Coordinated access to a file system's shared storage using dynamic creation of file access layout

• US 9,727,578 B2
• Filed: 09/28/2012
• Issued: 08/08/2017
• Est. Priority Date: 09/28/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Schmuck, Frank B., Eshel, Marc, Hildebrand, Dean, Wyllie, James C.
• Primary Examiner(s): Strange, Aaron
• Application Number:
  US13/630,684
• Publication Number:
  US 20140095582A1
• Time in Patent Office: 1,775 Days
• Field of Search: 709226
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 16/183 Provision of network file services by network file servers, e.g. by using NFS, CIFS network file access protocols H04L67/1097
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Embodiments of the invention relate to supporting coordinated access to a clustered file system'"'"'s shared disk storage subsystem by using dynamic creation of file access layout for different workloads and access patterns. Active file servers in the cluster have equivalent access to shared data of said file system. In one embodiment, a particular file server in the cluster is configured to coordinate the file system'"'"'s layout. The particular file server, in response to a particular client'"'"'s request for a layout of a file stored by said file system in said subsystem determines a file access layout for a parallel file access protocol to use when servicing an I/O request by said client, marks said layout as active for the client, and sends the layout to the client. The file access layout is determined based on one or more of current file system resources, a specific workload type and a specific byte-range of the file included in the client request, and properties of the file. The layout includes a set of file servers in said cluster and byte ranges of the file, and the byte-range being associated with a particular server among the set.

Claims

1. A method of supporting coordinated file access using NFS (Network File System) protocol for metadata and I/O to a clustered file system'"'"'s shared disk storage subsystem dynamic creation of file access layout for different workloads and access patterns, wherein running file servers in the clustered file system have equivalent access to shared data of the clustered file system, comprising:

in response to a NAS (Network-Attached Storage) protocol client'"'"'s request for a file access layout of a file stored by the clustered file system in the clustered file system'"'"'s shared disk storage subsystem, determining a file access layout identifying hostnames or IP addresses for a parallel file access protocol to use when servicing an I/O request by a client of the clustered file system associated with the request;

wherein, any file server in the clustered file system may receive the client'"'"'s request and the determining is based on a specific workload type and a specific byte-range of the file included in the client request, and the file access layout includes a set of hostnames or IP addresses of file servers in the clustered file system and byte ranges of the file, and the byte-range being associated with a particular server among the set;

marking the file access layout as active for the client;

adding the client to a list indicating what particular clients of the clustered file system are using the file access layout, wherein any client of the clustered file system can share a same file access layout and modify ranges of a file simultaneously; and

sending the file access layout to the client.

2. The method of claim 1, further comprising:

receiving a returned file access layout from the client;

marking the returned file access layout as inactive for the client; and

removing the client from the list of clients that are using the returned file access layout thereby revoking the client'"'"'s access.

3. The method of claim 1, further comprising:

adding a new file server to the clustered file system such that a client'"'"'s request for a file access layout may include the added file server in the determined file access layout; and

wherein, for clients that have mounted the clustered file system, if;

(i) a client has an active file access layout, and (ii) the cached file access layout is not active, invalidating the client'"'"'s file access layout and removing the client from the list of clients using the file access layout such that the next time the client performs I/O the client will retrieve a new file access layout from the clustered file system.

4. The method of claim 1, further comprising:

adding a new file server to the clustered file system such that a client'"'"'s request for a file access layout may include the added file server in the determined file access layout; and

wherein, for clients that have mounted the clustered file system, if a client has an active file access layout and the file access layout is currently being used to perform I/O to the clustered file system, invalidating the client'"'"'s file access layout after the client node has completed the I/O and returned the active file access layout to the clustered file system.

5. The method of claim 4, wherein the new server can include any layout for any file of any number of file access clients without any additional synchronization.

6. The method of claim 1, further comprising:

removing a file server from the clustered file system such that a client'"'"'s request for a file access layout that will no longer include the removed file server in the determined file access layout;

wherein, for clients that have mounted the clustered file system and have an inactive file access layout, invalidating the file access layout on the client such that the client must request a new file access layout that does not include the server that was removed or determining a new file access layout for the client that does not include the server that was removed; and

replacing the file access layout on the client; and

updating the list of clients that are using the new file access layout.

7. The method of claim 1, further comprising:

in response to a file server'"'"'s failure, choosing one of the remaining active file servers and having it service a client'"'"'s I/O request(s);

in response to the clustered file system detecting the file server'"'"'s failure, the clustered file system determines which clients have mounted the clustered file system and have an active file access layout; and

if the file access layout is active, the clustered file system either;

i) invalidating the client'"'"'s file access layout after the client node has completed the I/O such that the client must then request a new file access layout, ii) removing the file server from the client'"'"'s file access layout;

or iii) calculating a new file access layout and substituting it for the existing one and recording that the client is using the new layout;

or

if the file access layout is not active, the clustered file system either;

i) invalidating the client'"'"'s file access layout such that the client must then request a new file access layout;

ii) removing the file server from the client'"'"'s file access layout;

or iii) calculating a new file access layout and substituting it for the existing one and recording that the client is using the new layout.

8. The method of claim 1, wherein the determination of the makeup of the file access layout and if a file access layout should even be granted to a client are dependent upon the resources, wherein the resources are selected from the group consisting of:

(i) available file servers, (ii) available network bandwidth, (iii) available storage bandwidth, (iv) open file system connections, (v) processing unit resources, (vi) number of clients currently accessing the file, and (vii) number and type of existing active file access layouts for the file; and

wherein the client'"'"'s workload comprises its pattern of issuing NFS protocol requests.

9. The method of claim 8, further comprising:

revoking and/or updating client file access layouts in response to one or more of a change in the resources.

10. The method of claim 1, wherein the resources are determined by monitoring running file servers of the clustered file system and tracking each client'"'"'s workload across the clustered file system over time.

11. The method of claim 10, wherein the clustered file system, to perform the monitoring, runs an I/O performance optimization algorithm service to determine an aggregate number of file servers and byte ranges for each server to serve requests to the file based upon the client'"'"'s workload over time.

12. The method of claim 1, further comprising:

upon receiving a write request from a client, validating that byte-range to be written complies with the file access layout that was sent to the client; and

writing the data to the clustered file system'"'"'s shared disk storage subsystem and coordinating with other file system requests using a distributed lock protocol.

13. The method of claim 1, further comprising:

upon receiving a read request from a client of the clustered file system, validating that byte-range to be read complies with the file access layout that was sent the client; and

reading the data from the clustered file system'"'"'s shared disk storage subsystem, coordinating with other file system requests using a distributed lock protocol, and sending that data to the client.

14. The method of claim 1, wherein each file server in the clustered file system shares a common pool of available storage capacity such that the determining is based upon the common pool of available storage capacity.

15. The method of claim 1, wherein the file access layout may be the same for other clients of the clustered file system or comprise a different set of IP addresses and associated byte-ranges.

16. The method of claim 1, wherein the NFS protocol is pNFS (Parallel NFS).

17. The method of claim 1, further comprising:

upon receiving a write request from a client, if the byte-ranges to be written do not comply with the file access layout that was sent to the client;

then returning an error to the client, or revoking the file access layout on the client and force the client to request a new layout.

18. A system of supporting coordinated file access using NFS (Network File System) protocol for metadata and I/O to a clustered file system'"'"'s shared disk storage subsystem using dynamic creation of file access layout for different workloads and access patterns, comprising:

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

a plurality of file servers in a cluster, each of the file servers that are running having equivalent access to the shared data of the clustered file system, wherein the system, in response to a NAS (Network-Attached Storage) protocol client'"'"'s request for a file access layout of a file stored by clustered file system in the clustered file system'"'"'s shared disk storage subsystem, determines a file access layout identifying hostnames or IP addresses for a parallel file access protocol to use when servicing an I/O request by a client of the clustered file system that made the client'"'"'s request;

wherein, any file server in the clustered file system may receive the client'"'"'s request and the determine is based on a specific workload type and a specific byte-range of the file included in the client request, and the file access layout includes a set of hostnames or IP addresses of the file servers and byte ranges of the file, and the byte-range being associated with a particular server among the set; and

wherein the system, marks the file access layout as active for the client, adds the client to a list indicating what particular clients of the clustered file system are using the file access layout, wherein any client of the clustered file system can share a same file access layout and modify ranges of a file simultaneously, and sends the file access layout to the client.

19. A computer program product of supporting file access using NFS (Network File System) protocol for metadata and I/O to a clustered file system'"'"'s shared disk storage subsystem dynamic creation of file access layout for different workloads and access patterns, wherein running file servers in the clustered file system have equivalent access to shared data of the clustered file system, the computer program product comprising a computer readable storage medium having computer usable program code embodied therewith, the computer usable program code being executable to:

in response to a NAS (Network-Attached Storage) protocol client'"'"'s request for a file access layout of a file stored by the clustered file system in the clustered file system'"'"'s shared disk storage subsystem, determine a file access layout identifying hostnames or IP addresses for a parallel file access protocol to use when servicing an I/O request by a client of the clustered file system that made the client'"'"'s request;

wherein, any file server in the clustered file system may receive the client'"'"'s request and the determine is based on a specific workload type and a specific byte-range of the file included in the client request, and the file access layout includes a set of hostnames IP addresses file servers in the clustered file system and byte ranges of the file, and the byte-range being associated with a particular server among the set;

mark the file access layout as active for the client;

adds the client to a list indicating what particular clients of the clustered file system are using the file access layout, wherein any client of the clustered file system can share a same file access layout and modify ranges of a file simultaneously; and

send the file access layout to the client.

20. The computer program product of claim 19, the computer usable program code being further executable to:

receive a returned file access layout from the client;

mark the returned file access layout as inactive for the client; and

remove the client from the list of clients that are using the returned file access layout thereby revoking the client'"'"'s access.

21. The computer program product of claim 19 the computer usable program code being further executable to:

add a new file server to the clustered file system such that a client'"'"'s request for a file access layout may include the added file server in the determined file access layout; and

wherein, for clients that have mounted the clustered file system, if;

(i) a client has a cached file access layout, and (ii) the cached file access layout is not active, invalidate the client'"'"'s file access layout and remove the client from the list of clients using the file access layout such that the next time the client performs I/O the client will retrieve a new file access layout from the clustered file system.

22. The computer program product of claim 19, the computer usable program code being further executable to:

add a new file server to the clustered file system such that a client'"'"'s request for a file access layout may include the added file server in the determined file access layout; and

wherein, for clients that have mounted the clustered file system, if a client has an active file access layout and the file access layout is currently being used to perform I/O to the clustered file system invalidate the client'"'"'s file access layout after the client node has completed the I/O and returned the active file access layout to the clustered file system.

23. The computer program product of claim 19, the computer usable program code being further executable to:

remove a file server from the clustered file system such that a client'"'"'s request for a file access layout that will no longer include the removed file server in the determined file access layout;

wherein, for clients that have mounted the clustered file system and have an inactive file access layout, invalidate the file access layout on the client such that the client must request a new file access layout that does not include the server that was removed or determine a new file access layout for the client that does not include the server that was removed; and

replace the file access layout on the client; and

update the list of clients that are using the new file access layout.

24. The computer program product of claim 19, the computer usable program code being further executable to:

in response to a file server'"'"'s failure, choosing one of the remaining active file servers and having it service a client'"'"'s I/O request(s);

in response to the clustered file system detecting the file server'"'"'s failure, the clustered file system determines which clients have mounted the clustered file system and have an active file access layout; and

if the file access layout is being used to perform I/O to the clustered file system, the clustered file system;

i) invalidate the client'"'"'s file access layout after the client node has completed the I/O such that the client must then request a new file access layout, ii) remove the file server from the client'"'"'s file access layout;

or iii) calculate a new file access layout and substituting it for the existing one and recording that the client is using the new layout;

or

if the file access layout is not being used to perform I/O to the clustered file system, the clustered file system either;

i) invalidate the client'"'"'s file access layout such that the client must then request a new file access layout;

ii) remove the file server from the client'"'"'s file access layout;

or iii) calculate a new file access layout and substituting it for the existing one and recording that the client is using the new layout.

25. The computer program product of claim 19, the computer usable program code being further executable to:

upon receiving a write request from a client, validate that the byte-range to be written complies with the file access layout that was sent to the client; and

write the data to the clustered file system'"'"'s shared disk storage subsystem and coordinating with other file system requests using a distributed lock protocol.

26. The computer program product of claim 19, the computer usable program code being further executable to:

upon receiving a read request from a client of the clustered file system, validating that the byte-range to be read complies with the file access layout that was sent the client; and

read the data from the clustered file system'"'"'s shared disk storage subsystem, coordinate with other file system requests using a distributed lock protocol and send that data to the client.