0

Is there any method of creating a fault tolerant disk array with a mix of disks? I intend to use debian linux stable as a file server for home. The data is going to be cloud backed up but I don't really want to have to restore from the cloud as it would be slow so my aim is to be able to lose one of more of the disks. Performance is less of a concern as resilience.

I have 6 disks of varying capacity, brand, and speed which I want to create as fault tolerant as possible.

1 x 2 TB
2 x 1 TB 2 x 500 GB 1 x 40GB SSD

I don't think (but I could be wrong) I can use traditional RAID sets as none of the disks are paired.

Improve this question
3
  • by fault tolerant do you mean "I can lose one disk and still go on with no data loss" or "I can lose one disk, shrug it off, and replace the disk"?
    – Journeyman Geek ♦
    Feb 10, 2015 at 12:49
  • You could create a 2TB (RAID0) with your 1TB + 2x500GB and then RAID1 this with your 2TB. Your 40GB SSD is an odd-one-out so could be used as a swap space disk... if you lose any 2 disks of your RAID mirror at the same time (single 2TB and any from your RAID0) then you are in trouble...
    – Kinnectus
    Feb 10, 2015 at 12:54
  • I want to be able to keep running if a disk dies
    – Cookie
    Feb 10, 2015 at 13:06

2 Answers 2

Reset to default
1

There's a few possible options, each with its own advantages and disadvantages.

Its probably useful to consider a few concepts before picking what to use.

The simplest option I can think of is mhddfs - it doesn't do redundancy, but it basically concatenates multiple drives into a single drive. It handles where best to put any file, and will handle the loss of one or more drives, with only the loss of the data on that drive. Its also better than greyhole with many small files, though greyhole is somehting you should try too.

With an SSD and regular disks, ZFS may be an option - ZFS needs a SSD for caching and lots of ram for best results.I've not used it but its worth looking into as well.

Improve this answer
1
  • I'll have a look into zfs
    – Cookie
    Feb 10, 2015 at 13:36
0

You could create a mix of RAID levels to give you some fault tolerance. The mix you have, however, has a catastrophic point of failure in the event of two or more disks from either side of a RAID1 failing at the same time.

enter image description here

A RAID0 configuration for your 1x1TB and 2x500GB will give you a 2TB volume. This 2TB matches your existing 2TB hard disk. If you RAID1 (mirror) these devices you will have a fault-tolerant 2TB space.

However, if your single 2TB AND any of your 1TB or 500GB disks fail at the same time then your entire array will fail and you may lose all data.

Luckily you have a cloud-based backup strategy but you also have the ability to use data recovery on your single 2TB disk. A failed RAID0 is pretty much impossible to recover.

Your 40GB SSD you could use as your OS disk (for Linux) which would make the OS quite quick! You then store your important data on your RAID setup.

Ultimately:

  • 40GB SSD used purely for your OS
  • 2TB (RAID) used for your important files and other data.
Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .