1144800018 M * daniel_hozac looks sane.
1144800040 M * Bertl okay, great!
1144800055 M * Bertl any things we have still open/unfixed from kernel side?
1144800091 M * Bertl I just tried the sched_pause thingy Andreas mentioned
1144800118 M * Bertl but it works fine here .. i.e. hangs but can be easily unpaused
1144800176 M * daniel_hozac hmm, vc_set_ccaps?
1144800195 M * Bertl you mean the bcaps stuff?
1144800203 M * daniel_hozac yeah.
1144800226 M * Bertl we could try that ... but that requires that we move it into the capability check
1144800240 M * Bertl could you test such a modification with some guests?
1144800262 M * Bertl I currently have no real test hardware available ...
1144800281 M * daniel_hozac hmm, well, i just meant for new processes. i guess you're talking about the capable() mask thing?
1144800304 M * Bertl yeah, I'd move it there, so we would not need any change to the interface
1144800329 M * daniel_hozac how so? how would you raise the capabilities?
1144800354 M * Bertl bcaps would become a mask (well it basically already is)
1144800371 M * Bertl and the task caps would not be masked except for capable checks
1144800374 M * daniel_hozac ah, but apply it at runtime, ok.
1144800394 M * Bertl that should make _all_ processes happy
1144800404 M * daniel_hozac indeed.
1144800414 J * shedi ~siggi@inferno.lhi.is
1144800466 M * Bertl but it would require another userspace tool change
1144800491 M * daniel_hozac that would be rather racy.
1144800533 M * daniel_hozac i guess having a bmask and only modifying those capabilities in the mask would be the cleanest.
1144800569 M * Bertl okay, so another command/version with a slightly different structure
1144800623 M * Bertl shouldn't we move the stuff to a separate get/set bcaps command then?
1144800633 M * Bertl and leave the old stuff as is?
1144800643 M * daniel_hozac yeah, i always wondered what bcaps were doing in set_ccaps.
1144800678 M * daniel_hozac well, leaving the old stuff as is is probably not a good idea. having all the bcaps reset because you called vc_set_ccaps with a 0 bcaps is rather unexpected.
1144800692 M * daniel_hozac especially when you have a separate vc_set_bcaps.
1144800721 M * Bertl well, I'd suggest to do it this way:
1144800743 M * Bertl have the old ccaps, and handle it similar to now, .i.e &=
1144800756 M * Bertl have a new version which only has the ccaps and cmask
1144800771 M * Bertl and have a new bcaps command with bcaps and mask
1144800804 M * daniel_hozac right.
1144800994 M * brc bertl, sorry for taking too long. This machine is not as fast as i though
1144801004 M * brc I will leave the ssh opened and tomorrow i will have everything PERFECT to continue the tests
1144801010 M * brc i am still working on it but it seems to take sometime
1144801016 M * brc maybe i will remove unused stuff from the kernel now
1144801027 Q * ntrs Read error: Connection reset by peer
1144801033 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1144801035 M * Bertl don't worry, as you see we are working on something different anyway
1144801104 M * brc ok
1144802502 J * mkhl ~mkhl@200-148-40-145.dsl.telesp.net.br
1144802540 M * daniel_hozac i'll try to get my test machine up and running tomorrow, i'll need to swap some harddrives before that though.
1144802562 M * Bertl okay, should have something to test by then :)
1144802631 J * kingruedi ~king@p508B7DB2.dip.t-dialin.net
1144802640 M * Bertl welcome kingruedi!
1144803185 J * FireEgl Atlantica@Atlantica.Tcldrop.US
1144803611 Q * comfrey Ping timeout: 480 seconds
1144803961 M * Bertl daniel_hozac: what about this for a start: http://vserver.13thfloor.at/Experimental/delta-bcaps-feat01.diff
1144804001 M * Bertl getting something to eat now, brb
1144805271 M * Bertl back
1144805396 Q * id23 Ping timeout: 480 seconds
1144805483 M * daniel_hozac it looks fine.
1144805503 M * Bertl what do you think, shall we put that into rc16?
1144805555 M * daniel_hozac sure, it's not that much of a change.
1144805564 M * daniel_hozac when are you planning on releasing the finals?
1144805770 M * Bertl well, personally I would have already released a month ago or so
1144805778 M * daniel_hozac lol
1144805795 M * Bertl but I want some kind of PLM check on the code, as well as some testing which I didn't find the time yet
1144805815 M * daniel_hozac ah, is PLM still down?
1144805823 M * Bertl unfortunately PLM is still broken and it doesn't seem as if they are going to fix it soon
1144805840 M * daniel_hozac that's too bad.
1144805842 M * Bertl well, I have to try to submit the stuff via the web interface, maybe that works
1144806095 M * daniel_hozac what sort of testing are you looking for? on which branch?
1144806801 Q * kingruedi Quit: This computer has gone to sleep
1144808375 J * id23 ~id@p54A03F23.dip0.t-ipconnect.de
1144809556 Q * softi42 Ping timeout: 480 seconds
1144810096 J * softi42 ~softi@p549D4E5C.dip.t-dialin.net
1144811592 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at
1144811710 Q * f_ Quit: 
1144815645 Q * FireEgl Quit: Bye...
1144819697 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at
1144819761 Q * jkl helium.oftc.net oxygen.oftc.net
1144819792 J * jkl eric@c-67-172-156-116.hsd1.co.comcast.net
1144819838 Q * jkl Read error: Operation timed out
1144819842 J * jkl eric@c-67-172-156-116.hsd1.co.comcast.net
1144821099 J * Hmmmm ~Hmmmm@221.135.51.19
1144821733 Q * f_ Read error: Connection timed out
1144821959 J * Aiken_ ~james@tooax7-176.dialup.optusnet.com.au
1144822282 Q * Aiken Ping timeout: 480 seconds
1144823104 N * BobR_oO BobR
1144823398 J * f_ ~f_@83-215-237-2.seek.stat.salzburg-online.at
1144823607 Q * Hmmmm Quit: Ex-Chat
1144824702 J * |coocoon| ~coocoon@p54A0780C.dip.t-dialin.net
1144824707 M * |coocoon| morning
1144824776 M * BobR morning
1144826516 J * FireEgl Atlantica@Atlantica.IRCNut.Com
1144826811 Q * |coocoon| Read error: Connection reset by peer
1144827036 N * BobR BobR_afk
1144827719 N * Bertl Bertl_zZ
1144827795 J * |coocoon| ~coocoon@p54A0780C.dip.t-dialin.net
1144828005 N * morrigan morrigan_oO
1144828047 J * harti ~hw@83-215-237-5.seek.stat.salzburg-online.at
1144828477 J * domtec sp@83-215-237-4.seek.stat.salzburg-online.at
1144828505 N * domtec bertl_wecker
1144828939 N * bertl_wecker s_
1144829337 Q * mnemoc Ping timeout: 480 seconds
1144830042 J * mnemoc ~amery@200.73.54.5
1144830921 Q * |coocoon| Read error: Connection reset by peer
1144831562 Q * shedi Quit: Leaving
1144832338 N * morrigan_oO morrigan
1144832965 M * derjohn EHLO morrigan !
1144832977 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1144833055 Q * id23 Quit: Leaving
1144834018 J * redtux ~redtux@pc199.pub.univie.ac.at
1144835414 J * shedi ~siggi@tolvudeild-194.lhi.is
1144836162 Q * Aiken_ Ping timeout: 480 seconds
1144839071 J * kingruedi ~king@p508B5BFE.dip.t-dialin.net
1144840003 N * BobR_afk BobR
1144840450 T * * http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc15 | 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 ;)
1144840450 T * Bertl -
1144840483 Q * blizz helium.oftc.net xenon.oftc.net
1144840483 Q * trasher helium.oftc.net xenon.oftc.net
1144840483 Q * Medivh helium.oftc.net xenon.oftc.net
1144840483 Q * blackfire helium.oftc.net xenon.oftc.net
1144840483 Q * ray6 helium.oftc.net xenon.oftc.net
1144840483 Q * mugwump helium.oftc.net xenon.oftc.net
1144840483 Q * pusling helium.oftc.net xenon.oftc.net
1144840483 Q * sizo helium.oftc.net xenon.oftc.net
1144840483 Q * SNy helium.oftc.net xenon.oftc.net
1144840483 Q * lonewolff helium.oftc.net xenon.oftc.net
1144840483 Q * ag- helium.oftc.net xenon.oftc.net
1144840483 Q * h01ger helium.oftc.net xenon.oftc.net
1144840483 Q * morrigan helium.oftc.net xenon.oftc.net
1144840483 Q * Bertl_zZ helium.oftc.net xenon.oftc.net
1144840483 Q * tokkee helium.oftc.net xenon.oftc.net
1144840532 J * h01ger ~holger@bone.digitalis.org
1144840532 J * morrigan morrigan@212.16.62.52
1144840532 J * sizo ~janek@sushi.huggz.com
1144840532 J * SNy ef32795f88@bmx-chemnitz.de
1144840532 J * pusling pusling@195.215.29.124
1144840532 J * mugwump ~samv@watts.utsl.gen.nz
1144840532 J * lonewolff lonewolff@adleman.lonewolff.info
1144840532 J * ray6 ~ray@vh5.gcsc2.ray.net
1144840532 J * Bertl_zZ herbert@212.16.62.52
1144840532 J * tokkee tokkee@ssh.faui2k3.org
1144840532 J * blackfire blackfire@dp70.internetdsl.tpnet.pl
1144840532 J * Medivh ck@paradise.by.the.dashboardlight.de
1144840532 J * trasher daniel@derichs.info
1144840532 J * blizz ~blizz@evilhackerdu.de
1144840532 J * ag- ag@caladan.roxor.cx
1144842462 J * |coocoon| ~coocoon@p54A07D7A.dip.t-dialin.net
1144842465 M * |coocoon| hello
1144844506 Q * shedi Quit: Leaving
1144844518 N * BobR BobR_afk
1144844784 J * Dr4g ~Dr4g@80-195-133-218.cable.ubr06.uddi.blueyonder.co.uk
1144845156 J * fwl ~f_@83-215-237-1.seek.stat.salzburg-online.at
1144845253 Q * Dr4g Quit: Open Source Development :: http://dynamichell.org
1144845402 Q * f_ Ping timeout: 480 seconds
1144846071 Q * ntrs Ping timeout: 480 seconds
1144846725 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1144847297 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1144847305 N * BobR_afk BobR
1144847676 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1144848902 Q * mnemoc Ping timeout: 480 seconds
1144849815 Q * harti Quit: Leaving
1144850626 N * Bertl_zZ Bertl
1144850635 M * Bertl morning folks!
1144850655 M * Bertl daniel_hozac: sorry, lost connectivity yesterday
1144851219 M * daniel_hozac no problem, i fell asleep shortly after asking anyway :)
1144851279 J * mnemoc ~amery@200.73.54.5
1144851317 M * Bertl welcome mnemoc!
1144851568 M * ray6 morning Bertl!
1144851793 M * |coocoon| hello bertl
1144852375 J * Dr4g ~Dr4g@80-195-133-218.cable.ubr06.uddi.blueyonder.co.uk
1144852683 J * dearaujo ~dan@pixpat.austin.ibm.com
1144852812 M * dearaujo after reading the vsched+explained page, I get the feeling you can set/modify context scheduling (using the schedule file) as well as use vsched to apply limits to individual processes - that correct?
1144853099 M * Bertl yep, that's correct
1144853116 M * Bertl with latest devel you can do even more than the page shows
1144853129 M * dearaujo such as?
1144853132 M * Bertl well, you also need special tools to unleash newer functionality
1144853144 M * dearaujo heh - ok
1144853159 M * Bertl for example, you have not only rate/interval  but also rate2/interval2
1144853174 M * Bertl and you can assign values per cpu for example
1144853180 M * dearaujo ah - cool!
1144853190 M * dearaujo quite nifty 
1144853233 M * Bertl yeah, you can do a lot with that, but we put a limit to the options where they would add overhead
1144853249 M * Bertl (unnecessary overhead that is)
1144853273 M * dearaujo but with the current stable version I can create a context cpu limit of say 20% (of the system) and then limit - say my httpd server - to 10% of the 20%...
1144853277 M * derjohn hy, I bound 127.0.0.2 to eth0 within a guest. now postfix bind that adress too and - bad luck- it's the first address within the guest. so i have to switch it to the last posotion. I assume I have to restart the guest in order to do so?
1144853301 M * derjohn further: would there be a disadvantage to give loXY to the guest?
1144853367 M * derjohn funny: "  |Entire Message.eml |MAILHOST(RBL)= bl.spamcop.net|127.0.0.2"
1144853369 M * derjohn :)
1144853378 M * Bertl dearaujo: yes, you can
1144853401 M * Bertl derjohn: well, 127.0.0.2 _is_ already a disadvantage
1144853402 A * derjohn wonders why that worked at all ... several thousand mails went without hassle ...
1144853408 M * dearaujo Bertl: wow - excellent - when do the "new" tools plan to be stable?
1144853413 M * Bertl derjohn: but you can as well put it on lo
1144853455 M * derjohn simply add lo in the second line to /etc/vservers/gerstung3/interfaces/dev ?
1144853480 M * derjohn and /etc/vservers/gerstung3/interfaces/1/... < "dev" ?
1144853481 M * Bertl dearaujo: there is nothing which would make my 'hack' tools stable, but I guess I will upload a vsched shortly, and if I got that right, daniel_hozac is already working/thinking on extending util-vserver 0.30.210 for that
1144853513 M * Bertl derjohn: yep
1144853538 M * derjohn Bertl, can vsched _write_ the config to /etc/vservers/xxxx/... to make the settings persistent?
1144853540 M * dearaujo Bertl: great, thanks. 
1144853548 M * Bertl derjohn: nope
1144853724 M * derjohn Bertl, bad :( would it be difficult to read out the values and create the schedule file?
1144853753 M * Bertl no, but I'm not the one doing the userspace tools :)
1144853788 M * derjohn Bertl, read six lines above :) youre wrong. q.e.d. ;)
1144854048 M * derjohn Bertl, daniel_hozac or the "Host Optimalistic Zero additional Configuration" (HOZAC) Extension ?
1144854205 M * Bertl lol
1144854223 M * Bertl gdm: nice page .. you work on scheduling?
1144854580 M * derjohn Bertl, serious: cant you have the tool get the 4 values from the context and output it to stdout (so we can pipe it to a file) ?
1144854672 M * Bertl actually 8 values, but yes
1144854876 J * Viper0482 ~Viper0482@p54975C89.dip.t-dialin.net
1144855154 Q * s_ Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net ::
1144855167 Q * harry Ping timeout: 480 seconds
1144855406 Q * Dr4g Quit: Open Source Development :: http://dynamichell.org
1144855456 M * derjohn Bertl, 8? .... 14 32 500 200 1000 dummy ... thats 6-dummy=5 ... any new values appeared?
1144855497 M * Bertl actually we have now fillrate/interval fillrate2/interval2 tokens/min/max prio_base and flags
1144855512 M * Bertl and most of them are per cpu
1144855528 M * derjohn Bertl, what happens with a new kernel and an "old" schedule file?
1144855541 M * tokkee *argh* I should remove the hilight on tok+e+ ;-)
1144855545 M * Bertl works quite fine, that's why we have endless compatibility code
1144855549 M * derjohn are missing values filled by defaults?
1144855572 M * derjohn Bertl, well, that is part of the utils, so not you part :)
1144855590 M * Bertl the compatibility is _kernel_ code
1144855598 M * derjohn Bertl, why? can
1144855603 M * Bertl otherwise it would not work with older tools
1144855607 M * derjohn the utils not hand over defaults?
1144855618 M * derjohn older means? < 209 ?
1144855626 M * derjohn or non alpha?
1144855633 M * Bertl whatever you like, go back to 0.28
1144855648 M * derjohn (BTW: it gets time to rename 'alpha' to 'the only one worth trying')
1144855658 M * Bertl but even 0.30.210 will not handle the new interfaces (yet)
1144855706 M * derjohn Bertl, would it be possible to use defaults (i.e. 'max' or so?) or is the question wrong?
1144855720 M * Bertl derjohn: feel free to urge enrico to change it to devel or so
1144855766 M * Bertl derjohn: the ABI ensures that older syscall commands work and give reasonable results
1144855780 M * derjohn Bertl, i'll try .. but ... why 'endless compatibility code'? cant the kernel simply assign defaults to missign values? Doesnt sound to 'endless'?
1144855888 M * Bertl derjohn: well, it isn't that simple, but 'basically' that is what the compatibility code does
1144855902 M * derjohn ok. :) 
1144855919 M * derjohn Did you fix the algebra traps?
1144855939 M * Bertl algebra traps?
1144855950 M * derjohn I mean: The situatuionn where the scheduler can't get further ...
1144855962 M * derjohn (you told there might be some of them)
1144855967 M * Bertl ah, yes, that's fixed in 2.1.1-rc16
1144855981 T * Bertl http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc16 | 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 ;)
1144855985 M * derjohn in German "Algebrafalle" (Allg. Algebra)
1144855995 M * derjohn foo - rc16 !
1144855996 M * derjohn :)
1144856017 M * derjohn would you provide a minichangelog? I put it on the wiki
1144856121 M * Bertl no change to 2.0.2, two scheduler fixes and more scheduler improvements, a capability and history cleanup, the scheduling monitor and a new bcaps interface
1144856153 M * Bertl hmm, let me check that I actually added the schedmon :)
1144856172 M * Bertl yep, I did :)
1144856201 J * harry ~harry@d54C2508C.access.telenet.be
1144856221 M * Bertl welcome harry!
1144856227 M * sizo moin
1144856235 M * Bertl hey sizo!
1144856238 M * sizo ej
1144856241 M * derjohn no capability monitor yet?
1144856252 M * derjohn schedmon?
1144856268 M * Bertl everything there
1144856272 M * derjohn it removes those situations that should never oocur?
1144856288 M * Bertl who?
1144856305 M * derjohn can I see if a guest reqests a cap it isnt granted to use? 
1144856322 M * derjohn schedmon -  it removes those situations that should never oocur ?
1144856330 M * Bertl ad caps: ah, no, that's still a feature request
1144856335 M * derjohn "Scheduler monitor" ?
1144856344 M * Bertl ad schedmon: no, that is a debugging tool like the history tracing
1144856357 M * derjohn monitor like "looking into". k
1144856401 M * Bertl but the scheduler should be safe now except for special paus flag cases, which I didn't get to handle them yet
1144856403 M * derjohn well , lets hope those nasty situations never occur (I thought about s.th. like a "beast counter" for endless loops or so)
1144856492 M * derjohn no change to 2.0.2 rc16 == rc15 .... ?
1144856500 M * Bertl yep
1144856849 M * derjohn Bertl, what about new devel release?
1144856877 M * daniel_hozac "two scheduler fixes and more scheduler improvements, a capability and history cleanup, the scheduling monitor and a new bcaps interface"
1144856928 M * Bertl tx :)
1144857444 J * restill ~restill@c-24-11-171-10.hsd1.mi.comcast.net
1144857460 M * Bertl wb restill!
1144857467 M * restill hello
1144857512 M * restill whats new?
1144857579 M * Bertl some scheduler fixes, history and cap cleanup, a scheduling monitor :)
1144857698 M * derjohn Bertl, could you please have look to http://linux-vserver.org/ChangeLogExperimental ?
1144857784 M * matta Bertl: is that due to the swsoft tests?
1144857804 M * Bertl matta: they did some testing?
1144857840 M * matta yeah, they were communicating with sam on their list
1144857854 M * Bertl ah, the 1:2:3 setups
1144857911 M * Bertl no, not really related, but probably the reason why I started testing the new scheduler
1144857969 M * Bertl basically we do not do precise limits at all costs, as it seems the OVZ guys consider appropriate
1144857988 M * Bertl s/guys/folks/
1144858041 M * Bertl so, for example the following setup will always be somewhat inaccurate with linux-vserver:
1144858065 M * Bertl two cpus, 3 processes, 1:2:3 ratio
1144858120 M * Bertl well, that is even solveable by putting the first two on the first cpu
1144858139 M * Bertl but 1:2:4 will not
1144858252 J * Pazzo ~thomas@host130-250.pool8172.interbusiness.it
1144858271 M * Bertl welcome Pazzo!
1144858278 M * Pazzo hi Bertl!
1144858312 M * matta right.. if it's not a threaded app it will only ever use 1 CPU
1144858326 M * Pazzo Bertl: How are you? Everything fine?
1144858339 M * matta so the 1:2:3 does not work with a per-CPU aware scheduler
1144858348 M * matta it shouldn't work
1144858407 M * Bertl 1:2:3 can be done (with only 3 processes that is), by intentionally putting the 1 and 2 on one CPU and the 3 on the other
1144858797 Q * Pazzo Quit: Ex-Chat
1144858920 J * Pazzo ~thomas@host130-250.pool8172.interbusiness.it
1144858979 N * nokoya _nokoya
1144858985 N * _nokoya nokoya
1144859555 M * Bertl daniel_hozac: hmm, seems I should add some sanity checks to the scheduler interfaces:
1144859565 M * Bertl [   60.494092] divide error: 0000 [#1]
1144859570 M * Bertl [   60.497801] EIP is at vx_tokens_recalc+0x17d/0x300
1144859585 M * Bertl I accidentially set a rate of 0
1144859643 M * Hollow daniel_hozac: ping
1144859645 M * daniel_hozac hehe, yeah.
1144859648 M * daniel_hozac pong
1144859663 M * Hollow the clone-arch patch is the successor of the sparc64 patch right?
1144859675 M * daniel_hozac yeah.
1144859678 M * Hollow good :)
1144859819 M * daniel_hozac you know of the updated vcontext-uid patch, right?
1144860147 M * Hollow damn
1144860152 M * Hollow too late i read that
1144860153 M * Hollow :P
1144860224 Q * ||Cobra|| Remote host closed the connection
1144860299 M * Hollow hm, this is interesting..
1144860307 M * Hollow you uid patch never was in our svn
1144860353 M * daniel_hozac http://dev.croup.de/proj/gentoo-vps/browser/util-vserver/patches/0.30.210-r10/util-vserver-0.30.210-vcontext-uid.patch?rev=317 ?
1144860367 M * daniel_hozac or are you speaking of some other svn repo?
1144860381 M * Hollow hm, am i totally dumb?
1144860495 M * derjohn more, but maybe Hollow ;)
1144860498 M * Hollow i don't have this patch in my working copy
1144860509 P * dearaujo 
1144860541 M * Hollow i.e. it is also not in the patch tarballs
1144860547 M * daniel_hozac you removed it in rev 318 :)
1144860577 M * Hollow hm
1144860609 M * Hollow but the changeset doesn't show it
1144860648 M * Hollow strange things..
1144860729 M * Hollow hm, seems like i didn't svn up
1144860731 M * Hollow d'oh
1144860739 M * Hollow overwrote phreak`` s changes completely haha
1144860743 M * Hollow phreak``: owned!
1144860833 M * daniel_hozac haha.
1144860909 M * Hollow smells like r11 :o
1144861232 J * comfrey ~comfrey@h-64-105-215-74.sttnwaho.covad.net
1144861240 M * Bertl welcome comfrey!
1144861434 M * micah daniel_hozac: there is an update to the vcontext-uid patch?
1144862211 J * hw ~Hartmut@85-124-100-166.dynamic.xdsl-line.inode.at
1144862223 N * hw harti
1144862731 M * daniel_hozac micah: yeah.
1144862757 M * daniel_hozac https://savannah.nongnu.org/patch/?func=detailitem&item_id=4966
1144862881 J * shedi ~siggi@inferno.lhi.is
1144864059 M * Bertl wb shedi!
1144864115 M * Bertl off for now ... dinnertime!
1144864121 M * Bertl back later ...
1144864126 N * Bertl Bertl_oO
1144865261 N * BobR BobR_oO
1144865333 J * bonbons ~bonbons@83.222.37.206
1144865571 N * Bertl_oO Bertl
1144865582 M * Bertl back now ...
1144865766 M * ray6 ree bertl
1144865785 M * ray6 my vienna expedition is crashing :/ But I hope I can rescue parts of it...
1144866766 M * Bertl hmm?
1144866795 Q * phedny Ping timeout: 480 seconds
1144866887 M * ray6 Bertl: some of the guys I was planing to drive with canceled as the CNGW is too chaotic or something like that %->
1144866914 M * Bertl aha, hmm ... so no car?
1144866945 M * ray6 Bertl: looks like I'll be driving on Friday instead of Thursday now and only me and one or two others. I was thinking about flying again (was quite comfortable last weekend) but NOW the prices are up to 600 EUR %->
1144867203 M * ray6 But I meanwhile have unpacked my athlon. Could boot it up and veryfy that strange benchmark I got minutes before leaving at the hotel last week :) 
1144867299 J * phedny ~mark@volcano.p-bierman.nl
1144867422 M * Bertl welcome phedny!
1144867977 Q * Pazzo Quit: Ex-Chat
1144868195 N * BobR_oO BobR
1144868495 Q * comfrey Ping timeout: 480 seconds
1144868961 J * comfrey ~comfrey@h-64-105-215-74.sttnwaho.covad.net
1144869570 M * micah daniel_hozac: cool, is there a good way to get notifications of updates that are good to push out?
1144869682 M * daniel_hozac micah: other than adding yourself to CC on the patches/bugs?
1144869738 M * micah daniel_hozac: that would be a good way :)
1144869746 Q * harti Quit: Client exiting
1144869752 M * micah daniel_hozac: I'm just thinking of the other patches you've got also
1144869790 M * daniel_hozac micah: well, i could ping you :)
1144869820 M * micah daniel_hozac: that works if it doesn't bother you
1144869827 M * daniel_hozac not at all.
1144869904 M * micah is it possible to subscribe to the entire util-vserver page so you get emails of all bugs opened, etc?
1144869949 M * micah I'm not seeing it anywhere there, but it seems like it would be useful
1144869966 M * daniel_hozac i'm not aware of anything like that.
1144869969 M * daniel_hozac i'd like that too.
1144870139 M * micah you submitted an update to your bug on 4/11, the original on 3/11
1144870284 M * daniel_hozac hehe.
1144870375 M * micah I'll expect an update on 5/11
1144870444 M * micah daniel_hozac: there is another vlogin patch too it appears?
1144870687 M * daniel_hozac hmm, you should have that one already.
1144870713 M * daniel_hozac i lingered a bit with posting it to savannah, i wanted to make sure no other bugs popped up :)
1144870731 M * micah ah ok, yeah I was just comparing them and they seemed the same
1144870860 Q * kingruedi Ping timeout: 480 seconds
1144871599 Q * lilalinux_ Remote host closed the connection
1144871806 N * BobR BobR_oO
1144871820 J * lilalinux ~plasma@dslb-084-058-197-162.pools.arcor-ip.net
1144872832 Q * restill Quit: Leaving
1144873888 J * gianthome ~giant@213.146.112.138
1144873898 M * gianthome good evening
1144873928 M * daniel_hozac hello
1144874277 Q * bonbons Quit: Leaving
1144874352 M * micah does anyone know if there is a way to get nullmailer to eventually bounce messages, instead of requeuing them forever? 
1144875550 M * derjohn hm, i remember that access on a guest to 127.0.0.1 is remapped to it#s first ip. what, if the first ip _is_ 127.0.0.2? is 127.0.0.1 remapped to 127.0.0.2? what if the second ip the public one?
1144875687 M * derjohn hm, are the ip sorted in acending order within a guest? I just saw that ip '0' is my public one, and '1' is the 127.x .. but the 127.x is the first ones tat appears.
1144875850 M * daniel_hozac 127.0.0.1 would be remapped to 127.0.0.2.
1144875874 M * derjohn but only 127.x adresses are remapped?
1144875890 M * daniel_hozac only 127.0.0.1 is remapped.
1144875906 M * derjohn any why does te 127.0.0.2 appear as first one with 'ip addr' ?
1144875931 M * derjohn i assume te daemons be default take rhe first one as primary (sending) ip
1144875955 M * derjohn argh.  I assume the daemons by default .....
1144876041 Q * fwl Quit: This computer has gone to sleep
1144876081 M * daniel_hozac yes, they do.
1144876122 M * daniel_hozac the IP addresses should be ordered in the same way echo /etc/vservers/<name>/interfaces/* displays them.
1144876159 M * derjohn nope, that is not the case. as '0' i have 80.69..... as '1' is have 127.0.0.2
1144876161 M * daniel_hozac but the order in ip addr isn't important. look at /proc/virtnet/<nid>/info
1144876171 M * derjohn ip addr gives the 127... as first
1144876245 J * fwl ~f_@83-215-237-1.seek.stat.salzburg-online.at
1144876353 Q * Viper0482 Quit: bin raus, 
1144876412 M * Bertl okay, off for now .. back later
1144876495 N * Bertl Bertl_oO
1144876548 M * derjohn daniel_hozac, why do i see two <nid> with the same content?
1144876566 M * derjohn (differnet ones .. and yes: te 127.xx is on position two there)
1144876930 M * micah Bertl_oO: The blkmtd driver was removed - it's no longer present in Linus's tree, so that patch isn't needed anymore! hehe
1144877056 M * daniel_hozac lol
1144877058 Q * fwl Quit: This computer has gone to sleep
1144877065 M * daniel_hozac derjohn: hmm? i guess you're using old utils?
1144877078 M * daniel_hozac (and/or dynamic contexts)
1144877086 M * micah (thats what andrew morton told me)
1144877097 M * gdm Bertl_oO: thanks for the comment ;-)
1144877112 M * derjohn daniel_hozac, vserver-stats says ctx 200 
1144877113 M * gdm shame my wiki formatting knowledge is not that good, i'm afraid
1144877122 M * derjohn the nid is 49xxx
1144877130 M * daniel_hozac derjohn: so, old utils.
1144877147 M * derjohn # vserver --version
1144877147 M * derjohn vserver 0.30.209 -- manages the state of vservers
1144877159 M * derjohn *ouch* Debian ones ...
1144877190 M * daniel_hozac Debian's utils should be in pretty good shape now.
1144877198 M * derjohn micah, why 0.30.209-2 instead of 0.30.210-xx ?
1144877210 M * daniel_hozac 0.30.210-6 is the latest one, AFAIK...
1144877235 M * derjohn micah, in sid ... (didn't check experimental)
1144877246 M * micah derjohn: apt-get update?
1144877248 M * derjohn micah, sry
1144877254 M * derjohn micah, 0.30.210-6 !!! thx !
1144877255 M * micah derjohn: sid doesn't have 0.30.209
1144877285 M * derjohn micah, Preparing to replace util-vserver 0.30.209-2 /me eyes where blind ;)
1144877317 M * micah gdm: did you make a wiki page out of that email?
1144877343 M * gdm micah: no, not the mail, but the background to the mail
1144877358 M * gdm micah: i made the http://linux-vserver.org/vsched+explained 
1144877377 M * gdm micah: was considering doing the email, but need to anonymise ips etc first
1144877384 M * gdm micah: it made sense?
1144877426 M * micah gdm: I'm still trying to digest it :)
1144877459 M * gdm micah: me too :)
1144877570 M * micah gdm: I wanted to mention that you should move to static context ids
1144877600 M * gdm micah: i was wondering about that as i was writing that email. is it easy to do?
1144877608 M * gdm i mean, for already in use vservers?
1144877610 M * micah gdm: very easy
1144877627 M * gdm micah: and if i respond here?
1144877637 M * gdm micah: it's just like a restart, i presume
1144877644 J * dearaujo ~dan@pixpat.austin.ibm.com
1144877647 M * micah all you have to do is cat > /etc/vserver/<vsname>/context
1144877656 M * micah and put in the context number you want for that vserver
1144877666 M * micah and yeah, it requires a restart, but you can do that whenever is convenient
1144877694 M * dearaujo it is true a guest can be bound to only 16 ip addressed from the host?
1144877696 M * micah you can pick whatever number scheme is useful for you, I just started at 10 and started counting up
1144877706 M * dearaujo *addresses
1144877714 M * gdm micah: great, thanks
1144877729 M * daniel_hozac dearaujo: well, you can patch it to support more, but by default, yes.
1144877734 M * micah derjohn: yes, Bertl is looking for 6 legitimate complaints to change it :)
1144877743 M * gdm now, i am a little confused about vsched again, but i think that is a question for Bertl_oO when he returns
1144877764 M * gdm as he said that it doesn't write teh config file - but just makes changes
1144877769 M * dearaujo daniel_hozac: I assume there's some time latency after the increase?
1144877777 M * micah I actually was hoping it might be possible to look at vserver-stat to try and make an educated guess about vsched settings and memory limitations
1144877780 M * daniel_hozac dearaujo: see the recent mailing list posts.
1144877790 M * dearaujo ok
1144877794 M * dearaujo thanks
1144877796 M * gdm micah: vsched is just for the cpu from what i could tell
1144877808 M * micah gdm: yeah, hence the and :)
1144877818 M * gdm ahh, right!
1144877822 P * dearaujo 
1144877835 Q * comfrey Ping timeout: 480 seconds
1144877842 M * micah oh yeah, I see how you could interpret that another way, my fault for being ambigous
1144877887 M * daniel_hozac micah: /proc/virtual/.../limit would be better for that.
1144877897 M * daniel_hozac micah: as it also contains the maximum seen.
1144877913 M * daniel_hozac (for memory limits)
1144877916 M * micah vserver-stat shows you how many processes and the VSZ/RSS of each vserver plus the userTIME and sysTIME, it seems like it could be easily graphed with munin and you could (if you understood it) extrapolate some per-guest limits
1144877934 M * daniel_hozac vserver-stat uses ugly hack for its values.
1144877944 M * micah ah interesting
1144877946 M * daniel_hozac graphing should use /proc/virtual/.../limit.
1144877967 M * micah daniel_hozac: do you know what the different columns are in /proc/virtual/.../limit?
1144877982 M * micah current, max, -1, 0?
1144877984 M * daniel_hozac current, maximum, limit, hits.
1144877995 M * micah hits meaning how many times it has hit the limit
1144877999 M * daniel_hozac right.
1144878023 M * micah interesting
1144878525 M * gdm yeah, that's described somewhere on the wiki or in a paper, i think
1144878534 M * gdm i've been trying to read and understand it this week
1144878549 M * gdm my host was brought down last week but an application in a vserver consuming all the memory
1144878557 M * gdm making the machine unreachable
1144879093 M * derjohn micah, could you please elaborate on extrapolating? does munin support regression?
1144879116 M * micah derjohn: regression?
1144879132 M * micah derjohn: munin just polls values and then graphs them
1144879141 M * derjohn well, your msg reads like prediction/forecasting ...
1144879152 M * derjohn .... could (if you understood it) extrapolate some per-guest limits ....
1144879180 A * derjohn wrote a trafficforecaster in Java, thats why I am interestef in all forecast stuff
1144879208 M * micah derjohn: oh well, I was meaning that if you were graphing this data over time, you might be able to make adjustments based on trends if you understand them
1144879354 M * derjohn ah, ok, so only 'passive graphing'  ...
1144879368 M * micah yes, munin polls every five minutes
1144879376 M * micah its rrdtool/mrtg basically
1144879427 M * derjohn micah, btw: rrdtool knows to forecast one step with a holt/winters method for 'anomaly detetcion' (which is pretty cool )
1144879430 M * derjohn :)
1144880012 J * andrew_ ~andrew@host-87-74-59-116.bulldogdsl.com
1144880501 J * Aiken ~james@tooax6-119.dialup.optusnet.com.au
1144881465 Q * FireEgl Ping timeout: 480 seconds
1144882000 Q * matta Ping timeout: 480 seconds
1144882158 J * FireEgl Atlantica@Atlantica.CJB.Net
1144882478 M * Wonka will there be vserver patches for 2.6.17-rc, or only for 2.6.17?
1144882523 M * daniel_hozac IIRC Bertl_oO already has patches for 2.6.17-rc1.
1144882528 M * daniel_hozac just not public yet.
1144882554 M * ray6 daniel: you mean those he made for the Niagara last Weekend? :)
1144882562 M * daniel_hozac probably :)
1144883214 M * Wonka *want*
1144883241 J * lilalinux_ ~plasma@dslb-084-058-192-058.pools.arcor-ip.net
1144883322 Q * |coocoon| Ping timeout: 480 seconds
1144883426 Q * FireEgl Quit: Bye...
1144883672 Q * lilalinux Ping timeout: 480 seconds
1144883880 Q * andrew_ Quit: andrew_
1144883944 M * ray6 I think he wanted to fix a few things which currently are "oh conflict in XFS... disable XFS..." :)
1144884014 A * Wonka just needs ext2/ext3
1144884045 J * matta ~matta@71.224.125.126
1144885713 J * Dr4g ~Dr4g@80-195-133-218.cable.ubr06.uddi.blueyonder.co.uk
1144885804 J * FireEgl Atlantica@Atlantica.Tcldrop.Org