Anyone else noticing a trend? Servers offline? Data corrupt or inaccessible? Online backups not online?
How long should an SSD last?
Longer than a HDD?
With the increase in SSD adoption this last 5 years often for perceived reasons of reliability, transfer speed or because of the “belief” that there is higher risk of HDD failure, a post from those that are on the front line of data recovery and data protection.
An alternative reality… one that has to be factored into business plans and research projects.
SSDs can die within hours of being first switched on for exactly the same reasons HDD become inaccessible.
Power spikes, interruption and fluctuations affect any storage at any scale.
Observations
💾 The largest single disaster recovery engagements have been of magnitude 1.5PB+, one was a 15 years old SAN (500 x 300GB) The fibre channel HDD affected after a SAN power supply and UPS were unable to handle a lightening strike to the grid supplying a data centre - “Holes in data stores” filled 4 days after drives were brought to Security House for repairs, The SAN is still in operation.
📀 A 450TB NAS using HDD that were regularly crashing, it was assumed because so many of the HDD had bad sectors and some severe platter damage that the HDD were a faulty batch, the actual cause was an SSD RAID group used for O/S and HDD RAID controller. The SSDs own power management was the cause, VMs and Veeam backups declared corrupt and unrecoverable recovered from the HDD.
🧱 SSDs bricked after 26,000, 32,000 and 40,000 hrs has been a regular feature of 2020 and 2021 but also some after just a few months of intensive writes and deletions - many associated with Databases or video editing, most unrecoverable in an useful timeframe and requiring considerable research to develop solutions.
Macbook onboard SSD - logic board fails or NAND chips fail - Apple say “buy a new logic board or replace under warranty but data is gone”.
All data storage has its issues when it comes to a disaster. Consider that as a lab R3 engineers are the last line of defence for data if a backup fails.
R3 do recover from many types of SSD failure, overfull, bad sectors / blocks, component failures but if NAND flash is dead the chances of recovery plummet and is a slow process, worse if just one NAND dies in a group of 2 or more all data is affected.
Research - but who and how is that financed?
R3 funds R&D and investment in infrastructure by maximising success rates for HDD data recovery, hundreds of millions of HDD in use and some have accidents but most are recoverable.
Success rates keep the lights on 💡
The growth in adoption of SSD and increasing flow into Security House of failed ones is a concern not just to data recovery engineers who make their livings from the catastrophes of others but also to lab owners who care about data protection.
R3 Data Recovery Ltd Britains data recovery Hospital for Businesses and Consumers (VM VHD HDD SSD).