When do you actually replace a failing drive?

[Chris Moore]

But how is WD gold in comparison to the WD red?

The WD Gold drives are supposed to be better than the Red drives. At work, we bought a 20 pack of the WD Gold drives as spares for some of our servers that came with WD Red Pro drives. We have only needed to use 4 of the Gold drives so far, as replacements for failed Red Pro drives, but the function in the server appears virtually identical and the drives look visually identical except for the color on the label.
I wouldn't hesitate to take the Gold drives.

Should i modify the fan mounted plate in the middle of the case so the cables can go a little bit more straight through it instead of the bottom one in the middle and back up again?

If the cable is damaged, it would likely need to be replaced. For the future, they are somewhat fragile and they should not be tightly bent. The thinner the cable, the more delicate it is, in my experience.

 

[melloa]

You are missing the point

Don't agree. The statement that resilvering could cause other disk to fail is correct. I've never compared that with anything else, you did.

 

[MrToddsFriends]

All good advice, but it is no more stressful on the pool to resilver than it is to scrub.

I think I can safely say that in a typical SOHO environment scrubbing --- coming along with heavy read activity for hours if not for days --- is the operation that is the most mechanical stressful one for the HDDs during normal operation (apart from burn-in). Even regularly occurring if scheduled as recommended.

 

[Chris Moore]

I think I can safely say that in a typical SOHO environment scrubbing --- coming along with heavy read activity for hours if not for days --- is the operation that is the most mechanical stressful one for the HDDs during normal operation (apart from burn-in). Even regularly occurring if scheduled as recommended.

I think you are correct. I have (for example) a system at work that contains over 265TB of data and the scrub takes a little over 4.5 days. It takes about as long to do a resilver of a disk, but the only drives that are being worked that hard are the ones in the same vdev with the disk being resilvered and the drive that is working the most is the new drive...

 

[aadje93]

Great that this is getting a real good discussion about the work for the disks!

I think i'll order 24 WD gold 4tb and get them in striped mirrors (12x mirror vdev) so i get better iops for my VM workload, and also "quick" resilver times. The reds can then be re-purposed in a backup/replication target with 4 Z2 vdevs of 6 disks each so it has more capacity when in future i may upgrade the golds again to bigger sizes if needed.