I mentioned about installing a new SSD in the workstation, putting a clean install of Debian-stable on it, and starting to migrate my stuff? I'm sure I did. Would have been right at the end of December, after I spent more time than expected on the server migration (also not yet completed).
Well... I'm still actually using the old HDD; haven't quite gotten around to finishing up the migration stuff.
A little while ago, I was editing an architecture document for a client, hit SAVE, and got a blank dialog box with OK and CANCEL buttons. Neither of which worked. A moment later, my wallpaper disappeared. Hitting ALT-F1 got a blank screen instead of a text console. I could still ping the workstation from the server, but not ssh in.
Okay, so hit the RESET button. It boots from the SSD; did I leave it set up that way? Telling the bootloader to use the old install gets an error message: something about the kernel being missing. Trying again but booting the SSD normally gets error messages about a SATA drive being missing, but it does eventually reach single-user mode (no multi-user mode when there were serious errors in the init process).
Reboot, and go into BIOS setup. The SSD is visible; the HDD isn't.
Well, all the important data is stored on the server, and backed up daily, but losing the HDD this week rather than a couple of weeks out would still be a significant inconvenience.
Oh, well. Shut down. Turn power off. Take the cover off. Wiggle the SATA and power connectors. Turn power back on.
The HDD is back. SMART says it's been powered up for 5 years, 4 months and 20 days, so I suppose the warranty has probably expired by now. It also says the drive proclaims itself to be OK, though the error stats suggest that it's getting rather senile and is probably delusional about its health.
Guess I should finish the migration soon. And I hope all the bits of the old server hold out for a few more months; I have no plans for replacing its in-house functions until a late phase of Operation Bug Out.