On Sun, 15 Mar 2015 08:33:06 -0500 Dave Ihnat wrote:
It only allows errors to be seen by the host OS when it can't do this--meaning it's had enough bad blocks accumulating to exhaust its pool.
The host, in fact, hasn't seen an error. There is no trace of any I/O error reports in logs going back a year. Absolutely the only error I'm seeing is smart itself reporting 8 pending sectors over and over again, yet a long selftest doesn't find a single bad LBA.
At this point I suspect confused firmware in the SMART department and nothing at all actually wrong with the disk :-). (My crucial SSD drive was another one that had a SMART firmware bug so that it stopped working right after 5000 some-odd hours of operation not because there was anything wrong, but because the SMART firmware was busted - fortunately a firmware update fixed it).