Linux RAID-1 [4]

So the “new” RAID-1 array finished syncing and it was time to start testing it to see if it would work as planned. Unfortunately this did not prove to be the case. Attempting to mount the new array in /etc/fstab failed on a number of occasions because once again the system would stop dead during startup reporting the device is not ready etc. I have still not been able to work out whether this is because of the missing md module or some other issue. Because I eventually reached the outcome where fstab started stalling the startup even with all lines relating to /newhome removed completely from it, I have had to reinstall Mint again from scratch.
Here at least you do get a massive advantage over Windows. As your /home partition is on a separate volume, you can just remount it at install and start using it again as if nothing had happened. Whilst it is possible to change the profiles drive in Windows to a different drive, there are two problems with this. One is that this is unsupported to the extent that service packs will often not install, and secondly the system will create a new profile, in which all you can restore are files from directories like Downloads, Documents, Music and Pictures etc. You won’t have the experience of being able to start up Thunderbird and seeing it load all your accounts and emails, or being able to install Opera and seeing it recall your history and sync settings. So a reinstall is a much faster process; the actual Mint installation takes at most 10 minutes, and for other applications there isn’t this process of having to reinstall all the settings.
So at this point I went back to the test system to see if I could have more success in remounting its RAID-1 array at startup. Here I found similar to MainPC, the system has renamed the volume as /dev/md127 instead of /dev/md0. The main difference in this system was that I had created a partition on the new volume and labelled it “Data”. When I went to the terminal and ran the blkid command I got a UUID just for “Data” (as opposed to the UUID for the RAID-1 array). I then came up with the idea of mounting that UUID in /etc/fstab. As it turned out this worked and using the mount -a command prior to rebooting showed there were no errors. When the system was rebooted it came up normally and the partition was mounted as /newhome as expected. I then applied these lessons to MainPC. Again it came up with a partition called “Data” with its own new block ID. I rebuilt the RAID array from scratch with MDADM and then once it had finished syncing, formatted it then got the UUID of the new partition and put that into /etc/fstab. mount-a produced no errors and rebooting the system came up properly.

So that has been an epic journey up to this point and it ain’t over just yet. The next step is to restore /home to the RAID disks (previously it ran on a single one of the pair) and I am planning to let things stabilise for a week to show it all really is working well before I am prepared to move things back across again. In the meantime I want to have a look at mdadm’s monitoring system to see how it works or perhaps there is an alternative system that will monitor the state of the RAID array.