Jackrabbit startup repository with an index import can result in corrupted indexes

During startup, even changes made by the cluster node itself must be synced. This is because of the index export/import feature. During export/import the following situation can occur:

A cluster node (the container) gets restarted with a fresh new file system containing an index import on file system. During startup, the journal log is queried to sync the changes which came in after the index export was done. However, in case the restarted cluster node has the same cluster node id, it is likely that the Journal table also contains change records indicating the change was made by the current cluster node. If during startup these changes are skipped, the index misses all changes made by the cluster node after the index export. To avoid this situation, make sure to even sync the changes made by the cluster node during STARTUP! Note that changes made by a cluster node after startup never need to be synced as already processed by the cluster node.