Abstract: One or more techniques and/or systems are provided for incremental transfer with unused data block reclamation. For example, source volume of a source storage filer may have a physical replication relationship with a destination volume of a destination storage filer. Snapshots of a file system of the source volume may be used to identify new data blocks (e.g., data blocks allocated or reallocated to a new snapshot) and/or unused data blocks (e.g., data blocks of a deleted snapshot) used to incrementally update the destination volume so that the destination volume is a mirrored replication of the source volume. In this way, new data blocks may be updated and unused data blocks may be punched out (e.g., deallocated) of the destination volume during execution of an incremental update.

Abstract: One or more techniques and/or systems are provided for incremental transfer with unused data block reclamation. For example, source volume of a source storage filer may have a physical replication relationship with a destination volume of a destination storage filer. Snapshots of a file system of the source volume may be used to identify new data blocks (e.g., data blocks allocated or reallocated to a new snapshot) and/or unused data blocks (e.g., data blocks of a deleted snapshot) used to incrementally update the destination volume so that the destination volume is a mirrored replication of the source volume. In this way, new data blocks may be updated and unused data blocks may be punched out (e.g., deallocated) of the destination volume during execution of an incremental update.

Abstract: The present invention provides a system and method for comparing data sets, to ensure that they are accurate reflections of each other, without the need for performing O(N2) operations, in which N is the size of each data set. A hash table is generated for the first data set. For each of the second data set entries, should the entry not exist in the hash table, the entry is second data set unique. Otherwise, the entry is removed from the hash table. At the end of the pass through the second data set entries, only those entries that exist in the hash table are first data set unique. Alternately, two processes operate in parallel so that each selects entries from one of the data sets and determines if the entry exists in the hash table. If the entry does exist, it is removed. Otherwise, the entry is added to the hash table.

Abstract: A method and system for secure and reliable event data logging are disclosed. In accordance with an embodiment of the invention, event data is generated for events relating to backup, or archiving, operations. The event data is appended to the log file. However, during the operation to append the event data to the log file, the appending operation is monitored to determine whether the event logging operation is successful. If the event logging operation is unsuccessful, the event data is retained in memory until the event data can be successfully appended to the log file.

Abstract: A system and method for providing soft locks so applications can communicate their dependence upon PCPIs for a plurality of storage systems and applications. The soft locks, created and maintained by an application executing on one or more of the storage appliances, identify certain PCPIs that are to be retained by a storage appliance and its applications. An illustrative embodiment of this system and method is to use these soft locks to ensure that storage volume destinations along the cascaded chain of applications running on that or other storage systems retain the appropriate PCPI and state.

Abstract: A method and apparatus for operating a storage system is disclosed. A root node has a first pointers to a plurality of lower nodes, the lower nodes having reference through a plurality of second pointers to point to blocks of persistent storage, the blocks of persistent storage storing data of the storage system. A persistent consistency point of the storage system is created at a selected time, the persistent consistency point created by creating a partial consistency point for one or more of the first pointers and second pointers to selected blocks of data. A statutes indicator is set to indicate that an identified partial consistency point is completed. All partial consistency points are coalesced into a final completed consistency point for the selected time after all of the partial consistency points reach a status of completed.

Abstract: A system and method for fault-tolerant synchronization of replica updates for fixed persistent consistency point image consumption. PCPIs are consumed at a fixed level and each coalesced PCPI presents a valid version of the source file system from some point in time. The qtrees concurrently move through a coalescing cycle wherein some qtrees are updated, some may update partially and some may not update at all. Partially updated qtrees are rolled-back to the previous state, after which all qtrees are transitioned to a stable state.