Bug 4074710 10.1.0.3.0 RDBMS 10.1.0.3.0 PI/DISTRIB PRODID-5 PORTID-23 ORA-7445 (See ORA-7445.ora-code.com) Abstract: ORA-7445 (See ORA-7445.ora-code.com) [DDFNET7RESTORE()+320] DURING REFRESH OF INLINE MVIEW Severity: 2,Severe Loss of Service Fixed In Ver: 10.2.0
eos (end of section) . POTENTIAL SOLUTION(S) ====================== If the master table is remote, set event 10336 as a workaround for Bug 3278020.
If the master table is local, you seem to be hitting bug 4074710, which is fixed in 10.2. We have a backport for this bug for Solaris 64-bit. It is available from Metalink->Patches:
Patch 4074710 Description ORA-07445 (See ORA-07445.ora-code.com) [DDFNET7RESTORE()+320] DURING REFRESH OF INLINE MVIEW Product RDBMS Server Release Oracle 10.1.0.3 Platform or Language Last Updated 20-FEB-2005 Size 25K (26580 bytes)
eos (end of section) @(protected)
-- --Original Message-- -- From: oracle-l-bounce@(protected) [mailto:oracle-l-bounce@(protected)] On Behalf Of Michael Fontana Sent: Thursday, March 10, 2005 10:21 AM To: oracle-l@(protected) Subject: Oracle 10g Mview Refresh Performance Isses
We are currently testing Oracle 10g performance and functionality, and have run into severe performance degredation issues in performing fast refreshes of materialized views. In some cases, the refreshes fail with ORA-07445 (See ORA-07445.ora-code.com) errors. In opening at least two TARS with Oracle, their analysts have given us optional parameters that did not work, have pointed us to patchsets that are not yet available on our platform, and finally gave us a customized patchset. In light of these issues, we have decided we will NOT be migrating this particular application to Oracle 10g in production for some time to come.
Has anyone else run into such issues? We have noted that a typical fast database refresh, even when it does work, takes at least 50% longer compared to a typical Oracle 9i implementation.
