linux: running fsck hosed my drive?

[rainypickles]

hey all, i have a small samba server setup. i was logged in with putty and rebooted the machine. upon reboot, it said there were inconsistencies in the file system (or something similar) and told me to run fsck manually. so i did and it prompted me with things like "bla bla bla found. fix? (y/n)" i hit y to all of them.

after fsck "fixed" everything, i typed reboot, but the shutting down process got stuck at shutting down localhost loop (?) so i just hit the power button to reboot.

now my drive won't boot, and it is not recognized in the bios. the hard drive light stays on while the computer is on.

did i fsck up my drive by using fsck? (har har) what can i do about it? or did pressing the power button kill my drive?

thanks!

 

[bersl2]

Are you running smartd? You should.

 

[cleverhandle]

Sounds like the drive was already dying before you rebooted the first time. Powering off via the button shouldn't have done much more harm at that point - generally, abrupt shutdowns should only really cause problems at the filesystem level (due to losing disk buffers and such). I suppose that if the drive was in the middle of a seek and you cut the power that that could do some damage too, but that's not likely to be bad enough to cause the drive to be unrecognized by the BIOS. At that point, the damage was already done.

Certainly, fsck itself isn't going to directly cause a drive to die. All fsck does is check to make that the filesystem is logically consistent - it doesn't perform any low-level physical operations on the drive. Indirectly, maybe fsck would accelerate damage due to it generating more seeks. But again, at that point the damage was already done. Just serving some more files would have accomplished the same thing.

As bersl2 suggested, running smartd on a server is a good idea - it should listen to the drive's low-level SMART reporting and alert you in case of impending disaster. In theory, at least...

edit: typo