International Business Machines Corporation

Joon Chang

Janet Elizabeth Adkins

Manoj N Kumar

David Jones Craft

Ninad S Palsule

Andrew N Solomon

Jean-Philippe Sugarbroad

Amy Yi-mei Shi

Zachary Merlynn Loafman

Gokul Bhargava Kandiraju

Grover Herbert Neuman

Gerald Francis McBrearty

Neeta Garimella

Sreenivas Makineedi

Shashidhar Bomma

Matthew Todd Brandyberry

Robert K Gjertsen

Ramesh Veeramalla

Ashley D Lai

A method includes determining, by a processor, that a write request indicates a clone file that was cloned from a file in a first fileset. The clone file is of a second fileset and a clustered file system comprises the first fileset and the second fileset. A data block of the clone file affected by the write request is determined. The method includes determining that the data block at a current location cannot be modified. In response to a determination that the data block at the current location cannot be modified, the method includes implementing the write request with respect to the data block at a different location and indicating that the data block at the different location can be modified for subsequent write requests that indicate the clone file.

File cloning across different filesets

A cluster server manages allocation of free blocks to cluster clients performing writes in a clustered file system. The cluster server manages free block allocation with a free block map and an in-flight block map. The free block map is a data structure or hardware structure with data that indicates blocks or extents of the clustered file system that can be allocated to a client for the client to write data. The in-flight block map is a data structure or hardware structure with data that indicates blocks that have been allocated to clients, but remain in-flight. A block remains in-flight until the clustered file system metadata has been updated to reflect a write performed to that block by a client. After a consistency snapshot of the metadata is published to the storage resources, the data at the block will be visible to other nodes of the cluster.

In-flight block map for a clustered redirect-on-write filesystem

Methods, systems, and products for delayed deletion of extended attributes associated with a file in a file system are provided. Embodiments include identifying a file to delete, moving an extended attribute directory associated with the file to a background deletion directory, deleting the file, and deleting in the background the extended attribute directory. Some embodiments also include deleting an ACL associated with the file.

Delayed deletion of extended attributes

After a clone file is created from a file, a write request targeting the clone file can be handled. It is determined that the write request indicates the clone file that was cloned from a file in a first fileset. The clone file is of a second fileset and a clustered file system comprises the first fileset and the second fileset. A data block of the clone file affected by the write request is determined. It is then determined that the data block at a current location cannot be modified. The write request is implemented with respect to the data block at a different location. It is indicated that the data block at the different location can be modified for subsequent write requests that indicate the clone file.

Embodiments include a computer program product and an apparatus for cloning files across different filesets and maintaining the clone files. After a clone file is created from a file, a write request targeting the clone file can be handled. It is determined that the write request indicates the clone file that was cloned from a file in a first fileset. The clone file is of a second fileset and a clustered file system comprises the first fileset and the second fileset. A data block of the clone file affected by the write request is determined. It is then determined that the data block at a current location cannot be modified. The write request is implemented with respect to the data block at a different location. It is indicated that the data block at the different location can be modified for subsequent write requests that indicate the clone file.

A method includes initiating committing to persistent storage of a current consistency snapshot of a plurality of data objects in a redirect-on-write file system. Each of the plurality of data objects has a first copy of at least two copies of the data having a first context of at least two contexts. The method includes receiving, during committing to storage of the current consistency snapshot, an update to a data object of the plurality of data objects. The method also includes responsive to receipt of the update to the data object, creating a second copy of data of the data object from the first copy. The second copy of the data has a second context of the at least two contexts. The method includes responsive to receipt of the update to the data object, updating, based on the update, the second copy of the data of the data object.

Multiple contexts in a redirect on write file system

One embodiment is for using a sparse file as a clone. The method includes creating, by an application associated with a computer system, a snapshot of a file, such as an individual file, a collection of files, file system, directories, and so forth. Further, the method includes requesting, by the application, a clone of that created snapshot. Further still, the method includes generating, by the application, an empty sparse having at least one inode identifying the empty sparse file as the clone and pointing to the snapshot associated with the clone. Thereby, the requested clone is created using the created sparse file, which begins as an empty sparse file.

Using a sparse file as a clone of a file

Management of a file hierarchy for a clustered file system can be distributed across nodes of the cluster. A cluster file hierarchy is accessed to determine location of a file in response to a request to write to a file. A first node maintains the cluster file hierarchy. It is determined that management of a fileset object, which represents a fileset that includes the file, has been delegated to a second node based, at least in part, on said accessing the cluster file hierarchy. A node file hierarchy maintained by the second node is accessed responsive to determining the delegation. The cluster file hierarchy represents filesets of the clustered file system and the node hierarchy represents a subset of one or more of the filesets. Location of the file is determined based, at least in part, on said accessing the node file hierarchy.

Distributed file hierarchy management in a clustered redirect-on-write file system

In a cluster that distributes fileset management, fileset managers will maintain a free block map and an in-flight block map to allocate blocks to clients writing to a relevant fileset. A fileset manager managing free block allocation for cluster clients that write to a fileset managed by the fileset manager. Since a clustered file system manager manages an initial free block map for the cluster ('cluster free block map'), the clustered file system manager provisions free blocks to fileset managers from the cluster free block map. The fileset manager then allocates free blocks from the provisioned free blocks to cluster clients that write to the fileset managed by the fileset manager.

Distributed free block map for a clustered redirect-on-write file system

Maintaining a generation value for each fileset that is distinct from a corresponding fileset manager preserves the independence of nodes while also allowing distributed fileset management. A fileset manager can maintain a value that reflects consistency snapshots for that node ('node generation') separately from a value that reflects consistency snapshots for a particular fileset ('fileset generation').

Independent fileset generations in a clustered redirect-on-write filesystem

Some embodiments of the inventive subject matter allow a redirect-on-write file system to continue processing file transactions while writing a snapshot to persistent storage. Thus, embodiments do not block file transactions while flushing snapshots to persistent storage. To avoid blocking, some embodiments the file system can maintain two snapshot generations. The first snapshot generation is the snapshot being written to persistent storage. The second snapshot generation may reside in main memory, and store data for file transactions occurring after the file system began writing the first snapshot generation.

Inode management in redirect-on-write file system

A client node on which a file system is mounted can be configured to efficiently handle content access requests after a snapshot is created. The client node can maintain generation numbers at a data block, a file, and a fileset level to determine whether a data block is part of a snapshot generation and consequently whether the data block is writable in a current generation. Data block mappings and write access permissions associated with the data blocks need not be revoked at the client node prior to creating the snapshot. Cached data block mappings can be accessed to identify data blocks for servicing a read request. The data block can be updated in place if the data block is not part of the snapshot generation. The write request can be serviced at a new data block if the data block is part of the snapshot generation.

Managing data access requests after persistent snapshots

A system for managing a file system. In response to a user selecting an image of the file system to perform a rollback operation, each data block in the file system that is owned by the image is searched for. Then, it is determined whether a data block owned by the image is marked in a first map and a second map. In response to determining that the data block owned by the image is marked in the first map and the second map, the data block is marked in a third map. In response to determining that the data block owned by the image is marked in the third map, a new data block is allocated in the file system. Subsequently, data from the data block owned by the image is copied to the new data block in the file system prior to performing the rollback operation to avoid a collision that causes data corruption.

Growing community of inventors

Austin, TX, United States of America

Joon Chang