Storage area network attached clustered storage system

Patent metadata

Field	Value
Publication number	US-2016119427-A1
Application number	US-201514918033-A
Country	US
Kind code	A1
Filing date	Oct 20, 2015
Priority date	Sep 23, 2011
Publication date	Apr 28, 2016
Grant date	—

Abstract

A storage area network (SAN)-attached storage system architecture is disclosed. The storage system provides strongly consistent distributed storage communication protocol semantics, such as SCSI target semantics. The system includes a mechanism for presenting a single distributed logical unit, comprising one or more logical sub-units, as a single logical unit of storage to a host system by associating each of the logical sub-units that make up the single distributed logical unit with a single host visible identifier that corresponds to the single distributed logical unit. The system further includes mechanisms to maintain consistent context information for each of the logical sub-units such that the logical sub-units are not visible to a host system as separate entities from the single distributed logical unit.

First claim

1 - 18 . (canceled)

19 . A storage area network (SAN)-attached clustered storage system comprising:
a plurality of storage server nodes coupled to communicate with each other through an interconnect, each storage server node of the plurality of storage server nodes to include:
a protocol module, executed by a hardware processor, to provide a host with access to the SAN-attached clustered storage system, wherein the protocol module implements a first logical sub-unit of a single distributed logical unit in a SAN environment that is distributed across the SAN-attached clustered storage system, such that the first logical sub-unit and the single distributed logical unit appear as a single entity, wherein the single distributed logical unit comprises a plurality of logical sub-units including the first logical sub-unit; and
a cluster services module, executed by the hardware processor, to maintain consistency between the plurality of logical sub-units that make up the single distributed logical unit, the cluster services module including:
a cluster transaction manager to control cluster management operation requests and cluster control operation requests by using a multi-phase voting process, wherein the cluster transaction manager includes a group coordinator to direct the multi-phase voting process in a group to determine whether to commit an operation and to maintain a queue of entities in the group, such that each entity has an opportunity to propose the operation.

20 . The SAN-attached clustered storage system of claim 19 , the cluster services module including a quorum manager, wherein the quorum manager of one of the plurality of storage server nodes is a master quorum manager to monitor an in quorum/out of quorum status of each other quorum manager in the clustered storage system.

21 . The SAN-attached clustered storage system of claim 19 , wherein each logical sub-unit of the plurality of logical sub-units implements an instance of a global task set to queue commands and task management functions that modify context information, such that context information between each logical sub-unit remains consistent.

22 . The SAN-attached clustered storage system of claim 19 , wherein the protocol module implements a plurality of virtual SCSI target ports, wherein each of the plurality of virtual SCSI target ports is associated with a physical target port of a corresponding storage server node and presents the plurality of virtual SCSI target ports to the host.

23 . The SAN-attached clustered storage system of claim 19 , wherein the cluster transaction manager includes the group coordinator to direct the multi-phase voting process in a group to determine whether to commit an operation.

24 . The SAN-attached clustered storage system of claim 23 , wherein the group coordinator maintains a queue of entities in the group, such that each entity has an opportunity to propose the operation.

25 . The SAN-attached clustered storage system of claim 19 , wherein the master quorum manager determines whether each storage server node in the cluster is in-quorum or out-of-quorum based on whether a storage server node has not successfully responded to a heartbeat proposal a threshold number of times.

26 . A storage server node for use in a SAN-attached clustered storage system that includes a plurality of storage server nodes, the storage server node comprising:
a protocol module to implement a logical sub-unit of a single distributed logical unit that is distributed across the SAN-attached clustered storage system, such that the logical sub-unit and the single distributed logical unit appear as a single entity to the host; a cluster transaction manager to control cluster management operation requests and cluster control operation requests by using a multi-phase voting process; and a quorum manager to maintain consistency between the storage server node and each other storage server node of the plurality of storage server nodes.

27 . The storage server node of claim 26 , wherein the quorum manager maintains consistency between the storage server node and each other storage server node of the plurality of storage server nodes based on an in-quorum/out-of-quorum status for each of the plurality of server nodes.

28 . The storage server node of claim 26 , wherein the logical sub-unit implements an instance of a global task set to queue commands and task management functions that modify context information, such that context information between each logical sub-unis in each of the plurality of storage server nodes remains consistent.

29 . The storage server node of claim 26 , wherein the protocol module implements a plurality of virtual SCSI target ports, wherein each of the plurality of virtual SCSI target ports is associated with a physical target port of a corresponding storage server node and presents the plurality of virtual SCSI target ports to the host.

30 . The storage server node of claim 26 , wherein the cluster transaction manager includes a group coordinator to direct the multi-phase voting process in a group to determine whether to commit an operation.

31 . The storage server node of claim 30 , wherein the group coordinator maintains a queue of entities in the group, such that each entity has an opportunity to propose the operation.

32 . The storage server node of claim 26 , wherein the master quorum manager determines whether each storage server node in the cluster is in-quorum or out-of-quorum based on whether a storage server node has not successfully responded to a heartbeat proposal a threshold number of times.

33 . A method comprising:
in a SAN-attached clustered storage system that includes a plurality of storage server nodes, each storage server node: implementing, by a protocol module, a logical sub-unit of a single distributed logical unit that is distributed across the SAN-attached clustered storage system, such that the logical sub-unit and the single distributed logical unit appear as a single entity to the host; controlling, by a cluster transaction manager, cluster management operation requests and cluster control operation requests by using a multi-phase voting process; and maintaining consistency between each of the plurality of storage server nodes.

34 . The method of claim 33 , comprising maintaining, by the logical sub-unit, context information that is consistent with corresponding context information of a plurality of other logical sub-units that make up the single distributed logical unit.

35 . The method of claim 34 , wherein maintaining context information includes:
implementing an instance of a global task set for the logical sub-unit to queue commands and task management functions that modify context information; and coordinating entry of commands into the global task set such that each instance of the global task set is consistent with each other instance of the global task set.

36 . The method of claim 33 , comprising:
implementing, by the protocol module, a plurality of virtual target ports, wherein each of the plurality of virtual target ports is associated with a single physical target port of the storage server node; and presenting the plurality of virtual target ports to the host.

37 . The method of claim 33 , wherein maintaining consistency between each of the plurality of storage server nodes includes:
causing, by a quorum manager, the cluster transaction manager to originate a heartbeat proposal to each storage server node in the storage cluster; and receiving,

Assignees

• Netapp Inc

Inventors

• Rao Santosh Ananth
• Brown Geoffrey Stewart
• Natarajan Srikumar
• Patnaik Pranab
• Tan Kai
• Corbett Peter Frank
• Venkatesan Vivek

Classifications

• H04L43/10

  Active monitoring, e.g. heartbeat, ping or trace-route · CPC title

• G06F13/22

  using successive scanning, e.g. polling (G06F13/24 takes precedence) · CPC title

• H04L67/1097 Primary

  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] · CPC title