FIG. 1: Apple''s Disk Utility application allows for simple RAID creation and disk-repair functions.
Many people get a sinking feeling when asked about backing up their files. It''s one of those inevitable tasks like taking out the garbage. We know we should be doing it on a regular basis, but the reality is that we frequently find other more pleasurable things to do—like writing and playing music—that gets in the way of what we should do. However, nothing hits home like working on your latest masterpiece and ending up with a lost or corrupt file—or worse yet, a completely failed hard drive.
THE COLD, HARD TRUTH
It''s a fact of our technology-saturated lives that hard drives fail. The old saying goes, “It''s not if a hard drive will fail, but when.” I''ve been playing the technology game of life for many years now, and I''ve found that hard drives fail for many reasons. I''ve had a brand-new computer that apparently was imaged poorly. The fix was to reformat the drive, which resulted in wiping off all the original files on the drive and all the programs and associated files I had spent hours installing from backup floppy disks and CDs. Another time, I had my laptop suddenly freeze up while I was using it, only to find that the culprit was the hard drive. It simply stopped working. Luckily, this was literally minutes after I had finished backing up the entire drive before going on a trip. It could have been that I overworked my hard drive in the long, sustained backup process or perhaps it was just a coincidence. I''ll never know, but I was happy I had a backup.
The bottom line: You need to back up your files. Not only your music files, but anything irreplaceable like business documents, digital photos, etc.
SET THE WAY-BACK MACHINE TO 1984
Stepping back in the not-so-distant past, as electronic musicians, our early backups for most modern computers were floppy disks. These were slow and held small amounts of data, but it was a relatively inexpensive way to back up. The data files themselves were fairly small, so the relatively small size of these disks worked just fine. We clawed our way up to double-sided, double-density (DSDD) floppies and archived 1.2MB.
Over time, we acquired larger data backup options. Removable cartridge disk drives seemed too good to be true; 45 megs on a SyQuest cartridge seemed huge; and writable compact disks gave us a jump to a glorious 650MB, with larger removable drives as a quick, yet relatively expensive option.
Next, writable DVDs bumped that number to 4.7GB (4700MB), followed by writable Blu-ray discs showing up on newer computers. Now we can have 25GB of data on a single disk. But backing up on disks is unfortunately a slow process; it can take hours to write your files onto a DVD or Blu-ray Disc.
GETTING WITH THE TIMES
Technology continues to march on and our files keep getting bigger. Yesterday''s state-of-the-art digital recordings at 16-bit/44kHz needed about 10.1MB per stereo minute, while today''s standard of 24-bit/96kHz recordings (or greater in some circumstances) needs more than three times that amount for a 3-minute multitrack song, which may end up closing in on almost 100MB. A multitrack recording easily exceeds the gigabyte range; quickly, a 1TB drive doesn''t seem so big. Luckily, the mass production and fierce competition in the hard drive market has brought the costs down to a point where you can buy a 1TB (1,000GB or 100,000MB) drive in an external USB or FireWire configuration for less than $100, with prices falling daily. Currently, backing up a 1TB drive leaves you very few options except to use another 1TB drive as that archival target. This, however, brings us back to the earlier statement, “It''s not if a hard drive fails, but when.”
If we worry about our backup failing as well as our main disk, how do we get ourselves out of this mess? Back up the backup? The answer, surprisingly enough, is yes. Ultimately, if you really want to be fail-safe, you''ll even have a copy of your backup data offsite. The most common way to do this is with a bundle of DVD or Blu-ray Discs, or a hard drive for a faster approach locked away in your safe deposit box. But for local use, many users are migrating over to a multi-drive solution with inherent file protection: a RAID.
IT'S A RAID!
A RAID is a grouping of two or more hard drives that can be configured in a variety of ways into a real-time or redundant backup configuration. RAID stands for “Redundant Array of Independent (some say inexpensive) Disks,” and it is a way of formatting and controlling multiple hard drives that act as a single unit with built-in redundancy of data to ensure that your data is protected or to increase throughput, or both. A RAID can be configured with two or more drives and in a variety of ways. For simplicity, we''ll focus on the common formats.
RAID 0 is not a backup solution but a way of striping multiple drives for higher data throughout. Each drive has a finite amount of throughput that can be reliably achieved, and the combination of two or more drives sharing this data throughput increases performance. Unfortunately, if you are looking for a file-protection scheme, RAID 0 is not the solution. Just like a regular drive, a hard drive failure means your data is lost. This is further complicated with RAID 0, as your single storage target comprises multiple drives.
RAID 1 is a simple mirrored disk array, meaning that the data from one disk is copied exactly onto a second disk. This is a great first step to backing up and probably the most common approach. Your computer should be able to work nicely with a RAID 1 array. If one disk fails, you simply replace it with a new drive. Using your backup with a RAID 1 array, you can have everything copied to the new drive in a few hours to restore the mirrored data.
RAID 3 distributes parity (duplicate) data in bytes over multiple drives, so if one drive fails, no data is lost. RAID 3 uses less space than a RAID 1 or RAID 5, but performance is better on RAID 5.
RAID 5 is like RAID 3, but it uses parity data in blocks instead of bytes spread across multiple drives. RAID 5 has acceptable performance, but as you are writing the original data and the parity information, write time does suffer. RAID 5 uses three or four disks, and you lose about 20 percent or more of your combined storage to the formatting.
RAID 10 (RAID 1 + 0) uses multiple mirrored drive sets. Now we''re getting into the big boys'' systems. This provides the great performance of RAID 0 with the mirrored backup capability of RAID 1. This solution is costly per megabyte like RAID 1, where you''ve doubled the need for capacity per megabyte.
There is a balance on what your choice should be. The three things to consider are file protection, cost, and performance. RAID 10 offers the best of the file protection and performance by increasing throughput with the RAID 0 and adding the mirrored drives as in RAID 1. Cost-wise, this is the most expensive route, requiring the greatest number of drives and a robust hardware controller on an existing computer or large dedicated NAS device. More on this below.
FIRST THINGS FIRST
An easy step to creating your first RAID is with an old computer. As we are in the habitual race to keep up with technology, you''ll no doubt have an old computer from your studio that will work just fine. Keep in mind that the connection speed of the older computer should be at least 100Mbps. When I updated my studio computer, I realized the old G4 had very little value on the open market and these were selling for a couple hundred bucks on eBay. So I used it as a file server/backup machine. I built a small closet in the garage where the G4 computer and my Gigabit Ethernet switch live. Using an old computer is an inexpensive way to build your first RAID as the cost to do this is simply the cost of the new hard drives.
If your computer is already a few years old, start by replacing the main drive and then adding two brand-new internal or external hard drives for the RAID. These drives need to be the same size, and for best practice, they should be the same model. You could omit the main drive and put the OS on the RAID to keep things simple, but I started out simply by replacing the main drive before adding the RAID. This made creating the RAID much simpler because when you format the drive for the RAID configuration, it will wipe all data on that drive.
Windows XP''s disk management will allow you to create a RAID 0 only. Luckily for Mac owners, Apple''s Disk Utility application allows a simple way to configure these two drives in a RAID 1 configuration, and these two hard drives then appear as one (see Fig. 1). The downside is that 2TB instantly becomes 1TB because the drives are now mirrored copies of each other. All the data I store on these hard drives is simultaneously written to both hard drives. If one drive fails, like it did with me, the other drive is still intact with 100 percent of the data. I replaced the faulty drive and restored the downed RAID disk (sometimes referred to as leaves) with Disk Utility. Just a bit more than eight hours later, the data was again mirrored across the two drives.
In Windows 7, you have a bit more flexibility for RAID options. Windows 7 Home will allow you to format for RAID 0, and Windows 7 Professional Enterprise and Ultimate also allow RAID 1. If you want to set up a RAID 5, Windows 7 does not directly support that with software control only so you''ll need to purchase a dedicated RAID hardware controller. In either case, you''ll need to convert your existing disk to a dynamic disk first, then add the second drive. You''ll create the RAID in the computer management window. Go to Control Panel > System and Security > Administrative Tools > Create and Format Hard Disk Partitions. You''ll get the popup display for the Disk Management window. When formatting, be very careful you do not format the wrong drive as you can erase the data. I''d be sure I have an intact backup before performing any RAID setup.
FIG. 2: Synology''s Disk Station DS409 is one of many compact network-attached storage (NAS) devices that can stand alone on a network.
PLAYING IT SAFE
I spoke with a couple of local IT heavyweights in town, and they personally had found RAID 5 more stable than RAID 1 in their systems. Things got a little more complicated from here. I decided I wanted a very accessible and very safe drive for my music and sound-design work, but I needed a lot of storage. Typically, if you do this on a local computer, you would want to buy a RAID card. These cards have their own CPU for handling the management of a disk array that could be taxing on a system. This is particularly true with a computer that is working hard like my studio Mac. As I don''t really want to give up any of the CPU''s horsepower to the hard drive control, I opted to go with a RAID card. Apple sells a RAID card for the Mac Pro for about $700, and it connects to the Mac''s internal hard drives.
The solution for me was to install a network-attached storage (NAS) device. The NAS has a dedicated CPU controller that manages the overhead necessary for a large RAID array. However, you don''t need a dedicated computer. The NAS sits on the network and shows up like any other shared computer, and it has a Gigabit Ethernet connection to keep things running smoothly and quickly.
After some careful research, balancing cost, convenience and reliability, I purchased a Disk Station DS409 NAS by Synology (see Fig. 2).
The DS409 is a small box approximately 6.5-inches wide by 7.5-inches tall by 8.5-inches deep. It has bays for up to four 2TB hard drives, and it can be configured in a variety of ways. One of the clever features of this unit is that you can continue to add more disks, including disks of larger capacity. We all know that today''s $100 1TB drive will give way to 2TB, 3TB, and larger drives in a few years at the same price. The DS409 allows you to keep building your system and storage a little at a time, to the point of replacing older drives with newer and larger drives. The DS409 automatically copies the data to the newly added drive.
For my purposes, I wanted to use a RAID 5 configuration. My decision was based on achieving the large storage footprint and not spending too much on a NAS device. I could have formatted my DS409 for RAID 10, but I would have ended up with much less storage because of the full mirroring. For example, with RAID 5, my four 1.5TB drives provide about 4.2TB of storage. The same system formatted with RAID 10 would provide approximately 2.8TB of storage.
The RAID 5 formatting for this configuration ate up 1.5TB; my 6TB block quickly turned into 4.5TB. The formatting took about 12 hours to complete. I installed four brand-new Seagate 1.5TB Barracuda drives and powered on the DS409. I attached the DS409 to my network using a simple Ethernet cable attached to my Gigabit Ethernet switch. Then I launched my browser on one of my other computers and entered the default Internet Protocol (IP) address from the documentation from the NAS. An IP address typically looks something like this: 188.8.131.52. Most of the time, there''s no need to change the IP address from the default. The only case would be if you had a larger network or perhaps had multiple devices from the same manufacturer. Most computers will avoid collision by dynamically allocating an IP address. This avoids two devices having the same IP address and creating a problem on the network.
A page opened up in my browser asking me to choose between RAID 0, RAID 1, RAID 4, RAID 5, or RAID 10. I selected RAID 5, and the formatting process started. Twelve hours later, I had a fast, secure, redundant 4.5TB bucket of love, just waiting to be populated with all my digital goodness. Because the majority of this data is a backup from my studio Mac, I''m pretty happy with the level of safety in file protection I now have.
As mentioned above, having a RAID is not 100-percent failsafe, but a means to better protect data. If you have a corrupt file on your local drive, backing it up to your RAID only provides multiple copies of corruption. A solution I like to employ is what I call Progressive Backups. The concept is simple. When composing or creating sound and music, I start using the Save As command after every significant change to protect against corruption or accidental editing mistakes. My song “Psychotic Ant” will start as Psychotic Ant 1_0. The next save after a major addition, deletion, or other chainsaw massacre is Psychotic Ant 1_1, and so on. I''ll also tag a suffix to the name such as key change or drum double to identify where a significant change has occurred in case I need to revisit this later. This way, if there is any issue with the file, I can backtrack through previous saves and harvest a lost file or unintentional automation of mutes.
I should mention a couple of other points that will help keep your files safe. As I do not write directly to my NAS RAID 5 device while working, I routinely need to copy files from the studio Mac to this unit. To keep track of which files have made it onto the RAID, I use the color-coding utility in the finder. If you right-click (or option-click) on a folder, you can choose a color for your files or folders. When I''ve copied all the data from my Mac to the NAS, I color-code these folders green. When I''m working on a project, I remove that color so I know the file or folder is not current on the RAID drive. After backing up that folder, I return it to that happy color.
But in the end, a file saved offsite—DVDs or a hard drive in your safe deposit box, or one of the many online storage options—is the only 99-percent-sure way of protecting data.
Dedicated backup programs like Apple Time Machine on the Mac or Roxio Retrospect for the PC will automate backups for you. These programs vary in features and cost. Some, like SOS, even come with many gigabytes of online storage, which is convenient, but slow and typically inadequate for the data required to backup huge audio files and libraries.
UPS AND DOWNS
Another must-have is an uninterruptible power supply (UPS) unit. I have multiple UPS units scattered throughout the studio, server closet, and even on the Vonage phone. Simply put, a UPS unit has a big battery in it that can keep your system running for many minutes in the event of a power failure. Some UPS units can even power devices for hours, but those are typically reserved for large businesses. UPS units also work as great surge protectors and voltage regulators in case of a brownout or loss of full power that can horrendously damage your valuable electronics. The battery in the UPS will ensure your system never cuts out in the middle of writing a file to disk, which can damage your data. Having two copies of corrupt data is basically garbage.
If you purhcase a UPS for your studio computer, remember that the more devices you have plugged into it, the shorter the time you''ll have to shut down in case of an outage. Make sure that your computer monitor is connected to it, too (or, even better, to its own UPS), because the temporary power the UPS provides to your CPU if the power goes out will be useless if you can''t see how to safetly shut down. That''s another lesson I learned the hard way.
Reek Havok is a sound designer and interactive consultant in the Seattle area. His company, Sounds Amazing, has created audio for the Science Fiction Museum, Mötley Crüe, Yes, and Robert Palmer.