Dead drive: Best option for replacement?

[thatflashcat]

I had a drive die last night and want to get a replacement ordered ASAP. It's a Seagate ST4000NM0033 and the listing I purchased them from (new) no longer exists:

• Any suggestions for a replacement drive?
• Is swap size likely to be an issue during resilvering if I buy drives that aren't identical to the existing?
• Should I get 7200 RPM to match the existing, or is it ok to get a slower one? Performance isn't a huge deal.
• The ones I have now are SMR.

I'm not sure what my best option is and I don't want to run my pool in a degraded state for too long.

 

[Chris Moore]

The ones I have now are SMR.

Seagate Constellation ES.3 ST4000NM0033 4TB 7200 RPM 128MB Cache SATA 6.0Gb/s 3.5" Enterprise Internal Hard Drive Bare Drive

I don't think that is a SMR (Shingled Magnetic Recording) drive. SMR drives perform very, VERY poorly with ZFS.

This is the latest and greatest 4TB drive from Seagate:

Seagate Exos Enterprise 3.5'' HDD 4TB 7200 RPM Internal Hard Drive - Newegg.com

Buy Seagate Exos Enterprise Capacity 3.5'' HDD 4TB 7200 RPM 512n SATA 6Gb/s 128MB Cache Internal Hard Drive ST4000NM0035 with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.com

These are the model you have, but they are refurbished (used) so you don't know for sure what you are getting:

Seagate Constellation ES.3 4TB 3.5" Internal Hard Drive - Newegg.com

Buy Seagate Constellation ES.3 ST4000NM0033 4TB 7200 RPM 128MB Cache SATA 6.0Gb/s 3.5" Enterprise Internal Hard Drive Bare Drive with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.com

There is no problem replacing a drive with a larger drive. In fact if you replace them all, on at a time, your pool will expand to the larger capacity of the new drives. After they are all replaced.

If you have more questions, please ask, someone will be happy to help.

 

[thatflashcat]

I ended up purchasing an Ironwolf Pro 4TB. Had I realized the Exos line was so similar in price, I'd have gone that route, but I think the Ironwolf Pro will do fine.

I don't think that is a SMR (Shingled Magnetic Recording) drive. SMR drives perform very, VERY poorly with ZFS.

That could be.

There is no problem replacing a drive with a larger drive. In fact if you replace them all, on at a time, your pool will expand to the larger capacity of the new drives. After they are all replaced.

For now I plan to stick with 4TB since I'm still below 10% of my pool capacity, so even accounting for 20% overhead I have a long way to go.

The real question is why the drive died in the first place considering I had another drive die about a year ago. Temperatures are good so I know it's not that.

 

[Chris Moore]

Temperatures are good

What are they?

 

[thatflashcat]

What are they?

26-32 C.

 

[Chris Moore]

What about vibration. What kind of mounting are we talking here. Where I work, we put storage platforms out to sea on ships and the vibration causes premature drive failures. It could have also been a bad batch, I suppose, but getting two failed disks out of four total disks is a high failure rate. There is also the possibility that the delivery driver dropped the box. I have seen drives fail rapidly when they have been subject to the high force of a drop, even when they are still in factory packaging.

 

[thatflashcat]

What about vibration. What kind of mounting are we talking here. Where I work, we put storage platforms out to sea on ships and the vibration causes premature drive failures. It could have also been a bad batch, I suppose, but getting two failed disks out of four total disks is a high failure rate. There is also the possibility that the delivery driver dropped the box. I have seen drives fail rapidly when they have been subject to the high force of a drop, even when they are still in factory packaging.

*six drives total but still.

These drives were shipped from California to the Midwest where I am, so I guess it's possible they were damaged on the way. However, the drive that failed was a replacement from the previous failure a year ago, which seems suspicious and could indicate some issue with the environment. I don't remember if I installed the last replacement in the same slot or not, but I have made note of where the failure occurred this time (topmost slot).

Here's one of the mounting trays: