Users Guide

wait for eight-and-a-half minutes to determine if redundancy exists. Ensure that the redundancy state is healthy before initiating any
failover owing to the health changes.
NOTE: CMC does not perform a pre-check of these conditions when you change the redundancy policy to or from Grid
redundancy. So, configuring the redundancy policy may immediately result in redundancy lost or a regained condition.
Power Management After PSU Failure
When an insufficient-power event occurs, such as a PSU failure, CMC reduces power supply to the servers . After reducing the power,
CMC reevaluates the power needs of the chassis. If power requirements are still not met, CMC turns off the lower priority servers.
However, this is done on the basis of power redundancy policy that you set on your CMC. A redundant server can tolerate the loss of
power without impacting the performance of the servers.
Power for higher priority servers is restored incrementally, while power needs remain within the power budget. To set the redundancy
policy, see Configuring Power Budget and Redundancy.
Power Management After Removing PSU
CMC may begin conserving power when you remove a PSU or a PSU AC cord. CMC decreases power to the lower priority servers until
power allocation is supported by the remaining PSUs in the chassis. If you remove more than one PSU, CMC again evaluates the power
requirements when the second PSU is removed to determine the firmware response. If power requirements are still not met, CMC may
turn off the low-priority servers.
Limits
• CMC does not support automated power-down of a low-priority server to allow turning-on of a higher priority server; however, you
can perform user-initiated turn-offs.
• Changes to the PSU redundancy policy are limited by the number of PSUs in the chassis. You can select any of the two PSU
redundancy configuration settings listed in Default Redundancy Configuration.
New Server Engagement Policy
If a new server that is turned on exceeds the power available for the chassis, CMC may decrease the power to the low-priority servers.
This could happen if the administrator has configured a power limit for the chassis that is below what would be required for full power
allocation to the servers, or if insufficient power is available in case higher power requirements by all servers in the chassis. If enough
power cannot be freed by reducing the allocated power of the low-priority servers, the new server is not allowed to turn on.
This occurs if the administrator had configured power limit for the chassis lower than the full power allocation to the servers or if
insufficient power is available to servers requiring high power.
The following table provides the actions taken by CMC, when a new server is powered on in the scenario described earlier.
Table 38. CMC Response When a Server Power-On is Attempted
Worst Case Power is Available CMC Response Server Power On
Yes No power conservation is required Allowed
No Perform power conservation:
• Power required for new server is available
• Power required for new server is not available
Allowed
Not Allowed
If a PSU stops functioning, it results in a non-critical health state and a PSU failure event is generated. The removal of a PSU results in a
PSU removal event.
If either event results in a loss of redundancy, on the basis of power allocations, a loss of redundancy event is generated.
If the subsequent power capacity or the user power capacity is greater than the server allocations, servers have degraded performance
or, in an extrement case, servers may be turned off. Both conditions are in reverse-priority order; that is, the low-priority servers are
turned off first.
The following table provides the firmware response to a PSU turnoff, or removal as it applies to various PSU redundancy configurations.
Managing and Monitoring Power
165