There is a third option you did not mention. That is to deregister the ENF listener but hang onto the CSA storage for reuse. If recording is requested again, reuse the same CSA storage. This eliminates both the cycles used by the unnecessary listener and the possible fragmentation of CSA storage. It does tie up the CSA storage until the server terminates however.
Chuck Arney Arney Computer Systems -----Original Message----- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@bama.ua.edu] On Behalf Of Phil Smith Sent: Thursday, June 07, 2012 4:20 PM To: IBM-MAIN@bama.ua.edu Subject: ENF Listener usage We have a long-running Started Task that controls the use of an ENF listener for SMF interval record collection with an operator command. The ENF listener requires use of CSA storage. When the Started Task receives an operator command to start SMF interval processing it allocates CSA storage, registers the ENF listener, and WAITs for interval expiration. The question is: what should it do when the operator command says to stop SMF interval processing? The options seem to be: 1) Deregister the ENF listener and free the CSA storage, meaning an operator can switch SMF interval processing on and off, causing CSA storage to be allocated and deallocated, possibly fragmenting CSA storage. But when we aren't collecting SMF interval data, we won't have an ENF listener registered. 2) Leave the ENF listener registered, but stop writing records when the interval expires. Don't stop the ENF listener and free CSA until the Started Task terminates (or possibly a special operator command like /f stcname,SMF STOP). This approach is easier on CSA allocations for the case where the operator is switching SMF interval collection on and off. But it leaves an ENF listener registered when one is not needed. Which approach have you seen? Which is "better" (and why)? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: INFO IBM-MAIN