-- It appears that the "uniqueness" applied by the programmable field is limited to the sysplex. It has nothing to do with an LPAR (aside from that z/OS runs within one). There is no uniqueness across sysplexes. -- It appears that the current implementation uses the "system number" managed by XCF (so the limit is the number of systems within a sysplex). -- It appears that there is no guarantee of uniqueness until such time as that system number is determined (which is early in IPL but is later than when it is allowed to issue STCKE).
Perhaps someone could share how they can make productive use of this level of uniqueness (and, in effect, why they care). That would be good rationale for understanding the importance of documenting whatever guarantee of uniqueness z/OS implements. A timestamp "later" than another timestamp (when looking at all the bits, including the programmable field) does not mean that the event correlated with that later time happened after the even correlated with the earlier time, unless there is sysplex-wide serialization (as might be held when writing to the couple data set, but I'd think it likely that the programmable field is not needed for uniqueness in that case) Peter Relson z/OS Core Technology Design ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
