The Cluster Information Base is a replicated store of (primarily) two types of cluster[1] related information. The first type is mostly static configuration data and is a definition of all the resource[2]s, cluster nodes[3] and constraints[4] (or dependencies) that exist in the cluster. The other information stored in the Cluster Information Base is the current state of the cluster. This takes the form of a list of the known nodes with information from the LocalResourceManager[5] about currently active (NOTE: "active" is not the same thing as "running") resources.
The Cluster Information Base is stored and processed as whitespace insignificant XML and its layout and mandatory information is contained in the Annotated DTD[6].
The Cluster Information Base is PolicyFree[7]. During updates, all elements and attributes are copied verbatim. This enables the Cluster Information Base to be both forwards and backwards compatible.
The role of filling in defaults, enforcing the DTD listed above, and detecting integrity issues is that of the PolicyEngine[8] who is the primary user of the Cluster Information Base's contents.
ClusterInformationBase/UserGuide[9]
| [1] | http://en.wikipedia.org/wiki/Computer_cluster |
| [2] | http://www.linux-ha.org/resource |
| [3] | http://www.linux-ha.org/ClusterNodes |
| [4] | http://www.linux-ha.org/Constraints |
| [5] | http://www.linux-ha.org/LocalResourceManager |
| [6] | http://www.linux-ha.org/ClusterResourceManager/DTD1.0/Annotated |
| [7] | http://www.linux-ha.org/PolicyFree |
| [8] | http://www.linux-ha.org/PolicyEngine |
| [9] | http://www.linux-ha.org/ClusterInformationBase/UserGuide |
This information provided courtesy of the Linux-HA project at http://linux-ha.org/