Seagate drives hold an unusually high percentage of their operational code within a reserved service area on the disk surface. If one portion of that code fails to be readable, the drive will spin perfectly and report nothing — and the user data remains completely intact right behind it.
A large share of these units show up spinning perfectly, but when you check the parameters reported by the drive they will either report incorrect capacity, blank model strings, or simply report nothing at all. This is a service area failure — one of the better outcomes available.
Understanding this is the most important piece of information regarding Seagate recovery.
All hard disk drives contain firmware modules stored on the actual disk platter, as opposed to being stored solely within a circuit board component — the defect map, the address converter which translates logical addresses into specific locations on the disk platter, and adaptive calibration values specific to each individual unit. Seagate’s design does place a good deal of these elements on the same disk.
If a module in that region becomes unreadable, the drive will fail to complete its initialisation. The drive appears to spin-up as normal, appears to function perfectly, reports zero capacity, an unknown model name or an unrealistic size. This occurs since the drive is unable to tell the computer about itself. Your files were untouched.
These are accessed via the manufacturer's diagnostic interface. The interface provides direct access to the service area allowing for reading and repair of damage done to the service area: defective modules can be restored/replaced with good ones from a donor unit that matches the exact family and firmware revision, and translators can be rebuilt using surviving defect lists. This is precise and non-invasive work that doesn't require opening the drive.
A useful note if you have an older Barracuda sitting in a drawer somewhere.
There was a known problem with certain Barracuda 7200.11 units that resulted in a bad firmware bug. When an internal event log became full, the drive would lock when attempting to boot up, reporting a busy state and never becoming ready. Other affected units simply stated they had zero capacity.
The issue with this particular fault was the drives were completely unharmed. There was no mechanical damage, there were no bad sectors, and no physical wear on the disks; instead, a logic error in the firmware caused an otherwise perfectly operational disk to fail to present itself. Recovery is via the diagnostic interface, using either clear or rebuild functions on the affected structures.
These continue to arrive, usually from archives and drawers where a drive was put away years ago and recently required access. Because they never had anything physically incorrect with them, prospects are very good for recovering these disks.
It affects both how the drives operate in terms of failure behaviour and how quickly your image will complete.
When you use SMR technology, the overlapping of adjacent tracks allows for increased capacity, but when you write data to one track, you are essentially rewriting all of the adjacent tracks that also contain data. Once you have written to a track, the drive caches the write and continues to work in the background reorganising the data. When you write continuously for an extended period of time, the cache becomes exhausted and data throughput drops off significantly. At some point during this process, the drive appears to fail when it has not actually done so.
SMR has been used by Seagate in a number of their consumer level Barracuda drives. While it will work well enough for archival purposes and just about any type of general purpose usage, when you need to do any kind of write-intensive operation on your data then it is a poor choice, particularly when using it as part of a RAID setup. The process of rebuilding the RAID array is a continuous write operation which is something that SMR performs very poorly under.
If you have to recover some data off an SMR-based device then it will take longer to image the disk. Where a background reorganisation was interrupted, there could potentially be data still in the cache rather than at its final location.
Note whether it shows up in the BIOS, whether it shows up at zero capacity, whether it shows up with no model number assigned, or if it shows up but cannot be read. These are all four separate possible diagnostic results and they help narrow down the work prior to receiving the drive.
Updating the firmware will rewrite the service area on the drive. The service area is the one specific portion of the drive that needs to be read during diagnostics. When working with a drive that cannot initialise itself, updating the firmware could render this issue permanent.
Repeatedly restarting will provide you with no results for a firmware error, and there are risks involved of writing additional times to the same faulty areas.
Model and family, along with firmware revision, determine what type of donor can be used; thus, a photo of the labels speeds up identifying the diagnostic route for the device.
In most cases no. This is the most common failure mechanism of Seagate drives. Drives store firmware modules in a service area located on the disks themselves. If a firmware module becomes unreadable it cannot describe itself, so it reports nothing, but the data stored on those drives remains intact.
Reserved space on the disk that contains the drive's own operating code: the defect map, the translator that converts logical addresses to physical disk addresses, and calibration values specific to each individual drive. The diagnostic interface provided by the manufacturer reads and repairs this space.
No. The update rewrites the service area — and the service area is what a recovery needs to read. A failure of the update on a failing disk can result in a permanent fault.
There was a firmware defect in some 7200.11 generation disks with internal logs filling up and locking at power-up, reporting busy and never becoming ready. They did have no mechanical issues — they were perfect mechanically. These were all due to logic defects and thus there is an overall good prognosis.
Yes many consumer Barracudas do. SMR increases storage capacity but when the cache fills, you get a huge drop off in performance for sustained writes. Thus these drives are very poor choices for RAID members — a rebuild is exactly the type of continuous writing that SMR performs most poorly.
From £300 + VAT for a single drive after a free 48-hour diagnostic. Service area and firmware work does not involve physically opening the drive. Therefore it often avoids the deposit that physical work carries.