[00:00] <mgpcoe> trimeta: Pop open another terminal window and tail -f /var/log/syslog
[00:00] <Ninjix> what virt software are you using?
[00:00] <Roath> windows virtual PC
[00:00] <trimeta> mgpcoe: Oh, wait, I think I know what's happening: I've put public_html in peoples' home directories, but all home directories are 700.
[00:01] <trimeta> Well, I wanted the security of 700, but it probably isn't an option here; oh well.
[00:01] <mgpcoe> trimeta: :) That's probably at least part of it.
[00:01] <Ninjix> Roath: strange, I would assume the default vga 8-bit mode would work
[00:02] <Roath> well, after an install with all default, the display gets all garbled, like long lines going from top to bottom, i can see something is happening on the screen but not what
[00:02] <mgpcoe> trimeta: I know there's a way to set Apache to spawn its processes under the UID of the file owner; that will probably solve part of the problem.
[00:02] <Roath> as if it cant render in that bit depth
[00:04] <Roath> default as far as i know is 24, but i read somewhere that windows virtual pc only supports 16 when running linux on it
[00:04] <trimeta> Yea, that was it. I knew it was probably something stupid, but I hadn't thought it would be that stupid.
[00:04] <Ninjix> Roath:
[00:04] <mgpcoe> trimeta: I wouldn't sweat it; I know I've done worse.
[00:04] <Ninjix> Roath: here are some other vga boot mode codes: http://www.pendrivelinux.com/vga-boot-modes-to-set-screen-resolution/
[00:05] <Roath> they dont work, i keep getting the same "popup" called Boot loader
[00:05] <trimeta> Anyway, thanks.
[00:06] <Ninjix> is that after you hit Tab and exit the boot options?
[00:08] <Roath> tab does nothing
[00:08] <Ninjix> Roath: "Boot loader" sounds like Grub wanting more command arguments
[00:09] <Roath> Ninjix, i've never needed to do this, so dont know what to do, which is why i ask so much
[00:09] <Ninjix> Roath: are you still trying to boot from the ISO or have you gotten 9.10 server installed already?
[00:09] <Ninjix> Roath: no worries
[00:09] <Roath> no install, im on the screen where i can choose to install and type in boot options
[00:09] <Roath> on the iso
[00:09] <Roath> brb phone
[00:09] <Ninjix> ok
[00:11] <Roath> back
[00:12] <Ninjix> Roath: I'm testing on a spare machine
[00:12] <Roath> ok, thanks
[00:14] <Roath> Ninjix, this here explains what i used to do last time (thought that was a long time ago) where it had a VGA option during install: http://www.cosky.com/installing_ubuntu_lamp_server_within_microsoft_virtual_pc
[00:15] <Ninjix> Roath: I'm able to hit the tab button after the syslinux menu appears
[00:15] <Ninjix> then it let's me edit the "install" boot arguments
[00:16] <Ninjix> but the screen reverts to standard VGA when the Debian/Ubuntu installer launches
[00:16] <Roath>  one moment
[00:16] <Roath> hit tab on what menu?
[00:17] <Ninjix> the installer boot menu
[00:17] <Ninjix> with the Ubuntu logo
[00:17] <Ninjix> choose "Install Ubuntu Server"
[00:17] <Ninjix> the hit Tab instead of enter
[00:17] <Roath> aah, i never needed to initiate the install first
[00:18] <Ninjix> erase the -- at the end
[00:18] <Ninjix> and replace with your vga= mode of choice
[00:18] <Roath> top, tab does not work
[00:18] <Roath> nope*
[00:18] <Ninjix> perhaps your Virt PC is intercepting it?
[00:19] <Roath> probably, but removing the two -- made it work
[00:19] <Ninjix> :)
[00:19] <Roath> now i hope the vga setting was the one that will make it work
[00:20] <jmarsden> Roath: vga=ask   would be one way to see what the virtual screen has available...
[00:20] <Ninjix> correct
[00:20] <Roath> during install?
[00:20] <jmarsden> On the boot line, where you removed the --
[00:20] <Ninjix> yes
[00:21] <Ninjix> I'm not sure if syslinux/isolinux supports all of the extra framebuffer modes
[00:21] <Ninjix> my spare laptop has one of those wacky Intel video chipsets
[00:22] <jmarsden> BTW, what virtual machine software is this?  I use virtualbox-ose and have never had an issue with this here...
[00:23] <Roath>  the Windows Virtual PC that runs on win7
[00:24]  * mgpcoe grumbles something unsavoury about postfix and dovecot
[00:25] <jmarsden> Interesting.  I thought that was only intended to let you run Win XP.  If it continues to give you issues, look at virtualbox instead :)
[00:30] <Ninjix> yeah. virtualbox is a great alternative
[00:30] <Ninjix> too bad it's now part of Oracle/Sun
[00:30] <Roath> jmarsden, it can run anything like more virtual machine software, its just that its XP mode can embed itself in the system, so i can launch eg. IE6 as if it was a native app in win7
[00:31] <Roath> will try virtualbox if it still ends up garbled when booting after install
[00:31] <Roath> up... off to virtualbox
[00:32] <Roath> yup*
[01:29] <MTecknology> that survey was fun :)
[01:42] <twb> I like KVM because it's endorsed by Linux and Ubuntu, and because qemu has curses and serial I/O (which is much nicer than VNC or SDL for server VMs).
[01:42] <twb> And I don't trust virtualbox because they have a non-free version, which shows a... lack of commitment to FOSS.
[01:45] <MTecknology> twb: I like vbox because it's a fast and easy deployment. I want to play with kvm someday but right now I have yet to own a bios that will let me use kvm
[01:46] <twb> MTecknology: do you deploy via apt-get, or using third-party (i.e. not .deb) packages?
[01:47] <MTecknology> apt
[01:47] <twb> Cool.
[01:47] <MTecknology> If it's not in apt I tend to not bother
[01:47] <twb> I agree, but you get some idiots in here :-)
[01:47] <MTecknology> i do get the closed source vbox though...
[01:47] <twb> (BTW, you don't really need anything in the BIOS for kvm, but you need hardware virtualization support.)
[01:48] <MTecknology> I don't know why anymore. I used to use the USB support but not anymore
[01:48] <MTecknology> The BIOS needs to allow you to enable the hardware virt support though
[01:49] <twb> MTecknology: have you tried futzing around it with dmidecode or whatever?
[01:49] <MTecknology> cat /proc/cpuinfo <- I have the flags there
[01:49] <twb> There ought to be a way to just poke the necessary bit from linux, and then reboot.
[01:49] <MTecknology> I never tried that
[01:49] <twb> MTecknology: hm, if /proc/cpuinfo reports it, then surely you have it turned on?
[01:49] <MTecknology> nope
[01:49] <twb> Interesting.
[01:49] <MTecknology> some vendors are pricks like that :(
[01:50] <twb> So if you "modprobe kvm", it hangs the machine or something?
[01:50] <twb> Hm, weird.  I can modprobe kvm even on my celeron...
[01:50] <twb> Oh well.
[01:50] <MTecknology> it says it can't enable the module. I'm using a kernel that has that stripped out now though
[01:51] <twb> Oh.
[01:51] <MTecknology> twb: I custom compiled this puppy - only what I need/want is in it and no init* :)
[01:51] <MTecknology> but that's not why I can't use kvm - I'm not the only one that's stuck on that road block
[01:53] <twb> Now, *I* can't use KVM on my main box because vmware and openvz are installed there, and I'm not allowed to scrap them :-(
[01:54] <MTecknology> I'm not a particular fan of vmware anymore
[01:54] <twb> I have always hated vmware
[01:55] <MTecknology> I used to rather like it
[01:56] <MTecknology> hm... How can I make any package that is no longer in the repos get uninstalled? I disabled a repo and anything that was installed through it I want gone
[01:57] <twb> You want to purge packages from that source?
[01:58] <MTecknology> ya
[02:01] <pmatulis> MTecknology: i think you will need a script to query all installed packages and then come up with a list
[02:01] <MTecknology> pmatulis: how could I query them?
[02:02] <MTecknology> If I know how to check one, it's easy enough to check them all..
[02:03] <MTecknology> actually aptitude full-upgrade seems to be taking care of it...
[02:03] <pmatulis> MTecknology: off the top of my head:
[02:03] <pmatulis> aptitude show -v vim | grep Filename | cut -d / -f 2
[02:03] <pmatulis> main
[02:04] <MTecknology> nifty :)
[02:04] <MTecknology> of the top of my head .......
[02:04] <pmatulis> MTecknology: so need another command to list every installed package and pipe that into the above command
[02:05] <MTecknology> easy - dpkg --get-selections | grep -v deinstall | awk '{print $1}'
[02:05] <pmatulis> or
[02:06] <pmatulis> dpkg -l | cut -d ' ' -f 3
[02:06] <pmatulis> actually, need to check for install status (ii) first
[02:07] <pmatulis> dpkg -l | grep '^ii' | cut -d ' ' -f 3
[02:07] <MTecknology> for app in $(dpkg --get-selections | grep -v deinstall | awk '{print $1}'); do echo $app; aptitude show -v $app | grep Filename | cut -d / -f 2; done
[02:07] <MTecknology> when aptitude full-upgrade finishes I'll try that
[02:07] <pmatulis> ok
[02:08] <MTecknology> lol if that works then...
[02:08] <MTecknology>  ^^^ > packages; grep repo packages
[02:09] <MTecknology> I would like to trim out a lot of the installed packages though... for a few various reasons
[02:10] <twb> You should be able to simply use ~V and ~A patterns in aptitude
[02:10] <twb> (Sorry, I was afk.)
[02:11] <twb> http://algebraicthunk.net/~dburrows/projects/aptitude/doc/en/ch02s03s05.html
[02:11] <twb> e.g. ~VCURRENT ~Ahardy-backports
[02:12] <MTecknology> nice page
[02:12] <twb> Where "hardy-backports" is the relevant value from the Release file of the repo in question.  You may need to temporarily reactivate the deb entry in sources.list and apt-get update.
[02:12] <twb> MTecknology: it's just the aptitude manual
[02:12] <MTecknology> this is a whole stinking crap load of updates considering 9.10 isn't that old....
[02:12] <MTecknology> twb: I've never seen it - and I like it a lot
[02:13] <twb> MTecknology: yup, it's one reason I like aptitude
[02:13] <twb> It's a lot faster to use limits like that than to chain together grep-dctrl and apt-cache and friends.
[02:13] <MTecknology> please tell me I'm not on my own system again.....
[02:14] <MTecknology> my drive isn't spinning...
[02:14] <MTecknology> wtf is going on :S
[02:14] <MTecknology> I love aptitude - now I have more reasons
[02:15] <MTecknology> apt-get is nice but I've developed a fondness of aptitude
[02:16] <MTecknology> Current status: 0 updates [-258].
[02:16] <MTecknology> Lovely... that's a lot of updates
[02:20] <TheGreatToilet> How sexy is my ride?
[02:20] <TheGreatToilet> http://www.youtube.com/watch?v=uEO2eRw4y5Y
[02:22] <twb> TheGreatToilet: plonk
[02:25] <MTecknology> W: Duplicate sources.list entry http://www.ksplice.com karmic/ksplice Packages (/var/lib/apt/lists/w
[02:25] <MTecknology> ww.ksplice.com_apt_dists_karmic_ksplice_binary-amd64_Packages)
[02:25] <twb> MTecknology: haha, ksplice
[02:25] <MTecknology> twb: umm.. that only shows up once in sources.list..
[02:25] <MTecknology> hu?
[02:25] <twb> MTecknology: check your sources.list.d/
[02:26] <MTecknology> heh... ranning aptitude update again fixed it
[02:27] <MTecknology> I moved from jaunty to karmic on the sources.list line but they seem to be the exact same package so no updates
[02:27] <MTecknology> twb: you don't like ksplice?
[02:28] <twb> MTecknology: I'm reluctant to allow anything from foo.com to taint my kernel
[02:29] <TangentCollision> hello
[02:29] <twb> I'm also dubious as to the utility and reliability of "upgrade your kernel without rebooting... as long as no data structures have changed."
[02:29] <TangentCollision> I just installed imagemagick and I don't know where it is installed to...context of expublish
[02:29] <TangentCollision> ez
[02:29] <twb> MTecknology: btw, I see from http://www.ksplice.com/apt/dists/karmic/ksplice/binary-i386/Release that the Archive: is karmic, so you can't match its packages using ~A
[02:30] <twb> TangentCollision: dpkg -S imagemagick
[02:31] <TangentCollision> thanks
[02:32] <MTecknology> twb: I haven't used ksplice long enough to know what I think but I've been wanting to test it
[02:32] <MTecknology> I just like the thought of security updates and bug fixes being applied without reboot
[02:32] <twb> MTecknology: as long as it isn't on a production server :-)
[02:33] <MTecknology> it is
[02:33] <MTecknology> worst case; I need to reboot the thing and remove it
[02:33] <MTecknology> worstest case; I'm boned
[02:45] <HFSPLUS> !ops
[02:46] <HFSPLUS> !ops| MBTA IS GOD
[02:46] <HFSPLUS> !ops| Come on lazy ops ban me
[02:48] <HFSPLUS> !ops| Come on lazy ops ban me
[02:48] <HFSPLUS> !ops| twb
[03:20] <MTecknology> twb: so when my server is an hour away; only 22/80 is allowed in; I killed any ssh access; and nobody is around for me to give my password to; ........ any suggestions other than cry? :P
[03:22] <pmatulis> MTecknology: you disabled SSH access?  why?
[03:22] <twb> pmatulis: presumably accidentally
[03:23] <twb> MTecknology: I think you're hosed
[03:23] <twb> MTecknology: you need some waldos to reboot it
[03:23] <twb> MTecknology: unless it's a VPS...
[03:30] <MTecknology> pmatulis: AllowGroups sshlogin
[03:30] <MTecknology> pmatulis: I forgot to create the group and add myself to it
[03:32] <pmatulis> MTecknology: and you restarted the daemon?
[03:33] <MTecknology> yup.....
[03:33] <jmarsden> MTecknology: Only an hour away?  Could be worse... if it had been 3000 miles away you'd *really* be in trouble :)
[03:34] <jmarsden> I think you're stuck until you can get you or someone else at a console session of that machine.
[03:35] <jmarsden> well, if you run some really buggy webapp on port 80, you could try breaking into it... but it's probably faster to drive for an hour than to do that.
[03:37] <Hammit> i'd like to be able to restrict cpu and memory of a vm...any ideas?
[03:38] <jmarsden> VM memory limits are usually set when you create the VM...
[03:38] <Hammit> oh, that's right
[03:38] <Hammit> forgot temporarily
[03:38] <Hammit> cpu was the main one tho...
[03:38] <jmarsden> CPU... nice the relevant processes?  use ulimit ?  What have you already tried?
[03:39] <Hammit> i just threw mem in while i was at it :)
[03:39] <jmarsden> Run it on a machine with a multicore CPU and set it to only use one core?
[03:39] <Hammit> well, i haven't setup the server yet...but was hoping to allocate part of a core
[03:40] <Hammit> for VPS stuff
[03:40] <twb> jmarsden: some servers I babysit are in .za and .il o_O
[03:40] <Hammit> i saw OpenVZ and noticed that it can restrict cpu to that level
[03:40] <MTecknology> jmarsden: The only web app right now is It Works!
[03:41] <jmarsden> twb: Yup, I've done that from CA for servers in the UK and in Mali and India, at one point... not in my current job though.
[03:41] <MTecknology> :P
[03:41] <jmarsden> MTecknology: So... why you still on IRC... get in your car and drive to the datacenter ... right?
[03:41] <Hammit> jmarsden: doesn't look like kvm handles this sort of thing atm
[03:41] <Hammit> ?
[03:41] <MTecknology> jmarsden: datacenter?
[03:41] <jmarsden> MTecknology:  Where they host your server.
[03:41] <MTecknology> under the steps counts as a data center?
[03:42] <jmarsden> I guess... :)
[03:42] <MTecknology> :P
[03:42] <MTecknology> I have two systems that run on xen... linodes (those are my only two production systems) the rest are backup, dev, etc
[03:43] <Hammit> can xen handle cpu restriction?
[03:44] <Hammit> as in restricted to user specified mhz?
[03:44] <MTecknology> When we grow big enough I'll donate the systems to some prick I hate and let him kill himself trying to make them function correctly... then get real systems and put them in a real data center with real cloud computing technology
[03:45] <jmarsden> MTecknology: Maybe you should move "the rest" to Linode too, so they have console access etc etc foe times like this?
[03:45] <jmarsden> *for
[03:45] <MTecknology> idk.. - this is the cpuinfo output - http://paste.ubuntu.com/336294/
[03:46] <MTecknology> jmarsden: Linode really doesn't offer much in the way of >1TB data + >8GB RAM
[03:47] <jmarsden> And you keep that "under the steps"? :)  I hope the steps have a good lock on them...
[03:47] <MTecknology> it works great for the two web services we provide. Given the money it would be nice to move the dev system to linode - but we can't afford it now
[03:48] <MTecknology> the area they exist is at the security level that I consider 'well enough'  - I'd like something better but then we get back to that price issue
[03:49] <jmarsden> Understood.  Since you already have the hardware, finding a (relatively) cheap local colo that has serial console access and remote reboot might be worth trying.
[03:50] <jmarsden> But for now... get in the car :)
[03:51] <MTecknology> If I had an extra $40,000 to blow; I would get 7 awesome systems with real raid an dput them in a datacenter with all the great "cloud" technology and guarantee some uptime. Then try to expand very fast to help ensure that I can afford it when the money  dries up
[03:51] <MTecknology> for now I'm going to just wait for somebody to show up :P - I was there once today
[03:52] <MTecknology> I wonder how much it would cost to do that in reality.....
[03:52] <MTecknology> I hate dealing with old crappy hardware that needs to maintain reliability in order for me to
[03:53] <MTecknology> I'm trying to build a business
[04:44] <Leszeck> Noob question: apt-get install python-vm-builder tells me it couldn't find the package.  What do I need to change to get it to view it?  Ubuntu 9.10.
[04:45] <Leszeck>  /view/install/
[04:45] <MTecknology> !info python-vm-builder
[04:46] <MTecknology> Leszeck: enable your universe repository; update apt; install it
[04:46] <Leszeck> Sorry, how does one do that?  Is there a tutorial someplace?  I'm more used to centos and yum.
[04:48] <MTecknology> Leszeck: the way I do it is.....
[04:48] <MTecknology> actually maybe ubottu can say better
[04:48] <MTecknology> !sources
[04:48] <jtaji> did something change in 9.10? upgrade my linode, I have no getty on hvc0, I do have gettys on tty1-6, although both /etc/inittab and /etc/event.disabled show the tty1-6 should be disabled
[04:49] <MTecknology> Leszeck: follow those links to enable the universe repository; then apt-get update && apt-get install python-vm-builder
[04:49] <MTecknology> hm?
[04:50] <Leszeck> Thanks, I'll try it out.
[04:52] <Leszeck> Outstanding; that did it.  Thanks everyone.
[04:52] <MTecknology> jtaji: Linode guys are irc.oftc.net #linode - they can tell you a lot easier
[04:53] <jtaji> yeah I did ask their first
[04:54] <MTecknology> I have no idea - I don't use their getty interface much
[04:54] <MTecknology> sorry
[04:56] <jtaji> no prob.. it's more of a general upstart question... tty1-6 files have been moved from /etc/event.d/ to /etc/event.disabled/ yet they still start... odd
[04:56] <MTecknology> I know the kernel is handled by xen which could potentially affect that
[04:57] <Leszeck> Does anyone know what I have to install on a 9.10 EC2 instances to get ec2-ami-tools?  I've got an updated universe and multiverse, but it can't find the ec2-ami-tools package.
[04:58] <jtaji> MTecknology: indeed.. the one reply I've received so far thinks my kernel should be good...
[04:59] <jmarsden> Leszeck: per http://packages.ubuntu.com/lucid/ec2-ami-tools they are supposed to be in multiverse
[05:02] <jmarsden> Leszeck: You are running Lucid, right?
[05:02] <jmarsden> Wait, you said 9.10.  ec2-ami-tools does not exist in Karmic, as far as I know, only in Lucid.
[05:03] <Leszeck> Got it working; I'm running karmic.  When I copied the multiverse stuff from the webpage I googled to my sources.list, it had "hardy" instead of "karmic".
[05:03] <jmarsden> OK.
[05:03] <Leszeck> apt-get got it; I'll let you know in a sec if it works.
[05:04] <Leszeck> cli gives me usage; now trying to bundle an image.
[05:04] <jmarsden> Oh, my mistake, they do exist in earlier version of Ubuntu too.
[05:04] <Leszeck> Working like a champ.  Thanks, all.
[05:04] <jmarsden> :)
[05:13] <jtaji> figured it out... it's moved again from /etc/event.d/ to /etc/init/
[05:34] <spoonybooncoon> Hi all, i have a ubuntu 8.04 LTS host with server 2.0.2 installed, i have an issue with performance, my windows guest is running amazingly slow, yet this host machine used to have windows on it running vmware server, and it was fine... Any performance tweaks i should know?
[05:37] <jmarsden> spoonybooncoon: How much RAM is allocated to the Windows guest?  Is that RAM really available for it to use, or is it swapping?  Do general performance checks on the server as a whole and see what the bottleneck seems to be?
[05:38] <spoonybooncoon> The windows guest has 2024MB ram, on a system with 4GB installed, and linux-headers-server with PAE...
[05:38] <spoonybooncoon> And, no it's not swapping... According to TOP
[05:39] <jmarsden> So why is it slow... disk access?  Using 100% CPU?  network access???
[05:39] <jmarsden> Try to narrow it down, the fact that it is Windows or a VM doesn't really matter at this point, just determing what the bottlenecked resource is.
[05:40] <spoonybooncoon> The load average goes way up when it's doing anything.
[05:40] <jmarsden> That could be several things... when it is "doing anything", does top show a lot of CPU usage?  Does iostat show a lot of disk activity?  etc.
[05:41] <spoonybooncoon> I'll have a look
[05:43] <spoonybooncoon> It's weird because i'm sure it was alot more efficient when i had windows as the host os, running the same vm.
[05:45] <jmarsden> That's no way to troubleshoot a performance issue :)  Get read information instead...
[05:45] <jmarsden> *real
[05:46] <spoonybooncoon> Ok
[05:48] <spoonybooncoon> what do you make of this
[05:48] <spoonybooncoon> this is with it not doing much at all
[05:48] <spoonybooncoon> avg-cpu:  %user   %nice %system %iowait  %steal   %idle
[05:48] <spoonybooncoon>            1.12    0.00   11.21   34.46    0.00   53.21
[05:49] <spoonybooncoon> and it under load
[05:49] <spoonybooncoon> avg-cpu:  %user   %nice %system %iowait  %steal   %idle
[05:49] <spoonybooncoon>            1.12    0.00   11.20   34.64    0.00   53.03
[05:49] <spoonybooncoon> Device:            tps   Blk_read/s   Blk_wrtn/s   Blk_read   Blk_wrtn
[05:49] <spoonybooncoon> sda             127.23      1566.00      4563.68   35232588  102675360
[05:50] <jmarsden> Looks IO bound... but don't paste that much at once to the channel, use pastebin
[05:51] <jmarsden> Could something else (in Linux) be causing significant disk i/o ?
[05:51] <jmarsden> 34% iowait when it is "not doing much at all" seems unusual, to put it mildly.
[05:55] <spoonybooncoon> It's a fresh install.
[05:56] <jmarsden> That doesn't answer the question.  It might be doing an update of the locate database, it might be downloading and applying a pile of security updates...
[05:57] <jmarsden> Use the Ubuntu tools you already know to determine what the server itself is doing...
[05:57] <jmarsden> Then look at the VM and Windows after that.
[05:58] <spoonybooncoon> how can i tell if it's doing a locate update?
[05:58] <spoonybooncoon> and how long does that generally take?
[06:01] <jmarsden> Use top to see what processes are running and using CPU... might go on for a few minutes, shouldn't be hours.
[06:01] <jmarsden> Worst case, shut down the Windows VM and then look at top and iostat and vmstat etc and see if there is still a lot going on...
[06:01] <spoonybooncoon> I think it might have been doing that, because i ran iostat directly after installing iostat
[06:02] <spoonybooncoon> so it could have been updating indexes still
[06:02] <spoonybooncoon> But i've noticed my IO wait continues to rise, and it's not dropping....
[06:02] <jmarsden> CAn you kill off the VM and see if it then drops to zero?
[06:03] <spoonybooncoon> My vm is doing some rather intensive operations...
[06:03] <spoonybooncoon> IO intensive..
[06:03] <jmarsden> BTW I have a Windows 2000 VM running here and iostat shows 0% iowait when Windows is "not doing much"...
[06:04] <jmarsden> OK, so then it is probably that IO which is the bottleneck.  How did you create the virtual disk -- is a file or a separate partition?
[06:11] <spoonybooncoon> file based.
[06:16] <jmarsden> Do you have a spare drive on the server -- I'm no vmware expert, but in general one way to reduce I/O overhead can be to use a real partition or LVM volume rather than a file for the virtual disk.
[06:24] <spoonybooncoon> No, i don't it's a laptop...
[06:25] <jmarsden> You are running the server version of Ubuntu on a laptop?  Really?!
[06:26] <jmarsden> So you can't display the windows VM screen on the laptop, you can only remote in over the network from another PC to see it?
[06:46] <twb> It *would* be reasonable to deploy Ubuntu Server on an old laptop when repurposing it as a router...
[06:52] <jmarsden> twb: Yes.  But it seems odd to run Windows server in a VM on it, and then do I/O intensive stuff in that VM... and the "old" laptop has 4GB RAM... I suspect we're looking at a Ubuntu desktop install, not server.
[06:53] <twb> Shrug
[06:59] <syncrondi> does anyone know an apache library package substitute for libaprutil1  that will work with php4-mysql?
[07:00] <twb> Isn't PHP4 long gone?
[07:00] <syncrondi> I wish
[07:00] <twb> It is.  Unless you're running Dapper, you can't have PHP4
[07:00] <syncrondi> Why not?
[07:01] <twb> Because later versions of Ubuntu ship with PHP5, and attempting to downgrade that is an insane nightmare
[07:01] <syncrondi> Ah. Well, I didn't think it came with php5 installed, however I recently started over and am using Lenny currently
[07:02] <twb> Lenny isn't Ubuntu
[07:02] <syncrondi> Right, I do know
[07:02] <syncrondi> But I came across the same issue
[07:03] <syncrondi> The issue is that I can get most of the php4 packages to install but libaprutil1 conflicts with php4-mysql
[07:03] <arj>  
[07:03] <twb> syncrondi: I won't help you with that, because IMO it is an extremely Wrong Thing.
[07:05] <syncrondi> twb: The script I'm trying to run was written for a shared host that was and is still running php4 and I think the coder must not be competent because when I moved this to a VPS and installed it on php5, the script stopped working. It would be a lot of effort to rewrite the script and under a deadline it isn't feasable. I thought it might be possible to just install php4 for the time being.
[07:06] <twb> syncrondi: IME it'll be more effort to get PHP4 on a recent Ubuntu or Debian that it will be to fix the script.
[07:06] <twb> I have been in exactly the same position before, IIRC on Etch.
[07:07] <syncrondi> twb: I'll take your word for it.
[07:07] <twb> It also depends how large and stupid the script is, I guess
[07:08] <syncrondi> Well, I've already sunk enough time into fixing this problem infrastructure side, that the coder is going to need to take this from here out!
[07:14] <syncrondi> thanks for the advice twb
[07:17] <`jpg> Heya, anyone running Ubuntu Enterprise Cloud here?
[07:18] <twb> !anyone
[07:22] <`jpg> Well I am only looking for a general impresssion and a maybe a few technical questions regarding the redundancy of the control services etc.
[07:22] <`jpg> I want to implement a high availabily high performance cloud.
[07:23] <`jpg> Main concerns are the redundancy of the control services.
[07:23] <Jeeves_> Morning
[07:24] <`jpg> I need to know if I can run multiple cloud controllers for a single set of nodes as to provide redunancy if the phsyical host hosting one or more of the control services fails.
[08:08] <andol> ttx: Regarding the NIS package and a "LDAP setup" as an alternative. Must be plenty of environments where Ubuntu computers are primary clients, the NIS is already in place, and there really isn't an alternative to using the NIS "client" included in the NIS package.
[08:09] <ttx> andol: right, that's what I meant by "existing environments to support"
[08:11] <andol> ttx: ok, with you then.
[08:20] <andriijas> my ubuntu-server installation has upgraded the kernel but its not using server labled kernels anymore, it automaticly switched to generic. why?
[08:21] <twb> andriijas: -server flavour disappeared on i386, IIRC
[08:22] <andriijas> twb: ah
[08:22] <andriijas> thats resaonable
[08:23] <andriijas> i've discovered that i infact have a 64bit cpu so will probably swap out the system and reinstall from the ground
[08:23] <andriijas> if theres no easy way to just reinstall all packages as 64bit instead...
[08:24] <twb> You can't switch to 64-bit without reinstalling.  Sorry.
[08:25] <andriijas> is there any easy way to download 64bit iso, extract to usb stick and install ubuntu from it?
[08:25] <Jeeves_> is it 'safe' to upgrade to Lucid yet? Or is it in non-working state? (I know it's pre-Alpha)
[08:26] <andriijas> Jeeves_: wierd qustion. upgrading to anything thats pre-alpha is on your own risk.
[08:26] <twb> Nothing is safe.
[08:27] <twb> Perhaps you should qualify your question
[08:27] <andriijas> living is not safe ;)
[08:28] <`jpg> You can install from usb using Unetbootin.
[08:31] <twb> You can install from USB using d-i
[08:31] <Jeeves_> andriijas: It's not a weird question. There is a difference between 'expect daily breakage' (which I do) or 'its completely broken, upgrading will definitly cause total breakage'
[08:31]  * twb doesn't like this newfangled crap
[08:32] <andriijas> Jeeves_: It's a wierd question because it might be working now but it could be broken any sudden day
[08:32] <twb> Jeeves_: I don't know, but you could start by asking malone (launchpad) about release-critical bugs in lucid...
[08:33] <andriijas> `jpg: https://help.ubuntu.com/community/Installation/FromUSBStick   sounds pretty trivial
[08:33] <Jeeves_> twb: There will obviously be release-critical bugs..
[08:34] <Jeeves_> But, nevermind, I'll try and see..
[08:34] <andriijas> Jeeves_: if you just want to play with it virtualbox is nice
[08:35] <twb> Jeeves_: but some of those RC bugs will not affect you
[09:10] <soren> Nafallo: Well, ospf and rip at least work.
[09:10] <Nafallo> soren: let me find the backtrace.
[09:13] <Nafallo> 2009/12/05 22:01:49 BGP: Assertion `ret != ((void *)0)' failed in file bgp_attr.c, line 647, function bgp_attr_unintern
[09:13] <Nafallo> full backtrace in PM
[09:22] <jpds> Hmm, quagge.
[10:58] <Jeeves_> https://bugs.launchpad.net/ubuntu/+source/nfs-utils/+bug/489499
[10:58] <Jeeves_> Who can we bug about that bug? :)
[11:07] <shriedi> what is vps
[11:07] <_ruben> virtual private server
[11:08] <Jeeves_> kees: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/487010
[11:09] <Jeeves_> Isn't that enough of a security issue?
[11:10] <twb> You'd think that if iptables is around, iptables6 should be, too.
[11:11] <twb> i.e. both or neither
[11:11] <twb> (Unless you know a-priori that only on will be used...)
[11:12] <Jeeves_> twb: the binary ip6tables is available. The needed modules aren't
[11:13] <twb> Jeeves_: I meant the modules.
[11:14] <Jeeves_> Ah ok :)
[11:15] <_ruben> who uses ipv6 anyways :P
[11:16] <Jeeves_>  /kick _ruben
[11:16] <Jeeves_> ;)
[11:16] <_ruben> hmm .. crap .. my sarcasm failed .. how'd i end up on a ipv4 freenode server?!
[11:17] <Jeeves_> :P
[11:17] <_ruben> hrm
[11:18] <_ruben> that's better
[11:28] <twb> _ruben: I was gonna migrate to native IPv6 last weekend, but I got distracted
[11:28] <twb> Tunneling over IPv4 is for losers
[11:33] <_ruben> twb: the office complex we're in is scheduled to get native ipv6 Q1 2010
[11:33] <twb> Internode has trial native IPv6 for its ADSL customers, yaaaay
[11:33] <_ruben> and tunneling ipv6 over ipv4 beats using ipv4+nat
[11:34] <twb> Ugh, tell me about it.
[11:34] <_ruben> i'd switch back to xs4all (big dutch isp), if only they wouldnt be so expensive, they've been offering tunnels for years, and also started deploying native to dsl users
[11:34] <twb> My boss was trying to get NAT working for his OpenVZ VEs tonight
[11:35] <twb> _ruben: yeah, I know of them.  IIRC Russell Coker was telling me how they were a nice employer.
[11:36] <twb> I think he (my boss, not Coker) managed to get triangle routing
[11:36] <_ruben> triangle routing ftw! :)
[11:37] <twb> Here, I pay about an extra AUD10/mo (about 15% of the total) to go with Internode instead of <mainstream idiot ISP>.
[11:38] <twb> IMO it's worth it just for Debian mirror and not having to pretend to edit the registry if I ever need to call their tech support.
[11:38] <yann2> hi
[11:39] <twb> And no bullshit like them blocking ports below 1024 or a bad transparent http proxy, or those laughable default DNSs some ISPs seem to have.
[11:40] <_ruben> switching from current "crap" isp (telfort) to xs4all would mean E30/mo -> E60/mo for same speed (20mbit adsl)
[11:40] <arj> telfort is fine here
[11:41] <_ruben> telfort just works, usualy .. that's as far as i'll go :)
[11:41] <arj> and the people that can read and or talk on the phone aren't mutually exclusive like tweakdsl
[11:42] <arj> and "lie at every phone call" is not a requirement for all the people that work there
[11:42] <arj> so I'm happy with it
[11:42] <_ruben> hehe
[11:42] <arj> does xs4all allow more than one ip when bridging?
[11:43] <_ruben> they did untill a few years ago .. fast-adsl had 4 ips
[12:14] <Jeeves_> _ruben: Don't expect xs4all to deliver native v6 any time soon
[12:14] <Jeeves_> I do know someone who is thinking about starting a cool ISP in .nl again
[12:15] <Jeeves_> But, a cool ISP cannot deliver under costprice :)
[12:24] <_ruben> Jeeves_: hehe
[12:25] <_ruben> speaking of cool, reducing the number of pcs here means i'll have to go turn on the heater .. frozen fingers
[13:16] <aubre> hello
[13:18] <Hammit> hi aubre
[13:21] <cemc> is there a dpkg equivalent to rpm -V ?
[13:22] <cemc> which would verify all the files in a package and show if anything changed
[13:38] <sommer_> morning
[13:42] <pmatulis> cemc: investigate package 'debsums'
[13:43] <cemc> pmatulis: thanks
[13:43] <pmatulis> !info debsums
[14:17] <spiekey> Hi!
[14:18] <alex88> hi
[14:18] <spiekey> this is maybe a little OT, but....what component is responsible if i want a Hot-Plugable SATA Disk?
[14:18] <spiekey> the Raidcontroller or the disk case or both?
[14:18] <alex88> i think the controller
[14:18] <alex88> but i'm not sure
[14:19] <alex88> i'm not so informed about these things
[14:19] <spiekey> me neither :)
[14:19] <spiekey> i am quite sure, but not 100% :)
[14:19] <alex88> :)
[14:41] <_ruben> spiekey: the controller indeed (doesnt have to be a raid controller, 'simple' sata controller could support it too)
[14:42] <spiekey> _ruben, thanks!
[14:43] <_ruben> the driver probably needs to support it too i think
[15:15] <rj175> does anyone know about ubuntu cloud?
[15:21] <MTecknology> rj175: I tried to understand what they mean when you install a "cloud" server but it's beyond me
[15:28] <smoser> rj175, people here do know "about ubuntu cloud". what particularly do you want to know?
[15:31] <ScottK> MTecknology: Step one is have ssh access to a system so you can install stuff.
[15:31] <ScottK> ;-)
[15:32] <ssd5> i am able to ping from linux to windows but not from windows to linux. destination host unreachable. am using ubuntu 9.10 server on virtualbox.
[15:45] <MTecknology> ScottK: :P
[16:12] <EsatYuce> is .bin file about Linux or Windows?
[16:13] <MTecknology> EsatYuce: *nix
[16:13] <MTecknology> chmod +x file.bin; ./file.bin
[16:13] <MTecknology> or sh ./file.bin
[16:14] <EsatYuce> Can i run this file in Linux?
[16:14] <MTecknology> read above
[16:14] <EsatYuce> it is just name
[16:15] <EsatYuce> i dont understand
[16:15] <MTecknology> !bin
[16:15] <MTecknology> hm..
[16:15] <MTecknology> first google result - http://www.cyberciti.biz/faq/howto-unix-command-run-execute-bin-files-in-linux/
[16:16] <EsatYuce> i will try that
[16:17] <mdeslaur> soren: the OATH Yubikey is available now!
[16:18] <EsatYuce> MTecknology : The terminal said me this: chmod: cannot access
[16:19] <soren> mdeslaur: You you push it and it pretends to be a USB keyboard that sends your one-time password?
[16:19] <mdeslaur> soren: yes
[16:19] <MTecknology> !details | EsatYuce
[16:19] <mdeslaur> soren: they had their own algorithm, but they now sell OATH ones too
[16:20] <soren> mdeslaur: Excellent.
[16:20] <soren> I'm just wondering if there are reasonable cases where you don't actually have USB access, but these days, I guess there isn't really.
[16:20] <EsatYuce> i have Ubuntu 9.10 i want to install Google Erath application, i downloaded GoogleEarthLinux.bin file from site. How can i run this application?
[16:20] <soren> EsatYuce: Try in #ubuntu.
[16:20] <EsatYuce> soren: how?
[16:21] <Jeeves_>  /j #ubuntu
[16:21] <soren> EsatYuce: Ask your question in #ubuntu instead of in #ubuntu-server
[16:21] <EsatYuce> ok
[16:21] <EsatYuce> thnaks
[16:21] <soren> Sure thing.
[16:24] <Ng> zul: do the current whiteboard comments on the server-lucid-canonical-application-support spec mean that drupal6 is going to languish with asterisk? :(
[16:25] <zul> Ng: yeah for now
[16:25] <Ng> that's unfortunate
[16:39] <rj175> i keep getting this error when trying to launch a instance in ubuntu cloud "in MANAGED-NOVLAN mode, priv interface 'eth1' must be a bridge, tunneling disabled"
[16:47] <jeeves_Moss> ho do I find out what program is running on a port?
[16:48] <Pici> jeeves_Moss: sudo netstat -pant
[16:49] <jeeves_Moss> Pici, how can I grep out just what's on port 80?\
[16:49] <Pici> jeeves_Moss: sudo netstat -pant | grep :80\
[16:50] <Pici> jeeves_Moss: Theres a space after that slash
[16:50] <jeeves_Moss> how can I fix this error with apache2?  http://pastebin.ca/1705701
[16:50] <jpds> Pici: I prefer netstat -ltpn personally.
[16:51] <jeeves_Moss> Pici, did you get a chance to look @ that pastebin?
[16:52] <jeeves_Moss> jpds, did you get a chance to look @ that paste bin
[16:53] <genii> jeeves_Moss: You can also do something like: sudo lsof -i:80
[16:54] <jeeves_Moss> genii, thanks
[16:55] <Pici> jeeves_Moss: your site configuration file(s) are incorrect. Did you define a NameVirtualHost before the rest of your virtualhost entries? I don't think you can do that.
[16:56] <jeeves_Moss> Pici, I think it's the flashMediaServer that's causing a headache
[16:57] <jeeves_Moss> Pici, it worked well before I started playing with subsonic
[16:58] <jeeves_Moss> Pici, what's the apache command that looks through your config files again?
[16:59] <Pici> jeeves_Moss: er. doing a apache2 reload or restart will surely do it.
[16:59] <Pici> I'm not sure if thats what you're looking for though.
[17:00] <jeeves_Moss> Pici, well, the output from that paste bin is what I get when I try to restart the server
[17:01] <Pici> jeeves_Moss: Normally that would mean that you have a problem in one of your /etc/apache2/sites-enabled/ files
[17:02]  * genii thinks about apachectl configtest
[17:02] <jeeves_Moss> Pici, that's the issue, I don't know what file is hooped and WHY
[17:03] <Pici> jeeves_Moss: try genii's command and perhaps it will tell you which file.
[17:03] <jeeves_Moss> Pici, command again?
[17:03] <Pici> jeeves_Moss: apache2ctl configtest
[17:03] <jeeves_Moss> thanks.
[17:06] <jeeves_Moss> Pici, I think it's the execwrapper we're having issues with
[17:07] <jeeves_Moss> Pici, http://pastebin.ca/1705726
[17:07] <jeeves_Moss> genii, http://pastebin.ca/1705726
[17:16] <jeeves_Moss> well, now I don't have a MySQL connector!  ARRGGHH
[17:16] <alex88> the ending "Syntax OK" it's nice..=)
[17:21] <genii> jeeves_Moss: I've not seen that particular SUEXEC issue before, but this Debian bug seems possible cause: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=395828
[17:23] <jeeves_Moss> genii, lol, and to think everything was working untill I rebooted last night!
[17:32] <DyGyTyL> so i havent used ubuntu since like 2.x but decided to use 9.10 for my server.  at one point, it was up and running even had x installed then i rebooted it for the first time and bam - it wouldnt load just 99 99 99 99 99 99 99
[17:34] <tolbrino> hey there, I have just installed Ubuntu Enterprise Cloud with one node. I ran into problems starting up instances, so I'd like to inspect the nodes logs. but I can't find information on how I can connect to the node with ssh. Do you have any advice on how to do that?
[17:35] <DyGyTyL> after some research i found out somethin goofy with my boot loader so i blew up my install and the MBR now after the ubuntu installation completes it boots up and says missing operating system or nonsystem disk or disk error replace and strike key when ready =o
[17:36] <genii> DyGyTyL: There were no 1.x 2.x or 3.x versions of Ubuntu
[17:37] <DyGyTyL> maybe it was 5 was awhile back :P
[17:37] <DyGyTyL> 2.x must've been the kernel
[17:39] <DyGyTyL> shit just versioned the whole channel by accident >_>
[17:45] <genii> DyGyTyL: Is this machine one with some builtin RAID5 controller, which you're trying to install the system onto?
[17:46] <DyGyTyL> no raid but it does have multiple(3) hard drives
[17:47]  * zul kicks launchpad
[17:48] <tolbrino> During the installation of UEC on the node there is no prompt for creating a standard user, so I assume that ssh access is key-based. But I can't find a key which allows me to access the node without asking for a password or a passphrase. Any hints?
[17:50] <Hypnoz> does anyone know how to apt-get install nis without being prompted for NIS domain, so I can script it into an install?
[17:51] <jmarsden> Hypnoz: You could pre-seed the answer for the question.
[17:52] <Hypnoz> so i can't use apt-get then
[17:52] <jmarsden> Yes, you set the answer in the question database and apt-get the package.
[17:53] <Hypnoz> what and where is a question database?
[17:55] <jmarsden> I'm trying to find a web page about it... preseeding for a whole install is discussed at https://help.ubuntu.com/9.10/installation-guide/i386/appendix-preseed.html
[17:55] <jmarsden> But that's not quite what you need...
[17:57] <jmarsden> man 7 debconf   # seems very relevant
[17:59] <jmarsden> based on that man page, the database is under /var/cache/debconf/
[17:59] <jmarsden> Looks like you can use apt-utils and dpkg-reconfigure to do what you want.
[18:21] <Hypnoz> hmm ok I'll see if I can find more info
[18:27] <Xpistos|work> I need help adding a signing key from the command line. Could someone give me a hand or point me in the correct direction.
[18:38] <Xpistos|work> my sudo add-apt-repository command says it is not found?
[18:38] <Xpistos|work> in ubuntu 9.10
[18:38] <unit3> you probably don't have the package installed then.
[18:38] <unit3> run it without sudo, and it'll tell you what package to install.
[18:42] <genii> Xpistos|work: http://packages.ubuntu.com/search?searchon=contents&keywords=add-apt-repository&mode=exactfilename&suite=karmic&arch=any
[18:49] <Xpistos|work> not there
[18:50] <billybigrigger> the package is add-apt-key isn't it?
[18:51] <Xpistos|work> I found it
[18:51] <Xpistos|work> I needed to apt-get install python-software-properties
[18:51] <Xpistos|work> and it is there now
[18:51] <Xpistos|work> thanks genii
[18:51] <genii> Xpistos|work: Yer welcome
[18:55] <moonpup> hi, do caching nameservers typically need a search path?
[18:56] <Xpistos|work> maybe one more thing. I am trying to upgrade my deluge and I have the new PPA installed but it says that files are being kept back when I do an apt-get upgrade? Could this be becuase I manually installed the package instead of using the repo?
[18:58] <Xpistos|work> sudo apt-get dist-upgrade worked however
[19:03] <ruben23> hi
[19:04] <billybigrigger> anyone here familiar with awstats, should /etc/awstats be owned by root or www-data?
[19:05] <billybigrigger> i'm having permissions errors on my .conf file, saying it can't be located in /etc/awstats, while in fact it is, tried to chown -r /etc/awstats to both root, and www-data, both give the same error
[19:06] <ruben23> http://pastebin.com/m3768ef60 using ubuntu-server
[19:09] <billybigrigger> ruben23, https://help.ubuntu.com/community/ApacheTomcat5
[19:10] <billybigrigger> shows you right there how to set the variables for tomcat to work
[19:10] <billybigrigger> actually...here's an updated one for 9.10
[19:10] <billybigrigger> https://help.ubuntu.com/9.10/serverguide/C/tomcat.html
[19:11] <billybigrigger> set JAVA_HOME in /etc/default/tomcat6
[19:16] <Xpistos|work> okay how about this one. I have an sshfs share on my laptop  from my server (both 9.10) how do I set my permissions so I can delete items out of my share? Right now when I try it says i do not have perms.
[19:17] <unit3> as long as the user you're logging in as has permissions to delete, it should work.
[19:20] <Xpistos|work> its not
[19:21] <Xpistos|work> it's not working I mean
[19:22] <Xpistos|work> If I SSH in to the server I can delete all day
[19:22] <Xpistos|work> and I can create from my sshfs, but not delete
[19:23] <unit3> that's really weird. maybe a bug in sshfs?
[19:30] <cj> has anyone else noticed that /etc/debian_version has a somewhat strange value?
[19:30] <cj> it also has a somewhat strange name, but that's beside the current point ;)
[19:31] <Xpistos|work> is anyone using kslice on their server?
[19:43] <billybigrigger> cj, billybigrigger@timmy:/etc$ cat debian_version
[19:43] <billybigrigger> squeeze/sid
[19:43] <billybigrigger> looks ok to me
[19:51] <ssd5> just now I installed nfs-kernel-server package on ubuntu server 9.10. but now its saying "Not starting NFS kernel daemon: no suppport in current kernel". o/p of uname -a 2.6.31-14-generic-pae. what should I do?
[20:47] <cj> billybigrigger: are you running sid?  karmic?  lucid?  I'd expect something similar to lsb_release -c
[20:48] <billybigrigger> karmic
[20:48] <jpds> cj: The file is taken from Debian really...
[21:11] <majuk> Hey fellas. I'm looking for a way to get around using the Windows System Policy Editor. Anyone know a good howto for building your own NTConfig.POL or something similar? Googling has failed me. I found a large file of examples (http://is.gd/5fhID), but it appears to be only the configuration for the SPE program and not the policy itself... though I could be wrong.
[21:12] <majuk> If anyone can point me in a direction, I'd appreciate it.
[21:21] <majuk> Rawr, that link is to an adm file, I guess that's why it looks like it's in the wrong format. gg
[21:22] <kirkland> cjwatson: hi there ... just coming online now
[21:22] <kirkland> cjwatson: i'm in Portland this week
[21:23] <cjwatson> kirkland: you can disregard my questions in scrollback - ttx helped me out today
[21:23] <kirkland> cjwatson: ah, okay, i was just about to start responding to those
[21:24] <cjwatson> oh, actually
[21:24] <cjwatson> 21:15 <cjwatson> kirkland: if the eucalyptus CLC and CC are on separate machines, do they each need their own SSH key?
[21:24] <cjwatson> kirkland: ^- I didn't ask ttx about that, so if you happen to know ...
[21:25] <kirkland> cjwatson: hmm, i've never really considered sharing a common host key across different machines
[21:26] <cjwatson> hmm, no, I don't think that can have been what I meant
[21:26] <kirkland> cjwatson: okay, can you clarify?
[21:26] <kirkland> cjwatson: you mean eucalyptus@CLC and eucalyptus@CC ?
[21:26] <kirkland> cjwatson: user keys, you mean?
[21:26] <cjwatson> kirkland: I believe I meant, does the CLC need its own SSH key (not host key, I mean one in ~eucalyptus/.ssh/id_rsa)
[21:26] <cjwatson> right
[21:27] <kirkland> cjwatson: ah, right; i'd think a single private key would make the most sense
[21:27] <cjwatson> well, I'm not sure I was actually contemplating sharing
[21:27] <cjwatson> if I remember my state of mind correctly, I was actually trying to discover whether the CLC needed a key at all
[21:27] <cjwatson> sharing's probably tricky, I'm not sure I fancy putting private key text in a preseed file
[21:28] <kirkland> cjwatson: definitely not
[21:28] <kirkland> cjwatson: definitely not -> putting private key text into a preseed file
[21:28] <cjwatson> the reason I ask is that the only thing that runs ssh-keygen automatically right now is eucalyptus-cc.postinst
[21:28] <cjwatson> so if the clc needs a key too, it needs a better postinst
[21:28] <kirkland> cjwatson: i *believe* that it will
[21:29] <kirkland> cjwatson: nurmi will be here later tonight; i'll ask him then, as I'm not positive
[21:30] <kirkland> cjwatson: should we start testing the ISOs soon?
[21:31] <kirkland> cjwatson: ie, are there bits landing on the server ISOs that are ready for some exercise?
[21:32] <cjwatson> sure
[21:32] <cjwatson> I was actually hoping to get the whole installer spec in for a1, but I'm not sure whether I'll quite manage it
[22:19] <davidlenwell> anyone know of an apache 2.2.14 package for karmic ?
[22:31] <jeeves_Moss> I'm having this issue with my server.  it's a v-hosted install of Apache2, and now, after a reboot, I'm getting a "SUEXEC wrapper" error.  Nothing that I know has changed since the last reboot.  and now, I can't get any kind of connections to my database server
[23:04] <moldy> hi
[23:04] <moldy> i am (still) having very weird problems with raid+lvm on karmic
[23:04] <moldy> when i boot the machine with both mirror disks of a raid1 attached, i don't see the partitions on the disks
[23:05] <moldy>  /proc/partitions shows sda, sdb, dm0 and dm-[0..3]
[23:05] <moldy> when i remove one of sda/sdb, i see the partition (sda1 / sdb1). what is going on here?
[23:05] <unit3> are you using linux md raid, or fakeraid on your motherboard?
[23:06] <moldy> linux md raid, to the best of my knowledge
[23:06] <unit3> ok, that is weird. what does /proc/mdstat say?
[23:06] <moldy> fakeraid is disabled in bios setup
[23:06] <moldy> one second, pasting it
[23:06] <unit3> k.
[23:07] <moldy> when i plug put one of the disks and attach it via usb, all works normally
[23:07] <moldy> i see sda1 and sdb1 and can add the missing disk to md0
[23:08] <moldy> i then waited for the raid to sync. shutted down the system, attached disk using s-ata again: problem is there.
[23:09] <moldy> unit3: mdstat: http://pastebin.com/f3914e90c
[23:09] <unit3> oh yeah, it's creating the dm devices, and then getting confused. that's weird.
[23:09] <unit3> hmm.. check dmesg, see if it says why it's creating those dm devices.
[23:09] <unit3> since it shouldn't be, if you're just doing disks->md->lvm.
[23:09] <moldy> unit3: mdadm -D: http://pastebin.com/f2e24855a
[23:10] <moldy> why is it using that /dev/block alias instead of /dev/sda1?
[23:11] <unit3> no idea. It shouldn't be, really.
[23:11] <moldy> i used the installer to setup the raid/lvm, btw
[23:12] <moldy> so this might be an installer bug
[23:12] <moldy> dmesg|grep -i "dm-" -C5: http://pastebin.com/f48886f75
[23:13] <moldy> i'm not knowledgeable enough to make much sense of those log messages
[23:17] <unit3> No idea. It really shouldn't be setting up DM devices unless you're using bios / motherboard fakeraid. I can't really explain that log at all.
[23:17] <moldy> hm ok, thanks anyway
[23:18] <moldy> i guess this will take some time to fix... i probably should attach another mainboard to the system to rule out that the motherboard's fakeraid is somehow interfering
[23:18] <unit3> might be a good step, yeah. or at least doublecheck the bios settings, or try some alternate ones for the SATA controller.
[23:19] <moldy> nvidia fakeraid is disabled in bios setup, i already played with all the other settings
[23:19] <unit3> yeah, don't know what else to suggest.
[23:19] <moldy> another thing that is weird is that there are devices with the name "nvidia" in them present
[23:20] <moldy> /dev/mapper/nvidia_ddeaefgd /dev/mapper/nvidia_ddeaefgd1
[23:20] <moldy> it looks like something is trying to be too clever... but i don't know much about this dm stuff