The Saga Of The Crashed NAS - Take 2


In my last post about my NAS crash I thought I had found the cause and fix for my out of memory crash on my Netgear ReadNAS RN104. While true that I did find the cause, I did not actually fix it. It appears btrfs is just not happy in such a limited memory space.

However, I have now had my NAS stable for some days and I don’t anticipate any issues before I am able to implement a more permanent solution.

So, then, what is the solution?

I disabled quotas and snapshots. I’m not a fan of this. There are a couple of shares I’d rather leave snapshots enabled on in case I fuck up. There is one volume I’d prefer to keep quotas on because an automated process writes to it and I don’t want it blowing other things up.

So if you have a Netgear ReadyNAS RN104 with large 2+ TB drives in it, and a lot of small files, DO NOT ENABLE THESE FEATURES.

To be honest, I doubt many people would have a need for these features. Snapshots would be useful for most people as anyone is prone to commit a dumb, and snapshots can make up for that. Snapshots have always existed as a feature on the RN104, and the newer versions of btrfs may be fine in the limited memory space for managing snapshots.

However, btrfs quotas were enabled much later, perhaps after the release of the RN204 and RN214.

This will at the very least give me time to upgrade to a newer model on my own terms. On the bright side, upgrading to a newer model should be fairly easy.