A Network Attached Storage (NAS) device is essentially a small computer that contains disk drives, so if there is a failure anywhere between the network connection and the file system of the NAS device, it will appear the same at the user's workstation. It takes a few minutes to determine where in the NAS “chain” of connections data access is working, and once this determination is made, all subsequent steps depend upon this initial assessment.
The logical sequence of testing for NAS failures is network, share, volume, then array. Each step in this process is less expensive and less likely to result in loss of data than the step after it. Testing the disks first when you suspect a NAS failure is how you turn what would have been a simple permissions issue into a costly recovery operation.
You cannot reach it on the network. This means either an IP address has changed, there is a DHCP problem, or a network switch or physical Ethernet cable has become faulty. However, the NAS hardware itself and its associated data are intact. Connect a laptop directly to the NAS before assuming anything else.
The unit will not power on. Usually, it is the external power supply that has failed rather than the disks. Consumer NAS bricks fail regularly and the disks behind them are intact.
The volume is degraded but working. One of the disks has dropped out of the array but the redundancy (RAID) still allows everything to work. However, this situation means the margin for failure is gone. Get your data off now, before anything else.
The volume has crashed or unmounted. This is where things get serious; this is when you've hit the genuine data problem. In most situations, data recovery is possible. And it's much more likely to succeed if no rebuild has been attempted on the system after the crash occurred.
Regardless of how you interact with the vendor software, almost every consumer and small business NAS product uses standard Linux software RAID underneath (mdadm arrays), usually with LVM, and then an ext4 or Btrfs file system. Data is not written to the disks in a format that only that specific NAS device can read.
Therefore, if the electronic components have died, the disks can be read by different hardware completely and the RAID then re-assembled from their superblocks. There is no requirement to obtain a new unit of the exact same model, nor should they ever be sent to the manufacturer.
What adds steps is what's above it. Synology Hybrid RAID creates partitions on multiple disk sizes. QNAP uses LVM thin provisioning, which includes a mapping layer that needs to be rebuilt before you can even get into your file system. Both are very well understood.
Usually due to 2 failed disks in a RAID 5 or file system damage on top of an otherwise good RAID. Don't try to fix it. Don't try to rebuild it. Power it down, label all of the bays and ship off all disks.
Almost always recoverable. Any standard mdadm RAID can be read by other hardware and reassembled. There doesn't need to be a replacement unit to recover the data.
Yes, they should be labelled as per bay. The RAID will reconstruct using the full set of disks and the bay ordering is part of its make-up.
Often, a thin-provisioning metadata error rather than a hard drive failure. The mapping between the logical volume (LV) and where the blocks exist physically has become damaged. Don't let the unit try to correct the problem.
No, and this misunderstanding is behind most of the NAS jobs we see. RAID protects against a disk failure but does absolutely nothing regarding deletions, ransomware, fires or theft.
From £500 +VAT for multi-bay units and £300 +VAT for single-bay models. Once the free diagnostic is complete, the cost is fixed.