| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: Import Question
You can approximate the scan time on the file by using wc on it, which
should quickly and cheaply reveal whether you have a disk read problem.
MOST LIKELY, you have a lot of constraint checking and index recreation going on. You should be able to notice if that is the case, because you would have most of your delay after all the rows are in. Ameliorating that is the usual process of finding out where the time is being consumed, but the work of checking the constraints and building indexes most likely is required for sensible use of what you're reloading - though some import purposes do not require this, such as just needing the data for a full table scan report into a flat file to feed into something else. So then you play with the switches on import to avoid the work.
SECOND MOST LIKELY, you have piled up on your online redo logs and you're waiting for archiving so another one comes free. Add (even temporarily) enough (and big enough) on-line logs so that import does not have to wait for archiving and rotation. You should be able to observe relevant waits in the alert.log.
I don't have hard data on those frequencies, that's just my recollection of where the problem/delays that you can do something about usually are.
Your mileage may vary, and perhaps some other folks have quotations on their "most likelies" that are completely trivial to check.
If it is not one of these things you can either totally avoid or see quickly in a log, use another session to turn on the 10046 stuff for the import job and find out what the real problem is.
Regards,
mwf
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of Fred Smith
Sent: Thursday, August 18, 2005 9:47 AM
To: oracle-l_at_freelists.org
Subject: Import Question
Hello,
I'm going schema by schema by schema importing data from an export file.
Version: 9.2.0.3
OS: Solaris 5.8
The export file is approx ~ 14 GB.
I'm finding that even when I import the schemas (fromuser = x, touser =x) that only own a few packages etc., the import takes forever. It just sits there with "Importing X's objects into X" for 20-30 minutes, and then finally comes back with , import terminated successfully without warnings.
So my question is, is it simply taking so long just because it has to search through the entire export.dmp file?? Is there anyway to tell what it is doing?
Thank you.
-Fred S.
--
http://www.freelists.org/webpage/oracle-l
--
http://www.freelists.org/webpage/oracle-l
Received on Thu Aug 18 2005 - 09:07:41 CDT
 |
 |