Abstract: Embodiments are provided for balancing storage resources in a distributed database. According to certain aspects, various hardware components may facilitate a three-stage technique including a node balancer technique, a shard balancer technique, and a replica balancer technique. The node balancer technique may create a set of pods from a set of nodes residing across a set of storage racks. The shard balancer technique may redistribute, among the set of pods, a portion of a set of shards assigned to respective pods of the set of pods. The replica balancer technique may, for each pod, distribute the set of replicas so that the replicas reside across the storage racks of that pod.

Abstract: Embodiments are provided for balancing storage resources in a distributed database. According to certain aspects, various hardware components may facilitate a three-stage technique including a node balancer technique, a shard balancer technique, and a replica balancer technique. The node balancer technique may create a set of pods from a set of nodes residing across a set of storage racks. The shard balancer technique may redistribute, among the set of pods, a portion of a set of shards assigned to respective pods of the set of pods. The replica balancer technique may, for each pod, distribute the set of replicas so that the replicas reside across the storage racks of that pod.

Abstract: Method, system, and programs for resolving data inconsistency in a distributed system having a plurality of replica instances. In response to a read request of at least one data element, a key-specific probability value for the at least one data element of the read request is retrieved. Based on the retrieved key-specific probability value and a system load of the distributed system, a self-healing probability of the at least one data element of the read request is computed. A self-healing transaction that includes the at least one data element of the read request is then generated based on the self-healing probability of the at least one data element. The generated self-healing transaction is further applied to at least some of the plurality of replica instances in the distributed system to resolve data inconsistency.

Abstract: Method, system, and programs for resolving data inconsistency in a distributed system having a plurality of replica instances. In response to a read request of at least one data element, a key-specific probability value for the at least one data element of the read request is retrieved. Based on the retrieved key-specific probability value and a system load of the distributed system, a self-healing probability of the at least one data element of the read request is computed. A self-healing transaction that includes the at least one data element of the read request is then generated based on the self-healing probability of the at least one data element. The generated self-healing transaction is further applied to at least some of the plurality of replica instances in the distributed system to resolve data inconsistency.

Abstract: A technique for migrating (e.g., “copying” or “moving”) temporary data from one session to another session is provided. While executing a workload associated with a particular session, session state is created. A subset of the session state comprises large objects that are referred to as “temporary data.” Thus, temporary data is stored in a database (instead of in the database server) and is normally deleted when the particular session is deleted. To avoid this deletion when the session state is migrated to a different session, the temporary data, in one embodiment, is saved in an external storage that is separate from the database. An external table is created in the external storage and stores the temporary data. Location information to access the external table is provided to the other session.

Abstract: A technique for migrating (e.g., “copying” or “moving”) temporary data from one session to another session is provided. While executing a workload associated with a particular session, session state is created. A subset of the session state comprises large objects that are referred to as “temporary data.” Thus, temporary data is stored in a database (instead of in the database server) and is normally deleted when the particular session is deleted. To avoid this deletion when the session state is migrated to a different session, the temporary data, in one embodiment, is saved in an external storage that is separate from the database. An external table is created in the external storage and stores the temporary data. Location information to access the external table is provided to the other session.