• Article Type: General
• Product: Aleph
• Product Version: 16.02

Description:
We have done parallel indexing previously. The changes shown in the Parallel Indexing document, Step 9, Option 1- Logical Synonyms, seem to be for the case where the indexing is being done for the first time. What would the changes be if it's a subsequent indexing?

Resolution:
The "Option 1: Logical Synonyms" doesn't properly address the case where you have previously run parallel indexing.

In the case where you are doing the indexing in a second parallel library, rather than replacing TAB and IND entries in the xxx01 library file_list with "LS"s, you would change the LS's in xxx01 file_list from one parallel library to the second parallel library.

Let's say that the ABC01 file list has:

LS z01 ABC03
LS z02 ABC03
LS z05 ABC03
LS z11 ABC03
<etc.>

Then to make the new indexes in ABC07 be in effect, you would change the LS's to ABC07:

LS z01 ABC07
LS z02 ABC07
LS z05 ABC07
LS z11 ABC07
<etc.>

and do util a/17/5/2 to recreate the synonyms.

There would be no need to "Drop the relevant Z tables in the actual library" because this was done previously when you previously created the LS's to the other parallel library (abc03).

The beauty of it is, if there's a problem, you can change the file_list back to ABC03, rerun util a/17/5/2 and you're back to your previous indexes -- a change taking less than five minutes.

In order to save space, you may, several days after implementing the new index -- when you are quite certain that everything is working correctly -- , want to do delete the old index in the old parallel library. This could be done using util a/17/1 but that utility recreates the table (empty, with the primary allocation). To actually drop the table, you would do: SQL> drop table znn . (Make sure you are doing this in the correct library!)

In the above example, you would drop the z01, z02, z05, and z11 tables in abc03.

• Article last edited: 10/8/2013