Techniques for storing and distributing metadata among nodes in a storage cluster system

• US 9,838,477 B2
• Filed: 10/30/2014
• Issued: 12/05/2017
• Est. Priority Date: 10/30/2014
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Mu, Paul Yuedong, Sundararajan, Manoj, Ngan, Paul
• Primary Examiner(s): TOKUTA, SHEAN S
• Application Number:
  US14/528,166
• Publication Number:
  US 20160127467A1
• Time in Patent Office: 1,132 Days
• 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

Various embodiments are generally directed to techniques for reducing the time required for a node to take over for a failed node or to boot. An apparatus includes an access component to retrieve a metadata from a storage device coupled to a first data storage module of a first node during boot, the metadata generated from a first mutable metadata portion and an immutable metadata portion, and the first metadata specifying a first address of a second data storage module of a second node; a replication component to contact the second data storage module at the first address; and a generation component to, in response to failure of the contact, request a second mutable metadata portion from a network protocol module of the first node and generate a second metadata from the second mutable metadata portion and the immutable metadata portion, the second mutable metadata portion specifying a second address of the second data storage module.

Claims

1. A device comprising:

a memory containing machine readable medium comprising machine executable code having stored thereon instructions for performing a method; and

a processor coupled to the memory, the processor configured to execute the machine executable code to cause the processor to;

retrieve a first metadata from a first storage device of a first node during a boot procedure, the first metadata generated from a first immutable metadata portion specifying a file system selection and a RAID level selection of a second node and a first mutable metadata portion specifying a first IP address of the second node, wherein the file system selection is one of a common internet file system, a network file system, or a write anywhere file system;

attempt to contact the second node at the first IP address;

determine a failure to contact the second node at the first IP address;

obtain a second mutable metadata portion specifying a second IP address of the second node; and

generate a second metadata from the second IP address of the second node specified by the second mutable metadata portion and the file system selection of the second node specified by the first immutable metadata portion.

2. The device of claim 1, wherein the machine executable code causes the processor to:

generate the first metadata from at least the first mutable metadata portion and the first immutable metadata portion;

receive client device data from a client device coupled to the first node; and

store the first metadata in the storage space alongside the client device data.

3. The device of claim 2, wherein the machine executable code causes the processor to:

perform a test of an interconnect, between the first node and the second node, to obtain the first IP address;

repeat performance of the test of the interconnect to obtain the second IP address; and

translate a storage services request to store the client device data from the client device into a data access command to store the client device data within storage.

4. The device of claim 1, wherein the machine executable code causes the processor to:

attempt to contact the second node at the second IP address via an interconnect; and

transmit the second metadata to a partner node of the first node via a second interconnect in response to successfully contacting the second node via the interconnect at the second IP address.

5. The device of claim 1, wherein the machine executable code causes the processor to:

attempt to contact the second node at the second IP address via an interconnect; and

store the second metadata within storage in place of the first metadata in response to successfully contacting the second node via the interconnect at the second IP address.

6. The device of claim 1, wherein the machine executable code causes the processor to:

replicate a data access command to store client device data within storage to generate a replica data access command;

form an active communications session through an interconnect between the first node and the second node;

form an inactive communications session through the interconnect between the first node and a third node;

determine that the inactive communications session has become active and the active communications session has become inactive; and

generate a third metadata specifying a third IP address of the third node to which to transmit the replica data access command in lieu of transmitting the replica data access command to the second node.

7. The device of claim 1, wherein the machine executable code causes the processor to:

replicate a data access command to store client device data within storage to generate a replica data access command;

transmit the replica data access command to the second node to store the client device data within a second storage device coupled to the second node at least partially in parallel with storage of the client device data within the storage;

receive an indication of results of storing the client device data within the second storage device from the second node; and

generate a third metadata based at least on an indication in the results of failure to store the client device data within the second storage device, the third metadata specifying a third IP address of a third node to which to transmit the replica data access command in place of the second node.

8. A method comprising:

retrieving a first metadata from a first storage device of a first node during a boot procedure, the first metadata generated from a first immutable metadata portion specifying a file system selection and a RAID level selection of a second node and a first mutable metadata portion specifying a first IP address of the second node, wherein the file system selection is one of a common internet file system, a network file system, or a write anywhere file system;

transmitting a message to the second node at the first IP address;

determining a failure of the second node to receive the message at the first IP address;

obtaining a second mutable metadata portion specifying a second IP address of the second node; and

generating, at the first node, a second metadata from the second IP address of the second node specified by the second mutable metadata portion and the file system selection of the second node specified by the first immutable metadata.

9. The method of claim 8, comprising:

generating, at the first node, the first metadata from at least the first mutable metadata portion and the first immutable metadata portion;

receiving, at the first node, client device data from a client device coupled to the first node; and

storing the first metadata in the storage space alongside the client device data.

10. The method of claim 8, comprising:

transmitting the message to the second node at the second IP address via an interconnect; and

transmitting the second metadata to a partner node of the first node via a second interconnect in response to the second node successfully receiving the message at the second IP address.

11. The method of claim 8, comprising:

transmitting the message to the second node at the second IP address via an interconnect; and

storing the second metadata within storage in place of the first metadata in response to the second node successfully receiving the message at the second IP address.

12. The method of claim 10, comprising:

determining a second failure of the second node to receive the message at the second IP address;

transmitting a second request for a second immutable metadata portion; and

generating a third metadata from at least the second mutable metadata portion and the second immutable metadata portion, the third metadata portion specifying a third IP address of the second node.

13. The method of claim 8, comprising:

replicating a data access command to store the client device data within storage to generate a replica data access command;

forming an active communications session through an interconnect between the first node and the second node;

forming an inactive communications session through the interconnect between the first node and a third node;

determining that the inactive communications session has become active and the active communications session has become inactive;

generating, at the first node, a third metadata specifying a third IP address of the third node to which to transmit the replica data access command in lieu of transmitting the replica data access command to the second node yJe; and

storing the third metadata within storage in place of the first metadata.

14. A non-transitory machine-readable storage medium comprising instructions that when executed by a machine, cause the machine to:

retrieve first metadata from a first storage device of a first node during a boot procedure, the first metadata generated from a first immutable metadata portion specifying, a file system selection and a RAID level selection of a second node and a first mutable metadata portion specifying a first IP address of the second node wherein the file system selection is one of a common internet file system, a network file system, or a write anywhere file system;

transmit a message to the second node at the first IP address;

determine a failure of the second node to receive the message at the first IP address;

transmit a request for a second mutable metadata portion specifying a second IP address of the second node; and

generate a second metadata from the second IP address of the second node specified by the second mutable metadata portion and the file system selection of the second node specified by the first immutable metadata portion.

15. The non-transitory machine-readable storage medium of claim 14, wherein the instructions cause the machine to:

generate the first metadata from at least the first mutable metadata portion and the first immutable metadata portion;

receive client device data from a client device coupled to the first node; and

store the first metadata in storage alongside the client device data.

16. The non-transitory machine-readable storage medium of claim 14, wherein the instructions cause the machine to:

transmit the message to the second node at the second IP address via an interconnect; and

transmit the second metadata to a partner node of the first node via a second interconnect in response to the second node successfully receiving the message at the second IP address.

17. The non-transitory machine-readable storage medium of claim 14, wherein the instructions cause the machine to:

transmit the message to the second node at the second IP address via an interconnect; and

store the second metadata within storage in place of the first metadata in response to the second node successfully receiving the message at the second address.

18. The non-transitory machine-readable storage medium of claim 14, wherein the instructions cause the machine to:

replicate a data access command to store the client device data within storage to generate a replica data access command;

transmit the replica data access command to the second node to store the client device data within a second storage device coupled to the second node at least partially in parallel with storage of the client device data within storage;

receive an indication of results of storing the client device data within the second storage device from the second node; and

generate a third metadata based at least on an indication in the results of failure to store the client device data within the second storage device, the third metadata specifying a third IP address of a third node to which to transmit the replica data access command in place of the second node.