Hard drive failure

Preamble:- this Forum re hard-drive failures & wretched Mirror RAID got me thinking along new lines hence new thread.

On seperate desktop pc's using XP and 80GB optical internal hard drives of various brands wife & I have had nine HD failures in six years (we are octogenarians with home use only). So they fail, not IF but WHEN. Of course using Mirror RAID has doubled the HD in use but at least we have never had the ultimate catastrophe.

The Promise RAID Controller has given me much grief. Backing-up now done to a single migratable 5OOGB external HD, amply big enough --- but it is still optical.....

Not likely that the external AND an internal would fail at the same time I agree. But sod's law always applies and bread always falls buttered-side to carpet

I am considering a different arrangement & invite comments. I won't throw away my big external HD now I have it but I will add a desktop USB2 hub and a solid-state 64GB flashdrive (aka memory stick) of a known make say Verbatim. I will back up to that.

Just how reliable are these solid-state devices? Rock-solid never fail?

I won't need to run CHKDSK, Defrag nor AV scan with it cos already doing that regularly with the internal optical HD. It will be I hope fast enough on USB2.

Costs? Approx £1 per extra GB once you are on the ladder so £35 for 32GB & £90 for 64GB & £160 for £128GB so my choice of 64GB will be £90 (can be had generic for £50 though) --- 7-port triangular USB2 hub £25, 10-port chunky £30, simple 4-port £10. If I choose carefully won't need £9 USB2 cable to insert flashdrive as hopefully slim enough for direct insertion? Warning, make sure all suit 32bit as reviews indicate possible probz with 64bit which is not me anyway

I propose issuing my own commands to backup, not say a Genie auto but £30 spent in that direction does have its attractions which include rescue disc/flashdrive.

In reply to a post by Littleseen:

Just how reliable are these solid-state devices? Rock-solid never fail?

Not very. Well...the hardware part isn't so bad but there's a lot of complex management involved in what's called 'wear levelling'. These algorithms can go wrong and there's a lot of people encountered problems with it.

http://en.wikipedia.org/wiki/Wear_leveling

I had my own brush with an SSD a few weeks back.

http://forums.thinkbroadband.com/general/t/4088333-r...

Thankfully solved (and it's backed up anyway) but it makes a good example.

My current solution is to backup a disk image to an external HDD. My server has a small internal SSD and the data is on another external HDD. The backup is on a rolling cycle where every fifth week is a full backup and the others are just differential (the difference between that week and the full backup). But in my case there's almost nothing totally irretrievable. It's music and video I've ripped from my own CDs and DVDs plus some photos most of which are scanned from slides.

For me a drive failure would be a nuisance rather than a disaster. I've made sure I can restore the server to full working condition by restoring the SSD so total downtime should only be as long as it takes to order a replacement drive. I use Acronis to backup but although it works it's slow and clunky and I couldn't recommend it. I get the distinct impression that the company is only in it for the money and more interested in adding on bells and whistles than actually producing a quality product.

One comment I'll add:Test your recovery strategy. I got the chance to test mine when I moved from a 250GB HDD to the 64GB SSD. Although it's slow and klunky Acronis can indeed restore a disk volume and resize it in the process. It would also have had to relocate files since that was originally the data disk and nearly full. All credit there - it resized the volume and it was still bootable.

Edited by Andrue (Wed 22-Feb-12 10:22:42)

Strewth? 9 failures? What on Earth are you doing?

Don't drop your PC's onto a concrete floor from 20 feet and start again!

I think the "average" lifespan of a hard disk is generally reckoned to be in the region of five years. I've currently got about a dozen or more devices containing hard disks (some of them more than one disk) and I've had 3 that I can remember fail on me in the last 10 years. Experience at work, with thousands of PCs, is about the same rate of failure.

9 failures from 2 PCs in 6 years is far beyond any statistical expectation. Perhaps the OP has a habit of moving computers whilst they are switched on (and even that shouldn't be a problem with a laptop).

er, just what do you, kind Sir, mean by "move" as I do slide the active pc across the desktop to access its back to swop say a USB lead? Is that naughty? If so I'd best forget plug & play and switch off first?

Not bumped in above process but nevertheless "moved"

Moving in the way that you describe can certainly lead to premature disk failure. Laptop disk drives have some degree of shock protection built in, but most desktop ones do not. It doesn't take a great deal of movement to cause the disk heads to bounce against the disk surface.

If you need to plug and unplug equipment (USB only - most other equipment should not be unplugged whilst the computer is powered up) it would be better to run an extension cable from the back of the computer, perhaps to a USB hub.

In reply to a post by AEP:

I think the "average" lifespan of a hard disk is generally reckoned to be in the region of five years.

I think it's getting worse though. I've had more failures in the last four years than in the fifteen years before that. The two most recent failures were still within warranty - sent the second of them on Monday for a replacement.

What I found most puzzling about these last two were that they were sat in a server. That means a 24/7 duty cycle with no stop/start to stress them. It's a VM host so sometimes has a lot of drive activity but mostly it just sits in a corner doing nothing.

I believe there were some problems with the very high capacity drives. Having said that, I have 2 1TB drives and 2 2TB ones and - at the risk of tempting fortune - have had no problems with any of them.

It does seem that many drives run hotter nowadays than they used to. Perhaps this means that, unless the disks are very well ventilated, continuous use is not the advantage that it used to be.

I thought SATA drives were plug n play?

Could well be (with appropriate BIOS settings). The same is true of SCSI. Or indeed Thunderbolt. But none of these interfaces comes as standard on many PCs. And I wouldn't recommend moving a running PC to plug any of them in.