Lab46 Wiki

• the formular plugin is giving me errors, need to figure this out (email assignment form)
• use include plugin to include a page containing various prior month status pages
• can I install writer2latex on wildebeest herd without needing gcj??

Some links of interest:

• http://www.freelists.org/post/dokuwiki/invoke-mediamanager-in-a-plugin,2
• unrelated: http://infoworld.com/d/adventures-in-it/run-it-business-why-thats-train-wreck-waiting-happen-477
• laddergoat
• LLVM
• Fluxbox config
• GLglobe
• Heavens Above

Sometime during the past week, when all those storms came through, the DSLAB must have suffered enough of a blink to reboot non-UPS machines… the video wall and sparta machines had only been up for 5 days (5 days, 5 hours, and 48 minutes as I write this, at 1:52pm Wednesday, July 30th).

In a way this was good, as I accidentally reboot data.

The video wall machines did not all come back up after the outage… 2 of them sat at POST complaining about dead BIOS batteries (lovely). Might be time to investigate rebuilding the video wall.

The Xserve RAID was also wailing… one drive appeared bad… I pulled the power on it as nobody has seemed to use it for a while now (we pulled the power on the Xserve G5 weeks ago and no complaints).

I added both the 64-bit and 32-bit versions of Slackware 13.1 to the LAIR netboot menus.

Still need to do some testing to make sure they actually work. If/when working, users can do HTTP-based network installs.

I found I had made a typo in setting nfs module options in /etc/modules and /etc/modprobe.d/local… I had: “nfs callback_tcpport:2049”, and it didn't like that.. instead, it wants: “nfs callback_tcpport=2049” – note the equal sign instead of the colon.

Made this change and also noticed that pending a future package upgrade, all files in /etc/modprobe.d/ lacking a .conf extension will be ignored, so I renamed “local” to a more appropriate “lair.conf”.. the package does a check for the old “local” file and deletes it, leaving only lair.conf (this only affects systems with 2.6.32+ kernels).

Ok… so if running a pvops-based kernel (2.6.32+), you may *NOT* want the following line in the Xen config file:

extra  = 'clocksource=jiffies'

I was battling with entirely stopped clocks… this ends up being quite annoying… machine strangely lags (network times out because it can't pass time), and of course time-dependent operations go nowhere.

A big SUCK. I took it out, shut down, and started up the VM, and time seems to be working again… we'll see if it keeps up or not.

I discovered that the nfs kernel module in 2.6.32 has a parameter to specify the port number for NFS callbacks, so I updated lair-nfs to deploy this if available.

A link that was useful is:

• http://kernel-handbook.alioth.debian.org/ch-modules.html

What actually led to the eventual update of lair-nfs, I thought to poke around at resolving that “peth0: too many iterations (6) in nv_nic_irq” message we've pretty much always had with our ASUS boards.

This link:

• http://pogson.6k.ca/2009/06/15/son-of-a-geek-or-too-many-iterations-in-nv_nic_irq/

Had some additional options worth implementing:

options forcedeth optimization_mode=1  poll_interval=100

I added these options to the existing max_interrupt_work we already had in place. We'll see if this makes any difference.

modinfo is my command of the day. It shows lots of useful module information, including what parameters are available. Highly recommended.

modinfo nfs

Suddenly brought some sanity where before there had been chaos.

As it turns out, 2.6.32 (the kernel that's coming with squeeze) is a Xen pvops kernel. The Debian “bigmem” kernels are apparently also pvops.

pvops kernels do not have the “xen.independent_wallclock” sysctl key, so lair-vm failed on my squeeze VM.

From what I can see, this is somewhat no longer needed in pvops, and there's a variable that can be set in /etc/default/rcS to avoid a harmless error message from being displayed on boot. Simply do the following:

echo HWCLOCKACCESS=no >> /etc/default/rcS

I have added logic to lair-vm to symptomatically check for pvops and either do the sysctl.conf or the /etc/default/rcS route.

As people are starting to test lab46new, I had a request from Scott (ssmith85) to install the zsh package, as he's been independently using zsh for a year+ now. In the interest of functionality, I've added zsh as a prerequisite of lair-std, so it will get pulled down wherever lair-std is installed.

No other changes beyond that were performed in this lair-std update.

To test out some things, I created a new squeeze VM, ircnew.offbyone.lan. It will eventually become the production irc for the LAIR.

It has a system configuration wiki page: system_irc

I completed the LAIR-wide updates by getting to the wildebeest herd today. aptitude update && aptitude upgrade all around.

This includes both lair-std and lair-vm updates.

I updated the lair-vm custom package to include the Xen independent wallclock logic.

Having the thought of flexibility in mind, I thought to harden the flakes so that they could be booted up and work even if all the wildebeest herd were not present (need a minimum of 1 on for things to work though).

In /export/client/etc/ on nfs I edited rc.local and modified the following (at the bottom of the file):

FLAKEID="`ifconfig eth0|grep 'inet addr'|cut -d':' -f2|cut -d' ' -f1|cut -d'.' -f4`"
 
# determine available X resource hosts
HOSTCNT=0
for XSERVE in antelope gnu wildebai wildgoat; do
    HOSTCHK="`ping -q -w 1 -c 4 ${XSERVE}.offbyone.lan | grep 'received' | cut -d',' -f2 |
 sed 's/^.*\([0-9][0-9]*\).*$/\1/g'`"
    if [ $HOSTCHK -gt 0 ]; then
        let HOSTCNT=$HOSTCNT+1
        XRESLOT[$HOSTCNT]="$XSERVE"
    fi
done
 
XHOSTID="`echo $(($FLAKEID % $HOSTCNT))`"
 
# assign to available X resource host
if [ "$XHOSTID" -eq 0 ]; then
    XRESHOST="${XRESLOT[1]}"
elif [ "$XHOSTID" -eq 1 ]; then
    XRESHOST="${XRESLOT[2]}"
elif [ "$XHOSTID" -eq 2 ]; then
    XRESHOST="${XRESLOT[3]}"
else
    XRESHOST="${XRESLOT[4]}"
fi
 
X -ac -query ${XRESHOST}.offbyone.lan :0 vt7 &

Basically, I make the process more flexible by checking to see which hosts are up, and create an array containing only the members of the wildebeest herd that are up… X server selection is then made from this list (in a round robin per IP fashion), so the flakes will work even if some of the wildebeest herd are down.

This makes especially good sense now, when we're running in lower power mode… I only have one of the wildebeest herd on (antelope at the moment).

I keep forgetting this… sometimes I need to override something, I recall this… and for the life of me cannot ever think of it by name to install it. And searching for it never works either.

I'm trying to create a virtual package of sorts that provides the “java2-runtime” symbol, so any existing prerequisites will not be pulled in… package in question is writer2latex, which wants to install gcj, but I've already got the sun java jdk installed (from package database!)

Made a new package, installed it, but it didn't seem to make any difference (this is on antelope)… gonna poke at it some more.

So I finally made some headway utilizing pdq and printing from antelope.

Pertinent files follow.

In /etc/pdq/interfaces, I made a symlink:

ln -s bsd-lpd-2.0 bsd-lpd

And in /etc/pdq/drivers/misc/hp-color-laserjet, I made sure the top line was as follows:

driver "hp-color-laserjet" {

  # This PDQ driver declaration file was generated automatically by
  # foomatic-rip from information in the file /root/pdq/etc/foomatic/hp-color-laserjet.ppd.
  # It allows printing with PDQ on the HP Color LaserJet Foomatic/Postscript.

  requires "foomatic-rip"

  option {
    var = "PageSize"
    desc = "Page Size"
    default_choice "PageSize_Letter"
    choice "PageSize_Letter" {
      desc = "US Letter"
      value = " -o PageSize=Letter"
    }
    choice "PageSize_A4" {
      desc = "A4"
      value = " -o PageSize=A4"
    }
    choice "PageSize_11x17" {
      desc = "11x17"
      value = " -o PageSize=11x17"
    }
    choice "PageSize_A3" {
      desc = "A3"
      value = " -o PageSize=A3"
    }
    choice "PageSize_A5" {
      desc = "A5"
      value = " -o PageSize=A5"
    }
    choice "PageSize_B5" {
      desc = "B5 (JIS)"
      value = " -o PageSize=B5"
    }
    choice "PageSize_Env10" {
      desc = "Envelope #10"
      value = " -o PageSize=Env10"
    }
    choice "PageSize_EnvC5" {
      desc = "Envelope C5"
      value = " -o PageSize=EnvC5"
    }
    choice "PageSize_EnvDL" {
      desc = "Envelope DL"
      value = " -o PageSize=EnvDL"
    }
    choice "PageSize_EnvISOB5" {
      desc = "Envelope B5"
      value = " -o PageSize=EnvISOB5"
    }
    choice "PageSize_EnvMonarch" {
      desc = "Envelope Monarch"
      value = " -o PageSize=EnvMonarch"
    }
    choice "PageSize_Executive" {
      desc = "Executive"
      value = " -o PageSize=Executive"
    }
    choice "PageSize_Legal" {
      desc = "US Legal"
      value = " -o PageSize=Legal"
    }
  }

  option {
    var = "InputSlot"
    desc = "Media Source"
    default_choice "InputSlot_Default"
    choice "InputSlot_Default" {
      desc = "Default"
      value = " -o InputSlot=Default"
    }
    choice "InputSlot_Tray1" {
      desc = "Tray 1"
      value = " -o InputSlot=Tray1"
    }
    choice "InputSlot_Tray2" {
      desc = "Tray 2"
      value = " -o InputSlot=Tray2"
    }
    choice "InputSlot_Tray3" {
      desc = "Tray 3"
      value = " -o InputSlot=Tray3"
    }
    choice "InputSlot_Manual" {
      desc = "Manual Feeding"
      value = " -o InputSlot=Manual"
    }
  }

  option {
    var = "Duplex"
    desc = "Double-Sided Printing"
    default_choice "Duplex_None"
    choice "Duplex_DuplexNoTumble" {
      desc = "Long Edge (Standard)"
      value = " -o Duplex=DuplexNoTumble"
    }
    choice "Duplex_DuplexTumble" {
      desc = "Short Edge (Flip)"
      value = " -o Duplex=DuplexTumble"
    }
    choice "Duplex_None" {
      desc = "Off"
      value = " -o Duplex=None"
    }
  }

  option {
    var = "Resolution"
    desc = "Resolution"
    default_choice "Resolution_600x600dpi"
    choice "Resolution_150x150dpi" {
      desc = "150 DPI"
      value = " -o Resolution=150x150dpi"
    }
    choice "Resolution_300x300dpi" {
      desc = "300 DPI"
      value = " -o Resolution=300x300dpi"
    }
    choice "Resolution_600x600dpi" {
      desc = "600 DPI"
      value = " -o Resolution=600x600dpi"
    }
    choice "Resolution_1200x1200dpi" {
      desc = "1200 DPI"
      value = " -o Resolution=1200x1200dpi"
    }
  }

  option {
    var = "DRIVERDOCS"
    desc = "Print driver usage information"
    default_choice "nodocs"
    choice "docs" {
      desc = "Yes"
      value = " -o docs"
    }
    choice "nodocs" {
      desc = "No"
      value = ""
    }
  }

  language_driver all {
    # We accept all file types and pass them to foomatic-rip
    # (invoked in "filter_exec {}" section) without
    # pre-filtering
    filetype_regx ""
    convert_exec {
      ln -s $INPUT $OUTPUT
    }
  }

  filter_exec {
      if ! test -e $INPUT.ok; then
        sh -c "foomatic-rip --pdq --ppd=/root/pdq/etc/foomatic/hp-color-laserjet.ppd${PageSize}${InputSlot}${Duplex}${Resolution}${DRIVERDOCS} $INPUT > $OUTPUT"
        if ! test -e $OUTPUT; then
          echo 'Error running foomatic-rip; no output!'
          exit 1
        fi
      else
        ln -s $INPUT $OUTPUT
      fi

  }
}

Basically, that the driver is defined as what its filename is. They differed. PDQ is silly and won't work because it can't find a definition when you provide a filename.

The above file is basically the openprinting.org ppd for the HP Color Laserjet, converted to pdq format, which I was able to do as follows:

foomatic-rip --ppd hp-color-laserjet.ppd --genpdq myprinter.pdq

I placed the file in the /etc/pdq/drivers/misc and named it hp-color-laserjet, and updated /etc/pdq/printrc accordingly.

Speaking of /etc/pdq/printrc, it is as follows:

# Redefinitions are silently ignored.  Be careful to define last the
# choices you want.  The order of processing is /etc/printrc and 
# then ~/.printrc


########################################################################
#
# Configurable options
#
#

# Directory to store jobs
job_dir "~/.printjobs"

# Time (in seconds) for which job files will be saved.
# Jobs files will be cleaned up after new jobs finish.
#job_history_duration 259200
job_history_duration 1800

# Maximum number of times to try to connect to the printer.
max_send_tries 16

# Delay (in seconds) between attempting to resend 
delay_between_tries 12

# Default printer definition
default_printer LAIRprinter

# Path that gets passed to driver scripts
driver_command_path "/bin:/usr/bin:/usr/local/bin"

# Path that gets passed to interface scripts
interface_command_path "/bin:/usr/bin:/usr/local/bin"

########################################################################
#
# Extra pieces of this config file, that define drivers and interfaces.
#
#

try_include "/etc/pdq/interfaces/*"
#try_include "/etc/pdq/drivers/*/*"
try_include "/etc/pdq/drivers/misc/*"


########################################################################
#
# Local printer definitions - Note that xpdq will create entries 
#                             automatically.  Changes made by superuser
#                             will be made to global config files.  To add
#                             entries by hand, see man printrc(5) for
#                             examples.

printer "LAIRprinter" {
    location "LAIR"
    model "HP Color LaserJet CP1518ni"
    driver "hp-color-laserjet"
    interface "bsd-lpd"
    driver_opts ""
    driver_args ""
    interface_opts ""
    interface_args "QUEUE = raw, REMOTE_HOST = laserjet.lair.lan"
}

#printer "stylusc800" {
    ## Added by the wizard on Fri Apr 30 12:10:16 1999
    #location "216 Talbot Lab"
    #model "Epson Stylus Color 800"
    #driver "epson-stylus-1.0"
    #interface "bsd-lpd-1.0"
    #driver_opts ""
    #driver_args "720x720, fscmyk"
    #interface_opts ""
    #interface_args "QUEUE = raw, REMOTE_HOST = printer.foo.edu"
#}

I did the entry by hand, basing it off the stylusc800 entry… and then was able to print /etc/pdq/printrc by issuing the following:

antelope:~$ pdq /etc/pdq/printrc

Getting end-user printing will need some more tweaks… turns out iceweasel isn't providing any adequate printer options (only printing to file)… I started to uncap some unruly gtk config nastiness which I didn't feel like monkeying with now… but interestingly, there were entries for the Color Laser CP1518ni *in* my iceweasel about:config … so some more playing is in order.

I found the following links of value:

• http://www.linuxfoundation.org/collaborate/workgroups/openprinting/database/pdqdocumentation
• http://www.openprinting.org/driver/Postscript
• http://pdq.sourceforge.net/

Etch and Lenny repository trees have been updated… Lair-STD package updated to support squeeze.

As a result of LAIR repository restructuring, I went and performed updates on all affected running machines… tweaked the sources.list, then ran aptitude update && aptitude upgrade

All running VMs, the NFS servers, VM servers, etc. have been updated (note: the wildebeest herd is not currently running, so they still need to receive the updates).

Part of my package repository work was related to streamlining the new Lab46 VM… I've proceeded by migrating over some essential /usr/local data.

Additionally, I've gone ahead and added the following packages:

• alpine
• apg
• aspell
• aspell-doc
• aspell-en
• at
• build-essential
• nullmailer
• mailutils
• nmh
• lair-std
• lair-nfs
• lair-vm
• lair-ldap
• bsdgames
• bvi
• curl
• finger
• finger-ldap
• flex
• ftp
• gawk
• lftp
• gdb
• valgrind
• gprolog
• gnupg
• hexedit
• iamerican
• imagemagick
• inotify-tools
• iproute
• irssi
• ispell
• elinks
• links2
• lynx
• lsof
• mawk
• mc
• mysql-client
• nano
• nasm
• ncftp
• netcat
• nethack
• nullhttpd (*** need to create and install this using that package faker thing)
• php5-cli
• php5-mysql
• pilot
• php5-curl
• python3-all
• python3-doc
• python3-examples
• ruby1.9.1-full
• ruby1.9.1-examples
• traceroute
• unzip
• clisp
• clisp-doc
• sun-java6-jdk
• subversion
• ia32-libs
• libncurses5-dev
• zsh (now pulled in via lair-std!)
• zshdb
• zsh-doc

I decided to take another stab and mitigating the noise APT makes when dealing with our custom package repository.

machine:~$ sudo gpg --gen-key
gpg (GnuPG) 1.4.9; Copyright (C) 2008 Free Software Foundation, Inc.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

Please select what kind of key you want:
   (1) DSA and Elgamal (default)
   (2) DSA (sign only)
   (5) RSA (sign only)
Your selection? 1
DSA keypair will have 1024 bits.
ELG-E keys may be between 1024 and 4096 bits long.
What keysize do you want? (2048) 1024
Requested keysize is 1024 bits
Please specify how long the key should be valid.
         0 = key does not expire
      <n>  = key expires in n days
      <n>w = key expires in n weeks
      <n>m = key expires in n months
      <n>y = key expires in n years
Key is valid for? (0)

Key does not expire at all
Is this correct? (y/N) y

You need a user ID to identify your key; the software constructs the user ID
from the Real Name, Comment and Email Address in this form:
    "Heinrich Heine (Der Dichter) <heinrichh@duesseldorf.de>"
Real name: Matthew Haas
Email address: wedge@lab46.corning-cc.edu
Comment: LAIR Package Repository Key
You selected this USER-ID:
    "Matthew Haas (LAIR Package Repository Key) <wedge@lab46.corning-cc.edu>"

Change (N)ame, (C)omment, (E)mail or (O)kay/(Q)uit? O
You need a Passphrase to protect your secret key.

We need to generate a lot of random bytes. It is a good idea to perform
some other action (type on the keyboard, move the mouse, utilize the
disks) during the prime generation; this gives the random number
generator a better chance to gain enough entropy.
+++++++++++++++++++++++++++++++++++...++++++++++++++++++++++++++++++.+++++++++++++++++++++++++.+++++.++++++++++++++++++++..+++++..+++++.++++++++++......+++++
We need to generate a lot of random bytes. It is a good idea to perform
some other action (type on the keyboard, move the mouse, utilize the
disks) during the prime generation; this gives the random number
generator a better chance to gain enough entropy.
++++++++++.+++++++++++++++..+++++.++++++++++++++++++++++++++++++.+++++.+++++++++++++++.+++++++++++++++.+++++++++++++++.++++++++++.+++++.+++++..+++++^^^
gpg: key 610DD9EF marked as ultimately trusted
public and secret key created and signed.

gpg: checking the trustdb
gpg: 3 marginal(s) needed, 1 complete(s) needed, PGP trust model
gpg: depth: 0  valid:   3  signed:   0  trust: 0-, 0q, 0n, 0m, 0f, 3u
pub   1024D/610DD9EF 2010-06-22
      Key fingerprint = 849B 250D A979 747E 71FA  E57E B599 98D1 610D D9EF
uid                  Matthew Haas (LAIR Package Repository Key) <wedge@lab46.corning-cc.edu>
sub   1024g/E34A3097 2010-06-22

machine:~$ 

Password is a super secretive “packages”.

dpkg-scanpackages $dist /dev/null | gzip -9c > $dist/Packages.gz apt-ftparchive packages $dist > $dist/Packages cd $dist gpg –sign Packages

machine:~$ sudo apt-ftparchive release squeeze > squeeze/Release

machine:~$ sudo gpg -abs -o squeeze/Release.gpg squeeze/Release

You need a passphrase to unlock the secret key for
user: "Matthew Haas (LAIR Package Repository Key) <wedge@lab46.corning-cc.edu>"
1024-bit DSA key, ID 17D9FFD4, created 2010-06-22

Enter passphrase: 

pubring.gpg from the key source (the public key from the public/private key pair), needs to propagate itself to the client, and then added to the apt-key keychain as follows:

apt-key add pubring.gpg

In the end I was successful… I found the following link quite useful:

• http://blogs.cae.tntech.edu/mwr/2007/05/28/my-own-private-debian-repository/

I also generated a standard repository directory tree for squeeze, and have my test VM working successfully off of it (installed lair-std without any complaints about unauthenticated packages). Need to do a few tweaks to lair-std to make it fully squeeze ready (implants lenny sources.list, for instance)… but overall a success!

Should roll out a new lair-std release for all, and migrate both etch and lenny into proper trees as well… got a nifty new regenpackages script that does the bulk of the work (once you put the package in the right place). Need to write this all down.

ufo.dslab.lan was established, as a DD-WRT'ed low power wireless network serving the 10.81.5.x subnet to allow wireless devices access to the internal network in the DSLAB. Encrypted and password protected, and broadcasting turned off, we'll see if it survives the scrutiny of CIT.

The great battery swap finally took place… turns out one of the batteries was cracked.

To accommodate the UPS battery upgrade, VMs on halfadder were moved to sokraits via live migrations, which took place without issue.

Following the upgrades, VMs were migrated back. Again, no problems.

As it turns out, migrating greatly exaggerated the “clock went backwards” message, and resulted in irc going stupid. I happened to encounter a remedy– how to decouple the Xen VM's clocks, so they won't be bound to the dom0 any longer, and therefore more resilient when being migrated between hosts.

How I accomplished this is as follows…

On the dom0, edit /etc/sysctl.conf and add the following:

# Decouple clock in Xen VMs
xen.independent_wallclock=1

In the domU Xen config files, add the following line:

extra       = 'clocksource=jiffies'

In the domU itself, perform the similar /etc/sysctl.conf modification.

I shut down and started back up all the VMs, so these settings are in place.

I performed “aptitude update; aptitude upgrade; aptitude clean” on the wildebeest herd, www, and log.

I decided to commence work on rebuilding the Lab46 VM (and Lab46 “Family of Services”). I want to do a squeeze (testing) install for Lab46, so as to maximize the currency (hopefully squeeze will be released sometime before the end of time).

I did the following:

# Create disk images (main, tmp, swap)
dd if=/dev/zero of=lab46new.disk bs=1M count=4096
dd if=/dev/zero of=lab46tmp.disk bs=1M count=256
dd if=/dev/zero of=lab46.swap bs=1M count=256

# Format images
mkfs.ext3 -j -O dir_index -v ./lab46new.disk
mkfs.ext3 -j -O dir_index -v ./lab46tmp.disk
mkswap ./lab46.swap

# Mount the main disk image
mount -o loop lab46new.disk /mnt

# Install the base squeeze distribution via debootstrap
debootstrap squeeze /mnt http://mirror/debian

# Set some key config files
cp -f /etc/network/interfaces /mnt/etc/network
cp -f /etc/apt/sources.list /mnt/etc/apt    # rename lenny to squeeze

# Hop into the new filesystem
chroot /mnt
mount -t proc proc /proc

# Perform some necessary maintenance
rm -f /etc/hostname
passwd
aptitude update

# Install a kernel and necessary tools
aptitude install ssh linux-image-2.6.32-5-xen-amd64 locales console-data

# Fix locale settings
dpkg-reconfigure locales console-data
# For the above, be sure to select "en_US" and go for "en_US.UTF-8 UTF-8"

# Set timezone
ln -sf /usr/share/zoneinfo/US/Eastern /etc/localtime

# Hop out
umount /proc
exit
umount /mnt

Be sure (before unmounting) to copy out the kernel and initrd from /mnt/boot into /xen/boot and update the Xen config file accordingly.

Initial Xen config file appears as follows (note disk image locations and MAC address):

##########################################################################
# LAIR Xen VM configuration file
##########################################################################

#################################################
#  Kernel + memory size
#
kernel  = '/xen/boot/vmlinuz-2.6.32-5-xen-amd64'
ramdisk = '/xen/boot/initrd.img-2.6.32-5-xen-amd64'
memory  = '512'
vcpus   = '2'

#################################################
#  Disk device(s).
#
root    = '/dev/xvda1 ro'
disk    = [ 'file:/root/lab46new.disk,xvda1,w',
            'file:/root/lab46tmp.disk,xvda2,w',
            'file:/root/lab46.swap,xvda3,w' ]

#################################################
#  Hostname
#
name    = 'lab46new'

#################################################
#  Networking
#
dhcp    = 'dhcp'
vif = [ 'mac=00:16:3E:5D:88:D7' ]

#################################################
#  Behaviour
#
on_poweroff = 'destroy'
on_reboot   = 'restart'
on_crash    = 'restart'

I set up lab46new.offbyone.lan in DHCP and DNS, and for now am running the VM out of root's home directory (to reduce stress on the NFS mounts to existing VMs).

Booting the VM works, but the Xen console is frozen at boot (I could SSH in and use it just fine though).

Edit /etc/inittab, and change tty1 to hvc0 (along with commenting out getty's that will never be accessed):

...
1:2345:respawn:/sbin/getty 38400 hvc0
2:23:respawn:/sbin/getty 38400 tty2
#3:23:respawn:/sbin/getty 38400 tty3
#4:23:respawn:/sbin/getty 38400 tty4
#5:23:respawn:/sbin/getty 38400 tty5
#6:23:respawn:/sbin/getty 38400 tty6
...

I installed the following packages:

linux-image-2.6.32-5-xen-amd64
ssh
locales
console-data
vim
fetchmail  # gonna try playing with IMAP new mail notification

Some potentially informative links:

• http://lubl.de/blog/2009/09/creating-a-squeeze-amd64-domu/
• http://wiki.debian.org/Xen#XenonUnstable.2BAC8-Sid

In anticipation of a group advising session on Tuesday, June 15th, I reconstructed the LAIRwall.

<html><center></html>

<html></center></html>