[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: How to fix I/O errors?

On Monday 06 February 2017 10:22:54 Joe Pfeiffer wrote:

> David Christensen <dpchrist@holgerdanske.com> writes:
> > On 02/04/17 07:18, Ric Moore wrote:
> >> I'm looking at a Seagate 750 gig drive that went south on me with a
> >> pile of errors. Good luck getting Seagate to give a good gosh darn.
> >> In the past I have had mixed results replacing the drive
> >> motherboard. I saved two out of three. I doubt I will buy anything
> >> Seagate makes in the future.
> >
> > Everything electrical and mechanical fails.  It's just a question of
> > when, followed by whether or not you're prepared.
> >
> >
> > I've found (and heard) that the worst thing I can do to a HDD is put
> > it on the shelf and let it rot.  I've had more than a few that
> > failed shortly after being put into a computer.
> I hadn't heard this...  I've got a drive I've been keeping as a cold
> spare.  Am I better off (in the sense of "is it more likely to
> actually be useable when I need it") installing it and adding it to
> one of my RAID1 arrays?  Can you point me to an article about it?

That cold spare will eventually develop stiction, seizing the parked haed 
to the surface of the disk solidly enough that the disk motor cannot 
break it loose to spin the disk up.  Such is best treated by hooking up 
the cables, but holding the drive in your hand so that you can turn on 
the power, and within a couple seconds, give the drive a good sideways 
blow on a corner with the ball of the wrist so the drive housing/casting 
is caused to rotate a few degrees around the axis of the disk, breaking 
the stiction so the spindle motor can spin it up. The theory is that the 
drive frame rotates when you drive it by hitting the corner, but the 
disk doesn't, breaking the stiction seal.

Based on my experience here, with terabyte drives, they seem to be better 
off spinning even when not in active service. I have one old terabyte 
drive thats had 25 bad, re-allocated clusters, since the first time I 
had smartctl do an extended self test on it several years ago. It still 
has 25 reallocated sectors this day, with (word wrapped):
 5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  
Always       -       25
  9 Power_On_Hours          0x0032   030   030   000    Old_age   
Always       -       61430

And that drive gets beat on every night as its my backup disk containing 
all the virtual tapes amanda uses.

In terms of spin time, thats 6.997490303 years.  And thats a Seagate 
Barracuda drive, which has a horrible reputation according to these 
mailing lists.

One secret though. When that drive was new to me, I went to the seagate 
web site and downloaded a cd image for that model that updated its 
firmware. As I already had a linux install on it, I applied it to that 
drive with a bit of trepidation. But worry wasn't needed, I didn't lose 
a byte of the install. But a side benefit was that the drives speed was 
nearly doubled.

If you can afford the time, I highly recommend putting the latest 
firmware in it before putting it in service. Who knows how long its been 
on the dealers warehouse shelf, but the dealer bought 10,000 when they 
were announced. And the initial shipment can be guaranteed to have been 
bagged with Alpha rated firmware in it.  Always update new drives is the 
message from this elderly (82) user.  You won't regret it.

Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

Reply to: