Many of us seem to be building up piles of flash devices. And the concept of "bit rot" or wear out by any other name is accepted, Which raises 3 questions about handling the issue.
1. Are there utilities that can report the history or other metric data on future life span?
2. Is it possible to somehow log write to device events as a file entry? Since many devices have a Write indicator, some way of probing that indicator to increment a counter, then some estimation of drive wear would be captured as a human readable numerical. So the device could be reassigned from rewritable to archive if that is viable, or replaced as pre-emptive risk reduction.
3. Is there evidence to support a flash device used for say-70% of it's estimated "Write Cycles" being still permanently stable as Read Only for archiving?
On Jan 13, 2008 7:16 PM, Oren Beck [email protected] wrote:
- Is it possible to somehow log write to device events as a file entry?
Since many devices have a Write indicator, some way of probing that indicator to increment a counter, then some estimation of drive wear would be captured as a human readable numerical.
Doing so would seem to double the number of write-to-device events. How about a way to query the device for how many such events had occurred. We could set up a protocol. Call it "Systematic Tracking Using Performance Indication Data". Yeah, that should work.
Wow. That STUPID.
Doing so would seem to double the number of write-to-device events. How about a way to query the device for how many such events had occurred. We could set up a protocol. Call it "Systematic Tracking Using Performance Indication Data". Yeah, that should work.
-
Monty J. Harder -
Oren Beck -
Phil Thayer