For example, every IBM® Rational® Asset Manager server instance builds a local index for assets and a local index for artifacts. This optimizes searches, relieves database load, and offers scalability in a cluster environment. Building indexes might take a long time, depending on the number of assets in the repository and in the contents of the artifacts. Although the indexer runs at a low priority, it requires processor activity and disk bandwidth. In a normal operation, a complete index does not need to be completely rebuilt; Rational Asset Manager augments the index whenever an asset is uploaded, removed, or updated.
The following information pertains to repository actions that can result in an intensive effort by the servers to normalize data.
If a repository administrator clicks the Rebuild Index button, the index will be completely rebuilt at the next scheduled indexing event.
If an index file becomes corrupted, that server instance completely rebuilds the index. Artifact indexing is typically more intensive in terms of processor and I/O activity.
If you change the role of a community so that either a search, read, or download permission is added or removed, then all assets for that community are re-indexed. The artifact index is not updated.
If you delete a community and move its assets to another community, the asset index is re-indexed. The artifact index is not updated.
Submitting a new asset, removing an asset, or updating an asset will update the indexers for the given asset only.
If you update the URL of a server, the Rational Asset Manager server must update all assets and asset manifests in its database. Asset manifests have namespace URIs in them that point to the server to which they belong.
If you add a server instance to the server cluster, the new server rebuilds its index.
Problems with the license server can incur a timeout of up to 30 seconds.
If you use IBM Rational ClearCase®, make sure to use a local area network. If you store assets in Rational ClearCase, asset persistence and downloading might require more processor cycles than assets on a shared RAID require.
The first time that Rational Asset Manager synchronizes with IBM WebSphere® Service Registry and Repository (either according to schedule or manually invoked), many assets will likely be brought into Rational Asset Manager; this can impact bandwidth consumption.