Cache failure on otherwise perfect disk?
 

I have a Shuttle XV-35 with a 2.5" 320 GB hard disk. The computer became very, very slow. About 50% wait states in top. Disk light almost always on. Like it was swapping all the time. Memory use less that 50% of physical memory so swapping out of the question.

When booting the machine from a live USB stick, I think I saw a message flashing by which said that cache on /dev/sda was not available.

I had checked all log files, but no indication of a problem. Checked the disk with SMART, not a single error was reported. Checking the file system did not reveal any errors. Summarizing: no errors reported, ever.

I replaced the hard disk. Cloned the old disk. Access of the old disk was normal: no delay in copying while the system was booted from a live USB.

Placed the new hard disk, and all slowness had disappeared.

My only conclusion is that there had been a problem with the disk. But no errors reported. Is that possible. Could it have been a non-working cache? Would that not be reported by S.M.A.R.T.?

jlinkels

Well, run a SMART long test (not short or others) if you haven't already and post the attributes 'smartctl -A /dev/sda'.

Cool, wish I had read the complete help of smartctl and not just followed the directions I found somewhere in a forum.

I have to mount the disk back and test. Time permitting I will do so and report back. For the sake of research.

Thanks

jlinkels

You run:

where 'sda' is the device node. Then you wait for it to finish, and you run 'smartctl -a /dev/sda'.