I have a product that, due to how the original programmers have done file I/O, makes it really difficult for me to easily control it in CICS. So, with that, I have some questions:
I issue a lot of STARTBR commands, usually without any ENDBR commands. What are the performance implications? My product has seven files or so, but can issue hundreds of STARTBR commands over them. I don't mean performance implications in terms of time. Does it lock up records? Consume memory? Create chains of control blocks? In short, would there be any reason that issuing hundreds of STARTBR commands would be bad? Similarly, is there any actual reason to concern myself with ENDBR? All of these files are read only, so there are no update locking issues. When the transaction ends, afaik, all memory, locks, browses, etc. are released as well. So what problems might I experience there? I'm just not seeing any problem with the methodology being used to read data, other than the obvious problem of having to issues zillions of reads to the same block (or same set of blocks.) I don't see any reason it would consume memory, cause deadlocks, cause CICS to hang or crash, or anything like that. Am I missing something? Would I need to provide more information? Thanks! David Logan ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
