RFS-process on physical standby database fails with Ora-00600:[Kcrrrfswda.9]

In this Document
  Symptoms
  Changes
  Cause
  Solution
  References

---

Applies to:

Oracle Server - Enterprise Edition - Version: 11.1.0.6 and later   [Release: 11.1 and later ]
Information in this document applies to any platform.

Symptoms

Physical standby database

RFS-process fails on physical standby database with fallowing errors:

Errors in file d:\oracle\app\administrator\diag\rdbms\pihist\pihist3\trace\pihist3_rfs_5364.trc:
ORA-00272: error writing archive log L:\ARCHIVE\PIHIST\ARC00089_0731084629.001
ORA-00600: internal error code, arguments: [kcrrrfswda.9], [4], [368], [], [], [], [], []
RFS[29]: Possible network disconnect with primary database

On the primary database there may be fallowing errors reported:

Errors in file d:\oracle\app\administrator\diag\rdbms\pdonds\pdonds\trace\pdonds_arc7_3572.trc:
ORA-12582: TNS:invalid operation
FAL[server, ARC7]: FAL archive failed, see trace file.
Errors in file d:\oracle\app\administrator\diag\rdbms\pdonds\pdonds\trace\pdonds_arc7_3572.trc:
ORA-16055: FAL request rejected

Changes

This errors are network related, eg firewall is enabled between primary/standby database.

Cause

This ORA-600 is a network related issue:

1. A firewall-feature is most likely the culprit for this ORA-600. Please ensure there is no feature enabled that is modifying TCP packets like 'fixup' or 'inspect' on Cisco Firewalls for example. Those features modify the TCP-packets containing redo data resulting into the RFS-Process on the standby database refusing those packets. This can then end up in the ORA-600.

Please make sure the following firewall features are disabled:

- SQLNet fixup protocol
- Deep Packet Inspection (DPI)
- SQLNet packet inspection
- SQL Fixup
- SQL ALG (Juniper firewall)

2. There is also a known issue reported in Bug 8605213: ORA-00600: [KCRRRFSWDA.9]:
The ORA-600 didn't happen anymore after setting  dead connection detection in Oracle Net (SQLNET.EXPIRE_TIME) to a value below the firewall timeout threshold.

3. There are different non-public bugs, most of them are closed as not reproducable, or customer was not willing to disable firewall-feature to verify if the error still happens afterwards.

Solution

1, check network setup

2. disable the firewall-feature or disable firewall

3. check sqlnet.expire_time and set it below the timeout value of the firewall

References

BUG:10352413 - ARCHIVELOG WAS NOT TRANSFERRED TO STANDBY
BUG:8218162 - ORA-00600: INTERNAL ERROR CODE, ARGUMENTS: [KCRRRFSWDA.9], [4], [368], [], []
BUG:8605213 - ORA-00600: INTERNAL ERROR CODE, ARGUMENTS: [KCRRRFSWDA.9]
BUG:8967127 - LNS REPORTS ORA-12582 AND RFS PROCESS REPORT ORA-00600: [KCRRRFSWDA.9]
BUG:9063691 - ORA-00600: [KCRRRFSWDA.9], [4], [368] IN RFS ON STANDBY.
NOTE:386417.1 - Redo Corruption Errors During Redo Transport

最后和网管确认后,禁用了inspect sqlnet,该问题搞定.

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/11134237/viewspace-684529/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/11134237/viewspace-684529/