Adding a new VLAN with lowest/highest VLAN ID causes the ClusterXL member to go "Down"
||R77.10, R77.20, R77.30
||Gaia, SecurePlatform 2.6
|Platform / Model
State of ClusterXL member changes to "Down" due to Critical Device "Interface Active Check" in the following scenario:
- Monitoring of the lowest and highest VLANs is enabled (default; fwha_monitor_low_high_vlans=1)
- A new VLAN is added on the ClusterXL member with VLAN ID, which is lower/higher than any existing VLAN ID (Note: It does not matter if the new VLAN is configured on the switch)
After fetching the new topology in SmartDashboard and installing the policy on the cluster, state of this ClusterXL member changes to an expected "Active" / "Standby".
Additional possible symptom: Only lowest VLAN is monitored on Bond interface, instead of lowest and highest:
- VLAN interfaces were defined on Bond interface
- Value of kernel parameter fwha_monitor_low_high_vlans is 1 (default; refer to sk92826)
- If CCP packets are not able to pass on the highest VLAN, Bond interface is not declared as "down" by the cluster member
- If CCP packets are not able to pass on the lowest VLAN, Bond interface is declared as "down" by the cluster member as expected
- Issue 01415023: ClusterXL member is updated about the newly added or removed VLAN immediately, which causes a mismatch between the local configuration and the cluster configuration installed by the policy.
- Issue 01987109: The kernel parameter fwha_update_monitored_vlans is not applied to Bond interfaces.
Note: To view this solution you need to