Cluster system management

• US 8,516,089 B2
• Filed: 05/21/2008
• Issued: 08/20/2013
• Est. Priority Date: 06/22/2007
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Johnsen, Bjorn Dag, Torudbakken, Ola, Bechtolsheim, Andreas
• Primary Examiner(s): LE, CHAU D
• Application Number:
  US12/124,686
• Publication Number:
  US 20080320117A1
• Time in Patent Office: 1,917 Days
• Field of Search: 715/734, 709/221
• US Class Current: 709/221
• CPC Class Codes: G06F 11/30 Monitoring
• CPC Class Codes: H04Q 1/035 Cooling of active equipments, e.g. air ducts
• CPC Class Codes: H04Q 1/06 Cable ducts or mountings specially adapted for exchange installations
• CPC Class Codes: H04Q 1/064 horizontal management arrangements
• CPC Class Codes: H04Q 1/066 arranged on the front side
• CPC Class Codes: H04Q 1/145 with switches arranged in a matrix configuration

Abstract

In a system having a cluster of system components interconnected by a cluster interconnect fabric, the system components include out-of-band management network interfaces to an out-of-band management network. The system is configured to use the cluster interconnect fabric and the out-of-band management network interfaces to determine the overall connectivity and status of the cluster.

Claims

1. A system comprising:

a cluster of system components interconnected by a cluster interconnect fabric, the system components further comprising a processor and out-of-band management network interfaces to an out-of-band management network;

wherein the system components in the cluster are mechanically coupled to other system components in the cluster through a plurality of cables, wherein each cable of the plurality of cables comprises a cable connector at each end point of the cable, wherein at least one cable connector in each cable is located at an interface of the cluster interconnect fabric, and wherein the at least one cable connector further includes a data storage module and a connector logic;

wherein the system is configured to use the cluster interconnect fabric and the out-of-band management network interfaces to determine an overall connectivity and status of the cluster; and

wherein determining the overall connectivity of the cluster involves determining a physical view that identifies system components in the cluster, wherein the physical view also identifies how the system components in the cluster are mechanically coupled to the other system components in the cluster through the plurality of cables by using at least one unique cable identity that is stored in the data storage module in a cable connector in each cable of the plurality of cables and received from the connector logic in the cable connector in each cable of the plurality of cables.

2. The system of claim 1, further configured to use any associated side-band channels to determine the overall connectivity and status of the cluster.

3. The system of claim 1, further configured to use available status information about at least one of cable presence and cable identity to determine the overall connectivity and status of the cluster.

4. The system of claim 1, wherein information defining the status of the cluster includes information about one or more or all of hardware types and revisions, firmware types and revisions and software types and revisions.

5. The system of claim 1, wherein a management network address is made available via an in-band path whereby in-band connectivity information identifies corresponding out-of-band management access points.

6. The system of claim 1, wherein components can communicate selectively via an in-band path or an out-of-band path.

7. The system of claim 1, comprising at least one IP point to point link configured by sending Direct Routed SMP packets via at least one InfiniBand link segment.

8. The system of claim 1, wherein a system component is configured to observe local connectivity and to correlate observed in-band connectivity with configuration rules defined for the component and/or the system for determining whether a required physical connectivity has been established.

9. The system of claim 1, wherein a system component is configured to observe information about types and revisions of one or more or all of software, firmware and hardware and to correlate this information with configuration rules defined for that component.

10. The system of claim 1, wherein a system component is configured to observe local connectivity and to correlate observed in-band connectivity with configuration rules defined for the component and/or the system for determining a desired logical configuration.

11. The system of claim 10, wherein the configuration rules are at least one of predefined and defined dynamically.

12. The system of claim 1, wherein the system is configured to perform distributed checking of one or more or all of connectivity, hardware types and revisions, firmware types and revisions and software types and revisions, relevant rules being distributed to individual system components that are operable to compare local status against the relevant rules.

13. The system of claim 1, wherein the system is configured to perform centralized checking of one or more or all of connectivity, hardware types and revisions, firmware types and revisions and software types and revisions, a central management server being operable to collect connectivity information from relevant system components and to compare the collected information against relevant rules.

14. The system of claim 1, wherein identifying how the system components in the cluster are mechanically coupled together involves identifying a set of ports for the system components, wherein each identified port is an endpoint for a physical link for the system components that are mechanically coupled together.

15. The system of claim 1, wherein the data storage module for each cable connector comprises at least one of a serial number, a type, and a length for the corresponding cable.

16. A system comprising:

at least two subsystems, each of the at least two subsystems includes a processor, wherein the at least two subsystems are interconnected by a cluster interconnect fabric, the system being operable to perform IP based management traffic transparently over a range of media types;

wherein the at least two subsystems are mechanically coupled to each other through a plurality of cables, wherein each cable of the plurality of cables comprises a cable connector at each end point of the cable, wherein at least one cable connector in each cable is located at an interface of the cluster interconnect fabric, and wherein the at least one cable connector further includes a data storage module and a connector logic;

wherein the system is configured to use the IP based management traffic to determine an overall connectivity and status of the subsystems; and

wherein determining the overall connectivity of the at least two subsystems involves determining a physical view that identifies the at least two subsystems, wherein the physical view also identifies how the at least two subsystems are mechanically coupled to each other through the plurality of cables by using at least one unique cable identity that is stored in the data storage module in the cable connector in each cable of the plurality of cables and received from the connector logic in the cable connector in each cable of the plurality of cables.

17. The system of claim 16, operable to use the physical view to provide fail-over scenarios between different media types between two chassis instances.

18. The system of claim 16, wherein a subsystem is operable to export a management IP address for the subsystem via an interface provided by another subsystem, wherein the other subsystem includes a processor.

19. A method of configuring a system comprising a cluster of system components interconnected by a cluster interconnect fabric, wherein the system components further comprise out-of-band management network interfaces to an out-of-band management network, the method comprising:

using the cluster interconnect fabric and the out-of-band management network interfaces to determine an overall connectivity and status of the cluster;

wherein the system components in the cluster are mechanically coupled to other system components in the cluster through a plurality of cables, wherein each cable of the plurality of cables comprises a cable connector at each end point of the cable, wherein at least one cable connector in each cable is located at an interface of the cluster interconnect fabric, and wherein the at least one cable connector further includes a data storage module and a connector logic; and

wherein determining the overall connectivity of the cluster involves determining a physical view that identifies system components in the cluster, wherein the physical view also identifies how the system components in the cluster are mechanically coupled to the other system components in the cluster through the plurality of cables by using at least one unique cable identity that is stored in the data storage module in the cable connector in each cable of the plurality of cables and received from the connector logic in the cable connector in each cable of the plurality of cables.

20. A non-transitory computer readable medium comprising program instruction operable to control one or more processors to configure a system comprising:

a cluster of system components interconnected by a cluster interconnect fabric;

wherein the system components in the cluster are mechanically coupled to other system components in the cluster through a plurality of cables, wherein each cable of the plurality of cables comprises a cable connector at each end point of the cable, wherein at least one cable connector in each cable is located at an interface of the cluster interconnect fabric, and wherein the at least one cable connector further includes a data storage module and a connector logic;

wherein the system components further comprise out-of-band management network interfaces to an out-of-band management network;

wherein configuring the system includes using the cluster interconnect fabric and the out-of-band management network interfaces to determine the overall connectivity and status of the cluster; and

wherein determining the overall connectivity of the cluster involves determining a physical view that identifies system components in the cluster, wherein the physical view also identifies how the system components in the cluster are mechanically coupled to the other system components in the cluster through the plurality of cables by using at least one unique cable identity that is stored in the data storage in a cable connector in each cable of the plurality of cables and received from the connector logic in the cable connector in each cable of the plurality of cables.