Data Recovery Case File · NAS & Network Storage · Which Group They Are In
Three Failures Across Twenty-Four Disks May Be Survivable or Fatal Depending on Their Grouping
This trade enquiry asks for a budgetary figure before committing to anything, which is reasonable and also difficult to answer well. Twenty-four disks across an appliance and its expansion unit, of which "three have failed a media integrity scan but do power up", in a single-parity configuration. The count alone does not determine the outcome: what matters is whether those three sit in the same group, and single-parity tolerates only one per group.
| Media | Twenty-four disks across a network appliance and expansion enclosure in single-parity configuration — three members failing an integrity scan while remaining powered and responsive |
| Reported situation | Twenty-four disks deployed across an appliance and an expansion unit · single-parity redundancy in use · three members failing a media integrity scan · all three continuing to power up · indicative costing requested before despatch · final assessment accepted as requiring inspection |
| Fault class | Multiple members flagged by integrity scanning with grouping undetermined — survivability governed by distribution across parity groups rather than by total count |
| Equipment used | Group membership established before any survivability assessment was offered · no rebuild or repair permitted at the appliance · all members imaged individually write-blocked under strict per-sector timeouts · flagged members assessed for genuine read failure against scan reporting · group geometry reconstructed from member metadata and volumes assembled offline |
The decode: why the total is the wrong number to look at
What single-parity redundancy actually promises: one failure per group. A group survives losing any one member and cannot survive losing two, whatever the size of the wider installation.
Why twenty-four disks are not one group: large installations are subdivided. An appliance with an expansion unit commonly runs several separate groups rather than one enormous one, precisely so that a failure is contained.
What follows for his three: the arrangement decides everything. Three failures spread one per group is three survivable situations, and three in the same group is a group beyond recovery by redundancy.
Why that cannot be answered from the enquiry: it needs the configuration. The grouping is recorded in metadata on the members themselves and read from them directly.
Why the budgetary question is genuinely hard to answer in advance: the range is wide. Reading a handful of flagged members is a different proposition from imaging twenty-four, and which applies depends on the grouping.
What can be said usefully before anything is sent: the cost structure. Assessment is free, work is charged per drive, and the number of drives needing work follows from the grouping — which gives him a range rather than a figure.
Why the detail that they still power up matters: a scan flag is not a death. Media integrity scanning reports members returning errors, which is a threshold rather than a failure.
Why that is genuinely encouraging: a responsive member can be imaged. A drive that powers up and returns most of its content is far more useful to a reconstruction than one that does not respond at all.
Why even three in one group may not be final: partial members combine. Where each flagged member fails in different regions, a complete volume can frequently be assembled from the union of what they return.
Now the action to refuse, and it is what an appliance will suggest: replacing a flagged member and rebuilding. A rebuild reads every remaining member under sustained load, which is how a second marginal disk is pushed into failing.
On the bench
Group membership was established before any survivability assessment was offered — single-parity redundancy tolerating one failure per group regardless of installation size, while appliances with expansion units subdivide into several groups precisely to contain failures, so three flagged members spread one per group differ entirely from three within one. Scan flags report members returning errors rather than members that have died. All members were imaged individually before offline assembly.
The outcome
Group membership established before any assessment was offered, no rebuild permitted at the appliance, and all members imaged individually. Free assessment, one fixed written figure including VAT, charged per drive, with 50% of parts and labour upfront where a drive has to be opened. The decode: three out of twenty-four is not the useful number. Single-parity tolerates one failure per group, so where those three sit decides whether this is three ordinary situations or one serious one.
Several members flagged in a large installation
Refuse any offer to replace a flagged member and rebuild — a rebuild reads every remaining member under sustained load, which is exactly how a second marginal disk gets pushed over. Then find the grouping before drawing conclusions from the count: single-parity tolerates one failure per group whatever the size of the installation, and large appliances subdivide precisely so failures stay contained. Members that still power up and return most of their content are far more useful than the flag suggests.
Find the grouping first — call Cambridge Data Recovery on 01223 655015; membership established before any assessment, members imaged individually, geometry reconstructed and volumes assembled offline.
Request a quote online →
Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.