The Address Manager Failover Monitoring feature lets Gateway monitor an Address Manager server replication cluster system, making sure that Gateway authenticates with the new primary Address Manager server in the cluster on a failover event. This makes sure that Gateway user sessions with the previous Primary server are maintained.
primary-address-manager when authenticating
their Gateway session (that is, they authenticated with a specific
Address Manager server instead of the "primary" cluster alias), a
failure of the Address Manager server will simply terminate the
session.The following sections describe what happens when users are logged in to Gateway during a failover event, as long as they logged in using
primary-address-manager as their Address Manager
server in the BAM Selection list.
Gateway user session behavior during failover when logged in through the UI
When a failover event occurs, active user sessions on the Gateway interface are terminated and users are redirected to the login page. Gateway displays a message indicating that an Address Manager failover is in progress and suggests either refreshing the page or trying again later.
Subsequent behavior depends on what credentials the user entered when logging in:
If the user logged in with credentials that match those configured in the Failover Monitoring workflow page, when failover completes they can simply refresh the page. Gateway will restore the user session and redirect the user back to the page they were on.
If the user logged in with credentials that do not match those configured in the Failover Monitoring page, when failover completes, the previous user session will no longer exist. The user must log in again to start a new session
Gateway user session behavior behavior when logged in through the API
When a failover event occurs, active sessions are terminated. During the failover,
attempts to log in to Gateway via the API endpoint
/rest_login will raise the following exception:
{
"code": 503,
"description": "Address Manager failover in progress. Refresh or try again later.",
"status": "FailoverProcessing",
"time": "09:16:11.673283"
}Subsequent behavior depends on what credentials were entered to authenticate the user via the API:
After failover completes, if the credentials matched those configured in the Failover Monitoring workflow page, the user session is automatically restored and you can continue using the same authorization token for API calls.
After failover completes, if the credentials did not match those configured in the Failover Monitoring workflow page, the previous user session no longer exists. The user must log in again to create a new user session.