Mechanism for highly available rack management in rack scale environment

A rack management module (RMM), configured to be implemented in a rack including a plurality of pooled system drawers, each pooled system drawer associated with a respective power zone including one or more power sensors and one or more power control devices and a respective thermal zone including one or more thermal sensors and one or more thermal devices, the RMM comprising circuitry and logic configured to operate the RMM as master RMM and a slave RMM at separate points in time,

• wherein when the RMM is operating as a master RMM the RMM is configured to,
• manage the power and thermal zones for each of the plurality of pooled system drawers, including maintaining power zone and thermal zone state information;
• periodically synchronize power zone and thermal zone state information between the master RMM and another RMM operated as a slave RMM;
• and
• perform a fail-over operation under a master RMM role is transferred from the RMM to the slave RMM;
• and wherein when the RMM is operating as a slave RMM the RMM is configured to detect a fail-over condition for the master RMM, and in response thereto,
• implement the RMM as a new master RMM;
• and
• reset the master RMM for which the fail-over condition is detected.

A system comprising:

• a rack, including a plurality of pooled system drawers, each pooled system drawer associated with a respective power zone including one or more power sensors and one or more power control devices and a respective thermal zone including one or more thermal sensors and one or more thermal devices;
• first and second rack management modules (RMMs), linked in communication via an RMM-to-RMM link, each RMM linked in communication with at least one power sensor and the at least one power control device in each of the respective power zones and linked in communication with at least one thermal sensor and at least one thermal device in each respective thermal zone;
• wherein the system is further configured to.
• implement one of a first and second RMMs as an initial master RMM and other RMM as an initial slave RMM;
• monitor the power and thermal zones in the rack via a master RMM;
• periodically synchronize state information between the master RMM and the slave RMM;
• and
• detect a fail-over condition for the master RMM, and in response thereto,
• implement the initial slave RMM as a new master RMM;
• reset the initial master RMM;
• and
• implement the initial master RMM as a new slave RMM.