WASHINGTON, January 8, 2018: System hardware and software failures are inevitable in a typical IT environment. For that reason, redundancy has become a key element of this industry. Addressing this issue on the hardware side, RAID backup and storage plays a crucial role in preventing faults and ensuring the safety and usability of data in the event of hardware failures.
RAID (Redundant Array of Independent Disks) technologies combine and integrate an array of inexpensive drives into a higher-capacity virtual single drive. Improvements in this industry sector focus on the unique ways vendors put together the drives in an array to increase redundancy and improve performance.
Methodologies may also include expansion cards that make it easier to add such support to existing or legacy systems via PCI-E slots. Although most current motherboards come with built-in support for backup arrays, expansion cards and other adaptations are still an option should a company’s legacy motherboards not offer such an upgrade path .
Primary considerations for RAID implementation
Careful planning prior to installing new RAID arrays will assure optimum benefits. Before installation, IT staff should make sure that crucial data is backed up in full, as the new array or arrays will not replace that data.
Before purchasing and building out this type of backup and storage for existing systems, IT should also consider the following issues:
- Size – Consider how big is your database, and how much additional capacity is your company likely to require in the foreseeable future?
- Strength – What degree of fault tolerance is necessary for data integrity. In other words, how strong a system should IT plan to build?
- Speed – What read/write performance levels will the company require in order to adhere to its own internal standards?
To gain optimum benefits from the new array or arrays, IT must address additional essential issues such as the desired, optimal method of connection.
System administrators will need to determine the manner in which RAID devices will be connected to and integrated with an existing system. Typical choices including connection via a network or direct connection to a PC. Storage connections established via the network are described as Network Attached Storage – NAS for short. Storage connections made direct to a PC via USB, FireWire or other commonly deployed connections are called Direct Attached Storage, or DAS.
DAS connections may prove the faster choice, especially when established through eSATA cabling and ports. The downside: Such connections are limited to only one PC at a time, which is similar to connecting an external hard drive. On the other hand, NAS is ideal in a typical network environment that can function at multiple PCs and act as a file server.
As discussed above, when considering the performance of a RAID system for both DAS and NAS, IT staff should always keep in mind critical factors such as maximum I/O and throughput speed and redundancy levels.
This primary layouts just described are is the ones most frequently deployed. According to cloud service experts, there are only rare cases where the usage of what’s known as RAID0 implementation may be appropriate. TechTarget usefully defines this technology as follows:
“…also known as disk striping, [it] is a technique that breaks up a file and spreads the data across all the disk drives in a RAID group.
The primary usage of RAID0 is to store the temporary files as in case of video editing, graphics processing, or higher-end computations.
In the end, whatever array you or your company chooses to deploy, this multiple disk array performs with far greater efficiency than could be achieved by chaining together several separate disks. Among the most notable improvements will be a significant reduction in seek time, and better performance via drives serving to the I/O stream.
Fault tolerance is another major reason for installing a RAID array or arrays. Currently, the primary choices are among RAID 5, 6, and 10. RAID1 is also a fault tolerant array, which is mostly employed in systems requiring redundant storage with capacities less than a single disk, as in case of system partitions.
When making this decision, consider the following factors.
First, if the priority is for random writes, consider choosing RAID10.
However, if the priority is not for random writes, then:
- For fewer disks, consider RAID5.
- For more disks, considering RAID6.
IT will also need to decide when, where, and how often to back up the stored data on the disk array. Even if the aim is a highly fault tolerance array, it is essential to consider data backup seriously to confront a variety of unexpected situations.
Addiing a hot spare
An important thing you need to keep in mind with particular configurations of the RAID which consist of an array of five or more disks, it is essential to add a hot spare disk too. It allows the RAID device to replace a dropped disk with it in case of a need without external human interaction.
Uninterrupted power backup and write back caching
Write cache helps to improve RAID5 and RAID6 in terms of performance. On the other hand, if there is a power loss during the write, there is high chance that data may get damaged in various ways ranging from the file system and database corruption, which may ultimately end up with the condition called RAID write hole.
Correct shut down of the system will eradicate this issue, and for this, it is essential to have appropriate battery back or UPS system to ensure unrelenting power. UPS may be better as it allows proper shut down whereas the BBU (battery backup unit) will let you keep the data cache for up to 12 to 36 hours only and if the power is not restored, the cache will be lost.
Custom made or readymade RAID?
Feasibility of custom-made or readymade RAID needed to be decided based on the availability of skilled implementers and your size, strength, and speed needs. Both tailor-made and readymade servers will require the same amount of maintenance. Even if you rely on the readymade RAID, it should be ensured that there are appropriate maintenance and service support to upkeep the system capabilities over time.
Even when you implementsuch systems properly in place, one should remember that any system, including this one, may fail sooner or later. For this rason, first determine and then keep all possible alerts set on who will repair your system with minimum downtime when the need arises. Either your company will require internal administrators who know the ins and outs of various backup array implementations and who possess top-notch troubleshooting skills; or you will need to have an independent service provider on retainer to assist your company and its IT staff at the first sign of trouble.