The symptom of the fault:
The customer powered off the device at a site, and then powered on the device again, found the service was abnormal. While checking the fault, he/she found the ESC node’s DRDB was in the dual-active status.
1. DRDB’s status is abnormal, firstly we guess the network, in this case, the management nodes can “ping” through each other, so we exclude this probability.
2. We can’t “ping” through the gateway at the management node.
3. Login in the management convergence switch to check, we find there hasn’t configured the management plane gateway, we guess the problem is caused by the gateway configuration data has been lost after the switch restarting.
4. Configure the corresponding gateway IP on the management convergence switch again, restart the management node, and then the DRDB’s status is normal.
1. When the management nodes can “ping” through the gateway, but management nodes can’t “ping” through each other, the DRDB will become to be dual-active, this condition has been described in an older case.
2. When the management nodes can’t “ping” through the gateway (no matter whether the management nodes can “ping” through each other), restart the management node, the software dog will check if the gateway can “ping” through firstly, if it fails for three times, the node’s DRDB will upgrade to be master status compulsively, in this condition, the management node’s DRDB may be in the dual-active status.