Expanding RAID Array

[iostermann]

So when I built my PC some years ago, I got a pair of WD Blues 1TB to run in RAID 1. They worked fine for a while until I acquired a pair of 1TB WD Blacks and was trying to figure out my best optimization. The blacks were only SATA2, the blues were SATA3, and my mobo only has two SATA3 ports. One of the SATA3 ports was taken up by an SSD at that time. My eventual optimization was a RAID 10 array with a black and a blue in each mirror as I trusted the blacks less. This gave my 2TB and was fine until recently. I decided to go ahead and get a 4TB Seagate IronWolf NAS drive in order to start a switch to a more nas oriented machine in my desktop for college as my storage requirements increase.

 

This brings me to my current conundrum. What's my new optimization for storage space and redundancy as well as transferring the data? My current idea is to copy the RAID 10 array to the single 4TB disk to keep data. Then I would nuke the RAID10 and switch it to RAID0 to make a 4TB volume at SATA2 in order to not bottleneck the the SATA3 drive. I would then mirror the RAID0 volume to the IronWolf for a pseudo RAID10 setup with 4TB usable. This would leave me with still being protected in the case of a single drive failure without succumbing to the write performance tank in firmware RAID5. Any thoughts or ideas are welcome, Thanks!

[deleted_member_030719]

9 minutes ago, iostermann said:

 I trusted the blacks less.

Racist.

10 minutes ago, iostermann said:

This brings me to my current conundrum. What's my new optimization for storage space and redundancy as well as transferring the data? My current idea is to copy the RAID 10 array to the single 4TB disk to keep data. Then I would nuke the RAID10 and switch it to RAID0 to make a 4TB volume at SATA2 in order to not bottleneck the the SATA3 drive. I would then mirror the RAID0 volume to the IronWolf for a pseudo RAID10 setup with 4TB usable. This would leave me with still being protected in the case of a single drive failure without succumbing to the write performance tank in firmware RAID5. Any thoughts or ideas are welcome, Thanks!

Personally, I'd go for Raid 10 just because of the safety from drive failure.

[leadeater]

You could do the RAID 0 using the motherboard RAID yes then add that in to Storage Spaces and setup a two-way mirror to the 4TB drive. The array will only be the speed of that single 4TB disks but I'm thinking you want capacity more than the speed. Later you can buy another 4TB disk and replace the RAID 0 array.

[iostermann]

 

3 minutes ago, Jed M said:

Racist.

Personally, I'd go for Raid 10 just because of the safety from drive failure.

Thanks for the input, and in my defense, they have twice the power on count and 45,000 hours compared to 18,000. 

[iostermann]

2 minutes ago, leadeater said:

You could do the RAID 0 using the motherboard RAID yes then add that in to Storage Spaces and setup a two-way mirror to the 4TB drive. The array will only be the speed of that single 4TB disks but I'm thinking you want capacity more than the speed. Later you can buy another 4TB disk and replace the RAID 0 array.

I was under the assumption that read speeds for mirroring were about 2n since it reads from both mirrors, is this different for Storage Spaces?

[leadeater]

Just now, iostermann said:

I was under the assumption that read speeds for mirroring were about 2n since it reads from both mirrors, is this different for Storage Spaces?

For Storage Spaces it depends on the number of disks and the configuration. Read speed is the number of disks and write speed is half the number of disks in two-way mirror.

 

For 2 disks two-way mirror it is much like RAID 1, once you go above 2 disks it's much like RAID 10.

 

You actually have the option if you can backup all the data to disks not going to be used in the array to add all the disks to the Storage Spaces pool and setup a two-way mirror virtual disk. Storage Spaces supports differing sized disks but a single virtual disk in the pool is limited in size once it hits the capacity of the smallest disks.

 

For example if you have 3 disks of 1TB, 3TB, 3TB the first virtual disk you create will be 1.5TB maximum size, you could then create another virtual disk to use the remaining unused space on the two 3TB disks for another 2TB virtual disk.