Distributed file system having separate data and metadata and providing a consistent snapshot thereof

A method of creating a snapshot of a distributed file system comprising:

• quiescing requests affecting the distributed file system, the distributed file system including a first set of one or more metadata file systems storing metadata about files in the distributed file system and including a second set of one or more data file systems storing file content data for files of the distributed file system;
• obtaining a snapshot of each metadata file system in the first set;
• obtaining a snapshot of each data file system in the second set;
• and
• unquiescing requests affecting the distributed file system upon successful completion of said obtaining said snapshot for each metadata file system in the first set and upon successful completion of said obtaining said snapshot of each data file system in the second set, whereby the snapshot of the distributed file system is created as a result of successfully completing said first obtaining step and said second obtaining step and wherein the snapshot of the distributed file system includes the snapshot of each metadata file system in the first set and includes the snapshot of each data file system in the second set;
• and
• wherein said unquiescing includes resuming processing of new requests affecting the distributed file system, and the method includes receiving a first request for first snapshot metadata of a snapshot of a first file included in the snapshot of the distributed file system, wherein the first snapshot metadata identifies file content data of the first file in the distributed file system rather than a snapshot of the file content data included in the snapshot of the distributed file system and processing the first request includes:
• obtaining a snapshot file system identifier for a first data file system including file content data for the snapshot of the first file;
• generating, using the snapshot file system identifier, a globally unique identifier identifying file content data for the snapshot of the first file as included in the first data file system;
• determining a file handle for the globally unique identifier;
• and
• including the file handle in a response returned to the client including requested metadata for the snapshot of the first file.

A system for creating a snapshot of a distributed file system comprising:

• a computer readable medium comprising code stored thereon, wherein the code is executed by at least one processor and performs a method comprising:
• quiescing requests affecting the distributed file system, the distributed file system including a first set of one or more metadata file systems storing metadata about files in the distributed file system and including a second set of a plurality of data file systems storing file content data for files of the distributed file system;
• obtaining a snapshot of each metadata file system in the first set;
• obtaining a snapshot of each data file system in the second set;
• and
• unquiescing requests affecting the distributed file system upon successful completion of said obtaining said snapshot for each metadata file system in the first set and upon successful completion of said obtaining said snapshot of each data file system in the second set, whereby the snapshot of the distributed file system is created as a result of successfully completing said first obtaining step and said second obtaining step and wherein the snapshot of the distributed file system includes the snapshot of each metadata file system in the first set and includes the snapshot of each data file system in the second set;
• and
• one or more server systems, wherein code for each of the one or more metadata file systems and each of the plurality of data file systems is stored in a memory of any of the one or more server systems, said one or more server systems storing data on storage devices of a data storage system including a plurality of storage tiers, each of said plurality of storage tiers including physical storage devices having performance characteristics different from other physical storage devices of others of said plurality of storage tiers, each of said plurality of data file systems having data stored on physical devices included in one of said plurality of storage tiers, a quality of service parameter is specified for a first file of the distributed file system, said quality of service parameter specifying one of a plurality of different performance levels used to select which of the plurality of data file systems includes file content data for the first file;
• and
• wherein said quality of service parameter is specified for a directory of the distributed file system, said quality of service parameter specifying one of a plurality of different performance levels used to select which of the plurality of data file systems includes file content data for files created in the directory.