Configuration

Use the COORD_TIMEOUT parameter in the [CONFIG] section of the OpenSwitch server configuration file to specify the number of seconds OpenSwitch waits for a response before determining that a CM is not responding. The default COORD_TIMEOUT value is zero (0) seconds.

If a CM does not respond within the COORD_TIMEOUT period, the OpenSwitch server acquires the next CM ID from the internal list of inactive CMs. The previous CM ID becomes obsolete. If the CM becomes active again, it requests a new CM ID from the OpenSwitch. This occurs in the CM API and is transparent to applications.

If a connection is lost because OpenSwitch fails for some reason, the callback handler for COORD_CB_LOST is called for any CM that has lost the connection. When OpenSwitch restarts, if the CMs that lost their connection detect that the OpenSwitch has restarted, those CMs reconnect and are issued a new CM ID on a first-come basis.

You can use concurrent coordination modules in a mutually-aware configuration in OpenSwitch 15.1 and later.For example, you have coordination modules CM1 and CM2 connected to an OpenSwitch server (OSW1) in a mutually-aware configuration, and coordination modules CM3 and CM4 connected to the other mutually-aware OpenSwitch server (OSW2). If OSW1 is handling a failure, the clients connected to OSW1 send the failure notification to CM1 first and wait for any actions initiated by CM1 for the period you specify in COORD_TIMEOUT to end, while the clients connected to OSW2 also wait for any action initiated by OSW1. If CM1 does not respond to the failure within the COORD_TIMEOUT period, the clients connected to OSW1 send the failure notification to CM2 and wait for any actions initiated by CM2 for the COORD_TIMEOUT period. If there is no response from CM2, then OSW1 handles the failure. See Chapter 5, “Using the Configuration File,” in the OpenSwitch Administration Guide to set the value for COORD_TIMEOUT.