System and method for representing named data streams within an on-disk structure of a file system

• US 6,643,654 B1
• Filed: 06/25/2001
• Issued: 11/04/2003
• Est. Priority Date: 06/25/2001
• Status: Expired due to Term

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Muhlestein, Mark, Patel, Kayuri
• Primary Examiner(s): JUNG, DAVID YIUK
• Application Number:
  US09/891,159
• Time in Patent Office: 862 Days
• Field of Search: 707/1-10, 707/100-104, 707/200-206, 711/100-116, 711/117-146, 709/212-237
• US Class Current: 1/1
• CPC Class Codes: G06F 16/10 File systems; File servers
• CPC Class Codes: G06F 2003/0697 device management, e.g. handlers, drivers, I/O schedulers
• CPC Class Codes: G06F 3/0601 Interfaces specially adapted for storage systems
• CPC Class Codes: G06F 3/0622 in relation to access
• CPC Class Codes: G06F 3/0643 Management of files
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A technique provides on-disk representations of multiple named data streams for a file system of a network storage appliance. The network storage appliance or filer includes a file system that implements a Write Anywhere File Layout (WAFL) disk format, wherein files are described by inodes of which there may be various types, including directory and regular inodes. A named stream inode type is defined that represents named data streams in the WAFL file system. Multiple data streams may be stored on disk(s) of the filer as representations embodying the named stream inode type associated with a file. The names and file handles of the data streams are stored in a “hidden” directory within the file system that is referenced by the base mode. The hidden directory is represented as a stream_dir inode type.

Claims

1. A method for providing an on-disk representation of a named data stream associated with at least one of a directory inode and file inode of a file system of a network storage appliance, the method comprising the steps of:

allocating a first free inode of the file system;

inserting a stream_dir type into a first type field of the first free inode to thereby convert the first free inode to a stream directory inode, the stream directory inode having a first data section with a first pointer configured to reference a stream directory data block;

modifying a xinode field of the file inode to reference the stream directory inode;

allocating an entry of the stream directory data block to store a name of the named data stream, the allocated entry further allocating a second free inode of the file system; and

inserting a stream type into a second type field of the second free inode to convert the second free inode to a stream inode for the named data stream, the stream inode having a second data section with a second pointer that references a user data block associated with the named data stream.

2. The method of claim 1 further comprising the step of assigning a third pointer to the allocated entry of the stream directory data block that references the stream inode.

3. The method of claim 1 wherein the step of allocating a first free inode comprises the step of asserting a predetermined flag within the file inode to transform the file inode into a base inode having more than one data fork.

4. The method of claim 3 wherein the predetermined flag is a WAFL_flag_stream flag.

5. The method of claim 1 further comprising the steps of:

loading a root inode and its referenced data blocks from a disk into a memory of the network storage appliance; and

searching contents of the root inode data blocks for a directory name of the directory inode.

6. The method of claim 5 further comprising the steps of:

upon finding the directory name, loading the directory inode and its referenced data blocks into memory; and

searching contents of the directory inode data blocks for a filename of the file inode.

7. The method of claim 6 further comprising the step of, upon finding the filename, loading the file inode and its referenced data blocks into the memory.

8. A system for providing on-disk representations of multiple named data streams within a filer, the system comprising:

a processor;

a memory coupled to the processor and having locations addressable by the processor;

at least one disk coupled to the memory and processor; and

an operating system resident in the memory locations and invoking storage operations in support of a file system configured to logically organize information as a hierarchical structure of directory and file inodes on the disk, each named data stream stored on the disk as a representation embodying a stream inode associated with a file inode, each on-disk file inode having a default data stream and at least one named data stream inode.

9. The system of claim 8 further comprising a storage adapter interconnected with the processor, memory and disk, the storage adapter cooperating with the operating system to access the information stored on the disk.

10. The system of claim 8 further comprising a network adapter coupled to the processor and memory of the filer, the network adapter connecting the filer to a client over a computer network, the client interacting with the filer by exchanging packets encapsulating a record requesting file services from the filer using a file system protocol over the network.

11. The system of claim 10 wherein the file system protocol is a Common Internet File System (CIFS) protocol and wherein the record is a CIFS record comprising information pertaining to an operation directed to the named data stream.

12. The system of claim 11 wherein the operating system comprises a series of software layers, including a file system protocol layer configured to support the CIFS protocol and a Write Anywhere File Layout (WAFL) layer configured to implement the file system.

13. The system of claim 12 wherein the CIFS record is interpreted as directed to a named data stream associated with a file and transformed into a message structure by the CIFS layer, and further passed to the WAFL layer, where the operation is performed.

14. The system of claim 13 wherein the message is passed from the CEFS layer to the WAFL layer as a function call.

15. The system of claim 13 wherein the WAFL layer loads the stream inode from disk into memory and accesses the stream inode as instructed by the operation.

16. The system of claim 12 wherein the operating system further comprises a media access layer of network drivers, network protocol layers, a disk storage layer that implements a disk storage protocol and a disk driver layer that implements a disk access protocol.

17. The system of claim 16 wherein a storage access request data path through the operating system layers enables performance of data storage access for the client request received at the filer.

18. The system of claim 17 wherein the storage access request data path is implemented as logic circuitry embodied within a hardware circuit.

19. A computer readable medium containing executable program instructions for providing an on-disk representation of a named data stream associated with at least one of a directory inode and file inode of a file system, the executable program instructions comprising program instructions for:

allocating a first free inode of the file system;

inserting a stream_dir type into a first type field of the first free inode to thereby convert the first free inode to a stream directory inode, the stream directory inode having a first data section with a first pointer configured to reference a stream directory data block;

modifying a xinode field of the file inode to reference the stream directory inode;

allocating an entry of the stream directory data block to store a name of the named data stream, the allocated entry further allocating a second free inode of the file system; and

inserting a stream type into a second type field of the second free inode to convert the second free inode to a stream inode for the named data stream, the stream inode having a second data section with a second pointer that references a user data block associated with the named data stream.

20. Apparatus for providing an on-disk representation of a named data stream associated with at least one of a directory inode and file inode of a file system of a filer;

the apparatus comprising;

means for allocating a first free inode of the file system;

means for inserting a stream_dir type into a first type field of the first free inode to thereby convert the first free inode to a stream directory inode, the stream directory inode having a first data section with a first pointer configured to reference a stream directory data block;

means for modifying a xinode field of the file inode to reference the stream directory inode;

means for allocating an entry of the stream directory data block to store a name of the named data stream, the allocated entry further allocating a second free inode of the file system; and

means for inserting a stream type into a second type field of the second free inode to convert the second free inode to a stream inode for the named data stream, the stream inode having a second data section with a second pointer that references a user data block associated with the named data stream.

21. A storage system for providing on-disk representations of multiple named data streams within a computer, the system comprising:

an operating system resident in a memory and invoking storage operations in support of a file system configured to logically organize information as a hierarchical structure of directory and file inodes on the disk, each named data stream stored on the disk as a representation embodying a stream inode associated with a file inode, each on-disk file inode having a default data stream and at least one named data stream inode.

22. A method for accessing an on-disk representation of a named data stream associated with at least one of a directory inode and file inode of a file system of a server computer, the method comprising the steps of:

sending a packet from a client to the server over a network medium, the packet encapsulating a record requesting a file service directed to the named data stream;

creating a message from the record and passing the message to a file system layer of the server;

loading a stream inode and a referenced data block for the named data stream from a disk into a memory of the server; and

processing the stream inode in accordance with the requested file service.

23. The method of claim 22 wherein the requested file service is a write operation.

24. The method of claim 23 further comprising the steps of:

loading a base inode of the stream inode from the disk into the memory;

updating a modification time stamp stored in the base inode for the named data stream;

updating the referenced data block as instructed by the write operation;

marking the referenced data block, the base inode and the stream inode dirty; and

flushing the marked block and inodes to the disk.

25. The method of claim 22 wherein the requested file service is a read operation.

26. The method of claim 25 further comprising the steps of:

loading a base inode of the stream inode from the disk into the memory;

updating an access time stamp stored in the base inode for the named data stream;

retrieving data contained in the referenced data block for delivery to the client;

marking the base inode dirty; and

flushing the marked inode to the disk.

27. A method for accessing an on-disk representation of a named data stream associated with at least one of a directory inode and file node of a file system of a server computer, the method comprising the steps of:

sending a packet from a client to the server over a network medium, the packet encapsulating a record requesting a file service directed to the named data stream;

creating a message from the record and passing the message to a file system layer of the server;

loading a base inode referenced by a file handle of the message from a disk into a memory of the server;

determining whether a predetermined flag of the base inode is asserted and, if so, loading a stream directory inode referenced by the base inode, along with a stream directory block referenced by the stream directory inode, into the memory;

loading a stream inode for the named data stream referenced by the stream directory block into the memory; and

retrieving (i) a name of the stream inode from the stream directory block and (ii) a size of the stream inode from the stream inode for delivery to the client.

28. A multi-protocol data access storage system for providing on-disk representations of multiple named data streams within a computer, the system comprising:

an operating system resident in a memory of the computer and invoking storage operations in support of a file system configured to logically organize information as a hierarchical structure of directory and file inodes on the disk, the operating system including a file system protocol layer configured to provide data access in support of a plurality of file system protocols, each named data stream stored on the disk as a representation embodying a stream inode associated with a file inode, each on-disk file inode having a default data stream and at least one named data stream inode.

29. The multi-protocol data access storage system of claim 28 wherein one of the file system protocols is a Common Internet FileSystem (CIFS) protocol.