Methods, systems, and computer program products for providing access to shared storage by computing grids and clusters with large numbers of nodes

• US 7,676,628 B1
• Filed: 03/31/2006
• Issued: 03/09/2010
• Est. Priority Date: 03/31/2006
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Fridella, Stephen, Faibish, Sorin, Jiang, Xiaoye, Gupta, Uday K., Compton, James T., Clark, Roy E.
• Primary Examiner(s): Tsai; Sheng-Jen
• Assistant Examiner(s): Krofcheck; Michael C
• Application Number:
  US11/394,768
• Time in Patent Office: 1,439 Days
• Field of Search: 711/114, 711147-148, 711/154, 711202-203
• US Class Current: 711/114
• CPC Class Codes: G06F 3/0605 by facilitating the interaction with a user or administrator
• CPC Class Codes: G06F 3/0665 at area level, e.g. provisioning of virtual or logical volumes
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Methods, systems, and computer program products for providing access to shared storage by a plurality of nodes are disclosed. According to one method, at a node of a plurality of nodes sharing access to a disk array, an application input/output (I/O) operation and whether sufficient space in the disk array has been mapped at the node for the I/O operation are detected. In response to detecting that sufficient space in the disk array has not been mapped at the node for the I/O operation, a map for the I/O operation including physical disk access information is requested and obtained from a server. The physical disk access information included within the map for the application I/O operation obtained from the server is used to perform the I/O operation by accessing the disk array without intervention by the server.

Claims

1. A method for providing access to shared storage by a plurality of nodes, the method comprising:

at a client node of a plurality of nodes separate from and sharing access to a disk array;

(a) detecting a first application input/output (I/O) operation;

(b) detecting whether sufficient space in the disk array has been mapped at the node for the first application I/O operation;

(c) in response to detecting that sufficient space in the disk array has not been mapped at the client node for the first application I/O operation, requesting, from a server, and obtaining, from the server, a map including physical disk access information for the first application I/O operation, wherein the physical disk access information includes a physical address of a location within a disk of the disk array; and

(d) performing the first application I/O operation by accessing the disk array without intervention by the server using the physical disk access information included within the map for the application I/O operation obtained from the server, wherein the client node maintains the map and uses the map for subsequent application I/O operations including accessing physical disks of the disk array directly without interaction with the server.

2. The method of claim 1 wherein the first application I/O operation includes a write request and wherein the write request is performed by writing data to the disk array without intervention by the server and wherein the map for the first application I/O operation obtained from the server includes a write data access map.

3. The method of claim 2 comprising, at the server, allocating at least one data storage block in the disk array to create a data storage extent for a region of a virtual address space associated with the write request.

4. The method of claim 3 comprising, at the server, creating the physical disk access information and logical storage extent access information associated with the created data storage extent at a logical location for the region of the virtual address space associated with the write request and marking a state of the physical disk access information as invalid within the map for the first application I/O operation.

5. The method of claim 4 wherein accessing the disk array without intervention by the server to perform the write request includes storing data associated with the write request within the data storage extent using the physical disk access information associated with the data storage extent within the map for the first application I/O operation.

6. The method of claim 5 comprising performing a commit operation to update a state of the physical disk access information associated with the created data storage extent at the server to valid.

7. The method of claim 5 comprising performing a commit operation to update a state of the physical disk access information associated with the created data storage extent within the map for the first application I/O operation to valid.

8. The method of claim 7 comprising creating a read data access map and storing the physical disk access information associated with the created data storage extent within the read data access map and setting a state of the physical disk access information to valid.

9. The method of claim 7 comprising creating a read/write data access map and storing the physical disk access information associated with the created data storage extent within the read/write data access map and setting a state of the physical disk access information to valid.

10. The method of claim 9 comprising, in response to a second application I/O operation, wherein the second application I/O operation includes a read request, reading data associated with a region of a virtual address space associated with the read request from the storage extent in the disk array using the physical disk access information.

11. The method of claim 1 wherein the first application I/O operation includes a read request and wherein the read request is performed by reading data from the disk array without intervention by the server and wherein the map for the first application I/O operation includes a read data access map.

12. The method of claim 11 comprising, at the server, creating logical storage extent access information associated with a region of a virtual address space associated with the read request in the map for the first application I/O operation and setting a state of the logical storage extent access information associated with the region of the virtual address space associated with the read request within the map to none.

13. The method of claim 11 comprising, at the client node, creating logical storage extent access information associated with a region of a virtual address space associated with the read request in the map for the first application I/O operation and setting a state of the logical storage extent access information associated with the region of the virtual address space associated with the read request within the map to none.

14. The method of claim 1 wherein the first application I/O operation includes a read request and wherein the read request is performed by returning a zero-filled buffer in response to detecting that sufficient space in the disk array has not been mapped at the node for the first application I/O operation.

15. A system for providing access to shared storage by a plurality of nodes, the system comprising:

(a) a disk array;

(b) a data storage extent pool representing at least a portion of the disk array and including at least one data storage block;

(c) a server for tracking and providing physical disk access information associated with the at least one data storage block of the data storage extent pool to a plurality of client nodes sharing access to the disk array, the physical disk access information maintained within a file system with logical units (LUNS) associated with the disk array mapped to files; and

(d) a plurality of client nodes separate from and sharing access to the disk array, wherein the plurality of client nodes includes a first client node adapted to, in response to a first application input/output (I/O) operation;

(i) detect whether sufficient space in the disk array has been mapped at the node for the first application I/O operation;

(ii) in response to detecting that sufficient space in the disk array has not been mapped at the first client node for the first application I/O operation, request, from the server, and obtain, from the server, a map including physical disk access information for the first application I/O operation, wherein the physical disk access information includes a physical address of a location within a disk of the disk array; and

(iii) use the map for the first application I/O operation obtained from the server to perform the first application I/O operation by accessing the disk array using the physical disk access information without intervention by the server, wherein the first client node maintains the map and uses the map for subsequent application I/O operations including accessing physical disks of the disk array directly without interaction with the server.

16. The system of claim 15 wherein the first application I/O operation includes a write request and wherein the first client node is adapted to perform the write request by writing data to the disk array using the physical disk access information without intervention by the server and wherein the map for the first application I/O operation obtained from the server includes a write data access map.

17. The system of claim 16 wherein the server is adapted to allocate at least one data storage block in the disk array to create a data storage extent for a region of a virtual address space associated with the write request.

18. The system of claim 17 wherein the server is adapted to create the physical disk access information and logical storage extent access information associated with the created data storage extent at a logical location for the region of the virtual address space associated with the write request and to mark a state of the physical disk access information as invalid within the map for the first application I/O operation.

19. The system of claim 18 wherein the first client node is adapted to access the disk array without intervention by the server to perform the write request by storing data associated with the write request within the data storage extent using the physical disk access information associated with the data storage extent within the map for the first application I/O operation.

20. The system of claim 19 wherein the node is adapted to perform a commit operation to update a state of the physical disk access information associated with the created data storage extent at the server to valid.

21. The system of claim 19 wherein the first client node is adapted to perform a commit operation to update a state of the physical disk access information associated with the created data storage extent within the map for the first application I/O operation to valid.

22. The system of claim 21 wherein the first application I/O operation includes a read request and wherein the server is adapted to store the physical disk access information associated with the created data storage extent and to set a state of the physical disk access information to valid within the map for the first application I/O operation.

23. The system of claim 21 wherein the first application I/O operation includes a write request and wherein the server is adapted to store the physical disk access information associated with the created data storage extent and to set a state of the physical disk access information to valid within the map for the first application I/O operation.

24. The system of claim 23 wherein the node is adapted to, in response to a second application I/O operation, wherein the second application I/O operation includes a read request, read data associated with a region of a virtual address space associated with the read request from the storage extent in the disk array using the physical disk access information.

25. The system of claim 15 wherein the first application I/O operation includes a read request and wherein the first client node is adapted to perform the read request by reading data from the disk array using the physical disk access information without interaction by the server and wherein the map for the first application I/O operation includes a read data access map.

26. The system of claim 25 wherein the server is adapted to create logical storage extent access information associated with a region of a virtual address space associated with the read request in the read data access map and to set a state of the logical storage extent access information associated with the region of the virtual address space associated with the read request within the map to none.

27. The system of claim 25 the first client node is adapted to create logical storage extent access information associated with a region of a virtual address space associated with the read request in the read data access map and to set a state of the logical storage extent access information associated with the region of the virtual address space associated with the read request within the map to none.

28. The system of claim 15 wherein the first application I/O operation includes a read request and wherein the node is adapted to perform the read request by returning a zero-filled buffer in response to detecting that sufficient space in the disk array has not been mapped at the node for the first application I/O operation.

29. A system for providing access to shared storage and shared services in a redundant array of inexpensive disks (RAID) environment by computing grids and clusters with large numbers of nodes, the system comprising:

(a) a disk array;

(b) a data storage extent pool representing at least a portion of the disk array and including at least one data storage block;

(c) a plurality of client nodes separate from and sharing access to the disk array, the plurality of client nodes including a first client node adapted to, in response to an application input/output (I/O) operation;

(i) detect whether sufficient space in the disk array has been mapped at the node for the application I/O operation;

(ii) in response to detecting that sufficient space in the disk array has not been mapped at the first client node for the application I/O operation, request, from a server, and obtain, from the server, a map for the application I/O operation, wherein the physical disk access information includes a physical address of a location within a disk of the disk array; and

(iii) use the map for the application I/O operation obtained from the server to perform the application I/O operation by accessing the disk array without intervention by the server, wherein the first client node maintains the map and uses the map for subsequent application I/O operations including accessing physical disks of the disk array directly without interaction with the server.

30. A client node for accessing shared storage in a disk array using information provided by a server, the client node comprising:

(a) an application for performing an input/output (I/O) operation;

(b) an extent map cache for storing mappings between logical identifiers and physical storage blocks within a disk array, wherein the mappings include a physical address of a location within a disk of the disk array;

(c) a fragment mapping protocol client for, in response to the I/O operation and a determination that a mapping does not exist for the I/O operation in the extent map cache, requesting, from a server, and obtaining, from the server, a mapping for the I/O operation;

(d) a fragment mapping protocol pseudo-device for obtaining a physical disk identifier for the I/O operation from the extent map cache; and

(e) a disk interface for performing the I/O operation using the physical disk identifier, wherein the client node maintains the mapping and uses the mapping for subsequent I/O operations including accessing physical disks of the disk array directly without interaction with the server and wherein the client node is separate from the disk array.

31. A computer program product comprising computer-executable instructions embodied in a computer-readable medium for performing steps comprising:

at a client node of a plurality of client nodes separate from and sharing access to a disk array;

(a) detecting an application input/output (I/O) operation;

(b) detecting whether sufficient space in the disk array has been mapped at the client node for the application I/O operation;

(c) in response to detecting that sufficient space in the disk array has not been mapped at the client node for the application I/O operation, requesting, from a server, and obtaining, from the server, a map for the application I/O operation, wherein the physical disk access information includes a physical address of a location within a disk of the disk array; and

(d) using the map for the I/O operation obtained from the server to perform the I/O operation by accessing the disk array without intervention by the server, wherein the client node maintains the map and uses the map subsequent application I/O operations including accessing physical disks of the disk array directly without interaction with the server.