A reindex is only The projectdoc Site Indexer is required after the database has been restored, a space has been imported, or for some reason the index is corrupt. This reindex may also be required if the algorithm for storing information by the projectdoc Toolbox in the Lucene index has been changed. The fact that this procedure is required after an update will be provided in the release notes of each version. Caution Box |
---|
Not only needs the Lucene index be updated, but also the projectdoc caches need to be cleared and the projectdoc internal structure needs to be updated. Since the projectdoc Toolbox has no control over the standard Lucene Reindex Process, a projectdoc Reindex is required to be started via the projectdoc REST API. This can be done selectively on projectdoc spaces or on the complete site. Note that the regular Lucene Reindex Process does no harm on the projectdoc document information in the database. Since the document are processed in random order, the contents of the properties may not be up-to-date. A reindex with the projectdoc Site Reindex will update the internal structures, the Lucene Index, and the cache This is the recommended procedure for the projectdoc Toolbox version 2.0 and up. The procedure for earlier versions is not very different, but recursion detection for references is a feature for display property macros added to 2.0. Users usually do not define recursions, but may do so very rarely unintentionally. Recursions in references are no problem as long as the cache is in place. If the cache is removed, recursions are detected by Confluence on a stack overflow. This may drain on resources until all recursions are resolved. Version 2.0 will detect recursions based on display property macros more quickly and render an error box as a value to the properties instead (if and only if the document cache is empty). |