Changed files list with time buckets for efficient storage management

Abstract

There is provided, in a computer processing system, an apparatus for managing object data. The apparatus includes a changed objects manager for creating and managing a changed objects list that at least identifies the objects that have changed based on time of change. The changed objects list is associated with a plurality of time buckets. Each of the plurality of time buckets is associated with a respective date and time period and with object change records for objects having a timestamp falling within the respective date and time period. Each of the object change records is associated with a unique object identifier and the timestamp for a corresponding one of the objects. The timestamp specifies a date and a time corresponding to a latest one of a creation time or a most recent update time for the corresponding one of the objects.

Description

BACKGROUND [0001] 1. Technical Field [0002] The present invention relates generally to data file storage systems and, more particularly, to a changed files list with time buckets for efficient storage management. [0003] 2. Description of the Related Art [0004] Use of electronic data storage for long-term recordkeeping is increasing at an exponential rate. Much of this data is stored in file systems. Moreover, much of this data is write-once and is to be retained for long periods of time. [0005] The most commonly used disk storage devices are cheap, but not free and certainly not perfectly reliable nor absolutely durable. Accordingly, there is a need to migrate data to cheaper and / or more reliable media, a need to backup data, and a need to make replicas. [0006] The vast amounts of data and numbers of files maintained make manual management of data backup, replication, retention, and deletion burdensome, error prone, and impractical. Also, government regulations and business requirem...

AI Technical Summary

Problems solved by technology

The most commonly used disk storage devices are cheap, but not free and certainly not perfectly reliable nor absolutely durable.

The vast amounts of data and numbers of files maintained make manual management of data backup, replication, retention, and deletion burdensome, error prone, and impractical.

The overhead of searching and reading the file catalogs and directories (scanning the metadata of a file system) whilst performing policy or rule driven maintenance operations such as backup and data migration is chewing up a significant number of cycles, so much so that it is becoming a significant problem or expense in the operation of these systems, as exemplified by Tivoli Storage Manager(TSM) (data backup) and Tivoli Storage Manager for Space Management(HSM) (data migration, which is also known as hierarchical storage management).

However, this approach has some limitations.

For example, one limitation is that the change journal based backup feature is not crash proof.

Journal integrity is lost upon reboot.

Moreover, another limitation is that the change journal based backup feature can degrade file system performance.

Further, another limitation is that the change journal based backup feature is only supported on certain versions of the WINDOWS operating system.

Also, another limitation is that the change journal based backup feature does not address the meta-data scanning problem for HSM.

Additionally, another limitation is that the space required by the change journal based backup feature is (potentially) unbounded (or until it breaks).

However, the rate and amount of change can outpace the storage capacity of the journal and / or the processing cycles allocated to the storage management subsystem(s).

When this “breakage” occurs, change information is lost.

Images