Problem restoring z07h after parallel indexing

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

Description:
We have successfully finished parallel indexing our headings and keyword indexes, except for the very last step - dumping the z07h records back into z07 to reindex the records added and updated in the main database while indexing was going on in the parallel database. Eight days worth of activity amounted to just over 84,000 records. The util-e-5-2 routine to drop the trigger and restore the z07 records failed after about 9 hours with this message:
INSERT INTO ABC01.Z07 USING SELECT DISTINCT
*
ERROR at line 1:
ORA-01555: snapshot too old: rollback segment number 5 with name "_SYSSMU5$"
too small

I think I can figure out how to do this manually - extract the record keys from the z07h table, dedup them and then add them to the z07 table. I've done something like this in the past by using p_manage_37 and specifying a very simple fix routine that doesn't actually do anything.

Resolution:
Based on previous experience, it seemed that -- even though the z07 Oracle indexes were showing as present and valid in util a/17/14 -- this could be a problem with the z07 Oracle indexes.

I stopped ue_01 and did util a/17/4 for Z07. When I tried to do util a/17/2 to build, it found duplicate keys. So I ran util a/17/18 which deletes duplicates and builds a unique index -- which it did successfully.

I have now restarted util e/05/02 . It is processing 260 records per minute. This means that it should do the 80,000 in about 5 hours. I think that not having other things updating the z07 (that is, having ue_01 stopped and low activity on the pc_server) should help.