According to Test DBLOG Entire System Server counts DB-calls when e.g. performing PROCESS WRITE-FILE. Can this really be true? And what happens to the written records when doing an ET?
We have made a module that writes log records to a dataset, trying to monitor our Adabas calls, but ASY seems to interfear with these calls, affecting the 1009 counter…
calls to Entire System Server are in fact technically ADABAS calls, which are routed to Entire System Server. in case of WRITE-FILE these are S1 calls. There no real transaction handling in Entire System Server since you can have for example the Program to send an Operator Command to the Console, and if this has been done, you cannot say, Backout Transaction, the Operator Command was not good.
Regarding 1009 Error, pls tell us, how many records are written, and if the parameter MADIO = 0 was set in NAT.
I’ll ask my collegue tomorrow, when he is back on the office.
I know he was testing our standard procedure for controlling 1009’s and 1029’s by doing a screen IO before those errors appear. He made a log dataset to be able to trace the Adabas activities without affecting them - so we thought anyway 
.
The calls to the database have codes in the second byte of the control block. If the call is a standard call to ADABAS from Natural, it contains “N”, if it is a (PROCESS)call sent to Entire System Server, it has “P”. Your logging/monitoring program can filter them based on the contents of the indicator byte.