SAP Note 26966 - Background jobs do not start when transporting

Symptom:

Transports which were started with the transport control program tp do not exit. In the tp log file SLOG, the following messages are periodically written:
Background job not running properly. Function: ... Jobcount: ... Status: S
Please check the system. Use Transactions SM21, SM37, SM50.

Other terms

RDDIMPDP, RDDMASGL, RDDDIS0L, RDDGEN0L, RDDDIC0L, RDDDIC1L, RDDDIC3L, RDDVERSE, RDDVERSL, RDDEXECL
TR496

Reason and Prerequisites

When executing transports, programs are started within the SAP System. These programs are scheduled as background jobs.

• Background jobs are started as separate processes (type BTC in Transaction SM50). If all such background job process slots are occupied no new background jobs can be started. These jobs stay locked until there are process slots free.

• Sporadically it may also happen that the scheduled batch jobs hang in status "ready" or "scheduled" despite of there being free BTC processes. So far we have no plausible explanation to that phenomenon.
  

Solution

The solution to this depends on whether the background processes are just temporarily over-used, or whether a long-term bottleneck exists.
If the capacity overload is only caused by one or two very lengthy background jobs that are not regularly scheduled, a little patience is the best remedy. After these jobs terminate, the transport programs will also exit correctly.
The number of background processes or application servers which are used to process background jobs should be increased if the capacity overload for background processing is long-term. Details for this can be found in the "BC System Administration" manual, in chapter 4 (Configure Background Processing).
The Transactions SM50 (process overview of an application server), SM51 (overview of the application servers) and SM37 (overview of background jobs) can be used for analysis. Details for this can also be found in the "BC System Administration" manual in chapter 4.

If the batch jobs do not start despite of free BTC processes, you can work around the problem by scheduling those "hanging" jobs again using Transaction SM37 (Job -> Repeat scheduling -> Immediate). It is important that you repeat the job scheduling as user DDIC (the user under which the original job had been scheduled).
According to our experience, the batch system will work without problems afterwards.