[01:08] <snth> Why is there no way to load a different version of the kernel without rebooting?
[01:08] <twb> Erm, because that's how kernels work.
[01:09] <twb> It's like saying "why can't I refuel my cessna without landing?"
[01:09] <snth> twb: hmm .. they should work differently then.
[01:09] <twb> You are welcome to talk to Hurd or Mach about that.
[01:09] <twb> Microkernels are probably more "pluggable"
[01:10] <twb> The erlang people have also done some work into replacing components of a large, scary system while it's running, without the whole thing crashing down
[01:10] <snth> See, it is possible to refuel it without landing then ;).
[01:12] <twb> You can refuel some aircraft in mid-air, but they generally have to be designed from the ground up with that functionality in mind.
[01:12] <snth> I know that you can patch it without rebooting.
[01:12] <twb> Do you just.
[01:20] <jmarsden> snth: You may be interested in http://arstechnica.com/open-source/news/2009/07/ksplice-is-like-viagra-for-linux-server-uptime.ars
[01:21] <snth> jmarsden: Yeah, exactly. They can apply patches, but I am not sure that they can reload a whole new version.
[01:22] <jmarsden> snth: kspice is open source -- if this is important to you, learn all about it and improve it so it does what you need.
[01:22] <jmarsden> *ksplice
[01:32] <JanC> I think VMS kernel can update without rebooting too, I wonder why MS didn't port that VMS-feature to NT with all the rest they took from it  ;)
[01:46] <twb> IIRC ksplice also can't patch kernels arbitrarily.  Any change involving a data structure still requires a reboot.
[02:05] <crohakon> I need a good guide on setting up virtual hosts...
[02:05] <crohakon> anyone got a good link?
[02:08] <twb> The ubuntu admin guide should include that
[02:08] <twb> Unfortunately there's no bloody link in ubottu, nor in /topic, and ICBF digging it up yet again
[02:08] <crohakon> is the admin guide the same as the server guide
[02:08] <twb> Yeah
[02:08] <crohakon> ahh
[02:08] <crohakon> okay, I have the link booked marketed. Thanks
[02:08] <twb> Which is why it takes me so long to find
[02:08] <crohakon> and yeah, why is it not in the topic?
[02:09] <crohakon> Any operators here want to be nice and add the link to the topic?
[02:09] <twb> !ops
[02:09] <crohakon> lol
[02:10] <crohakon> Also, I am not sure creating virtual hosts is the best option...
[02:11] <crohakon> I have two friends that are going to college for digital media design... and part of the course study is building websites. I wanted to set them up with a play ground on my server. Give them an one FTP account each, and access to a sub directory each. I am not really worried about security...
[02:11] <crohakon> Should I just create the two sub domains and create two ftp accounts that only have access to the respective directories?
[02:12] <nalioth> twb: everything sorted?
[02:13] <ajmitch> nalioth: it was about adding the server guide to the topic
[02:13] <crohakon> https://help.ubuntu.com/9.04/serverguide/C/index.html <--- in topic
[02:13] <twb> And/or in ubottu
[02:14] <ajmitch> though when I look now, I see "server guide: http://tinyurl.com/65jzxw"
[02:14] <twb> Oh.
[02:14] <crohakon> oh...
[02:14] <ajmitch> so I guess it's sorted enough
[02:14]  * crohakon dislikes tinyurl...
[02:14] <twb> I guess I don't look at tinyurls because they only exist for rickrolling
[02:14] <ajmitch> rickrolling & cramming as much as possible into a channel topic
[02:15] <crohakon> although, that tinyurl is for 8.something... not 9.04... or .10 for that matter...
[02:15] <twb> You could lose the www. from esr's site :-)
[02:15] <twb> But also the other stuff in /topic wasn't tinyurl'd.
[02:17] <ajmitch> you could lose esr's site
[02:18] <ajmitch> crohakon: right, it goes to the server guide for hardy, the last LTS release
[02:26] <twb> Yeah, replace it with the copy on linuxmafia
[02:26] <twb> Since Moen isn't an NRA goon
[02:26] <twb> >duck<
[02:42] <snth> Does anyone know what's the equivalent to Redhat Kudzu in Ubuntu?
[02:50] <twb> snth: you could create one by adding an init script that does "sleep 10m"
[02:51] <snth> twb: what? Kudzu is a hardware detection tool.
[02:51] <snth> It simply probes for the hardware, looks up the driver/module in a hwdatabase. Then configures your kernel to load that module.
[02:51] <twb> IME what kudzu mainly does is say:
[02:52] <twb> "hi, headless machine with no display!  Apparently the mouse was unplugged, so I will sit here and wait for input for a while instead of booting!"
[02:52] <qman__> I have never needed anything to perform that function since I started using ubuntu
[02:52] <qman__> 4 years ago
[02:52] <twb> But if you're talking about hardware detection, the closest thing is probably udev (for NICs) and UUID/LABELs (for fstab).
[02:53] <qman__> it just works
[02:53] <snth> qman__: I am trying to understand how it works :)
[02:53] <twb> In theory what kudzu is supposed to do is notice when you replace one NIC with another, or add a NIC, and pop up a helpful GUI asking you what to do about it.
[02:53] <qman__> ubuntu just does something about it
[02:53] <qman__> and doesn't ask you
[02:53] <qman__> if it's wrong, you go in and fix it
[02:53] <twb> Mainly what it does is REALLY, REALLY annoy you
[02:54] <snth> Kudzu doesn't have to ask you about anything. Also, it doesn't only work for NICs. It literally probes for any hardware you add/change.
[02:54] <twb> For example, it once managed to remove all the nodes from my RAID (or was it LVM?) array when I swapped in a new HDD
[02:54] <qman__> the point is, that function is not necessary
[02:55] <qman__> the system handles hardware changes without that
[02:55] <snth> Infact, as far as I remember, kudzu doesn' automatically run. It only runs during boot .. then, you can run it manually, but it won't just detect stuff for you after booting.
[02:55] <twb> qman__: rather, Ubuntu acts the same as if you let Kudzu time out.
[02:55] <qman__> right
[02:55] <twb> snth: IME it runs automatically during boot.
[02:55] <qman__> it's handled by udev (I think it's still udev?)
[02:56] <qman__> it just takes care of it
[02:56] <twb> udev will notice new hardware, but not do anything about it.
[02:56] <qman__> no user interaction necessary
[02:56] <snth> *looking up IME*
[02:56] <twb> IME = In My Experience
[02:56] <twb> Install bsdgames, run "wtf ime"
[02:56] <snth> :)
[02:57] <snth> So, when udev detects hardware, how do you load the right module for that hardware?
[02:57] <snth> I mean, how do you know which module/driver to load?
[02:57] <qman__> normally, it just loads it
[02:57] <twb> snth: if modules aren't automatically loaded, that is a bug
[02:58] <twb> snth: the system includes a big database of magic numbers vs. modules to load
[02:58] <snth> What loads that module?
[02:58] <qman__> the only thing the user needs to do is possibly configure the system to use it, like adding a NIC and assigning the IP address
[02:58] <qman__> or adding a disk and specifying a mount point
[02:58] <twb> If you get a very new piece of hardware, it might not be in the database yet.  Then you need to edit /etc/modules by hand.
[02:58] <qman__> theoretically you should never have to load or unload modules
[02:59] <snth> twb: where is that database in the system?
[02:59] <qman__> of course the world isn't perfect and some stuff doesn't work right out of the box, but for the most part, it works great
[02:59] <twb> snth: a combination of /etc/udev/rules.d and /etc/modutils.d, IIRC
[02:59] <twb> snth: udev mainly says things like "oh, a sata disk!  I'd better load SATA disk support."
[02:59] <twb> Some of it is also handled within the kernel itself, I imagine.
[03:00] <qman__> but that's precisely the point
[03:00] <snth> twb: hmm ..that behavior of udev is what kuduz does.
[03:00] <crohakon> trying to setup virtual users in vsftpd is giving me a headache...
[03:00] <qman__> while in redhat you use kudzu, in ubuntu, it just works, and you never have to worry about it
[03:00] <twb> snth: then kudzu is obsolete and useless.
[03:00] <snth> :) fair enough.
[03:04] <snth> hmm udev man page says that udevd receives uevents directly from the kernel if a device is added or removed from the system. qman__ So I guess this is how it just works :)
[03:04] <snth> I wonder what module in the kernel sends these uevents.
[03:21] <snth> twb: man hald this has the database I was talking to you about.
[03:21] <twb> snth: hal is obsolete
[03:21] <twb> snth: it's functionality has been absorbed by udev
[03:21] <snth> hmm .. is this what is replaced by udev?
[03:21] <snth> Oh OK.
[03:22] <twb> https://wiki.ubuntu.com/Halsectomy
[03:23] <twb> Unfortunately, other parts of hal are being replaced by devkit, which is reputed to be just as broken
[03:24] <snth> Cool .. thanks. That's one more difference between rhel and ubuntu. Man, the list keeps growing :).
[03:24] <snth> rhel 5 still uses hal
[03:25] <twb> Ubuntu still uses hal too, for now
[03:26] <snth> I haven't heard of any plans for redhat to change using HAL, yet at least.
[03:26] <twb> Well, RHEL do tend to lag behind everyone else
[03:27] <snth> Yeah. (fortunately/Unfortunately) all of our servers are RHEL.
[03:27] <twb> Pick your flavour of cowboy :-(
[03:28] <snth> My company pays a ton of money for RHEL support and we have never used it -- not even once.
[03:28] <twb> Switch to CentOS, then
[03:28] <snth> I wish it was up to me.
[03:28] <twb> My sympathies.
[03:28] <snth> Management still thinks that it is better to have support for the one time that we will need it.
[03:28] <twb> I am still maintaining servers that run FC3
[03:29] <poningru> heh
[03:29] <poningru> that is... a bad idea
[03:30] <twb> Like you say: not my decision :-(
[03:30] <twb> The problem with linux is it doesn't fail each week, it fails once every few years -- spectacularly.
[03:31] <twb> So it's hard to convince management to fund a man-week to migrate to a newer release
[03:31] <ScottK> snth: If you just want support for the one time you need it, support contracts for Ubuntu Server are a lot cheaper.
[03:31] <twb> They'd prefer to spend a man-month fixing the spectacular failure while all the users mill around due to the unscheduled outage.
[03:31] <snth> I rarely have problems with the kernel itself. It is usually the services on top that can/may crash.
[03:32] <twb> snth: sorry, I meant GNU/Linux, not the kernel
[03:32] <snth> ScottK: I wish Redhat have the same model. The problem is that all of our servers are RHEL for the last 9 or 10 years or so.
[03:33] <snth> It doesn't seem that they have any plans on changing the platform.
[03:33] <snth> As twb said, it doesn't break often, so they don't see why they may need to change to something else.
[03:34] <twb> We need a routine-failure package
[03:34] <twb> It makes stuff go wrong occasionally, and lets you configure the frequency and severity
[03:34] <snth> haha .. then, they would think that you aren't doing your job right.
[03:36] <twb> I actually remember seeing a package a bit like that, which was intended for vocational training of sysadmins (like the RHCE)
[03:36] <snth> Maybe in a few years when ubuntu server gains reputation of stability and so forth, companies would start looking at it since it would be cheaper than redhat.
[03:37] <snth> Yeah, when I was working on my RHCE we had some random scripts that just kinda mess with your system and you have to go figure out what happened. It was kinda cool.
[03:39] <twb> haha "when
[03:39] <twb> You mean "if"
[03:39] <snth> No, I mean "when" :)
[03:46] <snth> Are the ubuntu server maintainers/developers/commiters come to this channel or they have another room?
[03:46] <ScottK> This is both for development and support
[03:48] <snth> ScottK: thanks :) .. I am liking the discussions here so far. You guys form/have a good community in here.
[03:48] <ScottK> snth: It's a distinguishing characteristic of Ubuntu in all it's flavors.  It doesn't happen by accident.
[04:28]  * crohakon smashes head on vsftpd
[04:39] <twb> crohakon: there's probably a #vsftpd
[04:39] <crohakon> there is...
[04:39] <crohakon> I had not thought about that...
[04:39] <crohakon> =)
[04:39] <snth> crohakon: what's the problem with vsftpd?
[04:40] <twb> snth: he's trying to configure it
[04:40] <twb> 14:00 <crohakon> trying to setup virtual users in vsftpd is giving me a headache...
[04:42] <snth> I was just wondering .. I might be able to help.
[04:43] <twb> snth: I was just telling you :-)
[04:43] <crohakon> snth; Every guide I have tried those far has not worked. I just tried setting it up to use Mysql database for user/pass but that does not seem to work either.
[04:44] <crohakon> those = thus
[04:44] <snth> twb: I was kinda hoping for a bit more accurate description of the problem :p.
[04:44] <twb> snth: fair enough
[04:44] <snth> crohakon: what's your config file look like?
[04:46] <crohakon> How is it that I send it to pastebin or what not from console?
[04:47] <twb> crohakon: w3m hpaste.org
[04:47] <twb> That's how I'd do it, anyway
[04:47] <twb> There are little dedicated paste utilities, but I can't recommend one
[04:48] <snth> pastebinit /etc/vsftpd
[04:48] <crohakon> ahhh
[04:48] <crohakon> there we go
[04:51] <crohakon> http://pastebin.com/f477bf642
[04:52] <snth> crohakon: and you want mysql authentication only ?
[04:53] <crohakon> yes
[04:53] <snth> crohakon: first mistake is in line 27. Change this to NO.
[04:54] <crohakon> okay, done
[04:54] <snth> crohakon: and where exactly in this file do you try to authenticate against mysql?
[04:55] <crohakon> I don't know... I was following a guide... I just did what it tells me to do... something to do with PAM?
[04:58] <crohakon> /etc/pam.d/vsftpd contains the lines for accessing the DB I believe...
[04:58] <snth> it shouldn't .. but can you show me the contents of that file.
[04:58] <snth> Also do you have the package libpam-mysql?
[05:01] <crohakon> I do now..
[05:02] <snth> sweet. have you created the database?
[05:02] <crohakon> yes, it is created and I have users added already
[05:03] <snth> crohakon: IM me, I need to send you config lines to add.
[05:12] <twb> You realize that libpam-mysql only does one of the four pam bits, don't you?
[05:13] <twb> (auth and not password/session/account, IIRC.)
[05:14] <snth> twb: Yeah, that's how I am configuring it with crohakon now. It should suffice, right?
[05:14] <twb> I don't know.
[05:14] <snth> I am pretty sure that it should be fine.
[05:14] <twb> I would never use mysql for anything.
[05:19] <crohakon> =(
[05:19] <crohakon> It just seemed like the easiest way...
[05:25] <qman__> mysql and easy don't belong in the same sentence
[06:03] <kosmic> hey guys
[06:04] <crohakon> qman__; I have worked with mysql for some time now... so I figured it would give me an advantage.... I was wrong.
[06:12] <twb> That's like saying that you're been a plumber for five years, so you thought you'd make a sports jacket out of turds.
[06:23] <crohakon> twb; I understand this now... but a few hours ago I was a tiny bit more ignorant... hehe. I am slowly working on curing that... one struggle at a time.
[06:58] <Classic> hello all
[06:58] <Classic> i get this error when ever i try installing mysql
[06:58] <Classic> errors were encountered while processing: fuse-utils, gvfs-fuse
[06:58] <matei> I have a question about the Karmic EC2 image: are Linux Containers (lxc) supposed to work on it? When I try to use lxc-execute or lxc-start, I get "lxc-execute: failed to clone(0x2c020000): Invalid argument"
[07:09] <jmarsden> Classic: Installing mysql how, on what version of Ubuntu server?
[07:09] <Classic> jmarsden: i used sudo apt-get install mysql. the version is 9.02
[07:09] <kblin> Classic: seems unconnected to mysql, though
[07:10] <kblin> ...
[07:10] <kblin> Classic: seems unconnected to mysql, though
[07:10] <jmarsden> Classic: Sounds like your apt-get database is inconsistent or otherwise unhappy?  Try sudo apt-get -f install      and then try the sudo apt-get install mysql again.
[07:10] <Classic> kblin: i get the same error when i uninstall mysql using sudo apt-get remove mysql
[07:10] <Classic> ok
[07:11] <Classic> I get the same kind of error
[07:11] <Classic> From running sudo apt-get -f install
[07:11] <Classic> Errors where encountered while processing:
[07:11] <Classic> fuse-utils
[07:11] <Classic> gvfs-fuse
[07:11] <crohakon> Thanks to snth I can stop hitting my head against vsftpd..
[07:12] <Classic> E: Sub-process /usr/bin/dpkg reutrned an error code (1)
[07:13] <jmarsden> Classic: OK...so it's *really* unhappy.. :(  Did you get those two packages in some unusual way, or have strange issues installing them?
[07:13] <snth> Alright guys -- having a long week :) gotta get some sleep. Later.
[07:13] <Classic> jmarsden: Wel I'm running on a VPS and I'm helping a friend install something. This was the very first command line I typed: sudo apt-get install mysql-server subversion binutils cvs cvsutils gcc gdb make libmysql++-dev libssl-dev libtool automake g++
[07:14] <snth> twb: mysql_pam supports auth and account :).
[07:14] <jmarsden> Classic: That doesn't really answer my question... did you try at some point to install fuse-utils or gvfs-fuse?
[07:14] <twb> snth: righto
[07:15] <Classic> jmarsden: Nope.
[07:15] <snth> Later
[07:15] <jmarsden> Classic: Did your friend, before you started helping?
[07:16] <Classic> jmarsden:  No.
[07:16] <Classic> jmarsden: I was the very first who accessed it. Idk if its because I was using debian commands at one point >< (I am more familiar with debian than ubuntu so yeahh).
[07:16] <kblin> debian commands?
[07:17] <twb> "apt-get -f install" is not a Debianism
[07:17] <twb> It's a "I broke the system, please try to guess the fix for me"ism.
[07:17] <jmarsden> Classic: Both sudo and apt-get work the same way in both Ubuntu and Debian... so what did you or someone else *really* do here?
[07:18] <twb> Classic: pastebin the *entire* output of apt-get -f install.
[07:18] <twb> Classic: I suspect the real information about the problem is during the dpkg --configure -a step.
[07:19] <kblin> at least the fuse parts are not part of a plain server install
[07:24] <Classic> root@sfwebhosting:/# sudo apt-get -f install
[07:24] <Classic> Reading package lists... Done
[07:24] <Classic> Building dependency tree... Done
[07:24] <Classic> 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
[07:24] <Classic> 2 not fully installed or removed.
[07:24] <Classic> After this operation, 0B of additional disk space will be used.
[07:24] <Classic> Setting up fuse-utils (2.7.4-1.1ubuntu4) ...
[07:24] <Classic> creating fuse group...
[07:24] <Classic> udev active, skipping device node creation.
[07:24] <jmarsden> Classic: Use pastebin please!
[07:24] <Classic>  * Reloading kernel event manager...                                            No /sbin/udevd found running; none killed.
[07:24] <Classic>                                                                          [fail]
[07:24] <Classic> invoke-rc.d: initscript udev, action "reload" failed.
[07:24] <Classic> dpkg: error processing fuse-utils (--configure):
[07:24] <Classic>  subprocess post-installation script returned error exit status 1
[07:25] <Classic> dpkg: dependency problems prevent configuration of gvfs-fuse:
[07:25] <omani> Classic, wtf?!
[07:25] <Classic>  gvfs-fuse depends on fuse-utils; however:
[07:25] <Classic>   Package fuse-utils is not configured yet.
[07:25] <Classic> dpkg: error processing gvfs-fuse (--configure):
[07:25] <Classic>  dependency problems - leaving unconfigured
[07:25] <Classic> No apport report written because the error message indicates its a followup error from a previous failure.
[07:25] <Classic>                           Errors were encountered while processing:
[07:25] <Classic>  fuse-utils
[07:25] <Classic>  gvfs-fuse
[07:25] <Classic> E: Sub-process /usr/bin/dpkg returned an error code (1)
[07:25] <Classic> Sorry
[07:25] <Classic> http://paste-it.net/public/wd13eb3/
[07:26] <jmarsden> If you are sure you don't need those packages, try removing them.
[07:26] <Classic> ok
[07:28] <Classic> that did the trick
[07:28] <Classic> thanks
[07:30] <jmarsden> Classic: No problem.
[07:38] <__ruben> hm .. rebooted a jaunty router/firewall due to power maintainance .. now its ipv6 stack seems totally busted (cant ping link local addresses, radvd cant send packets, ..)
[07:39] <kblin> __ruben: you sure ipv6 came up at all? tried loading the module?
[07:41] <__ruben> kblin: in jaunty its no longer a module
[07:41] <__ruben> and i do see link local addresses, but they're tentative
[07:42] <kblin> __ruben: certainly is a module for me
[07:42] <__ruben> kblin: on jaunty?
[07:43] <kblin> yeah
[07:43] <kblin> ubuntu armel port
[07:44] <kblin> though it's a beagleboard, your kernel might be different
[07:44] <__ruben> ah .. ports (can) use different .config's
[07:45] <__ruben> hmm .. attempt to boot previous kernel by altering menu.lst seems to have failed .. and the box is remote (at work)
[07:46] <kblin> __ruben: well, there's no official ubuntu kernel for my hardware
[07:49] <twb> Dunno why you'd use Ubuntu on armel instead of Debian or Emdebian.
[07:49] <twb> At leat armel is a first-class arch on Debian
[07:50] <kblin> not for the OMAP SOCs
[07:50] <twb> Huh.
[07:51] <twb> I'm using kirkwood here, so I guess I just assumed the OMAP targets were already supported
[07:57] <quizme> ssh ubuntu@stuff.com 'echo $PATH'  ..... my $PATH variable is not fully set .... i added a couple directories to my PATH variable in .bashrc.... anybody know why ?
[07:58] <quizme> why doesn't .bashrc get read when running commands with ssh
[07:58] <__ruben> perhaps it gets executed by sh instead of bash for instance
[07:59] <twb> Passing a command to ssh will cause that command to be run in a sh shell on the remote host.
[07:59] <twb> The only configuration file on the remote host that might be read in this circumstance is ~/.ssh/environment.  This file is not read by default.
[08:00] <twb> You can also configure sshd to allow, and ssh to send, additional environment variables.  $LANG and $LC_* are typically in this list; $PATH is (for obvious reasons) not.
[08:01] <twb> Additionally, if you control the script invoking ssh, you can simply do something like ssh x env PATH=z y
[08:02] <quizme> no it's running bash
[08:02] <quizme> the error is:   bash: rake: command not found
[08:02] <quizme> so it's comming from bash
[08:03] <twb> Hm.
[08:04] <twb> Maybe it invokes your login shell -- I didn't think it did.
[08:04] <jmarsden> twb: ssh jonathan@localhost 'echo $SHELL'  outputs /bin/bash, so it sure looks like bash is being run to me...
[08:04] <twb> You're right, it invokes the login shell.
[08:04] <twb> OK, this is even weirder
[08:05] <twb> I changed my shell to /bin/sh on the remote host, and it's still using bash
[08:05] <twb> Grr, NIS' ypchsh didn't do its job.
[08:05] <quizme> can i set PATH=$PATH:/opt/ruby19/bin:.  in ~/.ssh/environment ?
[08:06] <twb> TIAS
[08:08] <quizme> doesn't work
[08:08] <twb> Did you enable environment?
[08:08] <quizme> how do you do that ?
[08:08] <twb> PermitUserEnvironment    Specifies whether ~/.ssh/environment and environment= options in ~/.ssh/authorized_keys are processed by sshd(8).  The default is “no”.  Enabling environment processing may enable users to bypass access restrictions in some configurations using mechanisms such as LD_PRELOAD.
[08:08] <med\weed> quizme: try doing one of those path add commands as an export
[08:08] <twb> ...from sshd_config
[08:09] <med\weed> something like
[08:09] <med\weed> ssh "export $PATH=$PATH:/SOME/PATH; commands"
[08:10] <twb> med\weed: those should be single quots, and the leading $ is wrong, and ideally you'd use && instead of ;.
[08:10] <quizme> med\weed want to to configure it so that i don't have to do that every time
[08:10] <twb> But I already suggested that approach: 19:01 <twb> Additionally, if you control the script invoking ssh, you can simply do something like ssh x env PATH=z y
[08:12] <crohakon> how do I view a list of users in terminal?
[08:12] <twb> getent passwd
[08:13] <qman__> who
[08:13] <crohakon> thanks
[08:14] <crohakon> goodnight
[08:14] <quizme> so i enabled PermitUserEnvironment
[08:15] <quizme> but in ~/.ssh/environment    I put "duck=wow"  but when i did locally: "ssh ubuntu@stuff.com 'echo $duck'" it was empty
[08:16] <med\weed> twb: i know - like i said _something like_
[08:16] <twb> quizme: did you restart sshd?
[08:16] <med\weed> its only for conceptual ~_~
[08:16] <quizme> twb: no....
[08:17] <quizme> twb: /etc/init.d/sshd restart    like that ?
[08:17] <twb> That will do, yes.
[08:17] <twb> ssh, not sshd, actuall
[08:19] <quizme> i restarted ssh, but $duck is still the empty string
[08:20] <twb> quizme: I give up.  Ask #openssh
[08:21] <quizme> k
[08:21] <quizme> thanks
[08:22] <quizme> pretty tricky stuff
[08:35] <quizme> twb: it's set in /etc/environment
[09:07] <ycy> in order to use rsync between two machine, should one of them have rsync server (and thus a tcp port open) installed?
[09:09] <atomic_1> no
[09:10] <kwork> you can rsync over ssh i think
[09:10] <atomic_1> yeah
[09:11] <atomic_1> rsync -arvupz /source/folder --exclude-from '/from/here.txt' user@remotehost:/BackupFolder > /your/logfile.txt is what i use
[09:12] <atomic_1> shove it in crontab and you're done
[09:12] <kwork> thou bacula is fancier then rsync
[09:23] <cemc> or maybe BackupPC? which can also use rsync
[09:51] <knecht> hi there. i have to compress huge backups (>120GB). Some files are already compressed (rar, zip, tar.bz2 files), and i search for a way to get these files into an archiv file without double-compresse them. (i prefere 7z or tar.bz2)
[09:52] <knecht> in short words: is there a way to tell an archiv manager: Compress all Data except already compressed files, and put everything in a archive? (like a "do not compress these files" filter)
[09:52] <knecht> ?
[10:02] <knecht> in short words: is there a way to tell an archiv manager: Compress all Data except already compressed files, and put everything in a archive? (like a "do not compress these files" filter, but dont skip the files)?
[10:04] <kwork> you can tell archiver to udpate the archieve with the changed files
[10:05] <jmarsden> knecht: By definition a .tar.bz2 file is a single .tar file which is compressed using bzip2, so ... no, what you want can't be done with that archive format.
[10:06] <jmarsden> You could create a .tar file or all the already-compressed files, and a .bar/bz2 of everything else, maybe?
[10:07] <knecht> kwork: i know, but that does not help me out. i need to do full backups every day, and it tooks to long. i thought already compressed files don't need to be processed again.
[10:07] <jmarsden> knecht: Is the issue really CPU usage??  More likely the speed issue is that your backups are I/O bound, in which case trying to avoid compressing some files will not really help your speed.
[10:07] <knecht> jmarsden: to make two archiv is a way, but i will try to find a better one
[10:08] <kwork> knecht, try something like bacula
[10:08] <knecht> jmarsden: i have a raid, the cpu is for shure the problem
[10:08] <knecht> kwork: i take a look at bacula (never heard that before)
[10:14] <knecht> kwork: mabe the best way is to copy the already compressed archive every day, and update the clone . . . should work, or?
[10:15] <knecht> kwork: bacula is to big for my purpose
[10:16] <kwork> for me bacula is doing one full backup + incremental
[10:16] <kwork> from there
[10:16] <kwork> rather tehn taring whole world together and rsyncing it
[10:16] <kwork> i have move that that option
[10:16] <kwork> its pretty straightforward to set up aswell
[10:16] <jmarsden> knecht: If your backups are all on disk, use rsync to update them each day; something like rsnapshot could work well for you and is smaller/simpler than bacula
[10:20] <knecht> kwork: i agree, but my boss does not. There should be rotating full backups available of the last seven days. Not incremental cause he wants to be able to simply pick a backup via smba, and open it on his computer. So if it works, it is the way to go for me.
[10:36] <kwork> knecht, you can define rules
[10:36] <kwork> so that you make full backup weekly for exampel
[10:36] <kwork> and incremental otherwise
[10:42] <knecht> kwork: i also tend to do so, for now i had to do it like i mentioned. Thanks for your help!
[11:06] <Vertigo009> Hello everyone. I'm trying to install pear auth_SASL in my ubuntu server, but after i do the install, it keeps saying that it is not installed. Can anybody help me please? Thanks.
[11:07] <Vertigo009> root@neptuno:/# pear install auth_SASL
[11:07] <Vertigo009> downloading Auth_SASL-1.0.3.tgz ...
[11:07] <Vertigo009> Starting to download Auth_SASL-1.0.3.tgz (5,724 bytes)
[11:07] <Vertigo009> .....done: 5,724 bytes
[11:07] <Vertigo009> root@neptuno:/# pear list auth_SASL
[11:07] <Vertigo009> `auth_SASL' not installed
[11:07] <Vertigo009> root@neptuno:/#
[12:09] <kim0> Hi folks .. ubuntu server 8.04 LTS, fails to boot when a degraded raid happens
[12:09] <kim0> I understand passing the kernel option, "bootdegraded=true" .. should resolve this
[12:09] <kim0> but it seems to be ignored
[12:10] <kim0> any idea if kirkland fixes have been integrated into 8.04.2 ?
[12:24] <joeD2> get during system installation the follwing message: file:///cdrom/pool/main/u/util-linux/bsdutils_2.16-1ubuntu5_amd.64.deb was corrupt. downloaded the image already twice. Any idea? Thnx.
[12:27] <pmatulis> joeD2: why do you need to install from the cd?
[12:27] <joeD2> pmatulis: Where do you install from???
[12:28] <pmatulis> joeD2: the internet
[12:29] <joeD2> pmatulis: what is the command for that? Or do you have a link for a description?
[12:29] <pmatulis> joeD2: remove the cdrom line (or comment it out with a '#' mark) and then '$ sudo aptitude update; sudo aptitude install bsdutils'
[12:30] <pmatulis> joeD2: the file is /etc/apt/sources.list
[12:30] <joeD2> pmatulis: the system has not been up. it's during the groundsystem setup that I get the message.
[12:31] <pmatulis> joeD2: ah
[12:31] <pmatulis> joeD2: check the md5sum of the iso you d/l'd
[12:32] <joeD2> pmatulis: how?
[12:36] <pmatulis> joeD2: '$ md5sum /path/to/iso' and then compare it to the mirror you d/l'd from, do a search for "ubuntu md5sum"
[12:36] <gamla_kossan> umm, how can I check when a user account was created?
[12:36] <gamla_kossan> by the date of their homediw?
[12:36] <gamla_kossan> *dir?
[12:36] <joeD2> pmatulis: md5sum ubuntu-9.10-server-amd64.iso 14707e8847b9c9ba2dd1869fb5086e4f  ubuntu-9.10-server-amd64.iso is the result. downloades from http://releases.ubuntu.com/releases/9.10/
[12:41] <joeD2> pmatulis: found the md5 and it matches nevertheless I burned two cds (after two seperate downloads) and still got the sysm error.
[12:44] <pmatulis> joeD2: i recommend burning another cd but at a lower speed
[12:45] <joeD2> pmatulis: already on the way trying it, but what is strange that I bourned the two cds also on different burner
[12:48] <pmatulis> joeD2: wow, that *is* strange
[12:48] <zul> morning
[12:49] <pmatulis> zul: good morning
[12:49] <zul> hey pmatulis
[12:49] <joeD2> pmatulis: I thought it could be a problem in the iso, but seams no one else has the problem.
[12:50] <pmatulis> joeD2: maybe also test your ram, it might be using the exact place in ram during the install and then boom
[12:52] <joeD2> pmatulis: don't get it. what exact place in the ram?
[12:53] <pmatulis> joeD2: during the install, when it processes bsdutils
[12:53] <joeD2> pmatulis: by teh way started with a new cd
[13:11] <joeD2> pmatulis: ok it was the cd. changed the cd brand and downgraded the burning speed. now server is installed and running. thnx
[13:21] <pmatulis> joeD2: good stuff
[13:52] <kirkland> kim0: those fixes should have landed for 8.04.3
[13:52] <kirkland> kim0: though you would still need to manually run grub-install on your md device, to put a bootloader on each disk
[13:57] <Fenix1> if my domain ip is 67.20.30.1 what is the reverse zone file address
[13:58] <Fenix1> I made up the ip
[13:59] <Fenix1> Do u only but the network part of your address in a reverse zone file?
[13:59] <kblin> depends on your setup
[14:00] <Fenix1> what do u mean
[14:01] <kblin> what you put in the zone setup in named.conf.local
[14:01] <andol> Fenix1: The actually file name you specify yourself. The name of the zone depends on what is delegated from above.
[14:01] <kim0> kirkland: ah .. thanks
[14:02] <kblin> eg. if it's zone "30.20.67.in-addr.arpa", you'd only put the last octet into the database file
[14:02] <Fenix1> ok thanks
[14:19] <Sorell> Does any one know If I must have 2 machines to run a simple cloud
[14:19] <Sorell> I will be using this as a personal cloud
[14:19] <Sorell> ( I.E. just for me )
[14:21] <kblin> I haven't poked the cloud setup yet. for personal use, I just use a vbox or kvm directly
[14:21] <Sorell> I'm really just doing this right now, as an experiment.
[14:22] <Sorell> https://help.ubuntu.com/community/UEC/CDInstall
[14:23] <Sorell> I'm looking at this right now
[14:23] <Sorell> it says that I need to have a front end and a node
[14:23] <Sorell> but that seems kind of silly for what I'm doing as
[14:24] <PJiPhone> Can they not be virtual?
[14:24] <Sorell> the front end will be doing almost nothing.
[14:24] <Sorell> I was thinking that, however I'm not sure
[14:25] <kblin> well, the nodes run kvm, right?
[14:25] <kblin> I'm not sure if you can even run kvm on a kvm guest
[14:25] <Sorell> I think so.
[14:25] <Sorell> right
[14:26] <Sorell> because I think that it needs the CPU VM extensions
[14:26] <PJiPhone> You can do remote X though
[14:26] <Sorell> and the VMs wouldn'tdo that.
[14:26] <Sorell> :/
[14:29] <PJiPhone> Or VNC of course
[14:29] <Sorell> yeah I really didn't want to do that.
[14:30] <Sorell> the end goal here is to be able to access my system on whatever machine I'm on from a host
[14:31] <Sorell> with min impact on said host
[14:31] <Sorell> ( I'm looking into setting up a web front end. )
[14:34] <PJiPhone> VNC is not hard
[14:34] <PJiPhone> Sure it could be done via web
[14:49] <Sorell> grabs gaming rig pulls out HDD and slips in clean disk.
[14:50] <Sorell> * Starts node installation
[15:00] <Sorell> anyone know if ubuntu can take advantage of cuda?
[15:04] <Brian_H> how do you manage nfs share rights between users?
[15:04] <Brian_H> with samba its pretty easy, and acls just work, however with nfs it seems that its not so easy
[15:06] <__ruben> nfs uses the standard filesystem permissions
[15:07] <andol> Brian_H: Of course, if you don't have the same uids and gids, yes then I'm sure it can be confusing :)
[15:09] <Xpistos|work> Is there an FTP site to download Ubuntu Server? I can only find torrent and HTTP
[15:10] <jpds> Xpistos|work: https://launchpad.net/ubuntu/+cdmirrors
[15:11] <Xpistos|work> jpds: Thanks
[15:11] <jdstrand> soren, kirkland: fyi, https://wiki.ubuntu.com/Virtualization needs some love
[15:13] <kim0> Hi folks .. my preseed late_command seems to be not running some commands .. where do I look for hints why they're not running ?
[15:16]  * soren hands jdstrand the "Understatement of the day" award
[15:18] <jdstrand> :)
[15:20] <kirkland> jdstrand: agreed
[15:21] <kirkland> jdstrand: however, wiki.ubuntu.com contains all sorts of out of date old UDS spec cruft
[15:24] <soren> For starters, we could move it to something like /UDS/Hardy/Virtualisation.
[15:25] <soren> At least that way it doesn't make any promises it can't keep :)
[15:26] <zul> lol
[15:27] <jdstrand> kirkland: sure, but I don't think that was a spec
[15:28] <jdstrand> do with it what you will, I was looking for the documentation for libvirt in karmic, and naturally looked there, and it was, uhmm, not relevant ;)
[15:28] <kirkland> jdstrand: oh?  i thought it looked like the Hardy Virtualization spec
[15:28] <jdstrand> kirkland: maybe it was, I didn't look to closely
[15:28] <jdstrand> too
[15:29] <kirkland> jdstrand: i think this is the spec that was used to decide to go with KVM rather than Xen in Ubuntu
[15:29] <jdstrand> kirkland: ah
[15:30] <jdstrand> kirkland: well, one thing the security team has done to clean these things up with things it cares about is create a SecurityTeam/Specifications page, with links to specs in it, then create landing pages for important topics
[15:30] <jdstrand> kirkland: eg-- wiki.ubuntu.com/AppArmor was a spec for using AppArmor in Ubuntu
[15:30] <jdstrand> it was horribly out of date
[15:31] <jdstrand> so we moved the spec to SecurityTeam/Specifications/AppArmor, then created a new wiki.ubuntu.com/AppArmor page, with links all over
[15:31] <jdstrand> kirkland: food for thought, I'm certainly not saying you have to do something similar
[15:32] <kirkland> jdstrand: gotcha, thanks.
[15:32] <kirkland> jdstrand: it does need some housekeeping, i agree
[15:32] <jdstrand> and we didn't do this for *everything*, just things we notice as we go
[15:33] <soren> jdstrand, kirkland: It was basically a collective braindump from the virtualisation discussion at UDS in Boston.
[15:33]  * jdstrand remembers that session fondly
[15:34] <soren> Man, I had /no/ clue back then :)
[15:34] <jdstrand> heh
[15:44] <LyonJT> Hey
[15:44] <LyonJT> I'm having a few issues with my interfaces file and OpenVPN
[15:44] <LyonJT> anyone able to help?
[15:51] <mugginz> Anyone around who has info on Samba's installation policy?
[15:51] <mugginz> Wondering what the current position is with respect to sharing folders on a default desktop install
[15:58]  * soren calls it a day
[16:22] <StrangeCharm> i've decided that i want to move /home, /var, and /tmp onto their own disks. is there a good tutorial for this?
[16:23] <mushroomblue> StrangeCharm: cp -a * /path/to/new/var
[16:23] <mushroomblue> then mount that path as /var, and update fstab
[16:23] <mushroomblue> s/mount/remount/
[16:23] <StrangeCharm> thanks, mushroomblue
[16:24] <mushroomblue> yep.
[16:34] <kblin> StrangeCharm: actually you'll want to make sure that once the data is oin the new location, you remove it from the old location
[16:34] <kblin> otherwise you won't get back the disk space
[16:35] <StrangeCharm> kblin, gotcha
[16:40] <LyonJT> Does anyone have a interfaces file template?
[16:40] <LyonJT> mines gone wrong and i need the template back
[16:47] <__ruben> there's no "template", either start from scratch or find the error(s)
[16:48] <LyonJT> __ruben: i just need one from scratch
[16:48] <LyonJT> i.e. when u first install the OS
[16:48] <LyonJT> then i can amend it from there!
[16:48] <__ruben> it depends on the installation, whether or not dhcp was available or not for instance
[16:49] <LyonJT> 9.10 and it was static
[16:50] <LyonJT> i will amend the ip addresses etc
[16:50] <kblin> man interfaces has a simple one, I think
[16:52] <LyonJT> wiked cheers
[16:53] <LyonJT> Nope man doesn't
[16:56] <kblin> it does have a description, though
[16:57] <LyonJT> okay
[17:20] <arthurjohnson> Okay, I have the ubuntu mini.iso and just ran a cli install.  What should I apt-get to be sure to have the "server" release?
[17:26] <zash> arthurjohnson: "ubuntu-server" i think
[17:27] <zash> or what, no such package in karmic?
[18:06] <incorrect> is acl being replaced by something else?
[18:09] <blackxored> there's a git plugin for fusionforge?
[18:10] <googa> if i have bind installed, will it get answers for queries directed to it?
[18:11] <qman__> googa, only if you configure clients to ask it
[18:12] <googa> so it resolves automatically
[18:12] <googa> ?
[18:12] <googa> without seperate configuration
[18:12] <qman__> what do you mean
[18:13] <googa> well, if you have a authrative name serer for some area, you have to make all these settings for the area and *yawn*, but if you want to set up a resolver for a network you only need to install bind a assign clients to it
[18:13] <googa> ?
[18:14] <googa> ye i guess so
[18:14] <googa> it directs them to root servers or smoething, ye?
[18:14] <googa> defined in named.conf
[18:14] <qman__> you have to either enable root hints, or recursive queries, or a forwarder
[18:14] <qman__> root hints are enabled by default
[18:14] <googa> yesyes
[18:14] <googa> thats all
[18:14] <googa> thank u
[18:14] <qman__> yep
[18:14] <googa> :D
[18:24] <leonel> I see that  kqemu is no longer supported in qemu.. what's the way to go for a non  VMX cpu ??
[18:24] <leonel> this with karmic
[18:31] <caseyd> is there a gui that comes with ubuntu-server edition.. or is it just a terminal interface
[18:34] <leonel> caseyd: servers normally does not have/use  a monitor
[18:34] <leonel> caseyd: you can install  any gui you like
[18:35] <qman__> caseyd, you can install ubuntu-desktop if you want, but that is not supported in this channel
[18:36] <bogeyd6> caseyd servers by definition use no gui, but using one such as xfce might now be so bad
[18:36] <leonel> caseyd: sudo apt-get install ubuntu-destkop --no-install-recommends
[18:36] <leonel> caseyd: that will give you only gnome
[18:39] <arthurjohnson> As far as I can tell, there is no "ubuntu-server" metapackage.  I installed a cli only system, I'll just install the linux-image-server and openssh-server.  That should be enough, right?
[18:39] <qman__> arthurjohnson, linux-image-server is one of the kernel packages
[18:40] <qman__> if your system is running you already have it
[18:40] <caseyd> cool, thanks
[18:40] <arthurjohnson> qman__: I have the ubuntu mini.iso, and installed a cli system.  I'm trying to install the "ubuntu-server" edition, or as close as I can.
[18:41] <caseyd> I decided since im a beginner for linux, ill just install ubuntu desktop and then install lamp on it. Will be a little easier for me to configure. Thanks =)
[18:41] <qman__> arthurjohnson, there's not much installed by default
[18:41] <qman__> base system, and that's it
[18:42] <qman__> SSH is an option but is not installed by default
[18:42] <arthurjohnson> qman__: Okay, then I suppose I'm pretty close with the "cli" option and linux-image-server, maybe even a little leaner.
[18:50] <dinger1986> darkpixel: i got it working but decided aginst it anyways hope your uncle is ok
[18:59] <ruben23> hi
[18:59] <ruben23> how do i change timezone on my ubuntu-server
[19:01] <ruben23> anyone
[19:03] <qman__> http://lmgtfy.com/?q=ubuntu+change+timezone+command+line
[19:03] <Pici> qman__: Thats not helpful, this is a support channel.
[19:04] <qman__> it most certainly is helpful, the first result has his answer
[19:04] <ScottK> ruben23: IIRC it's sudo dpkt-reconfigure tzconfig, unless you are on Dapper
[19:04] <Pici> ruben23: I believe  its sudo dpkg-reconfigure tzdata or tzconfig
[19:04] <ScottK> qman__: It may be helpful, but it's not a very friendly way to provide help.
[19:06] <ruben23> Pici:  it will not chnage after reboot...?
[19:07] <Pici> ruben23: It shouldn't.
[19:29] <cemc> theoretical question: I have a md0 (raid1). one of the disks gets poof... what happens if I try to add a new _smaller_ disk to md0 ? error message when I try to add, error at the sync's end, or what?
[19:46] <bogeyd6> cemc different drives and speeds are fine
[19:46] <bogeyd6> cemc you will be limited to the size of the smallest drive
[19:47] <bogeyd6> cemc but what you can do is change the bad drive, then build, then change the smallest drive to the biggest and build
[19:48] <cemc> I know. but what happens if I try to add a smaller drive to an active array?
[19:48] <cemc> to an active md0 I mean
[19:49] <bogeyd6> it depends on if the current data usage exceeds the limits of the smallest drive
[19:50] <bogeyd6> my guess is you are trying to run it very close, and if you do, you will fail
[19:53] <cemc> root@u910-ipv6:~# mdadm -a /dev/md0 /dev/sdb1
[19:53] <cemc> mdadm: /dev/sdb1 not large enough to join array
[19:53] <cemc> this is what I wanted to know :)
[19:53] <cemc> thanks
[19:54] <bogeyd6> yw
[19:54] <bogeyd6> cemc is this software raid?
[19:55] <cemc> yes
[19:55] <cemc> and I made sdb1 smaller than sda1 to see what happens if I try to add it
[20:47] <incorrect> stupid upgrade process
[20:49] <Sorell> :(
[20:50] <Sorell> something break?
[20:50] <incorrect> libvirt didn't like the fact that i had users on ldap
[20:50] <incorrect> the system rebooted
[20:50] <Sorell> :(
[20:50] <kirkland> smoser: ping
[20:50] <smoser> hey
[20:50] <incorrect> seems ldap upgraded ok, now i need to finish off i guess a dist-upgrade should do it
[20:51] <smoser> kirkland,
[20:51] <kirkland> smoser: do you have the ami-* id of the karmic GA image in ec2?
[20:51] <smoser> si
[20:51] <kirkland> smoser: or a pointer to where I could find it?
[20:51] <smoser> http://uec-images.ubuntu.com/releases/karmic/release/
[20:53] <smoser> kirkland, ^^
[20:55] <incorrect> ok nothing major
[20:55] <incorrect> i really should setup up a second ldap server
[20:55] <Sorell> :)
[20:55] <incorrect> and all my kvm's starting back up
[20:56] <incorrect> sadly i only have one server in my garage
[20:57] <aubre> are those images the same ones you get when you install from the Store in Eucalyptus
[20:57] <aubre> ?
[20:57] <incorrect> aubre, talking to me?
[20:57] <Sorell> I think to smoser
[20:58] <aubre> incorrect: sorry
[20:58] <mathiaz> kirkland: hi!
[20:58] <incorrect> wow even my zimbra server came back to life!
[20:58] <mathiaz> kirkland: do we have a list of eucalyptus we're looking at for SRUs?
[20:58] <aubre> smoser: are those images the same ones you get when you install from the Store in UEC
[20:58] <mathiaz> kirkland: /SRUs/SRU/
[20:58] <Sorell> incorrect: congrats
[20:58] <smoser> aubre, they are, yes.
[20:59] <incorrect> phew! my heart sank when libvirt didn't upgrade
[20:59] <aubre> smoser: ok, they still say RC in the Store, but they work great for me
[20:59] <incorrect> can you rsync to ubuntuone?
[21:00] <incorrect> i think i will use that for my backups if i can simply just rsync
[21:00] <Sorell> I posted this in the forum, does it make sense?
[21:00] <Sorell> http://ubuntuforums.org/showthread.php?t=1311527
[21:00] <ScottK> mathiaz: From reading my bugmail, it looks like Courier is deeply broken.  I don't have time to look into it, but it's probably a good area for someone to dig into.
[21:01] <aubre> I'm thinking of writing a "client document" that I can use to hand hold users through the whole client process, since they don't really need to know all that we know about setting the system up
[21:01] <smoser> aubre, there be a couple noticable fixes in release compared to rc.  bug 458850 was the biggest.  then there is another fix that happened in 'mountall' that i think would cause a x86_64 image to not boot.
[21:01] <mathiaz> ScottK: hm ok. I'll keep that in mind then.
[21:01] <ScottK> Thanks.
[21:02] <aubre> smoser: hmm. I'll have to keep that in mind
[21:02] <aubre> smoser: I am running release on my front end and my nodes
[21:03] <incorrect> oh no my zimbra vm just vaporised
[21:04] <aubre> smoser: I did notice that in the SC while you can specify the device you want the volume to have, it will assign you the next available alphabetical letter no matter what, i.e. I put --device /dev/sdh , I get /dev/sdb. I noted this in StorageController documentation, I didn't think it was a big deal
[21:05] <aubre> Really when it comes to SC, I wouldn't be sad if it didn't pick my device name for me and just told me what to use.
[21:05] <kirkland> mathiaz: hmm, good question
[21:06] <kirkland> mathiaz: i think just the bugs that are targeted at karmic-updates
[21:06] <mathiaz> kirkland: right - ttx asked to keep an eye on these and prepare an SRU while he is away
[21:06] <kirkland> mathiaz: agreed, i'll help you with that
[21:06] <mathiaz> kirkland: I've worked on a bunch of them
[21:06] <kirkland> mathiaz: i think we should plan an SRU asap
[21:07] <aubre> smoser: kirkland: mathiaz: how will images get put in the store, can users build/sell them like on Amazon, and are there specific store images that the team need to create?
[21:07] <smoser> aubre, well, kit really can't do that. thats the thing. at least not without cooperation with the guest. i was told that it would work if you specified /dev/sdh.
[21:07] <aubre> smoser: give it a try. Maybe I'm doing something wrong.
[21:08] <incorrect> oh crud
[21:08] <smoser> my euc is busted right now :-( not letting me run instances.
[21:08] <aubre> smoser: doh :(
[21:08] <kirkland> aubre: you should talk to niemeyer about that
[21:09] <aubre> kirkland: will do
[21:09] <aubre> kirkland: tyvm
[21:09] <incorrect> weird none of the other vm's are going bang
[21:18] <incorrect> sigh
[21:27] <incorrect> oh wow i think it was acpid
[21:27] <incorrect> grr
[21:27] <incorrect> software with bugs! whatever next!
[21:34] <aubre> hmm, does anyone remember that command that shows all the IPs you are using that is preferred to ipconfig ?
[21:34] <aubre> I mean ifconfig?
[21:34] <mathiaz> aubre: ip addr show
[21:34] <aubre> mathiaz: ty
[21:36] <aubre> mathiaz: Ok, I have a VM I can ssh into at a public IP address, and it lets me in, but when I do a ifconfig or a ip addr show it doesn't show that IP address, it only shows a 172.19.1 address. So when I go to start apache2, it gets confused.
[21:36] <mathiaz> aubre: yes - that's normal.
[21:36] <mathiaz> aubre: VM don't know about their public IPs
[21:37] <mathiaz> aubre: they only know about their private IPs
[21:37] <aubre> mathiaz: ok, how can I tell an apache2 setup to respond?
[21:37] <qman__> aubre, configure apache to listen on the private IP
[21:37] <mathiaz> aubre: public -> private mapping is done on the CC
[21:37] <aubre> qman__: ok
[21:37] <aubre> mathiaz: ok
[21:37] <aubre> I love this - I learm more every day :)
[21:38] <qman__> the conversion is handled entirely by the router, or in this case, the host OS
[21:38] <qman__> it's the same as in NAT networking setups
[21:46] <phoenixz> sudo fuser -vki /var/lib/dpkg/lock;sudo dpkg --configure -a    gives me this : http://pastebin.com/m46588684     How can I fix this?
[21:53] <aubre> qman__: are UEC VMs not pingable?
[21:54] <aubre> qman__: I am getting the feeling my VMs aren't visible to the outside world for some reason, I can connect to them from the CC
[21:55] <aubre> hmm, I just tried to ping one and got no response
[21:55] <aubre> I can ping it from the front-end
[22:00] <jwindle> Is there anyone around that might be able to help with a networking interface bonding issue that has come up since using karmic that did not exsist in jaunty with the same configuration?
[22:04] <jwindle> It seems like a problem with upstart trying to configure the bonding interface to soon durning the boot process. I continually get "bonding: bond0: Warning: Found an uninitialized port" from dmesg until I run /etc/init.d/networking restart. Networking doesn't work at all until this is done either.
[22:20] <phoenixz> sudo fuser -vki /var/lib/dpkg/lock;sudo dpkg --configure -a    gives me this : http://pastebin.com/m46588684     How can I fix this?
[22:20] <phoenixz> This happened right after I did a software upgrade over SSH and the connection got severed while DPKG was showing me some text user interface
[22:21] <zoopster> mathiaz: list for eucalyptus sru's https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bugs?field.milestone%3Alist=12716
[22:43] <mathiaz> kirkland: why is there a patch directory in eucalyptus?
[22:43] <kirkland> mathiaz: great question
[22:43] <mathiaz> kirkland: it seems that some patches were directly applied to the source code?
[22:43] <kirkland> mathiaz: i have already filed a spec on this
[22:44] <kirkland> mathiaz: that we need to clean up the merging of eucalyptus
[22:44] <kirkland> mathiaz: and reduce the diff between us and upstream
[22:44] <mathiaz> kirkland: ok - so what's the current practice?
[22:44] <mathiaz> kirkland: patch directly or quilt?
[22:44] <kirkland> mathiaz: i think the patches dir is deprecated
[22:44] <mathiaz> kirkland: ok
[22:44] <kirkland> mathiaz: i have been modifying the source directly, since it's under rev control
[22:45] <mathiaz> kirkland: right - do you keep track of which patch have been cherrypicked from upstrea>
[22:45] <mathiaz> kirkland: ?
[22:45] <kirkland> mathiaz: i suspect that the patches predated the bzr branch
[22:45] <kirkland> mathiaz: what do you mean?
[22:45] <mathiaz> kirkland: well - I'm gonna cherry pick version 933 from upstream
[22:46] <mathiaz> kirkland: so I won't use bzr merge to do that
[22:47] <kirkland> mathiaz: right
[22:47] <kirkland> mathiaz: did you try the cherrypick command?
[22:47] <mathiaz> kirkland: well - bzr doesn't support cherrypick
[22:47] <kirkland> mathiaz: so i see...
[22:48] <kirkland> mathiaz: hmm, well, i usually just create a diff from upstream for that one revision
[22:48] <kirkland> mathiaz: try to apply it (hopefully it's clean)
[22:48] <kirkland> mathiaz: then, in the changelog, i quote the other commit message and id
[22:48] <mathiaz> kirkland: right - I just used bzr diff --old ../upstream ---new ../upstream -r 932..933 | bzr patch
[22:49] <mathiaz> kirkland: ok - so you add the upstream revno in the changelog entry
[22:50] <kirkland> mathiaz: yeah, for posterity
[22:56] <mathiaz> kirkland: I've pushed bzr branches for relevant bug fixes in eucalyptus
[22:57] <mathiaz> kirkland: https://bugs.launchpad.net/ubuntu/karmic/+source/eucalyptus/
[22:57] <mathiaz> kirkland: working from this list^^
[22:57] <kirkland> mathiaz: cool.  ready to roll an upload for -proposed?
[22:57] <mathiaz> kirkland: I'm going to create merge request so that you can review them
[22:57] <FirstSgt> is there a cool web-ui for configuring ubuntu services, i heard webmin is outdated/doesn't-work with debian-like systems
[22:57] <kirkland> mathiaz: okay
[22:57] <mathiaz> kirkland: well - some more work needs to be done to prepare the SRU in the bug
[22:58] <FirstSgt> I am trying to setup pptp, and can't remember where my users are suppose to go
[22:58] <FirstSgt> I thoguht it was in /etc/pptp.conf
[22:59] <FirstSgt> hmm, found chap-secrets
[23:00] <kirkland> mathiaz: the test instructions and such?
[23:00] <mathiaz> kirkland: yes
[23:00] <FirstSgt> there we go... starting to love ubuntu as a server too... debian is almost 100% adios
[23:01] <mathiaz> kirkland: although some of the bugs may be difficult to reproduce - bug 454405
[23:01] <kirkland> mathiaz: https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/444352
[23:01] <kirkland> mathiaz: i think we need to SRU fix that one
[23:02] <kirkland> mathiaz: it's won'tfix right now ... i think it should be promoted;  i'm seeing it a lot now
[23:05] <mathiaz> kirkland: right - I've seen it too in my testing
[23:05] <mathiaz> kirkland: hopefully a fix can be found :)
[23:06] <mathiaz> kirkland: I'm also going to create karmic branch
[23:06] <kirkland> mathiaz: great
[23:06] <mathiaz> kirkland: from bzr+ssh://bazaar.launchpad.net/~ubuntu-core-dev/eucalyptus/ubuntu/
[23:11] <phoenixz> Anybody who could tell me how to solve this one? I have 4 ubuntu servers stuck in this.. sudo fuser -vki /var/lib/dpkg/lock;sudo dpkg --configure -a    gives me this : http://pastebin.com/m46588684     How can I fix this?
[23:14] <aubre> what would cause my UEC instances to not be visible to the outside world?  they can see the outside world fine
[23:14] <crohakon> is there an application that will show you real time who is accessing your web server and ftp server?
[23:16] <aubre> I can't ping them, and I can't access their web services
[23:16] <aubre> and they are using public IP addresses
[23:17] <aubre> could it be some sort of firewall problem on the front-end?
[23:18] <FirstSgt> how can i diagnose pptp to see if its even hearing the connection?
[23:18] <crohakon> aubre; What is going on?
[23:19] <crohakon> aubre; Are they hosting the web server on a residential account?
[23:19] <crohakon> aubre; I know many ISPs (including my own) block incoming ports like 80, 8080, 21, etc... to stop people from running servers.
[23:20] <aubre> crohakon: I think it is an issue with euca-authorize
[23:20] <crohakon> the solution is to have the servers listen on other ports... for example my little play ground at home for web site design and such http://xxx.xxx.xxx.xxx:90/
[23:22] <kirkland> mathiaz: okay, so you're just looking for me to spot check your eucalyptus changes, or what?
[23:22] <FirstSgt> has anyone else configured a pptpd on a ubuntu server?
[23:23] <aubre> port 80 is not blocked on this subnet
[23:23] <mathiaz> kirkland: yes - if you could have a look at them and ack them that would be great.
[23:24] <mathiaz> kirkland: I'll do the actual merge once you've ACK'ed them
[23:24] <mathiaz> aubre: are your routes set correclty?
[23:24] <kirkland> mathiaz: postconf -# ... that's a new one on me :-)   /me just read the manpage
[23:25] <aubre> mathiaz: : I believe so - do I need to do a seperate euca-authorize for port 80 ? or the ping port?
[23:25] <kirkland> mathiaz: [ -x /usr/sbin/invoke-rc.d ] && INIT="invoke-rc.d postfix" ... i find this strange, as you test for the executability of a particular path, and then call the invoke-rc.d from the $PATH
[23:26] <kirkland> mathiaz: i'd expect you to either check the output of which, and call the one in $PATH
[23:26] <kirkland> mathiaz: or to test the executability of a full path, and call that full path
[23:26] <mathiaz> kirkland: right - makes more sense - I probably just copied'n paste from the postfix maintainer script
[23:26] <kirkland> mathiaz: cool
[23:26] <mathiaz> aubre: yes - you need to do a specific euca-authorize for port 80
[23:27] <mathiaz> aubre: and allow for ICMP as well
[23:27] <aubre> mathiaz: ok This might need to be in the docs somewhere.
[23:27] <aubre> mathiaz: thanks
[23:27] <kirkland> mathiaz: rest of https://code.edge.launchpad.net/~mathiaz/eucalyptus/k-fix-email-config/+merge/14328 looks fine
[23:27] <kirkland> mathiaz: i left a comment in the merge notes
[23:28] <mathiaz> kirkland: great - thanks.
[23:31] <kirkland> mathiaz: https://code.edge.launchpad.net/~mathiaz/eucalyptus/k-fix-network-index-values/+merge/14329 looks fine
[23:31] <kirkland> mathiaz: thanks for mentioning the revno in the merge
[23:31] <kirkland> mathiaz: in the changelog, i mean
[23:33] <qman__> oh no
[23:34] <qman__> something happened on my file server
[23:34] <qman__> some files disappeared and I'm getting read-only filesystem errors on a filesystem that is clearly mounted rw
[23:34] <kirkland> mathiaz: https://code.edge.launchpad.net/~mathiaz/eucalyptus/k-eucalyptus-conf-support/+merge/14332  <---- \o/  looking forward to seeing that one in action!
[23:38] <qman__> mdstat is showing two failed disks
[23:39] <FirstSgt> my netstat shows tcp        0      0 rico.local:1723         68-118-209-54.dhc:50837 TIME_WAIT
[23:39] <FirstSgt> so I know its getting by the router
[23:39] <FirstSgt> dmesg doesn't show anything pertenant
[23:41] <kirkland> mathiaz: regarding https://code.edge.launchpad.net/~mathiaz/eucalyptus/k-keep-network-state-on-restart/+merge/14331 ....
[23:45] <gp> hi guys pl help my server has run out of space (ec2 instance ) ......i created a new partition and mapped home folder to it but its NOT working
[23:45] <gp> pl have look my fstab ->>>>>>>>>>>> http://paste.ubuntu.com/308024/
[23:47] <gp> earlier it was mapped to mnt
[23:47] <gp> "/dev/sda2                                       /mnt            ext3    defaults        0       0"
[23:47] <mushroomblue> anyone know how to inject a stub for a package in apt?
[23:48] <gp> i copied to home folder to /mnt   and then renamed the home folder
[23:48] <mushroomblue> trying to build FreeNX on PPC, and the nxagent source file creates a package called nxagent-source instead.
[23:49] <gp> changed "/dev/sda2 /mnt ext3 defaults 0 0"  to "/dev/sda2                                      /home            ext3    nodev,nosuid    0       2"
[23:49] <gp> but fstab is not mounting it
[23:50] <gp> pl in the name of Gaint panda pl help me
[23:50] <FirstSgt> my vista machine gets caught on verrifying username and password
[23:52] <gp> its national emergency pl help me
[23:56] <kirkland> anyone here running Karmic and qemu-kvm wants to help a guy out and test something simple?