How to tell when a machine is fsck-ing and hasn't had a kernel panic?

Question

I have a server with Debian on it that I regularly reboot after upgrades. Sometimes (on schedule) fsck will check a disk when the computer is booting. With the exception of sitting in front of the console to observe the fsck, how can I determine the difference between a problematic halt and an fsck (besides waiting out the fsck, hoping it is an fsck)?

When I send the computer down to reboot, I will usually have a terminal window open pinging the computer so I know when it has come back up. My first thoughts drifted to fantasizing about hacking fsck to respond to pings with some special magic byte so you could tell via ping that a computer was fsck-ing, but I'm thinking there have got to be easier ways..

Answer 1

When Linux boots, it sends messages to the console, including the one about fsck checking the disks. Assuming you're purchasing your hosting from somewhere, ask your provider how you access the console during boot, that way you can monitor what the server is doing. If they don't offer an option for remote console access, consider another provider because you'll need it eventually.

Outside of console access, you'll just have to wait ...

(If this isn't a purchased hosting option, dedicated server, vps, whatever, and is instead your own physical server somewhere, then you need to work out how to give yourself remote console access which varies depending on the hardware, distance and network involved).

The other answers posted in the comment show you how to tell before you reboot the machine, but that's only useful if you a) remember and b) have time. If the machine reboots on it's own, remote console access is essential.

Answer 2

If your server has a serial port, and you have another computer within a reasonable distance, you could use a serial console.