Case Study
Two drives down on a RAID 10, a third about to follow.
An HPE ProLiant MicroServer running a business's VMware virtual machines lost its RAID 10 redundancy when two drives failed and a third began to go. The array was rebuilt and the VMs restored in our Belfast laboratory.
The case
A RAID 10 MicroServer, past its design life.
This HPE ProLiant MicroServer belonged to a small business in Northern Ireland, configured as a four-drive RAID 10 array: two Seagate IronWolf drives paired with two WD Red drives, giving 7.5TB of usable storage. The server ran the business's VMware virtual machines — the platform its day-to-day operations depended on.
RAID 10 combines mirroring and striping, and this configuration is built to tolerate up to two drive failures without losing data. It had been a dependable, cost-effective setup for running the company's VMs.
What went wrong
Two drives failed. A third was following.
The drives had passed their 3-year, 24/7 design life. Two of the four had already failed outright, and a third was showing clear signs of imminent failure — leaving the array's redundancy exhausted and its VMware VMs unreachable.
The server arrived at the lab in this state: a RAID 10 array with its data striped and mirrored across four ageing drives, two of them already dead.
The recovery
Rebuilding a striped, mirrored array from failing drives.
- 01
RAID structure assessed
The array's RAID 10 layout was identified — how the data was striped and mirrored across the two IronWolf and two WD Red drives — before any recovery work began.
- 02
Drives imaged individually
Each drive, including the third showing signs of failure, was imaged separately so nothing further was risked on the original media.
- 03
Array reconstructed virtually
The RAID 10 set was rebuilt from the images, reconstructing the striped and mirrored data needed to bring the VMware virtual machines back.
The outcome
The virtual machines, restored.
The recovery salvaged the business's critical data and restored its VMware virtual machines — the systems its daily operations ran on — avoiding prolonged downtime and further disruption to the business.
A RAID 10 array that has lost its redundancy is running on borrowed time. Get it imaged and assessed by a lab that rebuilds server RAID for a living.