How to run bleeding edge

[bodhi.zazen]

How to run bleeding edge

... And maintain your sanity

I like to run on the bleeding edge but as you know this sometimes leads to problems ranging from minor bugs to major problems to total system failure.

To file a bug Bug reporting

To work around this I suggest running two partitions, might call them stable and testing.

You may partition however you like, but here is an example:

Partition table:

/dev/hda1 = 10 Gb = Stable root partition
/dev/hda2 = 10 Gb = Testing root partition

/dev/hda3 = 1 Gb = swap

/dev/hda4 = the rest = data

I suggest you use a data partition rather then /home

I will use /dev/hda1 as the "master" menu.lst (grub). Alternately you may use a /boot partition.


In addition I use the gparted live CD

You can Download it here

=====================

First download and install Ubuntu 7.04 Feisty Fawn Herd 1:
Download Feisty

Install into /dev/hda1

Update

sudo aptitude update && sudo aptitude upgrade

Go ahead an customize Firefox, install additional packages, or your desktop or what have you.

Do not do anything bleeding edge (like beryl or the latest nVidia beta drivers) yet ....

=====================

Now go ahead and set up your testing partition.

1. Boot to the gparted CD
   
2. Format /dev/hda2 if you have not already done so (same format as /dev/hda1)
   
3. Now copy partition /dev/hda1 to /dev/hda2
   
4. Note: When you copy a partition, both partitions will have the SAME UUID
   
   To fix this:

   • Either revert to referring to partitions as /dev/hdxy (as below)
   • Or set a new UUID and use the new UUID rather then /dev/hdxy (not shown)
     
     To set a new UID (ext2/3) :

     tune2fs -U random <partition_to_set_UUID>

     To list you devices by UUID :

     blkid

When the copy is done we need to do some minor edits:

Open a terminal. In gparted you are root so no need for sudo.

1. Edit /boot/grub/menu.lst on /dev/hda1
   
   

   cd /tmp
   mkdir hda1 hda2
   mount /dev/hda1 hda1
   mount /dev/hda2 hda1

   Now:

   nano hda1/boot/grub/menu.lst

   Edit and add these lines:
   
   In the top of menu.lst you need to edit a few configuration defaults :
   

   ## ## Start Default Options ##
   ## default kernel options
   ## default kernel options for automagic boot options
   ## If you want special options for specific kernels use kopt_x_y_z
   ## where x.y.z is kernel version. Minor versions can be omitted.
   ## e.g. kopt=root=/dev/hda1 ro
   ## kopt_2_6_8=root=/dev/hdc1 ro
   ## kopt_2_6_8_2_686=root=/dev/hdc2 ro
   # kopt=root=/dev/hda1 ro
   
   ## default grub root device
   ## e.g. groot=(hd0,0)
   # groot=(hd0,0)

   Make sure the root device is properly identified
   
   

   
   title Feisty STABLE kernel 2.6.19-7-generic
   root (hd0,0)
   kernel /boot/vmlinuz-2.6.19-7-generic root=/dev/hda1 ro quiet splash
   initrd /boot/initrd.img-2.6.19-7-generic
   savedefault
   boot
   
   
   title Feisty TESTING kernel 2.6.19-7-generic
   root (hd0,1)
   kernel /boot/vmlinuz-2.6.19-7-generic root=/dev/hda2 ro quiet splash
   initrd /boot/initrd.img-2.6.19-7-generic
   savedefault
   boot

   Note: Initially you will have your root partition identified by UUID. Example: UUID=fe139e06-121d-4887-b74a-ac2ccc178e3a although you will obviously have a different (unique) number. Change this, as above, to /dev/hda1 (STABLE) and /dev/hda2 (TESTING).
   
   Save and exit nano (Ctrl-X, answer Y).
   
   Copy menu.lst to TESTING:

   cp hda1/boot/grub/menu.lst /hda2/boot/grub/menu.lst

   Note: You will need to manually update the kernel line for your testing partition (/dev/hda2 in this example) after kernel updates. Alternately you can use a boot partition and chainload both the stable and testing partitions. Just be sure to watch those configuration lines in menu.lst
   
   
2. Edit /etc/fstab on TESTING /dev/hda2
   

   nano hda2/etc/fstab

   Change the root partition to hda2
   Change

   # /dev/hda1
   UUID=fe139e06-121d-4887-b74a-ac2ccc178e3a / ext3 defaults,errors=remount-ro 0 1

   To

   # /dev/hda2
   /dev/hda2 / ext3 defaults,errors=remount-ro 0 1

   Note: Again, your UUID=fe139e06-121d-4887-b74a-ac2ccc178e3a will be a different (unique) number. In this case we need to change to /dev/hda2
   
   Save and exit nano (Ctrl-X, answer Y).

=====================

Done.

=====================

Now go ahead and boot to /dev/hda2 Feisty TESTING kernel 2.6.19-7-generic

Go ahead and install any bleeding edge stuff you like. If it works, use gparted to copy /dev/hda2 to /dev/hda1 as follows:

1. Boot the gparted CD
   
2. Copy /dev/hda2 to /dev/hda1
   
3. Update fstab:
   
   

   cd /tmp
   mkdir hda1
   mount /dev/hda1 hda1
   nano hda1/etc/fstab

   Change:

   # /dev/hda2
   /dev/hda2 / ext3 defaults,errors=remount-ro 0 1

   To:

   # /dev/hda1
   /dev/hda1 / ext3 defaults,errors=remount-ro 0 1

   Save and exit nano (Ctrl-X, answer Y).
   
   
4. Do not forget to update menu.lst:
   

   nano /hda1/boot/grub/menu.lst

   Update your kernel information and initrd.img for:

   title Feisty STABLE kernel 2.6.19-7-generic
   root (hd0,0)
   kernel /boot/vmlinuz-2.6.19-7-generic root=/dev/hda1 ro quiet splash
   initrd /boot/initrd.img-2.6.19-7-generic
   savedefault
   boot

   Save and exit nano (Ctrl-X, answer Y).
   
   If you are unsure of your kernel and initrd,

   ls hda1/boot

   Use the set with the highest number....

=====================

You should run updates and install new (bleeding edge) software in TESTING.

If, however, you experience total system failure, boot to the gparted CD and copy /dev/hda1 to /dev/hda2 and try again !


Enjoy,

bodhi.zazen