When Read Heads literally hit the platter [PICS]

So just to sum it up - this drive is 1 of 4 that failed almost simultaneously in our NAS Server at work. The NAS device in question is a Thecus N5200Pro - Quite a nice piece of work all said .

A number of array rebuilds took place in the weeks before the entire thing fell over, While rebuilding another HDD croaked it, And this last drive went out with a BANG... OK well not literally but damage to the platters was a lot more than I've seen in the past.

All 5 drives were Seagate 500GB that age almost 2-3yrs ago, The drives themselves have a tad over 2 1/4yrs Power On Time.

• In the pics that follow you'll notice that the platters are matted in some sort of smudge, the read heads we're jammed in these photos and put some nice gouged into the underside of the top platter and the top of the bottom platter.
  
• The read head themselves are worn to the metal except the one reading the underside of the bottom platter, it still had the little chip on it.
  
• Heaps of fine metallic like dusts covers the inside, and also the air filter? was packed witih it.
  
• The heads that made the gouges are almost gone, I can't get a photo of them as I don't have a good enough camera.

Sorry the photos may be a little fuzzy - iPhone camera .

Upon opening this is what greeted me


The platter is no longer shiny, just to the left of the heads is a clean section


I believe this is a air filter that captures air flung into it from the spinning platters


Metal shavings in anything mechanical is not good.




Bottom of the top platter, nice gouges.


Top of the bottom platter suffered the same as the first platter

 

I don't think you'll be able to recover any data... you've opened the drive.


()

If 4 failed simultaneously, could it have been some kind of controller/software error?

 

Wow!

I wonder how it happened. I can imagine it doing that over a period of several years - but I would have expected that it'd fail almost as soon as the process started rather than continuing to run for so long.

 

I believe the gouging of the platters happened over 3 days(Easter long weekend). My supervisor switched the NAS unit off on the Monday because it was making a racket. It was 2% into a Rebuild end of business Thursday.

A combination of both controller/software error and flaky HDD - Race condition?. Have not alot of experience with gear going out like this but I suppose its possible, but from what I've learnt the N5200 runs a software RAID through a custom Linux OS using mdadm so that could of been a factor.

This is something like 4 lines of the 20k in the System log also.

Code:

2010/04/2 20:30:18 N5200 : The system N5200 is suffering the severe disk problem. 
2010/04/2 20:30:20 N5200 : The system N5200 is recovering the RAID and rebuilding is in progress. 
2010/04/2 20:30:28 N5200 : The system N5200 change to degrade mode. 
2010/04/2 20:30:35 N5200 : The system N5200 is recovering the RAID and rebuilding is in progress. 

 

Probably from the same batch so it is possible.

 

Do you know the model numbers? I have a couple of Seagate 500GB drives that I bought about 2 years ago.

 

ST3500320AS. As Dave2972 mentioned it was probably a bad batch.

 

Bugger. That's the same model I have here that's been put aside because it was playing up. Tests fine after a zero fill, starts clicking and recalibrating as soon as it's put to actual practical use.

Firmware: SD15 Date code: 08447 Site Code: KRATSG (Thailand)

Here's a thread about how a zero fill "fixes" the SMART variables on this drive...

http://forums.overclockers.com.au/showthread.php?t=809412

 

Quite interesting, technically the Bad and Pending sector count should never change and its there for life - Hence why you be careful changing logic boards between same model drives.

Your drive is a tad bit earlier than mine I assume?
Firmware: SD15
Date Code: 08215 (5th day 21st week of 2008?)
Site Code: WUXISG

All the drives that stopped working also had 2048 Bad sectors according to the NAS software, the one pictured in this thread had 2031 Bad and 21 Pending.

Be interesting to know whats going in the drives, from what I found in this drive was the read heads wore away hence the dust.

 

I would be more inclined to argue the point that this kind of hard disk failure is more likely than others. Why? All the disks in this NAS were purchased in bulk. That means they probably came straight from the manufacturer, or via an intermediate, packed into the same crate. They would have come off the production line near each other or even one after the other, thereby ensuring that any problems with the manufacturing or materials when these drives were made would be present in all of them.

 

Getdataback can do software raid array rebuilds if you've still got N drives.

N is obsiously the rebuildable data.

ie
Raid 5 = N+1
Raid 6 = N+2

 

SD15, 3500320AS 7200.11's suffer from that nasty firmware bricking issue.
rowan194 have you made sure yours aren't affected?


I have got four 500GB ST3500320AS 7200.11, SD1A Thailand drives, shit I hope they dont suffer from whatever the f*ck ruined your disks