Techniques for distributing access to filesystems through multipe filesystem management nodes

• US 9,753,943 B1
• Filed: 03/31/2015
• Issued: 09/05/2017
• Est. Priority Date: 03/31/2015
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Bono, Jean-Pierre, Corniquet, Frederic, Srinivasan, Sudhir
• Primary Examiner(s): Jami, Hares
• Application Number:
  US14/674,732
• Time in Patent Office: 889 Days
• Field of Search: 707821, 707822, 707827
• 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 16/188 Virtual file systems
• 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

A data storage system has one or more processing nodes, each processing node comprising memory coupled to processing circuitry, the one or more processing nodes running a set of virtualized filesystem managers (VFMs) having a shared namespace. The set of VFMs includes (a) a master VFM for managing a configuration filesystem stored on the persistent storage, the configuration filesystem storing configuration information for the set of VFMs and (b) a slave VFM for remotely mounting the configuration filesystem from the master VFM, the slave VFM obtaining details of its configuration from the remotely-mounted configuration filesystem, the slave VFM sharing the shared namespace with the set of VFMs, the shared namespace including a set of filesystems, each filesystem of the shared namespace being managed by a single VFM, and each filesystem of the shared namespace being accessible by the remote host through any of the VFMs of the set of VFMs.

Claims

1. A system comprising:

persistent storage;

networking circuitry for connecting to a remote host over a network; and

one or more processing nodes, each processing node comprising memory coupled to processing circuitry, the one or more processing nodes running a set of virtualized filesystem managers (VFMs) having a shared namespace, the set of VFMs including;

a master VFM, the master VFM managing a configuration filesystem stored on the persistent storage, the configuration filesystem storing configuration information for the set of VFMs; and

a slave VFM, the slave VFM remotely mounting the configuration filesystem from the master VFM, the slave VFM obtaining details of its configuration from the remotely-mounted configuration filesystem, the slave VFM sharing the shared namespace with the set of VFMs, the shared namespace including a set of filesystems stored on the persistent storage, each filesystem of the shared namespace being managed by a single VFM of the set of VFMs, and each filesystem of the shared namespace being accessible by the remote host through any of the VFMs of the set of VFMs.

2. The system of claim 1 wherein the master VFM and the slave VFM run on a single processing node of the one or more processing nodes, the master VFM and the slave VFM sharing access to cache of the single processing node.

3. The system of claim 1 wherein the master VFM and the slave VFM run on separate processing nodes of the one or more processing nodes.

4. The system of claim 1 wherein the master VFM manages a first filesystem of the set of filesystems, the first filesystem being locally mounted by the master VFM, the slave VFM remotely mounting the first filesystem from the master VFM.

5. The system of claim 1 wherein the slave VFM manages a first filesystem of the set of filesystems, the first filesystem being locally mounted by the slave VFM, the master VFM remotely mounting the first filesystem from the slave VFM.

6. The system of claim 1 wherein the set of VFMs also includes another slave VFM, the other slave VFM remotely mounting the configuration filesystem from the master VFM, the other slave VFM obtaining details of its configuration from the remotely-mounted configuration filesystem, the other slave VFM sharing the shared namespace with the set of VFMs.

7. The system of claim 6 wherein the master VFM manages a first filesystem of the set of filesystems, the first filesystem being locally mounted by the master VFM, the slave VFM and the other slave VFM both remotely mounting the first filesystem from the master VFM.

8. The system of claim 6 wherein the slave VFM manages a first filesystem of the set of filesystems, the first filesystem being locally mounted by the slave VFM, the master VFM and the other slave VFM both remotely mounting the first filesystem from the slave VFM.

9. The system of claim 1 wherein the one or more processing nodes also run another set of VFMs having another shared namespace, the other set of VFMs including:

the master VFM; and

another slave VFM not within the set of VFMs, the other slave VFM remotely mounting the configuration filesystem from the master VFM, the other slave VFM obtaining details of its configuration from the remotely-mounted configuration filesystem, the other slave VFM sharing the other shared namespace with the other set of VFMs, the other shared namespace including another set of filesystems stored on the persistent storage, each filesystem of the other shared namespace being managed by a single VFM of the other set of VFMs, and each filesystem of the other shared namespace being accessible by the remote host through any of the VFMs of the other set of VFMs, the set of filesystems and the other set of filesystems differing in at least one filesystem member.

10. The system of claim 9 wherein the master VFM manages a local filesystem not within the set of filesystems and not within the other set of filesystems, the local filesystem being locally mounted by the master VFM, the local filesystem not being accessible through either of the slave VFM and the other slave VFM.

11. A method performed by a slave virtualized filesystem manager (VFM) running on processing circuitry of a data storage system, the method comprising:

remotely mounting a configuration filesystem stored on persistent storage of the data storage system from a master VFM running on the processing circuitry of the data storage system, the master VFM managing the remotely-mounted configuration filesystem, the remotely-mounted configuration filesystem storing configuration information for a set of VFMs running on the processing circuitry having a shared namespace, the slave VFM obtaining details of its configuration from the remotely-mounted configuration filesystem;

remotely mounting a remote filesystem stored on persistent storage of the data storage system from a manager VFM of the set of VFMs running on the processing circuitry, the manager VFM managing the remote filesystem, the remote filesystem being within the shared namespace; and

processing a first filesystem storage command directed at the remote filesystem received from a remote host over a network interface of the data storage system by sending a second filesystem storage command directed at the remote filesystem of the shared namespace from the slave VFM to the manager VFM and receiving a storage fulfillment reply at the slave VFM from the manager VFM in response.

12. The method of claim 11 wherein the manager VFM is the master VFM.

13. The method of claim 11 wherein the manager VFM is another slave VFM of the set of VFMs running on the processing circuitry.

14. The method of claim 11 wherein the method further comprises locally mounting a local filesystem stored on persistent storage of the data storage system, the local filesystem being within the shared namespace, the slave VFM managing the local filesystem.

15. The method of claim 14 wherein the method further comprises processing a third filesystem storage command directed at the local filesystem of the shared namespace received from the remote host over the network interface of the data storage system by sending a data storage command directed at the local filesystem to the persistent storage of the data storage system and receiving a storage fulfillment reply at the slave VFM from the persistent storage in response.

16. The method of claim 14 wherein the method further comprises processing a third filesystem storage command directed at the local filesystem of the shared namespace received from an intermediary VFM of the set of VFMs, the intermediary VFM having previously received a fourth filesystem storage command directed at the local filesystem of the shared namespace from the remote host over the network interface of the data storage system, wherein processing the third filesystem storage command includes:

sending a data storage command directed at the local filesystem to the persistent storage of the data storage system;

receiving a first storage fulfillment reply at the slave VFM from the persistent storage in response to sending the data storage command to the persistent storage; and

sending a second storage fulfillment reply from the slave VFM to the intermediary VFM in response to receiving the first storage fulfillment reply from the persistent storage.