On Fri, 21 Jul 2023 16:46:26 -0500, Wendell Lovewell wrote: >Hello Listers, > >I have a question about how additional extents for a dataset are chained >together for files on EAV volumes. > >If I understand correctly: >- The format 8 record has slots to keep track of the first 3 extents of a >dataset (DS1EXT1/DS1EXT2/DS2EXT3). > >- If there are more than 3 extents, the DS1PTRDS will point to a format 9 >record
For format-8, the DS1PTRDS will always point to a format 9 record. In DFSMSfp Advanced Services, describing DS1PTRDS, it says "In a format-1 DSCB this can be a pointer (CCHHR) to a format-2 or format-3 DSCB or be zero. In a format-8 DSCB this always is the CCHHR of a format-9 DSCB." Notice that for format-8 it doesn't say "or zero". Therefore every F8 has an F9. I have seen this in a VTOC dump, even for data sets with 1 extent. In that case the F9 was all hex 00's except the one hex F9 byte > >- The format 9 record contains pointers in the DS9F3 area to point to up to 10 >format 3 records > >- Each format 3 record can point to up to 10 extents 13 extents. > >- The DS9PTRDS description contains the text: > > "FORWARD CHAIN POINTER (CCHHR) TO THE NEXT FORMAT 9 DSCB >-> OR TO THE FIRST FORMAT 3 DSCB IF DATA SET HAS MORE THAN 3 EXTENTS. >OTHERWISE ZERO." > >There wouldn't be a format 9 record if there weren't more than 3 extents. So >I think the comment on the DS9PTRDS field is confusing. > There is always at least one format 9 for each format 8. >It seems there are 2 ways extents could be chained via the DSxPTRDS fields: > >DS9PTRDS -> the next FMT 9 record, whose DS9PTRDS -> another FMT 9 record, >or >DS9PTRDS -> the next FMT 3 record, whose DS3PTRDS -> another FMT 3 record > I don't think there are ever more than two FMT 9 records, but new F9 subtypes might be added some day. Two FMT 9 records are more than enough for the maximum 255 extents. I'm not sure if there is still a limit of 16 extents, even though the design allows more. >My question is: >If there are more than 13 extents, will the DS9PTRDS always point to another >format 9 record? Or are there circumstances when it will point to a format 3 >record? ITYM more than 16 extents. It would take more than 133 extents (3 + 10 * 13) to need a second F9. > >Or, maybe restated: >What will the DS3PTRDS field contain when the format 3 record has been pointed >to by a format 9 record, when there is another format 3 record? > >The graphics in >https://www.ibm.com/docs/en/zos/2.5.0?topic=components-data-set-control-block-dscb-types > show DS3PTRDS->another format 3 for non-EAV records, but they don't show >multiple format 9 records. > Bill ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
