Fast backup storage and fast recovery of data (FBSRD)

• US 20050216788A1
• Filed: 05/04/2005
• Published: 09/29/2005
• Est. Priority Date: 11/20/2002
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: Filesx Limited
• Inventors: Mani-Meitav, Irit, Sarfati, Assaf
• Granted Patent: US 7,596,713 B2
• Time in Patent Office: Days
• US Class Current: 714/6
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 11/1456 Hardware arrangements for backup
• CPC Class Codes: G06F 11/1458 Management of the backup or restore process
• CPC Class Codes: G06F 11/1464 for networked environments
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A Fast Backup Storage and fast Recovery of Data (FBSRD) method for a facility, preferably with a SAN, coupled to a network with servers and workstations, operating in both a storage mode and a recovery mode. Coupled to the network are a primary storage, a repository, and a Backup Appliance computer BA. The BA runs a Backup Computer Program in association with at least one computer. When in storage mode, data is retrieved out of primary storage for back up, by taking snapshots and saving the retrieved data into repository in block format. In recovery mode, backed-up data is retrieved from repository and recovered into primary storage in either one of both block format and file format. The BA runs interactively with an Agent Computer Program residing in each server and workstation, and with a Backup User Interface management computer program operated by a user.

Claims

1. A method for Fast Backup Storage and fast Recovery of Data (FBSRD) operating both separately and simultaneously in a first storage mode and in a second recovery mode in a computing facility, which includes:

a network for communication,

at least one computer, coupled to the network, for operation by a user and for managing user data, and

a primary storage coupled to the at least one computer and configured as a first data storage memory for storing therein data managed by the at least one computer, the method comprising the steps of;

coupling a Backup Appliance computer BA to the network, where the BA is a computer with an operating system having a file-system,

providing a repository coupled to the BA, and configured as a second data storage memory, and

operating FBSRD by running a Backup Computer Program in association with the at least one computer and with the BA, including;

in storage mode;

retrieving data out of primary storage for back-up, by taking snapshots of data at predetermined successive time intervals,

saving the retrieved data into repository in block format, and in recovery mode;

recovering backed-up data out of the repository for use by the at least one computer only, and

recovering data into primary storage in either one of both block format and file format.

2. The method according to claim 1, wherein the computer facility has at least one server coupled to the network for managing user data, the method including:

coupling a SAN to the at least one server and to the backup appliance computer BA,

providing at least one primary storage coupled to the SAN,

using the at least one server to store user data on primary storage via the SAN, and

using the BA to access primary storage via the SAN.

3. The method according to claim 1, wherein the primary storage and the repository comprise at least one non-volatile data storage memory selected, alone and in combination, from the group consisting of internal computer memory, direct attach storage memory, LAN-coupled memory, and SAN-coupled memory, including:

coupling the repository to the backup appliance computer via at least one communication link, and

configuring the primary storage and the repository as random access Read/Write disk storage devices for transferring data by disk-to-disk data transfer at disk-to-disk speed, both when in storage mode and when in recovery mode.

4. The method according to claim 1, wherein the backup computer program ( ) has:

a Backup Manager Computer Program (BMCP) running interactively with at least one Agent Computer Program (ACP) and with a Backup User Interface management computer program (BUI), the method including;

running the BMCP on the backup appliance computer for managing operation of the FBSRD,

running the ACP on the at least one computer, for monitoring and for protecting access by the at least one computer to primary storage, and

running the BUI, when desired, on a workstation coupled to the network, for enabling a user of the BUI-running workstation to select and define therewith a backup configuration of computing facility devices operating FBSRD, and for visualization of current FBSRD-process status of the computing facility.

5. The method according to claim 4, wherein operating the backup user interface management computer program-running workstation includes:

designating a data backup range as spanning from at least one volume out of a plurality of file-system volumes residing in primary storage to an entire primary storage,

selecting the data in the designated backup range as source data for backup,

defining a first snapshot initiation time and a time-interval between successive snapshots, and

defining additional parameters for controlling backup operations, such as total number of successive snapshots in a sequence, and further snapshot timing parameters.

6. The method according to claim 5, wherein backup is achieved in storage mode by snapshooting primary storage, including:

enabling protection of data comprised in an ongoing snapshot for the duration thereof, to prevent storage of inconsistent data caused by write instructions emanating from any computer, such as a backup client, coupled to the primary storage from which the ongoing snapshot is being taken, and

taking a snapshot for storing data in repository in block level data format, by copying data according to either one of both;

copying only selected data blocks which are detected as being valid data blocks when the ongoing snapshot is a first initiated snapshot in a sequence of snapshots, and

copying only selected data blocks, which are detected as being incremental and valid data blocks relative to a previous snapshot.

7. The method according to claim 6, wherein a snapshot of data backed-up in block-format in the repository is processed for recovery in file-level format to at least one computer, including:

running the backup manager computer program BMCP for building and for storing in repository of a catalog including file-level system information, and of a Look-Up Table (LUT) comprising block-level data location information, the catalog and the LUT information being mutually indexed for association with data in the snapshot,

running the BMCP for recovering a file selected from the snapshot by associating therewith indexed information saved in the catalog and in the LUT file, and

recovering the file from the backup appliance computer, via the network, to the at least one computer.

8. The method according to claim 7, wherein building the catalog includes:

running the backup manager computer program for access and for analysis of file-system information pertaining to a file system comprising a plurality of objects stored as a snapshot of data blocks in repository,

analyzing the snapshot for deriving therefrom and accordingly indexing file-system type and file-system meta-data,

analyzing the derived file-system meta-data for deriving therefrom and indexing file-system attributes of all objects out of the plurality of objects in the file-system, comprising file and directory information such as names and relative location, and

building a catalog in file-system catalog format, and comprising, file-system information format, file-system type, file-system structure, file-system data, and file-system meta-data.

9. The method according to claim 7, wherein, when in recovery mode, for both recovering into primary storage and for executing analysis by running an application computer program in file-level format of a file selected from a snapshot residing in repository in block format, includes:

a. reading file-system information from the catalog,

b. reading data block information from the look-up table,

c. associating and converting the data read in steps a. and b. into organized file-system format data,

d. sending the organized data through the network to the at least one computer, and

e. using the organized data received by the at least one computer as file-system information source for operations, such as for executing saving to primary storage and for executing analysis by running an application computer program.

10. The method according to claim 6, wherein, when in recovery mode, recovery of data blocks includes:

selecting a snapshot stored in repository in block-level format as source data to be recovered by a recovery operation,

allocating a region in primary storage for storage of data recovered in original location, the region being defined as a target,

running the backup computer program BCP for enabling protection against attempts by a backup client associated with a volume of data located in the target, to access the target, by detaching or dismounting the volume of data from the backup client'"'"'s file system,

reading data and location information from the selected snapshot, as recovered data, and writing the recovered data as data recovered to original location in the target, by running the BCP,

reattaching or remounting to the backup client'"'"'s file system, any file system volume currently located in the recovered region, if desired, by running the BCP, thereby providing access to backup clients associated with the recovered volume of data.

11. The method according to claim 6, for data block recovery when in recovery mode, including:

running a Read File System , or RFS computer program on a server coupled to a repository via a SAN,

selecting from the repository and attaching or mounting of a selected snapshot as a selected file system,

running the server'"'"'s file-system to access file data and file meta-data stored on and previously created from the selected snapshot,

running the RFS for accepting all data block read-requests from the server'"'"'s file system directed to the selected volume when either one of both file data and file meta-data is requested,

using data retrieved from a look up table ) pertaining to the selected snapshot and stored in repository to locate the requested data block in the repository, and

reading the requested data block(s) from the repository,

whereby the server may access file system information from a snapshot stored in repository in block-level format.

12. The method according to claim 5, including:

running the backup user interface computer program (BUI), when desired, on a workstation attached to the network and interactively coupled with the backup manager computer program, the BUI being configured for allowing an operator to manage and monitor computing facility information related to backup, and recovery, as well as for configuring and for scheduling FBSRD operations.

13. The method according to claim 6, wherein reading data blocks from SAN-coupled primary storage includes:

selecting data blocks to be read from primary storage, and

running the backup manager computer program on the backup appliance computer BA for reading the selected data blocks over the SAN from primary storage and into a BA memory coupled to the BA,

whereby selected data blocks are read from primary storage and written into BA memory.

14. The method according to claim 6, wherein reading data blocks by the backup appliance computer BA out of primary storage configured as Direct Attach Storage coupled to a backup client, includes:

selecting data blocks to be read from primary storage,

running the backup manager computer program BCP for sending a command over the network from the backup manager computer program BMCP to the agent computer program ACP for the latter to read the selected data blocks,

running the ACP to read the selected data from primary storage to the at least one computer memory, and,

running the BCP for sending the selected data blocks over the network, from the ACP to the BMCP, out of backup client-coupled memory for writing into the backup appliance computer BA memory,

whereby selected data blocks are read from primary storage and written into BA memory.

15. The method according to claim 6, wherein selecting valid data blocks for FBSRD includes:

selecting a desired scope of source data ranging from at least one file-system volume, and up to an entire primary storage device,

adding data blocks with layout meta data, which are defined as either one of both disk layout meta data and file-system volume layout meta data, to the list of valid data blocks,

selecting data blocks with layout meta data for reading into the backup appliance computer memory, by running the backup computer program , and analyzing the selected data blocks for detecting attributes such as name, type and extent of file-system volume associated therewith, and

adding, when the analyzed information matches the desired scope of backup data, valid data blocks associated with the selected file system volume to the list of valid data blocks.

16. The method according to claim 15, wherein selecting valid data blocks pertaining to a file-system volume for FBSRD includes:

selecting a file-system volume having disk layout meta data with file system attributes such as type and extent,

selecting file-system meta-data blocks listing valid data blocks associated with the file system, the selection method depending on type and extent of selected file-system,

running the backup computer program to read the selected data blocks into the backup appliance computer memory, analyzing the selected meta-data blocks for derivation therefrom of information listing valid data blocks, and

adding the derived information listing to the list of valid data blocks,

whereby valid data blocks associated with the selected file-system volume are added to the list of valid data blocks.

17. The method according to claim 6, wherein detecting incremental data blocks includes:

running an agent computer program ACP on each backup client,

sending a command from the backup manager computer program BMCP to the ACP, via the network, to start monitoring a selected region of the backup client-coupled primary storage,

running the ACP for creating an initially-empty list of all incremental data blocks residing in the selected region of the backup client-coupled primary storage, and associating a unique ID to the initially-empty list of incremental data blocks,

running the ACP for intercepting data write instruction directed by a backup client to the backup client-coupled primary storage associated therewith, and adding to the list of incremental data blocks the primary storage data location to which the write instruction is targeted,

selecting and associating a snapshot storage operation with the unique ID assigned to the list of incremental data blocks, and

running the ACP and the BMCP for copying the list of incremental blocks from backup client memory to the backup appliance computer memory via the network, when the selected snapshot enters backup storage mode,

whereby incremental data blocks are selected for backup storage operation.

18. The method according to claim 16, wherein selecting data blocks for backup storage includes:

selecting each one backup client having access to a primary storage selected for data backup storage,

running the backup manager computer program BMCP and the agent computer program associated with each one selected backup client for retrieving therefrom, over the network, of a list of incremental data blocks associated with the relevant snapshot,

running the BMCP for selecting valid data blocks pertaining to the primary storage being backed up, and

selecting for backup storage, all valid data blocks which are also incremental data blocks out of the selected backup clients,

whereby only valid incremental data blocks are selected for data backup storage.

19. The method according to claim 10, wherein data selected for a snapshot is protected against alteration by:

starting a snapshot by sending a command from the backup appliance computer BMCP, via the network to the agent computer program ACP, to start protecting a selected region of the backup client-coupled primary storage,

using the ACP for intercepting each block level data write instruction directed by the backup client to the backup client-coupled primary storage, and delaying execution of each write operation ante factum,

running the ACP for checking a possible match of the location into which data is to be written, with the selected region of primary storage being protected,

running the ACP for reading the data currently residing in the primary storage location defined as a target for write operation when the location matches, and for sending the data read from the defined primary location, via the network, from the ACP to the BMCP,

running the ACP for enabling completion of the write operation emanating from the backup client,

defining appropriate data for a snapshot as data residing in the primary storage before or prior to beginning of the snapshot,

running the BMCP for determining the appropriate data for the snapshot, out of data received from the ACP and out of data read from primary storage, and saving the appropriate data in the repository, and

sending a command from the BMCP via the network to the ACP, when the snapshot is completed, to stop protecting the selected region, whereby data in primary storage is protected against alteration by a backup client during snapshot storage operation.

20. The method according to claim 6, wherein the computing facility further has a server cluster, the method including:

running an agent computer program ACP in each one server out of the server cluster,

commanding the ACP in each one of the backup clients to enable snapshot protection, and

running the backup computer program for conflict resolution during an ongoing snapshot, when protection of data is required against attempts, emanating from more than one backup client pertaining to the server cluster, to write to a primary storage location comprised in the ongoing snapshot,

whereby data copied during a snapshot storage operation is protected against alteration.

21. The method according to claim 6, wherein writing a copy of selected data blocks to repository includes:

using the backup appliance computer BA for selecting and defining a region of primary storage residing in a BA-coupled storage device, as repository,

running the BA and the BA'"'"'s file-system to format the repository as an empty file-system structure,

selecting data blocks for backup from BA memory, and running the backup manager computer program BMCP and the BA'"'"'s file-system to write the contents of the selected sectors to a file in repository,

running the BCMP to prepare a Look-Up Table (LUT) associated with a snapshot and containing location information of data blocks selected for backup, the location information comprising location in source primary storage and location in repository, the location in repository further comprising the repository file name and location in the repository file, and

running the BMCP and the BA'"'"'s file-system, to write the LUT to a file in repository,

whereby a copy of the contents and the location of data blocks selected for backup is saved in repository.

22. The method according to claim 21, including:

selecting a copy of a snapshot for retrieval from repository,

selecting a data block for retrieval from the selected snapshot copy,

running the file-system of the backup appliance computer BA to read the repository-resident Look-Up Table LUT file associated with the selected snapshot copy into BA memory,

running the backup manager computer program to retrieve the location of information in primary storage and in repository of the selected data block from the LUT,

running the file-system of the BA to read the selected data block from the retrieved location in the repository,

whereby a data block is read from the repository.

23. The method according to claim 22, including:

selecting a first snapshot out of a sequence of snapshots from repository for the creation therefrom of a catalog,

running the backup manager computer program BMCP for reading disk layout meta data associated with the selected first snapshot into the backup appliance computer BA memory, and analyzing the disk layout meta data to determine type and extent of all file system volumes associated with the selected first snapshot,

selecting each one of the file system volumes associated with the selected snapshot for analysis, the analysis further comprising the steps of;

running the BMCP for reading a node table associated with the selected file-system volume into BA memory,

running the BMCP for analyzing a directory tree associated with the selected file system volume, building a catalog entry for each file and sub-directory in the directory tree and linking the catalog entries in a structure representing the file-system structure, and

saving the catalog to the repository when all file system volumes have been analyzed,

whereby a catalog of the first snapshot of a sequence is saved to the repository.

24. The method according to claim 23, wherein analysis of the directory tree includes analyzing all directories and files iteratively, starting at the first directory, or root directory, and ending when all files have been analyzed, and the analysis of each file including the steps of:

selecting a next file according to result of analysis of the file'"'"'s parent directory,

retrieving, as result from analysis of parent directory, the desired file'"'"'s attributes, such as node-table entry, name, type, size and last modification date,

creating a new catalog entry in the backup appliance computer BA memory and copying the desired attributes thereto,

linking the new catalog entry to the parent directory'"'"'s catalog entry in BA memory, and for a file type which is a directory, further performing the steps of;

analyzing the selected directory'"'"'s entry in the node table to determine data blocks in the snapshot containing directory data,

reading the selected directory'"'"'s data from the snapshot into BA memory, and

preparing for further analysis of all files aggregated in the directory,

whereby there is created a catalog containing a representation of all directories and files in the file system, as well as the file system structure.

25. The method according to claim 23, including:

saving a catalog of the first snapshot out of a sequence of snapshots into repository,

selecting each subsequent snapshot following the first snapshot of the sequence out of repository to create a catalog,

running the backup management computer program BMCP for building a list of valid data blocks residing in the selected snapshot,

correlating between incremental data blocks in the selected snapshot and the valid data block list, ignoring incremental data blocks not represented in the valid data block list and selecting for catalog analysis, incremental blocks which are represented in the valid data block list,

running the BMCP for reading the node table associated with the selected snapshot into the backup appliance computer BA memory,

correlating between selected incremental data blocks and the node table residing in BA memory, and selecting for catalog analysis only incremental blocks which present changes in file-system meta-data,

analyzing selected incremental data blocks for detecting changes in file-system structure, such as modifying, adding, and deleting files and directories, and

adding detected changes to the catalog entries representing changed files and directories,

whereby file-system history of changes is added to the catalog according to incremental snapshot analysis.

26. The method according to claims 25, wherein building the list of valid data blocks residing in the selected snapshot includes:

reading data blocks of the selected snapshots out of primary storage,

adding data blocks with layout meta data, which are defined as either one of both disk layout meta data and file-system volume layout meta data, to the list of valid data blocks,

selecting data blocks with layout meta data for reading into the backup appliance computer BA memory, by running the backup computer program BCP, and analyzing the selected data blocks for detecting attributes such as name, type and extent of file-system volume associated therewith, and

adding, when the analyzed information matches the desired scope of backup data, the valid data blocks associated with the selected file system volume to the list of valid data blocks.

27. The method according to claim 25, including:

creating a database of snapshot history having pertinent information such as snapshot status, name, execution time, associated backup clients and primary storage volumes,

updating snapshot status information as an entry to the snapshot history database when a snapshot is started, when a snapshot is completed, and when a snapshot catalog of snapshots is completed,

whereby a database containing history of all snapshots and their associated snapshot status is maintained.

28. The method according to claim 9, including:

operating the backup appliance computer BA to run a remote file system supporting network file-system protocols such as Common Internet File System (CIFS) and Network File System (NSF), the remote file system being coupled to the network through the backup appliance computer BA'"'"'s network support services, such as a Sockets TCP/IP interface, and

running the backup manager computer program for creating and maintaining a virtual file-system and an associated file-system structure, representing the snapshots in repository,

coupling the virtual file-system to the remote file system operated by the BA, and

perating a computer coupled to the network and running a compatible remote file-system,

whereby the contents of the snapshots residing in repository is accessible through the LAN for operations such as browsing and reading files.

29. The method according to claim 27, including:

operating a computer coupled to the network and running a remote file system supporting protocols compatible with the backup appliance computer BA'"'"'s remote file system, such as Common Internet File System CIFS and Network File System NFS,

operating the computer'"'"'s remote file system for selecting a file or a directory out of a virtual file-system associated with the BA, and for requesting contents of the selected file or directory over the network, running the remote file system in the BA for receiving requests from the computer over the network, for obtaining desired information of the virtual file-system,

running the virtual file-system in the backup manager computer program BMCP for obtaining desired information of the virtual file-system structure, and

running the remote file-system in the BA for returning the desired information in response to requests, via the network to the computer,

whereby a workstation attached to the network is able to navigate the virtual file system structure, to use the contents of a directory to access sub-directories, and to browse and read files from the BA virtual file-system.

30. The method according to claim 27, including:

running a virtual file system on the backup manager computer program BMCP and coupling the virtual file system to the remote file system operated by the backup appliance computer BA,

running the virtual file system for creating an empty file-system structure in BA memory,

running the BMCP for retrieving out of a plurality of entries in a snapshot history database, entries representing snapshots with a status indicating a completed catalog, and

adding virtual directories to the file-system structure in first level below the root directory, the virtual directories representing snapshots retrieved from the snapshot history database,

whereby a first level of virtual file-system structure is created in BA memory.

31. The method according to claim 29, including:

running the remote file system in the computer to select a file or to select a directory, and sending, to the backup appliance computer BA, a request to obtain access to contents of the selected file or to the selected directory, the request containing a full path designation to, respectively, the selected file, or the selected directory,

running the remote file system in the BA for enabling to receive a request to obtain access to contents of a desired file or of a desired directory, the request containing a full path designation leading to the desired file or to the desired directory, and selecting, respectively, the desired file or the desired directory in the virtual file-system,

running the backup manager computer program BMCP to analyze the path leading to a first-level selected directory name,

locating the virtual file-system structure entry matching the first directory name of the full path of the desired file or desired directory, and selecting the matching catalog in repository

running the BMCP to analyze further selected directories and selected file names associated with the full path designation, then locating and selecting entries in the catalog structure matching the selected directory and the selected file names derived from the analysis of the path, and

repeating the analysis and selection operation in the catalog until the whole path is analyzed,

whereby a catalog entry matching a desired directory or a desired file is selected.

32. The method according to claim 30, including:

running the backup manager computer program for;

selecting a catalog entry matching a request emanating from the remote file system running on the workstation,

selecting the snapshot associated with the catalog comprising the selected catalog entry,

retrieving and selecting, out of the selected catalog entry, an attribute representing the node table entry associated with the desired file or directory,

reading the selected node table entry out of selected snapshot in the repository, for analysis and for detection of the data blocks representing the contents of the desired file or directory, and for selecting data blocks for recovery,

reading the selected data blocks out of the selected snapshot in the repository into the backup appliance computer BA memory,

running the remote file system in the BA to send the data blocks representing the contents of the desired file or of the desired directory, from BA memory and over the network, to the remote file system of a computer, and

running the remote file system of a computer for receiving the data blocks over the network and making the data blocks available to the operating system OS or to an application computer program running on the computer,

whereby files and directories are prone to recovery from the repository.

33. The method according to claim 10, wherein writing data blocks into SAN-coupled primary storage includes:

selecting data blocks residing in the backup appliance computer BA memory and comprising contents to be written into primary storage,

running the backup manager computer program on the BA for writing the contents of the selected data blocks, over the SAN, respectively, out of BA memory and into primary storage locations associated with the data blocks,

whereby selected data blocks out of BA memory are written into SAN-coupled primary storage.

34. The method according to claim 10, wherein writing data blocks into backup client-coupled primary storage includes:

selecting data blocks residing in the backup appliance computer BA memory and comprising contents to be written into direct attach storage DAS,

running the backup computer program for sending a command to write the selected data blocks to DAS, from the backup manager computer program to the agent computer program ACP over the network, the command comprising primary storage locations associated with the data blocks, and the contents of the data blocks,

running the ACP and the backup client'"'"'s operating system OS for writing the contents of the selected data blocks from backup client memory to primary storage locations associated with the selected data blocks,

whereby selected data blocks out of BA memory are written into backup client-coupled memory.

35. The method according to claim 10, including:

coupling a workstation to the network,

running the backup user interface management computer program BUI via the workstation by help of an operator, for selecting a repository-residing snapshot comprising requested data, which is defined as source data,

running the BUI to designate a region of primary storage as destination, or target, for the source data to be recovered, and

running the BUI to initiate a fast block-level data recovery operation wherein the FBSRD reads and stores the source data,

whereby a block-level recovery operation is initiated.

36. The method according to claim 35, including:

a. running the backup management computer program BMCP to send a command to the agent computer program ACP, over the network, for disabling backup client access to the designated target in primary storage,

b. running the ACP to disconnect or dismount any backup client file system volume stored on the designated target in primary storage,

c. running the backup computer program BCP to read block information out of the selected snapshot and into the backup appliance computer BA memory, the information comprising locations and contents of data blocks,

d. running the BCP to write block information out of BA memory and into the designated primary storage, the operation comprising writing the contents of the data block to the location associated with the data block,

e. repeating steps c. and d. until copy completion of all data blocks out of a selected snapshot and into designated primary storage,

f. running the BMCP, if desired, to send a command to the ACP over the network, for enabling backup client access to the designated primary storage, and

g. running the ACP to connect or mount any client file system volume coupled to the designated target in primary storage,

whereby an exact copy of a selected snapshot is recovered within a designated target in primary storage.

37. The method according to claim 36, wherein a disk-creation recovery operation is initiated by running the backup user interface management computer program, including:

enabling creation of disk layout meta data on primary storage,

selecting a plurality of file system volumes from a plurality of snapshots as data sources,

designating a target in primary storage as data destination, and

initiating an FBSRD disk-creation recovery operation using the selected sources of data and the designated destination.

38. The method according to claim 37, wherein a designated primary storage is created with a plurality of file-system volumes recovered thereto by running the backup computer program BCP, including:

disabling or dismounting backup client access to the designated primary storage,

creating, within the backup appliance computer BA memory, of a disk-layout meta-data, which matches a chosen disk-creation recovery operation, writing the disk layout meta data into a designated primary storage,

fast-recovering of each snapshot selected out of the repository for storage into an indicated primary storage disk, and

enabling or mounting file-system volumes created on the designated primary storage disk by backup clients.

39. The method according to claim 11, wherein a computer has access via the SAN, to a repository, including:

running a read file system computer program, or RFS, on the computer,

using the RFS to create a representation of the data blocks stored in a selected snapshot in the repository,

running the file system to obtain file system structure and file data to convert file data requests to block requests, and

accepting block read requests from the file system of the computer, and retrieving requested data blocks from the selected snapshot,

whereby snapshot data in file format is recovered from a snapshot in the repository over the SAN.

40. The method according to claim 39, wherein creating a representation of the data blocks includes:

viewing snapshots in repository and selecting a snapshot,

creating a file system volume and designating a file system volume name associated with the selected snapshot,

mounting or attaching the selected snapshot as a file system volume carrying the designated volume name,

whereby the file system of the computer is associated with the file system volume carrying the designated volume name.

41. The method according to claim 40, including:

creating a virtual block device with a name associated with the designated volume and with a reported capacity identical to the capacity of the selected snapshot, and

associating the virtual block device with the file system of the computer.

42. The method according to claim 41, including:

running the file-system of the computer to read the file system structure and the data from the file system volume stored in the virtual block device,

running the file system of the computer to read data blocks containing file meta data and file data from the virtual block device,

running the read file computer program System RFS and the file-system of the computer to read the repository-resident look-up table LUT file associated with the data blocks requested by the file system, into the memory of the computer memory,

running the RFS to retrieve from the LUT the location information of the selected data blocks, the location information comprising a file name in the repository and location in the file, and

running the RFS and the file-system of the computer to read the selected data blocks from the retrieved location in the repository.

43. The method according to claim 39, wherein access control to the repository includes:

intercepting data read requests directed to the repository by the file system, and

accepting the data read requests originating from the repository only if the originator of the file I/O request is the read file computer program System RFS, but returning data indicating invalid file system structure to the file system when an originator of an I/O request is other than the RFS.