SAP Note 27617 - Syslog Q0G, scratch area too small

Symptom:

Syslog message Q0G / request (type NOWP) cannot be processed

In the dispatcher trace dev_disp you can find:
*** ERROR =>AdExecBlk:scratch area too small (10436) [adxx.c 0435]
*** ERROR =>DpAdmHandle:AdExecBlk [dpxxdisp 5782]
*** ERROR =>DpRqNoWpHandle:DpAdmHandle [dpxxdisp 1587]
***LOG Q0G=>DpRqBadHandle, bad_req (NOWP) [dpxxdisp 2160]
or also something like
*** ERROR +> MsISnd2: len/2 200/9924 [msxxi.c 0625]

This problem particularly occurs if many work processes are defined. When you go to Alert Details in Transaction RZ03, the syslog is displayed. In Transaction AL01 it is displayed regularly without the user's intervention.
The syslog message may also be generated by external monitoring tools. Test whether the messages disappear when the monitoring tools are switched off.

Cause and prerequisites

When alert information is received, more information is generated than there is space in the message buffer if more than approximately 25 work processses are configured. The number also depends on the kernel release.

Solution
If a lot of work processes are required, it can be more favorable for performance to create two instances on the same host to avoid bottleneck situations at critical points.
A second option is to decrease the amount of transferred information for the alert details (only up to Release 3.0C).
In program RSALINC1 deactivate the following lines:

clear req_tbl.
req_tbl-opcode = ad_profile. "Instance_Name
req_tbl-buffer = 'INSTANCE_NAME'.
append req_tbl.

clear req_tbl. <-deactivate
req_tbl-opcode = ad_wpstat. "Workprozess-Status <-deactivate
append req_tbl. <-deactivate

clear req_tbl.
req_tbl-opcode = ad_queue. "Queue-Status
append req_tbl.

Up to and including Release 3.0F, the following line should be deactivated in the programs rsrzalm1, rsrzadm1:
perform get_slog.


Additional key words

CCMS, Alert, Alerts, ADM Work processes, ADM buffer AD_SCRATCHLEN