How RM and the resource structure impact IMS activities
This topic provides information on how RM and the resource structure impact some IMS activities.
Conversations
Conversation status is shared within the IMSplex, and so the conversation's IDs must be unique for individual dynamic users or static nodes. To uniquely identify a conversation, qualify the conversation with the associated input LTERM.
To display conversation status, use the /DISPLAY CONVERSATION command for local and RM conversation information.
Initialization
IMS enforces consistency of single signon, SGN=M, in the startup parameters when RM and a resource structure are active. The first IMS active in the IMSplex sets the value for the entire IMSplex. If an IMS joins an IMSplex with a different value for single signon, IMS issues a warning message and uses the IMSplex value. You can change the specification when all the IMS systems leave and then rejoin the IMSplex (warm start). You can override the value by using the /NRE or /ERE commands.
During a warm start, when RM is active, the global signon value comes from the startup parameter of the first IMS to join the IMSplex. The signon value comes from the checkpoint log records when RM is inactive during the warm start.
Shutdown
When an IMS system shuts down, all nonrecoverable status owned by the system is deleted from the resource structure. SCI notifies the other IMS systems that the IMS is leaving because of either a normal or abnormal shutdown.
When an IMS system shuts down abnormally, another IMS in the IMSplex becomes the cleanup IMS. This IMS requests a list of resources without significant status. If a resource does have significant status and the resource recovery mode is GLOBAL, the owner is cleared and the resource becomes available to other systems. If the recovery mode is LOCAL, affinity for the failed system is enforced.
Checkpoint
Because terminal and user status are maintained by RM, if active, local DC control blocks must be cleaned up and deleted during IMS checkpoint if they are inactive. Checkpoint deletes local status in local blocks with GLOBAL status recovery mode. Checkpoint maintains status in local blocks with LOCAL status recovery mode.
These recovery modes are treated the same way during a simple checkpoint after a warm or emergency restart. However, if a resource is in LOCAL status recovery mode, DFSSGNX0 and DFSLGNX0 can request that another IMS steal the resource that is owned by the failed IMS system. In that case, the restarting IMS deletes the local status during its restart.