Smart drive status information / warnings

[kallisti5]

To alert users to potential hardware problems, having Haiku periodically monitor and report critical SMART faults would be a great feature.

General ideas:

• Kernel interface or ioctl for drive smart status
• Tracker SMART drive information. (right click drive, get info)
• Tracker warning icon on disks with elevated warning levels
• Notification on critical failures
• Add smart warnings to checkfs

Scope:

• AHCI
• NVMe
• ATA

Resources:

• ​https://www.smartmontools.org/
• ​http://www.intel.com/content/www/us/en/io/serial-ata/serial-ata-ahci-spec-rev1_3.html

[kallisti5]

example of smart information from gnome drive management app

[axeld]

While a Tracker integration would be nice, I think the first step would be to show that data in DriveSetup.

However, AFAIK, the data varies between manufacturers and model a lot, so it might not be easily possible to provide a Tracker integration like the one you are suggesting (ie. detecting a problem across manufacturers might prove problematic).

[kallisti5]

I know Ubuntu *used* to alert you of failing drives. This is partially where I got the idea from. I had it happen once and it was pretty amazing since it gave me a little time to make sure my backups were in order.

Here is someone documenting it: ​https://blog.al4.co.nz/2009/09/hdd-failure-warning-in-ubuntu-karmic-9-10/

I have a dedicated Haiku machine at my desk, and i've heard the heads clack occasionally in a way they shouldn't which spawned this ticket :-)

[kallisti5]

Just noticed a comment shows a neat research paper from google on predicting drive failures based on SMART data:

​https://static.googleusercontent.com/media/research.google.com/en//archive/disk_failures.pdf

[pulkomandy]

IIRC the SMART data varies, but for each field, you get the current value, max value ever observed, and warning/critical thresholds.

There is a tool on Linux to dump this data, which can give some hints about how to handle it.