[01:57] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [02:36] nathan_ (~nathan@207.44.202.162) left irc: Quit: My damn controlling terminal disappeared! [03:49] netrose (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [03:57] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [05:45] virtuoso (~shisha@ip114-115.adsl.wplus.ru) left irc: Remote host closed the connection [05:45] virtuoso (~shisha@ip114-115.adsl.wplus.ru) joined #vserver. [05:49] netrose (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [05:57] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [07:49] netrose (~john877@CC3-24.171.21.47.charter-stl.com) joined #vserver. [08:55] Nick change: Bertl_oO -> Bertl_zZ [09:34] netrose (~john877@CC3-24.171.21.47.charter-stl.com) left irc: Ping timeout: 480 seconds [10:33] noel- (~noel@pD952C93F.dip.t-dialin.net) joined #vserver. [10:41] noel (~noel@pD9E0963D.dip.t-dialin.net) left irc: Ping timeout: 504 seconds [13:20] serving (~serving@213.186.189.219) left irc: Ping timeout: 512 seconds [15:11] serving (~serving@213.186.191.84) joined #vserver. [16:24] Doener_aw (~doener@p5082DE9D.dip.t-dialin.net) joined #vserver. [16:32] Doener_zZz (~doener@pD9588786.dip.t-dialin.net) left irc: Ping timeout: 512 seconds [19:18] tanjix (ViRu_@c-180-201-54.n.dial.de.ignite.net) joined #vserver. [19:50] meebey (meebey@meebey.net) joined #vserver. [19:50] is vserver ready for 2.6 ? [19:53] meebey: jepp. there are experminatal patches. see mailinglist [19:54] how experimental is it? [19:56] see the mailinglist. its discussed yesterday [19:56] Nick change: noel- -> noel [19:56] ok thanks [20:04] hhmm sounds like its experimental :) [20:07] hhm I will still give it a try, [21:33] tanjix (ViRu_@c-180-201-54.n.dial.de.ignite.net) left irc: [21:52] Nick change: Bertl_zZ -> Bertl [21:53] a happy new year! [22:00] stubbsd (~stubbsd@public2-birk1-6-cust242.manc.broadband.ntl.com) joined #vserver. [22:23] &whois Bertl [22:23] its him personally :) [22:27] Bertl: I am trying the 2.6 patch [22:29] hey cool! [22:29] its really experimental, but ... [22:37] :) [22:41] Tamama (~Tamama@a62-216-20-152.adsl.cistron.nl) joined #vserver. [22:41] 'lo [22:48] i just noticed something ... odd [22:48] well probably not odd, i just dont see how or why it is doing it :) [22:49] let's hear! [22:50] apache + MPM=perchild [22:50] it starts a main server that spawns worker processes.. running as specified UID/GID [22:50] you can set a UID/GID pair for each vhost [22:51] stubbsd (~stubbsd@public2-birk1-6-cust242.manc.broadband.ntl.com) left #vserver (Leaving). [22:51] it just hangs, stopped serving pages :) [22:52] hmm, what patches? [22:52] vs1.22/quota [22:53] hm i should try and run this outside a vserver.. and see what happens :) [22:53] hmm, try to analyze the 'state' of the task/threads ... [22:54] maybe it is waiting for something ... [22:54] well i see it starts 2 httpd processes... but they quit really fast :) [22:54] (ps -ux lots of times :)) [22:58] i'm currently building it on the root server and going to see if it works there.. [22:58] make sure that the nameservice and other things are the same ... [22:59] the vserver is a complete copy of the root [22:59] same ip's too [23:04] Bertl: 2.6 patch looking nice [23:04] hmm, thanks ... [23:04] what about that log that got send to the ML? [23:05] http://www.paul.sladen.org/vserver/archives/200312/att-0425/01-vserver.log [23:05] the vserver code has improved a LOT over the last year [23:06] meebey: that _is_ a bug/race we'll have to investigate/fix ... [23:06] Bertl: something critical? [23:06] currently the proc stuff is the source of all locking issues ... [23:06] I didnt fire any vserver up yet on my 2.6 system [23:07] ic [23:07] if you do not use preemtion or SMP then those issues should be no concern ... [23:07] hm so if you disable /proc you wont get the problem either/ [23:07] yeah, that's right ... [23:07] or is it some other proc stuf? :) [23:08] SMP I dont have but preemption, should I disable it? [23:08] well, depends, if you want to do some testing, I would suggest enabling preemption _and_ SMP and send me the bug reports ;) [23:08] since I am a vserver newbie, I have to read alot before I can fire my first vserver :) [23:09] Bertl: SMP I may enable after my first tests.. [23:09] if you want a stable system, you have to stick with 2.4.23/vs1.22 [23:13] nah I dont need stable, I use 2.6.0 anyhow ;) [23:15] good ;) [23:15] (http://kernelnewbies.org/wiki/moin.cgi/MustFix) [23:15] :) [23:15] hm [23:15] same problem on root server.. hm... [23:15] Tamama: I think I had that problem too, its apache related [23:15] i was thinking that :) [23:15] Tamama: you need to spawn some apache servers with diff uids [23:15] Tamama: very ugly solution it is [23:15] hm? [23:15] it searches for parenst with that uid and uses them then [23:17] hrm [23:17] you make a little pool for each uid/gid [23:17] then the virtual hosts will be executed by those parents, if there is no parent with that uid, it hangs [23:17] gah [23:17] yeah, its ***** [23:17] that is welll. wank [23:17] back to suexec :) [23:17] Action: Tamama patches apache perchild ;) [23:17] :) [23:17] all it needs to do is spawn 1 parent if not found.. how hard can it be :) [23:17] its crap for webhosting providers [23:17] kinda useless [23:17] how so? [23:17] saves the headaches of 'how will that perl or php script run this time?' :) [23:17] I mean if the provider needs a apache parent for each homepage [23:17] each uid.. but, the parent is only what, 3MB and it uses shared libraries so... [23:17] 1000 sites on one server? [23:17] can get ugly [23:17] at the moment I still prefere suexec [23:18] but its shit for PHP, kinda slower than the module [23:18] loads slower [23:18] about 100ms per request [23:18] if you have a good compiled PHP you can get down to 50ms [23:18] (50ms for spawning the process with loading the php.ini) [23:18] depends on hardware too but its kinda slow.. :) [23:19] yeah right, as module, no overhead [23:19] directly executing the script in a thread [23:19] Happy new year 2 everybody! [23:19] i guess i could use suexec on perl, and use homedir lock on php [23:19] Happy new year virtuoso :) [23:20] homedir lock? PHP has some build in features for it, but they are not that great :) [23:20] safemode, basedir etc [23:21] anyhow with vserver running, its much more secure for the whole server! [23:21] ;) [23:21] heh i dont care about _that_ kind of security... i am more worried about user1 modifying user2 files ;) [23:21] hehehe, yeah thats fun though [23:22] dont ask me what kinda stuff I already did... [23:22] I think the best was "rm -rf /" [23:22] :-P [23:22] lol [23:23] next day the provider forum was full of "Hey my Bulletin forum is completly gone!!" [23:24] they chmod it to 777 [23:24] ... [23:24] heh [23:24] of course I couldnt delete sytem stuff, only gid of www-data :) [23:24] but for worldwriteable stuff.. [23:24] but enough forums, guestbooks and homepages was gone [23:25] enough OT, back to vserver + 2.6 [23:31] suexec is pants too really [23:31] ;) [23:31] since ith 2000 hosts you;d still need 2000 system-users [23:32] true [23:33] and you can't 'fake' them... [23:33] that is, just assign an id :D [23:34] yeah problematic with ftp daemon [23:35] whats the normal way of starting a vserver? [23:35] vserver $name enter start ? [23:35] -enter [23:36] stubbsd (~stubbsd@public2-birk1-6-cust242.manc.broadband.ntl.com) joined #vserver. [23:36] yeah [23:36] vserver name start [23:36] it stops running if no additional daemons are actually started through /etc/rc.d/rc though [23:36] that is, vserver $name status will say its not running [23:37] ah ic, it needs something running [23:37] like ssh [23:37] sshd [23:37] something anyway :) [23:37] the logger should always be running ... [23:37] true syslogd [23:37] infowolfe: Thanks for the tar, is emerging now :-) [23:38] Bertl: hm not for me.. [23:38] has someone a woody base system? I only got sarge [23:38] root@powerhouse:/# vserver dns1 start [23:38] Starting the virtual server dns1 [23:38] root@powerhouse:/# vserver dns1 status [23:38] Server dns1 is not running [23:39] Tamama: no syslogd? [23:39] nope [23:39] dunno why not.. [23:39] hhhmm dns only setup? [23:39] i have other vservers but they have the same (and no syslogd) [23:40] lol funny [23:40] root 676 0.0 0.2 2760 1032 ? S 20:17 0:00 /usr/local/sapdb/indep_prog/pgm/vserver start [23:40] they call it vserver too :) [23:41] i have syslogd running in root server though... hmm [23:42] also it seems the vserver only runs rc ... [23:43] not rc.local etc [23:43] hhmm [23:44] which is probably why syslogd didnt start [23:45] stubbsd (~stubbsd@public2-birk1-6-cust242.manc.broadband.ntl.com) left irc: Quit: Leaving [23:45] ;) [23:45] what should ip= be= [23:46] ip= ? [23:46] depends what you want i guess.. [23:46] I thought I tell there which IP the vserver is allowed to use [23:46] like 10.1.0.100 [23:46] right.. [23:46] Can't set the ipv4 root (Bad address) [23:46] it didnt like it though [23:47] i just set IPROOTDEV=eth0 and IPROOT=some_ip_address [23:48] right now i have them set to real 'world' ips so i dont have any weird problems.. but i used 10.x before, it worked [23:48] k [23:49] aha [23:49] my example was outdated [23:49] # Put the prefix of your network to save some typing [23:49] #WOLK ip= [23:49] # Network device for IP aliases setup [23:49] #WOLK ondev=eth0 [23:49] hehe -WOLK [23:49] (tab completion) [23:49] gehe [23:50] that was 1.0 ? [23:50] :) [23:51] dont know some newvserver.defaults file [23:51] ah i dont use newvserver [23:51] vserver $name build for me :) [23:51] I neither [23:52] and then later vs dup name1 name2 :P [23:53] hm suexec doesnt prevent someone from actually reading other peep's files.. if they know the exact path.. pluh [23:56] I am very exited fireing my first vserver [23:56] :) [23:57] I read so much about vserver, but never had the time to really use it [23:57] IPROOT="10.1.0.100" [23:57] IPROOTDEV="eth0" [23:57] this looks ok= [23:57] ? [23:57] no need for "" [23:58] at least i dont have them :P [23:58] hmm, this will create an alias, ist that what you want? [23:58] but i guess it cant hurt heh [23:58] Bertl: an alias? [23:58] I have eth0 with 3 ips on the main server [23:59] one of it is 10.1.0.100 [23:59] if you specify eth0 either in IPROOT or IPROOTDEV, it will create an alias on startup .. [23:59] if you want to use an existing ip, you have to eliminate the eth0 ... [23:59] is it possible to say that the eth0 in my vserver has IP X [23:59] bertl: nah [23:59] it works fine with an existing ip [23:59] using eth0:2 inside the vserver is not so nice .) [00:00] --- Fri Jan 2 2004