today, a table that gets hit a lot ran out of space when we were in the middle of moderately heavy testing. We started getting: Error Type: oci.error Error Value: (1043, 'ORA-01043: user side memory corruption [], [], [], []\012', "SELECT d.name as name,\012 d.value as value\012 FROM Session_data d, Sessions s\012WHERE \012s.session_name = 'mbC5HGphnQvCdFa/yKfXYg'\012AND\012s.session_id = d.session_id \012AND d.name = '__sessac'") followed by the DCOracle cursors being unusable. Is it possible that there's either a bug in the code in DCOracle that handles errors that isn't multithread safe? Or in the Oracle client libs? The above error means that the area of memory for the cursor/host bit in the client application got stomped. This is on Redhat Linux 5.2, Oracle 8.0.5.0.0 Anthony