1254962556 M * mugwump Marillion: well, Ubuntu are more bzr-focused really aren't they
1254962604 M * mugwump for all I know they are using bzr or hg internally.  But they publish the changes in git, which is best practice especially for kernel work
1254969262 N * Bertl_zZ Bertl
1254969303 M * Bertl Mark17: you want to get a more recent one for 2.6.31.x, preferably the latest snapshot
1254969460 M * Bertl Mark17: you also want to install one of the crypt packages, i.e. beecrypt or nss (devel versions) and dietlibc (if not already done so)
1254969780 Q * uva_ Ping timeout: 480 seconds
1254970283 Q * isodude Ping timeout: 480 seconds
1254970792 J * saulus_ ~saulus@d003027.adsl.hansenet.de
1254971201 Q * SauLus Ping timeout: 480 seconds
1254971205 N * saulus_ SauLus
1254971246 J * isodude dude@2001:6b0:2:2f0d:204:76ff:fe0d:65c3
1254971768 J * uva bno@220-136-228-238.dynamic.hinet.net
1254975471 Q * sardyno Ping timeout: 480 seconds
1254976924 M * Bertl off to bed again ... have fun!
1254976930 N * Bertl Bertl_zZ
1254977228 Q * Piet Remote host closed the connection
1254977331 J * Piet ~piet@04ZAABBI9.tor-irc.dnsbl.oftc.net
1254979627 J * sardyno ~me@pool-173-75-5-88.pitbpa.fios.verizon.net
1254981201 Q * Piet Remote host closed the connection
1254981321 J * Piet ~piet@04ZAABBKD.tor-irc.dnsbl.oftc.net
1254983037 Q * sardyno Ping timeout: 480 seconds
1254983543 Q * Wophon Remote host closed the connection
1254983884 Q * derjohn_mob Ping timeout: 480 seconds
1254984226 J * julius ~julius@217.20.127.15
1254984323 J * todo ~todo@89.72.37.91
1254984530 J * ghislainocfs2 ~Ghislain@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr
1254985227 J * derjohn_mob ~aj@80.85.196.112
1254986161 Q * nkukard Quit: Leaving
1254986442 J * sardyno ~me@pool-173-75-5-88.pitbpa.fios.verizon.net
1254986451 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1254986906 Q * Mr_Smoke Ping timeout: 480 seconds
1254987129 J * Mr_Smoke ~smokey@layla.lecoyote.org
1254988960 Q * mugwump Remote host closed the connection
1254989867 J * fleischergesell ~fleischer@dslb-088-076-055-136.pools.arcor-ip.net
1254989991 J * thierryp ~thierry@zanzibar.inria.fr
1254989996 M * fleischergesell hi folks, i got one thing to ask: i setup vserver on debian lenny having several external ips all DNATed to the webVserver which runs apache2 named-based virtual hosts - but never sees the "real", meaning the public ips of the hostsystem - how do i solve this? because otherwise i dont see any way to implement name-based virtual hosts on the apache vserver
1254990050 M * fleischergesell the webserver has to listen on publicip1:80, publicip2:80 etc - but how can i achieve this using iptables / whatever tool?
1254990108 M * fleischergesell okay, one solution would be to setup a dummy-ip for each public ip and then DNAT publicip1:80 to dummyip1:80 etc
1254990158 M * fleischergesell but this involves adding a new interfaces for every new public ip and one dummy interface respectively
1254990175 Q * yang Remote host closed the connection
1254990180 J * yang yang@static-ip-62-75-255-125.inaddr.server4you.de
1254990212 N * yang Guest584
1254990225 M * fleischergesell sounds complicated for that little task to achieve - anybody with an iptables-like-solution? anything i could add to my DNAT that lets apache vserver know to which public ip the packets were addressed originally 
1254990227 M * fleischergesell ??
1254990388 M * _Shiva_ as you say: they are _name_ bases vhosts... so why would an IP address matter at all?
1254990430 M * _Shiva_ just bind your Apache to the vserver guest ip rather than the public ip 
1254990604 M * fleischergesell thats what i did and what did not work
1254990643 M * fleischergesell but 
1254990647 M * fleischergesell you're right
1254990650 M * fleischergesell it doesnt matter
1254990663 M * todo fleischergesell, wich version of vserver patch you use on lenny?
1254990699 M * fleischergesell at least not for port 80 vhosts
1254990717 M * fleischergesell ssl will be problematic
1254990732 M * fleischergesell todo: latest version to get via apt
1254990747 M * fleischergesell the thing with ssl is: i can only have ONE vhost per ip
1254990749 M * _Shiva_ fleischergesell: not really... ssl is not an issue 
1254990765 M * fleischergesell so as soon as there is more than 2 ssl hosts, they need seperate IPs
1254990796 M * fleischergesell so i have to DNAT publicip1 to dumm1 und publicip2 to dummy2
1254990810 M * fleischergesell or am i confused again?
1254990861 M * _Shiva_ fleischergesell: first, your assumption is correct when using class1 certs 
1254990903 M * _Shiva_ fleischergesell: class3 certs offer the possibility to add "Aliases" to the CommonName
1254990937 M * fleischergesell well, that is not an option for my project
1254990938 M * _Shiva_ fleischergesell: so there _is_ a way to have Apache-SSL-vHosts... but that's not your problem ;-)
1254990946 M * fleischergesell hehe
1254991024 M * fleischergesell fact for me is: I can only have one ssl vhost per ip so for serving x-ssl-vhosts I have to make apache vserver listen on x IPs
1254991044 M * _Shiva_ fleischergesell: so..?
1254991079 M * fleischergesell is there a way to implement this with iptables? somehow letting the new destination know who the old destination was?
1254991118 M * fleischergesell i dont really like the option with x dummy-ifs for x-public-ips
1254991127 N * Guest584 yang
1254991167 M * _Shiva_ fleischergesell: actually, you may also have your guest's Apache listen to it's guest ip and do IP-based (actualy port-based) vhosting within your guest
1254991202 M * fleischergesell mh, thats true - since there is only one vserver listening on port80 anyway
1254991217 M * _Shiva_ fleischergesell: then dnat publicIP_1:80 -> guestIP:80,  publicIP_2:80 -> guestIP:81, ...
1254991247 M * fleischergesell oh thats what you mean
1254991260 M * fleischergesell mhh, that be interesting too
1254991348 M * fleischergesell actually, that sounds way less complicated and way less error-prone than the other thing where I would have to have x dummys for x public ips
1254991355 M * _Shiva_ or add an Apache/lighttpd/ngnix/whatever to the host and do mod_proxy_* to the guest's Apache
1254991396 M * fback _Shiva_: this won't work with HTTP/1.0 ;>
1254991415 M * fleischergesell well, I personally like the port-based approach more
1254991427 M * _Shiva_ fback: hmm? what exacly?
1254991490 M * _Shiva_ fback: . o 0 ( and who - except 3117-Script-Kid-Scripts - still do 1.0? ;-)  )
1254991516 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1254991692 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net
1254992521 M * harry_ todo: yes, the 2.6.22.19 patch 
1254992564 M * harry_ the others are quite stable, but someone has a bug that i can't really pinpoint... so you can try, but not sure it's bugfree
1254992572 M * harry_ (can one ever be sure...)
1254992766 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it
1254993600 M * todo ok, thanks harry, i try 2.6.22.19
1254994357 M * harry_ Linux ssh 2.6.22.18-grsec2.1.11-vs2.2.0.6 #1 Wed Feb 13 10:47:06 CET 2008 x86_64 x86_64 x86_64 GNU/Linux 11:32:30 up 251 days, 22:59,  1 user,  load average: 0.12, 0.06, 0.02
1254994372 M * harry_ since the last powercut at my hosting firm :)
1254994397 M * harry_ i didn't upgrade to the 2.2.0.7, because the patches didn't involve my setup :)
1254994507 J * mugwump ~samv@watts.utsl.gen.nz
1254994635 N * matthew-1 matthew-_
1254995005 J * kir ~kir@swsoft-msk-nat.sw.ru
1254995612 M * todo nice. harry_, wich hosting company is your?
1254995722 M * sid3windr  11:55:18 up 476 days,  4:12, 17 users,  load average: 0.08, 0.03, 0.01
1254995727 M * sid3windr since my last datacenter move :>
1254995967 M * harry_ todo: the datacenter where my server it
1254995968 M * harry_ is
1254995973 M * harry_ i don't have a datacenter myself
1254995974 M * harry_ ;)
1254996005 M * sid3windr yet!
1254996027 M * harry_ hehe
1254996033 M * harry_ naaaaah.... that's not my thing
1254996035 M * sid3windr :)
1254998031 J * davidkarban ~david@199.123.broadband11.iol.cz
1254998272 N * Bertl_zZ Bertl
1254998277 M * Bertl morning folks!
1254998837 M * harry_ heya Bertl 
1254998848 M * harry_ Bertl: how's stabilizing going?
1254998859 M * harry_ i'm working on a new grsec+vserver patch as we speak :)
1254998875 M * harry_ with latest and greatest 2.3.0.36.16 patch
1254998886 M * harry_ (did make it for 2.6.31.3 tough....
1254998888 M * harry_ though
1254998935 M * Bertl good, that's the kernel the current work is on
1254998961 M * Bertl patch will be up shortly ...
1254998961 M * harry_ it's a VERY minor upgrade (vanilla)
1254998977 M * harry_ only obvious reject is the Makefile
1254999002 M * harry_ 2 chars changed in the code from 2.6.31.1 to the 2.6.31.3 :)
1254999011 M * harry_ (were added actually)
1254999029 M * harry_ 0x42 (for the nerds... B's for humans)
1254999966 M * Bertl and I always thought 'nerds' are humans too :)
1255000260 M * harry_ myths..
1255000265 M * harry_ they are ubermensch :)
1255000706 M * harry_ understand the same things as normal humans, but more :)
1255000896 M * Bertl k, just uploaded patch-2.6.31.3-vs2.3.0.36.17.diff which now has btrfs support .. you might want to rebase your patch on that one
1255001005 M * harry_ mkay...
1255001008 M * harry_ (and damn ;))
1255001026 M * Bertl should be trivial though ...
1255001036 M * harry_ yeah :)
1255001040 M * harry_ it is :)
1255001202 M * harry_ still no switchover to the localversion-layout? ;)
1255001298 M * harry_ 05:56 < spender> but i'm waiting for pipacs to upload his test13 that fixes kernexec with paravirt
1255001301 M * harry_ 05:56 < spender> and also he has to fix some problem he caused with vm guests :p
1255001304 M * harry_ bloody hell
1255001309 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1255001751 Q * BenG Quit: I Leave
1255001901 M * Bertl harry_: is there a grsec-devel irc channel somewhere?
1255002090 J * dna ~dna@69-205-103-86.dynamic.dsl.tng.de
1255002120 M * harry_ yesh
1255002127 M * harry_ #grsecurity
1255002173 M * harry_ also on oftc
1255002399 M * Bertl k, tx!
1255002403 M * harry_ np
1255002773 M * Bertl off for now ... got some work in the basement, bbl
1255002781 N * Bertl Bertl_oO
1255002874 M * harry_ yoooooooo
1255003015 M * harry_ aparently, i still have to wait a bit for the grsec part...
1255003018 J * doener_ ~doener@i59F56F46.versanet.de
1255003121 Q * doener Ping timeout: 480 seconds
1255003905 M * Mark17 I did get the current version from linux-vserver.org for util-vserver, is there a better version for kernel 2.6.31.1?
1255004322 J * geb ~geb@earth.gebura.eu.org
1255004413 M * harry_ ?
1255004420 M * harry_ Mark17: there is a patch for 2.6.31.3 now...; )
1255004422 M * harry_ ;)
1255004609 J * halcyonCorsair ~quassel@leibniz.catalyst.net.nz
1255004824 M * halcyonCorsair can anyone tell me what problem might cause this to show up when starting a vserver? vsysctl: open("."): Permission denied
1255004836 M * halcyonCorsair the vserver in question fails to start after that
1255005280 M * fback harry_: have you changed something on people.linux-vserver recently?
1255005286 M * harry_ fback: yes
1255005308 M * harry_ i renamed the _README_ to README.html
1255005312 M * harry_ and html'ized it 
1255005317 M * harry_ (<pre> tags :))
1255005320 M * harry_ why
1255005329 M * harry_ parser prob?
1255005359 M * fback harry_: could you please repair this html page and move </pre> before </body>? ;-)
1255005372 M * harry_ ow... sorry
1255005372 M * fback hmmm
1255005390 M * fback and cut those additional <html> inside the body ;)
1255005391 M * harry_ lol... what a mistake..
1255005396 M * harry_ happens when you copy paste :)
1255005414 M * harry_ like this?
1255005451 M * fback harry_: there's still <body> before <pre> :)
1255005462 M * harry_ you don't want that?
1255005481 M * harry_ better?
1255005502 M * harry_ just pre tags...nothing more
1255005520 M * fback harry_: otherwise it's malformed html
1255005522 M * harry_ (without the pre, i need the </ br> tags...
1255005526 M * harry_ why's that?
1255005534 M * fback and standard parsers choke on this :)
1255005536 M * harry_ normal html starts with html and title/body tags)
1255005560 M * harry_ or am i too 1999 ;)
1255005571 M * fback there can be only one <body> at the beginning and </body> at the end of the document
1255005582 M * harry_ that's what i had
1255005603 M * harry_ <html><body><pre> then blabla, then </pre></body></html>
1255005664 J * vasko ~vasko@unreal.rainside.sk
1255005785 M * _Shiva_ it maybe some Apache magic that prepends a html header and appends a html footer by itself..? so maybe README.html only needs to be (read: _must be_ ?)  page _content_ and not a full page?
1255005825 M * BWare or README is already html'ified by apache 
1255006105 M * harry_ uhu
1255006691 Q * fleischergesell 
1255007318 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1255008827 Q * dna Quit: Verlassend
1255009122 Q * geb Ping timeout: 480 seconds
1255009792 J * geb ~geb@earth.gebura.eu.org
1255010892 Q * geb Ping timeout: 480 seconds
1255010929 M * Mark17 Now I get the following error: vserver joomla01 start - /usr/lib/util-vserver/vserver.functions: line 899: pushd: /etc/vservers/joomla01/vdir: No such file or directory - Failed to start vserver 'joomla01'  How can I fix this? (a '-' is used for a return/enter)
1255011237 Q * mrfree Quit: Leaving
1255011444 J * geb ~geb@112.4.82-79.rev.gaoland.net
1255012444 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it
1255012876 Q * davidkarban Quit: Ex-Chat
1255013100 J * mathx ~math@sizone.org
1255013150 M * mathx goog'd it, but dont see anything on the site about staggered startup/boot - what's the prescribed method? is there a way to do it load triggered or something? 30 vservers starting hammer the box.
1255013691 Q * mrfree Quit: Leaving
1255013769 Q * _nono_ Quit: Leaving
1255014688 J * _nono_ ~gomes@libation.ircam.fr
1255014741 M * PowerKe Mark17: You've set up something wrong: /etc/vservers/joomla01/vdir should be a symlink to /etc/vservers/.defaults/vdirbase/joomla01
1255014750 M * PowerKe and /etc/vservers/.defaults/vdirbase should point to /vservers
1255014813 M * PowerKe (so basically /etc/vservers/joomla01/vdir is pointing to /vservers/joomla01)
1255014856 M * PowerKe Normally the vserver installation and vserver create scripts take care of all that
1255015116 M * Mark17 reinstalling the vps (as it is a new installation) could be a solution?
1255015177 M * PowerKe check which of the symlinks is wrong/missing: 17:12 < PowerKe> Mark17: You've set up something wrong: /etc/vservers/joomla01/vdir should be a symlink to /etc/vservers/.defaults/vdirbase/joomla01
1255015186 M * PowerKe (wrong paste)
1255015191 M * PowerKe /etc/vservers/joomla01/vdir or /etc/vservers/.defaults/vdirbase
1255015244 M * PowerKe if the first one is wrong, you can probably recreate the guest, if it's the second one, something went wrong when installing the vserver tools
1255015739 M * Mark17 it was the first one that was wrong
1255016357 M * Mark17 new errors :( http://yourpaste.net/3430/ for the errors, for vserver.log see: http://www.streamservice.nl/vserver.log
1255016609 M * mathx any hints on staggered/low-impact startup guys?
1255016982 M * PowerKe Mark17: no idea what's causing that
1255017013 M * PowerKe mathx: Write your own script that uses vserver start for each guest?
1255017040 M * mathx yeah im thinking that, just wondering if it was already done.
1255017049 M * mathx or i could hack the init.d/util-vserver script
1255017060 M * mathx (or whatever it calls to start all the default servers)
1255017070 M * mathx zimbra is such a bloody pig :)
1255017083 M * mathx ive already renumbered it to the highest context #
1255017114 M * Marillion mathx: if you want know "How do I make a vserver guest start by default"?
1255017125 M * daniel_hozac mathx: what do you want to achieve?
1255017221 M * mathx daniel_hozac: seems the whole machien thrashes pretty hard at startup trying to launch 50 ssh'ds and mysqlds and whatever else
1255017244 M * mathx some servers can really stay offline for 10-20 minutes while the important one start, so perhaps load based or otherwise would be a good trigger to start each next vserver.
1255017297 M * Marillion CPU share is a good idea, i believe
1255017329 M * Marillion if i understand your problem correcly
1255017330 M * mathx well actually load is the problem. 8 core sunfire.
1255017332 M * daniel_hozac you can change NUMPARALLEL in /etc/init.d/vservers-default
1255017337 M * mathx ah, nice.
1255017342 M * mathx it's really disk io. (not enough disks!)
1255017366 M * mathx (which spikes load, cpus are not even 20% busy at startup)
1255017453 M * PowerKe maybe you can also hashify the sshd and mysqld, then you have to read it only 1 time from the disk for all the guests.
1255017889 M * mathx ah.
1255017903 M * mathx well, i think the problem with mysqld isnt loading the binary, it's the damn table check crap it does at the start.
1255017915 M * mathx (good luck to me converting all my customers to postgres...)
1255017951 M * mathx ill check out the NUMPARALLEL thing, i think i tuned that before. will check again...
1255018036 M * Mark17 PowerKe: next week I will reinstall that server with Debian Etch and use that for it
1255018063 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1255018081 J * ViRUS ~mp@p579B5B8F.dip.t-dialin.net
1255018119 M * mathx is it safe to change /etc/vserver/$name/context values while they're running?
1255018129 M * mathx does anything inside the vserver use that file during run?
1255018135 M * mathx or just outside commnads? (will stop work? :)
1255018139 M * mathx spose that i shouldnt... heh.
1255018238 M * daniel_hozac stop won't work.
1255018289 M * mathx figured. ill shut them down when i can and renumber.
1255018309 M * daniel_hozac Mark17: etch is oldstable. you don't want to use that for production.
1255018319 J * nkukard ~nkukard@196.210.204.155
1255018752 J * fleischergesell ~fleischer@dslb-088-076-055-136.pools.arcor-ip.net
1255018850 M * fleischergesell hi guys, how do you usually solve this: vserver1 running apache+php vserver2 mail (postfix etc) - php scripts on vserver1 should be able to send mails, but cant becasue obv there is no /usr/sbin/sendmail - what can I do to enable php to send mails (possibly through my mailserver vserver2)?
1255018877 M * daniel_hozac install ssmtp
1255018898 J * imcsk8 ~ichavero@148.229.1.11
1255018912 M * fleischergesell and how do i tell php how to use it?
1255018941 M * daniel_hozac it will automatically.
1255018948 M * daniel_hozac since it'll have a sendmail link.
1255018956 M * fleischergesell okay
1255018978 M * fleischergesell so i have to configure ssmtp then to send mails using my postfix on vserver2?
1255019020 M * daniel_hozac yes.
1255019044 M * fleischergesell okay, now comes the tricky part: what do i tell ssmtp to do?
1255019066 M * mathx you can also mail out from php directly.
1255019071 M * daniel_hozac to send emails through postfix on vserver2.
1255019073 M * fleischergesell what is this called what i'm about to implement?
1255019075 M * mathx but then spooling is a problem...
1255019091 M * fleischergesell ya but i dont know how.. so i want to google but i need to know what people call the thing i want to do :D
1255019104 M * mathx just install a postfix with a smarthost on the one with the php
1255019106 M * mathx set it up to SEND ONLY
1255019114 M * mathx might be a bit heavy, im just an anti-exim bigot ;) i HATE exim.
1255019123 M * mathx tho there are packages with smaller footprints that im not recalling
1255019124 M * fleischergesell mathx: i dont really want to install postfix on the webserver-maschine
1255019127 M * mathx somehting like simple smtp or something
1255019144 M * mathx look around there's like 50 mailers :) one is perfect for sending only spooling and gives a sendmail binary
1255019153 M * mathx (well, i mean an executable called sendmail for you to use :)
1255019184 M * fleischergesell so spooling is what i want to do?
1255019204 M * daniel_hozac ssmtp is just that...
1255019224 M * fleischergesell ya, but now i know the word i can google for on how to set this up...
1255019241 M * fleischergesell (using ssmtp)
1255019286 M * mathx you want ssmtp to use your mail server as a SMARTHOST
1255019301 M * mathx that term has been used in postfix and smail for a while, I think even sendmail refers to that
1255019315 M * mathx will send all outbound mail to it
1255019332 M * mathx but, php does send mail directly, you shouldnt even need a package on the webserver, it will use the mailserver as the spooler ....
1255019340 M * fleischergesell oh i recall that term
1255019353 M * mathx but that sounds like more than you need, ssmtp is an extra step
1255019363 M * fleischergesell ahh.. but php does NOT know that there is a second vserver acting asa mailserver..
1255019409 M * fleischergesell so how do i tell php to not use sendmail but my mailserver? (i suppose this should be easier than setting up a smarthost)
1255019425 M * mathx aha.
1255019445 M * fleischergesell and by "tell php" i do not mean "tell my scripts" because its not my scripts
1255019449 M * mathx i had a bin/sendmail so i jsut used mail(). i did see that there was a whole socket smtp sender in php though
1255019462 M * mathx so, yeah either install that in php (which could be a pain) or ssmtp (which might be easier :)
1255019473 M * fleischergesell hehe
1255019486 M * mathx if they're looking for sendmail already then yeah, sendmail executable form ssmtp...
1255019491 A * mathx finally gets it.
1255019497 M * fleischergesell hehe thanks
1255019510 M * fleischergesell will try my best to do it with ssmtp -> smarthost
1255019513 M * mathx hopefully ssmtp has a nice debian config that lets you specify smarthost.
1255019732 M * fleischergesell ya.. what do i have to tell postfix in order to act as the smarthost?
1255019740 M * daniel_hozac to listen on port 25.
1255019820 M * fleischergesell well, my postfix obv listens there - i meant any special thing - well, ill ask google 
1255019837 M * fleischergesell thanks for your help tho
1255019851 M * daniel_hozac just act as a relay for that guest's IP.
1255020369 J * nkukard_ ~nkukard@196.212.73.74
1255020483 M * fleischergesell mh that was easy
1255020496 M * fleischergesell mailhub=postfixvserver
1255020509 M * fleischergesell no more settings to change on postfixvserver
1255020522 M * mathx postfix just needs to have the ip your connecting from in its my_networks
1255020522 M * fleischergesell that sounds not very secure to me..
1255020531 M * fleischergesell well, apparently not :D
1255020548 M * mathx vserver might be coming in from localhost? not sure what it sees
1255020548 M * PowerKe it probably allows from a /24
1255020559 M * mathx that can get you nice and RBL'd, cuz this one time, I...
1255020572 M * mathx i did exactly that, and the !(*@#&(!*&holes on the next /29 used my /29 to spam
1255020574 P * gnuk NoFeature
1255020574 M * mathx (!)
1255020589 M * fleischergesell mhh
1255020595 M * PowerKe depends if you're using public or private IP's on your guests
1255020596 M * fleischergesell normally, i have to auth myself to do smtp
1255020610 N * Bertl_oO Bertl
1255020618 M * Bertl evening folks!
1255020674 M * fleischergesell my_networks from postfix is 127.0.0.1
1255020732 M * Bertl mathx: you might consider unification to improve startup, if your guests are resonably similar
1255020801 Q * nkukard Ping timeout: 480 seconds
1255020809 M * Bertl fleischergesell: make sure that you do not have single ip special casing on when you are using lback virtualization on your guest
1255020811 M * fleischergesell oh god, i'm a little...
1255020815 M * fleischergesell guess what i did..
1255020830 M * fleischergesell i used the postfix that has the address configured that i was sending to
1255020834 M * fleischergesell no wonder it accepts it
1255020871 M * fleischergesell hopefully it does not accept it for any other destination-adress
1255020876 M * fleischergesell *testing*
1255020945 M * mathx Bertl: they're not similar :/
1255020956 M * fleischergesell hehe.. relay access denied.. puh
1255020957 M * mathx some are ubuntu some are debian (some etch, some lenny, some even squeeze)
1255020967 M * fleischergesell for a moment i was in fear i had an open_relay running
1255020971 M * mathx really, my problem is pure IO on data, and backing up 6 million files.
1255020997 Q * thierryp Ping timeout: 480 seconds
1255021001 N * nkukard_ nkukard
1255021009 M * mathx well, actually the 6M files isnt a problem, its nicely preemptible with ionice -c3 rsync, my problem is when rsync gets into a large write thats not so preemptible. io goes into the toilet.
1255021024 M * mathx (not strictly a vserver problem.)
1255021061 Q * opuk Ping timeout: 480 seconds
1255021094 M * fleischergesell mathx: thanks for your help, everythings working fine now
1255021111 M * Bertl mathx: should be a lot better with 2.6.31+
1255021180 M * fback Bertl: really?
1255021193 M * Bertl daniel_hozac: what's the status of the 'COW' flag support? should I test with my own command tool or is there something on the way?
1255021228 M * fback Bertl: after upgrade to karmic (.31) unpacking kernel sources just kills responsivnes 
1255021229 M * Bertl fback: yes, basically it started going bad with 2.6.24 .. reached the worst state around 2.6.26 and started to improve on 2.6.30
1255021255 M * Bertl fback: server or workstation?
1255021267 M * fback workstation
1255021289 M * Bertl you should consider patching in Con's scheduler then :)
1255021332 M * Bertl (the BFS scheduler :)
1255021351 M * fback Bertl: I don't use debian-supplied kernels on servers, as opposed to ubuntu workstations ;-)
1255021395 M * daniel_hozac Bertl: hmm, so V_COW is set via vc_set_iattr?
1255021418 M * Bertl yep, it should be, the defines should be already there
1255021436 M * daniel_hozac they're in linux/fs.h
1255021446 M * daniel_hozac well, the kernel ones.
1255021450 M * daniel_hozac i don't see any in the IATTR API.
1255021511 M * Bertl sec, maybe I forgot about them, in any case, you can assume they are there
1255021525 M * Bertl #define IATTR_COW       0x00080000 
1255021536 M * Bertl in linux/vserver/inode.h
1255021553 M * fback otr, it took me more than a year to proove our lan team, that my workstation is hardware-broken
1255021570 M * daniel_hozac new in .16, i guess?
1255021578 M * daniel_hozac i just noticed i was looking at .15.
1255021579 M * daniel_hozac hehe.
1255021584 M * Bertl could be, yes
1255021634 M * fback do you have some news from Hollow?
1255021646 M * Bertl daniel_hozac: I still have to update the get/set_iattr, but I need to extend the test scripts for that too, thus I'm asking for util-vserver support :)
1255021662 M * daniel_hozac setattr is there.
1255021690 M * Bertl not completely, it is missing in the mask check :)
1255021702 M * Bertl i.e. won't work atm :)
1255021726 M * daniel_hozac i meant in util-vserver ;)
1255021736 M * Bertl ah, k, my bad ...
1255021932 Q * fleischergesell 
1255022216 M * daniel_hozac okay, showattr is there as well. c is for no CoW, C is for CoW.
1255022311 Q * derjohn_mob Ping timeout: 480 seconds
1255022337 M * Bertl excellent, should we do the migration from current IS_COW() to the new flag in the kernel or in userspace?
1255022389 M * Bertl probably for now it's best to report and handle both as COW
1255022427 M * Bertl the question is, should I report back COW_OLD as COW, or do we do that transition at some point in userspace?
1255022430 M * daniel_hozac can you add a bit to the vci info, or bump the version, or something, to say whether that bit should be used?
1255022449 M * Bertl sure, no problem there, what would you prefer?
1255022461 M * mathx Bertl: what should, writing large files without any preemption?
1255022463 M * daniel_hozac either one is fine by me.
1255022484 M * Bertl bit probably is more appropriate, we might end up with 'other' versions
1255022508 M * Bertl mathx: hmm?
1255022536 M * Bertl (read, I didn't understand the question :)
1255022538 M * mathx Bertl: you said it'd get better in 2.6.31+ - you referring to IO?
1255022565 M * Bertl yes, the starvations seen with small vs large file IO definitely got better
1255022586 M * Bertl (means, it performs better, not the startvations)
1255022600 M * mathx heh.
1255022606 M * mathx just want the small writes to get through
1255022674 M * Bertl for example, my devel machine here is often doing large writes (when I move backups around) and it happened almost every time that a cscope run (which scans the kernel tree) suddenly took ages
1255022707 M * Bertl that has definitley got a lot better with 2.6.30
1255022742 M * Bertl *grammar
1255023660 N * vServer_User vServer_User_offhomeskis
1255023675 Q * vServer_User_offhomeskis 
1255023738 J * derjohn_mob ~aj@c175246.adsl.hansenet.de
1255024133 M * mathx Bertl: cool.
1255024133 Q * xdr Read error: Connection reset by peer
1255024306 Q * Pazzo Quit: Bye!
1255024499 M * Bertl mathx: it also helps to adjust the I/O scheduler to the workload, i.e. use the different class levels as well as the idle class
1255024919 M * mathx well, i cant really. dont know what my customers are doing :)
1255024926 M * mathx my workload i ionice -c3 for my rsync backups
1255025123 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds
1255025551 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg
1255026764 Q * hijacker_ Quit: Leaving
1255027499 J * thierryp ~thierry@home.parmentelat.net
1255027510 Q * thierryp 
1255027635 M * Mark17 < daniel_hozac> Mark17: etch is oldstable. you don't want to use that for production. << have you a better idea with good documentation and that won't give real issues with creating and starting a vserver VPS?
1255027647 M * Mark17 and it has to be debian or ubuntu
1255027694 M * daniel_hozac http://linux-vserver.org/Installation_on_Linux_2.6
1255027705 M * daniel_hozac change the versions though.
1255027710 M * daniel_hozac to e.g. 2.6.31.
1255027716 M * daniel_hozac and 0.30.216.
1255028000 M * Bertl do we have a 0.30.216 release?
1255028083 M * daniel_hozac no :)
1255028106 M * Bertl for a moment I thought I missed a release :)
1255028980 P * mathx 
1255029252 J * [1]todo ~todo@89.72.37.91
1255029252 Q * todo Read error: Connection reset by peer
1255029252 N * [1]todo todo
1255030143 J * [1]todo ~todo@89.72.37.91
1255030143 Q * todo Read error: Connection reset by peer
1255030143 N * [1]todo todo
1255035299 J * DoberMann[ZZZzzz] ~james@cap31-6-88-180-72-76.fbx.proxad.net
1255035434 Q * bonbons Quit: Leaving
1255036402 Q * ViRUS Quit: If there is Artificial Intelligence, then there's bound to be some artificial stupidity. (Thomas Edison)
1255036547 Q * kir Quit: Leaving.
1255037442 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds
1255037654 J * DoberMann[ZZZzzz] ~james@cap31-6-88-180-72-76.fbx.proxad.net
1255038347 J * blues_ ~blues@aeu12.neoplus.adsl.tpnet.pl
1255038377 M * Bertl hey blues_! got a chance to test one of the newer kernels?
1255038467 Q * blues Ping timeout: 480 seconds
1255039172 P * halcyonCorsair http://quassel-irc.org - Chat comfortably. Anywhere.
1255039696 Q * geb Ping timeout: 480 seconds
1255041307 Q * imcsk8 Quit: Leaving
1255041984 Q * todo Read error: Connection reset by peer
1255042017 J * todo HydraIRC@89.72.37.91
1255042098 Q * todo Read error: Connection reset by peer
1255042129 J * todo HydraIRC@89.72.37.91
1255044180 Q * ghislainocfs2 Quit: Leaving.
1255045808 J * ghislainocfs2 ~Ghislain@adsl2.aqueos.com