Prereq query for Logical Standby runs for days

Question

Have had an SR open for a month that isn't getting anywhere, thought I'd check here and see if anybody has run into this.
We're looking at setting up a logical standby database, Oracle 11.2.0.3 on AIX running PeopleSoft.
There is a prereq query in the Oracle manual which is also run via the GC wizard to find tables without unique logical identifiers.
SELECT OWNER, TABLE_NAME FROM DBA_LOGSTDBY_NOT_UNIQUE
WHERE (OWNER, TABLE_NAME) NOT IN
(SELECT DISTINCT OWNER, TABLE_NAME FROM DBA_LOGSTDBY_UNSUPPORTED)
AND BAD_COLUMN = 'Y';
My problem is that the darned thing runs for 5.5 days.
The view in that query runs the function  logstdby$tabf, which runs the following SQL over and over and over and over:
SELECT * FROM LOGSTDBY_UNSUPPORT_TAB_11_2B
Each time it runs the LOGSTDBY_UNSUPPORT_TAB_11_2B query is about 8 seconds. Run it 100,000 times and that adds up. Looks like the big problem in that query is full table scans of TAB$ and COL$.
Anybody seen it? Anybody fixed it?

Answer 1

Good suggestion - in 15 minutes you got farther than support did in 2 weeks.
Rather than remove the sub-query, I can change it into a minus, which is functionally equivalent to the original but changes the executions on DBA_LOGSTDBY_UNSUPPORTED from many times to just once.
select owner, table_name from dba_logstdby_not_unique
where bad_column = 'Y'
minus
select distinct owner, table_name from dba_logstdby_unsupported;
The top part of the query still takes 9.5 hours to run though, with the time now being spent the following SQL which is called from the function logstdby$tabf which is called from the view DBA_LOGSTDBY_NOT_UNIQUE :
SELECT * FROM LOGSTDBY_SUPPORT_TAB_11_2B
Sorry for not putting that in the original post, but I want the original query to run if possible so the GC wizard will work - and I didn't want to clutter up the question.