Wednesday, August 8, 2012

Why you should wait more to upgrade to the next LTS..

Notify me for long-term support versions
There are still a few of us, who prefer to be safer. We often upgrade only when an LTS of Ubuntu is released. However, not most of us realize that we should wait till the point release is available, letting the critical bugs found by the users fixed. The point releases are usually made available around the end of July. Ubuntu 12.04.1 will be available this month. Update manager will list 12.04 as available for the 10.04 users, only when 12.04.1 is released. That means, it won't be shown to the users who have set their system to notify only when LTS is released, till the first point release is released.

However, if you are still in a haste to go 12.04 from your stable 10.04, you can do this by "update-manager -d" I tried above, and it was a failed attempt, and hence will discourage anyone else to go to 12.04, directly from their 10.04, for their mission critical systems. Once 12.04.1 is released, it will be listed as 12.04 in your update manager, and that will be the ideal time to upgrade. I tried as above, which successfully downloaded the upgrades (took 26 hours in a speed of 60 kB/s), installed, and restarted. But it failed to restart, even after attempting all the suggestions on the Internet. When I tried with the 11.10 DVD (I didn't have 12.04 DVD with me, and I didn't have time to download and burn one, or get one from someone else), it showed the current operating system as 10.04, and from there I picked the option to "Upgrade" using the DVD. My file system was safe, along with all my documents. Some of the applications I had in my 10.04, had to be reinstalled for my new 11.10. Anyway, I smoothly upgraded 11.10 to 12.04, as expected, within 5 hours. I then installed the missing applications. Everything is safe, and now I am with 12.04.

Now I am just analyzing what might have happened initially to cause failure in my upgrade. It seemed to be an issue that happened when upgrading the Operating System during the final restart, as the operating system was still detected as 10.04 when I inserted 11.10 DVD. (It considered it an upgrade from 10.04, and not an installation replacing 12.04). I like that installing on top of 10.04 option, without formatting the entire disk, or without leaving the previous data in another partition, wasting the space.

Post-mortem

The page started to show the below message after the restart.
The disk drive for / is not ready yet or not present. 
Continue to wait; or Press S to skip mounting, or M for manual recovery.

Clicking S complained missing /tmp. Further S lead me to an empty black screen that takes nowhere. Go for M 
Root filesystem check failed.
A maintenance shell will now be started.
CONTROL-D will terminate this shell and reboot the system.
Give root password for maintenance 
(or type Control-D to continue):

and the manual fsck attempts did no fix.
fsck -f /dev/sda1

vi /etc/fstab 
showed
/mnt/4096 Mb.swap none swap sw 0 0
and
vi /etc/mtab
showed
/dev/sda1 / ext3 rw 0 0

Copying the contents of fstab to mtab and restart did no help, though suggested in some forums.  

mount /dev/sda1 /
mount: /dev/sda1 already mounted or / busy
mount: according to mtab, /dev/sda1 is already mounted on /

Remounting with rw permissions to follow the below commands as suggested.
mount -w -o remount /
dpkg --configure -a
apt-get install -f

dpkg --configure -a
gave
dpkg: error: parsing file '/var/lib/dpkg/status' near line 28305 package 'xmind'.
blank line in value of field 'Description'.

Removed the paragraphs defining xmind from the files /var/lib/dpkg/status and /var/lib/dpkg/available. It also complained about virtualbox-2.2, which I eventually removed as well.

Further fix attempts took me to the below shell.
Gave up waiting for root device. 
Common problems: 722988907ca
-  Boot args (cat /proc/cmdline)
- Check rootdelay = (did the system wait long enough?)
- Check root - (did the system wait for the right device?)
ALERT
/dev/disk/by-uuid/c67309e6-6099-40db-8b1b-1722988907ca does not exist. Dropping to shell.

Removed the UUID definitions,
root=UUID=.....7ca
from the entries in
/boot/grub/menu.lst
Making 
/boot/vm......generic
root=UUID=.....7ca ro quiet splash
to 
/boot/vm.....generic
ro quiet splash 
in the entries, as discussed here.

However, these fixes eventually took me to the shell from all the kernel images, letting me explore further issues. I finalized, installing from DVD would be a smarter option, as nothing I will lose.

Conclusion 

Lack of documentation or help is surely a frustration for an average user, in case of a failure like this, though it is not so common. We should focus more on community documentation to avoid turning the users off. My suggestion for a normal user to take a backup before upgrading such, and in case of failure, try if you can fix it, and if not, go for reinstalling on the same partitions, without formatting.

No comments:

Post a Comment