I was troubleshooting an issue where the Oracle Database Result Cache did not get used when the RESULT_CACHE hint was specified inside a subquery. As the main query has a sysdate inside, the RESULT_CACHE Hint can not be specified for the main query, but only for the subquery.

1
2

SELECT SQ.*, sysdate FROM
(SELECT * FROM tableX, ....) SQ;

When using “/*+ RESULT_CACHE */ inside of Subquery, result cache was not enabled.

1
2
3

 
SELECT SQ.*, sysdate FROM
(SELECT /*+ RESULT_CACHE */ * FROM tableX, ....) SQ;

I found out that there is a trace event which can be set for diagnostics.

1

ALTER SESSION SET EVENTS '43905 trace name context forever, level 1';

This generated a trace file containing a clue to the reason:

1

QKSRC: ViewText[ecode=942] = SELECT  /*+ RESULT_CACHE */ ......

So, it seemed that there was an internal ORA-942 somewhere that prevented result caching from being used. The solution was to qualify the objects inside the subquery. Now the result cache was used correctly:

1
2

SELECT SQ.*, sysdate FROM
(SELECT /*+ RESULT_CACHE */ * FROM ownerY.tableX, ....)

During the Upgrade of Oracle Linux 7.2 to 7.3 with “yum update”, when updating the microcode_ctl package, the system crashed and firmware reported an uncorrectable CPU error after reboot.

It turned out that the update of the package in combination with specific Intel CPUs causes the issue. In the meantime, there is a red hat bug and solution available.

References:

• https://community.oracle.com/thread/3997160
• https://bugzilla.redhat.com/show_bug.cgi?id=1398698
• MOS: The system could not start since suddenly rebooted during upgrade to microcode_ctl package (Doc ID 2217876.1)