I have purchased 3 x 3TB WD reds (CMR) to have a spares as I want to ensure I have enough on hand to replace any issues in the foreseeable future. I got them pretty cheap, brand new etc.
I wanted to know the best approach to conducting a test on the new drives prior to assuming they’re good as I realise that a significant percentage of drives will fail early on in their lifetime. What is the best approach?
I was considering simply connecting them to my PC on a USB 3.0 HDD caddy and running Western Digital ‘dashboard’ software which allows an ‘extended SMART’ test to conducted. I could connect 2 drives at a time and run this application, however would that be a sufficient way of being reasonably sure that no DOA issues are at play?
The other option I had considered was connecting it to the spare SATA ports on the motherboard and running an extended SMART test or the multi-report script within TN Scale?
The PC/USB caddy idea is easier for me to do (just more convenient), if that is a valid option, however I would be interested in the community’s perspective on testing new drives.
The general idea is to “sufficiently” stress the drives, with “sufficient” being freely defined (usually a few days, but up to one month for some).
Solnet-test-array automates the reads, and is useful for the actual array (or at least drives in the actual NAS). badblocksautomates the read and write tests for you with different patterns.
Run badblocks -b 4096 -ws /dev/sdX for each drive in a tmux session.
I typically run a short SMART test, the badblocks suite with that test pattern, followed by a long SMART test. If the drive survives all that, it goes into the qualified spare bin, the rest go back to the seller.