Master baseboard management controller election and replacement sub-system enabling decentralized resource management control

Patent metadata

Field	Value
Publication number	US-9842003-B2
Application number	US-201414509014-A
Country	US
Kind code	B2
Filing date	Oct 7, 2014
Priority date	Oct 7, 2014
Publication date	Dec 12, 2017
Grant date	Dec 12, 2017

Abstract

An information handling system and method of a master baseboard management controller election and replacement sub-system (MBMCERS) enables decentralized resource management control via the elected master baseboard management controller (BMC). The information handling system includes a plurality of server nodes, each having a BMC capable of controlling a plurality of shared common resources among the plurality of server nodes. Each BMC has a unique BMC identification. A master register stores BMC identification that has been elected as the master BMC to control the shared common resources. The master BMC relinquishes control of the shared common resources when the master register is placed in the reset state. When the master register is in the reset state, any one of the BMCs can elect to become a replacement master BMC.

First claim

What is claimed is:

1. An information handling system comprising:
a plurality of server nodes, wherein each of the plurality of server nodes comprises a baseboard management controller (BMC) capable of controlling shared common resources among the plurality of server nodes, and wherein each BMC has a unique BMC identification;
a programmable logic (PL) switch coupled to each BMC and to a plurality of shared common resources and which enables an identified master BMC to communicate with and control the shared common resources;
a master baseboard management controller election and replacement sub-system (MBMCERS) that enables decentralized resource management control via the identified master BMC, the MBMCERS comprising:
a master register which stores the unique identification of the BMC that is the identified master BMC to control shared common resources; and
a master heartbeat register that receives periodic updates from the identified master BMC;
a watchdog timer that tracks an elapsed time since the master BMC last updated the master heartbeat register and triggers a reset of the master register to a reset state when the watchdog timer expires without the identified master BMC writing to the master heartbeat register, wherein the master BMC relinquishes control of the shared common resources when the master register is placed in the reset state, and wherein when the master register is in the reset state, any one of the BMCs can elect to become a replacement master BMC.

2. The information handling system of claim 1 further comprising:
logic that supports the watchdog timer and master heartbeat register and resets the master register to a reset state in response to the watchdog timer expiring without the master BMC writing to the master heartbeat register.

3. The information handling system of claim 2 , wherein in response to a pre-established condition, the logic forces the master register into the reset state whereby the current master BMC is forced to relinquish the mastership of the shared common resources and the master register is forced into the reset state to enable an election of the replacement master BMC.

4. The information handling system of claim 2 , wherein after receiving feedback indicating the master register is in the reset state, a next BMC writes to the master register an identification of the next BMC to attempt to become the replacement master BMC.

5. The information handling system of claim 4 , wherein the next BMC takes over as the replacement master BMC after verifying that the identification stored in the master register is the identification of the next BMC.

6. The information handling system of claim 5 , further comprising control interconnects communicatively coupled among all the BMCs, the master register and the master heartbeat register.

7. The information handling system of claim 1 , wherein the master register includes an output that provides select signals to the PL switch to select one set of control/status lines to receive control/status signals from, wherein the PL switch selects control/status lines of the identified master BMC, and wherein the selected control/status lines of the master BMC control the plurality of shared common resources.

8. The information handling system of claim 1 , wherein each BMC that is not the master BMC is identified as a slave BMC and each slave BMC polls the master register to read the unique identification of the master BMC stored in the master register and rely on the master BMC to control the shared common resources.

9. The information handling system of claim 8 , wherein in response to a slave BMC requesting the master BMC relinquish the mastership, the master BMC ceases to update the master heartbeat register, which triggers the master register to enter the reset state to enable an election of the replacement master BMC.

10. The information handling system of claim 1 , wherein the reset state includes a predetermined reset identification that indicates that the current master BMC is not active or has relinquished the mastership and that the master register is in condition to be written to.

11. The information handling system of claim 1 , wherein the master BMC polls at least one slave BMC and reads a status of the at least one slave BMC that affects one or more operations of the shared common resources, and wherein the master BMC controls the shared common resources in correlation with the polled status of the at least one slave BMC.

12. The information handling system of claim 1 , wherein the shared common resources include shared infrastructures comprising one or more of sensors, fans, power supplies, keyboard video mouse (KVMs), shared storage, and shared input/output (I/O) subsystems.

13. The information handling system of claim 1 , further comprising a collision detection and prevention module wherein each of a plurality of slave BMCs consecutively reads the contents of the master BMC one at a time.

14. The information handling system of claim 1 , wherein the master register is accessible to be written to only during a reset condition.

15. In an information handling system having a plurality of server nodes, each having a baseboard management controller (BMC), and shared common resources, a method of enabling decentralized management and control of the shared common resources, the method comprising:
electing a master BMC from among multiple BMCs of the plurality of server nodes, wherein each BMC is capable of controlling shared common resources when elected as the master BMC, and wherein each BMC has a unique BMC identification;
storing the identification of the elected master BMC in a master register;
wherein the master register includes an output that provides select signals to a programmable logic (PL) switch coupled to each BMC and the shared common resources;
performing the management and control of the shared common resources via the elected master BMC and the PL switch; and
the elected master BMC periodically updating a master heartbeat register to indicate that the elected master BMC continues to provide the management and control of the shared common resources, wherein the periodic update of the master heartbeat register refreshes the watchdog timer to enable the master BMC to maintain control of the shared common resources.

16. The method of claim 15 , further comprising:
verifying that the master BMC has been elected;
selecting control/status lines of the master BMC, wherein the selected control/status lines of the master BMC provides signals which control the shared common resources;
controlling the shared common resources through the use of the control/status lines;
identifying each BMC that is not elected as the master BMC as a slave BMC; and
each slave BMC polling the master register to read the unique identification of the master BMC stored in the master register and relying on the master BMC to control the shared common resources.

17. The method of claim 16 , further comprising:
detecting when the BMC master relinquishes control of the shared common resources based on a reading of the master register.

18. The method of claim 17 , further comprising:
in response to detecting that a previous master BMC has relinquished the master role, electing a replacement master BMC by:
resetting the master register to a reset state that enables the master register to be written to and to store a replacement master BMC identification;
writing to the master register the identification of the slave BMC elected as the replacement master BMC; and
verifying that t

Assignees

• Dell Products Lp

Inventors

• Rao Balaji Bapu Gururaja
• Vancil Paul W
• Lambert Timothy M
• Stuewe John R

Classifications

• G06F2209/505

  Clust · CPC title

• H04L47/783

  Distributed allocation of resources, e.g. bandwidth brokers · CPC title

• G06F9/5061 Primary

  Partitioning or combining of resources · CPC title

• H04L47/826

  Involving periods of time · CPC title