[00:12] <RoAkSoAx> kirkland: this is what I wanna do: First, have a loadbalancer that is listening in the network. Then start a webserver that will tell the loadbalancer "This is my IP address, register me". 3. Loadbalancer will grab the IP and add the webserver to its config.
[00:13] <RoAkSoAx> kirkland: so, as far as I understand, uec-component-listener, would be my listener in my first step, correct?
[00:18] <kirkland> RoAkSoAx: right, that's the listening piece, that handles registering a new component
[00:19] <kirkland> RoAkSoAx: see also the debian/*publication*
[00:19] <kirkland> RoAkSoAx: that's what broadcasts to the network that "i'm a eucalyptus thing, up and ready to be registered"
[00:21] <LowValueTarget> how do i run apt-get to only include security updates
[00:21] <RoAkSoAx> kirkland: awesome. That would be my webserver announcing that they are up and running and ready for registration...
[00:22] <RoAkSoAx> kirkland: now, how are the registration scripts in debian/registration/{cluster,common,node,etc} used or "called" whenever the listener obtains a broadcast
[00:23] <RoAkSoAx> (and I do believe those are the ones that actually register the nodes into the cluster, correct?)
[00:28] <coofbar> if I can choose between xen-hvm and xen-pvm for a linux guest - which one should I use?
[00:28] <LowValueTarget> are there any other secure alternatives to puppet?
[00:35] <bluethundr> I have added some of the repos that I saw on launchpad.net but every time I add one the result is the same : Could not resolve 'mirrors.ccs.neu.edu' or whatever.. what am I doing wrong here? http://pastebin.com/2HzVMnq7
[00:37] <RoAkSoAx> kirkland: i got how it runs the scripts, with ACTIONS_DIR. Now, however, if i *just* compile uec-component-listener.c and run it, and then from another server, i just send a avahi broadcast, it should work right?
[00:40] <thedonvaughn> hola, I have ubuntu 10.04 server installed as a KVM hypervisor.  I'm using vmbuilder, however i can't seem to use the options --firstboot and --tmpfs anymore
[00:40] <thedonvaughn> were these taken away?
[00:48] <thedonvaughn> vmbuilder: error: no such option: --firstboot
[01:36] <lyrae> Hi. Ubuntu desktop detected my NIC and internet worked fine, but the server edition didn't. what can i do
[01:39] <slackster> lyrae: 10.04 the both of them?
[01:39] <lyrae> slackster, correct
[01:40] <lyrae> the server gave error during DHCP configuration. so i clicked to skip and set it up later
[01:41] <slackster> do you have more than one NIC?
[01:42] <slackster> (just looking for something simple)
[01:42] <lyrae> slackster, no
[01:42] <lyrae> it's also a netbook, if that matters any
[01:43] <slackster> are you trying to connect wirelessly?
[01:44] <lyrae> slackster, yes
[01:44] <slackster> have you installed a desktop environment?
[01:44] <lyrae> slackster, gui? no
[01:45] <lyrae> i don't think the server edition comes with one anyhow
[01:45] <slackster> it doesn't
[01:45] <lyrae> right
[01:45] <slackster> but it would make connecting wirelessly easier
[01:45] <slackster> you have to do it all manually
[01:46] <slackster> the reason it didn't connect is it doesn't know what to connect to, and if there is authorization/password required
[01:46] <slackster> let me see if I can find someting..
[01:47] <lyrae> slackster, but i dont even think it detected a NIC
[01:47] <slackster> this should get you started: http://www.wirelessdefence.org/Contents/LinuxWirelessCommands.htm
[01:48] <lyrae> let me show you a screenshot first, and tell me if it looks right (will be able to make it work)
[01:48] <lyrae> of iwconfig
[01:48] <slackster> lyrae: what are the results of ifconfig? do you see somthing like "wlan0"
[01:48] <lyrae> not for ifconfig, for iwconfig yes
[01:48] <lyrae> showing you screenshot, one sec
[01:49] <lyrae> slackster, http://img37.imageshack.us/img37/5117/photogp.jpg
[01:50] <slackster> use wlan0
[01:50] <slackster> see my previous link
[01:51] <lyrae> slackster, i am. thank you
[01:51] <slackster> np
[01:53] <lyrae> slackster, the first part didnt work. tryijng out second one (manually)
[01:55] <slackster> you need to do something like "iwconfig wlan0 essid [wirelessnetworkname]
[01:55] <lyrae> i tried. still trying. one sec
[01:55] <slackster> then iwconfig key [key]
[01:55] <slackster> etc
[01:55] <slackster> iwconfig wlan0 key [key]*
[01:56] <slackster> not something you have to worry about with a netbook
[01:58] <slackster> if I were you I would just install desktop edition, then kill gnome when you don't want it
[01:59] <slackster> with something like sudo /etc/init.d/gdm stop
[02:01] <lyrae> yea i might because this isn't working
[02:01] <lyrae> i can't even iwlist scan. says netowrk is down
[02:01] <lyrae> but installing regular ubuntu with a DE adds more stuff to it
[02:02] <lyrae> wanted to keep the netbook ... 'light'
[02:02] <slackster> just stop gdm when you don't want it
[02:02] <lyrae> but it comes with a bunch of other packages too
[02:03] <lyrae> maybe i could try plugging the netbook to an ethernet cable during installation
[02:03] <lyrae> do an update and see if wireless work
[02:03] <slackster> install gnome-lite (think that's available) on top of the server install then
[02:04] <lyrae> slackster, or xfce?
[02:04] <slackster> oh.. looks like gnome lite is a FreeBSD thing
[02:05] <slackster> lyrae: you could.. I don't know if you're trying to avoid excess packages or what.
[02:05] <slackster> but yes xfce is light
[02:05] <slackster> there is a netbook edition of ubuntu also.. I haven't looked at it though
[02:05] <lyrae> yes i was trying to avoid as much as possible. but its alright. ill try xfce
[02:06] <lyrae> brb
[02:15] <tommy_> is there a proper way to upgrade from 8.04 to 10.04 using internal mirror?
[03:28] <enav> tommy_ sudo aptitude update
[03:29] <enav> tommy_ sudo aptitude upgrade  -d
[03:29] <enav> that it
[03:29] <enav> ho man sorry
[03:29] <enav> tommy_  do-release-upgrade
[03:30] <enav> tommy_  https://help.ubuntu.com/10.04/serverguide/C/installing-upgrading.html
[03:45] <KenjiPops> tried that, but keep getting no mirror found, i'm using internal repo.
[03:45] <KenjiPops> seems to be related to meta-release
[03:45] <KenjiPops> looked at meta-release-lts and lucid is not in there
[03:56] <bc> I just came into posession of this cutting edge notebook. http://www.buysellcommunity.com/uploads/120906/ww1/onzxeldbggdk.jpg
[03:57] <bc> Massive 1 GB of disk space and 16 MB of memory.
[03:57] <bc> Darn, I jumped into the wrong year again.
[04:27] <f1yback> bc
[04:27] <f1yback> nice thin client
[04:27] <f1yback> or serial terminal
[04:27] <f1yback> for router gear etc
[04:28] <f1yback> i'd could use a few more myself mine are all in pieces
[06:10] <Nonpython> What do PTR records mean in DNS?
[06:10] <twb> http://en.wikipedia.org/wiki/PTR_record ?
[06:11] <Nonpython> Page not found.
[06:11] <twb> "pointer record (RFC 1035) - Pointer to a canonical name. Unlike a CNAME, DNS processing does NOT proceed, just the name is returned. The most common use is for implementing reverse DNS lookups, but other uses include such things as DNS-SD."
[06:11] <twb> Nonpython: works for me.
[06:12] <twb> Maybe your country is blocking wikipedia?
[06:12] <Nonpython> I am in the US of A.
[06:12] <twb> My sympathies.
[06:12] <Nonpython> Where are you?
[06:13] <twb> .au
[06:13] <Nonpython> Thanks, I use a cached version of wikipedia, I wrote it and it is supposed to check if there is a new version of a page, I tried in the program, when I actually looked it worked.
[06:13] <Nonpython> We have a R18+ rating for games, so nyah!
[06:14] <twb> Technically we have such a rating, it's just that games rated thusly cannot be distributed.
[06:14] <Nonpython> Same thing.
[06:14] <enav1> maybe this is a big stupid question but can i put a wallpaper to my ubuntuserver terminal?
[06:15] <chrismsnz> enav1: i vaguely remember doing something along these lines back in my gentoo days
[06:15] <Nonpython> enav1: framebuffer or GUI?
[06:15] <chrismsnz> search around about "framebuffer"
[06:15] <enav1> just ubuntu seever terminal background
[06:16] <twb> enav1: what does "echo x$DISPLAY" print in your terminal?
[06:17] <enav1> i mean terminal backgrounnd like this   http://is.gd/ce8CV
[06:17] <chrismsnz> try and find an app called "fbdecor"
[06:17] <enav1> thanks
[06:18] <chrismsnz> if it's for the "console" framebuffer
[06:18] <Nonpython> I use that, it rocks!
[06:18] <chrismsnz> won't work if you're in X/Gnome
[06:19] <chrismsnz> ugh, it's not an app - it's a kernel patch
[06:19] <chrismsnz> no idea whether it's included with ubuntu or will apply
[06:20] <twb> If you're trying to put a background picture on the fbcon, you obviously need more homework.
[06:21] <Nonpython> If my server's IP is 69.175.115.18, what do put as my PTR record?
[06:22] <twb> Nonpython: if you don't know what to put, you probably don't need a PTR record.
[06:22] <Nonpython> I have seen two disagreeing setup guides.
[06:24] <twb> You should be reading the Ubuntu Server Guide as at your release.
[06:24] <twb> apt-get install ubuntu-serverguide && w3m /usr/share/doc/ubuntu-serverguide, or so
[06:27] <Nonpython> I am.
[06:41] <|eagles0513875|> hey guys how do i change the repos my server uses from the local mirror to another mirror
[06:45] <Nonpython> what do I do in the reverse data file for 2 sites with the same first ocelet?
[06:47] <Nonpython> I have the clusterfeck of switching to a ubuntu server running BIND from a windows server at the same time as I add a second website.
[07:59] <deslector> hi, what is the best way to install iFolder in ubuntu?
[07:59] <lambrecht> mornin
[11:38] <twb> Suppose I boot with "break", halting the in ramdisk before pivot_root'ing.
[11:39] <twb> When I call "halt" there, the system shuts down, but doesn't cut power -- the screen stays on, showing the console.
[11:39] <twb> If I let the system boot normally, "halt" will power down the machine.
[11:39] <twb> What's the difference, and how can I fake it from within the ramdisk?
[11:41] <twb> The application is for my automated disk flashing boot=install initramfs-tools script that basically just does "wget -O/dev/sda <URL>; halt".
[12:13] <Zider> is there a neater way to add things like gallery2 or phpbb to your vhosts than just copying the dir? I used to use webapp-config in gentoo but it doesn't seem to exist in ubuntu
[13:13] <stanman246> : hi in here i'm trying out ifolder on a 9.04 ubuntu server (openvz). Having troubles installing mono-complete, how can i install it?
[13:32] <zul> ttx: there is a tomcat6 bug there screaming at you
[13:33] <ttx> There are several... which one ?
[13:33]  * ttx should spend some time on tomcat6 today
[13:34] <zul> there is one new one
[13:34] <zul> complaining about the directory structure
[13:35] <ttx> ah right.
[13:41] <soren> ttx: I have a problem with the python-cloud{servers,files} change you made.
[13:41] <ttx> soren: ah
[13:41] <soren> I could have sworn I kindly asked to keep the source package name, by the way..
[13:42] <soren> Anyways, the package is still in Debian under the old name.
[13:42] <ttx> hrm
[13:42] <soren> "the package" == python-cloudservers. cloudfiles is a different kettle of fish.
[13:42] <soren> ..and my sponsor is reluctant to change it.
[13:42] <soren> Python packaging policy says that the module name should be used for the package name.
[13:43] <soren> Since we didn't (thank goodness!) change the module name (i.e. you still do "import cloudservers" to use it), the correct package name is "python-cloudservers".
[13:43] <ttx> the correct source package name or the correct binary package name ?
[13:44] <soren> binary.
[13:44] <stanman246> how do i install mono-complete on jaunty?
[13:44] <soren> stanman246: You upgrade to karmic.
[13:44] <stanman246> lol
[13:44] <stanman246> touche
[13:44] <soren> ttx: My sponsor in Debian would accept python-cloudservers-rackspace.
[13:45] <soren> ttx: He does not want to rename the source package at all. I agree with that. The upstream name for the software is python-cloudservers. We can't really change that.
[13:46] <ttx> soren: let me gather my thoughts
[13:46] <soren> *nod*
[13:47] <soren> It's going to be a hassle to maintain them in both Debian and Ubuntu this way, and I really do want to do that.
[13:47] <soren> python-cloudfiles has the same concerns, except it's not in Debian ATM, so it doesn't have the troublesome sycning problem.
[13:51] <diago> I have a karmic server I run by virts from. Is it possible to install suite lucid from the karmic vmbuilder?
[13:51] <diago> s/by/my/
[13:52] <bogeyd6> great, installed php5 and ubuntu screwed y virtual hosting
[13:52] <Japje> then you should fix it
[13:53] <bogeyd6> Japje, cool story bro
[13:53] <soren> diago: I'm afraid not. It's not super difficult to do, but I'm not sufficiently motivated to do it myself, and noone has sent a patch.
[13:54] <ttx> soren: I agree with you, you should open a bug about it on both packages
[13:54] <diago> ok, but iso install with minimal will be fine?
[13:54] <Japje> yeah i thought so myself. My glass ball and my thee leaves couldnt really determine the problem.. so this is the best anwser you can get
[13:57] <Japje> bogeyd6: but perhaps you can enlighten us with the exact problem, then who knows.. maybe someone has an awsner
[13:57] <ccheney> hi guys :)
[13:58] <bogeyd6> Japje, installed php5, it installed forked apache2, forked apache didnt like my sloppy virtual host style so I had to create a vhost for each "subdomain".
[14:01] <Japje> so your saying.. i had an ugly vhost configuration and now apache is whining about it
[14:01] <Japje> after installing php5
[14:07] <sommer> morning
[14:37] <zul> ttx: do you want me to take nagios off your hands?
[14:39] <ttx> zul: sure, please do
[14:51] <JanC> so, I found which package violates Ubuntu's "postfix is the default MTA" policy...
[14:53] <JanC> bug #582255
[15:08] <zul> SpamapS: ping when you are around
[15:24] <Japje> Any1 have some experience with preseeding postfix? My preseed skippes inet_protocols and i cant find a way to set it
[15:42] <Daviey> looking at bug #572388, that is as designed for 10.04, isn't it?
[15:42] <Daviey> One cluser per subnet.
[15:53] <sbalneav> I just installed Lucid Server x64, setting up software raid.  Won't boot kicks me to a busybox prompt.  http://ubuntuforums.org/showthread.php?t=1474950 seems to indicate that the only solution is to install 8.04?  Any other pointers?
[16:14] <Daviey> sbalneav: That certainly looks like a bug we should investigate, if it's pratical - can you try installing 8.04 - just to check it's a regression.. Then open a bug please.
[16:34] <ttx> zultriaged the Tomcat question
[16:34] <ttx> s/lt/l:t/
[16:37] <Spiralmatrix> hi does anyone know how to setup bind9 with dynamic dns on ubuntu server, or could point me in the right direction of how to do it please
[16:41] <RoAkSoAx> ttx: just saw you changed the status of cloud-loadbalancing blueprint to discussion. Where's the discussion being held?
[16:42] <ttx> it was held at UDS
[16:42] <ttx> "approved" would require the spec to be drafted first
[16:42] <ttx> the drafter should move it to "drafting" when he starts working on it
[16:43] <RoAkSoAx> ttx: oh ok thought it was moved to discussion cause something changed and something else was to be discussed about it
[16:44] <ttx> no -- was moved (back) to discussion because I think moving it to Approved was a mistake.
[16:45] <RoAkSoAx> ttx: oh ok :)
[16:58] <bogeyd6> oh lawdy
[17:28] <LowValueTarget> Does anyone know of an alternative to unison
[17:28] <zul> rsync?
[17:28] <LowValueTarget> Unison is currently eating up too much memory on my servers
[17:28] <LowValueTarget> i need directory mirroring
[17:28] <zul> rsync
[17:28] <LowValueTarget> Is it feasible to use rsync for mirroring
[17:29] <vraa> yeah why not, use --delete :)
[17:30] <LowValueTarget> cool ill try that
[17:30] <LowValueTarget> unison is slow and cumbersome
[17:35] <BrixSat> hello :)
[17:35] <BrixSat> how do i set up my ubuntu as a proxy server ?
[17:35] <mcas> BrixSat: normal http/https proxy?
[17:35] <BrixSat> mcas:  yes
[17:36] <mcas> aptitude install squid :-)
[17:36] <BrixSat> my hosting domain is blocked on my firewall and i want to use a proxy on another vps i have to make redirection of my browser to there
[17:36] <mcas> do you need user authentication?
[17:37] <RoyK> squid or varnish should be good for that
[17:37] <BrixSat> mcas:  that would be prefereble
[17:37] <BrixSat> squid is installed now what?
[17:37] <RoyK> varnish is way faster for reverse proxying than squid
[17:37] <RoyK> but it's somehow worse to configure
[17:37] <BrixSat> squid is good :)
[17:37] <RoyK> BrixSat: google for squid reverse proxy :)
[17:38] <BrixSat> i just need for some tips
[17:38] <joe-mac> upgraded one of my boxes to test from 8.04 to 10.04- rsyslogd is not logging anything except that it starts.
[17:38] <mcas> the main config is /etc/squid/squid.conf which is well documented
[17:38] <joe-mac> auth log, daemon log, all empty
[17:38] <RoyK> squid.conf has 20x more docs than settings :)
[17:39] <RoyK> or 100x
[17:39] <BrixSat> RoyK:  yes it is quite big :p
[17:40] <mcas> BrixSat: this should help you http://wiki.squid-cache.org/SquidFaq/ReverseProxy
[17:41] <RoyK> "/etc/squid/squid.conf" 4963 lines --0%--                                                                                           1,0-1         Top
[17:41] <joe-mac> anybody have any idea why rsyslogd just refuses to log anything now/
[17:41] <zul> mathiaz: ping
[17:41] <mcas> joe-mac: have you checked that rsyslogd is running?
[17:41] <joe-mac> yes mcas
[17:41] <joe-mac> this isn't amateur hour
[17:41] <BrixSat> mcas:  :)
[17:42] <mcas> ok sorry joe-mac
[17:42] <BrixSat> is there any web config for squid :p
[17:44] <RoyK> !ebox
[17:44] <RoyK> BrixSat: that might have a plugin for squid
[17:45] <mathiaz> zul: o/
[17:45] <RoyK> otherwise, the config file should be quite trivial
[17:45] <zul> mathiaz: nm...i just had a question about the mysql version but i answered my own question
[17:45] <mathiaz> zul: I'm glad I could help you so fast ;)
[17:46] <zul> mathiaz: you are awesome!
[17:47] <joe-mac> https://bugs.launchpad.net/ubuntu/+source/rsyslog/+bug/407862
[17:48] <joe-mac> long thread about it there
[17:50] <RoyK> joe-mac: could you update to lucid?
[17:50] <BrixSat> f**** ebox installed a lot of things i did not need :S
[17:50] <RoyK> BrixSat: heh - low on disk space? :)
[17:50] <BrixSat> no low  on ram memory
[17:50] <BrixSat> why do i need postgreesql for example :p
[17:51] <mcas> RoyK: i think joe-mac is still on lucid
[17:51] <BrixSat> jabberd :S
[17:51] <joe-mac> RoyK i have something like 70 8.04 boxes, just upgraded one that was for testing to 10.04
[17:51] <joe-mac> and having this syslog rpoblem
[17:51] <RoyK> ah
[17:51] <joe-mac> thereby making it impossible to know the results of these tests cause nothign is getting logged
[17:51]  * RoyK kind of likes 8.04
[17:52] <RoyK> it's bloody rock stable, for a start
[17:52] <joe-mac> rofl, sure
[17:52] <RoyK> we only have one 10.04 in production so far
[17:53] <joe-mac> yea well, coming from a rhel world ubuntu is a joke
[17:53] <RoyK> 16-core numbercruncher - the scientists wanted new lib versions
[17:53] <mcas> joe-mac: why?
[17:53] <RoyK> a joke??
[17:53] <RoyK> we're moving from rhel to ubuntu these days
[17:53] <RoyK> less hassle, just as stable
[17:54] <joe-mac> there's so many reasons and hacks that i couldn't even articulate any right now. i've just become borg'd by canonical due to the infrastructure i inherited
[17:56] <RoyK> joe-mac: I guess most of it boils down to taste and perhaps religion - use rhel if you like it - I don't :)
[17:56] <joe-mac> no- it boils down to what the overlords say
[17:57] <RoyK> one of the really nice things about ubuntu, is the amount of packages available - tried installing scipy with hdl libs on redhat?
[17:57] <RoyK> it's a PITA
[17:58] <therian> i would like to join in this ubuntu love in, by saying i love it
[17:58] <therian> and 10.04 is ever so fine
[17:59]  * RoyK doesn't like grub2
[18:00] <joe-mac> yea, all the unmaintained, broken packages
[18:00] <joe-mac> so nice
[18:01] <BrixSat> f***k :@ ebox is good but install a lot of crappy things
[18:02] <BrixSat> max memory usage
[18:02] <Zider> it screwed over my config files too :P
[18:04] <BrixSat> :@
[18:04] <joe-mac> if anyone is interested in tghe rsyslog thing, apparently 8.04 was running a super old version so you just add -c 1 to your /etc/default/rsyslog file. case closed.
[18:05] <BrixSat> thanls :) formating vps :S and then squid again
[18:05] <BrixSat> bye :)
[18:10] <JanC> joe-mac: not sure how you upgraded that system, but on all my lucid systems (both upgraded & newly installed) I have -c4
[18:10] <JanC> I don't have any system upgraded from 8.04 -> 10.04 though
[18:11] <joe-mac> JanC: yes... and 8.04 was running version 1
[18:11] <joe-mac> therefore backwards compatibility with version 4, doesn't really work with the config files for version 1
[18:12] <JanC> at least one of those systems was upgraded to every version since before 4.10
[18:12] <pmatulis> joe-mac: i'm surprised you were allowed to go from 8.04.1 to 10.04
[18:13] <JanC> pmatulis: there is nothing preventing you from doing that
[18:13] <joe-mac> pmatulis: version 1 of rsyslog
[18:13] <joe-mac> it was the latest rev of 8.04
[18:13] <pmatulis> joe-mac: oh
[18:13] <joe-mac> either way, if you want to stick to regular old syslog.conf syntax, you need compat with version 1 of rsyslog
[18:13] <danlii> After I upgraded my server from karmic to lynx, I have no console login. It is as if the getty sessions won't start, after the init the console flips over to tty7 and some stupid framebuffer text with console messages, but if i try to change to another tty I just get a blinking cursor and nothing else. What could have gone wrong?
[18:14] <pmatulis> danlii: can you log in remotely (SSH)?
[18:14] <ryoohki> i overwrote /etc/apt/sources.list by using apt-cdrom( which blithly destroys your sources.list and sources.list~( when run twice).  does anyone have a pristine install u.s.a. version i can wget?
[18:15] <JanC> heh, now he's gone
[18:15] <danlii> pmatulis: I could at first yes, but now I seem to have broken that option as well. I can boot with init=/bin/bash if there is some file that needs to be altered.
[18:16] <pmatulis> danlii: was the system specially configured re logins prior to the upgrade?
[18:20] <sh1ny> ryoohki, can you try removing "quiet splash" from the boot line ?
[18:20] <sh1ny> ryoohki, might give a better idea of what's going on
[18:20] <ryoohki> anyone in the u.s.a. i can get a copy of /etc/apt/sources.list from a pristine 10.04 install?
[18:21] <ryoohki> sh1ny: what?!
[18:21] <sh1ny> ryoohki, when you get to grub, press e on the kernel line and remove "quiet splash" from that line ( it's at the end )
[18:21] <ryoohki> sh1ny: this is an apt-get issue
[18:22] <ryoohki> sh1ny: it has nothing to do with booting
[18:22] <danlii> pmatulis: No. But actually, I just got it working. I just booted with the older kernel still left in grub... Thanks anyway. :)
[18:22] <sh1ny> ryoohki, i thought you said you can't get a shell after boot ?
[18:22] <ryoohki> sh1ny: wrong person
[18:23] <sh1ny> ah sorry, ryoohki
[18:23] <sh1ny> wtb reading comprehension and glasses :/
[18:23] <therian> anyone know the ctrl alt del equivent in synergy?
[18:23] <ryoohki> sh1ny: "wtb"?
[18:24] <therian> want to buy
[18:25] <pmatulis> danlii: make sure you open a bug
[18:25] <sh1ny> therian, well at least you can tell whos'a n mmo player :D: )
[18:25] <ryoohki> danlii: look for how the vts( virtual ttys( ctrl-alt-f1, ctrl-alt-f2, etc...)  are started up
[18:25] <therian> haha
[18:25] <therian> shhh
[18:27] <therian> actually i think thats the first time i gave an answer in this channel
[18:27] <sh1ny> therian, :D :)
[18:28] <therian> dont think i could have started any lower tho...
[18:29] <sh1ny> hey, at least you know stuff, that many don't ! :)
[18:30] <therian> haha
[18:35] <ryoohki> danlii: normally it's in /etc/inittab but ubuntu does it differntly
[18:35] <ryoohki> danlii: try looking at /etc/default/rc-sysinit and /etc/init/*tty*
[18:36] <ryoohki> can anyone share a pristine /etc/apt/sources.list from a u.s.a. system?
[18:36] <jpds> ryoohki: For?
[18:37] <sh1ny> ryoohki, http://paste.ubuntu.com/435657/
[18:37] <sh1ny> it's from bulgarian just replace bg.archive with archive
[18:37] <sh1ny> and it will work
[18:37] <ryoohki> jpds: i used apt-cdrom which overwrites /etc/apt/sources( rather than make use of "#" commenting)
[18:37] <billybigrigger> anyone familiar with tweaking nfs?
[18:38] <ryoohki> billybigrigger: a little.  what's the problem?
[18:38] <jpds> ryoohki: For which release, sorry?
[18:38] <sh1ny> ryoohki, other than that it's stock sources.list
[18:39] <ryoohki> jpds: 10.04
[18:39] <JanC> the default archive points to servers in the UK, maybe there are faster servers in the US ?
[18:39] <jpds> JanC: Yes; I'm getting at that. ;)
[18:39] <jpds> ryoohki: http://paste.ubuntu.com/435662/ should be fine.
[18:40] <billybigrigger> ryoohki, what is a good speed for an nfs transfer over gige network? transfers are from sata/sata disks
[18:40] <axisys> i had to use ethtool to force the interface to full duplex because the old switch at work does not support autoneg.. how do I make sure this change survives the reboot ?
[18:40] <billybigrigger> ryoohki, nautilus is reporting 62mb/s just wondering if i can squeeze anymore out of it
[18:40] <ryoohki> billybigrigger: i don't know
[18:41] <Citrate> Does anyone have experience with installing xen on 9.10?
[18:42] <jpds> ryoohki: The ANL mirror is based near Chicago, if you want a geographically closer mirror I can make suggestions.
[18:45] <ryoohki> so it's look like there is no u.s.a. pristine sources.list for 10.04 since i have a bulgarian one and an edited on
[18:45] <ryoohki> sh1ny: thanks
[18:45] <jpds> ryoohki: Well, you can use us.archive.u.c, but that's based in London.
[18:46] <sh1ny> ryoohki, you're welcome
[18:46] <ryoohki> jpds: i just want an unedited one from a u.s.a. install
[18:46] <zul> smoser: ^^^ hehehe
[18:46] <smoser> i opened it
[18:46] <zul> smoser: ah ok
[18:47] <smoser> was trying to write a watch file for http://s3.amazonaws.com/ec2-downloads/
[18:47] <smoser> but i dont think it can be done
[18:48] <ryoohki> sh1ny: i ran "sed -i -e 's#/bg.#/us.#g' sources.list" but it's not really what i wanted
[18:50] <JanC> ryoohki: us.archive.ubuntu.com is the default US config I suppose, but it's not necessarily the best  ;)
[18:51] <sh1ny> ryoohki, ?
[18:51] <ryoohki> sh1ny: yes?
[18:52] <sh1ny> ryoohki, sed -i 's/bg/us/g' sources.list worked fine here
[18:54] <ryoohki> sh1ny: but that's bad advice since you assume "bg" isn't anywhere in the file, which is probably right, but is not a safe assumption without going through file teddiously by eye.
[18:54] <sh1ny> ryoohki, i know
[18:54] <ryoohki> sh1ny: i almost did that but i like to be safe'
[18:55] <sh1ny> ryoohki, it is a bad advice indeed, but in this particular case it's not
[18:56] <ryoohki> sh1ny: since people do searchs for info and this channel's irc logs are available via google i prefer the advice be always good and safe bbut even i sometimes fail to do that
[18:57] <sh1ny> ryoohki, yes you are right, i always assume people are smart enough to do backups, and to check before they use... :(
[19:09] <axisys> should I just add this in in interfaces file ?
[19:09] <axisys> up ethtool -s eth1 speed 100 duplex full autoneg off
[19:09] <axisys> or there is a more elegent way to do it ?
[19:10] <axisys> i am trying to keep the ethtool change permanent
[19:15] <Theravadan> has anyone had success getting lsi's megacli working with ubuntu?
[19:17] <mcas> Theravadan: yes
[19:17] <mcas> there is a mirror with debs but i don't have the link here
[19:18] <Theravadan> mcas, that would be absolutely amazing to find, thanks... alien is erroring out on me trying to convert it to a .deb package
[19:18] <ryoohki> axisys: i agree but i haven't found it using google and don't know off the top of my head
[19:19] <Theravadan> I think if I find this package: dpkg-shlibdeps: error: couldn't find library libpegclient.so.1 needed by debian/megacli/opt/MegaRAID/MegaCli/MegaCli (ELF format: 'elf32-i386'; RPATH: '/opt/lsi/Pegasus:/opt/lsi/openssl:./')   .. I could do it
[19:21] <axisys> ryoohki: some mentioned using pre-up instead.. the redhat solution looks much nicer w/ configs
[19:22] <queso> What's the best way to analyze log files?  Is there something that can easily graph them, etc.?  Particularly I'm looking at nginx access logs.
[19:30] <mcas> Theravadan: ah yes ... i remember
[19:30] <mcas> there was a trick
[19:31] <Theravadan> mcas, what was the trick?
[19:31] <mcas> i think i don't install it with dpkg
[19:31] <ryoohki> axisys: true, i would bet bsd is even better than red hat
[19:31] <mcas> i just unpack the deb and copy the binarys to the right place
[19:31] <mcas> i use mc for that
[19:31] <mcas> please try it
[19:31] <axisys> ryoohki: with solaris i use the driver.conf
[19:31] <Theravadan> mcas, ah mc, never heard of that, i will try it
[19:34] <ryoohki> axisys: should be just one line right? /etc/network/eth0: if eth0 duplex=full auto=never provsion=dhcp,nick=myhost,prefer=192.168.1.112
[19:34] <sh1ny> Theravadan, for crying out loud don't do that :/
[19:34] <sh1ny> Theravadan, there's a repo with megacli for ubuntu
[19:34] <binBASH> Hi sh1ny
[19:34] <sh1ny> hey binBASH ! :)
[19:35] <sh1ny> Theravadan, http://hwraid.le-vert.net/
[19:35] <binBASH> I fixed the cloud yesterday btw.
[19:35] <binBASH> :)
[19:35] <sh1ny> great !
[19:35] <binBASH> reinstalled the cloud master then it worked fine
[19:35] <sh1ny> is it better than centos ?
[19:35] <binBASH> much
[19:35] <binBASH> ;)
[19:36] <sh1ny> what i never get to wrap my brain around is
[19:36] <sh1ny> what happens if one node goes down
[19:36] <sh1ny> and can i run the nodes on top of let's say heartbeat or rhcs cluster
[19:36] <sh1ny> so i get failover and stuff
[19:36] <binBASH> node is just a normal server
[19:36] <sh1ny> yes, but when it fails, then what ?
[19:37] <binBASH> then the image is gone ;)
[19:37] <sh1ny> exactly :/
[19:37] <sh1ny> Theravadan, and also are you sure you're not trying to install a 32bit binary on a 64bit os ?
[19:39] <Theravadan> mcas, heh "rpm2cpio package.rpm | cpio -dimv" worked but I'll follow sh1ny's advice preferably then use the files in the .rpm package as plan B
[19:39] <sh1ny> Theravadan, i got it working on 2 servers with LSI raids just fine
[19:39] <Theravadan> sh1ny, the .rpm file contains a 64 bit binary which worked
[19:39] <binBASH> sh1ny: The hardest thing here is, my provider restricts usage of ip addresses.
[19:39] <binBASH> they are bound to the mac address of the eth0
[19:39] <Theravadan> sh1ny, what did you get working exactly? megacli?
[19:39] <sh1ny> Theravadan, i even did configure the raid with it and it didnt make things blow up the sky
[19:39] <sh1ny> Theravadan, yes :)
[19:40] <binBASH> so I have a non standard Eucalyptus setup here
[19:40] <mcas> Theravadan: i use some debs for my way but have some dependencies which won't work
[19:40] <binBASH> one dhcpd per node
[19:40] <binBASH> ;)
[19:40] <Theravadan> sh1ny, that's a bonus.. how did you get megacli? download it from lsi?
[19:40] <sh1ny> Theravadan, nope, from that site i gave you
[19:40] <Theravadan> sh1ny, nice ok investigating
[19:40] <sh1ny> Theravadan, i installed jaunty packages on karmic and they worked fine, not sure with lucid, but you could try the debian/squeeze ones
[19:41] <sh1ny> Theravadan, still better than rpm's
[19:41] <Theravadan> sh1ny, i have a brand new lsi card so we will see
[19:41] <sh1ny> Theravadan, mine is new too
[19:41] <Theravadan> sh1ny, ok good that is more reassuring
[19:41] <sh1ny> Theravadan, /0/100/9/0            scsi0      storage    LSI MegaSAS 9260
[19:42] <Owner_> hello people
[19:42] <mcas> Theravadan: sh1ny mine comes with some fujitsu servers
[19:42] <Owner_> is there way to connect to ubuntu server from other pc and have gui?
[19:42] <Owner_> openssh is for command
[19:42] <Zider> vnc
[19:42] <Owner_> is there something so i can connect to server (installed kde)
[19:42] <Owner_> ?
[19:43] <Zider> yes, vnc
[19:43] <sh1ny> binBASH, so do you think it's worth to investigate a eucalyptus node on a rhcs cluster ?
[19:43] <Zider> :P
[19:43] <Theravadan> # lspci -nn | grep -i lsi
[19:43] <Theravadan> 05:00.0 RAID bus controller [0104]: LSI Logic / Symbios Logic LSI MegaSAS 9260 [1000:0079] (rev 03)
[19:43] <sh1ny> Theravadan, :)))
[19:43] <Theravadan> sweet, same cards
[19:43] <Owner_> oh VNC is pay version
[19:43] <Theravadan> this is why they have the internet
[19:43] <Owner_> any free software>?
[19:43] <sh1ny> Theravadan, yes works flawlessly here , even tho it's doing a raid6 :)
[19:43] <Zider> there are free vnc implementations
[19:43] <binBASH> sh1ny: Don't think so ;)
[19:43] <Theravadan> sh1ny, ahh, i'm doing raid 10
[19:44] <binBASH> rhcs cluster is quite nice I've heard
[19:44] <mcas> sh1ny: do you try megacli-sas-status?
[19:44] <binBASH> but you can try
[19:44] <binBASH> Don't have experience with it ;9
[19:44] <sh1ny> macs, i am using zabbix and did my own stuff for that :)
[19:44] <Theravadan> sh1ny, <-- uses zabbix too, nice
[19:45] <sh1ny> Theravadan, you dont have access to my servers, right ? :P
[19:45] <Theravadan> sh1ny, I don't see the d/l link here: http://hwraid.le-vert.net/wiki/LSIMegaRAID
[19:45] <Owner_> so with VNC i will be able to connect to ubuntu server with KDE?
[19:45] <sh1ny> Theravadan, http://hwraid.le-vert.net/ubuntu/
[19:45] <Theravadan> sh1ny, wait a minute, why are the tools already installed?
[19:45] <binBASH> sh1ny: Here I'll run 2 varnish infront
[19:45] <binBASH> not in the cloud
[19:46] <binBASH> so if node goes down with eg. webserver it's np
[19:46] <sh1ny> binBASH, ah, so you'll loadbalance with that ? :)
[19:46] <sh1ny> hm good thinking
[19:46] <sh1ny> Theravadan, and i'd recommend using gdebi to install stuff like that
[19:47] <sh1ny> Theravadan, dpkg does less checking and more breaking :)
[19:47] <Theravadan> sh1ny, ok thx
[19:47] <RoAkSoAx> SpamapS: let's discuss it tomorrow or the day after tomorrow so I can draft what I was thinking a little bit better
[19:47] <sh1ny> binBASH, unfortunately i can't get away like that :(
[19:48] <binBASH> sh1ny: Sad ;)
[19:48] <sh1ny> binBASH, and i think i'll go straight for a cluster and not a cloud
[19:48] <SpamapS> RoAkSoAx: sounds good.. I'm still looking at HAproxy and how feasible it is to do with it, what I was originally thinking to do w/ IPVS (which may prove too complicated, given the IPTUN requirement)
[19:48] <binBASH> sh1ny: Well we host some sites of gettyimages.com
[19:48] <binBASH> and they want cloud structure.
[19:48] <sh1ny> uhm no databases involved ?
[19:49] <binBASH> sure
[19:49] <RoAkSoAx> SpamapS: for what I know, HAProxy implementation is like IPVS DR implementation, so that means, in the same network
[19:49] <sh1ny> but they're not on the cloud ?
[19:49] <ryoohki> does anyone here have a pristine copy of a u.s.a. /etc/apt/sources.list from a fresh install of 10.04 i can get a copy of?
[19:49]  * sh1ny hides
[19:49] <binBASH> sh1ny: they'll be in the cloud
[19:49] <binBASH> but replicated
[19:50] <sh1ny> aha
[19:50] <RoAkSoAx> SpamapS: now, we need to define if we want a layer 4 or layer 7 loadbalancing. If it is layer 4, it might be faster (IPVS), if we go for layer 7 (HAProxy), it would not only mean loadbalancing, but proxying
[19:50] <SpamapS> RoAkSoAx: hm thats interesting.. the way I was reading it its more like a traditional proxy
[19:50] <binBASH> their main demand is to have much cpu power, for video processing
[19:50] <sh1ny> well i still think i'll try cloud over cluster, when i get my hands on enough hardware :P
[19:50] <binBASH> :)
[19:50] <binBASH> sh1ny: You run mysql?
[19:51] <mathiaz> SpamapS: RoAkSoAx: IIRC the use case is layer 7 http loadbalancer
[19:51] <sh1ny> i got a book on xenHA with heartbeat ( for free from the author ) and it doesnt seem so hard, so i should be able to apply it to eucalyptus nodes :P
[19:51] <sh1ny> binBASH, not if it's up to me
[19:51] <sh1ny> <- pgsql fan :/
[19:51] <SpamapS> RoAkSoAx: The biggest difficulty with using IPVS in the cloud is that they will not necessarily share physical LAN or even the same subnet.. so we have to assume that a loadbalancer <-> server connection must be established. With HAproxy, that is natural.. with IPVS.. IPTUN has to happen first.
[19:51] <binBASH> ok ;)
[19:51] <RoAkSoAx> SpamapS: and for example, in case we want to add a second loadbalancer in HAProxy, (for failover) we need to use keepalived, which can be used aswell with IPVS
[19:51] <sh1ny> but my company has a "homegrown" software
[19:51] <RoAkSoAx> SpamapS: indeed
[19:51] <sh1ny> something like ERP
[19:51] <binBASH> ok
[19:52] <sh1ny> that's just baddly written and uses mysql
[19:52] <Theravadan> sh1ny, did you use anything other than megacli from that link?
[19:52] <SpamapS> mathiaz: yeah, I think we got side-tracked w/ IPVS because of the persistence discussion, but HAproxy has excellent support for persistence.
[19:52] <sh1ny> so i'm forced to :P
[19:52] <binBASH> :))
[19:52] <RoAkSoAx> mathiaz: if we want layer7 loadbalacing, HAproxy is our solution
[19:52] <binBASH> How you achieve HA there sh1ny?
[19:52] <binBASH> using mysql cluster?
[19:52] <RoAkSoAx> SpamapS: indeed, I'd also think that if we need specialized HTTP loadbalacing, we should go for HAProxy
[19:52] <sh1ny> uhm no HA for it atm
[19:52] <mathiaz> SpamapS: RoAkSoAx: we're looking at providing an image that could do the same thing as AWS Elastic Load balancing
[19:53] <sh1ny> they still havent optimized their db calls
[19:53] <sh1ny> so they create a mess
[19:53] <RoAkSoAx> mathiaz: the loadbalancing is not the issue. The issue is autoscaling features
[19:53] <SpamapS> mathiaz: yes I think that was also lost in the discussion we had
[19:53] <Theravadan> mathiaz, eucalyptus hq is just down the street from me
[19:53] <SpamapS> mathiaz: looking into that tho, we have to create a tool to automatically edit haproxy's configs..
[19:53] <Theravadan> i could knock on their door for you
[19:53] <sh1ny> and i told them no HA until they fix the damn 99.9 loadavg every 3-4 hours :P
[19:54] <RoAkSoAx> SpamapS: I was thinking on autoregistration of a webserver to a loadbalancing domain, and when that happens, loadbalancer (HAProxy) adds it automatically its config. I need to look into HA Proxy first though
[19:54] <sh1ny> Theravadan, not really, tho megacli-status should work
[19:54] <SpamapS> RoAkSoAx: seen this: http://code.google.com/p/scalr/ ?
[19:54] <RoAkSoAx> SpamapS: since I don't really know how healthchecking works in HAProxy.
[19:54] <mathiaz> RoAkSoAx: elastic load balancing doesn't provide auto scaling
[19:55] <mathiaz> RoAkSoAx: auto scaling is another component
[19:55] <SpamapS> http://aws.amazon.com/elasticloadbalancing/ ....
[19:55] <RoAkSoAx> SpamapS: scalr insteresting will look at it.
[19:55] <SpamapS> Elastic Load Balancing can detect the health of Amazon EC2 instances. When it detects unhealthy load-balanced Amazon EC2 instances, it no longer routes traffic to those Amazon EC2 instances instead spreading the load across the remaining healthy Amazon EC2 instances.
[19:56] <sh1ny> SpamapS, which doesn't mean it launches new instances ? :)
[19:56] <RoAkSoAx> mathiaz: I do understand that, but as nijaba mentioned, a feature that is really desired is autoscaling because customers are asking for it
[19:56] <SpamapS> yeah not quite auto-scaling
[19:56] <Theravadan> sh1ny, http://hwraid.le-vert.net/ubuntu/pool-jaunty/megaclisas-status_0.5_all.deb ? that seemd to work but it's megaclisas-status
[19:56] <mathiaz> RoAkSoAx: right - auto scaling is much bigger and requires other components to be in place first
[19:56] <sh1ny> Theravadan, yes yes, that one, sorry i am lazy on typing details :(
[19:57] <mathiaz> RoAkSoAx: (ex all the monitoring infrastructure)
[19:57] <RoAkSoAx> SpamapS: healthchecking in regular cluster solutions, such as ldirectord or ipvsadm can also be done in several ways and in several degrees, if the healthchecking fails, the webserver is removed from the loadbalancing cluster
[19:57] <SpamapS> but later on that page..
[19:57] <SpamapS> "Auto Scaling with Elastic Load Balancing
[19:57] <SpamapS> Let’s say that you want to make sure that the number of healthy Amazon EC2 instances behind an Elastic Load Balancer is never fewer than two. You can use Auto Scaling to set these conditions, and when Auto Scaling detects that a condition has been met, it automatically adds the requisite amount of Amazon EC2 instances to your Auto Scaling Group."
[19:58] <sh1ny> SpamapS, if you're ready to spend money on a cloud, i'd recommend using rightscale with amazon, the guys over there are really nice and they saved me a lot of headache :)
[19:58] <RoAkSoAx> mathiaz: for example, if I wnated to implemented autoscaling in a regular cluster implementation (either keepavelid+ipvsadm or ipvsadm+ldirectord+pacemaker), the only thing I need to do is when webserver is started, it needs to tell the loadbalancer that it is up and running and ready to receive load, then loadbalancer would add that server to the loadbalancing domain
[19:58] <sh1ny>  /comercials off
[19:58] <sh1ny>  s/comercials/commercials/g
[19:59] <Theravadan> sh1ny, do you have a BBU? I want to see if it's actually running and caching writes.  megacli -AdpAllInfo -a0 says it's there but not sure how to see if it's caching.. i can call the manufacturer
[19:59] <SpamapS> sh1ny: you can do commercials, as long as they work like the commercials in The Invention of Lying .. "Coke.. its really just brown sugar water, but a lot of people drink it. Thats right, COKE, Its really famouse."
[19:59] <sh1ny> i don't have a BBU :/ i did turn read cache on tho
[19:59] <sh1ny> Theravadan, megacli -LDInfo -LAll -aAll
[20:00] <sh1ny> Theravadan, i believe that's what you're looking for
[20:00] <RoAkSoAx> mathiaz: in *regular* clsuter infrastructure, the monitoring (healthchecking) is already implemented... whjy implemente something else from scratch if it is already there... we can just adapt it for our needs or extend its functionality
[20:00] <sh1ny> SpamapS, :D :)
[20:00] <sh1ny> note taken :))
[20:01] <sh1ny> Theravadan, ok i lied, seems i have turned on the write cache without a BBU
[20:01] <sh1ny> Theravadan, i'll go cry in the corner
[20:01] <sh1ny> Theravadan, http://paste.ubuntu.com/435695/ :F
[20:02] <RoAkSoAx> mathiaz: so for example, in ldirectord+ipvsadm+pacemaker, ldirectord is the one who determines if the webservers are ready to receive load or not, in case the healthcheck failed, it tells IPVS that the server is not to receive load, and is removed from the node list that will receive load.
[20:03] <RoAkSoAx> mathiaz: that moniutoring is done in various ways, One way is just check if port 80 for example is open, another way is that it checks for a fail accessible through HTTP, other way is request specific data content in a file of the webserver
[20:03] <cybrocop> Hi All. I wonder if someone has gotten Eucalyptus to work on 10.04?
[20:03] <sh1ny> cybrocop, i believe binBASH had, unless he's been lying to me !
[20:04] <cybrocop> It is driving me crazy for a week now.
[20:04] <sh1ny> hm, sounds like my girlfriend
[20:04] <SpamapS> RoAkSoAx: UP/DOWN checks are easy. I think what Mathias is discussing is more collective checks. "If total system load is at 80%, spawn X instances"
[20:04] <cybrocop> :)
[20:04] <cybrocop> I get it installed OK and all seems fine, but when I try to create Windows 2003 images and try to upload them, I get this weird error...
[20:05] <Theravadan> sh1ny, you were faster than support, i see writeback cache is the current policy yay
[20:05] <RoAkSoAx> SpamapS: That's actually something that ivoks said as a cool feature for loadbalancing
[20:05] <cybrocop> "Image registration failed because the manifest referenced is invalid or unavailable."
[20:05] <sh1ny> Theravadan, i am a support guy actually...well top level but non the less ;P
[20:05] <cybrocop> Happens haphazardly, works one time then doesn't work another time.
[20:05] <ryoohki> SpamapS: or coke is really delicious...
[20:06] <ryoohki> SpamapS: and it has caffiene
[20:06] <ryoohki> SpamapS: and sugar!!!!!
[20:06] <Theravadan> sh1ny, here  is my output: http://dpaste.com/196213/
[20:07] <Theravadan> sh1ny, nice! small world.. i'm a programmer+sysadmin
[20:08] <sh1ny> ^_^
[20:08] <RoAkSoAx> SpamapS: and well that's totally something different from loadbalacing itself
[20:08] <SpamapS> ryoohki: in the referenced movie, people aren't able to say things that aren't true.. they haven't evolved lying yet. So they couldn't say "is really delicious" without qualifying it. In that parallel universe, you can't even imply something that is untrue.. so ugly people all dress poorly, while attractive people dress well...
[20:08] <sh1ny>  ^_^
[20:08] <sh1ny> sorry uvirtbot :(
[20:09] <SpamapS> RoAkSoAx: right, it is in essence a signal for an event engine that should drive actions, because you also want "Total system load is under 40%, reduce instances"
[20:09] <sh1ny> Theravadan, actually after looking again my zabbix script is using megasasctl
[20:09] <RoAkSoAx> SpamapS: yep. That I think is a work-item for the spec. Another work item is autoregistration
[20:09] <sh1ny> Theravadan, i do a megasasctl -H -B and it returns ok if no ouput and "YOU'RE ABOUT TO LOOSE YOUR JOB" if anything gets printed
[20:10] <Theravadan> sh1ny, how did you get "Disk Cache Policy: Enabled" and why did you say you would cry?
[20:10] <Theravadan> sh1ny, lol nice
[20:10] <RoAkSoAx> SpamapS: because given load 80%, signal can be sent to say "start webserver in domain A, mount point B" and then webserver will start and use autoregistration, and the same will happen if user would like to start another webserver instance manually
[20:10] <sh1ny> Theravadan, well usually it's a bad thing to enable write cache with writeback when no BBU, or so i've been told
[20:10] <sh1ny> Theravadan, power failure might cause data loss
[20:10] <SpamapS> RoAkSoAx: would make sense to copy elastic LB's M.O. with an  elb-register-instances-with-lb compatible command..
[20:10] <Owner_> does anyone uses tightvnc?
[20:11] <Theravadan> sh1ny, oh yeah, that could be bad. so you explictly enabled writeback?
[20:11] <ttx> SpamapS: questions about the spec writing ?
[20:11] <sh1ny> Theravadan, but i'm relying on that dual power supply i got connected to two separate ups's
[20:11] <ttx> SpamapS: for the sync/merging process, I'll let your mentor explain that to you
[20:11] <sh1ny> Theravadan, raid6 aint't the fastest thing around :(
[20:12] <RoAkSoAx> SpamapS: it will, but if there's a command no "autoregistration" would be needed from the webserver because you would be already passing the parameters to the webserver
[20:12] <RoAkSoAx> SpamapS: and telling the loadbalancer to add that webserver you are instancing
[20:12] <sh1ny> with that i can get 130Mb/s writes and around 200Mb/s reads
[20:12] <SpamapS> ttx: I'll prepare any questions and email them to you directly, since I'm sure you are about ready to end your day. :)
[20:13] <sh1ny> Theravadan, but unless you have *really* good backups and at least two power supplies i wouldn't recommend it without a BBU
[20:13] <ttx> SpamapS: that's a wrong assumption. I surrendered my Tuesday nightlife to my job.
[20:14] <ttx> SpamapS: so i'm looking for easy things to do for the next hour and a half, and explaining things is easy :)
[20:14] <sh1ny> Theravadan, but i have and i only run virtual machines on that, so restoring takes less than 20 minutes so i don't really care...and for 1 year i had 0 troubles so i'll just go along with this till it works ;)
[20:14] <SpamapS> ttx: very well then, let me just finish reading.
[20:14] <ttx> SpamapS: sure
[20:15] <Theravadan> sh1ny, my power is stable so I understand
[20:16] <Theravadan> sh1ny, i have "Disk Cache Policy: Disabled" ... i assume this is referring to the read cache?
[20:17] <sh1ny> Theravadan, yes
[20:17] <sh1ny> Theravadan, you can enable it with megacli
[20:18] <sh1ny> Theravadan, http://hwraid.le-vert.net/wiki/LSIMegaRAIDSAS it's here somewhere
[20:18] <zul> SpamapS: ttx has no life
[20:18] <sh1ny> zul, that was harsh :(
[20:19] <Theravadan> sh1ny, thx
[20:20] <sh1ny> btw Theravadan are you doing this on lucid ?
[20:20] <Theravadan> sh1ny, yes
[20:21] <sh1ny> hm good to know it works, my servers are still on karmic and will be till mid-summer ....lazy and careful on updating
[20:21] <Theravadan> sh1ny, i understand, it has taken me a month and i still havent migrated over from the old centos server to this new ubuntu server, i'm cautious
[20:22] <Theravadan> so glad not to use centos, yech
[20:22] <sh1ny> i'm using centos on some virtual guests
[20:22] <sh1ny> but i wouldnt install it on the host ....mostly because it will take time for me to get familiar with it
[20:22] <Theravadan> sh1ny, my problem w/ it is the packages are really old
[20:23] <SpamapS> ttx: ok so the spec explanation seems fairly straight forward. And from the discussion, as I understand it, the implementation should be stated as work items.. yes?
[20:23] <sh1ny> Theravadan, sometimes in ubuntu newer packages might cause a pain
[20:23] <Owner_> ok any good free VNC ?? can anyone help me set it up?
[20:23] <sh1ny> Theravadan, for example i did a test upgrade on my puppet server
[20:23] <sh1ny> Theravadan, so many things went wrong, i don't know even where to begin with
[20:24] <Theravadan> sh1ny, yikes.. i did a test upgrade from 9.10 and 8.04 to 10.04 without problems
[20:24] <Theravadan> sh1ny, what kinda problems? package problems, did you have 3rd party .deb's?
[20:24] <sh1ny> Theravadan, yes ubuntu itself is fine, just puppet breaks stuff , aaand it has a nasty bug when used with ruby 1.8.7
[20:24] <ttx> SpamapS: yes.
[20:25] <sh1ny> Theravadan, and lucid happens to ship with ruby 1.8.7
[20:25] <ttx> SpamapS: you should work on those where you are set as "Drafter"
[20:25] <ttx> SpamapS: https://blueprints.launchpad.net/~clint-fewbar/+specs?role=drafter
[20:25] <ttx> (quite empty now)
[20:25] <sh1ny> Theravadan, other than that i don't foresee any problems going 9.10 -> 10.04
[20:26] <ttx> SpamapS: I asked Jos to confirm some of those. I think you can safely start with   	 server-maverick-uds-web20-workloads
[20:26] <sh1ny> Theravadan, but there are occasional packages that break backwards compatibility
[20:26] <cybrocop> Owner_: try TightVNC
[20:26] <ttx> SpamapS: and   	 server-maverick-uds-cloud-loadbalancing
[20:26] <sh1ny> Theravadan, from release to release and depending on how important they are for you, it might cause major pain
[20:26] <wizardslovak> cybrocop : i tried i couldnt connect to it
[20:26] <RoyK> Theravadan: I have a couple of test systems upgraded from 9.10 to 10.04 - works well - I also have a production system on 10.04 - works well
[20:26] <ttx> SpamapS: for the others I'll soon confirm who drafts them
[20:27] <RoyK> Theravadan: I'd still wait a few months before upgrading critical servers from 8.04, though
[20:29] <Theravadan> RoyK well I have no problems going from 8.04 and 9.10 so I'm migrating everything over to 10.04 right now
[20:29] <sh1ny> RoyK, well my only 8.04's are the zimbra servers i have and i'm stuck with them until they release zimbra for 10.04
[20:29] <Theravadan> I'm aggressive like that, usually it pays off
[20:29] <cybrocop> wizardslovak/Owner_: If you installed tightvnc and it can't connect to the server that means there is something wrong with either the network or the server. Did you install a VNC server on the other end and is it configured properly?
[20:29] <RoyK> sh1ny: indeed
[20:30] <sh1ny> RoyK, we adopted ubuntu instead of debian right after 8.10, so we went the 6 months releases till 10.04
[20:30] <RoyK> http://karlsbakk.net/top-16-2.png <-- nice box
[20:30] <SpamapS> ttx: ok, and for the work items.. they don't have any assigned value per-se right? So each one in the burndown is considered equal to another?
[20:30] <wizardslovak> well i installed it and opened ports in ufw and router
[20:30] <wizardslovak> still i am getting "cannot conenct to server" error
[20:30] <wizardslovak> althoutgh when i tried "/etc/init.d/vncserver restart it says it didnt find it
[20:31] <sh1ny> but upgrading every 6 months is getting cumbersome
[20:31] <RoyK> sh1ny: use LTS releases
[20:31] <sh1ny> RoyK, that's the plan, starting with 10.04 ;)
[20:31] <ttx> SpamapS: yes, they should be worth 0.5-2.5 days of work
[20:31] <RoyK> 8.04 is LTS
[20:31] <RoyK> sh1ny: 8.04 should probably work well for at least another three years
[20:31] <sh1ny> yes, but we decided we will go for 8.10 and 9.04 for many reasons
[20:32] <Theravadan> how do i convert a 10.04 system to LTS, edit a file and have the value be "lts" instead of prompt?
[20:32] <cybrocop> RoyK: CPU1 is at 0.3% system utilization. Need to do something about that?
[20:32] <RoyK> Theravadan: 10.04 _is_ LTS
[20:32] <sh1ny> Theravadan, it's just lts, no need to do anything
[20:32] <RoyK> cybrocop: hehe
[20:32] <Theravadan> sh1ny, so when 10.09 comes out it won't upgrade?
[20:32] <RoyK> cybrocop: load varies, but the box chews ash quite well
[20:32] <Theravadan> what if I wanted a machine to upgrade to 10.09?
[20:33] <RoyK> Theravadan: you can set the machine to lts or normal upgrades
[20:33] <Theravadan> RoyK got it... that is what I was getting at
[20:33] <SpamapS> Theravadan: 10.10 btw. ;)
[20:33] <Theravadan> ah 10.10
[20:34] <RoyK> Theravadan: see /etc/update-manager/release-upgrades
[20:34] <Theravadan> RoyK: thx
[20:34] <sh1ny> RoyK, when i got hired by my current employer, they had problems with updating debian servers, since we have more than 150 of those, and they wanted latest stuff with auto-updates of security
[20:34] <sh1ny> RoyK, so i suggested sticking to 6 monthies till 10.04 then sticking with that for 2 years
[20:35] <RoyK> sounds reasonalbe
[20:35] <RoyK> reasonable, even
[20:35] <sh1ny> aaand they wanted me to replace xen with something and kvm isnt that good in hardy
[20:35] <RoyK> is it good in lucid?
[20:36] <sh1ny> way better
[20:36] <RoyK> we have a piece of crap xenserver running
[20:36] <RoyK> would be nice to replace that with something proper
[20:36] <Theravadan> sh1ny, that is what i'm doing, in order to get some new packages i need for bucardo. But I may need even newer packages in the future thus it's nice that I can go off of LTS if I really want to.
[20:36] <sh1ny> tho i still have to see if it improves with guests that are acting as samba file servers
[20:36] <sh1ny> Theravadan, from what i learned
[20:36] <RoyK> imho stuff like file servers shouldn't be VMs
[20:37] <sh1ny> RoyK, the client is always right :(
[20:37] <RoyK> file servers are i/o intensive, not a VM thing
[20:37] <sh1ny> Theravadan, don't go mixing distros too much , well depending on the number of machines you support
[20:37] <RoyK> numbercrunchers are cpu and sometimes memory intensive, not a vm thing
[20:38] <RoyK> general stuff like your little webserver or something, that's a vm thing
[20:38] <sh1ny> RoyK, well we found out how to make it bearable so it "just works" for now
[20:38] <Theravadan> sh1ny, yeah that's a headache, in a month or so all systems will be ubuntu 10.04 and i will be happy
[20:39] <RoyK> Theravadan: already in a month?
[20:39] <sh1ny> i got like 80 ubuntu 9.10's and ~100 debian unstable ( previous sysadmin's work >.> )
[20:39] <RoyK> I'd wait a few more if I were you
[20:39] <RoyK> sh1ny: Sid??
[20:39] <sh1ny> RoyK, no, they were unstable 3 years ago
[20:39] <sh1ny> RoyK, i guess now they're something after lenny , not very sure
[20:39] <RoyK> Sid in production isn't good
[20:40] <sh1ny> RoyK, i don't bother upgrading them, because they pop open like can of worms
[20:40] <sh1ny> RoyK, i'm just slowly reinstalling them with ubuntu
[20:41] <sh1ny> RoyK, it was the situation when i arrived, and to be honest i am having a hard time pushing new technologies + converting all the old crap that i inherited
[20:41] <RoyK> Sid == unstable
[20:41] <sh1ny> yes i know it's bad for production, but it was like this when i arrived
[20:41] <sh1ny> so >.>
[20:42] <RoyK> we have a bunch of old fedoras around
[20:42] <RoyK> not good either
[20:42] <sh1ny> yea well i should be done with all of those in a year or so
[20:42] <SpamapS> ttx: could you provide me with some examples of some of the specs that were completed in the last cycle?
[20:42] <RoyK> the old man that installed them has finally accepted, or getting close to, installing ubntu lts instaed
[20:42] <sh1ny> i often get "If it's working don't touch it" from customers
[20:43] <sh1ny> and it's kinda hard to explain what is "security" and why it matters
[20:43]  * RoyK isn't a consultant anymore - just an employee
[20:43] <sh1ny>  :)
[20:43] <RoyK> easier that way, somehow
[20:43] <sh1ny> you lucky guy
[20:43] <sh1ny> ok, guys was nice having a chat, but i gotta be up in 6 hours
[20:43] <RoyK> nite
[20:43] <sh1ny> so good night and good luck with whatever you try to do tonight :)
[20:49] <ttx> SpamapS: sure
[20:49]  * ttx rummages
[20:50] <ttx> SpamapS: https://blueprints.launchpad.net/ubuntu/+spec/server-lucid-euca-remote-autoregister
[20:50] <ttx> https://blueprints.launchpad.net/ubuntu/+spec/server-lucid-aws-client-libraries
[20:51] <ttx> SpamapS: see http://people.canonical.com/~pitti/workitems/lucid/canonical-server-lucid-alpha-3.html
[20:52] <ttx> ^ and specs linked from there
[20:52] <SpamapS> nice burndown. :)
[20:56] <SpamapS> ttx: somebody really should create something for the ubuntu wiki that lets you say   [[lp:blueprint:xxx:whiteboard]]  so they stay in sync.
[20:57] <ttx> SpamapS: you should not duplicate the work items list in the wiki
[20:57] <ttx> SpamapS: just point to it
[20:58] <MTecknology> how can i figure out what's causing a site offline message?
[20:59] <SpamapS> ttx: roger that.
[20:59] <guntbert> MTecknology: from what system?
[20:59] <MTecknology> guntbert: ubuntu
[20:59] <guntbert> MTecknology: sorry - I meant: from what system do you get that message?
[21:00] <enav> apparmor is not easy to configure really?
[21:00] <MTecknology> guntbert: ? when i go to the website
[21:00] <MTecknology> The site is currently not available due to technical problems. Please try again later. Thank you for your understanding.
[21:01] <MTecknology> enav: it's incredibly simple
[21:01] <enav> good
[21:02] <guntbert> MTecknology: if that is the exact message you should have gotten it from the target host -- but some messages your browser generates look very similar
[21:02] <enav> but ubuntu guide is to short and only explain few things
[21:04] <MTecknology> guntbert: wrong channel...
[21:04] <MTecknology> guntbert: i meant to ask in #drupal :P
[21:05] <guntbert> MTecknology: I was already wondering about the sparse info you gave  ... :-)
[21:05] <MTecknology> guntbert: ya - in the right channe; it woulda made sense :P
[21:05] <guntbert> MTecknology: definitely :-)
[21:16] <Theravadan> what is the accepted way to get a custom script in /etc/init.d to start / stop?
[21:18] <Theravadan> update-rc.d tomcat defaults worked
[21:24] <EtienneG> smoser, I just noticed we have desktop AMI at http://uec-images.ubuntu.com.  How does that work?  Do they have ssh installed so we can connect?  NX, perhaps?
[21:25] <smoser> neatx server installed
[21:25] <smoser> ubuntu-desktop task installed
[21:25] <smoser> in maverick this will hopefully be moved to x2go or a more stable solution
[21:26] <EtienneG> smoser, sounds cool, will look into it
[21:26] <smoser> the neatx server and non-commercial nx clients provide less than desireable results.
[21:29] <smoser> EtienneG, http://groups.google.com/group/ec2ubuntu/browse_thread/thread/e57316d0a0af4b2/f963bac158169725?lnk=gst&q=lucid+nx+server#f963bac158169725
[21:29] <soren> Wow, this sucks. I come back to my laptop, see that I have firefox open, pointing to the page for filing bugs against vm-builder. The bug summary just says: "Multiple ".. I have no recollection of this and no clue what the bug might have been about.
[21:30] <smoser> http://www.stgraber.org/2009/12/12/ubuntu-desktop-cloud-now-working-lucid
[21:30] <bogeyd6> i never give apache2 enough credit. you know it sits there and redirects traffic from a domain to the cluster all day long. never breaking or complaining.
[21:34] <EtienneG> smoser, thanks, it's just for testing anyway, so I will not sweat it
[21:34] <EtienneG> cool stuff nonetheless
[21:34] <smoser> x2go should provide improvement
[21:34] <soren> Gawd, I hate preparing SRU's.
[21:36] <bogeyd6> oh lawd soren
[21:49] <therianlunch> group question! whats your favorite game available in the ubuntu software center?
[21:53] <smoser> http://www.vim.org/scripts/script.php?script_id=172
[22:03] <osmosis> how do I change which kernel ubuntu is using. I want to switch to  linux-virtual
[22:05] <celeborn999> the ur-quan masters and battle for wesnoth
[22:07] <celeborn999> nethack
[22:18] <binBASH> How to delete stored images in uec?
[22:21] <binBASH> because I installed an image out of the store and node controller gets a 403 and cannot get a file
[22:22] <erichammond> I've received a couple inquiries about an Ubuntu mirror in EC2 Asia Pacific.  Are there plans for ap-southeast-1.ec2.archive.ubuntu.com ?
[22:23] <erichammond> Is elmo still running these mirrors?
[22:26] <binBASH> smoser: Any idea how to fix this? http://www.pastie.org/966528
[22:28] <osmosis> im not able to switch over to linux-virtual.  see  http://dpaste.com/196281/
[22:37] <sbeattie> osmosis: the virtual kernel (on i386) is the same as the generic-pae kernel, just with a bunch of the modules removed.
[22:37] <sbeattie> (on amd64, virtual is the same kernel as the -server kernel)
[22:39] <ryoohki> i would like, from a fresh 10.04 server install, a copy of a u.s.a. /etc/apt/sources.list
[22:58] <billybigrigger> anyone tried burning an .iso over nfs?
[22:58] <billybigrigger> is it recommended?
[22:59] <alvin> billybigrigger: I have done that many times. Never had troubles. (buffer underrun is a thing from the past anyway)
[23:00] <billybigrigger> fair enough
[23:00] <billybigrigger> thanks
[23:02] <billybigrigger> alvin, using growisofs
[23:02] <billybigrigger> sob
[23:02] <billybigrigger> 279576576/7838695424 ( 3.6%) @3.7x, remaining 23:25 RBU  88.8% UBU  71.4%
[23:02] <billybigrigger> ^C/dev/dvdrw: flushing cache
[23:02] <billybigrigger> lol
[23:03] <billybigrigger> that was an expensive verb DL
[23:04] <celeborn999> i've managed to bork my apache2 install, i'm trying to do a 100% complete wipe of the thing and reinstall but i guess i'm not getting rid of all of it. any way to completely undo the install and redo it, like i never had it before?
[23:05] <celeborn999> in particular, after i install, it doesn't automatically start, and when i try to start it it complains about a missing /etc/apache2
[23:05] <billybigrigger> apt-get purge apache2
[23:06] <celeborn999> thanks but that doesn't help. i've tried apt-get purge apache2, apt-get --purge remove apache2, some trick with dpkg -i
[23:07] <celeborn999> when i do a locate for apache2 after i "uninstall" it, it still finds all kinds of stuff
[23:07] <celeborn999> i'm not sure about going through and manually deleting everything with apache2 in the name. that's what i did to /etc/apache2
[23:10] <jamesgao> hi everyone, I've been trying to set up an LDAP/Kerberos environment for my lab
[23:10] <celeborn999> figured out the apache thing: you have to get rid of apache2-utils, too
[23:11] <jamesgao> unfortunately, the server docs are either out of date, or don't work... I have kerberos set up, but LDAP + ssl breaks
[23:12] <jamesgao> I keep getting the "main: TLS init def ctx failed: -1" error, and I've triple-checked both the locations and commands that generated all the key files
 put in the cf today
 ubuntu 10.04lts server took 30 seconds to boot not including the bios time
 so if I ran coreboot it would take even less time
 mini-itxx 533
[23:40] <f1yback> not bad
[23:40] <f1yback> I can live with that startup time
[23:41]  * slackster misread can for can not. :P
[23:51] <Nonpython> When I try to start Bind I get "rndc: connect failed: 127.0.0.1#953: connection refused", what is wrong?
[23:52] <SpamapS> kirkland: you there?
[23:54] <JanC> f1yback: depending on what services you are starting during boot that's somewhere between pretty bad & pretty good  ;)
[23:54] <f1yback> stock install
[23:54] <f1yback> only function I selected during install was ssh
[23:54] <f1yback> it's a 1st gen itx
[23:54] <f1yback> c3-800 samual core
[23:54] <f1yback> samual2 sorry
[23:55] <JanC> well, my Celeron 900MHz-based netbook boots faster into a graphical UI...
[23:56] <bc> Nonpython: probably a file permission problem. Are you running named in a chroot environment? That could give you some hints.
[23:56] <f1yback> i'm not that worried about it
[23:57] <f1yback> basically the box runs nut-ups then fires off 30 ssh logins and execs a shutdown script
[23:57] <f1yback> that's it
[23:58] <Nonpython> bc: I am new to bind, how do I know?