Incorrect switch port setting causes Solaris Dual-Machine Switchover exception

Current network environment:
Two Solaris 5.10 (Oracle) minicomputer run apps and DB, each with two computers, DB, array, app and other resources in the resource group, two small machines connected to two Cisco 2950 switches, and trunk aggregation is configured between the two switches.
One day found that the host app business interruption, troubleshooting and found that the firewall host failure caused the device to switch, log on two small machines can not ping the gateway (firewall floating IP), each small machine configured in the same subnet two IP address, When viewing ARP, it is found that the sequence of two IP addresses in the ARP list changes constantly, causing the host's source IP address to change, and the policy of viewing the address translation of the firewall is a one-to-two IP address mapping.
The final troubleshooting found that the port setting of the switch is the normal access port, the Solaris small machine switch on the time is very high, how much did not find the data, but after testing should be required within 1S, if not in the specified time, the success of the switchover will be repeated back and forth, resulting in network anomalies. The Cisco switch is designed to prevent switch loops from opening port protection by default, which results in a long port start-up time, typically around 10S, and does not require a dual-machine switchover. In switch port mode, the business returns to normal after the corresponding port is open in Port fast mode.

This article is from the "11101079" blog, please be sure to keep this source http://11111079.blog.51cto.com/11101079/1741523

Incorrect switch port setting causes Solaris Dual-Machine Switchover exception