When the delay does not clear itself in a timely enough fashion, take appropriate action by trying to redistribute work or changing the workload. Keep in mind that the connector encountering the delay might be a victim rather than the cause of the delay (for example, when resource contention occurs on another system and holds up processing on the system that discovers the delay, or the structure is quiesced during rebuilds such as during the process of establishing or breaking duplexing). Use system commands when possible to aid in determining where the contention is occurring (for example, commands to display GRS waiting queue contention). As a last resort, you can terminate the connector.

If it is necessary to terminate the connector, first collect diagnostic information for any applications related to the connector that might also be encountering a delay. The following information should be collected: system log, system LOGREC symptom record information that contains diagnostic information provided along with this message, application log, and an appropriate dump. Use the application's instructions to end the connection encountering the delay.