IBM Mainframe Discussion List <[email protected]> wrote on 04/27/2006 11:07:35 AM:
> No comments so far so let me stir the embers. Anyone have thoughts on > providing a generic DUMPMGR service where function and justification are > similar to RESMGR? Here is a couple good summary points but I want to > write a requirement for this and some of you do a much better job of > expressing yourselves than I do. So feel free to chip in an and I will > roll up something and submit it as a requirement. > > Wouldn't you like to not have ABEND-AID, DUMPMSTR, and friends mucking > about with SVC 51? Would you like to eliminate a hard USERMOD from the > IMS and other product installs? > > IBM should provide a service to dynamically add and delete dump managers > for abend formatting. This service would allow IBM program products or > vendor products which today require a USERMOD zap to IGC0805A IEAVADFM > to install and remove themselves without an outage. > > IBM should provide a service to dynamically add routines to pre- and > post- process in the existing SNAP and DUMP interfaces supported by SVC > 51. This service would allow IBM program products or vendor products > which today "hook" SVC 51 to co-exist and to avoid changing the flow of > control for abends in unexpected ways. This service could provide for > consistent exits/controls to bypass/suppress some of these routines for > specified tasks. We already have the general purpose dynamic exit service - CSVDYNEX. It would not be difficult to define a dynamic exit for SNAP/ABDUMP and call it from IEAVAD08 (which is the non-OCO module which calls the exits zapped into IEAVADMN). However, in SP2.1.5 (around 1983), a broadcast SSI call (for the "HELP" function code (decimal 48)) was added to IEAVAD08, and some comments in the module suggest that this was done for IMS. I talked to an IMS developer, and he was unaware of this SSI function (he has only been in IMS since 1985, and even I was unaware of it until today), but was willing to look into using it in a future release of IMS to remove the need for zapping IEAVADMN. It wouldn't be a bad idea to submit a requirement to IMS to this effect. As for the products which intercept SVC 51, it would likely be possible to provide a dynamic exits which could be used instead intercepting, but we (IBM) would need know a bit more about what the products require as far as what point in dump processing they need to be called, and what needs to be provided as parameters. So I would suggest that you get after the owners of the products which are intercepting SVC 51 and ask them present their interface requirements to IBM. Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY ---------------------------------------------------------------------- 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
