Failing hard drive?

[M0E-lnx]

I have a box with 3 hard drives. 2 x 40GB  and 1 x 250GB all sata.

Today, I restarted my box after an upgrade, and it just started going into a boot loop. right before reading fstab, it will act as if someone hit the reset button all on its own.

This has happened before a buch of times, but it had stopped doing it.
Today, I opened it up, unplugged the 250gb drive and restarted... booted up just fine.

turned it off, plugged the drive back in and restarted it, booted fine.

Does this sound like a failing hard drive to anyone else or is it just me?
FWIW, I'm backing up all my data from that drive into an external drive as I type this.

Any hints are appreciated. I just want to be sure the drive is the problem before I invest in a new one

[Andy Price]

I only have one Linux box and the HD hasn't failed yet, so I'm basing the following observations on my experience of lots of Windows machines.

It doesn't sound especially like a hard drive problem to me, though of course it could be. When a HD fails the machine tends to halt with an error message. But you could download something like the Maxtor HD disk checker to make a boot floppy/CD and test the disks.

When I get this kind of problem I take the RAM out and put it back in first, as well as unplugging/replugging all the other cables. Next I change the RAM, then the power supply (luckily I have spares kicking around to swap in). If I'm still not getting anywhere I take out all the cards (except the video), disconnect the mouse, keyboard etc and try booting. If the machine boots OK I start putting things back in one by one. Sometimes a bit of hardware fails and causes the PC to behave strangely - I had one case where a keyboard went faulty and stopped the machine from booting at all.

The other thing to look for is faulty capacitors on the motherboard. If the tops are bulging or leaking then you need to replace them as this can lead to the kind of problem you are experiencing.

Sounds like your fault is intermittent though, so it might be hard to track down.
Andy

[M0E-lnx]

FWIW, I've done the ram swapping, putting it in one module at a time and it does the same. I know this sounds like a hard problem to diagnose, but I'm also hearing noise coming from this one particular drive.

One other important piece of information is that I suspected this drive had problems some time ago, but wasn't sure, so I used it to store non-vital data. All my linux installs are done on the other 2 drives.

[nightflier]

Whenever I suspect a hard drive, I use The Ultimate Boot CD and run the manufacturer's long diagnostic for the drive.

http://www.ultimatebootcd.com/

[toothandnail]

FWIW, I've done the ram swapping, putting it in one module at a time and it does the same. I know this sounds like a hard problem to diagnose, but I'm also hearing noise coming from this one particular drive.

One other important piece of information is that I suspected this drive had problems some time ago, but wasn't sure, so I used it to store non-vital data. All my linux installs are done on the other 2 drives.

I would install Gsmartcontrol and see if anything is showing up in the SMART logs for the drive. I've had warning of impending drive problems from SMART several times in the past (OS/2 has a useful little monitor which can be run on the desktop and will pop up any time a SMART error is logged - wish I had the same for Linux).

Another option would be to download the evaluation copy of Dfsee (http://www.dfsee.com), install the Linux version and run a bad block scan on the drive.

Paul.

[MikeCindi]

I would agree with nightflier if you want to check the drive. If the drive is not your boot drive and does not have a swap partition it should not prevent booting. This is clear as the machine will boot without the drive. FWIW recently I had a bad cable as I did the same as you in disconnecting the device only reconnecting didn't help replacing the data cable did.

[M0E-lnx]

After checking the drive with ultimate boot cd, both the short and long test reported errors on the 250gb drive.

[Hiero2]

I'm going to make a note to myself about Gsmartcontrol. I'm using gnome-disk-utility (aka palimpsest) for that same purpose. It uses the SMART disc checking. When it works, it is lovely, but it seems buggy - doesn't always run. Maybe it's something else on my machine.