You can monitor the RFC resources on all application servers and thus find out the load incurred by parallel RFCs on a server.
You can also dynamically change the various quotas on all servers. For procedure for doing this is described in the section Dynamically Configuring RFC Quotas.
Execute the transaction SARFC.
You then see a list of all SAP servers with information for each server, stating whether resources for asynchronous RFCs are available at the time you executed the transaction (text Resources ok) . If no resources are available, a short text explaining the reason is displayed.
Note that this list only represents the situation at the exact moment that you executed the transaction. Choose Refresh to update the display.
This is a list of the possible results of the resource check with return values and meanings.
0: resources are available on the server. The text ‘Resources ok’ is displayed.
1: The resource check is deactivated (parameter rdisp/rfc_use_quotas is set to 0).
The other return values indicate that the server currently has no available resources for asynchronous RFCs.
The server may have no resources for one of the following reasons:
· The server is not running.
· The server has been wrongly configured.
· An unexpected error has occurred.
In these cases, the text appears on a red background. You can find out the exact cause of the problem (for example, where exactly the incorrect configuration is located) by briefly activating and deactivating the trace for the server (see below). See the dispatcher trace file dev_disp of the server for the required information.
The individual return values are described below. Each value has a text (appears on a yellow background).
2: The server has too few free dialog work processes.
4: the quota for the RFC communication channels (rdisp/rfc_max_comm_entries) has reached its maximum.
5: The local queue for asynchronous RFC responses is full. This queue retains the responses to asynchronous RFCs until they are sent back to the caller. You can increase the size of the queue by increasing the value of the parameter rdisp/max_arq.
7: the quota for a dialog work process occupied by an RFC user (rdisp/rfc_max_own_used_wp) is too small.
9: the quota for the RFC requests in the dialog queue (rdisp/rfc_max_queue) has reached its maximum.
10: An error occurred when the request queue length was being determined.
11: The quota for RFC logons to the server is too small. Increase the size of the value of the parameter rdisp/rfc_max_login.
12: the quota for RFC logons to the server (rdisp/rfc_max_login) has reached its maximum.
14: the quota for own RFC logons to the server (rdisp/rfc_max_own_login) has reached its maximum.
15: The server has been deactivated and cannot process any requests.
16: The server is being stopped.
17: The server has been stopped.
18: The server is being started.
19: The server has just been started and is in the initialization phase.
20: The server is in an unknown state.
Choose Goto → Activate Trace. The system then writes the detailed result of the check to the trace file of the dispatcher on the server in question. To read the file, you have to log on to the server in question. To do this, open transaction SM51, double-click on the server name, then choose Process → Trace → Dispatcher → Display File).
It may make sense to do this if, for example, a server constantly has too few resources, and you want to find out the exact cause.
Section System Configuration for Parallel RFCs describes how to configure your system.