How to Choose a RAID Controller by Configuration | TechMikeNY



Timelapse ASMR: Watch Us (Re)Build Servers...

Wed, Feb 28, 24

  As you might have guessed, configuring enterprise-grade servers takes some serious...

GPU Compatibility: What to Consider

Tue, Feb 06, 24

From rendering complex graphics to training neural networks, the specialized parallel processing...

All About RAID Batteries – Preventing Data...

Wed, Jan 17, 24

  Ever sat around and wondered about your RAID controller’s battery? Can’t...

WATCH: Installing a Rear Flex Bay in an R7...

Wed, Dec 20, 23

This tutorial will show you, step-by-step, how to install a rear flex...


RAID Basics. What to Consider When Getting a RAID Controller

If you are in the market for a server, a concept likely on your radar is RAID ("Redundant Array of Independent Disks"). RAID is an industry-standard when it comes to the utilization of multiple drives on a single system since it allows for – depending on the configuration – improved speed, redundancy of data, or both. Virtually all modern servers are designed to house multiple drives, so selecting the right RAID controller for your needs is essential. You don't want to have to upgrade your Controller – especially if you plan to add drives down-the-line – as it can be tricky (see our vlog post here on the challenges of installing a controller in 12th Gen Dell servers). Conversely, you don't want to go overboard with a controller that provides features you don't need.   

In this blog post, we'll cover the basic standard RAID configurations with their advantages and disadvantages. We'll also share some of the questions to ask yourself when considering a RAID controller and its configuration. Walking you through it is what we're all about at TechMikeNY!

A RAID controller – center – seated in a 12th Generation Dell Server.

Standard RAID Configurations

 The RAID configurations below are the most commonly utilized. The key terms here are striping (or parity) – where data is shared across multiple drives for speed; and mirroring – where data is copied, or mirrored, across multiple drives for redundancy/fault tolerance.  

RAID 0 - Increased speed but the risk of data loss.  RAID 0 takes all of the data and stripes it across the drives. It essentially converts them into one large volume that is the sum of all of the drives. The advantage of this is speed; however, the major disadvantage to RAID 0 is no fault tolerance. In other words, if one drive fails, the entire array fails. RAID 0 requires a minimum of two drives.  

RAID 1 - Data redundancy but reduced write speed.  RAID 1 takes all of the data and mirrors it across the disks creating a copy (or copies, depending on the amount of drives). The obvious advantage is redundancy and backup, as all data is backed up at least once, and as long as one drive is functioning in the array, the data is secure. The disadvantage is that the sum of the total storage decreases as the data is copied on all drives. RAID 1 requires a minimum of two drives.

RAID 5 - Increased read speed and fault tolerance. RAID 5 utilizes the striping of RAID 0 and the mirroring of RAID 1. As such, RAID 5 requires a minimum of three drives. Again, due to the mirroring of the data, the sum of the drives' total volume is decreased. The main advantage of RAID 5 is it gives striping and mirroring in an environment where disks are limited. 

RAID 10 - Increased speed and fault tolerance.  RAID 10 (or RAID 1+0) is similar to RAID 5 in that it combines striping and mirroring across the array; however, it does so by joining multiple RAID 1 arrays using the striping of RAID 0. The advantage of RAID 10 over RAID 5 is increased speed – in the event of a drive failure, RAID 5 will use the parity from the remaining functional disks until the failed drive is replaced. However, RAID 10 cuts your total capacity in half. RAID 10 requires a minimum of four disks and is often utilized in large Enterprise environments.    

Other RAID configurations exist, such as RAID 4, RAID 6 – they all are variations of the striping and mirroring principle, where striping is for speed, and mirroring is for redundancy. 


What should I Consider When Selecting a RAID Controller?

 Now that we have the RAID basics down, some key questions to ask yourself are:

  • What level of RAID do I want to set-up? Is speed the most important, like in a dedicated gaming server? Or is redundancy, e.g., making sure your data is backed-up and avoiding any downtime?
  • What if you need speed and redundancy? If both are a premium, you'll want a controller that can handle both without bottlenecking.  
  • How many drives do you intend to have on the server? Do you want to add drives down the line? Some RAID controllers have a cap on how many drives they can manage in the array.  

Here are some key terms when looking at Controller: 

  • Cache – this the memory that the Controller has built-in. Controllers with higher cache are faster since they compensate for the potential latency of the drives reading & writing the data.
  • Battery backup unit (BBU) – some controllers have dedicated batteries. This is in the spirit of redundancy – if there should be a power failure, the Controller can quickly wind down operations to avoid catastrophic data failure. The more drives you have in your RAID, the more you should consider a controller with a battery.  
  • Passthrough” is a term you might see on some controllers. It merely means that that Controller is not performing any RAID functions. This is also known as HBA Mode (some controllers, if they have the 'HBA'-prefix, e.g., HBA330, that means it has no RAID functionality).  
  • PERC ("PowerEdge RAID Controller") is Dell's brand acronym for its PowerEdge line-compatible RAID controllers. If you see "PERC," you are looking at a Dell controller.

Hot Tips 

  • Dell PowerEdge servers come with a native RAID controller – however, that Controller is limited as it can only manage a maximum of four SATA drives (ONLY SATA, no SAS). 
  • Always make sure you are using the same type of drives, e.g., SATA vs. SAS and also SSD vs. HDD. 
  • Volumes should be the same as well. While RAID can be done with drives of different storage sizes, the RAID size will be restricted to the smallest drive in the array. It's generally frowned upon to use different drive types.  

Final Thoughts 

RAID is a significant part of server builds and configurations and making sure you are getting a sufficient Controller for your needs will save aggravation – and needed upgrades – down-the-line. 

At TechMikeNY, we ship our servers configured with the following:

  • Servers with one drive will have a RAID 0
  • Servers with two drives will have a RAID 1 between the two drives
  • Server with more than two drives will have a RAID 1 between the first two drives and a RAID 0 or a RAID 5 with the remaining drives as (allowed by the number of drives).

Don't forget: if you want a configuration different from the above, make a note in your order, and our Tech's will set-up it up per your request. That's our TechMikeNY Connection!



Leave a comment

Name . . Message .

Please note, comments must be approved before they are published