[Bug-ddrescue] Tip: Improving recovery chances with Seagate "F3" models

[Franc Zabkar]

Seagate models since the 7200.11 incorporate what is referred to as "F3 
architecture". That is, they return the "F3" prompt when accessed via their 
serial diagnostic interfaces.

These drives often go offline when they encounter difficult sectors. A 
power cycle is then required to bring them back online. The problem is that 
the drive gets hung up trying to reallocate bad sectors and remains BUSY 
for too long, causing the software to timeout.

When this happens you will see "InitiateMarkPendingReallocateRequest for 
disc_lba" messages at the serial console. By modifying certain "Congen" 
parameters, one can ...

- disable read sparing, write sparing, and correction
- disable Background Media Scan
- disable offline sparing and Deferred Auto-Reallocation

One can also disable or limit the number of ReadRetries. The default is 0x10.

See http://malthus.zapto.org/viewtopic.php?f=83&t=557&p=2805

A serial TTL interface can be constructed in a variety of ways. There are 
also off-the-shelf USB-TTL_serial bridges costing less than US$10.

Note that the above commands can be dangerous if used incorrectly. In 
particular, be aware that F3 commands are CaSe SenSitive. Using the wrong 
case, or entering the commands at the wrong level, can have data 
destructive consequences.

A complete set of F3 diagnostic commands can be found at The HDD Oracle, as 
can the instructions for the serial diagnostic console.