Professional Documents
Culture Documents
Liang
=====Note 1118297===========
Symptom
You are using the Exchange Infrastructure and you are processing asynchronous messages. These messages are
scheduled in queues that are no longer processed. The inbound queue of the qRFC Monitor (transaction SMQ2),
for example, contains entries with the status "Running". However, you recognize in the process overview
(transaction SM50) that there is no active work process.
Other terms
QRFC, SXMS_ASYNC_EXEC, inbound, running, resource
Solution
The asynchronous messages are scheduled in the queues.
Check the following points of your middleware.
Queue configuration:
If the resources are configured for TRFC/QRFC, note the common usage of system resources. For further
details, see Notes
74141 Resource management for tRFC and aRFC
527481 tRFC or qRFC calls are not processed
Queue status:
Transaction SMQ2 contains queue entries with the following status
"Running"
If the J2EE engine is downloaded in production mode with the mapping engine or recipient Adapter
Framework, processing QRFC entries may not recognize this and remain in the "Running" status. A symptom
can be, for example, that the RFC destination "AI_RUNTIME_JCOSERVER" is no longer registered. You can
find this information in the Gateway Monitor (transaction SMGW).
Otherwise, the system may reach this status if the application server is downloaded in a controlled manner, and
long-running processes can no longer terminate.
The system no longer processes any message. In transaction SMQR, you can activate the queues manually
again. You can also schedule the report RSQIWKEX, which automatically restarts the queues. For further
information see Notes
620633 Status RUNNING in SMQ2 with XI queues
864333 RSQOWKEX & RSQIWKEX start Running(Executed) queues
"Sysfail"
A queue entry with the status "Sysfail" can have two possible causes.
1. Further processing cannot be carried out on the process for technical reasons (for example, the program
terminates with a dump or a system (mapping or recipient system cannot be reached).
2. Queue processing should be stopped, as the message cannot be processed, due to a configuration error.
Processing is not permitted for the subsequent messages, as the queue sequence displays the message sequence.
You can use the report RSXMB_RESTART_MESSAGES to restart the message processing again. See Note
813029 Automatic processing of failed XI messages
Oracle database:
When using an Oracle database see Note
742950 Performance affected on Oracle DB with supplement 11
if you discover performance problems when processing queue entries. This improves the scheduling process
only, and not the processing time of a queue entry.
See also Note
1020260 Delivering Oracle statistics
The XI log guarantees the uniqueness of the transfer between the client and server. A restart does not trigger a
double execution of messages. Note that a synchronization of the reorganization between client and server is
guaranteed. The EO recognition is executed with the message ID. The retention time determines the amount of
time this information is retained in the system for.
The individual messages are saved to the database in the system with a key from the message ID and version.
This prevents a parallel execution of the same message. A queue entry can be executed in parallel using the
report RSQIWKEX or by carrying out a manual start from the queue. We recommend to set the parameter
LOCK_MESSAGE (default value is active "1") of the category "RUNTIME" in the XI configuration to the
value '0'. More information is contained in Note
1058915 Outbound queue remains in status 'SYSFAIL'