Supporting coordinated access to a file system's shared storage using automatic alignment of a parallel file access protocol and metadata management

• US 8,799,335 B2
• Filed: 09/28/2012
• Issued: 08/05/2014
• Est. Priority Date: 09/28/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Eshel, Marc, Hildebrand, Dean, Schmuck, Frank B., Wyllie, James C.
• Primary Examiner(s): NGUYEN, CAM LINH T
• Application Number:
  US13/630,861
• Publication Number:
  US 20140095459A1
• Time in Patent Office: 676 Days
• Field of Search: 707/703, 707/764, 707/822, 707/825, 707/826
• US Class Current: 707/826
• CPC Class Codes: G06F 16/119 Details of migration of file systems migration mechanisms in storage systems G06F3/0647
• CPC Class Codes: G06F 16/1824 implemented using Network-attached Storage [NAS] architecture distributed or networked storage systems G06F3/067; protocols for distributed storage of data in a network H04L67/1097
• CPC Class Codes: G06F 16/2308 Concurrency control transaction processing G06F9/466

Abstract

Embodiments of the invention relate to supporting coordinated access to a file system'"'"'s shared disk storage subsystem using an automatic alignment of a parallel file access protocol and metadata management on a per object basis. A file server (among servers in the cluster) receives a client'"'"'s request to access a file stored by the file system in the storage subsystem using a parallel file access protocol. If the file system has selected a file server to serve as a metadata manager of said file, the file server uses the selected file server to coordinate access to the file'"'"'s metadata. If the file system has not selected a file server to serve as the metadata manager of the file, the file system selects the file server that received said request as the metadata manager for said file until a file close request operation is received.

Claims

1. A method of supporting coordinated access to a file system'"'"'s shared disk storage subsystem using an automatic alignment of a parallel file access protocol and metadata management on a per object basis of said storage subsystem, wherein active file servers in a cluster have equivalent access to shared data of said file system, the method comprising:

in response to a client'"'"'s request to access a file stored by said file system in said storage system using a parallel file access protocol;

if said file system has selected a metadata manager of said file, using said metadata manager to coordinate access to said file'"'"'s metadata,

if said file system has not selected a metadata manager of said file, selecting the file server that received said request as the metadata manager for said file until a file close request operation is received,

performing a file open request operation for said file to the selected metadata manager, said performing being done by a file server that received said request;

opening said file and creating a file handle of said file,

encoding a file server ID, of the file server selected as the metadata manager of said file, in the file handle such that said file system'"'"'s file servers are enabled to identify which file server is said file'"'"'s metadata manager,

returning the file handle to said client for performing I/O to said file;

if the file server that received said client'"'"'s request is not the metadata manager and a metadata manager already exists for the file but no clients have said file open, migrating the metadata management from the file server, previously selected as the metadata manager of said file, to a file server in the cluster that at least one client has said file open.

2. The method of claim 1, wherein having a metadata manager selected thereby allows said client to access different metadata managers for different files, such that metadata management can be distributed across the file system'"'"'s file servers.

3. The method of claim 1, further comprising:

receiving a metadata access request for said file'"'"'s metadata from said client; and

coordinating the access for said file'"'"'s metadata through said metadata manager.

4. The method of claim 1, further comprising:

in response said client'"'"'s I/O request to said file, the file server receiving the I/O request;

retrieving said file server ID of the metadata manager, by decoding said file handle,

retrieving any required metadata from the file server, selected as the metadata manager of said file, to determine if said client has permission to perform I/O to said file, and

if said client has permission to perform I/O to said file, caching the retrieved metadata and executing the requested I/O request.

5. The method of claim 4, wherein said client does not have said permission, if either:

(i) access control fails for a user of said client, or (ii) said client attempts to perform a type of operation not in accordance with file system security policies.

6. The method of claim 1, further comprising:

in response said client'"'"'s request to close said file, the file server that received said request, uses an inter-node communication mechanism to remove cached metadata on said file system'"'"'s file servers.

7. The method claim 1, further comprising:

in response said client'"'"'s request for a file access layout of said file, the file server receiving the I/O request, returns a file access layout of said file to said client for said protocol to use to service the I/O request; and

wherein said layout includes a set of file servers in said cluster and byte ranges of said file, the byte-range being associated with a particular server among said file system'"'"'s file servers.

8. A file system of supporting coordinated access to the file system'"'"'s shared disk storage subsystem using an automatic alignment of a parallel file access protocol and metadata management on a per object basis of said storage subsystem, said file system comprising:

a shared disk storage subsystem that stores said file system'"'"'s shared data;

a plurality of file servers in a cluster, each of the file servers that are active having equivalent access to the shared data, wherein a particular file server (among said servers) is configured to manage metadata for a particular file; and

wherein a file server, among said servers, receives a client'"'"'s request to access a file stored by said file system in said storage subsystem using a parallel file access protocol;

if said file system has selected a file server to serve as a metadata manager of said file, using the selected file server to coordinate access to said file'"'"'s metadata,

if said file system has not selected a file server to serve as the metadata manager of said file, selecting the file server that received said request as the metadata manager for said file until a file close request operation is received, and

said server that received the client'"'"'s request;

(i) performing a file open request operation for said file, to the file server selected to serve as the metadata manager, (ii) opens said file and creates a file handle of said file, (iii) encodes a file server ID, of the file server selected as the metadata manager of said file, in the file handle such that said file system'"'"'s file servers may identify which file server is said file'"'"'s metadata manager, and (iv) returns the file handle to said client for performing I/O to said file;

wherein if the file server that received said client'"'"'s request is not the metadata manager and a metadata manager already exists for the file but no clients have said file open, the file system migrates the metadata management from the file server, previously selected as the metadata manager of said file, to a file server in the cluster that at least one client has said file open.

9. The file system of claim 8, wherein if the file system receives a metadata access request for said file'"'"'s metadata from said client, said file system coordinates the access for said file'"'"'s metadata through the file server selected as the metadata manager.

10. The file system of claim 8, wherein in response said client'"'"'s I/O request to said file, the file server that received the I/O request:

retrieves said file server ID of the metadata manager, by decoding said file handle,

retrieves any required metadata from the file server, selected as the metadata manager of said file, to determine if said client has permission to perform I/O to said file, caches the retrieved metadata, and

executes the requested I/O request.

11. The file system of claim 8, further comprising:

in response said client'"'"'s request to close said file, the file server that received said request, uses an inter-node communication mechanism to remove cached metadata on said file system'"'"'s file servers.

12. The file system of claim 8, further comprising:

in response said client'"'"'s request for a file access layout of said file, the file server receiving the I/O request, returns a file access layout of said file to said client for said protocol to use to service the I/O request; and

wherein said layout includes a set of file servers in said cluster and byte ranges of said file, the byte-range being associated with a particular server among said file system'"'"'s file servers.

13. A computer program product of supporting coordinated access to a file system'"'"'s shared disk storage subsystem using an automatic alignment of a parallel file access protocol and metadata management on a per object basis of said storage subsystem, wherein active file servers in a cluster have equivalent access to shared data of said file system, said program product comprising a non-transitory computer readable storage medium having computer usable program code embodied therewith, said program code being readable/executable by a file system to:

in response to a client'"'"'s request to access a file stored by said file system in said storage system using a parallel file access protocol;

if said file system has selected a metadata manager of said file, use said metadata manager to coordinate access to said file'"'"'s metadata;

if said file system has not selected a metadata manager of said file, select the file server that received said request as the metadata manager for said file until a file close request operation is received;

perform a file open request operation for said file to the selected metadata manager, said performing being done by a file server that received said request;

open said file and create a file handle of said file;

encode a file server ID, of the file server selected as the metadata manager of said file, in the file handle such that said file system'"'"'s file servers may identify which file server is said file'"'"'s metadata manager; and

return the file handle to said client for performing I/O to said file.

if a file system file server that received said client'"'"'s request is not the metadata manager and a metadata manager already exists for the file but no clients have said file open, migrate the metadata management from the file server, previously selected as the metadata manager of said file, to a file server in the cluster that at least one client has said file open.

14. The computer program product of claim 13, said program code further configured to:

receive a metadata access request for said file'"'"'s metadata from said client; and

coordinate the access for said file'"'"'s metadata through said metadata manager.

15. The computer program product of claim 13, said program code further configured to:

in response said client'"'"'s I/O request to said file, the file server receiving the I/O request;

retrieve said file server ID of the metadata manager, by decoding said file handle,

retrieve any required metadata from the file server, selected as the metadata manager of said file, to determine if said client has permission to perform I/O to said file,

cache the retrieved metadata, and

execute the requested I/O request.

16. The computer program product of claim 13, said program code further configured to:

in response said client'"'"'s request to close said file, the file server that received said request, uses an inter-node communication mechanism to remove cached metadata on said file system'"'"'s file servers.

17. The computer program product of claim 13, said program code further configured to:

in response said client'"'"'s request for a file access layout of said file, the file server receiving the I/O request, returns a file access layout of said file to said client for said protocol to use to service the I/O request; and

wherein said layout includes a set of file servers in said cluster and byte ranges of said file, the byte-range being associated with a particular server among said file system'"'"'s file servers.