Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: Import 8.1.6 to 9.2.0.1 weirdness

Re: Import 8.1.6 to 9.2.0.1 weirdness

From: <Jared.Still_at_radisys.com>
Date: Fri, 16 May 2003 10:33:32 -0800
Message-ID: <F001.0059B12B.20030516103332@fatcity.com>


Please note the MetaLink doc states that this is for CONVENTIONAL PATH ONLY. So direct=y should not be affected.

There may also be a patch available, depending on platform and version.

There is a workaround supplied, but it looks suspiciously like something that
will make exp slower than it already is.

ie. set ORA_OCI_NO_OPTIMIZED_FETCH=1 in the environment.

This script was run as a cursory test. It doesn't appear that setting  ORA_OCI_NO_OPTIMIZED_FETCH=1 has an adverse affect.


# test export speed with workaround from doc ID 223399.1
# this is the export corruption bug.  When more than 32k
# rows are fetched at once, corruption can occur.  This is
# supposed to affect only conventional path, not direct.


DATABASE=dv01

USERNAME=jkstill
PASSWORD=$(pwc.pl -machine all -instance $DATABASE -username $USERNAME)
BUFFER=10485760         # 10 meg
RECORDLENGTH=65535      # maximum allowed - see note 1028959.6
TABLE=PDBAREP.PDBA_TAB_COLUMNS unset ORA_OCI_NO_OPTIMIZED_FETCH

echo Conventional Path export with normal optimization

time exp userid=$USERNAME/$PASSWORD \

        file=conventional_opt.dmp \
        file=conventional_opt.log \
        buffer=$BUFFER \
        compress=n \
        tables=$TABLE


export ORA_OCI_NO_OPTIMIZED_FETCH=1

echo Conventional Path export with no optimization

time exp userid=$USERNAME/$PASSWORD \

        file=conventional_no_opt.dmp \
        file=conventional_no_opt.log \
        buffer=$BUFFER \
        compress=n \
        tables=$TABLE

unset ORA_OCI_NO_OPTIMIZED_FETCH

echo Direct Path export

time exp userid=$USERNAME/$PASSWORD \

        file=direct.dmp \
        file=direct.log \
        direct=y \
        recordlength=$RECORDLENGTH \
        compress=n \
        tables=$TABLE

=============================================================

Conventional Path export with normal optimization

Export: Release 8.1.7.2.0 - Production on Fri May 16 10:29:30 2003

(c) Copyright 2000 Oracle Corporation. All rights reserved.

Connected to: Oracle8i Enterprise Edition Release 8.1.7.2.0 - Production With the Partitioning option
JServer Release 8.1.7.2.0 - Production
Export done in WE8ISO8859P1 character set and WE8ISO8859P1 NCHAR character set

About to export specified tables via Conventional Path ... Current user changed to PDBAREP
. . exporting table PDBA_TAB_COLUMNS 281353 rows exported
Export terminated successfully without warnings.

real    0m6.227s
user    0m5.710s
sys     0m0.370s

Conventional Path export with no optimization

Export: Release 8.1.7.2.0 - Production on Fri May 16 10:29:36 2003

(c) Copyright 2000 Oracle Corporation. All rights reserved.

Connected to: Oracle8i Enterprise Edition Release 8.1.7.2.0 - Production With the Partitioning option
JServer Release 8.1.7.2.0 - Production
Export done in WE8ISO8859P1 character set and WE8ISO8859P1 NCHAR character set

About to export specified tables via Conventional Path ... Current user changed to PDBAREP
. . exporting table PDBA_TAB_COLUMNS 281353 rows exported
Export terminated successfully without warnings.

real    0m6.232s
user    0m5.870s
sys     0m0.240s

Direct Path export

Export: Release 8.1.7.2.0 - Production on Fri May 16 10:29:43 2003

(c) Copyright 2000 Oracle Corporation. All rights reserved.

Connected to: Oracle8i Enterprise Edition Release 8.1.7.2.0 - Production With the Partitioning option
JServer Release 8.1.7.2.0 - Production
Export done in WE8ISO8859P1 character set and WE8ISO8859P1 NCHAR character set

About to export specified tables via Direct Path ... Current user changed to PDBAREP
. . exporting table PDBA_TAB_COLUMNS 281353 rows exported
Export terminated successfully without warnings.

real    0m1.826s
user    0m1.120s
sys     0m0.930s









Rajesh.Rao_at_jpmchase.com
Sent by: root_at_fatcity.com
 05/16/2003 08:36 AM
 Please respond to ORACLE-L  

        To:     Multiple recipients of list ORACLE-L <ORACLE-L_at_fatcity.com>
        cc: 
        Subject:        Re: Import 8.1.6 to 9.2.0.1 weirdness



Refer Note: 223399.1. A data corruption can occur if the buffer parameter has been set to a value so more than 32,768 rows will be fetched.

Effects version 8.1.7.3 and above. Has been fixed in 9.2.0.3. For other versions theres a workaround.

Regards
Raj   

                    Prem Khanna J  
                    <jprem_at_kssnet.       To:     Multiple recipients of 
list ORACLE-L <ORACLE-L_at_fatcity.com> 
                    co.jp>               cc:   
                    Sent by:             Subject:     Re: Import 8.1.6 to 
9.2.0.1 weirdness 
                    root_at_fatcity.c  
                    om  
  
  
                    05/15/2003  
                    11:28 PM  
                    Please respond  
                    to ORACLE-L  
  
  




Jared,

sorry to drag away from the topic.
i read in a metalink doc that large BUFFER size during EXP may result in corrupt DMP file.

is it no so ? is 10M (compared to 32768) not large ?

just curious to know !

Regards,
Jp.

16-5-2003 09:37:02, Jared.Still_at_radisys.com wrote:
>I don't know if this may have any bearing on the problem you are
>experiencing,
>but a 64k buffer is rather small. I like to use 10 meg on imports
eg.
>buffer=10485760
>When it is hanging, have you logged in to the target database and
checked
>to
>see what the import session is waiting on via v$session_wait?
>Jared

--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Prem Khanna J
  INET: jprem_at_kssnet.co.jp

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).





-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.net
-- 
Author: 
  INET: Rajesh.Rao_at_jpmchase.com

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).




-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.net
-- 
Author: 
  INET: Jared.Still_at_radisys.com

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).
Received on Fri May 16 2003 - 13:33:32 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US