Hey everyone,
Within the next few hours I’ll be receiving a Synology DS1019+. This is one of Synology’s higher-end models, with five drive bays by default, support for adding more, and all the bells and whistles that you can get at the high-end of their consumer gear. I won’t tell you what I paid for it though.
For the first couple weeks of the NAS’ life, I’ll be performing stress tests on the disks that I’ve bought to go with it. I won’t be able to use the NAS for its intended purpose while that’s going on.
After listening to Noah rap on btrfs and how it’s not production-ready, yet reading Synology’s documentation saying it is… I’m intrigued. Synology’s statement is “[…] it is known that Btrfs RAID is unstable and not suitable for production environments. For this reason, Synology chose Linux RAID over Btrfs RAID” (What is the RAID implementation for Btrfs File System on Synology NAS?).
From that, we can see that they’re using Linux’s Device Mapper RAID instead of btrfs native RAID. When btrfs detects a checksum error in a block, then, some Synology software must be used to correct that error!
I’d like to put this to the test.
So, using a few older 320GB drives, I’m going to stress out btrfs. I’ll try to write files to the array, then tamper with the original copy and read it back. Later I’ll try to simulate a disk that’s absolutely trashed – randomly flipping bits past its RAID header but in its data section – see if that causes any problems.
Do you have any other ideas for the tests that I could perform on the system? Ways that you think I could break the RAID?