A Failover Fails Due To No Gateway Device in a Cluster

Publication Date:  2015-08-28 Views:  470 Downloads:  0
Issue Description
Product and version information:

VTL6000 V100R003
VTL6900 V100R005

The VTL6900-HAS (two SIR engines) is deployed at the current site. The specific network shows as follows.



 

1.  Devices communicate with each other through an Ethernet switch. This switch is a layer-2 switch whose IP address is 172.30.92.48 and gateway address is 172.30.192.1.

2.  During a failover, after all services on a faulty node are recovered, this node cannot enter the Ready state. Run the sms command. The command output is as follows:




3.  Run the tail -f /var/log/messages command to check system logs. In the command output, a message is displayed indicating that the faulty node cannot communicate with the network.



4.  Check event logs on the VTL Console. The error information is displayed like this: Server A can not communicate with Server B.
Handling Process
1.  Add a gateway device to the onsite network.

2.  Perform the failover again. The problem is solved.

Root Cause
1.  The onsite switch is a layer-2 switch. Although it ensures the normal communication between two nodes in the cluster, it works at the data link layer and cannot serve as a gateway device.

2.  After the faulty node is recovered, the ping 172.30.192.1 command is executed on the node to determine whether it is recovered. Because no gateway device exists on the network, the node cannot obtain the recovery information. Therefore, this node cannot enter the Ready state.
Suggestions
When you configure the VTL6900 cluster, determine whether the onsite network meets cluster network requirements and whether gateway devices exist on the network. If no gateway device exists on the onsite network, communicate with customers at the site to solve the problem.

END