1145836812 J * mkhl ~mkhl@200-148-41-120.dsl.telesp.net.br
1145836848 Q * ntrs Ping timeout: 480 seconds
1145837027 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1145837098 M * h01ger micah, gdm: i've now commited what i had last week already :-/ didnt had time to properly work on it today :(
1145837106 M * h01ger but at least we have a base now :)
1145837269 Q * jkl Remote host closed the connection
1145837542 Q * ntrs Ping timeout: 480 seconds
1145837771 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1145837847 Q * doener Ping timeout: 480 seconds
1145838165 J * doener ~doener@i5387C752.versanet.de
1145839522 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net
1145840624 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145842200 Q * mkhl Ping timeout: 480 seconds
1145842823 J * matt1 ~matta@c-68-32-202-140.hsd1.pa.comcast.net
1145843250 Q * matta Ping timeout: 480 seconds
1145846751 J * kilgur user@p50811A04.dip0.t-ipconnect.de
1145846802 J * Aiken_ ~james@tooax6-073.dialup.optusnet.com.au
1145847118 Q * Aiken Ping timeout: 480 seconds
1145847167 Q * kilgur_ Ping timeout: 480 seconds
1145849663 J * ZLinux[] ~ZLinux@212.118.97.63
1145849753 Q * ZLinux_ Ping timeout: 480 seconds
1145852075 Q * mountie Read error: Connection reset by peer
1145852155 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145859891 Q * kilgur Quit: Trillian (http://www.ceruleanstudios.com
1145861270 J * _coocoon_ ~coocoon@p54A070A7.dip.t-dialin.net
1145861291 M * _coocoon_ morning
1145862072 J * click click@ti511110a080-3573.bb.online.no
1145862243 J * Dr4g ~Dr4g@82-40-203-15.stb.ubr06.uddi.blueyonder.co.uk
1145863262 Q * cehteh Remote host closed the connection
1145863438 Q * _coocoon_ Ping timeout: 480 seconds
1145864100 J * _coocoon_ ~coocoon@p54A05D04.dip.t-dialin.net
1145864360 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1145865120 J * pagano ~pagano@lappagano.cnaf.infn.it
1145865307 J * Aiken__ ~james@tooax7-103.dialup.optusnet.com.au
1145865653 Q * Aiken_ Ping timeout: 480 seconds
1145865856 M * harry wiiiiiiiiiii
1145866016 M * pagano hi ;)
1145869063 M * gdm h01ger: ahh, kool. where is that committed?
1145869413 N * otaku42_away otaku42
1145869515 Q * ||Cobra|| Remote host closed the connection
1145870131 M * h01ger gdm, in pkg-vserver/people/holger
1145870255 M * h01ger gdm, micah: i just realized the plugins need a proper name-scheme: vserver_* is used for plugins which monitor all vservers at once, while your vsmem_ is supposed to run per vserver. there will be more like this. maybe vservers_foo and vguest_foo ? (cause i think vservers_ and vserver_ is too indifferent)
1145871329 J * lilalinux ~plasma@h1-gw.of.net-lab.net
1145872491 Q * FireEgl Ping timeout: 480 seconds
1145872918 J * ksf ~krazy_sys@202.80.169.52
1145873056 N * meebey_ meebey
1145874798 Q * ksf Quit: 
1145875005 M * Hollow olla
1145875088 Q * _coocoon_ Ping timeout: 480 seconds
1145875736 J * ksf ~krazy_sys@202.80.169.52
1145875745 J * _coocoon_ ~coocoon@p54A063AE.dip.t-dialin.net
1145875849 Q * _coocoon_ Quit: 
1145876211 M * gdm h01ger: thanks. i'll try to get a look at it properly later
1145876263 Q * softi42 Ping timeout: 480 seconds
1145876466 J * Milf ~Miranda@ipsio25.ipsi.fraunhofer.de
1145876664 Q * lilalinux Remote host closed the connection
1145876881 J * softi42 ~softi@p549D5E93.dip.t-dialin.net
1145877390 J * lilalinux ~plasma@h1-gw.of.net-lab.net
1145877436 J * _coocoon_ ~coocoon@p54A077BF.dip.t-dialin.net
1145877933 Q * ksf Ping timeout: 480 seconds
1145877983 Q * _coocoon_ Ping timeout: 480 seconds
1145878653 J * _coocoon_ ~coocoon@p54A0759D.dip.t-dialin.net
1145878775 Q * Aiken__ Quit: Leaving
1145879433 Q * kir Ping timeout: 480 seconds
1145879551 J * hallyn ~xa@c-24-11-243-196.hsd1.in.comcast.net
1145880073 J * kir ~kir@swsoft-mipt-nat.sw.ru
1145880695 J * ||Cobra|| ~cob@146.50.22.204
1145881478 J * ksf ~krazy_sys@202.80.169.52
1145882677 Q * mountie Remote host closed the connection
1145882976 Q * ksf Quit: 
1145883586 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145883807 Q * lilalinux Remote host closed the connection
1145884565 Q * matt1 Ping timeout: 480 seconds
1145884789 J * lilalinux ~plasma@dslb-084-058-211-196.pools.arcor-ip.net
1145885884 N * Bertl_zZ Bertl
1145885888 M * Bertl morning folks!
1145885928 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1145885944 M * Bertl wb matta!
1145886199 M * _coocoon_ hello bertl
1145886216 M * Bertl hey _coocoon_! any progress?
1145887172 Q * _coocoon_ Quit: KVIrc 3.2.0 'Realia'
1145887514 M * matta hey!
1145888402 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1145888434 M * Bertl afk, getting something to eat, brb ...
1145888997 M * eyck   sooo, 
1145889010 M * eyck this is how you treat your users, 
1145889014 M * eyck tell them to 'stop whining'
1145889026 M * eyck  and 'this is not a bug, this is a feature'
1145889027 M * eyck goood.
1145889077 M * Bertl eyck: well, if you prefer to whine over 'improving' things, then go ahead ... no problem with that
1145889078 M * eyck I feel like my providing feedback is valued very high, you know
1145889134 M * eyck  well, if you admit then there are some problems, then we can start working on improving this, 
1145889138 M * Bertl you keep repeating arguments which are none, complaining about things folks explained to you even twice
1145889154 M * eyck  when project leaders lives in denial, how can you improve anything
1145889159 M * Bertl e.g. what about the 'add ips' stuff?
1145889173 M * Bertl that was not even possible before 2.1.x
1145889192 M * eyck you think? how have I been doing this then?
1145889211 M * eyck and no, noone explained anything about adding ips.
1145889212 M * Bertl no idea, but it just doesn't work with kernel earlier than that :)
1145889222 M * eyck interesting...
1145889261 M * Bertl you can easily try it with any older kernel or old tools, just start a server (e.g. apache) which binds to 0.0.0.0 inside a guest, then add an new ip to that guest ...
1145889266 Q * hallyn Quit: leaving
1145889271 M * eyck  it doesn't work with new-style, with that, I agree.
1145889295 M * Bertl no, it does not work with non devel (2.1.x) kernels and super-new tools
1145889327 M * eyck so, you admin that you removed that feature ;) 
1145889331 M * eyck admit ;) 
1145889431 M * dev :/
1145889485 M * Bertl eyck: it _is_ a _new_ feature, added around 2.1.0.4 or so
1145889500 M * Bertl eyck: it _was_ _not_ possible to add/remove ips before (on the fly)
1145889515 M * Bertl eyck: and older tools do _not_ support it anyway
1145889535 M * Bertl eyck: I have not remove a feature which did not exist :)
1145889633 M * Bertl eyck: same goes for entering a stopped context, which was never possible (naturally)
1145889762 M * Bertl but, to get back to the old-vs-new style discussion, do you really think that the IPV4ROOT= method was soo cool to specify a dozen IP addresses, some of them on vlans maybe or reused from the host?
1145889825 M * Bertl I even had to write a longer walkthrough, so that folks would get it right, because setting the IPROOTDEV would mess it all up (usually)
1145889839 M * dev Bertl: have you implemented virtualized IPv4 or you use something else?
1145889865 M * Bertl dev: what do you consider 'virtualized IPv4'?
1145889984 Q * matta Read error: Connection reset by peer
1145889997 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1145890005 M * dev Bertl: private routes, hashes, netfilters etc.
1145890022 M * dev Bertl: net devices...
1145890052 M * dev Bertl: I'm just wondering what other approaches can be used here and what are pros and cons.
1145890056 M * Bertl no, as you know, we avoid that overhead on purpose
1145890099 M * dev Bertl: recently we measured overhead of venet on 1Gbit ethernet. it is less than 2-4%.
1145890115 M * Bertl there is no _real_ need for that IMHO, except maybe for honeypots or so
1145890125 M * Bertl dev: and on 10GB?
1145890126 M * matta Bertl: hosting needs it...
1145890135 M * dev didn'
1145890147 M * Bertl matta: not really (need != want)
1145890152 M * matta Bertl: ie. certain control panels, firewall apps (shorewall, apf, etc)
1145890161 M * dev didn't measured on 10GB, but since even on 1Gb CPU is almost saturated, I don't forsee much difference.
1145890179 M * matta Bertl: some major control panels (ie. cPanel - the largest) use iptables / bandmin for tracking bandwidth usage
1145890210 M * Bertl well, tracking bandwidth _inside_ the guest is not the best idea, isn't it?
1145890227 M * matta it is when the guest is resold as an alternative to a dedicated :)
1145890255 M * matta dev: see, I'm not exclusive to who I'll argue with :)
1145890267 M * Bertl but I already suggested that somebody should do a wrapper, which relays the iptable commands to the host, applies some business logic/sanity checks and acts on behalf ...
1145890296 M * Bertl IMHO that would make _much_ more sense than virtualizing this and putting the overhead into the kernel
1145890297 M * dev matta: I believed it :)
1145890380 M * dev Bertl: maybe you are right. Though I don't see much overhead. But ok. But can you shed some light on how it is implemented in vserver then? I'm asking due to pure curiosity.
1145890422 M * matta as a user here I see highlights of each system
1145890424 M * Bertl it is pure isolation, similar to the processes
1145890427 M * dev is VPS owner able to snif traffic from others?
1145890433 M * Bertl nope
1145890455 M * dev isolation? i.e. each VPS gets it's own interface? or what?
1145890464 M * dev matta: can you bring some details?
1145890471 M * Bertl no, networking, as you know, is ip related on linux
1145890488 M * Bertl so no need to 'operate' on devices, IPs are sufficient
1145890490 M * matta vserver might benefit from taking the network (venet) code and state save/restore and openvz could benefit from taking vservers per-VPS CFQ scheduler
1145890496 M * dev is it possible to run tcpdump inside VPS or it simply doesn't work?
1145890517 M * Bertl dev: it is not permitted to put any dev into promisc
1145890533 M * Bertl (dev being device :)
1145890536 M * dev matta: it's not vserver code (CFQ). and it will be included very soon in OVZ/VZ as well.
1145890557 M * matta oh, and openvz could definitely benefit from the VServer CoW links which IMHO are much better than even Virtuozzo VZFS
1145890567 M * matta dev: to be per-context aware it is.
1145890582 M * dev matta: also, please, note, that CFQ doesn't help very much as it doesn't schedule AIO and journal requests :/ it works only for sync i/o
1145890606 M * Bertl matta: had a look at the snapshotting stuff, it's quite simple and what I expected, so once there is a demand for that, we can add that quite easily
1145890627 M * matta dev: ok, I see a contradiction here....
1145890647 M * matta dev: if it doesn't help then why is the OVZ team _wasting_ their time?
1145890648 M * Bertl dev: you reported some ptrace issues for x86 on x86_64 (mainline issue) did you send that upstream yet?
1145890651 M * dev matta: which contradiction? in that we will include CFQ, but it doesn't help very much? it is true :/
1145890685 M * Bertl dev: if not, could you point me to the relevant patches?
1145890689 J * FireEgl Atlantica@Atlantica.Tcldrop.Org
1145890700 M * matta Bertl: are they just using a glorified swsup?  I didn't look at the code at all.
1145890734 M * dev matta: the problem with vserver COW links are that IMHO it doesn't provide you normal unix filesystem symantics. Also AFAIK, you believe that VZFS is worse due to vzcache util. But vzcache doesn't correspond to vunify. there is another tool in VZ which is fast :p
1145890757 M * matta dev: which tool is that?
1145890760 M * Bertl matta: it's the straight forward approach, archiver/unarchiver for each structure/element
1145890767 M * dev Bertl: which issues?
1145890770 M * matta and yeah, vzcache is one of the worst programs I have ever used.
1145890789 M * matta dev: you say there is another tool in VZ 
1145890795 M * matta that is fast
1145890802 M * dev 1 sec
1145890808 M * Bertl dev: kernel ptrace x86 on x86_64 (mainline 2.6.16)
1145890823 M * matta and not a horribly written piece of shit like vzcache (no personal offense, it's how I really feel about that program)
1145890835 M * dev there are too many issues with ptrace. which one?
1145890839 M * dev matta: vzpkglink
1145890845 M * matta oh, so it's template based
1145890854 M * Bertl dev: well, all which cause kernel panics actually :)
1145890856 M * matta even worse.
1145890893 M * dev matta: I can't stand for the details (as not an author), but why do you believe it is worse?
1145890905 M * dev matta: I would appreciate any details from you
1145890934 M * Bertl Hollow: ping!
1145890938 M * Hollow pong
1145890949 M * matta here is the thing, the virtuozzo templating idea is good for some uses, ie. the enterprise where an admin needs to manage 100 VPS's and wants to keep them all up to date
1145890961 M * dev Bertl: I'm not quite getting what you are talking about. Maybe I have told you something before what you are referncing, but  my memory lost it already :/
1145890964 M * matta thing is, I have customers who were setup with redhat 9 years ago
1145890972 J * mkhl ~mkhl@200-153-153-183.dsl.telesp.net.br
1145890977 M * matta so in the ve.conf the templates all still point to redhat-9
1145890998 M * matta but they've since upgraded to FC or to RHEL themselves using yum/apt/etc
1145891028 M * dev matta: oh, I see... and vzpkglink doesn't try to link them to new templates?
1145891057 M * Bertl dev: okay, simple question: are there any patches in OVZ which fix ptrace issues on ia32 emulation (for x86_64)?
1145891058 M * matta the vserver CoW links w/ vunify just scans for identical files and creates hard links so there is no one 'true' file
1145891082 M * matta dev: but then... the admin is expected to keep every template ever released on all their host servers (migrations, etc)
1145891085 M * matta it doesn't scale.
1145891124 M * dev Bertl: in 2.6.16 AFAIK - no. but there are issues, which are being discussed in security@. they will be commited as sorted out to 2.6.16.y
1145891176 M * dev matta: so you don't like vserver approach as well?
1145891181 M * Bertl dev: didn't see any of them on LKML though ...
1145891199 M * dev Bertl: sure. it is private talks until made public.
1145891203 M * matta dev: the 'it doesn't scale' approach was meant to be in regards to VZ
1145891211 M * dev :)
1145891255 M * dev matta: I got it. thanks for details. I will provide this info to template team tomorrow. Really many thanks.
1145891290 M * matta dev: since vserver uses the link approach so suppose you want to do a migrate, the rsync just breaks the link and copies the file
1145891312 M * matta then the weekly cron on the other host will re-link the files it has in common with others on the new host
1145891332 M * dev Bertl: it all doesn't matter... Linux kernel nowdays is too bugfull and any novice can dig into git and find the ways to kill Linux... :/
1145891355 M * matta the VZ approach is completely broken since it requires every template, everywhere.
1145891368 M * dev matta: I suppose it doesn't break anything, but just copies the file and deletes the one which is hard link. correct?
1145891369 Q * mountie Quit: LUNCK!
1145891387 M * matta dev: yes, hence CoW 'links'
1145891393 M * dev matta: ok :)
1145891444 M * Hollow Bertl: second pong
1145891459 M * Bertl Hollow: ah, tx!
1145891481 M * dev matta, Bertl: I know that there was a problem with COW hardlinks: before changing template file, it should have been deleted. was it fixed already?
1145891505 M * Bertl Hollow: had some network issues yesterday. wanted to ask when/if you want to test the lo stuff?
1145891520 M * Hollow sure
1145891542 M * Bertl Hollow: btw, you should be able to enable klog inside a guest too with recent kernels and the syslog flag (regarding guest layout)
1145891560 M * Bertl Hollow: not that I consider it useful, but the virtualized klog is there
1145891563 M * Hollow klog i.e. dmesg?
1145891567 M * Bertl yep
1145891569 M * Hollow ok
1145891573 M * Hollow good to know
1145891582 M * Bertl but it's disabled by default
1145891586 M * Hollow yup
1145891587 M * matta dev: this vzpkglink is an upgrade to vzcache for sure, but VZ templates are not updated anywhere near as quickly as users upgrade their VPS's
1145891628 M * Hollow Bertl: btw, you'll get a new ppc tester..
1145891813 M * dev Hollow: can we get him also? :)
1145891860 M * Hollow yes, i can test ovz as well :P
1145891878 M * Bertl Hollow: sounds good! power5? :)
1145891887 M * Hollow no, unfortunately not
1145891894 M * Bertl just kidding :)
1145891904 M * Hollow i just freed 10G on my ibook for the linuxtag and i think i'll keep it ;)
1145891920 M * Hollow even if the hdd is just 40G :/
1145891931 Q * pagano Quit: Leaving
1145891973 M * Hollow well, i'd have access to a G5 machine.. but i doubt i'm allowed to install linux on it *hehe*
1145891990 J * Viper0482 ~Viper0482@p54977BFA.dip.t-dialin.net
1145892000 M * dev Hollow: would be really nice!
1145892033 M * Hollow i even just noticed 2.6.17-rc2 supports my airport extreme card, yay!
1145892035 M * Bertl dev: would also be really nice if you could open up one or the other tests you run on OpenVZ (or Linux-VServer :)
1145892129 M * dev Bertl: I'd like to, but this 99% depends not on me :/
1145892153 M * Bertl well, on whom does OpenVZ stuff depend nowadays?
1145892215 M * matta yeah
1145892220 M * matta i thought it was a community project
1145892233 M * dev there is management team in the company as anywhere. since most tests are written inhouse, there should be a good argument for these guys to open tests.
1145892252 M * dev matta: OpenVZ is community and open source. the talk is about tests suites.
1145892256 M * matta how about so no one is just taking your word on it?
1145892258 M * Bertl damn, we have no management team :)
1145892274 M * dev Bertl: can't say whether it is good or bad :))))))))))))
1145892281 M * matta ok, well, I tested vserver with my own internal tests and the network overhead was 0.06% on 10Gbit/s
1145892310 M * matta seems waaaay better than OVZ's venet
1145892324 M * Bertl that sounds good, could you do a test on the upcoming lo stuff too (sooner or later)?
1145892326 M * matta 2-4% on 1Gbit/s? hate to see it on 10Gbit/s
1145892335 M * dev matta: have you measured it?
1145892345 M * matta dev: yeah, I just said I did.
1145892354 M * matta you don't take my word for it? :)
1145892389 M * dev matta: it is written with "?". So I don't get it :) did you mean that OVZ gives you 2-4% on 1Gbit?
1145892401 M * matta that's what you stated.
1145892411 M * dev so you didn't test it.
1145892422 M * dev ok.
1145892427 M * matta (10:48:07) dev: Bertl: recently we measured overhead of venet on 1Gbit ethernet. it is less than 2-4%.
1145892432 M * Bertl matta: it might not be true, just marketing figures :)
1145892434 M * matta I'm saying I tested VServers
1145892442 M * dev ha-ha
1145892456 M * dev Bertl: lets be less offendive to each other, ok?
1145892464 M * Bertl just kidding :)
1145892465 M * matta and the network overhead on 10Gbit/s network was 0.06% 
1145892481 M * dev matta: what test did you use?
1145892486 M * matta my own internal testr.
1145892489 M * Bertl dev: don't take everything personal ... try to get a distance from swsoft if you want to handle an OS project
1145892502 M * Bertl dev: well, you and kir :)
1145892529 M * dev Bertl: I don't get anything personal, it is just hard to communicate with people which do talk about things which has nothing to do with technology itself.
1145892561 M * matta I cannot release the script for you, but it was written by one of our developers and management won't let me release it :(
1145892563 M * dev matta: at least what kind is it? web? file download?
1145892571 M * dev :)))))
1145892571 M * dev ok
1145892572 M * matta ipperf
1145892574 M * dev keep it private 
1145892630 M * matta yeah, i have to. Sorry.
1145892665 M * matta i'll test OVZ and let you know how it goes.
1145892897 J * harti ~Hartmut@85-124-100-166.dynamic.xdsl-line.inode.at
1145893011 M * matta dev...
1145893021 M * matta i have to say, I ran into a major problem with vzpkglink
1145893033 M * matta it was my initial observation confirmed.
1145893053 M * matta since it is template based the savings are marginal at best
1145893069 M * dev ?
1145893074 M * matta say a provider (such as one I am at) offers CentOS + cPanel
1145893079 M * dev if you do updates, it should catch them
1145893086 M * matta well, you know cPanel installs it's own Apache, MySQL, etc
1145893123 M * matta and templating cpanel is not a good idea since the install changes frequently and you would need to run upcp right after an install anyway
1145893130 M * dev got it
1145893136 M * matta so no memory/disk sharing for any of those apps
1145893166 M * dev matta: AFAIK, vserver doesn't help here very much until VPS is restarted. correct?
1145893197 M * matta dev: I do not believe so, but neither does virtuozzo
1145893222 M * dev matta: yep. I'm just asking.
1145893236 M * matta the data is in memory already
1145893264 M * matta the disk savings are immediate, but with disk sizes what they are today the memory sharing is the important part.
1145893295 M * matta a huge feature VServer has over OVZ
1145893300 M * dev matta: there are 2 instances in memory for 2 VPSs. if you hardlink the same files, you need restart apps to force them to reopen the same file for memory savings. yep, disk savings are immediate.
1145893643 Q * michal` Read error: Connection reset by peer
1145893643 Q * lilalinux Read error: Connection reset by peer
1145893649 J * michal` ~michal@www.rsbac.org
1145893892 Q * michal` Quit: 
1145893974 M * doener Bertl: hm, about the on-the-fly ip address stuff, that was possible (except for single-ip-address vservers) since we had the vc_net_add (or whatever functions), IIRC that was 2.0, tested that when the early libvserver stuff was created... I kinda doubt that eyck refers to them though, linux 2.6 after all ;) *scnr*
1145894098 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145894162 M * Bertl doener: yes, the interfaces are there for some time now, but nobody used them until recently
1145894182 M * Bertl doener: and thus they didn't work as expected :)
1145894209 M * doener hm, in what way did they fail? has always worked for me back then, but maybe my use-cases were to specific
1145894259 M * Bertl I could be wrong, but I remember fixing something when Hollow/bonbons first started using them
1145894312 M * doener anyway, just wanted to point out that that feature (started to) appear(ed) in 2.0 ;)
1145894365 M * Hollow Bertl: lo patch?
1145894372 M * Bertl doener: right, sorry that I got that wrong :)
1145894409 M * doener heh, I got so much stuff wrong lately...
1145894417 M * Hollow Bertl: i'm sorry, it's up already :)
1145894422 M * Bertl Hollow: uploaded, delta-lo0.04.diff it's basically a rewrite, so a lot of things might be different
1145894451 J * lilalinux ~plasma@dslb-084-058-252-040.pools.arcor-ip.net
1145894452 M * Bertl Hollow: let me know about compile warnings and such stuff too (if you encoutner them)
1145894478 M * Hollow ok
1145894561 M * Hollow Bertl: hm, fails to apply... do i need anything else then 2.6.16.8 and vs2.1.1-rc17?
1145894588 M * Bertl hmm, you should not, but maybe the diff is off ... sec
1145894617 M * Hollow in top-level Makefile and devinet.c and udp.c
1145894629 M * Bertl sec, checking
1145894727 M * Bertl I have a cleanup script running for 4 days now, it somewhat blocks my way to work, hope it will finish today though ...
1145894767 M * Hollow lol
1145894772 M * Hollow what does it clean?
1145894805 M * Bertl basically data mining through my kernel trees,r emoving unused stuff and reunifying the trees
1145894823 M * Hollow ah.. your 1000+ kernel dir.. ic ;)
1145894832 M * Bertl yup, those trees :)
1145895177 J * bonbons ~bonbons@83.222.37.206
1145895217 Q * ||Cobra|| Remote host closed the connection
1145895239 M * Bertl wb bonbons!
1145895255 M * bonbons hi Bertl!
1145895405 J * cehteh foobar@cehteh.homeunix.org
1145895412 M * Bertl welcome cehteh!
1145895422 M * cehteh moin
1145895451 N * cehteh GNUcifer
1145895456 M * GNUcifer mal bouncer configgn
1145895521 M * GNUcifer Last login: Fri Oct 28 01:22:45 2005 from 10.20.200.100
1145895521 M * GNUcifer  ... mhm my irc bouncer runs on a vserver where i really rarely log in :)
1145895558 N * otaku42 otak42_away
1145895778 Q * mountie arion.oftc.net neutron.oftc.net
1145895778 Q * GNUcifer arion.oftc.net neutron.oftc.net
1145895778 Q * harry arion.oftc.net neutron.oftc.net
1145895778 Q * teukka arion.oftc.net neutron.oftc.net
1145895778 Q * k3mper arion.oftc.net neutron.oftc.net
1145895778 Q * yang arion.oftc.net neutron.oftc.net
1145895778 Q * bubulak arion.oftc.net neutron.oftc.net
1145895778 Q * ddlp arion.oftc.net neutron.oftc.net
1145895778 Q * otak42_away arion.oftc.net neutron.oftc.net
1145895778 Q * thorkild arion.oftc.net neutron.oftc.net
1145895778 Q * dhansen arion.oftc.net neutron.oftc.net
1145895778 Q * redtux arion.oftc.net neutron.oftc.net
1145895781 J * bubulak ~bubulak@cicka.wnet.sk
1145895787 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145895788 J * otak42_away ~otaku42@legolas.otaku42.de
1145895791 J * thorkild ~Hiya@stray.priv.no
1145895792 J * k3mper ~maio@mail.progamers.cz
1145895793 J * harry ~harry@d54C2508C.access.telenet.be
1145895797 J * teukka ~tmatilai@backport.ri.fi
1145895798 J * ddlp ~ddlp@sarayi.kariva.org
1145895801 J * dhansen ~dave@sprucegoose.sr71.net
1145895829 M * Bertl Hollow: ah, just the patches we added to rc17, I'll update the diff
1145895852 M * Hollow ok
1145895861 J * cehteh foobar@cehteh.homeunix.org
1145895868 M * cehteh tadaa
1145895902 M * Hollow ah, this is so awesome.. my airport extreme card work with native drivers
1145895903 Q * weasel Quit: reboot
1145895985 Q * lilalinux Ping timeout: 480 seconds
1145896096 J * redtux ~redtux@pc199.pub.univie.ac.at
1145896197 M * Bertl Hollow: okay, lo0.04.1 uploaded
1145896232 M * Hollow thx
1145896483 J * _coocoon_ ~coocoon@p54A062A8.dip.t-dialin.net
1145896758 J * lilalinux ~plasma@dslb-084-058-249-074.pools.arcor-ip.net
1145897362 J * coocoon3 ~coocoon@p54A07D4F.dip.t-dialin.net
1145897370 Q * _coocoon_ Ping timeout: 480 seconds
1145897377 Q * coocoon3 Quit: 
1145897394 J * _coocoon_ ~coocoon@p54A07D4F.dip.t-dialin.net
1145897399 M * Hollow Bertl: for any build errors you might be interested in, the complete build log: http://phpfi.com/114575
1145897464 M * Hollow rebooting..
1145897467 Q * Hollow Quit: Konversation terminated!
1145897647 J * Hollow ~hollow@home.xnull.de
1145897656 M * Hollow boots at least ;)
1145897665 M * Hollow so, any specific tests? or the whole bunch again?
1145897713 M * Bertl we have some test scripts (nc based) you can run those
1145897718 M * Hollow ah, nice
1145897738 M * Bertl http://vserver.13thfloor.at/Stuff/TEST/
1145897749 M * Bertl basically you grab the nc_client and server scripts
1145897760 M * Bertl then start the server, wait for SERVERS READY
1145897765 M * Bertl and after that start the client
1145897782 M * Bertl it is supposed to output something similar to the *24.out
1145897802 M * Bertl but you might need to recompile the nc, daniel_hozac had issues with a different netcat branch
1145897911 M * Hollow hrm
1145897917 M * Hollow no gawk in my busybox image
1145897999 M * Bertl Hollow: you can run it on the host
1145898018 M * Bertl Hollow: and use vnet/chbind 
1145898032 M * Hollow mhm
1145898041 M * Hollow ah, awk did it.. ;)
1145898236 M * Hollow Bertl: http://phpfi.com/114578
1145898246 J * lilalinux_ ~plasma@h1-gw.of.net-lab.net
1145898284 M * Bertl Hollow: ah, I always forget to explain the test scripts :)
1145898302 M * Bertl Hollow: they assume that you have certain IPs available/not avail
1145898326 M * Bertl Hollow: first, they assume that 192.168.0.1 is there and local, usually you configure it on dummy0
1145898344 M * Bertl then they assume that 192.168.0.9 is lan, but not used
1145898370 M * Bertl of course, both are assigned to the network context too
1145898386 Q * lilalinux Ping timeout: 480 seconds
1145898391 M * Hollow hm..
1145898406 M * Bertl you can give those as arguments though
1145898414 M * Hollow since i don't want to reconfigure my network.. can i use just other ips? :)
1145898417 M * Hollow ok
1145898422 Q * _coocoon_ Ping timeout: 480 seconds
1145898438 M * Bertl just use 10.10.0.x or so
1145898442 M * Hollow i have 192.168.0.{50,123,124} on eth0
1145898454 M * Hollow and they are configured in lotest
1145898462 M * Hollow so can i use them?
1145898471 M * Hollow 50 is not in lotest
1145898488 M * Bertl you should first start on the host (no guest at all)
1145898506 M * Bertl and see that you basically get the same output as I do
1145898515 J * _coocoon_ ~coocoon@p54A05D7A.dip.t-dialin.net
1145898529 M * Bertl (it's not a test which is that easy to get right as testme.sh :)
1145898544 M * Hollow yeah, it's more like testfs.sh ;)
1145898570 M * Bertl worse :)
1145898583 M * Hollow so, the 0.9 is any box in my lan but not myself?
1145898599 M * Hollow but i still chbind to it?
1145898611 Q * mountie Remote host closed the connection
1145898631 M * Bertl Hollow: no, it's an 'unused' ip, which is assigned
1145898644 M * Bertl but it should be an IP which _could_ be assigned on the host
1145898648 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145898676 M * Hollow ok.. dinner time, brb
1145898789 M * Bertl enjoy your meal!
1145899613 Q * mkhl Ping timeout: 480 seconds
1145900207 M * Hollow arr, i'm too dumb...
1145900223 M * Hollow chbind --nid 123 --ip 192.168.0.123/24 --ip 192.168.0.124/24 -- /bin/bash
1145900227 M * Hollow Bertl: is this ok?
1145900411 M * Bertl should work
1145900437 M * Bertl (note: you need that for server and client)
1145900446 M * Hollow i started screen inside bash
1145900846 M * Hollow Bertl: doesn't look much better imo http://phpfi.com/114587
1145901000 M * Hollow note: i changed 123 and 124 to 3 and 4 because the ip was too long and uglified the output
1145901023 M * Bertl that's why I suggested 10.10.x.x
1145901191 M * Hollow so.. is it ok, or am i still doing it wrong?
1145901211 M * Hollow there are many strange chars ni the output..
1145901216 M * Hollow phpfi made the htmlentities
1145901221 M * Hollow s/the/them/
1145901311 M * Bertl do you use gawk now?
1145901323 M * Bertl if not, please try again with gawk instead of awk
1145901338 M * Hollow yes, it is gawk now, as i ran it on the host
1145901471 M * Bertl okay, then the nc is too different
1145901497 M * Bertl I don't know why nc seems to exist in at least three branches, with completely different output and options
1145901534 M * Bertl I tend to add it as simple ncb.c with a makefile :)
1145901572 M * Bertl so for now the only 'option' seems to compile the one I provided and use that for testing ... sorry ...
1145901585 M * Hollow maybe i need GAPPINGSECURITYHOLE enabled? :)
1145901593 M * Hollow ok
1145901594 M * Bertl no, doesn't even use that feature
1145901602 M * Bertl but you can do your normal tests as well
1145901614 M * Hollow well, no problem with installing another nc..
1145901620 M * Hollow just point me to the source
1145901623 M * Viper0482 hi
1145901649 M * Bertl Hollow: I'm more interested in guest behaviour than in those test actually, as they ran here too .. (the source rpm is in the dir, just use alien to convert it)
1145901672 M * Hollow well, aren't the tests more extensive then my manually testing?
1145901684 M * Bertl I'd hope so ...
1145901698 M * Hollow so, what's the point in testing manually then? :)
1145901711 M * Hollow or what did you mean by guest behaviour?
1145901725 M * Viper0482 i have a small problem creating a new vserver with the new kernel (2.6.16-1.2096_FC4.vs2.0.2.0.rc17.1) and vserver-utils (0.30.210-14.fc4)
1145901734 M * Bertl Hollow: well, the test cannot verify services, just protocols with nc
1145901762 M * Hollow ah, so you basically want RL examples like httpd ftpd and such?
1145901786 M * Bertl for example, and how it feels and looks for the admin (host and guest)
1145901875 M * Bertl and what is somewhat untested is the crosstalk between guests
1145902021 Q * mountie Remote host closed the connection
1145902156 M * Hollow Bertl: hm, i cannot bind on the host to a specific addr
1145902185 M * Hollow zeus ~ # nc -l -p 80 -s 192.168.0.50
1145902186 M * Hollow Can't grab 192.168.0.50:80 with bind : Cannot assign requested address
1145902186 M * Hollow zeus ~ # nc -l -p 80 -s 192.168.0.123
1145902186 M * Hollow Can't grab 192.168.0.123:80 with bind : Cannot assign requested address
1145902186 M * Hollow zeus ~ # nc -l -p 80 -s 192.168.0.124
1145902187 M * Hollow Can't grab 192.168.0.124:80 with bind : Cannot assign requested address
1145902198 M * Hollow EADDRNOTAVAIL
1145902215 M * Hollow but they are configured (ip addr)
1145902281 M * Bertl nothing is bound to that IP already?
1145902286 M * Hollow right
1145902286 M * Bertl IP/port?
1145902291 M * Hollow i can bind to 0.0.0.0
1145902328 M * Hollow 192.168.0.{50,123,124}/80
1145902334 M * Hollow does not work
1145902336 M * Bertl what does lsof -ni show at port 80?
1145902338 M * Hollow 0.0.0.0/80 works
1145902351 M * Hollow nothing
1145902368 M * Bertl okay, but a guest is using port 80, right?
1145902382 M * Hollow no guest is running at all
1145902400 M * Bertl uhum ... checking 
1145902411 M * Hollow ar
1145902413 M * Hollow wtf
1145902415 M * Hollow i'm sorry
1145902420 M * Hollow i still was inside that screen
1145902428 M * Hollow only with chbind
1145902469 M * Bertl hmm, but that should work too then, at least 123/124 was assigned, no?
1145902483 M * Hollow yes
1145902496 M * Bertl except when you have something bound on the host 
1145902500 M * Hollow yup, and it still doesn't work
1145902518 M * Bertl okay, let me check that
1145902618 M * Bertl works perfectly fine here ...
1145902625 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145902808 M * Hollow my networking is completely b0rked
1145902840 M * Hollow zeus ~ # ping 13thfloor.at
1145902840 M * Hollow ping: unknown host 13thfloor.at
1145902840 M * Hollow zeus ~ # ping 213.133.98.19
1145902840 M * Hollow connect: Invalid argument
1145902894 M * Hollow interesting enough that i can still talk in irc ;)
1145903135 M * Bertl really interesting, looks and works perfectly normal here
1145903234 M * Bertl Hollow: do you ahve debugging enabled?
1145903254 M * Bertl if so, we could check a few things, but won't work with much traffic
1145903258 M * Hollow vserver: yes, kernel: no
1145903291 M * Hollow but i'll give it a reboot in a minute and we'll try again..
1145903321 M * Hollow *sigh*
1145903325 M * Hollow not my day
1145903332 M * Hollow i grepped the ibook kernel config
1145903342 M * Hollow yes, i have kernel debug enabled
1145903570 M * Bertl okay, let's try the following:
1145903594 M * Bertl set the 'net' debug to 0x20 and retry the bind
1145903621 M * Bertl (sysctl, vserver.debug_net = 32 )
1145903627 M * Hollow ok.. first shutting down some network activity
1145903742 M * Hollow ok, done
1145903780 M * Hollow there is already some output in dmesg now
1145903807 M * Hollow but i can't post it haha
1145903813 M * Hollow i can't access any website
1145903845 M * Hollow well, about 20 lines (getting more and more) like this:
1145903847 M * Hollow [ 6202.981305] vxD: sk_match_addr(a47a5940[#0],192.168.0.50) 192.168.0.50 [#0]
1145903870 M * Hollow each 2-5 secs such a message appears
1145903883 J * mkhl ~mkhl@200-153-153-224.dsl.telesp.net.br
1145903919 M * Bertl that's okay, it's a check for that ip, i.e. it tries to bind to it
1145903945 M * Hollow mhm.. but noone tries to bind :)
1145903970 M * Bertl obviously something _is_ conencting to .50 :)
1145903996 M * Hollow well, connecting maybe.. but not binding
1145904011 M * Hollow i have a dnsmasq running for the local network usually
1145904036 M * Bertl okay, let's give strace -fF a chance on the nc
1145904037 M * Hollow and well, i'm talking in irc over .50 :)
1145904048 M * Hollow ok
1145904088 M * Hollow sorry for all the others, but i have to paste it here..
1145904090 M * Hollow bind(3, {sa_family=AF_INET, sin_port=htons(80), sin_addr=inet_addr("192.168.0.50")}, 16) = -1 EADDRNOTAVAIL (Cannot assign requested address)
1145904173 M * Bertl okay, do you have .51 in use?
1145904179 M * Hollow yes
1145904182 M * Bertl .52?
1145904184 M * Hollow no
1145904188 M * Bertl good :)
1145904192 M * Hollow (at least i hope so
1145904196 M * Bertl let's try the following:
1145904203 M * Hollow maybe dhcp gave it.. but i'll use a non dhcp ip
1145904213 M * Bertl okay, which one?
1145904223 M * Hollow nothing between .50 and .100
1145904234 M * Bertl so .42 is fine?
1145904236 M * Hollow yep
1145904258 M * Bertl okay, do ip addr add 192.168.0.42/24 dev dummy0
1145904270 M * Bertl (make sure to load the dummy module)
1145904280 M * Hollow done
1145904290 M * Bertl then let's try the following nc command
1145904303 M * Bertl (with and without strace)
1145904315 M * Bertl nc -l -s 192.168.0.42 -p 42
1145904332 M * Bertl does that give you the same error?
1145904338 M * Hollow yup
1145904372 M * Bertl okay, I'm out of ideas here, because the very same line works perfectly fine here 
1145904389 M * Hollow *shrug* strange things are happening
1145904403 M * Hollow ok.. brb then
1145904409 M * Bertl k
1145904423 Q * Hollow Quit: Konversation terminated!
1145904554 J * Hollow ~hollow@home.xnull.de
1145904576 M * Hollow ok, let's start again ;)
1145904617 M * Bertl I'd suggest I uplaod a complete patch against 2.6.16.8 
1145904672 M * Bertl (will revert some things, as my basetree is not 2.6.16.8 yet) but that should not hurt
1145904709 M * Wonka Bertl: why not against 2.6.16.10? :)
1145904731 M * Bertl well, that was the version hollow had, but 2.6.16.10 is fine for me too
1145904765 Q * _coocoon_ Ping timeout: 480 seconds
1145904792 M * Hollow Bertl: ok, i tried binding httpd to 0.0.0.0, 127.0.0.1 and 192.168.0.123 in a guest, and connect to it, getting index.html, looks fine
1145904801 M * Hollow now testing cross guest 
1145904949 M * Hollow ok, cross test, and connection from the host look fine too
1145904967 M * Bertl okay, what did you change?
1145905008 M * Hollow i did not run nc_{server,client}
1145905010 M * Hollow :)
1145905028 M * Hollow and i did not use chbind
1145905033 M * Hollow i just started up two guests
1145905056 M * Bertl okay, most likely you got some bound services somewhere .. somehow ...
1145905085 M * Hollow well, i had a lots of temrinals open.. so who knows
1145905116 M * Hollow anyway, looks great all in all :D
1145905130 M * Hollow going to try udp now
1145905261 M * Hollow vx1 binds to 127.0.0.1:1234 UDP, vx2 connects to 127.0.0.1:1234 UDP, then nc on vx1 fals with: no connection: operation not permitted
1145905297 M * Bertl did you use udp for the client too?
1145905304 M * Bertl (just checking)
1145905315 M * Hollow yes.. 0.0.0.0 and 192.168.0.123 work fine with udp
1145905326 M * Hollow (and cross test)
1145905334 J * _coocoon_ ~coocoon@p54A05213.dip.t-dialin.net
1145905347 M * Bertl okay, could you provide the two nc commands you tried/used?
1145905379 M * Hollow root@lotest ~ # nc -l -u -p 1234 -s 127.0.0.1
1145905388 M * Hollow root@lotest2 ~ # nc -u 127.0.0.1 1234
1145905389 M * Hollow foo
1145905389 Q * mountie Read error: Connection reset by peer
1145905401 M * Hollow then, lotest fails with the above message
1145905426 M * Bertl hmm hmm ...
1145905438 M * Bertl could you retry the former and 'just' wait a little?
1145905467 M * Hollow waiting...
1145905470 M * Bertl what I mean is, it takes roughly 10 seconds for nc to 'decide' whether it can bind to a port or not
1145905487 M * Bertl so, it could as well be that this was 'coincidence'
1145905526 M * Hollow no, it's reproducable without exception..
1145905543 M * Hollow no matter how fast or slow i am :)
1145905553 M * Bertl okay, what IPs are assigned to the guests?
1145905571 M * Hollow lotest: .123, lotest2: .124
1145905727 M * ray6 morning Bertl :)
1145905732 M * Bertl morning ray6!
1145905735 M * ray6 still debugging this nc issue?
1145905759 M * Bertl nah, nc is resolved, we are testing the lo patches
1145905883 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1145905910 M * Bertl welcome Blissex!
1145905926 A * Blissex lurks :-)
1145905958 Q * lilalinux_ Remote host closed the connection
1145905962 M * Bertl Blissex: lurk on ...
1145905986 A * Blissex hides behind the curtain
1145906124 M * Bertl Hollow: are you sure the nc in vx1 gives you the EPERM?
1145906147 M * Bertl (i.e. the server not the client)
1145906178 M * Hollow yes
1145906219 M * Bertl at least interesting :)
1145906311 M * Bertl yes, indeed, I can reproduce that
1145906403 M * Bertl okay, will investigate this, any other anomalies yet?
1145906541 M * Hollow no
1145906688 M * Bertl ah, seems we still have the checks in ip_route_connect
1145906797 J * ben_ ~B.Lukas@88.134.54.56
1145906807 M * ben_ hello
1145906816 M * Bertl welcome ben_!
1145906833 M * ben_ can i whisper to u?
1145906843 M * Bertl whisper, hmm ...
1145906917 M * ben_ its privat, the vserver php front end :D
1145906926 M * Bertl k, whisper on ...
1145906930 Q * nebuchadnezzar Read error: Connection reset by peer
1145906998 J * nebuchadnezzar ~nebu@zion.asgardr.info
1145907016 M * ray6 a private php vserver frontend? :)
1145907035 M * Bertl of course, and top secret, so *shh*
1145907099 M * ray6 oops. sorry :)
1145907116 A * ray6 also wouldn't talk about using php in the public :)
1145907177 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1145907185 M * Hollow ray6: agreed. :D
1145907270 M * Hollow Bertl: do you want to integrate the lo stuff in 2.0.2?
1145907418 M * Bertl definitely not, but it might make it into devel, or at least in the first exp after a devel release
1145907437 J * mef ~mef@targe.CS.Princeton.EDU
1145907443 M * Bertl welcome mef!
1145907456 M * mef it's that time again...
1145907460 M * mef hi bertl
1145907470 M * mef 8 hours to paper deadline.
1145907661 M * Bertl everything ready, I presume?
1145907663 M * Hollow Bertl: ok, if you need some more testing (x86, amd64, ppc) before the releases let me know..
1145907781 M * Bertl Hollow: okay, great, tia!
1145907793 M * Bertl btw, the issue should be fixed shortly ... trivial patch
1145907799 M * Hollow great!
1145907809 M * Bertl just testing if it works as expected in this regard
1145907874 M * Bertl nah, missed the isolation part, so will work later ...
1145908004 M * derjohn rc18 in sight?
1145908224 M * Bertl yes
1145908419 M * micah derjohn: it appears that Ola doesn't follow the svn commits or tags I add to bugs
1145908431 M * micah derjohn: but he has a point about the bash completion script, did you write it?
1145908657 M * derjohn micah, it not my work, but i modified it several times ....
1145908753 M * derjohn micah, I sent you the e-mail address of the upsteam author, but maybe I can help, too?
1145908754 M * micah derjohn: do you know who wrote it?
1145908764 M * derjohn see above :) chk you mail
1145908766 M * micah derjohn: basically its a license problem, its licensed under the GFDL
1145908778 M * derjohn *ouch*
1145908813 M * derjohn Do you like to contact the autor or should I do it ? as you like ...
1145908989 M * ben_ hello, derjohn :D
1145908999 M * derjohn bello hen__ ;)
1145909162 Q * Viper0482 Quit: bin raus, 
1145909480 Q * _coocoon_ Ping timeout: 480 seconds
1145909617 M * Bertl okay, off for dinner now ...
1145909624 N * Bertl Bertl_oO
1145909822 M * micah derjohn: have you had contact with them in the past?
1145909833 M * micah derjohn: I will CC you and the bug report
1145909892 M * derjohn micah, I had a little contact as being a 1st time user of the script ... but nothing more the two or three e-mails
1145910055 J * _coocoon_ ~coocoon@p54A06368.dip.t-dialin.net
1145910602 M * _coocoon_ ok bertl tell me which router can all do what we need to use telnet
1145910696 M * derjohn _coocoon_,  <Bertl> okay, off for dinner now ..
1145910708 M * _coocoon_ oh ok
1145910710 M * derjohn _coocoon_, telnet? router?
1145910713 M * _coocoon_ thanx
1145910724 M * micah anyone know if there is a way to increasse the amount of ram the disk cache uses? it seems to top out at 2gig
1145910758 M * micah i've got a system that has a lot of free ram available, and is a fileserver that could use to have more in ram
1145910776 M * derjohn micah, not really but you may set the readahead buffer of CFQ 
1145910812 M * micah derjohn: that is if you are using CFQ, right?
1145910870 M * ben_ gn8 all 
1145910878 Q * ben_ Quit: 
1145910909 M * derjohn micah, obvisously yes ;)
1145910982 M * micah derjohn: is a CFQ scheduler a good choice for a fileserver?
1145911091 M * Blissex micah: thats a very interesting question.
1145911100 M * Blissex micah: depends what the fileserver does.
1145911119 M * derjohn micah, dont know. isnt there a nfs process fir each user?
1145911133 M * derjohn CFQ schedules IO by process groups ...
1145911142 M * derjohn i.e. shares among them
1145911152 M * derjohn or are you talking about samba ?
1145911209 M * Blissex micah: check out http://WWW.sabi.co.UK/Notes/anno05-3rd.html#050907 and surroundings for discussion and links
1145911275 M * micah its defiantely NFS and rsync only, no samba
1145911639 Q * harti Quit: Client exiting
1145911663 M * micah derjohn: if there was a different nfs process per user and I was using CFQ, would that make it more efficient?
1145911742 M * micah i've got 16 NFS processes with different IDs
1145911756 M * micah different PIDs that is 
1145912062 M * derjohn does pstree or ps faxu show it as one process group
1145912163 M * derjohn or as one?
1145912539 M * Blissex micah: lots of NFS processor is not necessarily a good idea.
1145913166 J * weasel weasel@asteria.debian.or.at
1145913736 Q * doener Quit: leaving
1145914117 M * Skram anyone arround?
1145914126 M * Skram will "iftop" work in a vserver environment?
1145914140 M * Skram  # iftop -i eth0:0010
1145914140 M * Skram interface: eth0:0010
1145914140 M * Skram IP address is: 70.86.176.11
1145914140 M * Skram MAC address is: 00:14:22:72:a1:4d
1145914142 M * Skram pcap_open_live(eth0:0010): socket: Operation not permitted
1145914145 M * Skram woops
1145914156 M * Skram well, that is because of the kernel access/etc?
1145914164 M * Skram so I need to give him Xcappibilities?
1145914357 Q * bonbons Quit: Leaving
1145914580 J * shedi ~siggi@cpe-24-165-167-167.midsouth.res.rr.com
1145915130 M * daniel_hozac Skram: it uses raw sockets, so yes. granting that will allow sniffing (of all traffic).
1145915243 M * Skram ok
1145915248 M * Skram anyway to limit it to just one uo?
1145915440 M * daniel_hozac a raw socket? no.
1145915492 M * Skram *ip
1145915494 M * Skram okay :
1145915495 M * Skram (
1145915507 M * Skram so if I want the user to be able to do that, they will have full access to all traffic
1145915536 M * daniel_hozac yep.
1145916068 M * micah h01ger: maybe vservers_memory for all vservers, and vsmem_ for individual plugins?
1145916141 M * h01ger micah, vsmem_cpu ? 
1145917212 Q * matta Ping timeout: 480 seconds
1145917612 M * micah h01ger: no, the munin standard for wild-card plugins is to just have "plugin_" and then you symlink "plugin_whatever" in /etc/munin/plugins... this is how the if_ plugin works, so you can create /etc/munin/plugins/if_eth0, /etc/munin/plugins/if_eth1 and the plugin reads what is after the _ to determine what to operate on
1145917689 Q * daniel_hozac Quit: brb
1145917707 M * micah h01ger: so vsmem_ would just be called vsmem_ and then you would symlink the vserver name in /etc/munin/plugins... there would be a different one for cpu, vscpu_ 
1145917728 M * thorkild micah: isn't that mostly for those which are shell script based?
1145917731 J * daniel_hozac ~daniel@c-2d1472d5.010-230-73746f22.cust.bredbandsbolaget.se
1145917766 M * thorkild oh, you're maybe using that. :) My bad. I am mostly used to the perl ones
1145917780 M * micah thorkild: some of the perl ones also do that
1145917815 M * micah thorkild: for example cps_ is a perl script
1145917885 M * micah thorkild: mysql_isam_space_, ntp_ and ping_ etc. all are
1145917888 M * thorkild hmm, yes, and the ntp ones. 
1145917913 M * micah thats the point of munin, is that you can write the plugins in any language you prefer
1145917926 M * thorkild micah: sorry, I must have misunderstood. I thought they were not doing that any longer.
1145917944 M * thorkild micah: yeah, I know. (I work at the place where it originated from)
1145917957 M * thorkild micah: my bad.
1145917958 M * micah thorkild: that could be possible, but I have not seen anything on the mailing list about them not doing taht any more
1145917967 M * micah thorkild: but i would be interested to know if that is true!
1145917974 M * thorkild micah: it most probably isn't.
1145917976 M * micah thorkild: I guess you can ask :)
1145918007 M * thorkild micah: I must have gotten the wrong discussion. When I think of it, it doesn't make sense to drop it. 
1145918064 M * micah h01ger: actually, I'm wondering if we could put the wild-card and single plugins together, so there is only one
1145919345 N * Bertl_oO Bertl
1145919349 M * Bertl back now ...
1145920080 M * brc BERTL
1145920090 M * brc i am erally sorry for the delay on the scripts, have been REALLY busy
1145920096 M * brc are there any patches to try out ?
1145920114 M * Bertl not much yet, so do not worry too much
1145920216 M * brc ok
1145920229 M * brc after we finish with the quota stuff, i had a greate idea for iptables and want to discuss with you
1145920273 M * Bertl okay ...
1145920282 M * h01ger micah, good idea. i can implement this - if you're not in a hurry, i doubt i'll have much time til debconf. then, it should be an hour or so and i should be able to find that. now i'm about to go to bed.. :)
1145920309 M * brc I though about coding a daemon that will run on the host and accept connections from the guests. The guests will run another tool i will code as "iptables" and will be able to do much stuff. everything will be translated on the daemon to the guests' specific chain. what do you think ?
1145920355 M * Bertl well, we could even do that with unmodified iptables inside, I guess
1145920387 M * brc Cause i dont know much about translating stuff inside a guest to the host, so i would use TCP connections for that
1145920398 M * brc let's finish quota before. i really want to do the iptables thing.
1145920471 M * Bertl yes, one thing after the other
1145920628 M * micah h01ger: i might be in a hurry because I need to set some limits on some over-eager vservers
1145920661 M * micah h01ger: before debconf, so they dont crush boxes while I am having tequila with you
1145920720 M * h01ger :)
1145920797 M * h01ger micah, let's just keep communicating over this channel :)
1145920808 M * h01ger and svn.. :)
1145920881 M * micah h01ger: sounds good, I keep trying to yell to you out my window, but i get no response, this is more reliable
1145921514 M * mugwump h01ger: hey howsit
1145921560 M * mugwump today I'm playing with vserver+fai (finally) :)
1145921625 M * mugwump utsl.gen.nz is getting an upgrade
1145921677 M * mugwump PIII-500/192MB RAM -> 3xOpt.275/4GB RAM
1145921699 M * mugwump er, 4x, not 3x (2 dual core chips)
1145921783 Q * Dr4g Quit: Open Source Development :: http://dynamichell.org
1145922314 M * micah mugwump: i think he went to sleep
1145922328 M * mugwump ah, right, not just snobbing me :)