rhunt01
Well Known Member
We recently went live with JDE9. One of the things that happens during the table conversion is that many of the tables are dropped, recreated, and data is copied back in. Since many of the tables have been recreated, I'm finding that F0902, F0911, F0901, F03B11, etc. are all now set to reuse deleted records instead of the alternative (which is use arrival sequence (all new goes to the end of the table) and records are "marked" as deleted).
After a decade of XE I have all the maintenance jobs prebuilt that I need to manage auotmated reorg physical file member maintenance each weekend for files that pass a deleted percent threshold.
In fact, I kind of liked this approach. I felt like the latest data always stayed close together in the table/file, minimal index seeks needs to occur to find table space to insert into, and I kept indexes dense with RGZPGM as needed.
In the MS SQL world I can speak to exactly how table structure, index approaches, file definitions (like reuse or not) translate to physical query implementations, data density on disk, disk hardware use, etc.
Anytime I try to have this discussion with an AS400 lifer or IBM I typically get quoted on something about single level storage and how trying to manage AS400 dasd to minimize disk latency, head travel etc. are innefective and can hurt AS400 performance.
So, looking for any AS400 uber nerds that have an opinion on whether or not I should set this back to the way XE had it on the AS400.
Thanks
Ryan
After a decade of XE I have all the maintenance jobs prebuilt that I need to manage auotmated reorg physical file member maintenance each weekend for files that pass a deleted percent threshold.
In fact, I kind of liked this approach. I felt like the latest data always stayed close together in the table/file, minimal index seeks needs to occur to find table space to insert into, and I kept indexes dense with RGZPGM as needed.
In the MS SQL world I can speak to exactly how table structure, index approaches, file definitions (like reuse or not) translate to physical query implementations, data density on disk, disk hardware use, etc.
Anytime I try to have this discussion with an AS400 lifer or IBM I typically get quoted on something about single level storage and how trying to manage AS400 dasd to minimize disk latency, head travel etc. are innefective and can hurt AS400 performance.
So, looking for any AS400 uber nerds that have an opinion on whether or not I should set this back to the way XE had it on the AS400.
Thanks
Ryan