SOLVED Pool configuration advise

[Fred974]

Hi,

I have 6x 3TB HHD and 3x 60GB SSD to create a pool. The plan is to have the 3 ssd as metadata disk to create a fusion pool.
My question is: With the 6x springing HDD, will I be better to create 2vdev with 3 HDD in each vdev or will I be better with a single vdev with the 6 HDD?
I use TrueNAS as general file storage but also as storage for xcp-ng via nfs and iscsi.

Thank you

 

[RickH]

With the 6x springing HDD, will I be better to create 2vdev with 3 HDD in each vdev or will I be better with a single vdev with the 6 HDD?

This really depends on your requirements - In either a 2x 3-drive z1 or a 1x 6-drive z2 config you're going to loose 2 drives worth of space to parity information so the formatted size is going to be identical. Both will survive up 2 drive failures; however, in the 2x 3-drive z1 config, if you loose 2 drives in the same vDev your data is gone - the single 6-drive config provides better redundancy here as you could loose any 2 drives and your data would still be safe. If your array is easily accessible and you have spare drives on hand to quickly replace a failed drive this may or may not be a concern for you.

2x 3 drive z1 vdevs will provide better slightly performance for virtualization loads (assuming the rest of your config is up to par) - for general file storage operations I would guess you're not going to notice any difference. Without the rest of your system and network specs it's hard to say where your performance bottleneck is going to be.

To sum up - if you're sure that the storage pool is going to be the bottleneck (not likely unless you're running iscsi over a 10GbE network), the virtualization performance is your top priority, AND you have your data backed up, then you might be better served by the 2x 3-drive config. If you're still using Gig-Ethernet I would go with the single 6-drive z2 config for the better redundancy.

 

[Fred974]

@RickH Thank your for your reply, your explanation was really useful. I've decided for redundancy and use 1 vdev. In future I can always expend with another 6 HDD vdev and keep the same mentality for redundancy .

Fred

 

[sretalla]

If your priority is to have IOPS for block storage, you can really only do mirrors, RAIDZ is not a good option for that workload (even in a multi-VDEV pool).

The path to success for block storage

It seems like I haven't written a sticky for awhile, but just in the last week I've had to cover this topic several times. ZFS does two different things very well. One is storage of large sequentially-written files, such as archives, logs, or data files, where the file does not have the middle...

www.truenas.com