- Article Type: General
- Product: Aleph
- Product Version: 17.01
We have done parallel indexing using Library (schema) ABC07. We now wish to move the objects (tables, indexes) created during parallel indexing from ABC07 into the ABC01 schema.
We would like you to review the method we plan to use, and ask you to advise us about our accuracy and whether there are any shortcuts which can save us time. [Some objects contain 1.6 M rows and will take time to manipulate. We want to be careful here.]
Steps we will use:
1. [corresponds to Step 9, Option1 on p. 8 of the Parallel Indexing document]
Use LOGICAL SYNONYMS to point to the ABC07 indexes so that patrons will not be making using of ABC01 indexes and this allows us to update them without affecting searching.
2. [move data from schema ABC07 to ABC01. This method avoids using export/import, which is messier.]
alter table abc01.z00H rename to z00H_sav ; [we only do this just in case there are problems and we need to easily undo our work]
create abc01.z00H as select* from abc07.z00H ; [“copies” the table data from schema ABC07 to ABC01]
alter index abc01.z00H_ID rebuild ; [Question: is this step required or will the one above automatically update the index properly?
We believe it is.]
3. Do similar operations for each of the following tables/indexes in ABC07 as listed below (ignore objects which are SYNONYMS, TRIGGERS OR PROCEDURES):
If any of these are not needed in ABC01, could you advise?
4. Undo Step 1 where we made use of LOGICAL SYNOMNYMS while we were updating ABC01 with ABC07 schema data. Then all searches will use ABC01 indexes, which were newly updated from ABC07
We do not recommend the use of Step 9, option 2. (There have been problems trying to transfer the Z98 table in particular.) We strongly recommend the use Option 1 (synonyms). You would create synonyms in ABC01 for the index tables which you have generated in the ABC07 parallel library -- and those tables only.
Be sure to thoroughly test the indexes for which you are creating synonyms with the OPAC pointing to the parallel library (prior to the time the synonyms are added to ABC01).
This method will let you handle *subsequent* parallel indexing in the fashion described in KB 16384-2518 -- which is definitely the best way to do it. In this scenario the 2nd, 3rd, etc. parallel indexing will be *completely* risk-free.
See KB 8192-4574 to make sure that you have a version of the "Parallel Indexing" document which warns against a possible problem.
- Article last edited: 10/8/2013