[00:00] <RootChaos> but normally run into issues when those releases retire
[00:00] <RootChaos> so i'll wait for the lts release to come out
[00:00] <RootChaos> like with 7.x
[00:00] <RootChaos> etc.
[00:00] <sbeattie> um, err, 10.04 LTS ought to be out tomorrow.
[00:01] <soren> sbeattie: Today, really. It's Thursday already. You guys are just behind. :)
[00:01] <RootChaos> that's good news... just hope our bandwidth to SA holds up with the seacom cable maintenance... we've been having poor internet access from upstream providers running mostly on seacom
[00:02] <sbeattie> soren: based on sleep patterns I'm still on tuesday. So nyah!
[00:02] <RootChaos> perhaps our TENET mirror will have it sooner than later
[00:02] <RootChaos> anyways
[00:02] <RootChaos> thanks guys
[00:02] <soren> sbeattie: Based on time aged during this release cycle, it's 2014 for me, probably.
[00:02] <sbeattie> RootChaos: You could um start prefetching now
[00:03] <RootChaos> 10.x LTS RC ?
[00:05] <RootChaos> hhmm
[00:05] <RootChaos> would it be easy to upgrade from the RC to the release version ?
[00:07] <Insyte> RootChaos: Yes.
[00:07] <RootChaos> ok cool, i'm downloading :)
[00:08] <arrrghhh> hey all, havin issues with mpd.  used to run just fine, but i decided to wipe my server & start clean.  now it segfaults when it tries to play a song...
[00:12] <RootChaos> nice
[00:12] <RootChaos> :p
[00:12] <RootChaos> coming down at 40kb/sec
[00:13] <RootChaos> on a 1mbit line
[00:13] <RootChaos> guess it will be done in the morning
[00:13] <RootChaos> off to bed
[00:40] <mathiaz> ahasenack: hey
[00:40] <ahasenack> mathiaz: hey
[00:41] <mathiaz> ahasenack: I was wondering what you thought about the drop of cn=localroot,cn=config in slapd in lucid
[00:41] <mathiaz> ahasenack: and not using SASL mapping anymore
[00:41] <ahasenack> mathiaz: no strong opinion either way, because the client when using ldapi doesn't even see this
[00:41] <mathiaz> ahasenack: instead I just use directly the sasl name for root in the olcAccess rule
[00:42] <mathiaz> ahasenack: it seems to be simpler
[00:43] <ahasenack> mathiaz: I just hope they never change the way this sasl DN is constructed :)
[00:43] <mathiaz> ahasenack: :) - we'll another olcAccess rule then :)
[00:43] <ahasenack> yep
[00:48] <gop> hi can I install from the ubuntu server ubuntu on to a usb flash stick
[00:48] <gop> Just curious
[00:50] <mikelifeguard> How do you tell ubuntu what domain name it is? For example,
[00:51] <mikelifeguard> 'ping localhost' tells you what IP & domain it is pinging.
[01:00]  * mikelifeguard got a new domain name & wants to start using it
[02:40] <erichammond> smoser: How soon do you think you can release AMIs for the new Asia Pacific region (ap-southeast-1)?
[02:44] <erichammond> smoser: Not like you had any other important AMIs to release in the next day or so :-)
[03:43] <kzman> hello, what is the best ftp server for ubuntu?
[03:43] <twb> kzman: openssh-server.
[03:46] <ScottK> +1 for that from me.
[03:46] <twb> http://mywiki.wooledge.org/FtpMustDie
[03:48] <kzman> is the ftp and windows filesharing download speed alike?
[03:48] <twb> I don't understand the question.
[03:50] <kzman> do they have FTP and Windows filesharing the same download speed?
[03:50] <twb> I don't understand the question.
[03:51] <kzman> which is more efficient?
[03:51] <twb> What is "windows filesharing"?
[03:51] <kzman> when i share files in a windows network (samba on linux)
[03:52] <twb> You mean CIFS?
[03:53] <kzman> i think
[03:54] <twb> I don't know which would be faster.
[03:54] <twb> They're both shockingly bad protocols by modern standards.
[03:55] <twb> HTTP/1.1 with pipelining ought to beat both.
[04:23] <RoAkSoAx> kirkland, ping
[04:23] <kirkland> RoAkSoAx: howdy
[04:23] <kirkland> RoAkSoAx: we're trying to get your testdrive uds session scheduled
[04:23] <RoAkSoAx> kirkland, howdy :)
[04:23] <kirkland> RoAkSoAx: looks like it'll be on the desktop track
[04:24] <RoAkSoAx> kirkland, yeah that's what i was wondering
[04:25] <kirkland> RoAkSoAx: it might get renamed, or something
[04:25] <RoAkSoAx> kirkland, but which one are we gonna use: https://blueprints.edge.launchpad.net/testdrive/+spec/server-maverick-testdrive-frontend-gsoc or https://blueprints.edge.launchpad.net/testdrive-gtk/+spec/server-maverick-gsoc-testdrive-frontend
[04:25] <kirkland> but it should be an open forum for collecting ideas
[04:25] <RoAkSoAx> to pass the info from one to another
[04:25] <kirkland> RoAkSoAx: we/you won't have to implement all of them, for sure
[04:25] <ScottK> RoAkSoAx and kirkland: Not sure if you've already got it covered, but maybe a session on getting onther flavors integrated into testdrive.  I know there are Kubuntu people interested in that and probably other flavors too.
[04:25] <kirkland> RoAkSoAx: but it would be nice to drum up some community around this
[04:26] <kirkland> ScottK: agreed; that's part of what RoAkSoAx should be doing this summer as part of his GSoC project
[04:27] <RoAkSoAx> kirkland, of course. That's what i was thinking. That's why at first I said goals and first get a PyGTK and then the garage/car approach.
[04:27] <ScottK> Great.
[04:27] <kirkland> ScottK: rough story-board looks something like http://people.canonical.com/~kirkland/testdrive-wizard.html
[04:27] <kirkland> ScottK: where the "flavor" field would be Kubuntu/Xubuntu/Mythbuntu, etc.
[04:27] <RoAkSoAx> ScottK, you mean have a Testdrive PyQT interface?
[04:28] <ScottK> RoAkSoAx: That would be nice, but I was starting with being able to use different flavors in test drive.
[04:28] <RoAkSoAx> kirkland, and yes, one of the features is to be able to testdrive other flavors... that's something that Im gonna start working on right after having the modularization finished
[04:28] <kirkland> RoAkSoAx: so if you can get PyQT going, that's great, but PyGTK comes first;  I think ScottK just means getting Testdrive able to download and test other Ubuntu flavors
[04:28] <ScottK> In line with what kirkland just pointed out.
[04:28] <RoAkSoAx> ScottK,^^
[04:28] <ScottK> Yeah.
[04:28] <kirkland> ScottK: great, we're in sync
[04:28] <kirkland> RoAkSoAx: if your modularization goes well, bolting on a PyQT interface should be easy for someone else to contribute
[04:29] <RoAkSoAx> ScottK, btw.. can I bother you with python code review? in a few days for advice in the way how am implementing the OO?
[04:29] <RoAkSoAx> kirkland, yes that;s the idea... it should be possible now btw...
[04:29] <ScottK> kirkland and RoAkSoAx: I would suggest consulting with someont like rgreening_ who worked on the USB creator split to make sure you start with a design that will be easy to add a different front end to.
[04:30] <ScottK> He's been through it before and can give good advice on laying a foundation for expansion.
[04:30] <rgreening> yep
[04:30] <rgreening> :)
[04:30] <RoAkSoAx> ScottK, yeah indeed that;s why I decided proposed to modularize the code
[04:31] <RoAkSoAx> rgreening, ok then I;ll bother you for some code review/advice then in a few days
[04:31] <ScottK> rgreening: BTW, you pinged earlier.  I'm here now.
[04:31] <rgreening> sounds good RoAkSoAx
[04:31] <kirkland> rgreening: excellent, you'll be at UDS?
[04:31] <rgreening> ScottK: I figured it out.
[04:31] <rgreening> kirkland: yep
[04:32] <kirkland> rgreening: cool, would be nice if you could join us for a session on Testdrive then
[04:32] <rgreening> kirkland: I'll be bugging the server guys lots while im there :)
[04:32] <rgreening> kirkland: sure thing
[04:32] <rgreening> kirkland: you have a blueprint up for that?
[04:33] <kirkland> rgreening: ack, https://blueprints.edge.launchpad.net/testdrive-gtk/+spec/server-maverick-gsoc-testdrive-frontend
[04:33] <rgreening> Im getting into some real kvm server work now at the office. and building lots of vm's .. so Ill pick lots of server brains at uds :)
[04:33] <kirkland> hmm, link broken
[04:34] <RoAkSoAx> kirkland, so i shall migrate the info already posted in https://blueprints.edge.launchpad.net/testdrive/+spec/server-maverick-testdrive-frontend-gsoc
[04:34] <RoAkSoAx> to  https://blueprints.edge.launchpad.net/testdrive-gtk/+spec/server-maverick-gsoc-testdrive-frontend
[04:34] <kirkland> RoAkSoAx: oh ... hmm, sorry about that
[04:34] <ScottK> kirkland and RoAkSoAx: Small nit for your mockup: Kubuntu and Kubuntu Netbook Remix are two separate ISOs/subflavors.  They are just built from the same seed set (unlike Ubunut/UNE) and so may not have appeared on your list.
[04:34] <kirkland> RoAkSoAx: yeah, please converge on one
[04:35] <kirkland> ScottK: thanks ... this was an html napkin ;-)
[04:35] <ScottK> Sure.
[04:35] <RoAkSoAx> kirkland, you already proposed https://blueprints.edge.launchpad.net/testdrive/+spec/server-maverick-testdrive-frontend-gsoc so, I guess we should have everything in there instead of the one in https://blueprints.edge.launchpad.net/testdrive-gtk/+spec/server-maverick-gsoc-testdrive-frontend
[04:35] <kirkland> ScottK: i'm going to try and get IS and the cdimage guys to give us a good manifest
[04:35] <rgreening> I suggest drop the gtk
[04:36] <rgreening> to remain FE agnostic :)
[04:36] <kirkland> RoAkSoAx: yes, please just make sure there's just one blueprint
[04:36] <RoAkSoAx> rgreening, the gtk is part of the GSoC project
[04:36] <kirkland> RoAkSoAx: move all the pertinent info to one
[04:36] <kirkland> RoAkSoAx: subscribe me to it
[04:36] <rgreening> ah
[04:36] <kirkland> RoAkSoAx: mark the other one "superceded"
[04:36] <rgreening> okies
[04:36] <RoAkSoAx> rgreening, but the modularization will allow us to have the other front ends
[04:36] <RoAkSoAx> kirkland, will do
[04:36] <rgreening> cool
[04:36] <kirkland> rgreening: yeah, RoAkSoAx is tasked with putting a GTK frontend on it for Google Summer of Code
[04:37] <kirkland> rgreening: in doing so, he's modularizing it such that another FE can bolt on easily
[04:37] <kirkland> rgreening: we're interested in your advice/review of that
[04:37] <rgreening> well, I may be able to help getting a qt/kde up and running
[04:37] <rgreening> yep
[04:37] <kirkland> rgreening: and if you want to put together the qt FE in parallel, that would be great ;-)
[04:37] <RoAkSoAx> awesome then :)
[04:37] <rgreening> testdrive rules :)
[04:37] <rgreening> heh
[04:38] <kirkland> rgreening: ;-)  sweet, thanks.  i find it very useful
[04:38] <kirkland> rgreening: and it warms my heart to hear that it's useful to someone else
[04:38] <rgreening> me 2
[04:38] <rgreening> yeah. 1 line to launch a vm.. I run for coffee while it syncs :)
[04:39] <kirkland> rgreening: saved me many, many hours this release cycle, doing ISO testing
[04:39] <kirkland> was *way* worth the couple of hours it took to write
[04:39] <rgreening> kirkland: we need to figure out how to integrate checkbox into it to automate testing :)
[04:39] <kirkland> rgreening: and the secret agenda is that it gets people using and testing KVM without really knowing it ;-)
[04:40] <kirkland> rgreening: kvm-autotest is how ;-)
[04:40] <rgreening> heh. tru dat
[04:40] <twb> It's not a secret if you tell us
[04:40] <rgreening> shh...
[04:40] <kirkland> twb: you're just in the know, now
[04:40] <rgreening> :)
[04:42] <RoAkSoAx> kirkland, done in: https://blueprints.edge.launchpad.net/testdrive/+spec/desktop-maverick-testdrive-frontend-gsoc
[04:44] <rgreening> cool. subscribing now kirkland, RoAkSoAx. and now for some homemade ribs :)
[04:45] <RoAkSoAx> rgreening, :)
[04:46] <kirkland> rgreening: you're eating your kid's ribs?
[04:46] <kirkland> :-P
[04:46] <rgreening> heh.
[04:51] <kirkland> g'night all
[04:52] <rgreening> nn
[04:54] <RoAkSoAx> kirkland, nighty
[04:56] <twb> A new customer has sent me some notes on how to connect to their network.
[04:56] <twb> They start with "Download the Cisco VPN Client here Mac/XP/Vista <URL>".
[04:56] <rgreening> ha
[04:56] <twb> What would that be on Ubuntu, vpnc?
[04:57] <rgreening> one of the vpn sw supports the cisco client iirc
[04:57] <rgreening> not vpnc though I think
[04:57] <twb> Awesome, there's also openconnect, which has a Description that talks about a *different* Cisco VPN protocol
[04:58] <twb> rgreening: I did aptitude search ~dcisco~dvpn
[04:59] <jpds> twb: network-manager-vpnc ?
[04:59] <rgreening> twb: Cisco vpn has a few varieties...
[04:59] <twb> jpds: presumably that's a wrapper for people dumb enough to run NM
[04:59] <rgreening> did customer provide any settings/options required?
[05:00] <twb> rgreening: there's some auth details further down the page.
[05:00] <rgreening> vpnc may do it.. depending on requirements
[05:03] <twb> Oh, and they gave me a .pcf file, which is probably the equivalent of openvpn's cert files.
[05:03] <rgreening> possibly
[05:03] <twb> Haha, no, it's just an ini file
[05:03] <twb> I hate customers
[05:03] <rgreening> lol
[05:05] <twb> Grr, why doesn't thttpd Just Work
[05:06] <twb> Because some clown set ENABLED=no in it's /etc/default/thttpd file
[05:08] <twb> Hm, looks like the QA team did it (#142306).
[05:08] <arrrghhh> hey all.  i'm trying to run mpd and it segfaults whenever it starts to play a song...
[05:14] <twb> Play a different song, then?
[05:14] <arrrghhh> lol if it was only that simple.  it's not just one song
[05:15] <arrrghhh> i haven't been able to get it to play one song.  i've tried at least 20
[05:18] <twb> Are you running mpd in the foreground?
[05:18] <arrrghhh> not currently, but i've used the --verbose and what is it, --no-damon or something?  i believe it just ends.  i see the segfault in my /var/log/messages tho.
[05:19] <twb> Ideally you want to know what it does BEFORE the segfault
[05:20] <twb> Though if I were you I'd probably just say "bugger it, let's use xmms2"
[05:20] <arrrghhh> lol really?
[05:20] <arrrghhh> can i connect to xmms2 with many different clients?  and i'm assuming it's a console-based player?
[05:21] <twb> xmms2 is a modular mpd
[05:21] <arrrghhh> i really did love mpd on my old ubuntu-server.
[05:21] <arrrghhh> whatever it was up to, i guess it was 9.10 before i blasted it all away and installed 10.04 fresh.
[05:22] <arrrghhh> all i really want is a music server with a small footprint that just runs as a service... and i need to be able to control it from many different clients
[05:23] <twb> Shrug.  Your two choices there are mpd and xmms2, and I don't care enough to help you debug the former.
[05:24] <arrrghhh> i thought xmms2 was just really a self contained player
[05:27] <arrrghhh> hrm.  that's unfortunate.  i don't understand why it would segfault on libavcodec.
[05:38] <twb> OK, with vpnc I can convert the .pcf file into a vpnc .conf file, and run vpnc on it.
[05:38] <twb> It spits out this: ==> vpnc: hash comparison failed: (ISAKMP_N_AUTHENTICATION_FAILED)(24); check group password!
[05:45] <twb> Never mind, $customer is retarded.  It's workign now.
[08:05] <Eric^-> Ska jag installera FTP så är det sudo apt-get pure-ftpd elr hur?
[08:05] <Jeeves_> English? :)
[08:05] <Eric^-> oh shit xD
[08:06] <Eric^-> uhm, when installing pure-ftpd
[08:06] <Eric^-> It's suppose to be sudo apt-get pure-ftpd? or?
[08:06] <Eric^-> or is it apt-install?
[08:06] <Jeeves_> apt-get install
[08:06] <Eric^-> yeah :D
[08:07] <twb> Eric^-: http://mywiki.wooledge.org/FtpMustDie
[08:07] <Eric^-> twb: uhm, whata heck is that?
[08:09] <Eric^-> http://tuxtweaks.com/2009/10/install-lamp-on-ubuntu-9-10-karmic-koala/
[08:10] <Eric^-> When following this guide, how the heck does i test if apache works since i use ubuntu-server without accses to go with normal interference
[08:10] <twb> Eric^-: it means: "don't run pure-ftpd, or any other ftpd.  Use OpenSSH's SFTP component."
[08:11] <Eric^-> twb: heck,man im a noob. FTP is easy, i will use it! :D
[08:12] <twb> Eric^-: you should not do so for the reasons outlined in the above article.
[08:16] <artemglusankov> is anybody here?
[08:16] <twb> !anybody
[08:17] <twb> How do I make a vbox .vdi usable in qemu (or kvm)?
[08:18] <Eric^-> How do i create a php file with terminal only? ;p
[08:18] <Eric^-> and edit it?
[08:18] <artemglusankov> I'm a newby in ubuntu-server. Is anybody have workable config for dhcp3-server?
[08:20] <artemglusankov> touch something.php
[08:27] <twb> I'm a lazy so, so I just use dnsmasq for DHCP
[08:28] <twb> The ISC stuff is all enterprisey and verbose and confusing
[08:28] <twb> s/so/sod/
[08:30] <artemglusankov> How can I get an information about version of my Linux?
[08:30] <twb> artemglusankov: what kind of information?
[08:31] <ajmitch> 'lsb_release -a' should tell you the distro version, at least
[08:31] <artemglusankov> thanks I will try
[08:32] <artemglusankov> It woks!!! )))))
[08:34] <artemglusankov> I have found a program named Vuurmuur. Does anybody work with this one?
[08:40] <twb> artemglusankov: it appears to be in the repository.
[08:52] <twb> From Karmic onwards, there's no qemu, only qemu-kvm.
[08:52] <twb> What am I supposed to do if I can't use KVM, and installing the kvm package (which modprobes kvm without asking) causes the kernel to hang completely, due to it fighting with vmware?  (And management won't let me remove vmware.)
[08:53] <_ruben> file a bug i assume
[08:53] <twb> I was afraid of that :-/
[08:54] <twb> For qemu 0.11, I rolled Debian's package into a ppa backport (for hardy).
[08:54] <twb> Now I need 0.12, so looks like I'll be doing that again
[08:54] <_ruben> does blacklisting the kvm modules help? not sure if blacklisting works for modprobe invokations though
[08:54] <twb> I'm a bit scared to try that on a production machine
[08:54] <_ruben> that's what test setups are for ;)
[08:55] <twb> I can't afford the $300 for a second VT CPU
[09:07] <soren> twb: You can blacklist the kvm modules.
[09:07] <soren> twb: The upstart job should respect that.
[09:07] <twb> soren: hardy doesn't use upstart much :-)
[09:08] <soren> twb: Well, you could just comment out the call to modprobe.
[09:08] <twb> But any call to modprobe will honour blacklists in modprobe.d
[09:08] <soren> I forget if the init script passed -b to modprobe in Hardy.
[09:08] <_ruben> twb: but hardy isnt "karmic onwards", so now i'm confused :)
[09:08] <soren> It's been a while, after all :)
[09:08] <soren> twb: No.
[09:08] <soren> twb: Only if you pass -b.
[09:09] <twb> Sorry, by blacklist I mean "install foo /bin/true" or so, not "blacklist foo".
[09:09] <twb> _ruben: sorry, I'm stuck on hardy but I want (to backport) qemu 0.12.
[09:10] <soren> twb: Anyways, commenting out the call to modprobe is perfectly acceptable as well. init scripts are conffiles as well.
[09:10] <_ruben> twb: ah
[09:10] <twb> soren: yes, but less reliable if some OTHER part of the kvm package tries to modprobe.
[09:11] <twb> I'd at least have to use policy-rc.d to prevent it doing it at install time, before I could get to the init script to edit it.
[09:11] <twb> Not that it matters much...
[09:15] <soren> I didn't think just loading the modules caused problems. I thought it wasn't until you actually ran some virtual machines things started to go South.
[09:17] <twb> Oh, sorry, I just remembered
[09:17] <twb> It's not kvm + vmware, it's kvm + openvz that makes it hang
[09:17] <twb> Or possibly all three together
[09:18] <twb> I just like to blame vmware because it's proprietary
[09:23] <smoser> erichammond, stuff is on its way up to ap-southeast-1
[09:23] <erichammond> smoser: Sweet!
[09:23] <smoser> kernels are there (not public)
[09:23] <smoser> hardy and karmic are on their way
[09:23] <smoser> and once karmic is there i can try lucid
[09:23] <erichammond> I'm migrating my S3 based AMIs and will copy EBS boot once your kernels are public.
[09:23] <smoser> lucid requires karmic for ebs publish
[09:33] <erichammond> smoser: Do you have a link handy for your instructions for connecting to lucid desktop with nxclient?
[09:34] <erichammond> smoser: hm, found this: https://lists.ubuntu.com/archives/ubuntu-cloud/2010-March/000163.html
[09:34] <smoser> thats what i would dig up
[09:35] <smoser> the other is instructions on how to use x2go
[09:35] <smoser> around that same time frame
[09:38] <janislaw> hello
[09:39] <janislaw> I am an average linux user - mostly from user experience
[09:39] <janislaw> In a moment our team is buing a new server and I am going to set up a system
[09:40] <janislaw> it's amd64 with 64GB of RAM and would be serving basic services for about 20 users
[09:40] <janislaw> I am looking for a comprehensive guide to set up the server
[09:40] <janislaw> I wonder if you could provide some links
[09:44] <twb> I'd avoid speccing hardware to last more than three, or at most five, years.
[09:45] <twb> Consider what would, today, be a five-year-old machine.  It's probably PATA or SCSI, DDR1, and a Pentium 4 or so.  You would have to pay MORE for spare parts than to buy a whole new machine three times faster.
[09:46] <twb> Oops, ignore me completely.  I read "20 users" as "20 years"
[09:46] <janislaw> ;)
[09:46] <twb> Because 64GB is pretty overkill for most applications of a twenty-user system
[09:47] <janislaw> well, we are replacing an about-9-year-old machine, for which the original admin is long gone
[09:47] <janislaw> oh, for an average user
[09:47] <janislaw> we will be doing computations which require 4GB+ memory
[09:47] <twb> File sharing, printing, and mail would only need 1GB
[09:47] <twb> Ah, OK
[09:48] <janislaw> almost a cheap, home-blown HPC
[09:48] <twb> Nod
[09:48] <janislaw> we're considering a queuing system if some users were too agressive on the processor time
[09:50] <janislaw> I am looking for a guide to help with such stuff like user quotas, hooks on user creation, packages considered as a must-have
[09:50] <twb> Some of that is just ulimits
[09:51] <janislaw> ok
[09:52] <janislaw> among others, there would be: LAMP, ssh, ftp, svn, bzr and mathematic packages on FORTRAN and Matlab
[09:53] <twb> janislaw: /etc/security/limits.conf
[09:53] <janislaw> could you recommend http://ubuntuguide.org/wiki/
[09:53] <twb> Hmm?
[09:54] <janislaw> twb: /etc/security/limits.conf is exactly what I need for the start
[09:55] <twb> limits.conf is basically the default values; the sh "ulimit" builtin allows users to voluntarily restrict themselves further.
[09:55] <twb> For disk quotas, you want the "quota" package, which requires some extra setup.
[10:02] <janislaw> ok!
[11:30] <echosystm> anyone here had a play with the cloud stuff?
[11:31] <twb> !anyone
[11:32] <zorzar> echosystm: which cloud stuff?
[11:37] <echosystm> the ubuntu ECS thing
[11:37] <echosystm> sorru
[11:37] <echosystm> UEC
[11:37] <echosystm> ubuntu enterprise cloud
[11:37] <zorzar> echosystm: i'm running ubuntu on amazons ec2
[11:37] <echosystm> im interested to know how it works
[11:37] <echosystm> the website istn very detailed
[11:38] <echosystm> i see a lot of places offering "cloud services"
[11:38] <echosystm> but its really just a VPS with ram/storage that can be expanded and contracted
[11:38] <echosystm> to me, a cloud service is something like google app engine, not rackspace "cloud servers"
[11:39] <echosystm> so, my question, which camp does EC2/UEC belong in?
[11:39] <zorzar> "cloud servers"
[11:40] <zorzar> the big differenc between a "cloud server" and a normal VPS is that you have an api to spawn and terminate servers and it happens in minutes
[11:41] <erichammond> ...and you pay for it in small units like by the hour.
[11:42] <echosystm> thats crap :(
[11:42] <echosystm> its hardly really a cloud is it
[11:43] <echosystm> if someones going to offer me some cloud service, i expect scalability to be automagic
[11:43] <echosystm> like appengine
[11:43] <zorzar> echosystm: then perhaps you have to rethink your definition of a cloud service
[11:44] <erichammond> echosystm: There is no widely accepted definition of "cloud" so it is difficult to say what is and isn't "cloud".  EC2 = "Elastic Compute Cloud" and is one of the few services which self proclaimed "cloud experts" would never say is "not cloud".
[11:45] <echosystm> the whole idea of the word "cloud" is that you just throw your app out there and it somehow just works
[11:45] <echosystm> thats hardly what EC2/rackspace/ etc. provide
[11:45] <echosystm> lol
[11:45] <echosystm> but anyway
[11:45] <erichammond> echosystm: You're free to add your own definition of cloud to the mess :)
[11:45] <echosystm> i guess i got my answer, thanks guys:)
[11:46] <zorzar> echosystm: np ;)
[11:46] <zorzar> echosystm: ec2 does offer autoscaling
[11:46] <echosystm> yeah, but its like scaling one VPS, isnt it?
[11:47] <echosystm> as opposed to having your app automatically deployed to multiple servers
[11:47] <lil_cain> scalability is never going to be automagic.
[11:47] <echosystm> it is on app engine
[11:47] <lil_cain> It may be easier, but if you want a scalable system, you have to design for it.
[11:47] <zorzar> echosystm: you can have your app automatically deployed to ec2 instances
[11:48] <echosystm> app engine you just upload to one place, and google works out where best to put it
[11:48] <zorzar> echosystm: yeah but app engine has other big limitations
[11:48] <echosystm> true
[11:50] <zorzar> echosystm: just use the tool that fits best, but just because you can hardly put a nail in the wall with a screwdriver, the screwdriver istn't bad by design
[11:52] <\sh> application scalability has nothing to do with the underlaying server...app engine just works like a simple web host .. if they wouldn't have shared storage and shared session management for the apps it wouldn't work as expected...
[12:24] <echosystm> ok guys
[12:25] <echosystm> ubuntu server has 5 years
[12:25] <echosystm> but desktop only has 3
[12:25] <echosystm> (LTS, i mean)
[12:25] <echosystm> is it like a subset of the repo has 5 years, while the rest has 3?
[12:25] <echosystm> or what
[12:25] <echosystm> because ubuntu and ubuntu server obviously use the same repo
[12:25] <echosystm> i dont get it
[12:26] <lil_cain> Yeah, it's a subset of the repo is on a three year support cycle.
[12:29] <echosystm> that sucks :(
[12:29] <JanC> echosystm: why?
[12:29] <echosystm> would have been nice to have the whole repo on that support cycle
[12:30] <echosystm> sometimes on servers you need to install gnome and so on
[12:30] <JanC> it's not like those packages will be removed
[12:30] <echosystm> yeah but they wont get security updates
[12:30] <echosystm> cant have that
[12:31] <lil_cain> I have never found a decent reason to install gnome on a server.
[12:31] <twb> What about "to convince $boss the hardware is too slow"
[12:32] <twb> "Gee, boss, look, after I installed tomcat, we get OOMed every few minutes"
[12:32] <twb> "Maybe you should give me $30 to upgrade to 1GiB of RAM"
[12:36] <echosystm> small business lil_cain
[12:36] <echosystm> they want a GUI
[12:36] <lil_cain> use a web thingy
[12:37] <lil_cain> !ebox
[12:37] <twb> echosystm: rather, they THINK they want bug-for-bug compatibility with MS SBS
[13:04] <pmatulis> zul: good game last night ;)
[13:05] <zul> pmatulis: i didnt watch it....hockey season is over
[13:05] <pmatulis> zul: ha ha
[14:56] <Guest13577> Hello... how do i make a ftp server where its possible to give directory permissions to different users/groups?
[14:56] <Guest13577> like an ACL kinda thing
[14:57] <peterlh> i cant seem to find any ftp server that does that on linux?
[14:57] <Slidey> you're talking virtual users and permissions? pretty sure proftpd will do what you want
[14:58] <peterlh> yes virtual users
[14:59] <peterlh> iam using pureftpd-mysql now
[14:59] <peterlh> but that can handlt it
[14:59] <peterlh> handle
[15:00] <JanC> can't vsftpd do that (as that's the default FTP server for Ubuntu)?
[15:05] <volksman> trying to nfs export a directory with a bind mount within it...the export works and I can see everything in the directory but when I look in the folder that contains the bind mount it's empty.  Is it possible?  Why doesn't it work?  http://paste.ubuntu.com/424618/
[15:27] <mrmojo> hi there
[15:27] <mrmojo> i'm using mdadm to create a raid1 array over two hdds
[15:28] <mrmojo> it all works fine until i reboot, the array is not automatically reloaded
[15:28] <mrmojo> am i missing something obvious here?
[15:34] <Slidey> what does your mdadm.conf say ?
[15:35] <mrmojo> where is mdadm.conf normally stored?
[15:35] <Slidey> i dont have it on an ubuntu box, but debian has it at /etc/mdadm/mdadm.conf
[15:35] <Slidey> the man page on there shows how to generate a new one
[15:35] <mrmojo> ok, cheers, i think i may of overlooked this step
[15:35] <Slidey> mdadm --examine --scan --config=mdadm.conf
[15:35] <mrmojo> cheers
[15:36] <Slidey> np
[15:36] <mrmojo> i still can't seem to find mdadm.conf
[15:40] <linxeh> is there any way to easily install the Dell OpenManage / OMSA etc software on ubuntu systems?
[15:41] <akincer> So we're close to full release, yes?
[15:42] <akincer> I figured the ISOs would have landed by now
[15:42] <Jeeves_> akincer: Yes, close. But not there yet
[15:42] <Jeeves_> akincer: Some are still missing
[15:42] <akincer> that's cool. I'm patient, but I would like to get the full release before lunch so I can load up a server for a new project :)
[15:43] <ttx> mathiaz: o/
[15:43] <akincer> by the way, is alfresco in main?
[15:44] <mathiaz> bug 571057
[15:45] <akincer> also, any movement on getting 10.04 as a supported version for their commercial release?
[15:46] <Jeeves_> akincer: who is 'their'
[15:46] <leonel> JUST SHOUTING  BIG  THANK YOU  FOR  LUCID !!   keep the great work
[15:47] <akincer> Unless I missed it, the devs for Alfresco offers a community version and a commercial version
[15:47] <linxeh> akincer: why not just install the server and then upgrade ?
[15:47] <akincer> linxeh: I'm doing that if the full release doesn't hit when I plan on setting up the server. was just hoping it would
[15:48] <linxeh> I'm  doing that now - can easily upgrade the pakcages that have changed
[16:03] <PerfectSine> hello all!
[16:03] <jiboumans> hggdh: eucalyptus call now, please join?
[16:10] <uiuiui> hi, i am using ubuntu to share internet with many people, is there a way to list who is connected, inorder to later block off unwanted users?
[16:10] <uiuiui> it being used as a  hotspot...
[16:14] <aurigus> uiuiui: you should be able to get a list of mac addresses and ip addresses
[16:15] <aurigus> running an in place upgrade of 9.10 ubuntu server right now.
[16:17] <uiuiui> i can get a list from /var/lib/dhcp...  but that is too much redundant information
[16:17] <RoyK> wasn't ubuntu 10.04 supposed to be released today?
[16:18] <RoyK> uiuiui: the dhcp clients file can be easily parsed with perl or something :)
[16:18] <hggdh> jiboumans: I am in
[16:19] <jiboumans> hggdh: please keep track of your gcalendar; bugs were discussed in the first 20 miis
[16:19] <hggdh> jiboumans: will do
[16:22] <maek> I have a bunch of DD images of entire disks. I have 1 partition per disk thats an LVM slice. I have the slice setup to a loop device and positions by offset. is there a way to mount an lvm partition or make lvm tools look at the loop device for a viable source?
[16:23] <uiuiui> so, i'd have to do scripting on it... no other way out
[16:23] <RoyK> maek: seems you need zfs :)
[16:23] <uiuiui> do u know of any utility that will list + block clients using iptables?
[16:23] <maek> RoyK: dont we all want zfs ? ;)
[16:24] <RoyK> maek: but yes, you can use losetup to connect a file to a loopdev - see the -o option
[16:24] <RoyK> maek: I use opensolaris when I need zfs :þ
[16:24] <maek> RoyK: thats where I'm at. I have the loopdevice setup and the -o option to tell it where it start
[16:25] <maek> RoyK: file -s /dev/loop3 shows me /dev/loop3: LVM2 (Linux Logical Volume Manager) , UUID: 00rP4mVL27fTaEM5mXBY1AB3DF3TA2P
[16:25] <maek> know I'm wondering how do I start the recovery of that lvm?
[16:25] <RoyK> maek: vgimport?
[16:26] <maek> oh just like that. the lvm tools will know the treat the loop mounted "device" as a legit source. sorry i've never used lvm before.
[16:29] <RoAkSoAx> jiboumans, howdy! I was wondering if the cluster stack blueprint was going to be accepted for UDS-M or what;s the status of its review?
[16:30] <jiboumans> RoAkSoAx: we'll do the review next week basically. cluster-stack will most likely get a UDS session
[16:30] <RoAkSoAx> jiboumans, awesome then! Thanks for the info
[16:33] <RoAkSoAx> kirkland, howdy!! Hey I was wondering now that the blueprint for Testdrive Front-end is under desktop, shouldn't we change the reviewer to  Rick Spencer?
[16:45] <kirkland> RoAkSoAx: sure, go ahead and do that
[16:46] <RoAkSoAx> kirkland, will do then :)
[17:00] <RoyK> any idea how far 10.04 release is?
[17:02] <PerfectSine> you can jump into #ubuntu-release-party
[17:02] <PerfectSine> those guys are going nuts :)
[17:03] <RoyK> lol
[17:06] <kirkland> hggdh: howdy
[17:09] <kirkland> hggdh: the eucalyptus sru package made it into the lucid-proposed ... could you re-test the Java OOM with that package?  (should be almost identical to the PPA one you tested)
[17:09] <kirkland> hggdh: fire it up and run a few thousand instances and confirm that you don't get the OOM errors
[17:14] <hggdh> kirkland: no prob, will start on it
[17:14] <kirkland> hggdh: awesome, thanks a lot
[17:15] <kirkland> hggdh: this is how we move an upload from -proposed to -updates ...
[17:15] <uiuiui> does dhcp3-server check the subnet mask of a client before sending a DHCPACK/DHCPNAK ?
[17:15] <kirkland> hggdh: oh, before you upgrade to the -proposed package, run it once with the current package, and reproduce the problem
[17:15] <kirkland> hggdh: then upgrade the package (and restart the service or reboot the machine, whatever)
[17:15] <kirkland> hggdh: and then confirm that the OOM problem is solved
[17:20] <aurigus> RoyK: I'm doing an in place upgrade now, seems to be working fine
[17:22] <aurigus> am i crazy or something? It seems like it is out but everyone is asking when its going to be released?
[17:22] <hggdh> kirkland: will do :-)
[17:22] <uiuiui> i cant seem to get my subnet to be fixed automatically, it behaves as if it only checks the ip addr
[17:24] <uiuiui> if a user connects to network1 w subnet 255.255.255.0, gets ip 10.255.255.4, then connects to network 2 which has subnet 255.255.0.0, the second network will allow the ip 10.255.255.4, but then if the user connects back to network1, the ip gets accepted and subnet remains at 255.255.0.0, not allowing it to connect to the gateway
[17:25]  * RoyK whines a little more about 10.04
[17:26] <cloakable> heh
[17:26] <aurigus> nm, i see these are all RC... im a fool :D
[17:27] <RoyK> aurigus: a few hours won't make much difference
[17:27] <RoyK> or even a few weeks
[17:28] <RoyK> works for me (tm)
[17:32] <brianherman> anybody know how to get ubuntu and comcast to work together like I have this box im not using and i would like to make it a router
[17:41] <RoyK> brianherman: isn't it just cable?
[17:41] <RoyK> DOCSIS?
[17:42] <brianherman> I tried ubuntu and I couldn't get an ip address when I plugged in directly
[17:42] <brianherman> Do i need to reset the cable modem and take out the two batteries?
[17:43] <RoyK> I guess they might be holding the IP for the old MAC address
[17:44] <RoyK> brianherman: try forcing the old mac address in /etc/network/interfaces - add hwaddress ether 00:01:04:1b:2C:1F or something at the end of the eth0 section
[17:44] <brianherman> i think i tried that
[17:44] <brianherman> i guess i did it wrong
[17:45] <RoyK> or try ifconfig eth0 hw ether x:x:x:x:x:x
[17:45] <RoyK> or just call them and tell them to release the address from the dhcp scope :)
[17:50] <brianherman> oh
[17:51] <brianherman> ok thanks bye
[19:00] <electro_> I am trying to register the Walrus service for UEC and I am getting a generic error on the web portal that says "Failed to save! (Check hostname and path.)"  Where does it write to?  I checked perms on /var/lib/eucalyptus/bukkits and it looks good
[19:13] <RoyK> 10.04 out :)
[19:14] <smoser> woohoo
[19:14] <smoser> just fixed http://uec-images.ubuntu.com/releases/10.04/release-20100427.1/
[19:14] <smoser> so output doesn't state that there is armel arch on ec2 :)
[19:19] <smoser> erichammond, lucid, hardy, karmic all on ap-southeast-1
[19:19] <smoser> (lucid not recieved much testing there)
[19:19] <erichammond> smoser: Thanks.  I've added the AMI ids to the top of http://alestic.com
[19:20] <erichammond> smoser: I migrated all of the Alestic S3 based AMIs to ap-southeast-1 (also listed)
[19:20] <erichammond> smoser: Remaining todo: Register EBS boot AMIs in ap-southeast-1 for Karmic and Hardy using your images.
[19:22] <erichammond> smoser: I was using http://uec-images.ubuntu.com/query/lucid/server/released.current.txt which is very convenient.
[19:22] <smoser> yep
[19:22] <smoser> i'm kinda stoked. 2 releases in one day :)
[19:22] <erichammond> :)
[19:24] <jiboumans> As you probably heard, 10.04 is out now: http://www.ubuntu.com. A big thanks to all of you for making this an awesome release.
[19:24] <jiboumans> Don't forget to go see your loco team tonight for the Release party :)
[19:25] <npope> jiboumans++
[19:26] <npope> ubuntu++
[19:28] <maek> anyone know how to tell what raid setup or type of lvm was created ?
[19:28] <npope> maek: lvdisplay for LVM
[19:29] <maek> npope: does that make any sense to you ? http://gist.github.com/384008 if you dont mine. thanks
[19:29] <maek> I had a readynas and something in it (drives I think) failed
[19:30] <maek> I have 4 disks, 2 of which I got DD images of, 1 has not partition table and the 4th makes a grinding noise when plugged in. trying to determine is the 3 disks were raid 5 or just concated.
[19:33] <incorrect> in libvirt is there a way to stop more vm's being started than the host can handle?
[19:40] <ChmEarl> what kernel version is final on 10.04 Server?
[19:52] <incorrect> ChmEarl, .32 iirc
[19:53] <incorrect> i will upgrade in a week i think
[19:55] <ChmEarl> incorrect, thanks
[20:03] <aubre> is there a nice command line way to check for the fastest upgrade server ?
[20:03] <jpds> aubre: No, but use a local one.
[20:03] <persia> There are a couple, but they are all very slow (because they try to download from lots of places).  Manually selecting a nearby mirror is almost always better than the tools.
[20:04] <aubre> ok , coworker was asking
[20:06] <npope> if your near wisconsin use ubuntu.mirrors.tds.net
[20:06] <RoyK> I had perfectly good download speed until five minutes ago
[20:06] <RoyK> I guess this thing can be left overnight, though :)
[20:08] <jpds> aubre: Where are based geographically?
[20:09] <aubre> Auburn, AL - I usually use gatech.edu
[20:09] <aubre> since ATL is only 90 mins from here
[20:09] <aubre> and we both have Internet 2 pipes
[20:17] <kirkland> dendro-afk: soren: hey guys, doubt you're terribly interested, but thought I'd mention ...  I released a new version of byobu last night with a new feature from jbernard, "rcs_cost", a little status applet that estimates cost of the current session in Rackspace
[20:32] <akincer> were there any changes to the final server packages? Update/upgrade on an RC server is showing no updates available
[20:38] <RoyK> akincer: same here, but I upgraded that a few hours ago
[20:39] <incorrect> now that sun java is in the partner repo, will it be updated when new jdk's are out?
[20:53] <soren> kirkland: You can usually deduce the instance creation time by checking the timestamp of /etc/crontab, and of network traffic, perhaps you could add a shutdown script that saves the traffic stats and count that as well?
[20:53] <soren> s/of network traffic/for network traffic/
[20:53] <soren> kirkland: The "since last reboot" notion doesn't seem very useful to me?
[20:54] <soren> I don't see when one really cares about that (except when you think you don't have a way to find the totals).
[20:55] <kirkland> soren: fair enough, thanks for the info....
[20:55] <kirkland> jbernard: can you take a look at that ^
[20:55] <kirkland> jbernard: and edit the rcs_cost script accordingly?
[20:56] <RoyK> I had perfectly good download speed until five minutes ago, but now it's dropped rather low
[20:56] <kirkland> soren: you're not still in Texas, are you?
[20:56] <kirkland> soren: seems a crew from Rackspace is heading up to Austin for the Lucid release party tomorrow evening
[21:00] <soren> kirkland: ah, cool. No, I'm not.
[21:01] <soren> kirkland: I'm on the proper side of the ash cloud^W^Wbig pond now :)
[21:02] <kirkland> soren: heh
[21:04] <soren> I don't remember, actually, if the billing cycle follows the month, but if it does, being able to determine the cost in the current billing cycle would probably be a nice addition.
[21:08] <soren> It doesn't. :(
[21:12] <Italian_Plumber> can someone recommend some type of live CD I can use to completely wipe a hard ddrive?
[21:12] <cloakable> Any livecd with dd on it
[21:13] <Jeeves_> cloakable: No
[21:13] <Jeeves_> Italian_Plumber: dban
[21:13] <Jeeves_> http://www.dban.org/
[21:13] <cloakable> Jeeves_: dd if=/dev/zers of=/dev/
[21:13] <Jeeves_> cloakable: That's not really wiping
[21:13] <cloakable> Jeeves_: Oh god, I'm typoing like crazy today
[21:13] <Jeeves_> If you really want to wipe, that's not good enough
[21:13] <cloakable> Jeeves_: it is if you do it three times :)
[21:14] <Jeeves_> Not even than, afaik
[21:14] <cloakable> What do you think dban does?
[21:16] <cloakable> http://blog.taragana.com/index.php/archive/how-to-rapidly-clean-wipe-hard-disk-drive/
[21:20] <Italian_Plumber> Jeeves, what is good enough?
[21:20] <Jeeves_> cloakable: overwriting with random data, several time
[21:21] <cloakable> Jeeves_: so dd if=/dev/ransom of=/dev/disk ?
[21:21]  * persia notes that one better have a HW RNG if doing that for reasonable sized disks.
[21:22]  * soren mumbles something about randomsound
[21:22] <cloakable> hell, just spamming /dev/zero over it once should work to overcome basic file utility programs.
[21:23] <soren> Also, why do I suddenly think about http://xkcd.com/538/ ?
[21:23] <cloakable> soren: pretty much :P
[21:24] <soren> If you really care that much about keeping your data safe, torch the bloody thing.
[21:25] <persia> soren: Ooh, nifty.  I've been using trousers.  Mind you, one has to actually generate audio for that, presumably by a random-skip algorithm.
[21:26] <soren> persia: Just plug in microphone. The background noise in any environment should provide plenty of entropy.
[21:26] <persia> soren: Ah, that's a good source indeed :)
[21:27] <Italian_Plumber> hmm... so I'm not srue what my answer is
[21:27]  * persia installs randomsound on the machine where the manufacturer *removed* the TPM chip in the newer models
[21:28] <persia> Ooh, shred does work on partitions (or entire disks).
[21:29] <Italian_Plumber> this page http://blog.taragana.com/index.php/archive/how-to-rapidly-clean-wipe-hard-disk-drive/ gives the same command for overwriting with "zeros" and "Random"
[21:30] <persia> Italian_Plumber: That's a typo on the page.  The second call should have been written with /dev/zero
[21:31] <Italian_Plumber> well I do have a drill press.  I can always drill holes in the bloddy thing
[21:31] <MTecknology> I can't figure out why logcheck won't send me an email..
[21:32] <MTecknology> 2010-04-29 15:29:16 1O7aMB-0006q8-RW => admin@kalliki.com <logcheck@emplar.kalliki.com> R=dnslookup T=remote_smtp H=aspmx.l.google.com [74.125.93.27]
[21:32] <Italian_Plumber> This says that shred doesn't work well on ext3, which is what my drive in question is:  http://linux.about.com/library/cmd/blcmdl1_shred.htm
[21:32] <MTecknology> I get the feeling gmail is screwing something up
[21:33] <npope> Italian_Plumber: whats wrong with dban?
[21:34] <RoAkSoAx> kirkland, shall we set *meeting* times when I can bother you with testdrive questions?
[21:34] <persia> Italian_Plumber: Read the *whole* manpage.  For certain sorts of journals, shred does pooly for individual files for ext3.  If you shred /dev/sda it'll do fine.
[21:35] <kirkland> RoAkSoAx: sure ;-)
[21:35] <RoAkSoAx> kirkland, what would be better time for you?
[21:35] <kirkland> RoAkSoAx: though i'm generally open here too
[21:35] <persia> (because then it's shredding the block device, not a file on the filesystem)
[21:35]  * soren doesn't really understand this whole problem
[21:35] <kirkland> RoAkSoAx: you're in Miami, right?
[21:35] <RoAkSoAx> kirkland, yep
[21:35] <soren> If your data is secret enough that you can be bothered to go through this hassle, isn't it really so secret that you want to just properly /destroy/ the disk?
[21:36] <kirkland> RoAkSoAx: you want a phone call, or irc, or what?
[21:36] <RoAkSoAx> kirkland, what we can do is that once per week we can *meet* to provide updates/status/questions... and whenever I have a doubt I can just ask you like now
[21:36] <RoAkSoAx> kirkland, whatever works for you
[21:36] <kirkland> RoAkSoAx: sure
[21:36] <kirkland> RoAkSoAx: want to do a phone call around lunchtime on Wednesdays?
[21:36] <kirkland> RoAkSoAx: middle of the week
[21:36] <RoAkSoAx> kirkland, sure
[21:36] <RoAkSoAx> kirkland, skype?
[21:37]  * persia agrees with soren, and recommends destruction.  It's fun, irreversible, and meets all known regulatory requirements.
[21:37] <kirkland> RoAkSoAx: or phone; do you have a US number?
[21:37] <npope> persia: dban then napalm?
[21:37] <npope> :)
[21:38] <persia> npope: last time I had a need to destroy disks, we used a special device that first soaked them in a strong magnetic field, then electrocuted them, then smashed them onto several spikes, then crushed them into small cubes, and electrocuted them again.  The results made handy paperweights.
[21:38] <ajmitch> that sounds like fun
[21:39] <persia> It was fun, although they only let me press the button once.
[21:39] <npope> that sounds like a blast
[21:39] <persia> It was about 2 quare meters of dedicated disk destructing machine.
[21:39] <persia> Err, 2 *cubic* meters.
[21:41] <soren> Every year at CeBIT there's a company (or more) that bring their disk annihilation machine. You can bring you disk and get it destroyed in a number of ways. The one I saw in use had a strong magnetic field and three big spikes that it poked through the disk. If people can get data from it afterwards, they deserve it, IMO.
[21:46] <Italian_Plumber> do you have to have some special status on launchpad to file a bug?  I don't se ehow to do it
[21:46] <MTecknology> Anyone know where to request support for google apps?
[21:47] <RoyK> asking google?
[21:47] <Insyte> I destroy my disks the old fashioned way:  Thermite.
[21:48] <MTecknology> RoyK: ya.. I just found gmail forums - might be what I need :) - I can't figure out why logcheck email isn't coming to me - exim logs seem to say that things are being sent just fine
[21:48] <soren> Insyte: That's probably overkill. A few hours at 280 degrees celcius in a regular oven should do it.
[21:49] <soren> Insyte: (A few hours to make sure it's hot all the way through.
[21:50] <RoyK> Insyte: even overwriting them once with zeros will kill stuff
[21:50] <RoyK> and you can use them for porn later
[21:50] <npope> RoyK++
[21:50] <Insyte> Yes, but thermite is much more fun.
[21:50] <lifeless> why bother with the zeros
[21:50] <lifeless> shove em in a ceph array, and fill it with porn straight up
[21:50] <RoyK> /dev/urandom, then
[21:51] <lifeless> Italian_Plumber: you have to be logged in, thats all.
[21:51] <RoyK> /dev/porn
[21:51] <Insyte> I read an article a while back, I forget where, that basically said all of this "multiple overwrites" nonsense was pointless.
[21:51] <RoyK> Insyte: today it is
[21:51] <Insyte> No one was able to demonstrate that they could recover data after a single pass of overwrites.
[21:51] <RoyK> Insyte: today's drives are so tightly packed, a single overwrite destroys the data
[21:52] <kopfweh> is there a boot spash on lucid server or is it just me?
[21:52] <RoyK> Insyte: in the eighties and early ninetees, they could find the old data by using electron microscopes
[21:52] <Insyte> But PCI mandates that we render retired disks unreadable.  I choose to interpret that as meaning: Melt them into a puddle with 4000F thermite.
[21:53] <amereservant> I'm curious about better understanding how cloud servers work.... Say I have two computers in a private cloud and I'm wanting to run Apache Server (overkill, I know) and host websites on them, what role does each computer in the cloud serve?  Do they both simultaneously run Apache and hold the data?
[21:53] <RoyK> Insyte: how boring - just dd a bunch if shit over the partition table and they won't find out
[21:53] <Insyte> Boring?!?
[21:53] <Insyte> It's a blast!
[21:53] <RoyK> wasting good drives
[21:54] <RoyK> good for porn
[21:54] <Insyte> That's very old school.  Isn't the new approach to just store your porn "In The Cloud(TM)"?
[21:56] <lifeless> amereservant: there are many deployments you could do
[21:56] <amereservant> lifeless: I gotcha.  The main thing I'm trying to understand is the role of each computer in the cloud.
[21:57] <lifeless> amereservant: the simplest would be two differently configured apaches running different sites, getting the data off elastic mounted volumes, or a cluster fs
[21:57] <lifeless> amereservant: a more 'cloudy' approach would be a load balancer, and as many apache instances as you want, all dynamically configured and reading data from elastic / cluster fs
[21:58] <amereservant> lifeless: The idea of Cloud Servers/Computing is the computers "share" resources, right?
[21:59] <lifeless> amereservant: no
[21:59] <lifeless> amereservant: the idea is simply that everything is a virtual machine, so you can stop and start and move it around really easily.
[21:59] <Insyte> amereservant: Cloud servers are just servers.
[22:00] <Insyte> amereservant: You just don't have to manage the infrastructure.
[22:00] <Insyte> Nothing magickal.
[22:00] <lifeless> amereservant: one consequence of that is that when you need only a fraction of a machine, you can put several vms on one machine [that the sharing], but its not usually a key element.
[22:00] <persia> Typically they are short-lived virtual servers with scripted configuration, rather than normal servers, but that's not that different.
[22:01] <amereservant> So then it's like a group of VPSs then?
[22:01] <Insyte> amereservant: Yes.
[22:01] <amereservant> Ahhhh, gotcha.
[22:01] <Insyte> With an API for setting them up / tearing them down.
[22:01] <amereservant> Yeah, I was seeing them all wrong.
[22:02] <amereservant> I was thinking it meant the servers were linked together and collaboratively working to serve content....
[22:03] <Insyte> They can be, but it's not automatic.
[22:03] <Insyte> Just like normal servers can be all linked together and collaboratively serving content.
[22:04] <amereservant> Yeah, but that's complex I'm sure.  A lot of factors and variables in that cookie jar.
[22:04] <Insyte> "The Cloud" is a BS term that's used to mean all sorts of things.
[22:04] <amereservant> I didn't realize it was also the same as using a machine to run multiple VPSs.
[22:05] <Insyte> The more specific terms are SaaS and IaaS.
[22:05] <Insyte> It's not *exactly* the same... there's generally a pretty significant software layer on top of that to allow for automatic provisioning / configuring / deprovisioning.
[22:05] <Insyte> So a smart IaaS rollout will expand as needed during periods of heavy use and then contract as things ramp back down.
[22:06] <Insyte> (To pick a popular example.)
[22:07] <amereservant> Insyte: So then what role does something like OpenVZ play?
[22:07] <amereservant> Because I was about to look into that and do some experimenting with it.
[22:07] <Insyte> That's one of the many technologies that could be used underneath all of the fancy APIs.
[22:08] <Insyte> Very approximately:  OpenVZ == Xen == KVM == VMWare
[22:08] <electro_> I am trying to register the Walrus service for UEC and I am getting a generic error on the web portal that says "Failed to save! (Check hostname and path.)"  Where does it write to?  I checked perms on /var/lib/eucalyptus/bukkits and it looks good
[22:08] <Insyte> *Very* approximately.
[22:08] <amereservant> Right, I've seen the VPS hosting where they offered both OpenVZ or Xen.
[22:09] <Insyte> Yeah, that's the software that lets you run virtual machines.
[22:09] <Insyte> I assume you're using Ubuntu... If so, KVM is the officially supported virtualization tech.
[22:09] <Insyte> Which means it's well tested and works out of the box.
[22:10] <Insyte> And Eucalyptus can sit on top of it as a "cloud" API.
[22:10] <amereservant> So is that something different/separate from the Cloud part or used "Instead of"?
[22:10] <Insyte> "As part of."
[22:10] <electro_> you guys configuring UEC?
[22:10] <amereservant> Ahhh, so it's a two-part ordeal then.
[22:10] <Insyte> To provide cloud-like services, you need a virtualization layer.
[22:10] <Insyte> That can be OpenVZ, KVM, Xen, etc...
[22:11] <amereservant> electro_: Nah, I was just trying to better understand how UEC works.
[22:11] <Insyte> Heh... more than two, depending on how complex you want to get.
[22:11] <electro_> amereservant: you and me both
[22:11] <electro_> Insyte: have you installed UEC?
[22:11] <Insyte> Nope.
[22:11] <amereservant> LOL.
[22:11] <amereservant> He definitely understands it better than I did.
[22:12] <electro_> I have one right now with a front end running CLC, SC, CC, Walrus, and 4 NCs
[22:12] <electro_> but i cant seem to register the walrus service
[22:12] <electro_> I can add it at the cli with the euca_conf
[22:12] <electro_> but not from the web portal
[22:12] <Insyte> amereservant: Walk through the KVM and UEC HOWTOs.  You'll understand it a lot better after you've played with it for an afternoon.
[22:13] <amereservant> Insyte: Thank you very much!  You're explaining it will definitely help me comprehend what it's telling me and not just leave me confused.
[22:14] <morrowyn> how do i upgrade from 10.04 beta to release ?
[22:14] <amereservant> morrowyn: Just run your updates.
[22:15] <morrowyn> apt-get update && apt-get dist-upgrade should do the trick right?
[22:16] <amereservant> I'm not 100% certain, but I'm pretty sure it's not a distribution upgrade.
[22:16] <amereservant> Should just need apt-get update
[22:18] <mikelifeguard> In /etc/apache/sites-enabled, the filename is 000-default so it is loaded first right? So if I want to put new VirtualHosts in there, I should name the files in order. 000 for the catch-all; 001 for the default, >002 for the rest and so on, yes?
[22:19] <RoyK> mikelifeguard: NameVirtualHost should sort that out
[22:19] <RoyK> or ServerName
[22:19] <RoyK> just put new files in there
[22:20] <mikelifeguard> Well, it does matter what order they're in - Apache uses the first matching ServerName, AFAIK
[22:20] <amereservant> mikelifeguard: Yeah, that's just the default in the event no other ServerNames are matched.
[22:20] <mikelifeguard> ok, right
[22:20] <amereservant> mikelifeguard: It will only match the first one if none of the other VirtualHosts match.
[22:21] <mikelifeguard> yes, exactly :)
[22:21] <mikelifeguard> \o/
[22:21] <mikelifeguard> I was right! I was right!
[22:21] <amereservant> The order is irrelevant.
[22:21]  * mikelifeguard writes home to mom
[22:21] <RoAkSoAx> kirkland, are you able to sync any image from testdrive?
[22:21] <RoAkSoAx> right now?
[22:21] <kirkland> RoAkSoAx: what do you mean "any image" ?
[22:22] <RoAkSoAx> kirkland, like: run testdrive, select any image to run, and see if its syncing or not
[22:22] <kirkland> RoAkSoAx: now == lucid's package?
[22:23] <kirkland> RoAkSoAx: http://people.canonical.com/~kirkland/Screenshot-Terminal.png
[22:23] <RoAkSoAx> kirkland, yes. It doesn't seem to be testdrive issue, but it get's stuck here with no further message: http://pastebin.ubuntu.com/424849/
[22:24] <hggdh> RoAkSoAx: just found an issue with testdrive (and, perhaps, the way images are put available): I downloaded Xubuntu desktop, and then run td -- and td clobbered the xubuntu with standard Ubunut
[22:25] <kirkland> RoAkSoAx: new code, old cod?
[22:25] <kirkland> hggdh: yep, it does that because cdimage calls the images the same thing
[22:25] <RoAkSoAx> kirkland, new code... it was working few mins ago... now it doesn't so that's why Im pretty sure its not testdrive issue, it is rsyncing :S
[22:25] <hggdh> kirkland: indeed. We should try to find a way out of it
[22:25] <kirkland> hggdh: i've been meaning to talk to cjwatson or IS to see if there's something that could/should be done on cdimage's end, or if we need to hack it into testdrive
[22:25] <kirkland> hggdh: on target for maverick
[22:26] <hggdh> kirkland: I ran the OOM, got them, and am now running the -proposed
[22:26] <RoAkSoAx> kirkland, i was planning to add support to list different flavors and syncing them from cdimage
[22:27] <hggdh> RoAkSoAx: makes sense, the the basic problem is different images have the same name. If cjwatson buys it (which I hope he does) this can be fixed at the source
[22:28] <hggdh> otherwise we have to hack it in (directory structure, probably)
[22:28] <kirkland> hggdh: right, i'm hoping it's an easy, non-controversial change on the cdimage server end
[22:29] <kirkland> hggdh: if not, we can work our way around it
[22:29] <RoAkSoAx> yep, i was just planing on hack the way around it
[22:30] <kirkland> RoAkSoAx: so when's your next code drop?
[22:30] <kirkland> RoAkSoAx: i'd like to see bits at a time;  makes it easier to review
[22:30] <kirkland> RoAkSoAx: try to keep it functional, though, all along the way
[22:30] <RoAkSoAx> kirkland, anyways, my issue is fixed. Seems it was issue with cdimage.
[22:31] <kirkland> k
[22:31] <RoAkSoAx> kirkland, I'm just cleaning up a little bit, doing some improvements to the modularization so that you can review it
[22:31] <kirkland> RoAkSoAx: great
[22:31] <kirkland> RoAkSoAx: we should make a list of "acceptance tests", the things we make sure continue to work, from release to release
[22:32] <kirkland> RoAkSoAx: have you tried it with virtualbox yet, for instance?
[22:32] <RoAkSoAx> kirkland, not yet... I'm just doing everything with KVM for now. since I wanna first test all the current features with KVM and then do the same with virtualbox
[22:34] <RoAkSoAx> kirkland, would a netbook run virtualbox? It should right?
[22:35] <erichammond> smoser: I'm trying to build the EBS boot Karmic/Hardy AMIs and think I'm running into the Lucid download craze today.
[22:36] <erichammond> smoser: Image files are taking forever to download.  E.g., http://uec-images.ubuntu.com/releases/karmic/release/unpacked/ubuntu-9.10-server-uec-i386.img.tar.gz
[22:36] <erichammond> smoser: Any chance you have a copy somewhere else I could use?
[22:38] <hggdh> RoAkSoAx: if it runs KVM, it will run VBox.
[22:38] <RoAkSoAx> hggdh, netbooks can't run KVM AFAIK
[22:39] <hggdh> RoAkSoAx: still a bit of hope: vBox is not as stringent. But I sort of doubt, though
[22:40]  * hggdh does not have a netbook...
[22:40] <RoAkSoAx> hggdh, well I can normally run VBox on machines without VT, which is good for old machines... anyways, I'll just give it a try since I'm gonna take one isntead of my laptop
[22:47] <kirkland> RoAkSoAx: definitely
[22:47] <kirkland> RoAkSoAx: you can run it along side kvm
[22:47] <kirkland> RoAkSoAx: just sudo stop qemu-kvm
[22:54] <kopfweh> how can i deactivate the boot splash in lucid server?
[23:03] <pmatulis> kopfweh: what version of GRUB are you using?
[23:03] <kopfweh> installed lycid lynx right now
[23:03] <pmatulis> kopfweh: edit /etc/default/grub.  see https://wiki.ubuntu.com/Grub2
[23:04] <kopfweh> thanx
[23:06] <kopfweh> another question: i have the problem that the boot stops when there is a problem in the fstab and waits till i press s. but this is a little problem for me because i'm using the server with ssh
[23:06] <jaypur> how can i upgrade my server to 10.4 with sudo apt-get ???
[23:06] <jaypur> i think it's not working
[23:09] <pmatulis> jaypur: don't do that.  use 'do-release-upgrade'
[23:09] <jaypur> pmatulis, now its working
[23:09] <jaypur> thanks
[23:10] <MTecknology> I think this is funny.. I get >50 UFW blocks a second from somebody that seems to think that this server has SMTP. They're apparently not even smart enough to check MX records.... Seems like they think changing IP's changes the fact that there's no SMTP server there.
[23:10] <pmatulis> kopfweh: what is the error?
[23:11] <MTecknology> This has been goin on for weeks, I'm starting to think I should maybe call the offenders ISP which isn't even masked..
[23:12] <kopfweh> pmatulis: i disconnected the datadisk and have an entry in the fstab for the disk
[23:12] <kopfweh> pmatulis: i just want it to ignore such errors
[23:12] <pmatulis> kopfweh: what is the error?
[23:13] <MTecknology> Is there any easy way to add a UFW rule that drops the packet and doesn't log but keeps logging everything else?
[23:13] <KristianDK> Hello - is there an amazon image of the new 10.04 available at the moment?
[23:13] <jdstrand> MTecknology: what version of Ubuntu are you using?
[23:13] <MTecknology> KristianDK: yup
[23:13] <MTecknology> jdstrand: 10.04
[23:14] <KristianDK> MTecknology, i couldnt find anything official
[23:14] <MTecknology> jdstrand: err... 9.10 on that system - I'll be upgrading saturday
[23:14] <kopfweh> pmautils: The disk .... is not ready yet or not present. I get this error during boot and the boot stops there till i press S to skip the error
[23:14] <KristianDK> MTecknology, do have any id or something so i can find the image you are reffering to?
[23:15] <kopfweh> pmautils: is it possible that such errors are ignored during boot?
[23:15] <jdstrand> MTecknology: add an explicit deny rule rather than letting it hit your policy
[23:15] <pmatulis> kopfweh: it really depends on the exact error.  normally you should not have errors stemming from fstab
[23:15] <MTecknology> KristianDK: Give me a minute so I can try to find that email - I thought in the announcement it said that was available
[23:16] <jdstrand> MTecknology: that will work going back to hardy
[23:16] <MTecknology> jdstrand: how do I do that?
[23:16] <jdstrand> ufw deny ...
[23:16] <KristianDK> MTecknology, thanks :)
[23:16] <MTecknology> KristianDK: https://lists.ubuntu.com/archives/ubuntu-announce/2010-April/000133.html
[23:16] <MTecknology> jdstrand: that's it... way too easy :)
[23:16] <jdstrand> :)
[23:17] <MTecknology> awesome..
[23:17] <MTecknology> KristianDK: I'm gonna run now - any other questions (or if I was wrong) - much smarter people than I in here to answer
[23:20] <kopfweh> pmatulis: i know that the problem is that disc for the entry in the fstab is not present. i just want the server not to stop at this error and boot anyways
[23:21] <pmatulis> kopfweh: so comment out the line if the disk isn't there
[23:21] <pmatulis> kopfweh: or put option 'noauto' i believe
[23:22] <kopfweh> pmatulis: but i want it to be automatically mounted on boot when it is connected
[23:23] <kopfweh> pmatulis: server 9.10 just ignored the error and booted
[23:23] <pmatulis> kopfweh: i see
[23:23] <pmatulis> kopfweh: maybe if you paste the exact error
[23:24] <pmatulis> kopfweh: what kind of device is it?
[23:26]  * pmatulis trots off to the local Lucid release party
[23:27] <kopfweh> pmatulis: 3 usb disks. during boot it stops with a purple screen which says that this usb disk can't be mounted. (because it isnt connected) but my only problem is that the boot stops there and i cannot login with ssh. so i just want to skip this error message
[23:28] <tag> --devel-release ?
[23:28] <tag> so my 8.04 server is trying to upgrade to intrepid for some reason.
[23:50] <kopfweh> huh giving up
[23:59] <failover> After a fresh install of Karmic my system uses about 50 mb of ram, a fresh install of lucid is using 160 mb, someone knows why ?
[23:59] <failover> both in 64bits version !