Dynamically changing members of a consensus group in a distributed self-healing coordination service

Patent metadata

Field	Value
Publication number	US-9690675-B2
Application number	US-201414334162-A
Country	US
Kind code	B2
Filing date	Jul 17, 2014
Priority date	Jul 17, 2014
Publication date	Jun 27, 2017
Grant date	Jun 27, 2017

Abstract

Systems, methods, and computer program products for managing a consensus group in a distributed computing cluster, by determining that an instance of an authority module executing on a first node, of a consensus group of nodes in the distributed computing cluster, has failed; and adding, by an instance of the authority module on a second node of the consensus group, a new node to the consensus group to replace the first node. The new node is a node in the computing cluster that was not a member of the consensus group at the time the instance of the authority module executing on the first node is determined to have failed.

First claim

What is claimed is:

1. A method for managing a consensus group in a distributed computing cluster, the method comprising:
reaching consensus, by at least two members of a consensus group of nodes in the distributed computing cluster, to elect an instance of an authority module executing on a first node, of the consensus group of nodes, to serve as a master instance of the authority module for the consensus group of nodes, wherein each member of the consensus group of nodes executes a respective instance of the authority module, wherein the master instance of the authority module is configured to assign locks to processes executing in the distributed computing cluster;
reaching consensus, by at least two members of the consensus group of nodes, that an instance of an authority module executing on the first node, of the consensus group of nodes, has failed;
reaching consensus, by the remaining members of the consensus group of nodes, to elect the instance of the authority module executing on a second node, of the consensus group of nodes, to serve as the master instance of the authority module for the consensus group;
receiving, by the master authority module on the second node, an advertisement from a new node in the computing cluster via a zero-configuration networking protocol, wherein the new node comprises a node in the computing cluster that was not a member of the consensus group at the time the at least two members of the consensus group of nodes reached consensus that the instance of the authority module executing on the first node failed; and
adding, by the master authority module on the second node, the new node to the consensus group to replace the first node.

2. The method of claim 1 , further comprising:
transmitting to the new node:
a set of nodes in the distributed computing cluster eligible to join the consensus group; and
a state of a set of locks issued to nodes in the distributed computing cluster.

3. The method of claim 1 , further comprising:
transmitting, by the instance of the authority module on the second node, to each node in the distributed computing cluster, an indication that the instance of the authority module on the second node has been elected as the master authority module for the consensus group,
wherein the locks are required to access data in a storage device, wherein the distributed storage cluster comprises a secondary storage environment.

4. The method of claim 1 , further comprising:
adding, by the master authority module on second node, the new node to a set of available nodes in the computing cluster; and
sharing the set of available nodes with members of the consensus group.

5. The method of claim 1 , further comprising:
removing, by the second node, the first node from the consensus group.

6. A distributed computing cluster, comprising:
a plurality of nodes configured to provide a service to clients, each node having a processor and a memory;
a consensus group of nodes formed from a subset of the plurality of nodes, the consensus group of nodes, each executing an instance of an authority module which performs operations for managing the consensus group, the operation, comprising:
reaching consensus, by at least two members of the consensus group of nodes, to elect an instance of an authority module executing on a first node, of the consensus group of nodes, to serve as a master instance of the authority module for the consensus group of nodes, wherein the master instance of the authority module is configured to assign locks to processes executing in the distributed computing cluster;
reaching consensus, by at least two members of the consensus group of nodes in the distributed computing cluster, that an instance of an authority module executing on the first node has failed;
receiving, by the master authority module on the second node, an advertisement from a new node in the computing cluster via a zero-configuration networking protocol, wherein the new node comprises a node in the computing cluster that was not a member of the consensus group at the time the at least two members of the consensus group of nodes reached consensus that the instance of the authority module executing on the first node failed; and
adding, by the master authority module on the second node, the new node to the consensus group to replace the first node.

7. The distributed computing cluster of claim 6 , wherein the operation further comprises:
transmitting to the new node:
a set of nodes in the distributed computing cluster eligible to join the consensus group; and
a state of a set of locks issued to nodes in the distributed computing cluster.

8. The distributed computing cluster of claim 6 , wherein the operation further comprises:
transmitting, by the instance of the authority module on the second node, to each node in the distributed computing cluster, an indication that the instance of the authority module on the second node has been elected as the master authority module for the consensus group,
wherein the locks are required to access data in a storage device, wherein the distributed storage cluster comprises a secondary storage environment.

9. The distributed computing cluster of claim 6 , wherein the operation further comprises:
adding, by the master authority module on second node, the new node to a set of available nodes in the computing cluster; and
sharing the set of available nodes with members of the consensus group.

10. The distributed computing cluster of claim 6 , wherein the operation further comprises:
removing, by the second node, the first node from the consensus group.

11. A non-transitory computer-readable storage medium storing instructions, which, when executed on a processor, perform operations for managing a consensus group in a distributed computing cluster, the operation comprising:
reaching consensus, by at least two members of a consensus group of nodes in the distributed computing cluster, to elect an instance of an authority module executing on a first node, of the consensus group of nodes, to serve as a master instance of the authority module for the consensus group of nodes, wherein each member of the consensus group of nodes executes a respective instance of the authority module, wherein the master instance of the authority module is configured to assign locks to processes executing in the distributed computing cluster;
reaching consensus, by at least two members of the consensus group of nodes, that an instance of an authority module executing on the first node, of the consensus group of nodes, has failed;
reaching consensus, by the remaining members of the consensus group of nodes, to elect the instance of the authority module executing on a second node, of the consensus group of nodes, to serve as the master instance of the authority module for the consensus group;
receiving, by the master authority module on the second node, an advertisement from a new node in the computing cluster via a zero-configuration networking protocol, wherein the new node comprises a node in the computing cluster that was not a member of the consensus group at the time the at least two members of the consensus group of nodes reached consensus that the instance of the authority module executing on the first node failed; and
adding, by the master authority module on the second node, the new node to the consensus group to replace the first node.

12. The computer-readable storage medium of claim 11 , further comprising:
transmitting to the new node:
a set of nodes in the distributed computing cluster eligible to join the consensus group; and
a state of a set of locks issued to nodes in the distributed

Assignees

• Cohesity Inc

Inventors

• Madduri Sashikanth
• Aron Mohit
• Reddy Vinay
• Gupta Vipin

Classifications

• G06F11/2028

  eliminating a faulty processor or activating a spare · CPC title

• G06F11/1658

  Data re-synchronization of a redundant component, or initial sync of replacement, additional or spare unit · CPC title

• G06F11/1425

  by reconfiguration of node membership · CPC title

• H04L41/5009

  Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF] · CPC title

• H04L41/0668

  by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure · CPC title