Multiple nodes or chassis overheating

Use the information in this section to troubleshoot multiple nodes or chassis that might be overheating.

Action

Complete the following steps until the problem is solved:

  1. Verify that the room where the system is located is properly cooled.
  2. Check the chassis air filters for dust accumulation and replace if necessary (see Replacing the filter media for more information).
  3. Check the event log on the Chassis Management Module for rising temperature events and fan module events. If there are no rising temperature or fan module events, the node is running within normal operating temperatures.
    Note: Some variation in operating temperature is to be expected.
  4. Make sure that no stickers or labels have been placed on the perforated section of the compute node front panel. This is an air inlet to the node, stickers or labels will hinder airflow leading to high component temperatures.
  5. If a fan zone is partially populated with compute nodes, make sure that the number of installed 80 mm fans is sufficient to cool the zone. Two or less nodes require only 2 fans, three to four nodes require 3 fans, five to seven nodes require all 4 fans (see Installing components for more information).