902922 J * Punkie ~punkie@235-105-207-85.bluetone.cz
1181903458 J * ema ~ema@rtfm.galliera.it
1181903520 J * dna___ ~naucki@115-247-dsl.kielnet.net
1181903924 Q * dna_ Ping timeout: 480 seconds
1181904832 J * Piet hiddenserv@tor.noreply.org
1181905135 Q * ema Quit: leaving
1181906143 Q * chand Quit: chand
1181907481 Q * FireEgl Quit: Bye...
1181909152 J * chand ~chand@212.99.51.254
1181909182 Q * pusling Remote host closed the connection
1181909212 J * pusling pusling@88.212.70.38
1181909738 Q * Piet Remote host closed the connection
1181910456 J * Piet hiddenserv@tor.noreply.org
1181910499 Q * Piet 
1181912311 J * HeinMueck ~Miranda@host-88-217-199-211.customer.m-online.net
1181912455 M * HeinMueck Hi all!
1181912469 J * meetra45 ~xtreme@bl7-90-209.dsl.telepac.pt
1181912485 M * daniel_hozac hello HeinMueck!
1181912495 M * HeinMueck Anyone here who is a king of migrating existing servers to vservers on debian edge? :)
1181912514 M * daniel_hozac etch?
1181912529 M * HeinMueck I did a fresh install of a debian server - yeah, etch :)
1181912555 M * HeinMueck I created two fresh vservers with the debian tools and they are running fine
1181912602 M * HeinMueck then I rsynch'ed an existing suse server to another the new server and created the configuration
1181912668 M * HeinMueck when I start it I get "vcontext: chroot(): Permission denied"
1181912717 M * daniel_hozac where did you rsync it to?
1181912733 M * daniel_hozac what kernel are you using?
1181912742 M * HeinMueck I checked the rights on /var/lib/vservers and lsattrib did not show any weird stuff
1181912772 M * daniel_hozac you mean showattr?
1181912772 M * HeinMueck debian kernel 2.6.18-3-vserver-686
1181912789 M * daniel_hozac IIRC, that one is known to be broken, you want at least 2.6.18-4.
1181912868 M * HeinMueck Ah - just made the update this morning - so let me reboot :)
1181912904 M * daniel_hozac interesting that you managed to get that one at all.
1181912913 M * daniel_hozac i thought etch shipped with 2.6.18-4.
1181912919 Q * meetra Ping timeout: 480 seconds
1181912975 N * meetra45 meetra
1181913272 M * HeinMueck Hm
1181913316 M * HeinMueck now having 2.6.18-4-vserver-686 it still says permission denied - that wasn't it
1181913317 M * HeinMueck pitty
1181913347 M * daniel_hozac no, but 2.6.18-3 will crash when you stop guests...
1181913361 M * HeinMueck I tried to wrap the vserver stuff in strace, but it just shows me that something ends with permission denied :)
1181913410 M * daniel_hozac so what does showattr -d /path/to/rsynced/guest show?
1181913418 M * HeinMueck I just wonder which file or directory it has  a problem with
1181913421 M * HeinMueck just a sec
1181913436 Q * Punkie Quit: Leaving
1181913487 M * HeinMueck marvin:~# showattr -d /var/lib/vservers/*
1181913487 M * HeinMueck ---buiX /var/lib/vservers/arch
1181913487 M * HeinMueck ---buiX /var/lib/vservers/balu
1181913503 M * HeinMueck arch is the fresh one, balu the rsynched
1181913583 M * daniel_hozac what xid have you set for that guest?
1181913592 M * daniel_hozac and are the files/directories tagged with that?
1181913678 M * HeinMueck I did nothing special on this at all
1181913705 M * HeinMueck can I see if there is one set on arch?
1181913714 M * daniel_hozac lsxid /var/lib/vservers/*
1181913778 M * HeinMueck ah
1181913784 M * HeinMueck arch has 0 on all dirs
1181913792 M * HeinMueck but balu has 65535
1181914015 M * HeinMueck cool
1181914033 M * HeinMueck I did reset the xid to 0 for balu too and now it seems willing to start
1181914073 M * HeinMueck at least it can chroot and starts the rc script!
1181914297 Q * kaner Remote host closed the connection
1181914302 M * HeinMueck hmm, now I get "vshelper.init: can not determine xid of vserver 'balu'; returned value was ''
1181914327 M * HeinMueck I've seen this when I tried to migrate some legacy servers to the new config layout.
1181914895 M * HeinMueck I did set the init style to plain now and at least I can enter the vserver and start some services manually
1181914918 M * HeinMueck which is good - at least :)
1181916104 Q * Guest1953 Quit: Guest1953
1181916323 M * HeinMueck but anyway, thanks daniel for getting me this far
1181916363 M * HeinMueck it seems that the rc script just exits because it expects RUNLEVEL and PREVLEVEL to be set
1181916400 M * HeinMueck If I get those defined I might be able to find my way through
1181916972 M * matti :)
1181917026 Q * Aiken Quit: Leaving
1181917085 N * dna___ dna
1181917137 J * kaner kaner@strace.org
1181918577 J * sauron ~zeus16384@12.20.4.100
1181918583 M * sauron hi all
1181918630 M * sauron guys, anyone heard of OpenVZ
1181918763 N * Bertl_zZ Bertl
1181918774 M * Bertl sauron: sure 
1181918786 M * Bertl morning folks!
1181918952 M * Bertl HeinMueck: what exactly are you trying to do?
1181919055 M * Bertl HeinMueck: xid is a value between 2 and 49151 (you have to choose an unique one from that range, 0 and 1 are special)
1181919096 M * Bertl HeinMueck: very recent tools (nothing in default Debian, I guess :) will select the xid for you on guest creation
1181919124 Q * chand Quit: chand
1181919279 M * daniel_hozac (if not specified)
1181919282 M * daniel_hozac morning Bertl!
1181919292 M * Bertl hey daniel_hozac! how are you?
1181919311 M * daniel_hozac fine thanks, and you?
1181919331 M * Bertl I think ... okay :)
1181919342 M * Bertl are you already in the US?
1181919350 M * daniel_hozac no, i'm leaving on monday morning.
1181919364 M * Bertl ah, i.c. so some time left then :)
1181919388 M * daniel_hozac yep :)
1181919400 M * Bertl could you motivate yourself to have a look at the cleanups yet?
1181919501 M * daniel_hozac sure.
1181919506 M * HeinMueck Hi Bertl! I tried to switch a Suse 10 from dedicated hardware to a vserver on a different machine. 
1181919530 M * HeinMueck So I rsynched it from a to b and now I try to make it run there :)
1181919546 M * Bertl okay, first, how did you create the guest config?
1181919604 M * daniel_hozac HeinMueck: you might want to look at using /etc/vservers/<guest>/apps/init/environment
1181919609 M * HeinMueck using a working server as a template (read only :)) and having the great flower open :)
1181919642 M * HeinMueck I did now set the environment file and give the two runlevel vars there
1181919645 M * daniel_hozac you're aware of vserver ... build -m rsync/skeleton, right?
1181919661 M * Bertl is that in debian?
1181919670 M * HeinMueck I removed all links from rc3.d apart from the sshd and I can now start the server as expected
1181919670 M * Bertl (the reync I mean)
1181919675 M * daniel_hozac yeah.
1181919685 M * daniel_hozac clone is the only one not in Debian, IIRC.
1181919693 M * Bertl ah, good to know
1181919755 M * HeinMueck it just seems that I miss some little bits; the hostname is not set and when I stop the server it does not catch the new runlevel 0 - so gives some errors
1181919897 M * Bertl why not run it with 'plain' init style if suse has a special init running? (but listen to daniel_hozac, he should know better :)
1181919942 M * HeinMueck I tried plain, but then it does not start anything
1181919960 M * Bertl that sounds strange ...
1181919990 M * Bertl are you sure that it didn't start anything? how did you check?
1181920031 M * HeinMueck enter the vserver and do ps ax 
1181920044 M * HeinMueck I could start processes with /etc/init.d/xxx start
1181920085 M * Bertl hmm, suse must be doing strange things then ...
1181920100 M * HeinMueck thats for sur e.)
1181920112 M * HeinMueck I have to leave to get home before it starts raining
1181920119 M * Bertl okay, cya
1181920121 M * HeinMueck so thanks again and I may be back :)
1181920134 Q * HeinMueck Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1181920751 M * Bertl doener: have you seen massive 'section mismatch' warnings on 2.6.22-rc4?
1181921944 M * sauron Bertl, what would you say , go for openVZ or for vserver? I dont know much abouyt openVZ but I read that it has a steeper learning curve. I am still at a loss how these two came about
1181921949 J * lada ~kvirc@area1-133-gw.hkfree.org
1181922024 M * Bertl welcome lada!
1181922041 M * lada hi :-)
1181922057 M * Bertl sauron: it is quite simple, SWsoft is selling a commercial product called Virtuozzo(tm) for some time now ... 
1181922132 M * sauron Bertl isn't some version of it also free>
1181922146 M * Bertl sauron: half a year ago (or maybe more) they decided to release essential parts to the open source community, but they are still the only contributors/maintainers to their OpenVZ branch (which is told to be the core of Virtuozzo(tm) too)
1181922167 M * sauron i see... so its really not fully open source then
1181922201 M * lada Could someone tell me a little about Linux-VServer and SELinux? I'd like to know if it is theoretically possible to have them both in one kernel...
1181922205 M * Bertl sauron: IMHO the motivations for doing so could be: pending GPL infringements, out-sourcing of development and testing to the community, competition :)
1181922223 M * Bertl lada: sure, they are not exclusive
1181922245 M * sauron Bertl got you
1181922285 M * lada well, every guide says "turn off SELinux" :-) i know that it causes kernel panic...
1181922307 M * lada but I am just interested why this happens?
1181922411 M * sauron I have a question. I read that vserver nicely shares program text (executables, libraries loaded into memo). but what happens if I start an exe up in a few vservers, then change its configuration on one of them and start it up again so it uses different memory allocation etc. will that be mirrored to the other vservers' memory, or is that any way possible (or they all have to be restarted)?
1181922463 M * sauron and I guess the exe has to be unified for the sharing to happen right?
1181922472 M * Bertl lada: if a mainstream kernel, patched with Linux-VServer panics with SELinux turned on, and this doesn't happen without the Linux-VServer patch, then we consider this a bug, and you should report it :)
1181922530 M * Bertl sauron: yes, files have to be unified, and the mapping has absolutely no relation to your guest configuration
1181922557 J * sladen paul@starsky.19inch.net
1181922563 M * sauron Bertl by the way, in terms of executable memory sharing....
1181922566 M * Bertl sauron: so, there will be no effect on other guests too
1181922573 M * sauron does it only share the program text really?
1181922581 M * sauron none of the data part, even if that's say static?
1181922598 M * Bertl it shares whatever is mapped from files
1181922605 A * sauron doesn't know details about how executable is structured in memoruy
1181922627 M * Bertl so if a big read-only data section comes from the file, it will be shared too
1181922639 M * sauron Bertl ok... so there is really no memory sharing of say, some area that the exe created and it will always be static. just what was "mapped" from the exe on loadup
1181922661 M * Bertl and from other files mapped into memory of course
1181922673 M * sauron interesting
1181922713 M * daniel_hozac lada: the reason every guide says to turn it off is because nobody has bothered to create the needed policy yet.
1181922731 M * sauron because yesterday I brought my system to its knees because I had to run 10 servers and each 2-3 executables that are each about 210MB VS, 30MB RS, and my sys is only 256MB ram and 256MB swap. and I ran out of swap even
1181922760 M * sauron and was thinking how to optimize to get the needed extra 20-30MB / vserver
1181922777 M * Bertl well, unification can help there a lot
1181922788 M * sauron Bertl yes its already unified
1181922793 M * slacker403 anyone up yet who knows where those debian packages would be for vserver ?
1181922815 M * sauron slacker I think you can download them with apt-get just fine
1181922835 M * sauron apt-cache search util-verver  (or vserver-utils, I think they swapped the name)
1181922852 M * Bertl IIRC, he is looking for updated packages
1181922858 M * lada daniel_hozac: ok, so that means that SELinux should work, but the default policies cause problems?
1181922880 M * Bertl lada: on the host, no, on the guest, very likely
1181922899 M * daniel_hozac slacker403: sid.
1181922946 M * lada I also noticed that there is a Linux-VServer + grsecurity patch so I thought that a somekind of a SELinux patch would be needed... I looked in the mailing lists and nobody asked about SELinux :-) so I am just curious
1181923005 M * Bertl lada: well, SELinux is part of mainline (at least something called SELinux) so no patch required
1181923013 M * Bertl grsec OTOH, is not
1181923028 M * lada would you know about a website where I could find more? like which SELinux policies might cause problems etc?
1181923043 M * lada Bertl: I see, it makes sense :)
1181923051 M * daniel_hozac there is none tailored for use on a host...
1181923065 M * daniel_hozac at least, not AFAIK.
1181923090 M * daniel_hozac if you create one, do share.
1181923108 M * lada well, i am not a SELinux expert :)
1181923120 M * slacker403 daniel_hozac: sid ?
1181923200 N * ensc Guest1983
1181923210 J * ensc ~irc-ensc@p54B4EBAB.dip.t-dialin.net
1181923283 M * daniel_hozac slacker403: http://www.us.debian.org/doc/manuals/debian-faq/ch-ftparchives.en.html#s-sid
1181923317 Q * Guest1983 Ping timeout: 480 seconds
1181923347 M * lada thank you for the info, very much appreciated. you both were very helpful :-)
1181923366 M * Bertl you're welcome!
1181923388 M * lada good bye :-)
1181923407 P * lada Time makes no sense
1181923452 M * slacker403 daniel_hozac: does sid have the newest packages ?
1181923452 J * fatgoose ~samuel@76-10-152-94.dsl.teksavvy.com
1181923462 M * daniel_hozac slacker403: yes.
1181923464 M * Bertl welcome fatgoose!
1181923484 M * fatgoose bonjour!
1181923559 J * HeinMueck ~Miranda@dslb-088-064-029-026.pools.arcor-ip.net
1181923569 M * Bertl wb HeinMueck!
1181923571 M * slacker403 daniel_hozac: i meant newest of vserver to be more precise 
1181923583 M * HeinMueck Hi Bertl!
1181923589 M * HeinMueck Made it before the storm :)
1181923595 J * stefani ~stefani@tsipoor.banerian.org
1181923602 M * Bertl morning stefani!
1181923611 M * stefani evening!
1181923612 M * daniel_hozac slacker403: well, that's doubtful as Bertl just uploaded that a day or two ago...
1181923654 M * Bertl slacker403: but recent patches/versions should be there
1181923675 M * slacker403 i jus want a relatively new versio
1181923704 M * daniel_hozac should be the somewhat broken 2.6.21 version, IIRC.
1181923782 M * slacker403 sounds promising 
1181923786 M * slacker403 somewhat broken
1181923963 M * fatgoose I want to script something that watch resources consumption of my vservers, is reading /proc/virtual/ the right way to do it?
1181924029 M * Bertl as human, yes
1181924054 M * Bertl if you want periodic readings (maybe even at low intervals)
1181924083 M * daniel_hozac you want to use the APIs.
1181924088 M * Bertl you should better use the syscall interface 
1181924105 M * fatgoose ok
1181924128 M * daniel_hozac Bertl: the kernel/sched_hard.h:vx_schedule hunk, should there be a space after flags?
1181924129 M * fatgoose I want to use the apis
1181924247 M * Bertl daniel_hozac: no, tx!
1181924443 M * slacker403 daniel_hozac: whats broken in the debian ones ?
1181924478 M * Bertl I think they are missing a mainline fix handling the sysfs issues and one vserver patch
1181924483 M * daniel_hozac slacker403: nothing Debian-specific, just the general mainline issues (i guess, i don't know if those are fixed in Debian's packages), and the tty issue in vservr.
1181924486 M * daniel_hozac +e
1181924525 M * Bertl daniel_hozac: I found a few other of those 'flags ,' please consider them fixed
1181924542 M * daniel_hozac okay.
1181924668 M * daniel_hozac btw, are the device mapping patches in 2.3 yet?
1181924693 M * Bertl yes, we have something there, but I'm not sure how complete it is
1181924720 M * daniel_hozac okay.
1181924772 M * Bertl I see the cmd interface ....
1181924867 M * daniel_hozac that appears to be all of it though.
1181924903 M * Bertl could be, I still haven't updated 2.3 yet
1181924911 M * daniel_hozac right.
1181924959 M * Bertl the xfs issue with sendfile .. did you investigate it a little?
1181924976 M * daniel_hozac not any further than finding the non-generic sendfile.
1181925011 M * Bertl okay, we had a report of somebody using joe? for editing, and that caused the files to become zeroed out when edited
1181925029 M * daniel_hozac yeah.
1181925032 M * Bertl (on xfs too, and I think that is related, yes?)
1181925058 M * daniel_hozac i guess so, as the COW-link breaking is broken.
1181925068 M * daniel_hozac (and opening a file in an editor should break the link)
1181925085 M * Bertl okay, when I test xfs on recent kernels with debugging enabled
1181925093 M * Bertl it bails out with a double locking
1181925116 M * Bertl could that be the reason?
1181925165 M * daniel_hozac i guess that would make sense...
1181925183 M * Bertl I'll upload the trace, sec
1181925233 M * Bertl http://paste.linux-vserver.org/2579
1181925240 J * FireEgl FireEgl@2001:5c0:84dc:1:ff:39ff:fed1:1f2d
1181925486 M * daniel_hozac so are we abusing VFS, or is this an XFS issue?
1181925499 M * daniel_hozac i mean, is it at all reproducible without COW?
1181925524 Q * lilalinux Remote host closed the connection
1181926496 M * Bertl well, the copy via sendfile is nothing mainline provides
1181926509 M * Bertl so, yes, we are doing something nobody expected :)
1181926542 M * Bertl note that Al Viro is actively blocking all kernel side copy mechanisms for a long time now :)
1181926601 M * Bertl but I think the main issue is from the attribute change
1181926628 M * Bertl would be interesting to compare a CoW link break of 3 links, vs 2 links
1181926806 N * fatgoose samueltc
1181926832 M * daniel_hozac that trace is of 3 links, no?
1181926881 M * Bertl not sure, it is from testfs.sh
1181926926 M * daniel_hozac or, hmm. what's the difference between 2 and 3 links?
1181926948 M * daniel_hozac i thought we handled 1 specially, and the rest were all handled the same?
1181926955 M * Bertl not sure my theory holds ... but we change the flags on 2 links
1181926972 M * Bertl 2 links get broken to 2 files with 1 link each
1181926984 M * Bertl IIRC, we change the flags there
1181926989 M * daniel_hozac "change the flags" means what?
1181927011 M * daniel_hozac as i recall, we simply remove the flags if there's just 1 link (i.e. nobody else is depending on this file).
1181927061 M * daniel_hozac for 2+ links, the process should be the same (break the link for this dentry).
1181927135 M * sauron thanks guys, later
1181927201 M * Bertl daniel_hozac: we mean the same thing, just call it different
1181927220 M * Bertl daniel_hozac: 2 links for me is  a-b linked together
1181927227 M * Bertl (before they are broken)
1181927228 M * daniel_hozac right.
1181927240 M * daniel_hozac same here.
1181927271 M * Bertl okay, so we leave a 'unified' file around when we break that?
1181927302 M * daniel_hozac yeah.
1181927312 M * daniel_hozac and when that is written to, the attributes are removed.
1181927314 M * Bertl okay, then I remembered that wrong ...
1181927347 P * sauron 
1181927789 M * samueltc hey, why you did not tell me about VCD?
1181927963 M * Bertl we didn't?
1181928308 J * awk ~awk@kia.inet-corp.com
1181928317 M * Bertl wb awk!
1181928902 M * Bertl daniel_hozac: do you have 2.6.20/21 Linux-VServer system somewhere?
1181929393 M * Bertl nevermind ... found a way to recreate the issue with QEMU
1181929393 P * marcfiu 
1181929623 Q * sladen Ping timeout: 480 seconds
1181929889 M * bXi i'm gonna start development of a good tool for administrating vservers soon
1181929902 M * bXi even the option to create them from an interface
1181930046 M * bXi even better i'll start now
1181930059 J * eSa| ~kvirc@ip-87-238-2-45.adsl.cheapnet.it
1181930098 M * Bertl bXi: sounds good, maybe take a look at existing solutions first though, just to know what you shouldn't do :)
1181930105 M * Bertl wb eSa|!
1181930150 M * eSa| hi all
1181930187 M * bXi Bertl: i was thinking of being able to upload the tarballs :P
1181930451 J * sladen paul@starsky.19inch.net
1181930902 M * eSa| hi
1181931030 M * Bertl eSa|: hi hi! :)
1181931120 M * Bertl daniel_hozac: how do I create a new ipc space with 0.30.213?
1181931229 N * eSa| eSa|away
1181932564 Q * samueltc Ping timeout: 480 seconds
1181932833 J * onox ~onox@kalfjeslab.demon.nl
1181932865 J * duckx ~Duck@tox.dyndns.org
1181932875 M * Bertl okay, off for now ...
1181932880 N * Bertl Bertl_oO
1181933819 J * ktwilight_ ~ktwilight@236.197-66-87.adsl-static.isp.belgacom.be
1181934228 Q * ktwilight Ping timeout: 480 seconds
1181935099 Q * eSa|away Ping timeout: 480 seconds
1181935121 J * eSa| ~kvirc@ip-87-238-2-45.adsl.cheapnet.it
1181935871 J * fatgoose ~samuel@76-10-152-94.dsl.teksavvy.com
1181937454 J * Vudumen ~vudumen@perverz.hu
1181937866 M * daniel_hozac bXi: you've seen openvcp and openvps, right?
1181938008 M * daniel_hozac Bertl_oO: it's implicit in the context creation, to better mirror the older versions.
1181938459 J * fatgoose_ ~samuel@76-10-152-94.dsl.teksavvy.com
1181938459 Q * fatgoose Read error: Connection reset by peer
1181938522 M * Bertl_oO daniel_hozac: so no way to explicitely create an ipc namespace and/or to join an ipc namespace of a guest?
1181938619 M * eSa| is there some facilty to search in the logs save @ http://irc.13thfloor.at/ ?
1181938808 M * Bertl_oO eSa|: google?
1181938815 M * daniel_hozac Bertl_oO: API-wise or command line? the former should work, the latter is missing right now.
1181938833 M * Bertl_oO daniel_hozac: okay, tx, will use vcmd for now
1181938853 M * daniel_hozac okay.
1181939030 M * eSa| I'll post my question: today i made a 2.6.20.11 kernel +vs2.3.0.12 development
1181939041 M * eSa| I want to try this interesting vserver thing
1181939059 M * Bertl_oO okay, then please start with the stable branch
1181939075 M * Bertl_oO i.e. get 2.6.20.14 and vs2.2.0 
1181939096 M * eSa| you don't want to give me a chance? :)
1181939123 M * daniel_hozac 2.3 is missing quite a bit right now.
1181939134 M * Bertl_oO http://vserver.13thfloor.at/Experimental/patch-2.6.20.14-vs2.2.0.diff
1181939247 M * eSa| hmm.. I see your point
1181939278 M * Bertl_oO nevertheless, please go on with your question
1181939293 M * eSa| this time I've been the cock 
1181939309 M * eSa| well, the error is like this one i found googlin'
1181939313 M * eSa| http://paste.linux-vserver.org/2564
1181939326 M * eSa| except my machine name is 'test0'
1181939349 M * Bertl_oO how was the guest created?
1181939361 M * daniel_hozac did you check the FAQ?
1181939386 M * daniel_hozac http://linux-vserver.org/Frequently_Asked_Questions#I_now_get_errors_like_.22ncontext:_vc_net_create.28.29:_Invalid_argument.3B_dynamic_contexts_disabled..22_on_startup._Vservers_are_not_started
1181939397 M * eSa| I did it with newvserver @ command line. I've followed instructions on the Debian wiki
1181939416 M * Bertl_oO which debian wiki? newvserver is deprecated
1181939439 M * matti Hi Bertl
1181939524 M * eSa| http://wiki.debian.org/LinuxVserver
1181939544 M * eSa| btw debian packages seems to be a bit outdated
1181939562 M * Hollow i have backported 0.30.213 from sid
1181939577 M * Bertl_oO eSa|: we know, but please tell them Debian folks!
1181939595 M * daniel_hozac Hollow: is it on backports.org?
1181939618 M * Hollow dunno, my backport is at http://debstager.newthinking.de/dists/etch/main/binary-amd64/net/util-vserver_0.30.213-1_amd64.deb
1181939698 M * Bertl_oO is there a slight chance that we can get all debian packages on ftp.linux-vserver.org or so?
1181939715 M * Bertl_oO I mean the Linux-VServer related and updated ones :)
1181939745 M * Hollow yeah, we could do so, but even etch has 0.30.212...
1181939751 M * Hollow that's pretty good for debian :)
1181939754 M * Bertl_oO micah, waldi: maybe you can help us here?
1181939784 M * daniel_hozac is there any point in doing a separate repository?
1181939812 M * Bertl_oO IMHO it would save _us_ quite some time on arguing and searching those packages ourselves
1181939828 M * Bertl_oO and it would definitely ease the life of those folks using debian, no?
1181939854 M * Hollow i think they do quite a good job lately.. they just need to remove this vserver-debiantools package ..
1181939873 M * daniel_hozac i suppose, but we don't have buildd's for all the arches available, AFAIK.
1181939905 M * Bertl_oO I'm not saying the maintainers are doing a bad job, but folks keep coming here with 0.30.210 and earlier and kernels as old as they are :)
1181939912 M * daniel_hozac whereas e.g. backports.org has all of the required infrastructure ready.
1181939933 M * daniel_hozac there's no reason anybody would be running 0.30.210 anymore.
1181939941 M * Bertl_oO then maybe make a wiki page called 'HowToFixYourDebian' ?
1181939946 M * Hollow Bertl_oO: well, do we want to backport for any ancient debian release because it cannot be updated without pita?
1181939981 M * Bertl_oO AFAIK, all older debian (distros) can handle recent tools, no?
1181939996 M * daniel_hozac wasn't there a beecrypt problem on sarge?
1181939997 M * Bertl_oO and the same goes for (most) of the kernels, at least for 2.6.19
1181940003 M * Hollow you'd still need a seperate *.deb ..
1181940010 M * Hollow for each release
1181940022 M * Bertl_oO I'm pretty sure they are already there ...
1181940032 M * Bertl_oO we just have to puzzle them together
1181940095 M * Hollow well, setting up a debian pool is rather easy, but i'm not sure we want that :)
1181940153 M * Bertl_oO let's wait for a comment from the debian maintainers
1181940319 Q * dna Quit: Verlassend
1181940330 M * daniel_hozac ensc: ping, any way i could convince you to branch util-vserver for EL-5 (and maybe 4, if that's allowed now?) in Fedora?
1181940427 M * waldi Bertl_oO: what?
1181940476 M * Bertl_oO waldi: hey, we are thinking about having a central place for recent debian packages (tools and kernels)
1181940498 M * Bertl_oO waldi: or at least some wiki page pointing to them
1181940565 M * waldi hmm
1181940596 M * Bertl_oO at least in my experience, it always goes like this:
1181940605 M * daniel_hozac do the kernels have any versioned userspace requirements? or can you install those packages pretty much anywhere?
1181940621 M * Bertl_oO - hello, I'm using debian xy ... and I installed a guest with newvserver
1181940625 M * daniel_hozac (i.e. do they need backports due to dependencies, or just for the different compiler?)
1181940637 M * Bertl_oO - don't do that! what kernel/tools do you have?
1181940641 M * waldi daniel_hozac: they have
1181940655 M * waldi and the kernel team have some ideas to handle that
1181940658 M * Bertl_oO - kernel <ancient> and tools 0.30 or so ...
1181940671 M * Bertl_oO - okay, please update to something more recent ...
1181940680 M * Bertl_oO - but there is none in stable debian xy
1181940701 M * Bertl_oO and then I would like to suggest a place where they could get the packages from
1181940749 M * Bertl_oO IMHO it also would not hurt to list the broken packages and kernel issues on some page, so that we can simply point folks there
1181940771 M * Bertl_oO i.e. you have kernel x and tools y, so expect the following ...
1181940802 M * Wonka debian stale... *sigh*
1181940813 A * Wonka always uses self-built kernels
1181940816 M * arachnist debian museum
1181940817 M * arachnist ;>
1181940829 M * dilinger Bertl_oO: hey, any news on that patch?
1181940866 M * Bertl_oO dilinger: there is a test version available, but it is basically 2.2.0 for your branch
1181940901 M * dilinger Bertl_oO: ok.  did you happen to talk to coderanger?
1181940905 M * Bertl_oO dilinger: if you like, you can get that, and an incremental patch with the missing features and/or the removals
1181940922 M * dilinger Bertl_oO: i'm not in a rush (i've got plenty of other stuff to do), but he was
1181940924 M * Bertl_oO the final version should be complete this weekend
1181940950 M * Bertl_oO yeah, when he shows up, today, I'll get him the test patch
1181940957 M * dilinger ok, cool
1181940965 A * dilinger doesn't see him in the office yet, i'm sure he'll be in later
1181941035 M * Bertl_oO np, maybe you want to skim over the latest patches and comment on kernel coding style issues?
1181941069 M * daniel_hozac ah right, thanks for reminding me.
1181941077 M * Bertl_oO dilinger: http://vserver.13thfloor.at/Experimental/patch-2.6.22-rc4-vs2.2.0-rc3.diff
1181941170 M * Bertl_oO (any feedback would be appreciated)
1181941341 M * waldi Bertl_oO: common place for latest debian kernels is http://kernel-archive.buildserver.net/, but no deps are currently done
1181941459 M * dilinger Bertl_oO: thanks.  i need to run out for a bit, but i'll check it out later
1181941471 M * Bertl_oO dilinger: np, thank you!
1181941481 M * daniel_hozac Bertl_oO: btw, does the CPU limiting work with CONFIG_NO_HZ?
1181941497 M * Bertl_oO yes, last time we checked it did
1181941514 M * daniel_hozac okay, good, thanks.
1181941535 M * Bertl_oO but I will check that this weekend anyways
1181941635 M * daniel_hozac Bertl_oO: the rest of delta-mainline-clean02 looks good to me.
1181941647 M * Bertl_oO excellent! tx!
1181941678 M * Bertl_oO daniel_hozac: btw, if you feel like it, the 2.6.22-rc4 version is leaking nsproxies
1181941711 M * Bertl_oO can be triggered by vnamespace ... chcontext ... sleep 10
1181941731 M * daniel_hozac hmm, okay.
1181941732 M * Bertl_oO doesn't happen without the vnamespace ...
1181941755 M * daniel_hozac what vnamespace command?
1181941759 M * daniel_hozac just --new?
1181941764 M * Bertl_oO yep
1181941798 M * daniel_hozac that sounds like mainline to me, unless i'm missing something obvious.
1181941799 M * Bertl_oO note that the vnamespace itself isn't triggering it either
1181941819 M * Bertl_oO yeah, it is definitely the result of the nsproxy changes in mainline
1181941851 M * Bertl_oO they do strange things now, like returning with a reference to the old proxy and such
1181941881 M * Bertl_oO I think we have to drop that when switching a task over or so
1181941921 M * daniel_hozac but that command shouldn't invoke any of our namespace touching code, IMHO.
1181941948 M * Bertl_oO the chcontext will copy it, no? and the task when migrated will get it
1181941963 M * daniel_hozac or, wait, chcontext will do IPC and uts.
1181941985 M * daniel_hozac i assume it doesn't happen with vcmd, right?
1181941992 M * Bertl_oO used older tools, they only affected the mnt space
1181941997 M * daniel_hozac oh.
1181942001 M * Bertl_oO haven't had the time to check with vcmd yet
1181942016 M * Bertl_oO currently not in reach of my normal workspace :)
1181943248 M * HeinMueck Hey Guys!
1181943272 M * HeinMueck Perhaps you might want to solve an easy one for a change :)
1181943289 M * HeinMueck just managed to get back to the Suse server again
1181943333 M * HeinMueck It seems they rely on the both environment vars PREVLEVEL and RUNLEVEL to run the rc script
1181943347 M * HeinMueck man init says it's set by init ...
1181943386 M * HeinMueck so the problem is that if I define them in apps/init/environment it works for the startup, but not for shutdown
1181943386 M * daniel_hozac but, not so much? or?
1181943395 M * daniel_hozac ah.
1181943458 M * daniel_hozac and if you used a plain initstyle, services wouldn't start anyhow?
1181943477 M * HeinMueck what would be the best way to make the rc script find those? in the great flower I read about cmd.prepare - would that be a place?
1181943524 M * daniel_hozac with the sysv initstyle, that should already be doing the right thing (i.e. if you execute runlevel inside the guest, that should have the correct information)
1181943595 M * Bertl_oO HeinMueck: tool version?
1181943700 M * HeinMueck Debian 0.30.212-1  ;-)
1181943791 M * HeinMueck when I do vserver balu start I can see that both of the vars are not defined
1181943815 M * daniel_hozac what do you have in /etc/vservers/balu/apps/init/environment?
1181943862 M * HeinMueck I had
1181943862 M * HeinMueck PREVLEVEL=0
1181943862 M * HeinMueck RUNLEVEL=3
1181943888 M * HeinMueck with that, the vserver starts up fine
1181943901 M * daniel_hozac but you can't stop it?
1181943901 Q * onox Quit: leaving
1181943946 M * HeinMueck when you do vserver balu stop, the current runlevel become the PREVLEVEL, so it's three. But the RUNLEVEL is empty
1181944012 M * HeinMueck that makes part of the scripts fail, no services get stopped
1181944068 J * coderanger_ ~laptop@wireless-251.media.mit.edu
1181944070 N * coderanger_ coderanger
1181944079 M * daniel_hozac okay, makes sense.
1181944095 M * daniel_hozac coderanger: about that API, are the calls to the functions implicit in the assignment?
1181944103 M * HeinMueck eval $(startpar $startopt -M stop  -P $PREVLEVEL -R $RUNLEVEL)
1181944119 M * daniel_hozac and if so, wouldn't it make more sense to call it .add and .remove instead of .flagword and .mask?
1181944149 M * HeinMueck thats what fails
1181944157 M * coderanger daniel_hozac: Wort of
1181944161 M * coderanger er, Sort of
1181944177 M * coderanger daniel_hozac: If you just do ctx.foo.blah = 0, it does the syscall right then
1181944193 M * Bertl_oO coderanger: if you want something to test, there is a patch available 
1181944218 M * Bertl_oO coderanger: it is not final, but it should do the stuff you are working on
1181944218 M * coderanger daniel_hozac: But if you want to combine multiple sets, you can wrap the block in the with ctx:, and it only calculates and executes the syscalls at the end of the block
1181944311 M * daniel_hozac hmm, okay.
1181944445 Q * fatgoose_ Quit: fatgoose_
1181944483 M * coderanger Bertl_oO: I'm not the one doing the kernel merge, and I would guess dilinger would want to wait for a final version of some form
1181944514 M * coderanger daniel_hozac: I am also going to add some kind of transactional semantic to the with block, so if one syscall in the middle fails, it makes some attempt to rollback one layer
1181944518 M * Bertl_oO okay, just wanted you to know that it would be available for you
1181944534 M * coderanger Bertl_oO: Thanky :)
1181944553 M * daniel_hozac coderanger: sounds nice.
1181944682 Q * tokkee Ping timeout: 480 seconds
1181945440 J * huiz ~gerrit@207.47.10.130.static.nextweb.net
1181945550 Q * HeinMueck Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1181945708 J * Aiken ~james@ppp121-45-192-48.lns1.bne1.internode.on.net
1181945731 M * Bertl_oO morning Aiken!
1181945763 M * Aiken hi
1181946596 M * eSa| daniel_hozac: thanks for your hint. Adding the context avoided the ncontext: vc_net_create() error 
1181946653 M * eSa| btw the debian related documentation on vserver wiki  seems outdated too
1181946662 M * daniel_hozac it is.
1181947298 M * eSa| should I use something else instead of newvserver to create guests?
1181947330 M * Bertl_oO  eSa| yes, vserver - build 
1181947345 M * Bertl_oO  eSa| check out 'vserver - build --help'
1181947364 M * Bertl_oO  eSa| '-' would be replaced by the guest name, of course
1181947430 M * eSa| ah ok, now I see those information following the 'generic' linux 2.6 installation pages
1181947473 M * Bertl_oO eSa|: the advantage of using the mainline 'vserver' command is that you can build all kind of guests with it, and it works on all distros
1181947495 M * Bertl_oO (while the newvserver script is quite debian specific, and often outdated)
1181947508 M * eSa| this is definetely what I needed to know http://linux-vserver.org/Building_Guest_Systems
1181947550 M * eSa| Bertl_oO: that's nice. so I may finally try out those pesky gentoo and arch linux
1181947574 M * Hollow eSa|: be sure to use 0.30.213 with gentoo :)
1181947601 M * Hollow see also http://www.gentoo.org/proj/en/vps/vserver-howto.xml
1181947955 P * click [IRSSI]
1181948047 M * dilinger Bertl_oO: why is the vserver syscall called sys_ni_syscall?
1181948421 M * Bertl_oO it isn't :)
1181948434 M * Bertl_oO sys_ni_syscall means not implemented
1181948450 M * Bertl_oO we change that to sys_vserver in our patches
1181948471 M * dilinger ah, ok
1181948779 Q * huiz Quit: Client exiting
1181949502 Q * ensc Ping timeout: 480 seconds
1181950072 P * stefani I'm Parting (the water)
1181950311 M * coderanger What does vc_net_migrate do?
1181950329 M * Bertl_oO  move a process into a network space/context
1181950360 M * Bertl_oO (similar to vc_ctx_migrate for the pid/process space/context)
1181950360 M * coderanger It only takes a single arg, so I take it that it only operates on the current PID?
1181950370 M * Bertl_oO yep
1181950499 M * coderanger Tell me if any of this is wrong re: the basic way to create a container:
1181950540 M * coderanger an admin-level program or daemon gets a request to containerize a program (probably usually some kind of init script that launches a system)
1181950545 M * coderanger It forks
1181950555 M * coderanger makes a new context/net context
1181950561 M * coderanger configure them
1181950571 M * coderanger migrates to both of them
1181950577 M * coderanger and execs the program in question
1181950625 M * daniel_hozac sounds like a plan.
1181950640 M * Bertl_oO yep, though fork is usually clone or requires unshare
1181950667 M * coderanger okay, just to make sure nothing an cross the chroot barrier no?
1181950686 M * daniel_hozac no, to get a new IPC/uts/filesystem namespace.
1181950709 M * daniel_hozac nothing in a non-admin non-watch context can ever cross the barrier.
1181950744 M * coderanger okay
1181950772 M * Bertl_oO in the distant future, we will end up with unsharing most of the spaces
1181950797 M * daniel_hozac right.
1181950811 M * daniel_hozac so you may want to use vc_get_space_mask to figure out what spaces to unshare/clone.
1181950830 A * coderanger looks that up
1181950913 M * coderanger Is the namespacing stuff explained anywhere?
1181950982 M * daniel_hozac what do you mean?
1181951064 M * coderanger How you use the namespace calls to control things and such
1181951082 M * coderanger There are no comments in vserver.h, look at the kernel headers now
1181951193 M * daniel_hozac vc_enter_space enters the spaces specified in flags, vc_set_space registers the current process's space(s) as those belonging to the context.
1181951322 M * coderanger I don't see where the flags are defined though
1181951362 M * daniel_hozac same as for clone/unshare, i.e. CLONE_NEWUTS etc.
1181951377 M * coderanger Ahh, okay
1181951538 M * coderanger It would be really nice if ubuntu actually include the man pages for things like clone
1181951919 J * stefani ~stefani@tsipoor.banerian.org
1181951958 P * stefani