Benjamin Smith <lists@xxxxxxxxxxxxxxxxxx> writes: > Aha, yep. Sorry: > Program received signal SIGSEGV, Segmentation fault. > 0x000000000043c82c in heap_modifytuple () > (gdb) bt > #0 0x000000000043c82c in heap_modifytuple () > #1 0x000000000043c8f5 in slot_getattr () > #2 0x000000000047a50a in FormIndexDatum () > #3 0x00000000004ebee3 in ExecInsertIndexTuples () > #4 0x00000000004e5265 in ExecutorRun () > #5 0x0000000000564312 in FreeQueryDesc () > #6 0x0000000000565287 in PortalRun () > #7 0x0000000000560f8b in pg_parse_query () > #8 0x0000000000562e0e in PostgresMain () > #9 0x000000000053d316 in ClosePostmasterPorts () > #10 0x000000000053ea59 in PostmasterMain () > #11 0x00000000005033c3 in main () Oh, so this is happening during index entry creation? (The reference to heap_modifytuple is misleading, but in a debug-symbol-free backend it's not so surprising.) This suddenly looks a whole lot like a known bug: http://archives.postgresql.org/pgsql-hackers/2005-11/msg01016.php Which version did you say you were using exactly? That bug is fixed in 8.1.1 ... regards, tom lane
