[smartmontools-support] Short/long tests vs internal hdd stats

[David Hajes]

Hi guys,

I have few data servers. One is running 24/7 with sensitive data.

I used to have 60 disks experimental server and default smartd parameters killed IO.

Too frequent tests...issue was resolved by staggering smart tests.

I can't figure out a point of running smart tests.

Smartd does polling every 30 min. If I understood correctly, it asks storage controller what has been changed in disk health.

Disk FW basically does SMART tests constantly.

Isn't enough to get notification if some critical counter change reported by storage controller?

Like secror relocation count for example.

What is point of running short/long tests, storage controller inside the HDD does that already, and on the fly.

I have short tests weekly, long tests annually.

Could anyone enlighten me what is point of running self-tests at all, please? It seems to me as overkill, and slowing down of RAID array.

Regards

David