I’ve had a day full of disasters, compounded because of this continuing problem (but in spite of this continuing installer problem, my immediate problem is now thankfully fixed).
I wanted to clone my system OS+boot pool SSD. My reasoning was that ordinarily, a quick re-install to a new drive ought to be painless but given these ongoing problems I decided I’d make a “backup” spare disk just in case.
I took my existing system disk out, intending to use gparted on my desktop machine to copy the partitions to a new drive of similar (but it turns out, slightly smaller) size, also connected to my desktop machine.
It didn’t work and, having been careful not to mount the system drive during this cloning attempt, I put it back into my TueNAS box only to find the TN wouldn’t boot. I don’t know why and I was too anxious to do much experimenting. This was turning into a disaster.
I decided to reinstall the OS on this original SSD, installed back in my TN box. Using iLO 4 I mounted the latest TrueNAS-SCALE-24.04.2 ISO which I downloaded today and attempted to install the OS.
The first problem I hit was the inability to select the drive I wanted to target - see screenshot1 below where I could move the blue bar up and down and I could select OK or cancel but no matter how many times I pressed the space bar or clicked with my mouse, the asterisk wouldn’t turn up [*].
Trying to proceed, I removed the RAIDED drives so the only disk was this original and now-broken system SSD. Perhaps the installer would see sense and install to the only disk available? That didn’t work. Screenshot2 below:
I started the process again - same problem.
I started the process yet again and this time, for reasons I have no idea about, the selection worked. By now I had spent the best part of 45 minutes trying to reinstall.
It got to the point where it says “do you want a UEFI (select yes) or a legacy boot (select no)”. I took the advice and I selected “no” and was horrified to see the familiar “this is not a block device”. Screenshot3:
Reasoning that the installer might be different, I tried using a newly downloaded ISO of TrueNAS-SCALE-23.10.2 and got the same failure message at the same point.
I read that someone had successfully installed “Core” and then upgraded so I tried the same approach with TrueNAS-13.0-U6.2 and I got a different problem in that I couldn’t proceed Screenshot4 says something about a full path or shorthand device name.
I then made a bootable USB stick to try and speed things up and also use a slightly different installation method. I tried booting from that. I got to the same point and got the same “not a block device” error.
I tried again using the USB stick and this time selected “UEFI (select yes)” on account of having run out of ideas. The process continued to completion I reinstalled my config+secrets file and all was well. Screenshot5.
Only about three and a half hours and a loss of a pint of sweat and tears!
I now have a working system, fully reinstalled and a spare blank SSD on which to attempt installing a backup copy with secrets so that, in the unlikely event of an SSD failure, I can power down, switch SSDs and be working again within ten minutes. Ten minutes is about how long it takes to install trueNAS on a fresh SSD when the installer does work, so once the problem is fixed I can simply go down that avenue to create a “backup”.
In the mean time, If anyone can advise a safe way of cloning my SSD bearing in mind its ZFS nature, please let me know. I should have looked at CloneZilla but I wanted to do something which avoided rebooting my main linux box (for no particularly good reason to be honest).