Platter Damage - Chances of Any Recovery

[alfarom]

My last post seems to have disappeared so I'll try again.

I run a computer repair company and refer any advanced data recovery to a specailist company. In return I get a 15% referral fee whether I deal with the customer or the data recovery company does. I'm trying to up my knowledge to save wasting the customer's, mine and the recovery company's time but weeding out lost causes, etc. One thing I can't seem to find out online concerns physical damage to a platter. If only one side of a platter is damaged does that mean that it should still be able to recover the data from all the others in the drive? I know there is the chance of data not being complete due to fragmentation when the data was written to the drive but does the nature of the technology mean that one damaged platter means all the data is lost?

Apologies for such a basic question but it would really help to get a few pointers on this..

[mhp666]

That a dish is damaged does not mean that you lose all the HDD information, or even the complete information of that dish.
Normally when the heads touch the plate or a head breaks it can deteriorate a wide area of the plate, but some of the information can be obtained.

In general, if the files are not large, you can recover a large part of files without faulty information.

If I have not explained myself well or have a question within my limited knowledge, I will try to answer it.

Greetings.

[alfarom]

No, I appreciate your answer. Thank you..

[pcimage]

We do this almost on a daily basis

Depending on the drive model, level of damage and which surface is damage we get varying levels of results. Also the nature of the data plays a part, with photo recoveries being mostly successful and larger files like video and large databases for example being less so.

Just this week had a Hitachi 1Tb 2.5” with H0 black with dust, so “physically disabled” H0 and persuading it to initialise from H1 I managed to image 75%. But as it was HFS th vast majority of the catalog file was missing

So I had to find a way of getting at least some part of at least the catalog file from H0.... after a couple of donors and a lot of platter cleaning etc. I got a reasonable read of the catalog file and got a pretty good recovery. There are extensive but not deep scratching to H0 surface.

Of course this wasn’t cheap and most clients think that because they are only getting a 75% Recovery it should be cheaper, when in reality it’s a much tougher job and costs of multiple donors increases costs further.

So it’s only relatively small % of clients who agree to the higher costs for a less than perfect recovery.

[Jidis]

after a couple of donors and a lot of platter cleaning etc.

I had a related question for a while if alfarom doesn't mind-

In most head crash cases, once new donor heads are in place, do you typically have to totally avoid whatever area the crash took place at, to keep from ruining the new heads, or is that usually just an area you won't be able to read? Does it also block head access to the corresponding area on the opposite undamaged side of the platter, or do you work around that somehow? (I realize there are varying levels of surface damage, I'm just wondering about the majority)

Thanks!

[mhp666]

In my opinion.
The Zone where you have touched the heads can have three possibilities:
a) You can continue reading without problems. Rarely.
b) A series of defects are produced in the area and there are usually enough defective sectors.
c) Deep scraping of the plate occurs. There are many possibilities that every time you pass your head it breaks.

In theory the opposite area should not be canceled because it is in the process of reading a sector the head approaches the plate to read, but the opposite is not close and should not be damaged.

When I replace the heads I apply in the DE a timeout of 800 ms and a jump of 1 million sectors to avoid those areas as soon as possible in the first pass, afterwards I analyze and make decisions.

[Jidis]

Much thanks for the explanation mhp666!

[Th3_uN1Qu3]

With healthy drives reading at good speed after a head swap, I start with even more aggressive timeouts, down to 150ms even. The idea is obviously to grab as much data as I can, as fast as I can. Then I progressively increase the timeouts on subsequent passes if the drive hasn't yet been able to read the files the client needs - 300, 500, 1000, 1500ms. I've had good success with this workflow.

For skip size I usually start with 100.000, then 10.000, then go down as needed on the final passes. On an average drive I get the needed data in 4-5 passes with these different settings.

[mhp666]

With healthy drives reading at good speed after a head swap, I start with even more aggressive timeouts, down to 150ms even. The idea is obviously to grab as much data as I can, as fast as I can. Then I progressively increase the timeouts on subsequent passes if the drive hasn't yet been able to read the files the client needs - 300, 500, 1000, 1500ms. I've had good success with this workflow.

For skip size I usually start with 100.000, then 10.000, then go down as needed on the final passes. On an average drive I get the needed data in 4-5 passes with these different settings.

I'll try this copy pattern.
I think that how you do it can be quite interesting so as not to damage the heads.
Thanks for the input.

[alfarom]

Thanks for all the replies, much appreciated.