Lothar,
I have an example showing
parameters ('sync (on commit)')
as the syntax for sync on commit, and a note that there was a bug with this
until 12.2 (bug #24589262 : $R CONTENTION DURING COMMIT)
Regards
Jonathan Lewis
________________________________________
From: l.flatz@xxxxxxxxxx <l.flatz@xxxxxxxxxx>
Sent: 25 November 2019 14:53
To: Jonathan Lewis
Cc: oracle-l@xxxxxxxxxxxxx
Subject: Re: Re: Text Index Create takes forever
HI Jonathan,
yes, I thought of it too. Of course the index will become less efficient. BTW:
was there not any preference to start a sync job on commit?
I can' t find it ...
Regards
Lthar
----Ursprüngliche Nachricht----
Von : jonathan@xxxxxxxxxxxxxxxxxx
Datum : 25/11/2019 - 15:37 (MN)
An : oracle-l@xxxxxxxxxxxxx
Betreff : Re: Text Index Create takes forever
Since it's going to be a killer anyway. hash partition about 16 or 32 ways
while you move it.
Regards
Jonathan Lewis
________________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx <oracle-l-bounce@xxxxxxxxxxxxx> on behalf
of l.flatz@xxxxxxxxxx <l.flatz@xxxxxxxxxx>
Sent: 25 November 2019 13:17
To: oracle-l@xxxxxxxxxxxxx
Subject: Text Index Create takes forever
Hi,
we have a 8oo million rows table that we want to migrate to an other Tablespace.
The key point is the migration of a Text index based on a BLOB (Secure file)
that contains PDF.
The recreation of the Index takes weeks. It is CPU bound and the time is spend
parsing the BLOB.
So far it seems the best Strategy to gfo as high parallel as possible.
Any idea to speed this up? DB version is 12.2.0.1.
Regards
Lothar
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l