A Gutsy Upgrade Story Part 3

David Vincent dvincent at sleepdeprived.ca
Fri Oct 26 19:48:58 UTC 2007


This is my last installment in this series.  This time I had to upgrade 
my MythTV Backend and my laptop.

The Laptop:

Alienware Area 51m 770, P4-3.2, 1gb ram, integrated atheros-based 
wireless, NVidia Go5700 video.

Upgraded from the Internet because my DVD drive is dying and couldn't 
read the disc properly.  Only two things didn't work:

1.  The upgrade installed a 386 kernel AS WELL AS the generic one but 
put the 386 one first in the boot order.  I recall one other thread on 
this list about someone ending up with a 386 kernel - is this how the 
installer treats hyperthreaded CPUs?  I noticed before the reboot so I 
uninstalled it straightaway and booted directly into the new generic kernel.

(That thread was [Re: Gutsy using only one CPU on dual core? [SOLVED]])

2.  My power settings were overwritten.  This has happened on every 
machine I upgraded.

Everything else just worked.  Compiz took no fiddling.  Wireless worked 
right away.  This was the easiest upgrade of all the ones I have done.

The MythTV Backend:

P3-800, 1gb RAM, generic pc with 2x120gb drives in a mirror for mp3s, 
1x120gb USB drive attached as video storage, 1x80gb drive as MythTV 
storage, and 1x20gb drive for the Ubuntu install itself.  Hauppage 
PVR-150 card, and a very low-end ATI graphics card.

When I installed Feisty on this machine and setup the USB drive to 
automount, I used /dev/sdb1 as the pointer to it in my fstab.  My other 
drives were detected as /dev/hda etc. and mounted via the UUID so this 
made sense at the time and worked fine.

After the upgrade - which I performed from the Alternate Install CD - 
the box didn't boot and gave me a "fsck died with exit status 8" error.

http://ubuntuforums.org/showthread.php?t=291890 solved the problem.

Gutsy reordered my drives and changed the USB from /dev/sdb1 to 
/dev/sdb5 so when processing my fstab the system tried to mount 
/dev/sdb1 twice and stopped.  A quick fstab edit and the server came 
back up perfectly.

Btw you can use "sudo blkid" to find the UUID of any attached drive and 
then use that in your fstab to mount partitions to avoid this problem 
completely.

Now all my machines are upgraded and everything seems to be running 
along just fine.  I need to wait on ATI to release a working driver for 
their hardware and I need Ubuntu/VMWare to hurry up and get VMWare 
Server 1.04 up in the partner repo - as far as I am aware it isn't there 
yet.  Otherwise I'll forget and just before doing some important work in 
VMWare I'll upgrade my kernel and then break VMWare again and have to 
recompile...

/me shakes it off

I haven't noticed any problems or slowness with the Tracker search tool 
(because I use the desktop edition of Ubuntu for my "servers" I turned 
it off on them right away and it isn't there in Xubuntu which I use on 
my slowest hardware).  I lost Azureus as my BitTorrent client (and I 
don't really like Deluge) but it was an easy install to get it back (and 
a leap forward in version too).  Truecrypt is also now in the repo so I 
didn't have to do anything fancy there.

Congratulations to the Ubuntu teams on their hard work and great 
results.  You folks rock.

-d





More information about the ubuntu-users mailing list