Can you explain why ECC is great for ZFS in particular as opposed to any other filesystem? And if the data leaves the NAS to be modified by a regular desktop computer then you lose the ECC assurance anyway, don't you?
Can you explain why ECC is great for ZFS in particular as opposed to any other filesystem? And if the data leaves the NAS to be modified by a regular desktop computer then you lose the ECC assurance anyway, don't you?
ZFS is about end-to-end integrity, not just redundancy. It stores checksums of data when writing, checks them when reading, and can perform automatic restores from mirror members if mismatches occur. During writes, ZFS generates checksums from blocks in RAM. If a bit flips in memory before the block is written, ZFS will store a checksum matching the corrupted data, breaking the integrity guarantee. That’s why ECC RAM is particularly important for ZFS - without it you risk undermining the filesystem’s end-to-end integrity. Other filesystems usually lack such guarantees.
The oversimplified answer is that ZFS’ in-memory structures are not designed to minimize bitflip risk, as some file systems are. Content is hashed when written to memory cache, but it can be a long time before it then gets to disk. Very little validation is done at that point to protect against writing bad data.