In part 1 of this post, I differentiated between archiving footage from a shoot and keeping footage safe during post. I wanted to get editors to think about what would happen if you lost your media.
More importantly, consider how media loss could affect the deadline. How long will it take to get back up and operational?
It’s certainly something I consider. Although there are many ways to do it and you’ll have to consider your own situation (layers of protection vs. costs), here’s how I protect my media with an eye to keeping data safe and making recovery fast if there’s a problem.
To begin with, my main media volume is a RAID – Redundant Array of Independent Disks. I use an array of disks to get the speed I need to edit 4K and beyond. For some projects I use SSDs, but I have to consider cost versus quantity of storage.
While using a RAID can help with speed, the “R” is important too. But not all RAIDs are redundant. RAID configurations are labeled with numbers—typically 0, 1, 5, 6 and beyond.
RAID 0 isn’t redundant: It takes all the drives and turns them into 1 big volume, which is useful if you need a lot of space, but all the drives have to be working or you’ll lose all your data.
RAID 1 is used with two drives and mirrors the data. As far as your computer is concerned, it looks like you have only one drive because the second drive is a mirror image of the first one. RAID 1 is a simple way to get redundancy: You lose a drive and you haven’t lost your data, but it takes half of your possible storage away.
RAID 5 takes multiple drives and configures them so you can lose a drive and still operate with less of a performance hit than RAID 1. RAID 6 is similar, but it allows you to lose two drives.
With any of these setups, it’s important to replace a bad drive and have the array rebuild so you still have protection. Most arrays are designed so you can edit while the array is rebuilding but with reduced performance.
I run with a RAID 6 array. In my experience, when you lose a drive in an array, sometimes the very act of replacing the bad drive can cause a second drive to go down. With RAID 6, I have that second drive covered.
Of course, with any redundancy you lose space, but it’s not as bad as RAID 1. If you have an 8 TB array using 8 disks, with RAID 0 you end up with 8 TB of usable space, with RAID 5 about 7 TB and RAID 6 about 6 TB.
There’s also speed loss by having to write “parity data” in RAID 5 and 6. (Parity is how RAID keeps the data recoverable when you lose drives—by placing parity data across all drives.)
But what about if the whole RAID system goes down? Stay tuned.