Cluster management in large-scale storage systems

Patent metadata

Field	Value
Publication number	US-12099719-B2
Application number	US-202218090792-A
Country	US
Kind code	B2
Filing date	Dec 29, 2022
Priority date	Dec 29, 2022
Publication date	Sep 24, 2024
Grant date	Sep 24, 2024

Abstract

Techniques are provided for implementing a distributed hierarchical cluster management system. A system comprises a data storage system and a cluster management system. The data storage system comprises a cluster of storage nodes that is partitioned into a plurality of subclusters of storage nodes. The cluster management system is deployed on at least some of the storage nodes of the data storage system, and comprises a global management system and a plurality of local management subsystems. Each local management subsystem is configured to manage a respective subcluster of the plurality of subclusters of storage nodes, and communicate with the global management system to provide subcluster status information to the global management system regarding a current state and configuration of the respective subcluster of storage nodes. The global management system is configured to manage the cluster of storage nodes using the subcluster status information provided by the local management subsystems.

First claim

What is claimed is:

1. A system, comprising:
a data storage system comprising a cluster of storage nodes that is partitioned into a plurality of subclusters of storage nodes; and
a cluster management system deployed on at least some of the storage nodes of the data storage system, wherein the cluster management system comprises a global management system and a plurality of local management subsystems;
wherein each of the local management subsystems is configured to manage a respective subcluster of storage nodes of the plurality of subclusters of storage nodes independent of other ones of the local management subsystems, and is further configured to communicate with the global management system to provide subcluster status information to the global management system regarding a current state and configuration of the respective subcluster of storage nodes; and
wherein the global management system is configured to manage the cluster of storage nodes using the subcluster status information provided by the local management subsystems.

2. The system of claim 1 , wherein each subcluster of storage nodes of the plurality of subclusters of storage nodes comprises a logical protection domain.

3. The system of claim 1 , wherein the global management system is configured to manage and communicate with client drivers of an application layer to provide connectivity information to the client drivers which enables the client drivers to connect to storage nodes of the cluster storage nodes and to access storage volumes of a virtual storage layer comprising aggregated storage capacity of the cluster of storage nodes.

4. The system of claim 1 , wherein configuring each local management subsystem to manage a respective subcluster of storage nodes of the plurality of subclusters of storage nodes comprises configuring each local management subsystem to manage operations associated with the respective subcluster of storage nodes, wherein the operations include at least one of: managing a rebuild operation within the respective subcluster of storage nodes; managing a rebalance operation within the respective subcluster of storage nodes; adding a new storage node to the respective subcluster of storage nodes; removing an existing storage node from the respective subcluster of storage nodes; and performing a migration operation within the respective subcluster of storage nodes.

5. The system of claim 1 , wherein:
the data storage system comprises a scale-out software-defined storage system comprising software components that are deployed on each storage node of the cluster of storage nodes to implement a storage control system on each storage node; and
each local management subsystem is configured to manage and communicate with the storage control systems of the storage nodes within the respective subclusters of storage nodes.

6. The system of claim 1 , wherein:
the global management system comprises a cluster of global metadata manager (MDM) nodes deployed on different storage nodes of the data storage system, wherein the cluster of global MDM nodes comprises a global primary MDM node and at least one global secondary MDM node; and
the local management subsystem for a given subcluster of storage nodes comprises a cluster of local MDM nodes deployed on different storage nodes within the given subcluster of storage nodes, wherein the cluster of local MDM nodes comprises a local primary MDM node and at least one local secondary MDM node;
wherein the local primary MDM node of the local management subsystem communicates with the global primary MDM node of the global management system.

7. The system of claim 6 , wherein the cluster management system is configured to cause a local secondary MDM node of a given cluster of local MDM nodes to assume a role of the global primary MDM node of the cluster of global MDM nodes.

8. The system of claim 6 , wherein the cluster management system is configured to automatically reallocate an MDM role of one or more MDM nodes of one or more of the cluster of global MDM nodes and a given cluster of local MDM nodes, in response to detection, by the cluster management system, of an event that warrants the MDM role reallocation.

9. The system of claim 8 , wherein the event comprises at least one of: detection of a loss of one or more network paths to one or more storage nodes; detection of congestion on one or more network paths; and reduced availability of storage or compute resources on one or more storage nodes on which the MDM nodes are deployed.

10. A method, comprising:
partitioning a cluster of storage nodes of a data storage system into a plurality of subclusters of storage nodes;
deploying a cluster management system on at least some of the storage nodes of the data storage system, wherein the cluster management system comprises a global management system and a plurality of local management subsystems;
configuring each of the local management subsystems to manage a respective subcluster of storage nodes of the plurality of subclusters of storage nodes independent of other ones of the local management subsystems, and to communicate with the global management system to provide subcluster status information to the global management system regarding a current state and configuration of the respective subcluster of storage nodes; and
configuring the global management system to manage the cluster of storage nodes using the subcluster status information provided by the local management subsystems.

11. The method of claim 10 , wherein partitioning the cluster of storage nodes of the data storage system into the plurality of subclusters of storage nodes comprises logically partitioning the cluster of storage nodes into a plurality of protection domains.

12. The method of claim 10 , further comprising configuring the global management system to manage and communicate with client drivers of an application layer to provide connectivity information to the client drivers which enables the client drivers to connect to storage nodes of the cluster of storage nodes and to access storage volumes of a virtual storage layer comprising aggregated storage capacity of the cluster of storage nodes.

13. The method of claim 10 , wherein configuring each local management subsystem to manage a respective subcluster of storage nodes of the plurality of subclusters of storage nodes comprises configuring each local management subsystem to manage operations associated with the respective subcluster of storage nodes, wherein the operations include at least one of: managing a rebuild operation within the respective subcluster of storage nodes; managing a rebalance operation within the respective subcluster of storage nodes; adding a new storage node to the respective subcluster of storage nodes; removing an existing storage node from the respective subcluster of storage nodes; and performing a migration operation within the respective subcluster of storage nodes.

14. The method of claim 10 , wherein deploying the cluster management system on at least some of the storage nodes of the data storage system comprises:
deploying a cluster of global metadata manager (MDM) nodes on different storage nodes of the data storage system to implement the global management system, wherein the cluster of global MDM nodes comprises a global primary MDM node and at least one global secondary MDM node; and
deploying a cluster of local MDM nodes on different storage nodes within a given subcluster of storage nodes to implement a given local management subsystem for the given subcluster of storage nodes, wherein the cluster of local MDM nodes comprises a local primary MDM node and

Assignees

• Dell Products Lp

Inventors

• Trudel Christopher
• Shah Maunish A
• Rastvortsev Alexey

Classifications

• G06F3/067

  Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS] · CPC title

• G06F3/0631 Primary

  by allocating resources to storage systems · CPC title

• G06F3/0604 Primary

  Improving or facilitating administration, e.g. storage management · CPC title