linuxhw
S.M.A.R.T. errors should not be reported as malfunc for all HDD
Report https://linux-hardware.org/?probe=59b41381ed showing possible hdd issues, those issues related to automatically marked SMART issues and becoming malfunc for all HDD, it should not be so.
- https://linux-hardware.org/index.php?probe=38e0ad48b1
- https://linux-hardware.org/index.php?probe=38e0ad48b1&log=smartctl#sda
Why?
The status of your particular drive is green. The warning tells us that other users experience some problems with the same drive model.
I'd like to get notices only about issues with drivers or design hardware issues (bugs, crashes, hangs, stability, quality). SMART issues only important when a lot of devices with specific model (model + firmware) has issues known to be broken.
For example if all devices with specific model get bad blocks within 1 year, I'd like to know it. But when 1 device has SMART report issues, I rather do not have any related information until issue get above average failure rate.
Maybe relevant to this issue … I do like to see malfunctions flagged at (for example):
- https://bsd-hardware.info/?probe=1f3fa432dc (2020-11-21)
- https://bsd-hardware.info/?probe=60d9540d35 (2020-12-31)
- https://bsd-hardware.info/?probe=c2e361eeff (2021-01-23)
Amber is fine, although the text could be better. In lieu of:
The S.M.A.R.T. errors are detected. It's recommended to replace the drive soon.
– consider:
Red alert for at least one S.M.A.R.T. attribute.
– then it's for the end user to interpret each error or alert in the context of:
- https://www.smartmontools.org/wiki/FAQ
Side notes for the example above:
- the significant increase of
Reallocated_Sector_Ctfrom168 (0 21)to208 (0 25)(2020-12-31) was probably the result of extraordinary thrashing + consumption of free space on 2020-12-29 –https://github.com/ncw/stressdisk/issues/14 - the hardware is to be written off after I gain a new notebook.
https://github.com/linuxhw/hw-probe/issues/82#issuecomment-752769666
… when a lot of devices with specific model (model + firmware) has issues known to be broken. …
Currently at https://bsd-hardware.info/?probe=6149ab50a8#d5:
This device model is known to have problems
Following the link to https://bsd-hardware.info/?id=ide:hgst-hts725050a7e630 it seems that three of the six malfunc listings are for a single computer:
If the "known to have problems" alert is based on just four disks: the sample size is probably too small.
Yes, we need to show some reliability metrics for drives instead of a warning. From linuxhw/SMART