The Network Automation NT-4000 uses a RAID (Redundant Array of Independent Disks) 10 array to provide the optimum mix of high database performance and redundant data storage with recovery features in the event of disk failures.
RAID 10 uses a minimum of four disk drives to create a RAID 0 array from two RAID 1 arrays, as shown in Figure 18. It uses mirroring and striping to form a stripe of mirrored subsets. This means that the array combines — or stripes — multiple disk drives, creating a single logical volume (RAID 0). RAID 10 combines the performance of RAID 0 and the high fault tolerance of RAID 1. Striping disk drives improves database write performance over a single disk drive for large databases. The disks are also mirrored (RAID 1), so that each disk in the logical volume is fully redundant. Figure 8 illustrates the principle.

Figure 8 RAID 10 Array Configuration

Note: The replacement IB-Type 3 disk assembly part number is IB-4000-R2-DISK-HDD300.

Disk Array Guidelines

Infoblox designs the disk array to be completely self managed. No maintenance procedures are required for a normally functioning disk array. Mishandling the disk array may cause an unrecoverable error and result in a failed appliance. Infoblox recommends that you observe the following guidelines:

• Remove only one disk at a time. Do not remove two or more disks from the appliance at the same time. Removing two or more disks at the same time may result in an appliance failure and require an RMA of the appliance. This rule applies to both powered and powered-down appliances.
• If you inadvertently remove the wrong disk drive, do not then immediately remove the disk drive that you originally intended to remove. Verify the status of the array and replace the disk drive that you initially removed before removing another drive.
• If the status of the array is degraded, remove the failed or failing disk drive only. Do not remove an optimally functioning drive.
• If your acceptance procedure requires a test of the RAID hot swap feature, remove only one disk drive at a time. You can remove a second disk only after you replace the first disk and the array completes its rebuilding process.
• Do not remove a disk drive while the array is rebuilding. This may result in an appliance failure. Verify the status of the array before removing a disk drive.
• You can hot swap a drive while the appliance remains in production.
• Some conditions may require powering down the appliance to replace a failed unit. This normally happens if the RAID controller detects an error that could damage the array. If you insert a replacement drive into a live array and the controller doesn't recognize the drive, power down the appliance.
• All disks in the RAID array must have the same disk type for the array to function properly.
• In the unlikely event that two disk drives fail simultaneously and the appliance is still operational, remove and replace the failed disk drives one at a time.
• Rebuild time can vary. The rebuild process takes approximately two hours on an idle appliance. On very busy appliances (over 90% utilization), the disk rebuild process can take as long as 40 hours.
• Replace a failed or mismatched disk only with a replacement disk shipped from Infoblox. When you request a replacement disk, report the Infoblox part number on the disk.