Testing failure safety
To verify that failure safety is working correctly, perform the following steps:
- Log in via the nevisProxy.
- See the log file on both instances as follows and check on which instance your request has actually authenticated —the AuthAudit trace group shows login/logout/stepup/stepdown/unlock requests:
# grep AuthAudit /var/opt/nevisauth/default/log/audit.log
- Shut down the instance that performed the login.
- Log out via the nevisProxy. The call should now be delegated to the fallback instance.
- The fallback instance performs the logout and tries to inform the master instance about the logout —this sync call will now fail, as the master server is down. Verify this by checking the log file /var/opt/nevisauth/default/log/esauth4sv.log of the fallback instance for "ERROR" entries containing the text "sync_kill to failsafe esauth failed".