Location-Based Resource Availability Management in a Partitioned Distributed Storage Environment

• US 20170318092A1
• Filed: 04/29/2016
• Published: 11/02/2017
• Est. Priority Date: 04/29/2016
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Maredia, Sahir, George, Martin, Srinivas, Gargi, Popuri, Sriram
• Granted Patent: US 10,205,782 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources
• CPC Class Codes: H04L 45/28 using route fault recovery
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation
• 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]
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]

Abstract

In some embodiments, a cluster computing system notifies a host system that a first path to a resource in the cluster computing system is optimized and that a second path to the resource is non-optimized. The resource is owned or managed by a first computing node of the cluster computing system. The first path includes the first computing node. The second path includes a second computing node and an intra-cluster connection between the second computing node and the first computing node. A disruption in the intra-cluster connection, which prevents communication between the first and second computing nodes via the intra-cluster connection, is identified. During a time period in which the disruption exists, the host system is notified that the first path is optimized and that the second path is unavailable, and input/output operations between the host system and the resource via the first path are continued.

Claims

1. A method comprising:

notifying, by a cluster computing system and via a data network, a host system that a first path to a resource in the cluster computing system is optimized and that a second path to the resource is non-optimized, wherein the resource is owned or managed by a first computing node of the cluster computing system, wherein the first path includes the first computing node, wherein the second path includes a second computing node and an intra-cluster connection between the second computing node and the first computing node;

identifying a disruption in the intra-cluster connection, wherein the disruption prevents communication between the first computing node and the second computing node via the intra-cluster connection;

during a time period in which the disruption exists, notifying the host system that the first path is optimized and continuing input/output operations between the host system and the resource via the first path; and

notifying, by the cluster computing system and during the time period in which the disruption exists, the host system that the second path is unavailable, wherein notifying the host system that the second path is unavailable comprises transmitting a message to the host system identifying the second path and identifying a state of the second path as unavailable.

2. The method of claim 1, wherein the cluster computing system comprises a distributed storage system, wherein the second computing nodes is included in a set of nodes from the distributed storage system, wherein the method further comprises:

identifying paths within the distributed storage system from the first computing node to the set of nodes;

identifying tasks assigned to at least some of the nodes, wherein the tasks include input/output operations involving resources that are accessible via the paths and that are owned or managed by the first computing node rather than the at least some of the nodes;

aborting, on the at least some nodes in the set of nodes, the tasks having the input/output operations that involve the resources;

notifying, by the cluster computing system, the host system that the paths involving the set of nodes are unavailable for accessing the resources.

3. The method of claim 2, wherein identifying the disruption comprises determining that the cluster computing system has been partitioned into a first partition comprising the first computing node and a second partition comprising the set of nodes, wherein the method further comprises determining that the set of nodes is included in an active partition for the cluster computing system and that the first computing node is excluded from the active partition with the set of nodes, wherein the cluster computing system notifies the host system that the paths are unavailable during the time period based on determining that the first computing node is excluded from the active partition with the set of nodes and that the resources previously accessible via the paths are owned or managed by the first computing node rather than the at least some of the nodes.

4. The method of claim 2, further comprising:

receiving a first command from the host system that identifies one of the paths and that specifies a first action involving a modification of a storage resource owned or managed by the first computing node;

determining that the first command is included in a set of media access commands and that the paths are to be advertised as unavailable in response to any command in the set of media access commands;

based on determining that the first command is included in the set of media access commands, responding to the first command with an error message indicating that the one of the paths is unavailable;

receiving a second command from the host system that identifies one of the paths and that specifies a second action involving a read-only operation for the storage resource owned or managed by the first computing node;

determining that the second command is included in a set of non-media access commands; and

based on determining that the second command is included in the set of non-media access commands, performing the second action in response to receiving the second command.

5. The method of claim 1, wherein the method further comprises determining that the host system requires access to the resource via the first path during the time period in which the disruption exists, wherein the cluster computing system notifies the host system that the first path is optimized during the time period based on determining that the host system requires the access to the resource.

6. The method of claim 1, further comprising:

rebooting the cluster system;

determining, during the reboot operation, that the second computing node is available to host systems prior to the first computing node; and

during an additional time period in which the second computing node is available prior to the first computing node, notifying at least one of the host systems that the second path is unavailable.

7. The method of claim 1, further comprising:

notifying, by the cluster computing system, the host system that a first additional path to an additional resource in the cluster computing system is optimized and that a second additional path to the additional resource is non-optimized, wherein the additional resource is owned or managed by the second computing node, wherein the first additional path includes the second computing node, wherein the second additional path includes a third computing node and an additional intra-cluster connection between the second computing node and the third computing node;

identifying an additional disruption in the additional intra-cluster connection, wherein the additional disruption prevents communication between the second computing node and the third computing node via the additional intra-cluster connection;

during a time period in which the additional disruption exists, notifying the host system that the first additional path is optimized and continuing input/output operations between the host system and the additional resource via the first additional path; and

notifying, by the cluster computing system and during the time period in which the additional disruption exists, the host system that the second additional path is unavailable.

8. A system comprising:

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

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

notify, via a data network, a host system that a first path to a resource in a cluster computing system is optimized and that a second path to the resource is non-optimized, wherein the resource is owned or managed by a first computing node of the cluster computing system, wherein the first path includes the first computing node, wherein the second path includes a second computing node and an intra-cluster connection between the second computing node and the first computing node,

identify a disruption in the intra-cluster connection, wherein the disruption comprises a prevention of communication between the first computing node and the second computing node via the intra-cluster connection, and

during a time period in which the disruption exists;

notify the host system that the first path is optimized,

continue input/output operations between the host system and the resource via the first path, and

notify the host system that the second path is unavailable, wherein the processor is configured to notify the host system that the second path is unavailable by performing operations that comprise transmitting a message to the host system identifying the second path and identifying a state of the second path as unavailable.

9. The system of claim 8, wherein the cluster computing system comprises a distributed storage system, wherein the second computing nodes is included in a set of nodes from the distributed storage system, wherein the processor is further configured to execute the machine-executable code to cause the processor to:

identify paths within the distributed storage system from the first computing node to the set of nodes;

identify tasks assigned to at least some of the nodes, wherein the tasks include input/output operations involving resources that are accessible via the paths and that are owned or managed by the first computing node rather than the at least some of the nodes;

abort, on the at least some nodes in the set of nodes, the tasks having the input/output operations that involve the resources; and

notify the host system that the paths involving the set of nodes are unavailable for accessing the resources.

10. The system of claim 9, wherein the processor is configured to identify the disruption by performing operations that comprise determining that the cluster computing system has been partitioned into a first partition comprising the first computing node and a second partition comprising the set of nodes,

wherein the processor is further configured to execute the machine-executable code to cause the processor to determine that the set of nodes is included in an active partition for the cluster computing system and that the first computing node is excluded from the active partition with the set of nodes,

wherein the processor is configured to notify the host system that the paths are unavailable during the time period based on determining that the first computing node is excluded from the active partition with the set of nodes and that the resources previously accessible via the paths are owned or managed by the first computing node rather than the at least some of the nodes.

11. The system of claim 9, wherein the processor is further configured to execute the machine-executable code to cause the processor to:

receive a first command from the host system that identifies one of the paths and that specifies a first action involving a modification of a storage resource owned or managed by the first computing node;

determine that the first command is included in a set of media access commands and that the paths are to be advertised as unavailable in response to any command in the set of media access commands;

based on a determination that the first command is included in the set of media access commands, respond to the first command with an error message indicating that the one of the paths is unavailable;

receive a second command from the host system that identifies one of the paths and that specifies a second action involving a read-only operation for the storage resource owned or managed by the first computing node;

determine that the second command is included in a set of non-media access commands; and

based on a determination that the second command is included in the set of non-media access commands, perform the second action in response to receiving the second command.

12. The system of claim 8, wherein the processor is further configured to execute the machine-executable code to cause the processor to determine that the host system requires access to the resource via the first path during the time period in which the disruption exists, wherein the processor is configured to notify the host system that the first path is optimized during the time period based on determining that the host system requires the access to the resource.

13. The system of claim 8, wherein the processor is further configured to execute the machine-executable code to cause the processor to:

reboot the cluster system;

determine, during the reboot operation, that the second computing node is available to host systems prior to the first computing node; and

during an additional time period in which the second computing node is available prior to the first computing node, notify at least one of the host systems that the second path is unavailable.

14. The system of claim 8, wherein the processor is further configured to execute the machine-executable code to cause the processor to:

notify the host system that a first additional path to an additional resource in the cluster computing system is optimized and that a second additional path to the additional resource is non-optimized, wherein the additional resource is owned or managed by the second computing node, wherein the first additional path includes the second computing node, wherein the second additional path includes a third computing node and an additional intra-cluster connection between the second computing node and the third computing node;

identify an additional disruption in the additional intra-cluster connection, wherein the additional disruption comprises a prevention of communication between the second computing node and the third computing node via the additional intra-cluster connection; and

during a time period in which the additional disruption exists;

notify the host system that the first additional path is optimized,

continue input/output operations between the host system and the additional resource via the first additional path, and

notify the host system that the second additional path is unavailable.

15. A non-transitory machine-readable medium having stored thereon instructions for performing a method, the instructions comprising machine-executable code that, when executed by one or more machines, causes the one or more machines to perform operations comprising:

notifying, via a data network, a host system that a first path to a resource in a cluster computing system is optimized and that a second path to the resource is non-optimized, wherein the resource is owned or managed by a first computing node of the cluster computing system, wherein the first path includes the first computing node, wherein the second path includes a second computing node and an intra-cluster connection between the second computing node and the first computing node;

identifying a disruption in the intra-cluster connection, wherein the disruption prevents communication between the first computing node and the second computing node via the intra-cluster connection;

during a time period in which the disruption exists, notifying the host system that the first path is optimized and continuing input/output operations between the host system and the resource via the first path; and

notifying, during the time period in which the disruption exists, the host system that the second path is unavailable, wherein notifying the host system that the second path is unavailable comprises transmitting a message to the host system identifying the second path and identifying a state of the second path as unavailable.

16. The non-transitory machine-readable medium of claim 15, wherein the cluster computing system comprises a distributed storage system, wherein the second computing nodes is included in a set of nodes from the distributed storage system, wherein the operations further comprise:

identifying paths within the distributed storage system from the first computing node to the set of nodes;

identifying tasks assigned to at least some of the nodes, wherein the tasks include input/output operations involving resources that are accessible via the paths and that are owned or managed by the first computing node rather than the at least some of the nodes;

aborting, on the at least some nodes in the set of nodes, the tasks having the input/output operations that involve the resources;

notifying the host system that the paths involving the set of nodes are unavailable for accessing the resources.

17. The non-transitory machine-readable medium of claim 16, wherein identifying the disruption comprises determining that the cluster computing system has been partitioned into a first partition comprising the first computing node and a second partition comprising the set of nodes, wherein the operations further comprise determining that the set of nodes is included in an active partition for the cluster computing system and that the first computing node is excluded from the active partition with the set of nodes, wherein the host system is notified that the paths are unavailable during the time period based on determining that the first computing node is excluded from the active partition with the set of nodes and that the resources previously accessible via the paths are owned or managed by the first computing node rather than the at least some of the nodes.

18. The non-transitory machine-readable medium of claim 16, further comprising:

receiving a first command from the host system that identifies one of the paths and that specifies a first action involving a modification of a storage resource owned or managed by the first computing node;

determining that the first command is included in a set of media access commands and that the paths are to be advertised as unavailable in response to any command in the set of media access commands;

based on determining that the first command is included in the set of media access commands, responding to the first command with an error message indicating that the one of the paths is unavailable;

receiving a second command from the host system that identifies one of the paths and that specifies a second action involving a read-only operation for the storage resource owned or managed by the first computing node;

determining that the second command is included in a set of non-media access commands; and

based on determining that the second command is included in the set of non-media access commands, performing the second action in response to receiving the second command.

19. The non-transitory machine-readable medium of claim 15, wherein the operations further comprise determining that the host system requires access to the resource via the first path during the time period in which the disruption exists, wherein the cluster computing system notifies the host system that the first path is optimized during the time period based on determining that the host system requires the access to the resource.

20. The non-transitory machine-readable medium of claim 15, the operations further comprising:

rebooting the cluster system;

determining, during the reboot operation, that the second computing node is available to host systems prior to the first computing node; and

during an additional time period in which the second computing node is available prior to the first computing node, notifying at least one of the host systems that the second path is unavailable.