Redundancy for Linux

I’ve been searching around and I’m not sure exactly how to get this working.

What is the best to use the Drobo with linux?

Currently, we have 2 TB drives formatted for a single 2TB ext3 volume - but this is obviously not configured for redundancy and we would like to utilize the dynamically expandable capabilities of the drobo with the redundancy.

We are using drobo-utils on Ubuntu 9.04 - Linux 2.6.31-14-generic headers.

Any ideas of how to configure this for redundancy using Drobo’s software?

Do I have to reformat as NTFS and them mount on Ubuntu?

No matter how the drobo is formatted it still uses BeyondRaid so that if a drive fails, your data is on the remaining drives in your drobo.


Thanks for the quick reply.

Is there a way to tell if I configured it correctly? How can I verify that BeyondRaid is active and that it is maintaining redundancy?

I don’t think I did, because… The 2 x 1 TB drives have a total available disk size of 1023G (based on df -h).

It seems that at least 20% of the drive would be used by BeyondRaid to create striping, no?

Because BeyondRaid is built into the drobo. There is no way to turn it off so thus it’s always on.

Your OS is showing that size because you formatted to a 2TB volume size.

Ok. So, is it the case BeyondRaid doesn’t use ANY space on the device for striping? I’m still a bit confused as to how this works.

I misspoke earlier. The DROBO has 2 x 500GB drives and it has 1TB available.

What happens if I want to add additional drives? Must I reformat?

BeyondRAID does use space for fault-tolerance (parity or whatever scheme it uses internally). However, Thin Provisioning controls what the OS “sees” in terms of volume size.

You don’t need to reformat when you add drives - that’s the beauty of Thin Provisioning.

Thin Provisioning tells the OS that you have a 2TB, 4TB, 8TB, or 16TB volume, even though you may not actually have that much physical storage.

Because it “looks” like a larger drive, when you add actual storage to the Drobo, the OS just treats it exactly the same and you don’t to reformat, repartition or do any other annoying things.

Drobo Dashboard and the front capacity lights give you the real story in terms of physical space available and used.

If your physical storage capacity exceeds the volume size you chose, Drobo will simply present another volume of the same size to the OS.

For example, I have a 4x 2TB Drobo which has ~5.5 TB capacity. I chose 2TB volumes on this particular Drobo, so the Drobo shows my OS three 2TB volumes. The free physical space is shared between the three volumes and managed by Drobo. If I were to upgrade to say, 4x 3TB drives, then I’d have over 6TB of usable storage, so Drobo would add another 2TB volume and I’d have four 2TB volumes presented to the OS.

On my other Drobo with another set of 4x 2TB drives, I have it formatted as 16TB volumes, so I have one massive 16TB volume, of which I can fill 5.5 TB worth until I upgrade to larger drives.

In your case, since you have open bays, you could just add more drives and you’d instantly increase your physical storage, while your OS is none-the-wiser.

Thin Provisioning saves a lot of migration headaches for me. In the past, I would have to migrate my data from the old array to some temporary storage, set up the new array, then migrate the data back. It was a time-consuming process.

@bhiga: gen1 is only certified to 2 TiB volumes, later models as only supported to 8 TiB. No matter the model, 16 TiB is an unsupported configuration likely to result in data loss.

Thanks philobyte… forgot that we were talking Linux.

Rules are different for Linux vs Windows/Mac.