Properly configuring your RAID systems helps you to get the most out of your investments on the storage hardware and guarantee
planned service level agreements. It also reduces your maintenance efforts and avoids potential problems that might cause data loss or discontinued operations. It is especially true for a powerful and flexible RAID system like the one you have now. This section provides some basic steps and guidelines for your reference. The initial
configuration has the following tasks:
1. Understanding your users’ needs and environments
2. Configuring the hardware settings and doing health check 3. Organizing and presenting the storage resources
4. Installing and launching bundled software (optionally) 5. Getting ready for future maintenance tasks
• Understanding your users’ needs and environments
The first step for procuring or deploying any equipment is to know the users’ needs and environments, assuming you’ve already known much about your RAID systems. Users’ needs include the capacity, performance, reliability, and sharing. The environment information
clustered), host systems, host adapters, switches, topologies (direct-attached or networked storage), disk drives (enterprise-class, near-line, or desktop) and management networks. Extra cares are needed if you are installing the RAID systems to an existed
infrastructure under operations. Check your RAID system supplier to ensure good interoperability between the RAID system and the components in your environments. You will also need to know the potential changes in the future, like capacity growth rate or adding host systems, such that you can have plans for data migration and reconfigurations. The quality of your configurations will largely depend on the information you collect. It is advised to write down the information of users’ needs and environments as well as the configurations in your mind, which can be very helpful guidance through the all the lifetime of the RAID systems.
• Configuring the hardware settings and doing health check
After installing your RAID systems with necessary components, like hard disks and transceivers, to your environment, enabling the user interfaces is a prerequisite if you want to do anything useful to your RAID systems. The only user interface that you can use without any tools is the LCD console, by which the settings of the RS232 port and the management network interface can be done to allow you to use the GUI and CLI (see 3.3 Menu on page 3-8).
Now, do a quick health check by examining the GUI monitoring page to locate any mal-functioning components in the chassis or suspicious events (section 2.2). Follow the hardware manual to do troubleshooting, if needed, and contact your supplier if the
problems still exist. Make sure the links of the host interfaces are up and all installed hard disks are detected. Since your hard disks will be the final data repository, largely influencing the overall performance and reliability, it is advised to use the embedded self-test utility and SMART functions to check the hard disks (see 2.8 Hardware
Configurations on page 2-64 ). A better approach would be to use benchmark or stress testing tools.
• Organizing and presenting the storage resources
The most essential configuration tasks of a RAID system are to organize the hard disks using a variety of RAID settings and volume management functions, and eventually to present them to host systems as LUNs (LUN mapping). This is a process consisted of both top-down and bottom-up methodology. You see from high-level and logical perspectives of each host system to define the LUNs and their requirements. On the other hand, you will do configuration starting from the low-level and physical objects, like grouping the disk drives into disk groups.
Tradeoff analysis is required when choosing RAID levels, like using RAID 0 for good performance but losing reliability, or using RAID 6 for high reliability but incurring performance penalty and capacity overhead. The appendix provides information about the algorithms of each RAID level and the corresponding applications. You can also use the embedded volume management functions to build LUNs of higher performance and larger capacity. The RAID system offers much flexibility in configurations, like
independently-configurable RAID attributes for each logical disk, such that capacity overhead can be minimized while performance and reliability can still be guaranteed.
You might need to pay attentions to a few options when doing the tasks above, like initialization modes, cache settings, alignment offset rebuilding mode, and etc. Please read the GUI chapter to know their meanings and choose the most appropriate settings, because they are directly or indirectly related to how well the RAID system can perform (see 2.6 RAID Management on page 2-23 and 2.7.15 Miscellaneous on page 2-61).
• Installing and launching bundled software (optionally)
The RAID system is equipped with host-side software providing solutions for multi-path I/O, VDS-compliant management, and centralized management console on multiple platforms. You can locate their sections in the chapter 5 and know their features and benefits, as well as how to do the installation and configuration.
Contact your RAID system supplier to know the interoperability
Note
When planning your storage resources, reserving space for snapshot operations is needed. Please check chapter 5 for information about the snapshot functions.
• Getting ready for future maintenance tasks
The better you’re prepared, the less your maintenance efforts would be. Below are the major settings you’ll need for maintenance.
Event logging and notification
You can have peace only if you can always get timely notifications of incidents happening to your RAID systems, so completing the event notification settings is also a must-do. You might also need to set the policies for event logging and notifications (see 2.9 Event Management on page 2-69).
Data integrity assurance
For better system reliability, you are advised to set policies for
handling exceptions, like to start disk cloning when SMART warning is detected or too many bad sectors of a hard disk are discovered (see 2.8.1 Hard disks on page 2-64), or to turn off write cache when something wrong happens (see 2.9.5 Miscellaneous on page 2-75).
You may also schedule periodic maintenance tasks to do disk scrubbing(see 2.7.9 Scrubbing on page 2-54) for defected sectors recovery or to do disk self-tests (see 2.7.11 Performing disk self test on page 2-56).
Miscellaneous settings
There are also minor settings that you might need to do, like
Note
Installing multi-path I/O driver is a must for redundant-controller systems to support controller failover/failback. Please check Chapter 5: Advanced Functions for more information about MPIO and redundant-controller solution.