{"id":2104,"date":"2016-03-10T10:28:53","date_gmt":"2016-03-10T13:28:53","guid":{"rendered":"http:\/\/dev.dbarj.com.br\/en\/?p=2104"},"modified":"2017-06-21T16:13:46","modified_gmt":"2017-06-21T19:13:46","slug":"ora-00600-internal-error-code-arguments-kpdbswitchprerestore-txn-oracle-database-12c","status":"publish","type":"post","link":"https:\/\/dev.dbarj.com.br\/en\/2016\/03\/ora-00600-internal-error-code-arguments-kpdbswitchprerestore-txn-oracle-database-12c\/","title":{"rendered":"ORA-00600: internal error code, arguments: [kpdbSwitchPreRestore: txn] in Oracle Database 12c"},"content":{"rendered":"

On several Oracle DB in 12.1.0.2 version I started receiving some strange ORA-00600 and ORA-07445 errors in the alert file.\u00a0In ORA-00600, the first parameter was\u00a0“<\/strong>[kpdbSwitchPreRestore: txn]”<\/strong>:<\/p>\n

Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0xFAC8] [PC:0x7A1FA6E, qksanEndSession()+30] [flags: 0x0, count: 1]\r\nErrors in file \/u02\/diag\/rdbms\/cdbtim01\/cdbtim01\/trace\/cdbtim01_j004_21773.trc  (incident=57639):\r\nORA-07445: exception encountered: core dump [qksanEndSession()+30] [SIGSEGV] [ADDR:0xFAC8] [PC:0x7A1FA6E] [Address not mapped to object] []\r\nORA-00603: ORACLE server session terminated by fatal error\r\nORA-00600: internal error code, arguments: [k2gteget: pdbid], [3], [1], [], [], [], [], [], [], [], [], []\r\nORA-00600: internal error code, arguments: [k2gteget: pdbid], [3], [1], [], [], [], [], [], [], [], [], []\r\nORA-00600: internal error code, arguments: [kpdbSwitchPreRestore: txn], [0x2D976B388], [0], [2685403136], [], [], [], [], [], [], [], []\r\nIncident details in: \/u02\/diag\/rdbms\/cdbtim01\/cdbtim01\/incident\/incdir_57639\/cdbtim01_j004_21773_i57639.trc\r\nUse ADRCI or Support Workbench to package the incident.\r\nSee Note 411.1 at My Oracle Support for error and packaging details.<\/pre>\n
The curious thing was\u00a0that this error appeared in your alert log just after the beginning of the Maintenance Window.<\/p>\n
When I checked the first trace file pointed by the alert, I\u00a0saw\u00a0that the current SQL statement was related to some SQL Tuning task.<\/p>\n
Unfortunately this is a bug (you can check on references section) that will be only be fixed in Oracle Database 12.2<\/strong><\/span> version (not release until today March\/2016).<\/p>\n
Workaround:<\/strong><\/p>\n
The best workaround to\u00a0avoid this errors is to disable the Automatic SQL\u00a0Tuning task (on both CDB and PDB)<\/strong>:<\/p>\n
BEGIN\r\n   dbms_auto_task_admin.disable(client_name => 'sql tuning advisor', operation => NULL, window_name => NULL);\r\nEND;\r\n\/<\/pre>\n
After a fix is provided by Oracle and you apply it, don’t forget\u00a0to enable it back (on both CDB and PDB)<\/strong>:<\/p>\n
BEGIN\r\n   dbms_auto_task_admin.enable(client_name => 'sql tuning advisor', operation => NULL, window_name => NULL);\r\nEND;\r\n\/<\/pre>\n
References:<\/strong><\/p>\n
Here are some docs in Oracle that describe this bug.<\/p>\n