segmentation violation - more info

(Apologies if you've all seen this mail already, but my news server is playing up) many thanks to contributions received on this ( esp. Mats, Harold, Josef), they have all allowed me to progress this. But particularly thanks for the AFDEBUG tip, it has allowed me to trap the offending sql statement, and here it is: Current SQL statement : SELECT cd.call_data_t,min(handler_id), min(state), min(last_pri), min(curr_svc), min(cli_sub), min(tli_sub), min(ld_first.cli), min(ld_first.tli), min(ld_first.setup_utc), min(ld_first.answer_utc), min(ld_first.last_utc), count(ld.call_data_t) FROM call_data cd, outer (leg_data ld, leg_data ld_first) WHERE cd.call_data_t =ld.call_data_t AND ld.call_data_t = ld_first.call_data_t AND ld_first.index = (SELECT min(index) FROM leg_data WHERE call_data_t = ld.call_data_t ) GROUP BY 1 ORDER BY 1 ASC Last parsed SQL statement : SELECT host_name,ipl_proc,pkg_lbl,proc_lbl,ipl_recno,command,args,wait_rep, max_tries,restrt_slp,restrt_cyc,recvry_pln,proc_offset,pid,retry_cnt,state_tm,state, report_it FROM act_prcs WHERE command matches '*fsdlexe *' OR command matches '*callctrl *' I assume the current statement is the relevant one. Can anyone tell me if this would generate a hash table? If so, any suggestions on reworking it so that it doesn't? When I try onstat -g ses 2550 shmem.9dd4f77.0 or onstat -g ath shmem.9dd4f77.0I get an error returned, but I wasn't able to catch this and now the evidence is gone, but I think the error was to do with shmat() many thanks Pete P.S. how should I allow the system to carry on when it says *oninit 2282 waiting for debugger*?