How to tell why drive is remounting read-only

Question

I have a Lenovo laptop with an SSD and Ubuntu 18.04. The root partition keeps remounting read-only, according to the console messages. Except it won't read when it happens, either, just giving I/O errors. Unfortunately, it's the root partition, and I don't have any other Linux partitions on the disk, so this means I can't do dmesg (or virtually any other command) when it happens. When I force it to reboot, everything is fine again for a few days until it randomly happens again. (Although, recently I did discover there is one particular setup script in a Git repo I'm working with that somehow always causes it to happen.)

Without any kind of logging or dmesg, the best info I can find is what is on the console before I reboot, which is always something like this, with more of the same scrolling across the screen until I force shutdown:
(Pardon the cell phone quality photo)

I've booted from a live USB and used e2fsck to force a check for errors including bad blocks (-c twice) and it found nothing. I've booted Lenovo's diagnostics and it found nothing wrong with the hardware. I've checked the S.M.A.R.T. data for the disk and everything was normal. I don't know what else to check.

Is there any way to determine what is going on?

Answer 1

This isn't a very complete answer to the question but it's the best I can do given a lack of other information and my knowledge of Linux thus far.

As suspected, the remount read-only and input/output errors were the result of an intermittent hardware failure. In this case, the connection between the system board and hard disk was flaky. By the time I rebooted the laptop, the problem would go away, which explains why it would work again and diagnostics didn't find any issues. Over time, the problem got more severe, where I would have to reboot multiple times (or shutdown and wait a few minutes) before it would go away. This meant it lasted long enough for diagnostics to catch the issue, which was how I figured it out.

The exact details of the hardware issue aren't terribly relevant, but for those who are curious, reseating the hard drive cable and replacing the drive didn't help. It turned out the system board itself was defective. I suspect thermal expansion played a role, since it tended to happen when I was running a heavy load and cleared up when the computer sat idle after losing its hard disk.

In any case, if anyone else sees similar errors, it's probably a good idea to suspect a hardware issue related to the hard disk (even if not the disk itself).