[ale] disk drive lubricant

Mark Wright mpwright at speedfactory.net
Mon May 15 10:32:13 EDT 2006


I used to work for a mainframe disk manufacturer.  One of the issues  
that all disk manufacturers face is the lubricant.  one line of disks  
we produced failed because we were using new (in the 80's) sputtering  
technology to put the oxide on the platters.  The engineers were so  
proud of the fact that they could fly the heads at 7 microns because  
the media surface was so smooth.  The only problem was that after  
about six months the lubricant would migrate out across the platters  
and get on the low flying heads.  It would change the aerodynamics of  
the head and lift it up too high to read the data.

Any way the point of the story is that lubricant on the media is a  
known factor in disk design (one the above mentioned engineers over  
looked).  One of the ways it can bite you is when the drive is  
powered off the head comes to rest on the platter, the lubricant  
cools and the head gets stuck.  I have some mid 90's technology  
controllers that used 100 Meg laptop disks current at the time, to  
boot from.  They only get booted maybe once a year.  If you power one  
off for 30 minutes you have a fifty- fifty chance it won't come up.   
Our fix has been to take the drive out and bang it on a table top  
pretty hard.  This pops the heads loose and the thing boots.

If you are going to use hard disks for archival storage  make it  
battery supported RAID and never power them off.  Then write a script  
or a program to incrementally read and re-write the data periodically  
so that every few months you know that all the data on the array is  
fresh.  (sometimes data written a year ago is hard to read due to  
mechanical wear like in the unusual story above.)

I wouldn't worry about the temp or the humidity except to make sure  
that it is relatively constant.  Drives can handle most anything but  
change.

What fun,

Mark




On May 14, 2006, at 6:44 PM, Dow Hurst wrote:

> Greg,
> Not so urban legend as an DEC engineer told me, about a decade ago,  
> that
> a SCSI drive can lock up due to returning to room temperature for too
> long.  He explained that the normal running state had the disk drives
> warmed to a normal running temperature.  When a drive wouldn't restart
> after a machine failure that he had fixed, he would try warming the
> drive with a hairdryer and then try restarting the server.  I lived in
> fear of our SGIs being down too long with the drives getting cold!!
>
> I am interested in the lubricant vapor pressure used in disk  
> drives.  If
> you are correct that lubricant can get deposited on the platter  
> surfaces
> and heads during storage, then that is a disk design issue.  There
> should be white papers from manufacturers on that topic available.  I
> wonder if humidity, pressure, and temperature changes of the  
> environment
> where the drives are stored affect the lubricants?  Should they be
> stored under nitrogen in a constant humidity environment.  At that
> point, the durability of tape medium and storage costs should come  
> into
> play.
>
> I think mirrored RAID 5 servers are a wonderful thing.  My personal
> dream is to have two separate servers with multiple times the storage
> capacity of the current amount of data that are mirrored to each other
> with power protection and in separate locations with fat bandwidth
> between them.  Then, the wet part of the dream is a third secret  
> server
> that only I know about in a hidden location that is the ace in the  
> hole.
> Dow
>
>
> Greg Freemyer wrote:
>> Are you planning to power-down the drives between uses?
>>
>> If so, there is NO data or gaurentee on disk drives maintaining data
>> with power off.
>>
>> We do it a lot around here, but we also make a tape backup before we
>> put the drives on the shelf.  Personally I think data loss starts to
>> happen in the 1 to 2 year range.  I do know some people pull drives
>> off the shelf every once and a while just to power them up and give
>> the grease, etc. a good workout.
>>
>> I've even heard a guy one time say there were disk drive preheaters
>> that could be used to warm a disk drive (and associated grease etc.)
>> up.  Treat that as urban legend.
>>
>> Greg
>>
>> On 5/10/06, Allan Metts <ametts2 at mindspring.com> wrote:
>>
>>> Hi everyone,
>>>
>>> I have several hundred Gigabytes of data that's in my way.  I  
>>> need to keep it, but I don't need regular access to it.  More  
>>> data is coming in now that will eventually need the same treatment.
>>>
>>> So I'm looking for a good solution for off-line data storage.   
>>> Here's what I'm looking for:
>>>
>>> --*-- Rack-mountable
>>>
>>> --*-- Non-proprietary hardware, no special "backup" software
>>>
>>> --*-- Supports Linux distributions.  I choose which one.
>>>
>>> --*-- Uses readily-available and inexpensive IDE drives.
>>>
>>> --*-- Easy to set up and administer.  We have no time to fiddle  
>>> with this -- I'm thinking 1) mount the drive, 2) copy the files,  
>>> 3) unmount the drive, 4) remove and replace the drive.  The  
>>> machine still needs to work when I pop in an unformatted or  
>>> unrecognized drive.
>>>
>>> --*-- Hot-pluggable drives would be nice if it doesn't make  
>>> things complicated.  I'm willing to power down the machine to  
>>> swap the drives if need be.
>>>
>>>
>>> I think what I'm looking for is a bare-bones rackable server,  
>>> with an internal IDE drive to boot Linux, a CD-ROM drive to  
>>> install Linux, and two removable drive bays that I can dump the  
>>> data to.
>>>
>>> Suggestions?  Recommendations?
>>>
>>>
>>> Best Regards,
>>>
>>> Allan Metts, VP -- Technology & Operations
>>> AirSage, Inc.
>>> ametts at airsage.com
>>> (404) 861-3404
>>>
>>> _______________________________________________
>>> Ale mailing list
>>> Ale at ale.org
>>> http://www.ale.org/mailman/listinfo/ale
>>>
>>>
>>
>>
>> --
>> Greg Freemyer
>> The Norcross Group
>> Forensics for the 21st Century
>> _______________________________________________
>> Ale mailing list
>> Ale at ale.org
>> http://www.ale.org/mailman/listinfo/ale
>>
>>
>
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale




More information about the Ale mailing list