This is usually indicative of an Oracle error - i.e. a bug in Oracle is causing your application to terminate on a signal (eg SEGV). Have you had a look in the alert log and/or the 'udump' directory to see if there is a dump from your process ? If this is happening 'all of a sudden' I would be suspicious of a DB upgrade introducing a bug.
Steve Ian Harisay wrote: >Hi All, > >I've got a problem with code that has been running successfully for months. >I've talked to my DBAs about it and haven't found resolution. What I am doing >is getting the currval for a sequence after inserting a record. below is the >trace information and the oracle error being thrown. > >Thanks for any help I can get. > >-Ian > > > >the sql statement: > > SELECT mondo.status_order_id_seq.currval FROM dual > >My trace info: > > DBI 1.21-nothread dispatch trace level set to 3 > Note: perl is running without the recommended perl -w option > -> prepare for DBD::Oracle::db (DBI::db=HASH(0x886ef1c)~0x886b7d8 'SELECT >mondo.status_order_id_seq.currval FROM dual') > New DBI::st (for DBD::Oracle::st, parent=DBI::db=HASH(0x886b7d8), id=) > dbih_setup_handle(DBI::st=HASH(0x8ac309c)=>DBI::st=HASH(0x8ac757c), >DBD::Oracle::st, 8ac7588, Null!) > dbih_make_com(DBI::db=HASH(0x886b7d8), DBD::Oracle::st, 208) > dbd_st_prepare'd sql SELECT > dbd_describe SELECT (EXPLICIT, lb 80)... > fbh 1: 'CURRVAL' NULLable, otype 2-> 5, dbsize 22/134, p0.s0 > dbd_describe'd 1 columns (row bytes: 22 max, 4 est avg, cache: 858) > <- prepare= DBI::st=HASH(0x8ac309c) at Database.pm line 212 via >/home/mondo/lib/Feed/Database.pm line 205 > -> execute for DBD::Oracle::st (DBI::st=HASH(0x8ac309c)~0x8ac757c) > dbd_st_execute SELECT (out0, lob0)... > dbd_st_execute SELECT returned (SUCCESS, rpc0, fn4, out0) > <- execute= '0E0' at Database.pm line 227 via >/home/mondo/lib/Feed/Database.pm line 221 > >Oracle Error: > > ORA-03113: end-of-file on communication channel (DBD ERROR: OCIStmtExecute) >