OT: Re: ER error defining template

OK, maybe not totally off topic. Last week I've had somewhat similar situation. Here are the details. ER setup in primary-target scenario. HP-UX B.11.23 ia64 + IDS 10.00.FC4 Solaris 9 (Sparc) + IDS 10.00.FC4X5 Customer tried to "cdr check" the replicates (checksum routines successfully registered before that). But, on some tables it fails with the next message: $ cdr check replicate -m g_hzzotest -r repl_idenos_1_10 g_epra -vError at line:4273 - Select index parts for zoroh.idenos ***ABORTING*** Segmentation fault (core dumped) Same error with "cdr sync" on the same replicate. It doesn't matter if the command is run on primary instance or the target. I tried to reproduce the behavior on Linux (on Intel x86), but everything's OK there, I can't reproduce it (I don't have HP or Sun machine to play with). It always reproduces on HP-UX and Solaris. There are other tables in the same database which fail the same way too. The rest of the tables succeed. And now the similarity with this post's problem: tables that fail have unique indexes created before PK constraints that use them are added. For example, the schema (only the relevant part) for table in question (idenos) is: create table "zoroh".idenos ( jmbg char(13) not null , zdrmb char(9), ... ); create unique index "informix".idenos_3 on "zoroh".idenos (zdrmb) using btree ; alter table "zoroh".idenos add constraint primary key (zdrmb); Laurie's table also has user-defined unique index which gets used when PK is added. Maybe that could be the cause. Thanks P.S. I have those core dumps available if needed.