natural rpc server behavior

Hi peope. I’m using natural (6.3) rpc server under linux red-hat and abadas database (6.1).
Rpc Servers are called from a java application using Entire X 8.
I see that when an rpc server is called, it start his works and do ( for example ) a lot of find or read to the adabas database.
According to time-out settings on Entire X, after 1 minute, the conversation is broken and Entire X send an EXX.74 error to the java application.
My problem is that Rpc server still runs on the linux machine and it consumes a lot of cpu without reason.
Limit natural cycle using madio or maxcl doesn’t works on rpc server becouse it sum all the adabas calls since his startup.
Is there a way to avoid this behavior?

I haven’t searched for ways of checking Madio etc…
But from the description you give there is certainly reason to believe that one or more of the underlying subprog(s) have problems in their loops.
Before spending too much time on the otherwise noteworthy search for parameters I would dig into the read/find loops !
Perhaps post the source here ?
Finn

:slight_smile: you’re right! of course. We are doing an optimization works…but it’s a long project, so I’m looking for a work-around…

See your point !

Actually - and now I did check the documentation - madio and maxcl are (supposed to be) reset for every callnat !
Have a look here:
http://techcommunity.softwareag.com/ecosystem/documentation/natural/nat6312unx/rpc/intro_rpc.htm#intro_rpc_limit_handling

Finn