RAID 1 is the simplest way to protect a computer, server or small NAS against a single failed drive: two drives hold identical copies of the same data. Plenty of home offices and small businesses rely on it. But when a mirror fails, it often fails in ways people do not expect, and the most natural reaction, swapping a drive and letting it rebuild, is sometimes the step that loses the data. Here is what goes wrong with RAID 1 and how to handle it safely.
How a RAID 1 mirror works
Every write is sent to both drives at the same time, so each drive holds a full copy. If one drive fails, the array keeps running on the other one in a “degraded” state. When you install a replacement, the controller or NAS copies everything from the surviving drive onto the new one. That copy is called a rebuild, or resync.
The setup protects against one thing: one drive failing. It does not protect against anything that changes the data itself.
Common ways a mirror fails
- The second drive fails before anyone notices the first. Degraded warnings are easy to miss if email alerts are not set up. Both drives are often the same model, bought on the same day and used in the same way, so they wear out at a similar pace.
- The rebuild pushes the surviving drive over the edge. A rebuild reads every sector of the remaining drive. If that drive has weak heads or bad sectors, this is when it is most likely to fail.
- The rebuild runs in the wrong direction. If an old, out-of-date drive is put back in, or the wrong drive is marked as good, some systems will copy the stale or blank drive over the good one.
- The controller or NAS loses the configuration. After a power outage, a motherboard replacement or a controller swap, the array may show as missing, foreign or offline even though both drives are fine.
- The data is damaged on both drives at once. Accidental deletion, a bad software update, file system corruption or ransomware is copied to the mirror instantly. A mirror is not a backup.
What to do first
- Stop and find out which drive failed. Check the RAID controller utility, the server management console or the NAS storage page, and the logs. Note the drive serial numbers and which bay each one is in.
- Copy off important data while you can. If the array is degraded but still working, back up your most important files to separate storage before installing a replacement drive.
- Use only a new or known-good replacement. Do not reuse an old drive that was previously part of this array.
- Watch the surviving drive. If it shows errors, makes clicking noises or becomes very slow, shut the system down rather than starting a rebuild.
What not to do
- Do not swap drives back and forth to “see which one works.”
- Do not create a new array or initialize the drives to clear an error. On most controllers this overwrites the RAID metadata, and some options wipe the drives.
- Do not start a rebuild if you are not certain which drive holds the current data.
- Do not run file system repair tools on a degraded or failing mirror.
- Do not keep restarting a server that will not boot from its mirror. Each attempt adds wear to drives that may already be failing.
The good news about RAID 1
Because each drive holds a complete copy, there is often more than one route to the data. Even when both drives have problems, the damaged areas are rarely in exactly the same places, so the readable parts of each drive can be combined. Recovery starts by making a sector-by-sector copy of each drive, after any physical repairs in a cleanroom, and the work is done from those copies rather than from the original drives. You can read more on our RAID data recovery, server data recovery and NAS data recovery pages.
Get your mirror evaluated
With Data Retrieval, an engineer evaluates your drives and gives you a fixed quote, and the recovery only starts once you approve it. We work on-site across Los Angeles, the Inland Empire and Ventura County, and accept mail-in drives from anywhere in the US. If your RAID 1 mirror has failed, call (213) 986-4855 or book your data recovery online.