I had a working R6.01 that I wanted to upgrade. I followed the directions on the wiki, making a backup before I did anything.

When I upgraded, I accidentally got myth 0.23, which I didn't want. Since my most recent backup was from my R6.01 system, I figured it would be okay to use the R6.02 installer disk and select "upgrade" to get the myth 0.22 system that I wanted.

The installation seemed to go fine, but after rebooting the system hung at "parent daemon running."

Undeterred, I tried the R6.01 installer disk. This time it said there was a problem with the database upgrade, so a new database was created instead. After reboot, the system hung again.

I was devastated, but I noticed something peculiar. The installer mounted a new /myth at /sd3. My old /myth, the one with about a hundred recordings I want to keep, remains intact, unmounted, on /sda4.

So two questions. Why did my upgrade attempt fail so miserably, and, more importantly, what can I do at this point to preserve those recordings and get back to a working system? I'm desperate for help.

Hello

I believe that the installer is not aware /sda4. It normally maps

/dev/sda1
/dev/sda2
/dev/sda3

This is the default setup. It would be helpful if you could post the
results of this command.



This is issue one you need to get a handle on the partitions first.

Great idea to make a backup ..)
Second - Your backup is .21 so you need to roll back to 6.01. Then run the restore to get you database back to normal. Then upgrade again to R6.02,

Is there any way to tell the installer to look in the right place for the backup?

I think you are a little confused. You run the system restore after you are inside of Knoppmyth. In your case r6.01 needs to be running and then run the restore. Assuming the partitions are correct. Run DF and check if you have and extra partition. Then it will restore the (database) back to the way it was

Okay, I see what you're saying now.

I reinstalled R6.01. The database was not restored, of course, but I then changed the fstab to mount /myth on sda4 rather than sda3. Then I ran mythrestore. Now I have what appears to be a working myth 0.21 system with my database working properly and all my recordings available. Phew!

So you're saying that if I now run the following code, then I'll be able to safely upgrade to myth 0.22?

Yes you should be safe to upgrade. Just make sure your /etc/pacman.conf is NOT pointing to the testing repos.

*sigh*

I was careful to keep only the regular core and extra repositories, and exclude the testing ones. But the "upgrade" apparently just gave me myth 0.21 again. How did that happen?

I would try the upgrade again, but my network stopped working after the other one.

Is there a way to specify the repositories so that I get myth 0.22 rather than .23 or .21?

Okay, here's the problem in a nutshell. Despite my strange partition table, with no sda3, and sda4 containing my recordings, I can get back to my original system using the R6.01 installer disk.

However, I still don't know how to upgrade to myth 0.22. If I use the R6.02 installer disk, it always fails to find my database and creates a new one on /sda3 instead.

If I try to upgrade over the net, one of two things happens. If I leave my repositories on core-testing and extra-testing, I get upgraded to 0.23, which I don't want. If I switch my repositories to core and extra, then the "upgrade" process keeps me at myth 0.21.

Can anyone suggest ways to make either upgrade method work for me?

The only thing left to try was:Some errors were logged, but none that seemed to matter much. The command actually upgraded my system to myth 0.22, and, after a week of struggle, I've accomplished what I set out to do.