Take the 2-minute tour ×
Ask Ubuntu is a question and answer site for Ubuntu users and developers. It's 100% free, no registration required.

I have a machine located 500 miles away that's running Ubuntu 8.04. I figure it's just about time that I upgrade to the latest LTS. However, there's a software raid (md_raid) in there, and I'm afraid that just a dist-upgrade when I switch over the sources.list will end with catastrophic failure. Like a panic on boot that the raid'd disk can't be read, or something else.

First, hoping that's not the case, however, if it ends up happening I'm wondering if there's a means of having someone drop in a Ubuntu 10.04 server install disk, and flip on ssh, and some means for me to hop on and re-run the installer remotely. Is this feasible? If so, what would one need to do aside run apt-get install ssh on the target machine? I do have friends who can be in front of the target machine to initiate the process, just not execute it out.

share|improve this question
3  
Changing sources.lst and running a dist-upgrade is not a recommended procedure to do a release upgrade and is likely to break your system, on a server you use "do-release-upgrade". –  João Pinto Dec 15 '10 at 0:06
    
Word, i forgot that's how ubuntu does it. I've come to use the update-manager -d when I have a gui. –  David Walker Dec 16 '10 at 14:52
add comment

4 Answers

Review the instructions on the upgrade. I have had to uninstall programs before the upgrade. I have done a number of installations over SSH without a problem. If there is a problem, booting a CDROM installation to complete the installation is an option. You may need someone to configure the IP address and start the ssh server for you.

Usually, you will be able to reconnect via ssh, although I would keep a second connection open in case you accidentally kill the first one. The installation process is very good at not killing existing session when running the upgrade.

You may need to mount the partitons and chroot into them to complete the installation. You may need to bind some devices from the live system into the directories to get the installation to complete. If you do, the missing device is usually obvious from the error. Just exit the chroot, do the bind, and begin again.

This is the kind of situation where IP access to the console can be extremely useful. An IP accessible KVM would be ideal.

share|improve this answer
add comment

What you want to do in order to do an upgrade is

# do-release-upgrade

Going by personal experience, this has broken my computer in some cases.

If that does happen, you can use the mini.iso image to start an install. It will be a fresh install, but you will need somebody on the other end to configure it for you. If you want this to happen automatically, check out the question on preseeding an ubuntu-server install

share|improve this answer
add comment

If you want to accomplish this on multiple Ubuntu server instances set them all up with the same SSH key and then simply enumerate through the hosts in a shell scripts having them all execute 'do-release-upgrade'.

share|improve this answer
add comment

The Ubuntu documentation explains the procedure well.

The most important thing is to install update-manager-core; this package finds every problem before updating.

share|improve this answer
    
Whilst this may theoretically answer the question, it would be preferable to include the essential parts of the answer here, and provide the link for reference. –  Ringtail Nov 25 '12 at 0:49
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.