Thourough HDD drive health check?

[Sleegemi]

Hi,

I'm new to nas4ree.

Installed the os and have added two HDD's.
I've searhed the documentation and thi forum, but I seem incapable of finding how to do a thourough check on t he health of both secondhand drives.
Already I went to Advanced | Execute and ran "smartcrl -t long /dev/ada0". But then nothing seems to happen and I have no clue whether thsi script is actually running now...

Is anyone willing to share with me what is the best an easy way to do a check?

Thank you.

Sleegemi.

[Sleegemi]

Oops!

My apologies for the misspellings. Well, it's allmost 1:30AM and am quite tired...

[kenZ71]

I assume and plan for random hardware failures.

So I simply install the drives and check SMART if it passes then the drive stays in my redundant pool. Also, I backup once a month. So I have multiple copies allowing for multiple drives to fail without data loss.

[Sleegemi]

Hi Ken, thank you for your answer.

The thing is, from several people I read that solely relying on S.M.A.R.T. doesn't give one a good impression on the healtiness of a drive. I really would like to somehow for at least one time do a full scan like checkdisk under Windows.

i.e. b0ssman wrote earlier this year: "Those smart test are mostly meaningless. They will report a drive to be OK long after I would consider the drive to be unusable. Analysing the raw data is much more useful"

[kenZ71]

Yes, smart can be meaning less which is why we have redundant pools and backups

[b0ssman]

if you have no data on the drive a good check for the health of the drive can be done using the linux tool badblocks

badblocks -wvsb 4096 /dev/sdb

depending on the speed of the drive this will take about a day.
save the raw smart values before and compare them afterwards.

the values to especially look at are
SMART 5 – Reallocated_Sector_Count.
SMART 187 – Reported_Uncorrectable_Errors.
SMART 188 – Command_Timeout.
SMART 197 – Current_Pending_Sector_Count.
SMART 198 – Offline_Uncorrectable.

[ChriZathens]

Badblocks is also available for FreeBSD.
Not sure it is there on N4F, though...(currently NAS is off)
If not, then it is a reasonable feature request.
If it gets added, a wiki entry for hdd burn in tests using badblocks, will also be a nice addition....(hopefully I will have time to create it)
EDIT:Using 9.3.0.2 - Nayla (revision 1310) on my main NAS and just verified that badblocks is there... Nice

[ChriZathens]

i.e. b0ssman wrote earlier this year: "Those smart test are mostly meaningless. They will report a drive to be OK long after I would consider the drive to be unusable. Analysing the raw data is much more useful"

IIRC, what b0ssman meant, was not that you should not use smart, or rely on smart. The overall "Passed" result is something to not rely on. But actually the raw data will do tell you what is going on with your drives.

[b0ssman]

yes. a smart test will give a pass to a drive that has reallocated sectors. this is a sign that the surface of the platter has problems.
however if that is your system drive you would normally not notice anything wrong and the drive will prob work fine for a bit longer (time to get out of warranty). its in the manufactures interest not to let customers return those drives to early.