[kwlug-disc] array of SSDs - what to consider

[Andrew Kohlsmith (mailing lists account)]

Hey everyone,

I’ve got four 2TB SSDs I’d like to put into an array and use as a single device. I do this already with spinning rust to great success using regular old mdadm, LVM and XFS. In fact that array’s been up and in use for the better part of a 20 years, moving from RAID1 to RAID5 to RAID6 without any data loss.

With SSDs my understanding is that TRIM is rendered ineffective when “behind” mdadm without some hacking, and I’m thinking of trying out the RAID5-ish equivalent of zfs since the filesystem would be aware of the underlying hardware and could properly work with the physical media to ensure longevity. I’ve been reluctant to use zfs in the past because I kind of like the separation of duties that mdadm+lvm+xfs affords me.

I am *not* concerned with throughput; what I’m most interested in is the ability to arbitrarily carve up the combined space, (cold) recover from any single drive failure without data loss and of course ensure data integrity and longevity of the underlying storage media.

With that in mind, what should I be looking at to achieve these goals? The array would only be directly accessed from Linux, although it would likely be an embedded ARM system, not your traditional beige box desktop, so the use of proprietary hardware RAID is not appealing.

Thanks,
Andrew

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: Message signed with OpenPGP
URL: <http://kwlug.org/pipermail/kwlug-disc_kwlug.org/attachments/20221122/3fd18af9/attachment.sig>