Clustered file service

• US 9,652,469 B2
• Filed: 06/04/2011
• Issued: 05/16/2017
• Est. Priority Date: 06/04/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Kuznetsov, Vyacheslav, D'Amato, Andrea, Petter, Vladimir, Warwick, Alan, Aloysius, Henry
• Primary Examiner(s): Spieler, William
• Application Number:
  US13/153,416
• Publication Number:
  US 20120311003A1
• Time in Patent Office: 2,173 Days
• Field of Search: 707827
• CPC Class Codes: G06F 16/176 Support for shared access to files; File sharing support
• CPC Class Codes: G06F 16/178 Techniques for file synchronisation in 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

Abstract

A cluster based file service may operate on a cluster of two or more independent devices that have access to a common data storage. The file service may have a namespace definition with each device in the cluster, but may be modified by any device operating the file service. Each instance of the file service may identify and capture a command that changes the namespace structure and cause the change to be propagated to the other members of cluster. If one of the devices in the cluster does not successfully perform an update to the namespace structure, that device may be brought offline. The cluster based file service may permit adding or removing devices from the cluster while the file service is operating, and may provide a high throughput and high availability file service.

Claims

1. A system comprising:

a plurality of devices, each of said devices having a file service operable on each of said plurality of devices;

a data store comprising files, said data store being accessible to each of said plurality of devices;

a namespace definition defining an organization for said files into a share, said namespace definition being stored in a cluster database on a master node accessible to each of said plurality of devices, said share being made available to client devices, the namespace definition, in addition to defining the organization, also defines which files are to be included within the share;

said file service that identifies changes to said namespace definition and updates said namespace definition on said cluster database on said master node to an updated namespace definition on said cluster database on the master node, said file service further updates a locally cached version of said namespace definition on at least one of the plurality of devices.

2. The system of claim 1 further comprising:

a load balancer that receives requests from said client devices and, for each of said requests, determining one of said plurality of devices to process a request.

3. The system of claim 2 wherein the load balancer includes a common cluster name that the client devices may use to address the cluster database.

4. The system of claim 2 wherein the load balancer is configured to determine that one of the plurality of devices will soon be disabled, and does not forward any further requests to that one of the plurality of devices.

5. The system of claim 1 further comprising:

a leader application executing on a first device of said plurality of devices, said leader application that identifies a new device and adds said new device to said plurality of devices.

6. The system of claim 5, said leader application that further identifies a first device and removes said first device from said plurality of devices.

7. The system of claim 1 further comprising:

a second plurality of devices being a subset of said plurality of devices; and

a second namespace definition defining a second organization for at least a subset of said files into a second share, said second namespace definition being stored said a cluster database accessible to each of said second plurality of devices, said second share being made available to client devices.

8. The system of claim 7, said second plurality of devices being the same as said plurality of devices.

9. The system of claim 1, said file service that further:

detects a failure when attempting to update said locally-cached version of said namespace definition for a first device and disables said file service for said first device.

10. The system of claim 9, said file service that further:

retries updating said locally cached version of said namespace definition and adds said first device to said plurality of devices when said updating is successful.

11. The system of claim 1, wherein updating by the file service the locally cached version of the namespace definition on the at least one of the plurality of devices includes:

detecting, by the file service, a failure when the at least one of the plurality of devices attempts to update the locally cached version of the namespace definition.

12. The system of claim 11, wherein the file service, after detecting the failure, causes the at least one of the plurality of devices to disable the locally cached version of the namespace definition.

13. The system of claim 11, wherein the file service, after detecting the failure, causes the at least one of the plurality of devices to retry to update the locally cached version of the namespace definition.

14. The system of claim 1, wherein a first instance of the file service operates on a first group of nodes to service a first request, wherein a second instance of the file service operates on a second group of nodes to service a second request, and wherein a particular node belonging to the first group of nodes also belongs to the second group of nodes such that the particular node is usable to service both the first request and the second request.

15. A method comprising:

defining a namespace definition for an organization for a data store comprising a plurality of files accessible by a plurality of devices, the namespace definition, in addition to defining the organization, also defines which files are to be included within the data store;

storing, in a cluster database on a master node that is accessible to each of the plurality of devices, the namespace definition;

identifying, with a files service, changes to the namespace definition;

updating the namespace definition on the cluster database on the master node to an updated namespace definition in response to an identified change to the namespace definition; and

updating by the file service a locally cached version of the namespace definition on at least one of the plurality of devices.

16. The method of claim 15 further comprising:

receiving at least one request from a client device;

determining one of the plurality of devices to process the at least one request; and

processing the at least one request on the determined one of the plurality of devices.

17. The method of claim 16 further comprising:

determining that one of the plurality of devices will soon be disabled; and

blocking further requests to that one of the plurality of devices.

18. The method of claim 15 further comprising:

identifying by a leader application executing on a first device of the plurality of devices, that a new device that has been added; and

adding the new device to the plurality of devices.

19. The method of claim 18 further comprising:

detecting a failure when updating a locally cached version of the namespace definition on one of the plurality of devices; and

disabling the file service for the one of the plurality of devices.

20. The method of claim 19 further comprising:

retrying the update of the locally cached version of the namespace definition;

adding the one of the plurality of devices back to the plurality of devices when the retrying the update is successful.