RAID Enclosure question

sactown024

Distinguished
Mar 19, 2010
36
0
18,530
Hey guys, I don't post here much or at least not in a long time but I need some help.

I am interested in buying a external RAID enclosure, possibly a 2 bay or maybe 4 bay. I have been reading that when you buy a RAID system the computer will read all 4 drives as a single drive, is this true? Basically, what I am after is to have this enclosure with multiple drives and I want my computer to register each drive separately (e.x. C drive, E drive, F drive etc...) is this possible?

from what I want above, is it best to be running a RAID 0, 1, or JBOD?

Thanks!
 

Paperdoc

Polypheme
Ambassador
If you want to access each HDD unit in the enclosure as a separate drive, you don't want RAID at all.

RAID has several versions all quite different, but they all have one thing in common - they COMBINE two or more HDD units into what appears to be ONE drive. You cannot address those component units individually. (Oh, by the way, JBOD stands for "Just a Bunch of Drives" and it is not technically a RAID system, but that feature often is included in systems that actually do RAID arrays. What JBOD does is combine two or more HDD units sequentially into what looks like one big continuous drive.)

What you appear to want is just an enclosure with 2 to 4 bays in which to mount separate drives. Now, such an enclosure MAY be sold as "RAID Enclosure", meaning that its internal electronics include a RAID controller system that can use the units inside in one of several RAID ways. But most of these will allow you NOT to create a RAID array from the units, and just use them as individuals as you appear to want. However, in your case, MAKE SURE that mode is provided in the enclosure's controller.

You MAY be able to find an enclosure that does exactly what you want without having a RAID controller in it, too. And that MAY end up a little cheaper, since you are not buying the RAID controller system.
 

sactown024

Distinguished
Mar 19, 2010
36
0
18,530
Thanks for all that info. So when you buy a PC with a C drive, D drive etc, is that a raid system inside the desktop or is that usually just one drive partitioned into 3 drives?
 


Get away from this RAID idea. Each drive letter is either a separate drive or a partition. It's not a RAID setup.

If you buy a computer with two drives, and you only see one drive in My Computer, then it's setup as a RAID where the two drives are setup to be seen as one, RAID 1 or RAID 0.

What you want is just a box with either multiple drives that are separate or just one drive, or a RAID system that you would then format and partition manually into multiple drives. Just a RAID setup will get you a single drive letter for all the drives.
 

Paperdoc

Polypheme
Ambassador
Let's review a few concepts and terminology.

For a starting point (which will get clearer later), note that the word "drive" has become confused because it has been used two different ways. Originally, a "Hard Drive" was a single piece of hardware that you store info on. More recently we have been using very much larger "Hard Drives" and people have begun using a feature that's been there for a long time - you can divide one Hard Drive physical unit (I'll call it an HDD Unit) into more than one region, and use each as if it were a separate storage device.

So, here's where we are now. In order to use any HDD unit, you must be able to connect a power supply to it (a connector from the Power Supply Unit, or PSU, of your computer) and a data communication cable. That cable, in turn, plugs into a HDD port on the motherboard (mobo). All modern mobos have several HDD ports, and several power connectors available from the PSU, so you can connect several HDD units inside one computer. Each of them normally is handled as a completely separate storage device. In today's mobos, it is common to have 4, and often more, ports for HDD units of the SATA type. (And recently, there is a newer version of SATA that works faster for certain types of devices, so you will read of SATA 3 Gb/s and SATA 6 Gb/s, and sometimes people will refer to these by an older term, SATA II and SATA III.) Some mobos also still have an older type of HDD port called IDE.

For EACH HDD unit connected, some organizing needs to be done before it can be used for data. At the time it is manufactured, the maker does what is called a "Low-Level Format" which writes as magnetic signals a series of circular tracks on each disk surface, and each track is broken up into a series of Sectors. One Sector can hold 512 bytes of data plus a bit of extra info we call "overhead" to help keep things clear. (The new HDD units now being sold change this to 4,196 bytes per Sector). But at this point there is no data in those tracks. You will see posts from time to time suggesting that you wipe a HDD unit clean by doing your own "Low-Level Format". This is impossible today - nobody has the hardware and software tools to do this at home.

When a user installs a HDD, two operations need to be done on it to ready it for use by your Operating System. The first is to create one (or more) Partition(s) on the HDD. What this really means is that a utility writes a little bit of data to a particular location at the start of the HDD that defines a region of the HDD platters (in terms of Sector Numbers) that are to be used as one separate storage device called a "drive". Originally, you see, all HDD units were organized with only ONE Partition, and so "drive" became synonymous with the whole unit. But not today. In the most common system used for this in Windows (MBR), the system has limits that it can have no more than 4 Primary Partitions on one HDD unit, and the total size cannot exceed just over 2 GB. If you choose to Create more than one Partition on the HDD, the appropriate entries for each will be added to the Partition Table at the start.

Once the Partition(s) is (are) created, each needs to have a File System "installed" on it. For one Partition, this means another utility writes info to the beginning of the space defined for this Partition to create a Root directory and a set of special files used for keeping track of which Sectors in the Partition are being used by which files. The older system for this was called FAT32 (there were earlier FAT16 and FAT12 versions), but it is too limited so most people use the newer NTFS File System. This process is more commonly known as Formatting the Drive, and there are two fundamental version of this - Quick and Full Formats. We won't go into details of those here, but either one will finish getting the "drive" ready to use. As part of this last process, the system assigns a unique letter name to the "drive" (really, one Partition) like D: and that's how your OS (Windows) will access it. If there are more Partitions, each of them need to be Formatted.

In many current OS's, the utilities used for these two preparation steps are combined into one to make it easier, called Initialization.

So, that puts the basic info and organization on one HDD unit and it is ready for use by the OS. The same is needed for each HDD unit installed. Now, when you buy a computer all ready to turn on and use, all this stuff and more already has been done for you. When you buy a new External Hard Drive, very often the work has already been done on that one, too, so you can start to use it right away. But sometimes not.

So now you have a computer with one or more HDD units installed and working inside. Depending on how the Partitions were set up, you will have one or more "drives" with letter names like C:, D:, E:, etc. Moreover, the same system of using letters to identify storage units is used for optical drives and USB memory sticks. Thus the system called "My Computer" in which you find and use all your storage devices may have many things with letter names, and the number does not necessarily match the number of physical storage devices you have. For example, the machine I'm using has one floppy drive called A:, two HDD units (each with only on Partition on it) called C: and D:, two optical drives called E: and F:, an external enclosure containing one HDD unit divided into two Partitions called G: and H: Then if I plug in a couple of USB sticks they become I: and J:, and when I turn on my printer that has a card reader in it, My Computer shows me a "drive" called K: that may contain a memory sick from a camera.

Note that, so far, ALL of these storage devices are being treated as individual units accessible singly. That is why each needs its own letter name. I am not using any RAID system at all here.

Now, RAID is a whole other field. There are many versions of RAID that operate very differently. They all have on thing in common. They use specialized HDD controllers that take several HDD physical units plugged into the specific ports they control, and organize these several units in their own way. Then they make them appear to the OS to be only ONE "drive", and the OS knows nothing more. The OS "sees" ONE "drive" with a single letter name, and that's it. It is certainly possible to create MORE than one RAID array in one machine, using different groups of HDD units, but not many do that.

In almost all current computers, the mobo has "built-in RAID". What does that mean? Well, the mobo has a HDD controller chip with its associated ports. (In fact, it may have more than one such chip, and hence two groups of ports). What those chips do is controlled by programming code in a mobo chip called the BIOS. The BIOS does a LOT of stuff besides this, but let's focus on HDD control. "Normal" operation means using all the HDD units separately as described above. But you have the option of turning on the RAID system, which means that the BIOS starts to use a different set of coding for handling the HDD units involved. In most cases, you get to specify which HDD units will be used for a RAID array run by the BIOS, and which will not. So you actually can have BOTH separate "normal" HDD units and a RAID array operating on the same mobo. It is also possible to use more advanced systems that employ a dedicated RAID controller, BIOS and RAM on a card plugged into the PCI bus to run a RAID array.

I'll briefly describe some RAID systems just for info, but don't get bogged down with these - you do not appear to need any of them. RAID0 combines at least 2 HDD units into one array in which the many Sectors of data of each file are spread over all the HDD units involved, one Sector at a time. The main advantage is that it reduces the access time delays that come from the physical response times of the mechanical components of each HDD, but its risk is that it requires that ALL of the HDD units keep working. If any one HDD unit fails, it's likely all the data on the RAID array will be lost. RAID1 uses 2 (or more) HDD units in a different arrangement: it keeps TWO copies of all the files on separate HDD units, so it is often called "mirroring". The idea is that, if ever one HDD unit fails, the other is still working just fine and the system can keep on working with no delays until the faulty unit can be replaced and the entire RAID1 array restored. Some people falsely view this system as an automatic backup system which it is NOT. RAID5 uses 3 or more HDD units (often 5) and splits up all data into 5 pieces: four are quarters of the data stream, and the fifth is a mathematical calculation from them. These are all written to separate places. The advantage is that, if any one of the data chunks for a file becomes faulty (or if one HDD unit fails completely), all four of the original data quarters can be completely reconstructed error-free from the remaining good pieces. So the storage system never loses a file, and can keep on working when a failure occurs, even during the replacement and restoration of the failed HDD unit. This system has good and bad points and is most commonly used in professionally-run large server systems that cannot accept downtime. Its major weak spot is a rare event, but I actually saw one happen. If TWO HDD units in a RAID5 array fail at the same time, no recovery is possible. At that time, after the array is repaired, the only recovery is from reliable backups. There is a more advanced system, RAID6, that can recover itself from simultaneous failure of TWO HDD units.

Sactown024, your original question says your interest is simply in adding more HDD units to your system and using them as individual drives with their own unique letter names. As you can see, that does NOT need RAID in any form (nor JBOD). So you need an external enclosure that can hold 2 or 4 HDD units and NOT combine them into a RAID array. As I said in an earlier post, sometimes you get this in what is called a "RAID Enclosure" and just do not use the RAID feature in it, but ONLY if the enclosure allows you to use it that way.

Given some of your recent posts, you may not yet have considered the simpler option of installing more HDD units INSIDE your current computer. If you have a desktop system (not so likely for a laptop) there is surely space for more HDD units, but how many more is unknown to us. Maybe you need a knowledgeable person to help you look inside and determine what can be done.
 

sactown024

Distinguished
Mar 19, 2010
36
0
18,530
Wow huge help paperdoc, thanks!

I should have mentioned that I am running MAC OS on a macbookbook pro, the only option for adding more space is to remove my optical drive which I don't want to do, or replace my 250GB SSD with a spinning disk which i also dont want to do.

That being said, it seems my only option is an external enclosure, I am a photographer and I need a lot of space. The more I read about raid the more I think I would be better off with one, reason being is because since I am a photographer I need everything backed up immediately so I am thinking a mirrored system would be perfect for me.

So another question I have with a RAID system is, if one drive goes bad on a mirrored system A. will the system let me know one went bad? B. Can I just pull the bad one out and replace it with a new one?
 


This last paragraph is exactly how RAID 1 works. It will let you know that a drive is degraded and when you replace it, the raid controller will rebuild the array - i.e. copy the data back over to the new drive.

RAID however is not a backup. You will still need backups. A virus/malware or accidental file deleteion has the same effect on RAID as it does on a single drive.
 


If you are going to be using the drive to store live files on (ones you work on), make sure whatever enclosure you get has either a gigabit network port, or USB3 or eSATA or Firewire (if your MAC has that). You do NOT want to be loading large files though a USB2 connection.

You probably would want a 3rd drive for backups, as Hawkeys said, RAID is there to keep your data safe in case of a drive failure of one drive, a backup is there so you can restore deleted files, corrupted files, things like that. If your files go bad somehow, a RAID1 setup will just copy the bad files over to the second drive so you have garbage there, swapping drives won't help you. You'd want a second place to store backups on to keep a clean copy of your files.
 

Paperdoc

Polypheme
Ambassador
Echo Hawkeye and Hang-the-9. There is NO substitute for a complete backup system.

RAID1 makes a mirror copy of everything at the very moment it writes the first copy. It main use is in places where you can't afford downtime - things just have to keep operating normally. RAID1 gives you that so you can TEMPORARILY keep running until a proper restoration can be done at a less critical time. For example, I use a RAID1 array in a Point-of-Sale system in a retail clothing store we run. We can't have the system collapse in the middle of a busy day of sales. BUT after hours I could repair a failed RAID1 array and get it ready for tomorrow.

How would I know if it failed? Well, the RAID management system built into the mobo BIOS monitors the array and puts out a message if either of the HDD units in it fails or has a significant error. Then it reverts to operating like a single HDD - it uses only the good drive and ignores the bad one. When I get to responding to the error alert, it has tools that will identify which HDD has failed and allow me to replace that unit. Then on command it will re-establish the array by making a fresh copy of the good drive's data to the new replacement, and we're back to normal operation.

Here's what is important to note, however. A RAID1 system cannot possibly protect against these catastrophes:
1. Malware (virus or whatever) corrupts the files on your HDD, or even wipes them all out, or locks up your "drive", etc. Whatever the malware does, it will happen to BOTH units in the RAID array simultaneously, so you have two good copies of useless files!
2. Another component of your system fails and burns out, requiring that you replace the mobo or something similar. RAID systems are NOT standardized, so often you cannot read the data from a RAID array written in one system on a different system. Now, there are ways to guard against this, and there are software tools to recover data from RAID arrays. In fact, RAID1 is the simplest to recover data from. BUT it is a BIG hassle and requires special skills and tools.
3. A theft, flood or fire destroys your entire system. You no longer have any media to recover files from.

For these and several other possibilities, the only reliable way to recover is through a backup system that is USED properly. This means regularly making backups of your data and verifying that you CAN recover the data from them, then storing the data offsite to avoid things like fires and floods. It involves a number of other procedures, too. You should read up on Backup systems and processes. For a photographer such as you, OP, whose livelihood depends heavily on archived files preserved for long times but immediately accessible, a reliable backup system is invaluable and necessary!

As a small example, I said in my paragraph outlining RAID systems that I had seen a two-HDD failure of a RAID5 array, which made automated recovery impossible. This was in a professionally-run shop serving a large manufacturing plant. Their system involved daily backups to TWO locations - one to a local tape system with tapes stored on-site for quick access, PLUS simultaneously writing the backup files electronically to a remote service in a different province of the country. After the RAID array was repaired with the installation of two new HDD units, they did a minimal restore to get the system running again, then an on-line background restore process to complete the job. It took over 3 days (working 24/7) for the whole thing. But basic functionality was restored in half a day, and over 99% of files after the full time. Only a few files modified just before the crash were lost. You can't achieve that without a good backup system that is USED properly.

By the way, especially for your work, OP, long-term storage of data is important. For that reason, as you learn about backup systems, also consider the normal reliable lifetime of storage media and how to guarantee ability to retrieve data. This may involve a process of waiting a specified time, then making a fresh backup of the existing archived files to fresh media, and no longer relying on the old storage unit. And that might involve also changing the type of media used, as technology advances over the years. I have a friend who is a professional librarian; he and his colleagues are very aware of the limitations of information storage and archiving in all forms. Their expertise forms a great match with people who work in the technology of data storage.