1210380107 Q * bronson Quit: Ex-Chat
1210380249 J * mire_ ~mire@28-173-222-85.adsl.verat.net
1210380644 Q * mire Ping timeout: 480 seconds
1210381509 J * mire__ ~mire@185-172-222-85.adsl.verat.net
1210381872 Q * mire_ Ping timeout: 480 seconds
1210382893 Q * mire__ Ping timeout: 480 seconds
1210383323 J * mire__ ~mire@242-175-222-85.adsl.verat.net
1210384183 J * bronson ~bronson@adsl-68-122-117-135.dsl.pltn13.pacbell.net
1210384413 M * snooze # /usr/sbin/setattr --barrier /vservers
1210384413 M * snooze /vservers: Function not implemented
1210384417 M * snooze whats wrong with that?
1210384433 M * daniel_hozac you're not running a Linux-VServer kernel, or your filesystem is not supported.
1210384443 M * snooze lol of course
1210384455 M * snooze thats on a test box without the vs kernel
1210384461 M * snooze stupid me .. :p
1210384492 J * mire_ ~mire@109-173-222-85.adsl.verat.net
1210384878 Q * mire__ Ping timeout: 480 seconds
1210385116 Q * nkukard Ping timeout: 480 seconds
1210385687 J * mire__ ~mire@249-172-222-85.adsl.verat.net
1210385748 Q * mire_ Ping timeout: 480 seconds
1210386520 J * mire_ ~mire@232-175-222-85.adsl.verat.net
1210386889 Q * mire__ Ping timeout: 480 seconds
1210387873 J * mire__ ~mire@78-173-222-85.adsl.verat.net
1210387974 Q * mire_ Ping timeout: 480 seconds
1210388859 J * hparker ~hparker@linux.homershut.net
1210389113 Q * mire__ Ping timeout: 480 seconds
1210389398 J * mire__ ~mire@85-172-222-85.adsl.verat.net
1210392474 Q * mire__ Ping timeout: 480 seconds
1210395816 J * Aiken ~james@ppp118-208-54-233.lns4.bne1.internode.on.net
1210398768 J * ktwilight_ ~ktwilight@87.66.195.80
1210399087 J * kriebel ~kriebel@216-164-160-36.c3-0.eas-ubr10.atw-eas.pa.static.cable.rcn.com
1210399147 Q * ktwilight Ping timeout: 480 seconds
1210403008 J * nkukard ~nkukard@196.212.73.74
1210403012 Q * nkukard Read error: Connection reset by peer
1210403599 J * nkukard ~nkukard@196.212.73.74
1210403709 J * yarihm ~yarihm@164-42-239-77-pool.cable.fcom.ch
1210403803 Q * pmenier Quit: Konversation terminated!
1210404092 J * larsivi ~larsivi@205.80-202-29.nextgentel.com
1210404169 Q * FireEgl Ping timeout: 480 seconds
1210404789 J * FireEgl FireEgl@adsl-226-58-240.bhm.bellsouth.net
1210405028 J * larsivi_ ~larsivi@144.84-48-50.nextgentel.com
1210405237 Q * larsivi Ping timeout: 480 seconds
1210407243 Q * yarihm Quit: Leaving
1210407609 J * dennis_ ~dennis@dslb-084-059-101-185.pools.arcor-ip.net
1210407646 M * dennis_ hi :)
1210407992 J * JonB ~NoSuchUse@77.75.164.169
1210408892 J * bfremon ~ben@lns-bzn-59-82-252-191-245.adsl.proxad.net
1210409165 Q * hijacker_ Remote host closed the connection
1210409829 J * geb ~geb@AOrleans-151-1-46-31.w90-21.abo.wanadoo.fr
1210409834 M * geb hi
1210410524 Q * balbir Ping timeout: 480 seconds
1210410829 J * dna_ ~dna@150-229-dsl.kielnet.net
1210411129 J * hijacker ~hijacker@213.91.163.5
1210411895 Q * dna_ Ping timeout: 480 seconds
1210412764 J * dna ~dna@150-229-dsl.kielnet.net
1210412835 N * Bertl_zZ Bertl
1210413274 N * Bertl Bertl_oO
1210413318 J * AStorm ~astralsto@chello089077102031.chello.pl
1210414217 M * AStorm hello
1210414236 M * AStorm any new or reintroduced features in 2.6.25-vs?
1210415265 J * pmenier ~pmenier@ACaen-152-1-68-50.w83-115.abo.wanadoo.fr
1210415875 J * MatBoy ~MatBoy@wiljewelwetenhe.xs4all.nl
1210416254 Q * JonB Quit: This computer has gone to sleep
1210418325 J * JonB ~NoSuchUse@77.75.164.169
1210418377 M * pmjdebruijn AStorm: as far as I know, it's just a port of the 2.6.22-vs... and it's not declared stable yet
1210418607 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1210418620 J * mire__ ~mire@85-172-222-85.adsl.verat.net
1210419072 Q * Aiken Quit: Leaving
1210419796 M * AStorm mhm
1210419805 M * AStorm including cpu limiting?
1210420282 M * dennis_ i have one host with many ips on eth0. whats the easiest way to forward the different ip to the internal vserver-guest-ip?
1210420373 M * transacid just assign them to the vserver not the host
1210420387 M * transacid or what do you mean?
1210420438 M * transacid s/vserver/guest
1210420439 M * dennis_ i like to controll the ips via iptables and there is only available on host.
1210420510 M * dennis_ at the moment the guests have internal ip like 192.168.0.x/8
1210420572 Q * bonbons Quit: Leaving
1210420572 M * transacid you can still use iptables
1210420645 M * transacid you can either use internal ips or public
1210420651 M * transacid doesn't matter
1210420665 M * dennis_ i tried this: iptables -t nat -A PREROUTING -s $EXTIP -m tcp -p tcp -j DNAT --to-destination $INTIP
1210420675 M * dennis_ but it dosnt work
1210420741 M * transacid example:
1210420742 M * transacid iptables -t nat -A PREROUTING  -p tcp  -s ! 10.0.0.0/8 -d 62.141.52.87 --dport 6667 -j DNAT -i eth0 --to-destination 10.1.1.123:6667
1210420819 M * transacid yours can't work
1210420854 M * transacid -s $EXTIP doesn't make sense
1210420874 M * transacid then it only routs traffic from the host to the guest
1210420904 M * dennis_ ahh... okay... i understand.
1210421174 M * dennis_ it works! :-)
1210421183 M * dennis_ thanks a lot =)
1210421190 M * transacid and don't forget the iptables -A INPUT -p tcp -d 10.1.1.123 --dport 6667 -j ACCEPT
1210421234 M * transacid and iptables -t nat -A POSTROUTING -s 10.0.0.0/8 -o eth0 -j SNAT --to-source 62.141.52.87
1210421353 M * dennis_ whats the reason for the both iptables?
1210421369 M * dennis_ the first accept the general input?
1210421394 M * dennis_ the second defines ip for outgoing traffic?
1210421511 M * transacid zap
1210421513 M * transacid yap
1210421542 M * transacid the input is mandatory otherwise the traffic will be routed but then droped
1210421653 M * dennis_ okay. :)
1210421727 M * transacid and postrouting should be slefexplaining
1210421745 M * transacid shure you have to set the right source ip and destination ip
1210421748 M * transacid arg
1210421755 M * dennis_ but the postrouting doesn't work for me, because the outgoing ip depends on internal-ip  (192.168.0.3 => 78.47.48.3, 192.168.0.4 => 78.47.48.4, etc.)
1210421760 M * transacid damn keyboard layout
1210421772 M * dennis_ hehe
1210421778 M * transacid dennis_: for each ip a postrouting statement
1210421817 M * transacid e.g.:
1210421822 M * dennis_ iptables -t nat -A POSTROUTING -s 192.168.0.4 -o eth0 -j SNAT --to-source 78.47.48.4  ?
1210421827 M * transacid yes
1210421834 M * transacid and so on
1210421874 M * dennis_ okay. :)
1210421929 M * dennis_ its possible to use the prerouting without define any ports?
1210421993 M * transacid sure
1210422015 M * transacid but i do it that way for each port, for extra security
1210422047 M * dennis_ mhh... okay, good argument :)
1210422047 M * transacid but then you could just assign the public ips to the guests
1210422053 M * transacid that's easier then
1210422102 M * dennis_ you're right.
1210422113 M * transacid then you do the routing via the interfaces
1210422121 M * transacid eth0:foo
1210422122 M * transacid and so on
1210422127 M * transacid gotta go
1210422133 M * transacid weather is too nice
1210422176 M * dennis_ time to barbecue. ;)
1210422233 M * transacid :)
1210422335 M * dennis_ thanks a lot for you help. :)
1210422359 M * dennis_ +r ;)
1210423106 Q * pmenier Quit: Konversation terminated!
1210423361 Q * hijacker Remote host closed the connection
1210423556 Q * esa` Ping timeout: 480 seconds
1210423890 Q * doener Quit: leaving
1210424409 J * esa bip@ip-87-238-2-45.static.adsl.cheapnet.it
1210424425 J * marl ~marl@84.13.35.31
1210424545 J * Piet ~piet@86.59.21.38
1210424823 M * marl hi folks, just a reayl quick one! if i give my host an ip of 192.168.0.1/24 and i give 3 of my guests ips in the range 10.0.1.[1-3]/24, and one guest 2 ips 10.0.1.4/24 and 192.168.0.2/24, should all the guests be able to comunicate via network with each other, and only the 4th guest be able to get access to the internet without having to setup routing/iptables in the guests or the host?
1210425106 J * hijacker ~hijacker@213.91.163.5
1210425324 Q * nebuchadnezzar Ping timeout: 480 seconds
1210425540 Q * JonB Quit: This computer has gone to sleep
1210425620 J * mire_ ~mire@207-168-222-85.adsl.verat.net
1210425903 Q * mire__ Ping timeout: 480 seconds
1210426262 Q * dennis_ Remote host closed the connection
1210426961 J * mire__ ~mire@133-172-222-85.adsl.verat.net
1210427343 Q * mire_ Ping timeout: 480 seconds
1210428333 J * nebuchadnezzar ~nebu@zion.asgardr.info
1210428475 Q * Piet Quit: Piet
1210428797 J * dna_ ~dna@244-223-dsl.kielnet.net
1210428800 M * daniel_hozac marl: yes
1210428902 J * dna__ ~dna@245-224-dsl.kielnet.net
1210428912 J * mire_ ~mire@251-170-222-85.adsl.verat.net
1210429201 Q * dna Ping timeout: 480 seconds
1210429224 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1210429291 Q * dna_ Ping timeout: 480 seconds
1210429310 Q * mire__ Ping timeout: 480 seconds
1210429385 Q * dna__ Ping timeout: 480 seconds
1210430667 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1210430795 J * mire__ ~mire@3-168-222-85.adsl.verat.net
1210431190 Q * mire_ Ping timeout: 480 seconds
1210432137 J * mire_ ~mire@202-170-222-85.adsl.verat.net
1210432511 Q * mire__ Ping timeout: 480 seconds
1210434093 J * mire__ ~mire@38-170-222-85.adsl.verat.net
1210434343 Q * mire_ Ping timeout: 480 seconds
1210437438 J * Linus ~nuhx@bl7-129-45.dsl.telepac.pt
1210438396 J * pmenier ~pmenier@ACaen-152-1-68-97.w83-115.abo.wanadoo.fr
1210440095 M * marl im sure i read someware that it is posible to configure a guest ip via dhcp, but i cant find it now, was i dreaming or is it posible?
1210440293 Q * pmenier Quit: Konversation terminated!
1210440682 M * marl ok, found the details on it, but its a big faig on what file to but the option in, it says to endbale enter IPROOT=dhcp in the configuration file, but it doesnt say which config file to enter it into :(
1210440962 M * daniel_hozac that's for jacques utils. they only have one configuration file, /etc/vservers/<guest>.conf.
1210441129 Q * ktwilight_ Ping timeout: 480 seconds
1210441208 J * JonB ~NoSuchUse@77.75.164.169
1210441233 M * marl :( is it posible to get a guest to use dhcp for getting its ips?
1210441276 J * esa` bip@ip-87-238-2-45.static.adsl.cheapnet.it
1210441289 Q * esa Ping timeout: 480 seconds
1210441300 M * daniel_hozac sure, it's just that nobody's done the work and shared it.
1210441386 M * daniel_hozac i've got it on my TODO.
1210441536 J * ktwilight ~ktwilight@87.66.195.80
1210441843 J * esa bip@ip-87-238-2-45.static.adsl.cheapnet.it
1210441956 Q * esa` Ping timeout: 480 seconds
1210442188 Q * ktwilight Quit: dead
1210442660 J * ddub ~ddub@213.219.143.229.adsl.dyn.edpnet.net
1210443583 A * kiorky_ is listening to Hammerfall - Secrets
1210443587 M * kiorky_ rah <3
1210443597 M * kiorky_ oups wrong window
1210443612 Q * kiorky_ Quit: Reconnecting
1210443624 J * kiorky ~kiorky@cryptelium.net
1210443983 Q * JonB Quit: This computer has gone to sleep
1210444325 J * hparker ~hparker@linux.homershut.net
1210445063 N * Bertl_oO Bertl
1210445069 M * Bertl back now ...
1210445118 J * Piet ~piet@tor.noreply.org
1210445212 J * simon ~simon@diogenes.anarcho.dk
1210445229 M * simon what's the preferred way in debian to autostart a vserver?
1210445252 M * Bertl withe the vserver-default runlevel script and a 'default' mark
1210445454 J * JonB ~NoSuchUse@77.75.164.169
1210445879 M * simon Bertl, on some distros that's apparently accomplished with "rc-update add vservers default", but does this differ in debian? I haven't got rc-update.
1210445911 M * daniel_hozac update-rc.d
1210445918 M * simon ah!
1210445925 M * Bertl well, on other distros it is done with chkconfig --add vserver-default, I don't have rc-update either :)
1210445953 M * simon maybe update-rc is a gentoo thing
1210445960 M * Bertl looks like
1210446156 M * simon I haven't got any /etc/init.d/vserver* files yet, are they supposed to be created dynamically?
1210446174 M * daniel_hozac look for util-vserver.
1210446231 M * simon I've got that.
1210446598 M * Bertl and on debian, make sure to get the utils from backports
1210447046 J * rgl ~rgl@lx2-84-90-9-93.netvisao.pt
1210447049 M * rgl hello
1210447056 M * Bertl hi
1210447068 M * rgl its a bad ideia do use a dynticks kernel in a server?
1210447092 M * Bertl you mean, as option for the host kernel?
1210447106 M * rgl yup
1210447122 M * Bertl no, should work with all recent versions
1210447165 M * rgl how about a non preemptable kernel?
1210447179 M * Bertl no problem with that either
1210447188 M * rgl 100_HZ timer?
1210447206 M * Bertl instead of dynticks?
1210447221 M * rgl its seems they are complementary :|
1210447244 M * rgl I mean, if they weren't, I was expecting I could not choose the timer freq.
1210447885 M * AStorm use 1000 Hz with dynticks
1210447900 M * AStorm lower Hz are only meant to reduce power usage
1210447954 M * AStorm maybe slightly less overhead, not good for interactive usage
1210447970 M * AStorm (same with preempt)
1210447975 M * Bertl you know what the purpose of a tickless kernel is?
1210447993 M * AStorm safe power
1210447997 M * AStorm *save
1210448009 M * AStorm while not reducing interactivity
1210448009 M * Bertl well, yes, and how is that accomplished?
1210448035 M * AStorm irqs are fired less often, so system stays in high C states more
1210448048 M * Bertl less often means?
1210448055 M * AStorm not every 0.1s, but when neede (event driven)
1210448067 M * JonB Bertl: that the CPU has more time to run the idle process
1210448068 M * AStorm 0.1ms :P
1210448076 M * AStorm Bertl: use powertop
1210448080 M * Bertl 100Hz is 0.01s and 1kHz is 1ms
1210448088 M * AStorm it prints who wakes system up
1210448099 M * AStorm ok, my bad :P
1210448109 M * Bertl so how does 100Hz or 1kHz timer match a tickless kernel?
1210448109 M * AStorm 10 ms iisn't good enough, btw :P
1210448132 M * AStorm it's the maximal rate in case of dyntick kernel
1210448173 M * AStorm also some older driver use HZ for calculations (deprecated)
1210448181 M * AStorm *drivers
1210448184 M * Bertl so how does it affect interactivity?
1210448243 M * AStorm dynticks shouldn't
1210448250 M * AStorm if hz is high
1210448268 M * AStorm low hz limits scheduler granularity and number of interrupts
1210448290 M * AStorm if too low, you may get stuttering or lag in some cases
1210448359 M * AStorm additionally, cpu accounting becomes less precise
1210448426 M * AStorm the drawback of dynticks is that they don't work properly on certain (broken, old) hardware
1210448450 M * AStorm dynticks need a timer that's running when the rest of the system is sleeping
1210448488 M * AStorm usually either APIC or HPET don't go to sleep
1210448498 M * AStorm rarely PIT doesn't too
1210449305 J * mrfree ~mrfree@host74-181-dynamic.13-79-r.retail.telecomitalia.it
1210449777 Q * mrfree Quit: Leaving
1210450191 M * rgl I'll have to probablly test with different HZ *G*
1210450219 M * Bertl IMHO that setting should only affect timer resolution (for low res timers)
1210450236 M * rgl and not the scheduler timer?
1210450291 M * rgl humm but AStorm said it does affect scheduler *G*
1210450348 M * AStorm it will
1210450358 M * AStorm for low value of HZ
1210450366 M * AStorm that is, HZ=100
1210450412 M * AStorm as the scheduler wants very high precision, normally 6ms
1210450445 M * rgl why 6ms?
1210450481 M * AStorm because that value provides nice interactivity feel w/o too much overhead
1210450484 J * Aiken ~james@ppp118-208-54-233.lns4.bne1.internode.on.net
1210450587 M * rgl for desktop?
1210450668 Q * mire__ Ping timeout: 480 seconds
1210450963 M * rgl humm, I have a 100_HZ;  but when I cat /proc/timer_list I have a timer called tick_sched_timer  which seems to infer that the scheduler is setting up its own timer.
1210451214 J * mire__ ~mire@79-172-222-85.adsl.verat.net
1210451235 M * AStorm nah
1210451251 M * AStorm it may , if you have hrtimers enabled
1210451276 M * AStorm that timer is max. the same HZ as te main one anyway
1210451283 M * AStorm unless with hrtimers
1210451301 M * AStorm but then, you should be using dynticks anyway
1210451307 J * ktwilight ~ktwilight@87.66.195.80
1210451322 M * rgl I'm using synticks and 100_HZ
1210452773 M * AStorm pointless a bit :P
1210452781 M * AStorm use high hz with dynticks
1210452864 M * rgl why pointless?
1210452891 M * AStorm because the savings will be tiny
1210452932 M * AStorm I can drop, to 27 ticks on an idle machine in text console
1210452943 M * AStorm with wifi on
1210452950 M * AStorm and ethernet
1210452984 M * AStorm mouse can grab 60 ticks alone when moving
1210453016 M * rgl ticks == intr?
1210453023 M * AStorm add graphics and you see why 100 isn't enough
1210453032 M * AStorm rgl: yes and no
1210453046 M * AStorm hardware and software irqs
1210453062 M * AStorm plus timer events
1210453068 M * rgl ah, so a tick is one of those?
1210453078 M * AStorm a run of one
1210453143 M * rgl but a hardware/software irq will be handled when its generated, so the 100 HZ has any meaning?
1210453194 M * AStorm yes, it's the maximum number of ticks
1210453208 M * AStorm irqs are checked every that often
1210453213 M * rgl huh?
1210453222 M * rgl you mean linux polls for irq?
1210453275 M * rgl irq aren't handled asap?
1210453362 M * Bertl irqs are always handled asap
1210453391 M * AStorm Bertl: yes and no
1210453395 M * AStorm not sw irqs
1210453404 M * AStorm and hw irqs are maskable
1210453423 M * AStorm it's a funny business out there
1210453436 M * Bertl yes, but nobody masks off hw irqs when sleeping :)
1210453446 M * AStorm most affected are timers
1210453451 M * Bertl that would kind of defeat the purpose
1210453455 M * AStorm yes
1210453526 M * Bertl so, sw interrupts are triggered by software
1210453535 M * AStorm they are
1210453537 M * Bertl (means that something is running alread :)
1210453541 M * Bertl +y
1210453543 M * AStorm yes
1210453553 M * Bertl hardware interrupts are triggered by external events
1210453556 M * AStorm by hw irqs most often :p
1210453566 M * AStorm e.g. sound
1210453570 M * AStorm or usb
1210453570 M * Bertl and, if a driver uses hw irqs, then they won't be masked unless
1210453580 M * AStorm hw irqs have to be short
1210453601 M * Bertl a higher priority interrupt is active, or the irq is serviced, or some irq off section is executed
1210453619 M * AStorm kernel threads often regiter "events" too
1210453625 M * AStorm *register
1210453625 M * rgl you are saying that the hw irq handler will just queue that into some kernel queue, and later (100HZ) it will handle it?
1210453633 M * Bertl in the irq off case, the interrupt is serviced asap (i.e. right after the irq off section)
1210453648 M * Bertl rgl: no, that's definitely not how it works
1210453669 M * AStorm Bertl: you're right, but only with hw irqs
1210453684 M * Bertl but the hw irqs 'delegate' some work to tasklets, workqueues or soft irqs
1210453693 M * AStorm yes
1210453701 M * AStorm thus are affected by hz
1210453707 M * Bertl it is correct that you strive to keep the hw irqs as short as possible
1210453725 M * Bertl no, not on a tickless kernel and not necessarily
1210453738 M * AStorm not on tickless, yes
1210453743 M * rgl so, it works like I said?   kernel queues, and at 100HZ intervails it will handle the heavy duty?
1210453782 M * AStorm no
1210453783 M * Bertl on a tickless kernel, IMHO the HZ value is purely for compatibility and timer (posix) precision
1210453793 M * AStorm Bertl: yes
1210453807 M * Bertl it will neither affect the scheduling decisions, nor the hi-res timer (if enabled)
1210453807 M * AStorm but it limits compat timer precision
1210453812 M * rgl AStorm, so all you were saying does not make sense to me.
1210453817 M * AStorm thus high one is advisable
1210453860 M * AStorm some apps have problems with low hz (e.g. java :P)
1210453870 M * AStorm (broken threading, dumdedum)
1210453884 M * rgl bumbedum? *G*
1210453978 M * rgl Bertl, eg: for gettimeofday(2)?
1210454004 M * AStorm rgl: gettimeofday uses new hrtimers
1210454013 M * AStorm at least with glibc
1210454031 M * rgl how do I known if hrtimers are being used?
1210454034 M * AStorm and even if not, it's not that slow
1210454039 M * AStorm rgl: dmesg
1210454057 M * rgl it should say "hrtimers"?
1210454092 M * AStorm abt. high resolution timers
1210454121 M * AStorm gettimeofday is limited in accuracy on non-hrtimer, low hz system
1210454136 M * AStorm if dynticks work, you must have hrtimers on
1210454255 M * rgl odd, I don't "timers" anywhere on dmesg
1210454327 M * AStorm what about singular?
1210454334 M * rgl nope.
1210454345 M * rgl I can only see it on the kernel config file as CONFIG_HIGH_RES_TIMERS=y
1210454477 M * AStorm hmm
1210454527 M * rgl grep hrtimer /proc/timer_list  works *G*
1210454533 Q * bonbons Quit: Leaving
1210454552 M * AStorm powertop is better anyway
1210454563 M * AStorm just needs a bit of kernel support
1210454598 M * rgl it does not mention hrtimer either
1210454606 M * AStorm it won't
1210454626 M * AStorm it'll say how many ticks/wakeups there are
1210454627 M * rgl oh, a task showed up there saying hrtimer_wakeup
1210454788 M * rgl I gtg.  cya later.  and thx for the tips!
1210455181 Q * JonB Quit: This computer has gone to sleep
1210455379 Q * rgl Ping timeout: 480 seconds
1210460535 Q * snooze Ping timeout: 480 seconds
1210462707 Q * bfremon Quit: Leaving.
1210462707 J * snooze ~o@1-1-4-40a.gkp.gbg.bostream.se