Patent Application
20110184907
A typical operating system includes a file system. The file system provides a mechanism for the storage and retrieval of files and a hierarchical directory structure for the naming of multiple files. More specifically, the file system stores information (i.e., data) provided by the client (i.e., a local or remote process) and information describing the characteristics of the data (i.e., meta-data). The file system also provides extensive programming interfaces to enable the creation and deletion of files, reading and writing of files, performing seeks within a file, creating and deleting directories, managing directory contents, etc. In addition, the file system also provides management interfaces to create and delete file systems. File systems are typically controlled and restricted by operating system parameters. For example, most operating systems limit the maximum number of file names that can be handled within their file system. Some operating systems also limit the size of files that can be managed under a file system.
An application, which may reside on the local system (i.e., computer) or may be located on a remote system, uses files as an abstraction to address data. Conventionally, this data is stored on a storage device, such as a disk. To access a file, the operating system (via the file system) typically provides file manipulation interfaces to open, close, read, and write the data within each file.
In some instances, the files need to be migrated from the current file to a new file system. In such instances, the data (and meta-data) currently stored in the current file system must be moved to a new file system. Such a migration is typically achieved by initially taking the current file system offline (i.e., preventing clients from reading or writing to the current file system). Once offline, various techniques may be used to transfer each directory and file in the current file system to the new file system. Depending on the amount of data in the current file system, the migration may take a significant period of time, during which the data in the current file system and the new file system are inaccessible to the clients.
In general, in one aspect, the invention relates to a computer readable medium comprising software instructions, which when executed by a processor, perform a method, the method including receiving a first file system (FS) operation request for a target FS from a client, making a first determination that migration for a source FS is not complete, making a second determination that the first FS operation request specifies a directory and that a directory level attribute for the directory on the target FS specifies that the directory on the target FS is un-migrated, in response to the first and second determination: obtaining meta-data for content in the directory from the source FS, creating, using the meta-data for content in the directory, a directory entry for a file in the directory on the target FS, wherein the directory entry for the file is associated with a file level attribute that specifies the file is un-migrated, adding unique identification (UID) for the file to a pending list, adding unique identification (UID) for the directory to a removed list, servicing, after the creating, the first FS operation request using target FS.
In general, in one aspect, the invention relates to a computer system that includes a processor and a virtual file system layer (VFS) operatively connected to a source file system (FS) and a target FS. The VFS, when executed by the processor, performs a method. The method includes receiving a first file system (FS) operation request for a target FS from a client, making a first determination that migration for a source FS is not complete, making a second determination that the first FS operation request specifies a directory and that a directory level attribute for the directory on the target FS specifies that the directory on the target FS is un-migrated, in response to the first and second determination: obtaining meta-data for content in the directory from the source FS, creating, using the meta-data for content in the directory, a directory entry for a file in the directory on the target FS, wherein the directory entry for the file is associated with a file level attribute that specifies the file is un-migrated, adding unique identification (UID) for the file to a pending list, adding unique identification (UID) for the directory to a removed list, updating, after the creating, the directory level attribute to indicate that the directory on the target FS is migrated, and servicing, after the creating, the first FS operation request using target FS.
Other aspects of the invention will be apparent from the following description and the appended claims.
Specific embodiments of the invention will now be described in detail with reference to the accompanying figures. Like elements in the various figures are denoted by like reference numerals for consistency.
In the following detailed description of embodiments of the invention, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to one of ordinary skill in the art that the invention may be practiced without these specific details.
In general, the invention relates to migration of files and directories from a source file system to a target file system. More specifically, the invention enables file system migration while allowing clients to continue to access the files and directories during the migration. Further, embodiments of the invention include a mechanism to guarantee traversal of the source FS during the migration of the source FS to the target FS.
For purposes of this invention, each file system represents directories, sub-directories, and files as directory entries. Each directory entry includes the name of the corresponding entity, i.e., directory, sub-directory, or file, and is associated with the corresponding meta-data and data (if applicable). Accordingly, all references to directories, sub-directories, and files are intended to include the corresponding directory entry, meta-data, and data (if applicable).
In one embodiment of the invention, the operating system (100) is configured to interface with clients (not shown) and with the file systems (106A, 106N) via the system call interface (102) and the VFS layer (104). In one embodiment of the invention, a client is any remote or local process (including operating system processes) that includes functionality to issue a file system (FS) operation request. In one embodiment of the invention, FS operation requests include, but are not limited to, read( ), write( ), open( ), close( ), mkdir( ), rmdir( ), rename( ), sync( ), unmount( ), and mount( ). Examples of operating systems (100) include, but are not limited to, MAC OS®, Solaris™, Linux, Microsoft® Windows®. (MAC OS is a registered trademark of Apple, Inc; Microsoft and Windows are registered trademarks of the Microsoft Corporation; Solaris is a trademark of Sun Microsystems, Inc; Linux is a registered trademark on Linus Torvalds.)
In one embodiment of the invention, the system call interface (102) is configured to receive FS operation requests from the operating system (100), forward the FS operation requests to the VFS layer (104), receive responses to the FS operation requests, and forward to the corresponding responses to the operating system. Those skilled in the art will appreciate that while the system call interface (102) is represented as a distinct component from the operating system (100), the system call interface (102) may be located within the operating system (100).
In one embodiment of the invention, the VFS layer (104) is an abstraction layer interposed between file systems (106A, 106N) and the operating system. In one embodiment of the invention, the purpose of the VFS layer (104) is to allow the operating system (100) to access different types of file systems (106A, 106N) in a uniform way. For example, the VFS layer (104) may be used to access local and networked file systems transparently without the operating system (100) being aware of the difference. Further, the VFS layer (104) enables the operating system (100) to access different file systems (e.g., ZFS, Network File System (NFS), Unix File System (UFS), New Technology File System (NTFS), Hierarchical File System (HFS), etc.) without requiring the operating system (100) to be aware of the type of file system it is accessing.
In one embodiment of the invention, each file system (106A, 106N) includes a method for storing and organizing files (including the corresponding file meta-data and file data). Further, each file system (106A, 106N) may include functionality to associate meta-data with directories. The meta-data (associated with directories and files) may include regular attributes and extended attributes.
In one embodiment of the invention, regular attributes are defined and interpreted by the file system, examples of meta-data stored in regular attributes may include but are not limited to access permissions for the file or directory, date/time file was created and/or modified. In one embodiment of the invention, extended attributes may be used to associate meta-data with files and/or directories; however, the meta-data stored in the extended attributes is not defined or interpreted by the file system. In one embodiment of the invention, the root of a target file system (FS) (discussed below), directories in the target FS, and files in the target FS each include at least one extended attribute used to indicate whether the root, directory or file has been migrated. In one embodiment of the invention, the extended attributes are interpreted by the VFS layer (104).
In one embodiment of the invention, each file system (106A, 106N) is configured to store meta-data and data on one or more storage medium (108A, 108N). Each storage medium (108A, 108N) corresponds to a physical storage device configured to store data and meta-data. Examples of the storage mediums (108A, 108N) include, but are not limited to, magnetic media (e.g., disk drives, tape drives), solid-state drives (e.g., NAND flash devices, NOR flash devices), optical media (e.g., compact disks (CDs), digital versatile disks (DVDs), Blu-ray® Disks, etc.), or any combination thereof. (Blu-ray is a trademark of the Blu-ray Disk Association).
Those skilled in the art will appreciate that while
In on embodiment of the invention, one or more file systems are associated with a persistent pending list (not shown). In one embodiment of the invention, the persistent pending list is initially empty and is located in persistent memory (e.g., on a persistent storage medium). In one embodiment of the invention, the persistent pending list is located in the root of the target FS. Those skilled in the art will appreciate that the persistent pending list may be located within other locations within (or accessible to) the target FS. In one embodiment of the invention, the persistent pending list is configured to store unique identifications (UIDs) for files and directories (or their corresponding directory entries) encountered during the migration of the source FS. The UID may be any value (numeric, alpha, or alpha-numeric) that may be used to uniquely identify a file or directory. Those skilled in the art will appreciate that persistent pending list may be implemented using a list data structure or any other data structure with functionality to store UIDs (or equivalent data).
In one embodiment of the invention, the computer system(s) upon which the operating system includes volatile memory (e.g., in Random Access Memory or any other non-persistent memory). Further, volatile memory is configured to store a removed list, which is configured to store UIDs (or equivalent data).
In Step 200, clients are disconnected from the source FS. Those skilled in the art will appreciate that Step 200 may include both disconnecting clients currently accessing the source FS and denying new FS operation requests. In one embodiment of the invention, clients currently waiting for a response from a previously submitted FS operation request are permitted to remain connected to the source FS until the response is received, after which they are disconnected from the source FS.
In Step 202, the source FS is set to read-only. In Step 204, the target FS is created. Those skilled in the art will appreciate that creating the target FS may be performed using any known method(s). Further, once the target FS is created, the target FS may be accessed via the VFS layer. In Step 206, the migration attribute of the root of the target FS is set to “un-migrated.” In one embodiment of the invention, the migration attribute is an extended attribute. In Step 208, the target system upon which the target FS is located is granted access, optionally, full access, to the source FS. Those skilled in the art will appreciate that the target system may not require full access of the source FS in order to perform the steps in
In Step 212, the target root UID (i.e., the UID of the root of the target FS) is added to the persistent pending list. In Step 214, a removed listed is created. In Step 216, clients initially requesting files from the source FS are redirected to the target FS. Said another way, FS operation requests for the source FS are now directed to the target FS.
In Step 300, a FS operation request is received from a client (via the VFS). In Step 302, a determination is made about whether the FS operation is a read operation. If the FS operation is a read operation, the process proceeds to Step 304. If the FS operation is not a read operation (e.g., the operation is a write operation), the process proceeds to Step 324.
In one embodiment of the invention, if the FS operation request includes a partial write of file data (i.e., one portion of the file data is to be overwritten), then the corresponding directory entry (including file metadata and file data) is migrated (see e.g., ST 330-ST336) prior to proceeding to Step 324.
In Step 304, a determination is made about whether the migration of the source FS is complete. The migration of the source FS is complete when all data and meta-data from the source FS has been copied (or otherwise transmitted) to the target FS. If the migration is not complete, the process proceeds to Step 306. If the migration is complete, the process proceeds to Step 324.
In Step 306, a determination is made about whether the FS operation request is a read request for a directory. If the FS operation request is a read request for a directory, then the process proceeds to Step 308. If the FS operation request is not a read request for a directory (i.e., the FS operation request is for a file), then the process proceeds to Step 326. In Step 308, the directory entry corresponding the requested directory in the target FS is located. In one embodiment of the invention, the directory entry includes at least meta-data associated with the directory and a migration attribute (which may be marked or unmarked, depending on whether the directory has been migrated).
In Step 310, a determination is made about whether the migration attribute in the directory entry is marked (i.e., has the directory been migrated). If the migration attribute is marked, then the directory has not been migrated and the process proceeds to Step 312. If the migration attribute is un-marked, then the directory has been migrated and the process proceeds to Step 322. In one embodiment of the invention, a directory is considered migrated when there is a corresponding directory entry for each file and sub-directory in the directory on the target FS for each file and sub-directory in the corresponding directory on the source FS. Those skilled in the art will appreciate that the directory entries in the directory on the target FS do not need to include corresponding file data in order for the directory to be migrated.
In Step 312, the meta-data for content in the directory (e.g., file meta-data and/or sub-directory meta-data) is obtained from the source FS. In Step 314, a UID for each file and sub-directory (as identified from the meta-data obtained in Step 312) is added to the persistent pending list. In Step 316, a marked directory entry is created in the directory on the target FS for each file and sub-directory using the meta-data obtained in Step 312. In one embodiment of the invention, a marked directory entry corresponds to a directory entry for the file with a marked migration attribute or directory entry for a sub-directory with a marked migration attribute. Further, in one embodiment of the invention, the path to each sub-directory and/or file identified in the meta-data obtain in Step 312 is stored in an extended attribute for the corresponding directory entry. In Step 318, the migration attribute for the directory entry is unmarked. In Step 320, the UID for the directory is added to the in-memory pending list.
In Step 322, periodically, the persistent pending list is updated using the UIDs from the in-memory removed list. In one embodiment Step 322 occurs once per minute; however, Step 322 may occur more or less frequently depending on the implementation of the invention. In Step 324, the FS operation request is processed using the target FS.
In Step 326, the directory entry (DE) for the file corresponding to the requested file in the target FS is located. In one embodiment of the invention, the directory entry for the file includes at least meta-data associated with the file and a migration attribute (which may be marked or unmarked, depending on whether the file has been migrated).
In Step 328, a determination is made about whether the migration attribute in the directory entry for the file is marked (i.e., has the file been migrated). If the migration attribute is marked, then the file has not been migrated and the process proceeds to Step 330. If the migration attribute is un-marked, then the file has been migrated and the process proceeds to Step 324. In one embodiment of the invention, a file is considered migrated when the file data for the file has been obtained from the source FS. In Step 330, the file data for the file is obtained from the source FS. In one embodiment of the invention, the file data is obtained using the path stored in the extended attribute for the corresponding directory entry for the file. In Step 332, the file data is stored in the target FS. In Step 334, the migration attribute for the directory entry for the file is unmarked. In Step 336, the UID for the file is added to the in-memory pending list.
In one embodiment of the invention,
In another embodiment of the invention, the method shown in
In one embodiment of the invention, the background process(es) asynchronously performs a depth-first traversal of the file system hierarchy. Those skilled the art will appreciate that other traversal schemes may be used. Each directory and file encountered during the traversal triggers a migration (i.e., the copying of meta-data and data (if applicable) to the target FS). Further, each directory or file that is encountered during the traversal is added to the persistent pending list and, once migrated, to the in-memory removed list. As described above, the persistent pending list is periodically updated with information on the in-memory removed list.
Once the normal asynchronous migration is complete, the pending list will be empty if all directories and files have been successful migrated. To guarantee traversal of the file system, at the end of the hierarchical traversal, any remaining entries on the persistent pending list may be processed individually and explicitly migrated. If additional files or directories are discovered during this process, they are appended to the persistent pending list and subsequently processed individually. Once the pending list is empty, it is guaranteed that every file and directory from the source file system has been migrated to the target file system.
The following examples are provided to illustrate various aspects of the invention and are not intended to limit the scope of the invention.
In one embodiment of the invention, the client(s) (400) cease to send FS operation requests to the source FS (404) and instead re-direct (or re-issue) all FS operation requests to the target FS (402). The client(s) (400) are unaware that the target FS (402) may not (at the time of the FS operation request) include a copy of the file, which is the target of the FS operation request.
As shown in
With respect to migrated directory entries, once the directory entries have been migrated, the client(s) (400) interact with the migrated directory entries in the same manner as the client(s) (400) interact with the new directory entries.
Referring to
In addition, the persistent pending list is populated with Sub-Dir-B UID, File A UID, and File B UID signifying that the aforementioned sub-directory and files have been accessed as part of the migration process but has not been migrated. In one embodiment of the invention, the persistent pending list is updated by appending the new UIDs to the end of the list. Further, the in-memory removed list is updated to include the Dir-A UID signifying that Directory A has been migrated. At this stage, the Directory A UID is present in both the persistent pending list and the in-memory removed list as the persistent pending list has not been updated using the in-memory removed list (i.e., ST 322 has not been performed). However, the Target Root UID is not present in either the persistent pending list or the in-memory removed list as ST 322 was performed prior to the FS operation request for Directory A.
In addition, in-memory list is updated to include File-A UID signifying that File A has been migrated. At this stage, Directory A UID and File-A UID are present in both the persistent pending list and the in-memory removed list as the persistent pending list has not been updated using the in-memory removed list (i.e., ST 322 has not been performed).
In addition, persistent pending list is updated with the in-memory removed list from
Embodiments of the invention provide a mechanism to ensure that each of the files and directories in the source FS is migrated to the target FS. Further, embodiments of the invention maintain two separate lists (i.e., the persistent pending list and the in-memory removed list) in order to guarantee successful migration. Specifically, in one or more embodiments of the invention, the use of the two lists results in minimal impact on performance of the migration as the lists independently track files and directories that have been accessed (but not yet migrated), and files and directories that have been migrated. Further, should computer system upon with which the target FS is connected to fail, the persistent pending list will continue to track (even during a power loss) the files and directories that have been accessed but not yet migrated; thereby preserving the information necessary to ensure a complete and successful migration of the source FS.
Embodiments of the invention may be implemented on virtually any type of computer regardless of the platform being used. For example, as shown in
Further, the computer system (600) may include output means, such as a monitor (612) (e.g., a liquid crystal display (LCD), a plasma display, or cathode ray tube (CRT) monitor). The computer system (600) may be connected to a network (614) (e.g., a local area network (LAN), a wide area network (WAN) such as the Internet, or any other type of network) via a network interface connection (not shown). Those skilled in the art will appreciate that many different types of computer systems exist, and the aforementioned input and output means may take other forms. Generally speaking, the computer system (600) includes at least the minimal processing, input, and/or output means necessary to practice embodiments of the invention.
Further, those skilled in the art will appreciate that one or more elements of the aforementioned computer system (600) may be located at a remote location and connected to the other elements over a network. Further, embodiments of the invention may be implemented on a distributed system having a plurality of nodes, where each portion of the invention (e.g., storage devices, operating system, etc.) may be located on a different node within the distributed system. In one embodiment of the invention, the node corresponds to a computer system. Alternatively, the node may correspond to a processor with associated physical memory. The node may alternatively correspond to a processor or micro-core of a processor with shared memory and/or resources. Further, software instructions to perform embodiments of the invention may be stored, temporarily or permanently, on a computer readable medium, such as a compact disc (CD), a diskette, a tape, memory, or any other computer readable storage device.
While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this disclosure, will appreciate that other embodiments can be devised which do not depart from the scope of the invention as disclosed herein. Accordingly, the scope of the invention should be limited only by the attached claims.
