| 1 |
On 27/04/2020 17:59, Rich Freeman wrote: |
| 2 |
> Really though a better solution than any of this is for the filesystem |
| 3 |
> to be more SSD-aware and just only perform writes on entire erase |
| 4 |
> regions at one time. If the drive is told to write blocks 1-32 then |
| 5 |
> it can just blindly erase their contents first because it knows |
| 6 |
> everything there is getting overwritten anyway. Likewise a filesystem |
| 7 |
> could do its own wear-leveling also, especially on something like |
| 8 |
> flash where the cost of fragmentation is not high. I'm not sure how |
| 9 |
> well either zfs or ext4 perform in these roles. Obviously a solution |
| 10 |
> like f2fs designed for flash storage is going to excel here. |
| 11 |
|
| 12 |
The problem here is "how big is an erase region". I've heard comments |
| 13 |
that it is several megs. Trying to consolidate writes into megabyte |
| 14 |
blocks is going to be tricky, to say the least, unless you're dealing |
| 15 |
with video files or hi-res photos - I think the files my camera chucks |
| 16 |
out are in the 10MB region ... (24MP raw...) |
| 17 |
|
| 18 |
Cheers, |
| 19 |
Wol |
Archives