hi! I have a desktop computer with two internal hard drives (sda = 1TB and sdb = 500GB) plus an external drive connected by USB. I've been running Linux Mint Tara in a dual-boot configuration with Win10 for several years, no problems. I'm in linux about 95% of the time.
Yesterday Mint started acting a bit funny, culminating in the updater refusing to work. When I logged off, it dropped me to a black screen of death with flashing cursor. i used REISUB to shut down, then tried to start back up.
After another black screen of death, eventually some error messages began to scroll by. They said (paraphrasing from memory)
There was also a long number in there, toward the beginning.

My research suggests that these I/O errors mean my drive is failing, but why is it pointing to "fd0"? I don't have a floppy drive. I saw a linux forum mentioned that Microsoft Hypervisor can cause this error because it assumes a floppy drive, but I don't have Hypervisor.

How should I go about figuring out which drive, if any, is failing? I ran on my home and root partitions, and in each case I got back something like
Does this mean these partitions are in trouble? That particular drive is not very old. It's the newest of the three.
Thanks for any thoughts.
Mary

Run a SMART test on each drive:

sudo smartctl -l selftest /dev/<device>
sudo smartctl -a /dev/<device>

The USB adapter (internal to the drive) may or may not pass SMART commands it may not be possible to check its status.

You should not run fsck on an active file system which is why you see the modified warning and 2.6% is the percentage of file fragmentation which is ok.

Great, I will try that when I get home, thanks.

I should have mentioned that I ran e2fsck from a live Mint DVD, so I think (?) the drives were not mounted. It's good to hear that 2.6% is not an abnormal result though. Based on looking at the manpage, it sounds like those non-contiguous blocks have now been somehow walled off so there is nothing more I need to do about them.

Most of the major hard drive manufacturers have diagnostic test utilities available for download and can be in the form of a bootable iso. Usually, you have to run them before they will give you a warranty replacement. If you post the manufacturer of the suspect hard drive, I'll try and find the download link.

Other than that, it's not clear from your post if you can still boot into Mint. Even if you can, I would get a linux livecd and boot with that to run any diagnostic tests. There's smartctl and gparted for starters which is generally available on most distros. I really like Parted Magic:

https://partedmagic.com/

It costs $11 and has a lot of hard drive tools. For free, I believe Knoppix has a bunch of hard drive utilities as well.

Thank you, that is very helpful. I have a linux Live CD, and I think i may already have the tools for Seagate. I need to figure out what brand the other drive is.

I haven't been trying to boot into linux because I am afraid of further damaging my system. Should I be?

Windows boots fine, for whatever info that provides.

What is on the USB drive? If only data then you could likely disconnect it while testing the other.

smartctl should be able to give you the details (brand, model, serial, etc.) of each drive with smartctl -a /dev/sd?

I have a USB connected drive that I use for backups and the detail data returned seems the same for both the internal and the USB connected drive.

You did not say whether you tried booting to linux after running fsck. If not then the fsck likely found and fixed errors as reported and you may be entirely good to go.

Update:
My system is still unbootable. Here's what I've done so far:

I ran the smtctl long test on both hard drives. It came back with no errors and claims general good health for each drive.

From a system rescue CD, I have run e2fsck on each of the linux and windows partitions. There were only a few non-contiguous blocks, which appear to have been repaired.

I began to suspect the / partition might be full, so I mounted my / partition from a rescue CD and found a log file that was several gigabytes in size and I deleted it. The root partition is now only 80% full after doing that. The system still would not boot.

I went to the BIOS and removed the "floppy drive" (I don't have a floppy drive) from all menus. I did this because my I/O error was specifying "fd0, sector 0" as the error source. As discussed below I no longer have any sort of error, just a black screen.

I ran the Seagate disk utility, which looked very similar to smartctl tools but I tried it anyway. No errors.
I ran gparted and used the "check" function on all Windows and linux partitions. Again, everything ran fine with no errors.

How, then can I troubleshoot this black screen of death? When my system tries to boot, the Mint logo comes up for a few seconds, then I get the flashing cursor on black background. Every two minutes or so, I see my mouse cursor and the Mint background will flash for about a second, then black screen of death. No errors or messages of any kind ... just a black void. It's almost like Linux is trying to start, but only every 2 minutes or so.

Where should I look next? When I mount / from the rescue CD, it seems to behave normally, so I'm wondering if maybe it's not a hardware problem.
Thanks!

It sounds like a graphics issue. Can you get to a console via ctrl-alt-F2, or another function key? That would give you a chance to do an update, and perhaps troubleshoot the errors further.

Thank you-I tried the “nomodeset” suggestion here https://community.linuxmint.com/tutorial/view/842
But still the same result. Ctrl Alt F2 does not bring up a prompt or anything. However I am able to boot into recovery mode (command prompt only).
I have a nvidia card that’s worked fine for years.any idea on how to troubleshoot would be welcome. I guess I can try reinstalling the driver as a start.

Further update: I’m trying to troubleshoot from recovery mode, but I can’t because the network’s not enabled. When I try to run commands, there is a lag, so the command to start the network times out. This is starting to get frustrating. I hate to reinstall an operating system over a small graphics error, but is there any other way?

Final update: on browsing the Linux Mint forums, apparently Mint had a colossal mess-up with an update on 9/30 for my version of Mint that makes systems utterly non-functional. See here https://forums.linuxmint.com/viewtop...f=143&t=331605
You run two commands from recovery mode and it's 100% fixed. Oh my goodness, I was about to retire this computer but that's no longer needed.
Thanks everyone for your help!

That happens with Mint from time to time, along with other issues. I continue to avoid using it. But it is rather popular, so continue to have fun with it.

Try LMDE - the Debian based version of Mint - very similar except that
Over the years Mint has had about 3 or 4 dodgy updates whilst LMDE has had zero.

Presumably, LMDE has to go through Debian grade QA.

LMDE is, or at least used to be, based on Debian Testing. Mint itself is based on Ubuntu, which in turn is based on Debian Unstable. LMDE doesn't have to be, and isn't, AFAIK, controlled or checked at all by Debian. It's just based on a more thoroughly tested version.