Redundant Storage Unit Balance Restoration

When voluntary actions such as a filestore migration cause an imbalance in system redundancy, manually invoke rebalancing using the Optimize System Storage REST API endpoint. Calling this endpoint raises a flag for Artifactory to run rebalancing after the next Garbage Collection. To expedite rebalancing, invoke garbage collection manually from the Artifactory UI.

Optimizing System Storage

After deployment, duplications of files bigger than the redundancy are expected and the garbage collector removes them after the optimization run. On each deploy operation to a shard, Artifactory checks if the checksum exists more than once. If the checksum exists more than once and the count of such repetitions is bigger than the number of redundancy, Artifactory triggers the optimization flag. When the next full garbage collection runs, the duplicates are removed.

By default, the full garbage collection runs after 20 iterations and each iteration runs after four hours resulting in six iterations per day. For more information on garbage collection, see Garbage Collection.

Artifactory REST API provides an endpoint that allows you to immediately trigger balancing between redundant storage units of a sharded filestore. If balancing is already running, the endpoint skips the process. For information on the API, see Optimize System Storage.