3 ms·
If there was 800GB of static data, it was almost full, so the wear-leveling process wouldn’t be optimal and it would write over a small part of the flash cells
by BearOso 5y ago
If there was 800GB of static data, it was almost full, so the wear-leveling process wouldn’t be optimal and it would write over a small part of the flash cells over and over. In that case you could exceed the program-erase lifetime, but not necessarily the TBW. We know the total TBW was high, though, so it was even more excessive.
From what I’ve read, Apple has fixed this problem in a recent Mac OS version. Apparently, it was a swapping bug.
- lawnchair_larry 5y agoWhere did you read that it was fixed or acknowledged as a bug? It definitely was not fixed.
- BearOso 5y agoThis thread: https://forums.macrumors.com/threads/ssd-swap-high-usage-of-terabytes-written.2284893/ https://forums.macrumors.com/threads/ssd-swap-high-usage-of-... Posters had narrowed down the excessive writes down by limiting memory usage, so they knew it was a swap problem, and some users said the latest OS fixed it. Now that I’ve checked it again, they’re saying it’s better but not fixed.
- 542458 5y agoI believe that modern SSDs shuffle data around so that wear levelling does still work even with nearly full drives.
- pkaye 5y agoIts the difference between static and dynamic wear leveling. Static wear leveling will shuffle around the data in static blocks to even out the program-erase cycles.