| 1 |
On Thu, Dec 23, 2021 at 12:39 PM Mark Knecht <markknecht@×××××.com> wrote: |
| 2 |
> |
| 3 |
> I'll respond to Rich's points in a bit but on this point I think |
| 4 |
> you're both right - new SSDs are very very reliable and I'm not overly |
| 5 |
> worried, but it seems a given that forcing more and more writes to an |
| 6 |
> SSD has to up the probability of a failure at some point. Zero writes |
| 7 |
> is almost no chance of failure, trillions of writes eventually wears |
| 8 |
> something out. |
| 9 |
> |
| 10 |
|
| 11 |
Every SSD has a rating for total writes. This varies and the ones |
| 12 |
that cost more will get more writes (often significantly more), and |
| 13 |
wear pattern matters a great deal. Chia fortunately seems to have |
| 14 |
died off pretty quickly but there is still a ton of data from those |
| 15 |
who were speculating on it, and they were buying high end SSDs and |
| 16 |
treating them as expendable resources - and plotting Chia is actually |
| 17 |
a fairly ideal use case as you write a few hundred GB and then you |
| 18 |
trim it all when you're done, so the entirety of the drive is getting |
| 19 |
turned over regularly. People plotting Chia were literally going |
| 20 |
through cases of high-end SSDs due to write wear, running them until |
| 21 |
failure in a matter of weeks. |
| 22 |
|
| 23 |
Obviously if you just write something and read it back constantly then |
| 24 |
wear isn't an issue. |
| 25 |
|
| 26 |
Just googled the Samsung Evo 870 and they're rated to 600x their |
| 27 |
capacity in writes, for example. If you write 600TB to the 1TB |
| 28 |
version of the drive, then it is likely to fail on you not too long |
| 29 |
after. |
| 30 |
|
| 31 |
Sure, it is a lot better than it used to be, and for typical use cases |
| 32 |
I agree that they last longer than spinning disks. However, a ZIL is |
| 33 |
not a "typical use case" as such things are measured. |
| 34 |
|
| 35 |
-- |
| 36 |
Rich |
Archives