It’s rare that a server fails completely as a server; typically, it has lost two or more members in an array behind working redundancy. The need for urgency in order to get the downtime resolved creates tremendous pressure to take action. In our experience, the majority of the damage we see occurs during that first hour. Here are the steps to follow instead.
An unsuccessful restore operation that is writing data back on top of the original location will lose all of the information that a successful recovery would have restored. To avoid losing potential recoverable information, if there is uncertainty about the integrity of your backups (and typically there will be), create an image copy of the entire array first. Then you can decide whether or not to proceed with restoring.
The longer a server failure remains unresolved, the greater amount of financial loss it causes per hour. This generates a high level of urgency to resolve the issue. The majority of the damage we see was done in that first hour.
We recommend this for more than safety's sake alone. An array operating under degradation conditions does not operate under normal levels of "margin" – each hour that passes while the system continues to operate increases the likelihood of further data loss.
Either manually initiated or by an automatic process when a replacement drive arrives. It will read each sector of all survivors and write as it goes. This is the commonest way a recoverable array ceases to be one.
"Just to find out if it returns" – creates new metadata over the recovery information that would be used to describe a potential recovery.
You don't want to do this until you can confirm that your backup is both complete and up-to-date. If you are restoring over the original then you will lose the safety net if the backup turns out not to be complete and up-to-date.
In order, including everything previously attempted. The log entry will change the path more so than the hardware configuration; no judgment will be placed upon whatever is logged.
What can be replaced in an assembly: chassis, PSU and operating system. What can’t be replaced is what was stored on the hard drives.
The part separating a server job from a drive job and where file-level recovery quietly falls short.
A SQL Server, MySQL, or Exchange database pulled out of an array that failed is not automatically usable. These types of databases contain embedded structure integrity components (page check-sums, allocation maps, index trees, log sequence numbers); therefore even if all the bytes of the database have been restored into a file it may still refuse to attach.
Recovered databases are therefore evaluated for structural integrity rather than simply extracted. When data corruption occurs within a page level, we identify which specific table(s), record(s) are impacted and report them by name; instead of returning a database that fails to attach and requires additional effort to determine the reason for its failure. Similarly, when recovering virtual machine images, e-mail archives and other data sets containing some type of internal format, the distinction between a file count and a functional system exists.
The price is based on the reconstruction, not on the capacity.
Recovery of your server, RAID and NAS starts from £500 +VAT, confirmed after a free 48-hour diagnostic. Recovery of enterprise SAN systems is priced separately from £1,250 +VAT. Where members need physical drive-level work there is a 50% deposit; the remaining amount is due on success.
NDAs are signed as standard. All communication regarding your project will occur via one person, from the point where the initial diagnostic is completed to when the final deliverables are provided. We can invoice you instead of requiring payment prior to beginning work. Your data will remain in the UK. We include a priority fee quotation alongside the total fees for recovering your data, so you know exactly what you will have to pay prior to making an informed decision.
Send the disks to us and provide information about your server type, RAID configuration, and intended use. The diagnostic that follows costs nothing, and we will send you one figure in writing prior to starting any actual work.
No work can begin until we receive either all the drive units (labelled in bay order) or the full server. Make sure you have packed it safely and include your contact details as well. The diagnostic is free, and one written estimate will go out prior to any work being done.
Posting it? Use a tracked and insured service. The value of what is on the drive far exceeds the cost of postage. Bringing it in? Weekdays from 9am to 5:30pm, and it still needs packing as described above for the journey.
Not sure if it's worth shipping? Let us know what model server you have, how it was configured with RAID, how it failed, and what you've tried so far — an engineer reviews everything we get, and gets back to you with a sincere evaluation of the prospects of success, and a cost range, before you even post anything.
We’ll be in touch shortly. If it’s urgent, call 0161 871 0788.
The most frequently asked questions regarding recovering a server.
Turn off your server and leave the disks in their bays. Do not allow a rebuild to run, do not recreate the array and do not restore a backup over the original volume until your data is safe. The majority of the damage we have seen occurred within the first hour after the failure.
Yes, we check for structural integrity when recovering databases rather than simply extracting them. A byte-for-byte accurate file may still fail to attach because of a capture of a database while it is in mid-transaction, or because the file contains damaged pages. Therefore, we locate which tables and records were impacted instead of trying to provide you with a "store" that will fail to attach.
Yes, all disks from the array including any previously labelled as failed — an ejected member frequently retains readable information and is quite often the difference between successful partial and complete recovery. Label each disk by its bay before removing it.
The diagnostic is free. Most diagnostics are back with you within 48 hours. Normal array work usually takes a few working days to complete once approved; however, if we need to use donor parts or perform extensive imaging, the time frame will be included in your estimate.
Yes. Both, as part of our normal process. All work goes through one designated point of contact, all data remains within the United Kingdom, and established businesses are invoiced rather than paying up front. We also include purchase order reference numbers when required by financial departments.
Data recovery prices start at £500 plus VAT for servers, RAID systems, NAS devices etc., fixed in writing after a free 48-hour diagnostic. Enterprise SAN system recoveries are priced differently, from £1,250 plus VAT.
Server chassis vary by manufacturer... failure modes don't. Cache battery failures, blackout losses and rebuild attempts started on marginal disks account for almost all of what we see from every single platform.
From £500 +VAT, and we'll quote separately for enterprise SAN systems starting at £1,250. In addition, we offer out-of-hours collections throughout Greater Manchester where a room full of people is waiting on it. Also, we'll assign a named contact to assist you from diagnostic completion to delivery back into production.