Frequent CHASSIS OVER TEMPERATURE alerts on AFF-A900

Applies to

•    AFF-A900 
•    Network interface card (NIC) X91152A
•    QSFP X6589-R6 

Issue

• EMS log frequently triggers a Call Home for CHASSIS OVER TEMPERATURE alert.

  Tue Mar 31 13:21:31 +0900 [node-01: env_mgr: callhome.chassis.hitemp:error]: Call home for CHASSIS OVER TEMPERATURE

• The e3d port on the Slot 3 card, and its associated VLAN port, repeatedly cycle between link-up and link-down states:

  • Link up events:
    Wed Apr 01 09:57:11 +0900 [node-01: kernel: netif.linkUp:info]: Ethernet e3d: Link up. Wed Apr 01 09:57:11 +0900 [node-01: intr: netif.linkUp:info]: Ethernet e3d-641: Link up.
  • Link down events:
    Wed Apr 01 09:57:18 +0900 [node-01: kernel: netif.linkDown:info]: Ethernet e3d: Link down, check cable. Wed Apr 01 09:57:18 +0900 [node-01: intr: netif.linkDown:info]: Ethernet e3d-641: Link down, check cable.
  • Port not assigned to any broadcast domain:
    Wed Apr 01 09:57:39 +0900 [node-01: vifmgr: vifmgr.reach.unassn:notice]: Network port e3d on node node-01 is not assigned to any broadcast domain. No broadcast domains appear to be reachable from this port. Wed Apr 01 09:57:39 +0900 [node-01: vifmgr: vifmgr.reach.skipped:notice]: Network port e3d on node node-01 was not scanned for reachability because it was administratively or operationally down at the time of the scan.

• EMS log frequently shows chassisTemperature.warm alert:

  Wed Apr 01 09:58:06 +0900 [node-01: env_mgr: monitor.chassisTemperature.warm:alert]: Chassis temperature is too warm: Port temps at max and fans at max speed

• In PLATFORM-SENSORS, the sensor value for MLX Highest Temp occasionally shows a status of crit-high.

  • Example sensor output:
    MLX Highest Temp thermal crit-high 85 C 65