The disk array controller as the core component of the RAID(Redundant Array ofIndependent Disks) system, it’s performance has become the critical factor whichdetermines the performance of the whole system.At the same time, the storage system’ssingle point of failure has begun to transfer to the controller. Controller redundancymechanism guarantee that the system doesn’t go down and can still provide continuousservice in the condition that when one single controller fails. However, it merely providesa single point of failure to avoid the application of. Now there is higher requirement fordata security, especially in certain sectors need to provide7*24uninterrupted service. Thesystem should also ensure that the reliability of the data storage side.The dual controller RAID system based on DRBD(Distributed Replicated BlockDevice) is a double solution, it can provide network storage service based oniSCSI(Internet Small Computer System Interface).Through the logical volumemanagement, the system achieves the device-level storage virtualization, make themanagement of the disk device more flexible. Redundancy mechanism both on thecontroller and the date which is used by DRBD that provides the function of mirror makethe system avoiding the failure from both of the two sides, and this improves theavailability. Through dividing the storage devices into two kinds, ordinary volume andhigh available volume, the system can provide two different level service according to therequirement.The management system is a B/S architecture, including the front Web managementinterface and the background daemon. We implementation the visual management ofdual-controller, and achieved timely and effective monitoring and recovery. The testresults show that the system has reached the design requirements. |