CDR Sync for ER

Hi All, Well, it's been a bit of a journey (thanks to those that I've reached out to privately for assistance), but I finally seem to have ER/GRID established between IFMX v12.10.FC4 on Solaris and IFMX v12.10.FC12 on RedHat! However, I came across something that has me absolutely baffled -- after a 'cdr sync replset ....' to populate the RedHat instance, a check of rowcounts shows that the target server has MORE rows than the master. How is that even possible? ----------------- select tabname, nrows, (select nrows from master@master_tcp:systables st where st.tabname = t.tabname) from target@target_tcp.systables t where tabid >= 100 and tabname not like "%ctr" order by 1 ; tabname address nrows 19702483.00000 orig nrows 19698516.00000 tabname address_typ_dom nrows 5.000000000000 orig nrows 5.000000000000 tabname attachment nrows 346530.0000000 orig nrows 346422.0000000 tabname batch_error nrows 450792.0000000 orig nrows 450780.0000000 ------ That is just a sampling. The smaller tables seem to be perfect, but tables with over 300K rows all have more target rows. Anyone have experience with this? Any ideas what can cause this? Thanks, Michael Hoffman

Hi Michael. Some ideas here of how this could be happening, but... 1) Could you please send one example of replicate definition on one of these big tables? 2) Do these unconsistent tables have some SERIAL column kind? 3) Have you defined your ER as a primary target architecture? Hope so.... Just for easing us helping you, ok? Best regards Alexandre Marini

Nevermind! I made a bad assumption: that statistics were updated on the Source side. I can see Art sorrowfully shaking his head. :-) Thanks, Michael