Storage virtualization by layering virtual disk objects on a file system

• US 7,107,385 B2
• Filed: 08/09/2002
• Issued: 09/12/2006
• Est. Priority Date: 08/09/2002
• Status: Active Grant

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Pawlowski, Brian, Srinivasan, Mohan, Rajan, Vijayan, Lewis, Blake H., Bolen, David Brittain
• Primary Examiner(s): Peugh, Brian R.
• Application Number:
  US10/216,453
• Publication Number:
  US 20040030822A1
• Time in Patent Office: 1,495 Days
• Field of Search: 711/711, 711/112, 711/114, 711/4, 709/219, 709/217, 707/100
• US Class Current: 711/4
• CPC Class Codes: G06F 16/10 File systems; File servers
• CPC Class Codes: G06F 3/0605 by facilitating the interaction with a user or administrator
• CPC Class Codes: G06F 3/0614 Improving the reliability of storage systems
• CPC Class Codes: G06F 3/0644 Management of space entities, e.g. partitions, extents, pools
• 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: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A storage virtualization selection technique “automates” a virtualization selection process to create virtual disk (vdisk) storage objects over a volume of a file system implemented by a storage operating system of a multi-protocol storage appliance. The file system provides a virtualization system that aggregates physical storage of a set of disks or portions (e.g., extents) of disks into a pool of blocks that can be dynamically allocated to form a vdisk. The file system also provides reliability guarantees for the vdisks in accordance with its underlying architecture. That is, the file system organizes its storage within volumes created among the managed disks. The vdisk is thereafter created as a storage object within a volume and, thus, inherits the underlying reliability configuration associated with that volume. The portions are aggregated and allocated as a vdisk with reliability guarantees in response to a request to create the vdisk from a user of the storage appliance and without further user involvement.

Claims

1. A method for creating a virtual disk (vdisk) on a storage system, the method comprising the steps of:

supporting at least one file system by the storage system;

aggregating physical storage of portions of disks into a pool of blocks;

dynamically allocating the portions of the disks to form a vdisk; and

providing reliability guarantees for the vdisk in accordance with an underlying architecture of the file system,

wherein the steps of aggregating, dynamically allocating and providing are performed in response to a user request to create the vdisk and without further user involvement.

2. The method of claim 1 wherein the portions of the disks are extents.

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

organizing storage of the file system within volumes created among the disks;

creating the vdisk as a storage object within one of the volumes; and

inheriting, on behalf of the vdisk, the underlying reliability configuration associated with the volume.

4. The method of claim 1 further comprising the step of one of growing and shrinking the vdisk under user control while preserving block and network attached storage (NAS) multi-protocol access to application data stored in the vdisk.

97. A computer readable media, comprising:

said computer readable media having instructions written thereon for execution on a processor for the practice of the method of claim 1, or claim 17, or claim 21, or claim 27, or claim 31, or claim 32.

5. A system for creating virtual disks (vdisks) on a storage system, the system comprising:

a storage operating system resident in a memory of the storage system and invoking storage operations in support of a file system configured to logically organize information as a hierarchical structure of vdisks within a volume of the storage system, each vdisk stored in the volume as a representation embodying a logical unit number (lun) inode functioning as a main container for storing data associated with the vdisk and at least one stream inode that functions as a persistent store for storing various attributes, wherein the lun inode and at least one associated stream inode are managed as a single encapsulated storage object within the file system.

6. The system of claim 5 wherein the at least one stream inode is an attributes inode.

7. The system of claim 6 wherein the lun inode comprises a data section that includes one of data and pointers referencing data blocks on the disk used to store the data.

8. The system of claim 7 wherein the lun inode further comprises a metadata section containing a type and size of the vdisk, the metadata section further containing a flag stream flag that identifies the lun inode as a data container and one or more stream sections.

9. The system of claim 6 wherein the attributes inode comprises a data section that functions as a persistent store for holding various named attributes associated with the vdisk, the attributes including information that allows the vdisk to be exported as a lun to clients of the storage system and that enable management of the vdisk as a single disk object in relation to the clients.

10. The system of claim 9 wherein the attributes further comprise at least one of serial number, space reservation, state and geometry information.

11. The system of claim 5 wherein each vdisk further embodies a lunmap inode that includes a data section functioning as a persistent store for holding a list of name-value pairs and a persistent reservations inode that includes a data section functioning as a persistent store for holding a list of persistent reservation records.

12. The system of claim 11 wherein the persistent reservation records provide ownership and access information relating to the vdisk.

13. The system of claim 5 further comprising a per-volume data structure stored in a metadata file that is used to optimize location determination and initialization of the vdisks in a volume.

14. The system of claim 13 wherein the per-volume data structure is a vdisk table of contents data structure comprising one or more records including file entry information that can be dynamically reconstructed from information stored in an encapsulated storage object representing the vdisk.

15. The system of claim 14 wherein each record includes file entries comprising:

a file identifier on the volume;

a generation number of the lun inode; and

directory information including a file block number in a parent directory containing an entry for the vdisk, along with an index of directory entries in a parent directory block.

16. The system of claim 5 wherein the vdisk encapsulates security information that restricts access to the vdisk, thereby providing multi-protocol access over either network attached storage or storage area network protocols while preserving data integrity.

17. A method for automating a selection process to create a virtual disk (vdisk) in a storage system using at least one storage disk, the method comprising the steps of:

issuing a lun create command through an administrative interface of the storage system, the lun create command specifying a size of the vdisk and a path descriptor of a location of the vdisk; and

creating the vdisk in response to the lun create command, the vdisk created on the at least one storage disk.

18. The method of claim 17 further comprising the step of resizing the vdisk using information pertaining to geometry of the vdisk.

19. The method of claim 18 wherein the step of resizing further comprises the step of algorithmically converting the specified vdisk size to the geometry information, the geometry information providing a limiting factor on how large the vdisk can be re-sized.

20. The method of claim 19 wherein the geometry information is cylinder size.

21. A method for creating a virtual disk (vdisk) in a storage system using at least one storage disk, the method comprising the steps of:

issuing a vdisk create command through an administrative interface of the storage system, the vdisk create command specifying a size of the vdisk and a path descriptor of a location of the vdisk;

creating a file inode associated with the vdisk, the file inode having a regular file type;

creating a stream inode for storing attributes associated with the vdisk;

populating the stream inode with said attributes; and

converting the file type of the file inode from regular to vdisk, thereby creating the vdisk on the at least one storage disk.

22. The method of claim 21 further comprising the steps of, after the step of issuing:

checking the specified size and path descriptor at a virtualization system of the multi-protocol storage appliance;

if the path descriptor and size do not check, aborting creation of the vdisk; and

if the path descriptor and size do check, proceeding to the step of creating a file inode.

23. The method of claim 21 wherein the step of creating a file inode further comprises the step of setting the specified size for the file inode.

24. The method of claim 21 wherein the attributes include a state of the vdisk, permissions for sharing the vdisk over network file system protocols and inquiry information.

25. The method of claim 24 wherein the step of populating further comprises the steps of:

setting a file owner and group identification parameter for use when the vdisk is exported over the network file system protocols; and

setting space reservation guarantees so that write operations to the vdisk do not fail.

26. A system for automating a selection process to create virtual disks (vdisks) on a storage appliance, 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

a storage operating system resident in the memory and invoking storage operations in support of a file system configured to logically organize information as a hierarchical structure of vdisks on the disk, each vdisk stored on the disk as a representation embodying a logical unit number (lun) inode functioning as a main container for storing data associated with the vdisk and an attributes inode that functions as a persistent store for storing various attributes, wherein the lun inode and attributes inode are managed as a single encapsulated storage object within the file system.

27. A method for automating a storage virtualization process to create a virtual disk (vdisk) on a storage appliance, the storage appliance using at least one storage disk, the method comprising the steps of:

issuing a vdisk create command through an administrative interface of the storage appliance, the vdisk create command specifying a size of the vdisk and a path descriptor of a location of the vdisk;

creating a file inode associated with the vdisk, the file inode having a vdisk file type;

creating a stream inode for storing attributes associated with the vdisk; and

populating the stream inode with said attributes, the vdisk supported by the at least one storage disk.

28. The method of claim 27 wherein the vdisk create command is a lun create command.

29. Apparatus for layering a virtual disk (vdisk) on a file system implemented by a storage operating system of a multi-protocol storage appliance, the apparatus comprising:

means for aggregating physical storage of portions of disks into a pool of blocks;

means for dynamically allocating the portions of the disks to form a vdisk; and

means for providing reliability guarantees for the vdisk in accordance with an underlying architecture of the file system,

wherein the means for aggregating, dynamically allocating and providing are performed in response to a user request to create the vdisk and without further user involvement.

30. A computer readable medium containing executable program instructions for storage virtualization of a virtual disk (vdisk) of a file system implemented by a storage operating system of a storage system, the executable program instructions comprising program instructions for:

issuing a lun create command through an administrative interface of the storage system, the lun create command specifying a size of the vdisk and a path descriptor of a location of the vdisk; and

creating a vdisk in response to the lun create command.

31. A method for storage virtualization of a virtual disk (vdisk) on a file system implemented on at least one storage disk, the method comprising the steps of:

providing an encapsulated storage object of the file system, the encapsulated storage object implemented on the at least one storage disk, the encapsulated storage object including a container for storing data associated with the vdisk and a persistent store for storing attributes associated with the vdisk; and

accessing the encapsulated storage object as a logical unit number.

32. A method for implementing a virtual disk (vdisk) on one or more physical disks, comprising:

issuing a create vdisk command, the create vdisk command specifying a logical unit number (lun) and a size for the vdisk; and

establishing the vdisk in response to the create vdisk command, the vdisk having a physical storage allocated over the one or more physical disks, and the vdisk being established without user involvement beyond issuing the create vdisk command.

33. The method as in claim 32, further comprising:

selecting the one or more physical disks from a pool of available physical disks.

34. The method as in claim 32, further comprising:

selecting the one or more physical disks from a pool of available physical disks by a storage operating system.

35. The method as in claim 32, further comprising:

allocating portions of the one or more physical disks for the vdisk by a storage operating system.

36. The method as in claim 32, further comprising:

organizing the one or more physical disks into a file system having volumes; and

creating the vdisk as a storage object within one of the volumes.

37. The method as in claim 32, further comprising:

organizing the one or more physical disks to have underlying reliability guarantees; and

providing the vdisk with reliability guarantees in response to the underlying reliability guarantees.

38. The method as in claim 32, further comprising:

growing a size of the vdisk under user control while maintaining access to data stored on the vdisk.

39. The method as in claim 32, further comprising:

shrinking a size of the vdisk under user control while maintaining access to data stored on the vdisk.

40. The method as in claim 32, further comprising:

associating a lun inode with the vdisk as a container for storing data associated with the vdisk.

41. The method as in claim 40, further comprising:

including a data section in the lun inode, the data section including data referencing blocks of data on a disk used to store the data.

42. The method as in claim 41, further comprising:

including in the data section in the lun inode pointers referencing blocks of data on a disk used to store the data.

43. The method as in claim 32, further comprising:

associating a stream inode with the vdisk, the stream inode functioning as a persistent store for storing attributes associated with the vdisk.

44. The method as in claim 43, further comprising:

including in the attributes information that allows the vdisk to be exported as a lun to clients of the storage system, and attributes that enable management of the vdisk as a single disk object by the clients.

45. The method as in claim 43, further comprising:

including in the attributes at least one of serial number, space reservation, state, and geometry information.

46. The method as in claim 32, further comprising:

including a lunmap inode that has a data section functioning as a persistent store for holding a list of name-value pairs, and a persistent reservation inode that includes a data section functioning as a persistent store for holding a list of persistent reservation records.

47. The method as in claim 32, further comprising:

providing persistent reservation records to provide ownership and access information relating to the vdisk.

48. The method as in claim 32, further comprising:

storing a data structure in a metadata file to optimize location of vdisks stored in a volume.

49. The method as in claim 32, further comprising:

storing a data structure in a metadata file to optimize initialization of vdisks stored in a volume.

50. The method as in claim 32, further comprising:

storing a vdisk table of contents of data structures having one or more records holding file entry information that can be dynamically reconstructed from information stored in an encapsulated storage object representing the vdisk.

51. The method as in claim 50, further comprising:

having a file identifier in the file entry information.

52. The method as in claim 50, further comprising:

having a generation number of the lun inode in the file entry information.

53. The method as in claim 50, further comprising:

having a directory information including a file block number in a parent directory containing an entry for the vdisk in the file entry information.

54. The method as in claim 50, further comprising:

having an index of directory entries in a parent directory block in the file entry information.

55. The method as in claim 32, further comprising:

encapsulating in the vdisk security information that restricts access to the vdisk thereby providing multi-protocol access over either network attached storage or storage area network protocols while preserving data integrity.

56. The method as in claim 32, further comprising:

using a lun create command as the vdisk create command.

57. The method as in claim 32, further comprising:

storing the vdisk on a RAID organized disk storage for the physical disks.

58. The method as in claim 32, further comprising:

creating the vdisk by a virtualization system without explicitly configuring the physical storage, so that the creation step is transparent to the user.

59. The method as in claim 32, further comprising:

dynamically reconstructing file entry information from information stored in an encapsulated storage object representing the vdisk.

60. The method as in claim 32, further comprising:

issuing the create vdisk command through a command line interface (CLI).

61. The method as in claim 32, further comprising:

issuing the create vdisk command through a graphical user interface (GUI).

62. An apparatus to implement a virtual disk (vdisk), comprising:

means for issuing a create vdisk command, the create vdisk command specifying a logical unit number (lun) and a size for the vdisk; and

means for establishing the vdisk in response to the create vdisk command, the vdisk having a physical storage allocated over one or more physical disks, and the vdisk being established without user involvement beyond issuing the create vdisk command.

63. The apparatus as in claim 62, further comprising:

means for selecting the physical disks from a pool of available physical disks by a storage operating system.

64. The apparatus as in claim 62, further comprising:

means for selecting the one or more physical disks from a pool of available physical disks by a storage operating system.

65. The apparatus as in claim 62, further comprising:

means for allocating portions of the one or more physical disks for the vdisk by a storage operating system.

66. A storage system apparatus for computer data, comprising:

a processor executing a storage operating system, the storage operating system accepting a create virtual disk command (create vdisk command), the create vdisk command specifying a logical unit number (lun) and a size for a virtual disk (vdisk); and

at least one physical disk, a physical storage for the vdisk being allocated over the at least one physical disks by the storage operating system in response to the create vdisk command, and the vdisk being established without user involvement beyond issuing the create vdisk command.

67. The apparatus of claim 66, further comprising:

a pool of available physical disks, the at least one physical disk selected from the pool of available physical disks, without user involvement beyond issuing the create vdisk command.

68. The apparatus of claim 66, further comprising:

a pool of available physical disks, the at least one physical disk selected from the pool of available physical disks by the storage operating system, without user involvement beyond issuing the create vdisk command.

69. The apparatus of claim 66, further comprising:

the at least one physical disk being divided into portions, selected ones of the portions being allocated for the vdisk, without user involvement beyond issuing the create vdisk command.

70. The apparatus of claim 66, further comprising:

a file system organized on the at least one physical disk, the file system having volumes, and the vdisk created as a storage object within one of the volumes.

71. The apparatus of claim 66, further comprising:

the at least one physical disk having underlying reliability guarantees, and the vdisk having reliability guarantees in response to the underlying reliability guarantees.

72. The apparatus of claim 66, further comprising:

user control to grow a size of the vdisk, while maintaining access to a data stored on the vdisk.

73. The apparatus of claim 66, further comprising:

user control to shrink a size of the vdisk, while maintaining access to a data stored on the vdisk.

74. The apparatus of claim 66, further comprising:

a lun inode serving as a container for data associated with the vdisk.

75. The apparatus of claim 66, further comprising:

a lun inode having a data section, the data section including data referencing blocks of data on a disk used to store the data.

76. The apparatus of claim 66, further comprising:

a lun inode having a data section, the lun inode having pointers referencing blocks of data on a disk used to store the data.

77. The apparatus of claim 66, further comprising:

a stream inode associated with the vdisk, the stream inode functioning as a persistent store for storing attributes associated with the vdisk.

78. The apparatus of claim 77, further comprising:

the attributes including information that allows the vdisk to be exported as a lun to clients of the storage system, and attributes that enable management of the vdisk as a single disk object by the clients.

79. The apparatus of claim 77, further comprising:

the attributes including at least one of serial number, space reservation, state, and geometry information.

80. The apparatus of claim 66, further comprising:

a lunmap inode that has a data section functioning as a persistent store for holding a list of name-value pairs, and a persistent reservation inode that includes a data section functioning as a persistent store for holding a list of persistent reservation records.

81. The apparatus of claim 66, further comprising:

persistent reservation records to provide ownership and access information relating to the vdisk.

82. The apparatus of claim 66, further comprising:

a metadata file to store data to optimize location of vdisks stored in a volume.

83. The apparatus of claim 66, further comprising:

a metadata file to store a data structure to optimize initialization of vdisks stored in a volume.

84. The apparatus of claim 66, further comprising:

a vdisk table of contents storing data structures having one or more records holding file entry information that can be dynamically reconstructed from information stored in an encapsulated storage object representing the vdisk.

85. The apparatus of claim 84, further comprising:

the file entry information including a file identifier.

86. The apparatus of claim 84, further comprising:

the file entry information including a generation number of the lun inode.

87. The apparatus of claim 84, further comprising:

the file entry information including directory information.

88. The apparatus of claim 84, further comprising:

the file entry information including a file block number in a parent directory containing an entry for the vdisk.

89. The apparatus of claim 84, further comprising:

the file entry information including a file block number in a parent directory containing an index of directory entries.

90. The apparatus of claim 66, further comprising:

security information encapsulated in the vdisk, the security information restricting access to the vdisk, thereby providing multi-protocol access over either network attached storage or storage area network protocols while preserving data integrity.

91. The apparatus of claim 66, further comprising:

a lun create command serving as the vdisk create command.

92. The apparatus of claim 66, further comprising:

the at least one physical disk organized as a RAID system to store the vdisk.

93. The apparatus of claim 66, further comprising:

a virtualization system to create the vdisk without explicitly configuring the physical storage, so that the creation step is transparent to the user.

94. The apparatus of claim 66, further comprising:

an encapsulated storage object representing the vdisk, information stored in the encapsulated storage object to dynamically reconstruct file entry information.

95. The apparatus of claim 66, further comprising:

a command line interface (CLI) for issuing the vdisk create command.

96. The apparatus of claim 66, further comprising:

a graphical user interface (GUI) for issuing the vdisk create command.