On Fri, 19 Aug 2005 23:27:17 -0400, Bruce Black <[EMAIL PROTECTED]> wrote:
>... I gather that it only occurs if an application calls "resolver" >which is a TCP/IP service. Resolver calls USS which fails if the >calling application does not have the OMVS segment. ... Obviously, your >applications are at risk only if they use Resolver. >... But just about any TCP/IP function invokes Resolver code. The code is named "Resolver" because it performs the first steps in IP name resolution but the code also handles dynamic dataset and file allocation for TCP/IP (dynamic allocation of TCPDATA, etc.). >Prior to z/OS 1.6, TCP recognize the error but they changed some code in >1.6. Now it tries to pick up an address of something, and if the >failure occured that something may sometimes be a UCB address, leading >to the overlay. I have no idea if other storage could be overlaid by >the bug. >... I sort of doubt the code goes out of it's way to locate a UCB. :-) I suppose it could be following a chain of addrs starting in low storage that always lands it on a UCB. >... >if you are less that 1.6, you should not have the bug. ... And we have just finished our migration to 1.6. of course. I've opened a Tech Q&A incident on IBMLink asking for more detailed information. So far I've had the APAR text quoted to me ... twice. <sigh> Pat O'Keefe ---------------------------------------------------------------------- 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
