On 3/3/25 02:03, Dan Purgert wrote:
On Mar 02, 2025, Eben King wrote:[...] ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000f 082 064 006 Pre-fail Always - 146369262146 million read-errors.7 Seek_Error_Rate 0x000f 084 060 045 Pre-fail Always - 232382570230 million seek errors9 Power_On_Hours 0x0032 093 093 000 Old_age Always - 6346h+20m+46.297s~9 years on-time
6346 hours / 24 hours/day = 264.42 days 6346 hours / 24 hours/day / 365.2425 days/year = 0.72395 years
195 Hardware_ECC_Recovered 0x001a 082 064 000 Old_age Always - 146369262Well, at least those read errors were all corrected ;) None of the first three bits are absolute proof that the drive is going, but they're certainly cause for suspicion.
I have been watching my HDD's and SSD's with smartctl(8) intermittently over the years, and am still confused. HDD's with "bad" statistics can still work, while failed drives can produce statistics above the "bad" drive values (!). It's a conundrum.
I have always wondered if the decimal numbers in the smartctl(8) "RAW_VALUE" column are actual event counts, or a decimal representation of some binary bit field whose correct interpretation only the manufacturer knows (?).
I typically look at the "VALUE" column, as it usually represents an integer percentage that starts at 100 (%) and diminishes towards 0 (%) as the drive degrades. That said, I have seen starting values of 120, 200, and possibly 255 (?). So, again, only the manufacturer knows.
I own about a dozen and a half 2.5" SSD's, and one M.2 PCIe NVMe SSD. None have failed (yet). AIUI, SDD's go from working to choppy to brick in rapid succession.
I awoke one night to the smell of fried electronics. I quickly found the source -- a USB flash drive connected to a MacBook Pro as a Time Machine backup drive. I did a little testing on a Debian machine the next day. The drive was marginally readable at a very slow transfer rate (timeouts?). I did not want to damage my computer, so I smashed the USB flash drive and recycled the husk.
David