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

Re: 3 minor issues

On Sat, May 22, 2004 at 09:47:59AM -0500, Marc Williams wrote:
> On Sat, 2004-05-22 at 09:26, ne... wrote:
> > On May 21, 2004 at 22:25, Marc Williams in a soothing rage wrote:
> > 
> > >1) My IDE HDD spins down on a reboot.  It's never done that with any
> > >other distro that I've run on this machine.  I don't think I like it. 
> > >Is this now normal?
> > Yes, it seems so.
> > 
> Rats.  I would think that this behavior would put unnecessary stress on
> the HDD.

As far as placing unnecessary stress on the disk I would not give it a
second thought.  

It is unlikely that you will be rebooting a large enough number of
times that this matters at all.  I expect your uptime will be measured
in terms of weeks once you have the OS installed.  98% of all
configuration changes do not require a reboot.  In most cases a simple
software command "service XYZ restart" is all you need after a change
to a config file for it to be activated.

As far as the spin down and seek to the landing zone this is a
desirable behavior for many disks.  With the very low flying height
of modern heads and that heads and media surfaces are very flat and
'stiction' is a common issue. Depending on the vendor and the drive
minor amounts of stuff can build up on the heads and a drive that is
halted and allowed to cool has a chance of sticking.  A touch and go
restart sequence tends to clean the surface of the heads and minimize
(scrape off) this buildup of crud.  About eight/ten years ago almost
all the disk vendors had this problem with a new generation of media
and heads and the microcode of the disks was enhanced to permit a
touch and go cycle to minimize this as a problem.

Since the internals of most disks are designed for the consumer market
(media, heads, motors) the designs expect a power cycle a day or more
because that is how people use them.  It is also very common in places
like Japan to turn off equipment for the night (plays havoc with cron
jobs and system housekeeping).

Some readers will correctly note that this stiction thing is OLD news
but the key point is that historically all disk vendors suffer risks
in this regard. A spin down land restart cycle is considered by some
to be important in routine operation.  There may be folks out there
that design/support large RAID systems with racks of disks (100s).
They may be able to comment about the importance of uninterruptible
power and touch and go cycles with current devices.  I know that in
planning the physical move of of large disk farms that I have been
involved in there was a step where the drives were conditioned (spun
down/up) a number of times prior to power-off and letting them cool for
the move.  I have seen multiple disks fail in this conditioning
cycles.  Since the RAID was active it was simple to replace the drive
and rebuild, no harm done.  

	T o m  M i t c h e l l 
	/dev/null the ultimate in secure storage.

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