ST4000DM005 For Raid 10?

[zubbs1]

I currently have 4x4TB Seagate ST4000DM005 hard drives. These are singlets, that I manually duplicate files between to create redundant copies. I am planning a dedicated machine for raid 10. I was planning on getting two additional ST4000DM005 drives, then making the array for an effective 12TB setup.

This will be a home media server. It will almost exclusively stream to only one system at a time. Writing to the drives will be minimal once array is established, as I will only add new files periodically to it. Most of the workload is simple reads.

Is there any issues anyone can see with using these hard drives? From what I can tell, Seagate's ERC is not present on these drives. Is that a big deal?

Appreciate your help.

 

[you2]

Depends on how you are creating the raid set. I use software raid (used to use mdadm; but now I use zfs). I think if you use hardware raid the issue might be different as some drives have explicit support (firmware support) but i don't know the details because I never delt with hardware raid.
-
Btw I would not use seagate drives for raid simply because their fail rate is higher than most other brands/models. I personally been using hgst though one has to consider model as well as brand.

 

[zubbs1]

Well I plan on utilizing software raid (mdadm). I like this approach because its easier than zfs vdev pools for expanding the array in the future.

I currently have 4x4TB, 1x3TB, and 1x1TB Seagate hard drives. The oldest are five years, and the newest are 1.5-2 years. I've been very lucky so far. Part of why I want this array setup is to get a proper data security program. Raid for some protection, then a pair of 6TB drives for 2x formal backups.

I just wanted to be sure the 4TB non-red/nas versions wouldn't be troublesome mirrored and striped into an array.

Thanks for your response.

 

[XavierMace]

I've got 16 ST2000DM's in a ZFS array. No issues.

 

[you2]

To be honest I was never overly happy with mdadm; brtfs has the best of both options but it isn't totally stable yet with more advance features (like growing arrays).
-
The big advantage of zfs over mdadm is that it check sums the data and if error is detected it will repair the bad data. For example I had a bad controller in one of my system and zfs was able to recover all the data by repairing the bad blocks across various drives. mdadm would never be able to do this because it can't detect bad data only bad blocks.
-
However, you did name the flaw with zfs - enlarging an array is very painful and expensive. My hope is in another few years brtfs will mature. My current array is 4+2 and it is running low on space and I don't see a good way to enlarge it.