One of the nice things about Linux, is if you have a problem, the chances are that someone has either written something on how to resolve the problem or an application to give you the information you need!
On my 64-bit Athlon, strange things were happening on my SATA data hard drive, my main drive on which Kubuntu 9.10 runs on is SCSI drive. I was busy setting up a CentOS 5.4 ISP server following the instructions on www.howtoforge.com (excellent site) and I started getting I/O errors in my Virtualbox VM. then my Data drive disappeared in Dolphin, showing as unmountable!.
So I ran a fsck on reboot by doing the following in terminal:
cd /
sudo touch /forcefsck
sudo reboot
The Touch command creates an empty file in the root of the system volume. On reboot Linux checks for the existence of this file and automatically runs fsck. On completion of fsck, the file is deleted by Linux. Neat!
fsck did find errors on the data disk, fixed them, and then started Kubuntu.
However, the errors continued to get worse and the virtual CentOS server refused to boot into a login screen due to multiple I/O errors. Not good.
Now I know my SCSI hard disk gets very hot, after all it does run at 10,000 rpm! So I wanted to see how hot my disks were. A quick search on the Internet produced http://www.cyberciti.biz/tips/howto-monitor-hard-drive-temperature.html and the Ubuntu package hddtemp.
A simple sudo apt-get install hddtemp
and I was ready to go.
From terminal just enter sudo hddtemp /dev/xxx where xxx is the device name. Often tis is either sda or sdb if you have two hard disks but can be hda or hdb. In may case I entered:
$ sudo hddtemp /dev/sda and got:
/dev/sda: Maxtor 6Y060M0: 45°C
Doing the same for my system disk but using sdb gave me:
/dev/sdb: COMPAQ BD03685A24: 44°C
Monitoring my data drive over time saw the temperature rise to 48°C, with the system disk remaining unchanged.
However, this is within specification (Max is 55 Deg), so tried this command which checks the disk for bad blocks, fsck only checks file structure:
sudo badblocks -s -v /dev/sda1
Time for a new disk?
=-=-=-=-=
Powered by Bilbo Blogger
MyLinuxRamblings 
I hope you have up to date backups of anything important on those disks! 45C is rather hot in my view, I have disks that run closer to 23C (although in quite a cool room).
Installing smartmontools is also a good idea – won’t pick up on all failures but still worthwhile.
The other thing is to look for disk errors in /var/log/messages of course – often this is the best clue to disk errors.
Finally, consider turning off write caching using hdparm -W 0 /dev/sdX (from memory) – this is a separate issue to hard disk health and may slow things down a bit, but avoids lots of potential filesystem and LVM corruption.
In the end I replaced a SATA cable and this seems to have resolved the problem. No data was lost and my virtual box images are intact. I’ve also dropped in another SCSI drive and this is now my data volume until my 300Gb SATA drive is delivered. It was interesting noticing the performance difference between the SCSI and SATA drives. The SATA drives really hit the CPU hard.
All my PCs (I have quite a few) have hard disks which run 40+ Deg. I take the view that as long as they are within the manufactures specification, then I’m okay!
Yes Smartmontools look like it would be worthwhile