 |
|
|
Cluster Platform and Cluster ManagerCluster platforms depend on the cluster management program to track the cluster node status (whether or not nodes in the cluster are available for work). Essentially there are two types of cluster managers, vendor supplied and Oracle supplied. On HP (Tru64 TruCluster, HPUX MC/Service Guard), AIX (HACMP/ES CRM), and several other platforms the vendor will supply the cluster manager. Make sure the vendor-provided cluster manager is supported with Oracle9i RAC. For Linux and Windows, Oracle provides the cluster management software. RAC is also certified for use on SUN Solaris using Sun Cluster version 3.0, Veritas DBE/AC version 3.5, and Fujitsu-Siemens Prime Cluster versions 4.1 and 4.0. RAC is also available for HP Alpha OpenVMS (VMS is cluster aware out-of-the-box so no clusterware is needed). To use RAC on IBM-OS/390, you need to have the XCF clusterware that is compatible with your release of OS/390. On IBM 390 or the IBM zSeries using Linux, Oracle provides the Oracle OSD clusterware for your use with RAC. Clusterware, regardless of who supplies it, provides node monitoring for the other nodes in your system. By means of a "heartbeat" signal sent over the cluster interconnect, all nodes in a RAC cluster keep track of what nodes are available, which ones are unavailable, and whether or not a node becomes available. The Global Cache Service (LMSx) and Global Enqueue Service (LMON) operate independently of the Cluster Manager. They rely on the Cluster Manager for timely and correct information about the status of other nodes. If the Global Cache Service and Global Enqueue Service cannot get the information they need from a particular instance in the cluster, they will shut down the instance that is out of communication. This ensures the integrity of Real Application Clusters databases, as each instance must be aware of all other instances to coordinate disk access. If the GSD (Global Services Daemon) service cannot get the status of other nodes in the instance (it gets this information form the Cluster manager), the instance on its node cannot be mounted as a RAC instance. You will receive errors at the time of startup stating that the status of other instances is not available. The Oracle or vendor supplied cluster manager must be running at all times. RAC receives all of its functional benefits from its ability to run on multiple interconnected nodes (standalone or subsetted from a pool of CPUs). RAC relies heavily on the nodes underlying the high-speed inter-process communication (IPC) component to facilitate this. The IPC defines the protocols and interfaces required for the RAC environment to transfer messages between instances. Messages are the smallest units of communication in IPC. The core IPC functionality is built around an asynchronous queued messaging model. IPC is designed to send and receive discrete messages as fast as your hardware allows. With an optimized DSC (Disk Subsystems Communication) layer, various services can be implemented above IPC. This is how the Global Cache Service and the Cluster Manager perform their communication duties. Obviously, the faster nodes can communicate, the better your RAC environment will perform. Various interconnect types and speeds are shown in Figure 4.12. Oracle requires at least 100BaseT for the interconnect protocol. Figure 4.12: Interconnect Protocols and Speeds Currently, Oracle supports the Ethernet and FDDI interconnects, but IEEE1394 (Firewire) is also mentioned in several sources. Cache Fusion uses the high speed IPC interconnect to provide cache-to-cache transfers of data blocks between instances in a cluster (called data block shipping). This eliminates the disk I/O in Oracle's previous shared server technology known as pinging (which is inherently slow, since it is a mechanical process), and optimizes read/write concurrency. Block reads take advantage of the speed of IPC and an interconnecting network. For more information, see the book Oracle 11g Grid and Real Application Clusters - 30% off if you buy it directly from Rampant TechPress . Written by top Oracle experts, this RAC book has a complete online code depot with ready to use RAC scripts.
|
|
|
