2100 Chassis Failure

I would assume that since this is a production system, that you maintain a current and complete backup. If so, that does give you a Plan B in case of problems. If your backup is not current, this would be a good time to get it so, if you don't have a backup this would be an excellent time to create one and maintain it going forward.

Now, I am not an IT professional, nor am a user of the 2100, but I have several desktop ReadyNAS units and the process should be similar. Now you are saying that you have a degraded array. That term is used to describe an array that has no redundancy and has encountered problems. On the other hand, a non-redundant array is one that has lost a single drive in what otherwise would be a single redundant array. The 2100 has 4 bays, so would be at best a redundant array. (Dual redundancy requires units with 6 or more bays). Unless you have other problems that you have not stated, this is your situation.

At a point when the activity is non-existant or very low (normally after hours and of course tell everyone left to shut down from the network so they don't lose any data), power down your existing 2100 normally. Remove the drive from the failed slot and set it to one side (if you have not already removed the drive). Then one at a time remove the other three drives and note the bay in which they were installed. This can be easily done with post-it notes if not otherwise labeled with stick on labels. Remove the failed 2100 and install the replacement 2100. Do not apply power yet. I am not sure if the 2100 automatically powers up when the cord is plugged in as the desktop x-86 models do or not, but leave that step to last just in case. Now install the three drives from the array in the same bays in which they were installed. Then apply power. The unit should boot up in the same state as the failed 2100.

Once it has booted, and is stable and everything on the system is back the way it was, you can hot add the drive from the failed bay. The unit should recognize the addition of the drive and start a resync which will take anywhere from 8-12 hours depending on the size of the drives. This is actually the stressful part of the process. During this time if another drive is weakened and fails, this is when the array will be compromised.

Going forward, I would highly recommend you work your way to a second 2100 with the same or greater capacity and make one your production machine and one your backup. You can then schedule automatic overnight backups via NFS fro the first backup and rsync after that to synchronize the files on the two machines. I would still recommend an off site backup in case of disaster, be it one that is hand carried to another location or what. You would have to decide your strategy based on what the data is worth to you. Just put your self in the shoes of some businesses that have had their physical location totally devastated by a disaster and what you would have to do to resume operations in that situation.

One thing I would suggest is that before you move the disks across you get a spare disk (must not be from your array) and put it in the replacement 2100. Ensure the firmware is updated to the same version (or newer) that you were running on the dead unit.