The Wanderer (12023-02-16): > That is exactly what I've always been told *does* happen, ever since > first reading about how SSDs et cetera work, more than a decade ago. > This is the first time I've seen a suggestion to the contrary.
This is surprising to me, since I have had the exact opposite information: hand-held devices with flash memory would require special, entirely-journalized filesystems like JFS2 to work properly, and that requires special MTD devices, but even the first and cheap USB sticks and CF cards would have some amount of wear-levelling. And my version is confirmed by experience: if the same whole block were written each time a sector in it is changed, then the block that contains the FAT, the superblock or the inodes for /var/log would be erased extremely frequently and die very fast, making the whole drive useless. Also, look at Google Trends or equivalent: “wear levelling” is not a new concept. > I'm also not sure that I'd have chosen to take the trade-off of that > added complexity for the presumed added performance and lack of need to > keep track of block size handling. The issue is not performance (although write performance would be impacted), and it is absolutely not the anecdotal question of aligning partition. The issue is that without wear-levelling your drive would die in less than a week of normal operation. Regards, -- Nicolas George
