Anda di halaman 1dari 18

Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...

User
Enterprise Strategies (/index.php/enterprise-
Velocity (/) strategies)
Enterprise Competencies (/index.php/enterprise-
competencies)
Project Types (/index.php/project-types-all)
Velocity (/index.php/velocity-content)

Enterprise Strategies (https://velocity.informatica.com
/index.php/enterprise-strategies)
Data Governance & Enterprise Information
Management (https://velocity.informatica.com
/index.php/enterprise-strategies/data-governance)
Enterprise Architecture & Business Transformation
(https://velocity.informatica.com/index.php/enterprise-
strategies/eim)
ICC/Lean Integration
(https://velocity.informatica.com/index.php/enterprise-
strategies/icc-lean-integration)
Enterprise Competencies
(https://velocity.informatica.com/index.php/enterprise-
competencies)
Business Process Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/business-process-management)
Data Quality Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/data-quality-management)
Enterprise Architecture
(https://velocity.informatica.com/index.php/enterprise-
competencies/enterprise-architecture)
Financial Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/financial-management)
Information Lifecycle Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/information-lifecycle-management)
Integration Methodology
(https://velocity.informatica.com/index.php/enterprise-
competencies/integration-methodology)
Integration Systems (https://velocity.informatica.com
/index.php/enterprise-competencies/integration-
systems)
Metadata Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/metadata-management)
Modeling Management
(https://velocity.informatica.com/index.php/enterprise-
competencies/modeling-management)
Project Types (https://velocity.informatica.com
/index.php/project-types-all)
Big Data (https://velocity.informatica.com/index.php Return to Top
1 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
/project-types-all/big-data-projectype) Search ...
Business to Business (B2B)
(https://velocity.informatica.com/index.php/project- Search
types-all/b2b-project-type)
Cloud Customer 360
(https://velocity.informatica.com/index.php/project-
types-all/pt-cloud-360)
Data Archiving (https://velocity.informatica.com/)
Data Archive for Performance
(https://velocity.informatica.com/index.php/project-
types-all/data-archiving/ilmarchive-project-type)
Data Archive for Retirement
(https://velocity.informatica.com/index.php/project-
types-all/data-archiving/ilmretire-project-type)
Data Integration (https://velocity.informatica.com/)
Data Migration (https://velocity.informatica.com
/index.php/project-types-all/data-integration/data-
migration-project-type)
Data Synchronization
(https://velocity.informatica.com/index.php/project-
types-all/data-integration/data-synchronization-
project-type)
Data Warehousing
(https://velocity.informatica.com/index.php/project-
types-all/data-integration/data-warehousing-
project-type)
Data Privacy (https://velocity.informatica.com
/index.php/project-types-all/dataprivacy-project-type)
Data Quality (https://velocity.informatica.com
/index.php/project-types-all/data-quality-project-type)
Master Data Management (MDM)
(https://velocity.informatica.com/index.php/project-
types-all/mdm-project-type)
Velocity (https://velocity.informatica.com/index.php
/velocity-content)
Roles (https://velocity.informatica.com/index.php
/velocity-content/roles-all-pb)
Project Phases (https://velocity.informatica.com
/index.php/velocity-content/project-phases-pb)
Best Practices (https://velocity.informatica.com
/index.php/velocity-content/best-practices-all-bp)
Sample Deliverables
(https://velocity.informatica.com/index.php/velocity-
content/sample-deliverables-all-bp)
Templates (https://velocity.informatica.com
/index.php/velocity-content/velocitytemplates-bp)
Accelerators (https://velocity.informatica.com
/index.php/velocity-content/accelerators-bp)
Home (/index.php) Best Practices (/index.php/best-practices-all) PowerExchange Configuration
(/index.php/best-practices-all/158-powerexchange-configuration) PowerExchange for Oracle CDC
Print (/index.php/best-practices-all/158-powerexchange-configuration
Best /442-m1bp423?tmpl=component&print=1&page=)
Return to Top
2 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...

Practices, Vote 5 Rate

Project
Details
Guidance and Hits: 10989

Tools
PowerExchange for Oracle
Roles (/index.php
/roles-all) CDC
Project Phases
(/index.php/project-
phases)
Challenge
Phase 1: Configure the Oracle environment for optimal performance when using
Manage PowerExchange Change Data Capture (CDC) in a production environment.
(/index.php
/project-phases
/phase1-all)
Description
Phase 2: There are two performance types that need to be considered when dealing
Analyze with Oracle CDC; latency of the data and restartability of the environment.
(/index.php Some of the factors that impact these performance types are configurable
/project-phases within PowerExchange, while others are not. These two performance types
/phase2-all) are addressed separately in this Best Practice.
Phase 3:
Architect Minimize Data Latency
(/index.php
/project-phases The objective of data latency is to minimize the amount of time that it takes
/phase3-all) for a change made to the source database to appear in the target database.
Phase 4: Some of the factors that can affect latency performance are discussed
Design below.
(/index.php
/project-phases Location of PowerExchange CDC
/phase4-all)
The optimal location for installing PowerExchange CDC is on the server
Phase 5: Build
that contains the Oracle source database. This eliminates the need to use
(/index.php
the network to pass data between Oracles LogMiner and PowerExchange.
/project-phases
It also eliminates the need to use SQL*Net for this process and it minimizes
/phase5-all)
the amount of data being moved across the network. For best results, install
Phase 6: Test
the PowerExchange Listener on the same server as the source database
(/index.php server.
/project-phases
/phase6-all)
Phase 7:
Volume of Data
Deploy The volume of data that the Oracle Log Miner has to process in order to
(/index.php provide changed data to PowerExchange can have a significant impact on
/project-phases performance. Bear in mind that in addition to the changed data rows, other
/phase7-all) processes may be writing large volumes of data to the Oracle redo logs.
Phase 8: These include, but are not limited to:
Operate
(/index.php Oracle catalog dumps
/project-phases Oracle workload monitor customizations
/phase8-all) Other (non-Oracle) tools that use the redo logs to provide proprietary
information
Best Practices Return to Top
3 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
(/index.php/best- In order to optimize PowerExchanges CDC performance, the amount of
practices-all) data these processes write to the Oracle redo logs needs to be minimized
Sample (both in terms of volume and frequency). This includes minimizing the
Deliverables invocations of the LogMiner to just a single occurrence. Review the
(/index.php/sample- processes that are actively writing data to the Oracle redo logs and tune
deliverables-all) them within the context of a production environment. Monitoring the redo
Templates log switches and the creation of archived log files is one way to determine
(/index.php how busy the source database is. The size of the archived log files and how
/velocitytemplates) often they are being created over a day will give a good idea about
Accelerators performance implications.
(/index.php
/accelerators) Server Workload
Optimize the performance of the Oracle database server by reducing the
Filter By number of unnecessary tasks it is performing concurrently with the
PowerExchange CDC components. This may include a full review of the
Project Type backup and restore schedules, Oracle import and export processing and
other application software utilized within the production server environment.
Project Types
PowerCenter also contributes to the workload on the server where
PowerExchange CDC is running; so it is important to optimize these
workload tasks. This can be accomplished through mapping design. If
possible, include all of the processing of PowerExchange CDC sources
within the same mapping. This will minimize the number of tasks generated
and will ensure that all of the required data from either the Oracle archive
log (i.e., near real time) or the CDC files (i.e., CAPXRT, condense) process
within a single pass of the logs or CDC files.

Continuous Extraction Mode Considerations


Continuous Extraction Mode is slightly different than Batch Extraction Mode
and allows PowerCenter to use the Oracle Real-time CDC connection with
an override for the CAPI-CONNECTION. In this mode, PowerExchange
initiates the extraction process of changed data from the Oracle archive
logs through PowerExchanges condense process. PowerExchange first
checks the change registrations for all of the CDC data sources associated
with the collection id. Once this process has completed successfully,
PowerExchange initiates an Oracle LogMiner session. The type of
LogMiner session is controlled by a configuration file parameter. This
configuration file parameter should be set to continuous mode. This will
establish an Oracle LogMiner continuous mining session to extract all of
changes from the Oracle archive and the online redo logs. Oracle LogMiner
will first process all of the archive logs that have not yet been processed
and then it will process the Oracle online redo logs. This process runs
continuously until it is manually stopped. PowerExchange flushes the write
buffers periodically for the open condense data files for consumption by
PowerCenter.
When PowerExchange receives a request for change data source records
from PowerCenter it retrieves the unprocessed records from the closed
condense data files and passes them to PowerCenter. Once all of the
condense files have been processed, PowerExchange starts processing the
condense files that are still open and being written to periodically.
Below is a high-level depiction of the processes associated with continuous
extraction. Return to Top
4 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
Change Data Capture Component

1. At startup time, the PowerExchange CDC process (i.e. condense) reads


the dbmover.cfg and dtlca.cfg files for configuration information.
2. The PowerExchange CDC process retrieves all of the change
registrations for the collection id specified in the dtlca.cfg file.
3. The PowerExchange CDC process retrieves restart tokens from the
CDC control files.
4. PowerExchange CDC initiates an Oracle LogMiner session to retrieve
changed data records.
5. LogMiner passes all the changed data records to the PowerExchange
CDC process.
6. PowerExchange CDC periodically flushes the write buffers to the
condense data file based on the FILE_FLUSH_VAL parameter.
PowerExchange CDC closes the condense data file and opens a new
file based on the FILE_SWITCH_CRIT and FILE_SWITCH_VAL
parameter settings.
7. The PowerExchange CDC process updates restart tokens from the
CDC control files.
Data Consumption Component

Return to Top
5 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...

1. The PowerCenter session issues a request for CDC data to the


PowerExchange client from a certain location.
2. The PowerExchange client checks the dbmover.cfg file for the IP
address associated with the location.
3. The PowerExchange client sends the request to the PowerExchange
Listener.
4. The PowerExchange Listener sends the request to the PowerExchange
CDC process.
5. The PowerExchange CDC process retrieves the requested change data
records from the condense data files.
6. The PowerExchange CDC process passes only the change records for
the requested sources to the PowerExchange Listener.
7. The PowerExchange Listener passes the requested change records to
the PowerExchange client.
8. The PowerExchange client passes the change records to the
PowerCenter session.
The best PowerExchange CDC extraction method to be used in most cases
involving Oracle is the Continuous Extraction Mode which has the following
benefits.
Not dependent on PowerCenter to initiate the change data capture
process.
Minimal impact on the Oracle database server resources.
Captures changes for all change registrations for the Oracle database
SID with a single pass of the available Oracle archive logs.
Once all outstanding Oracle archive logs have been processed, Oracle
LogMiner starts mining the online redo logs.
Lower risk of unprocessed Oracle archive logs being swept from disk
prior to being processed by Oracle LogMiner
The latency of the data is relatively small as long as the requesting
PowerCenter Workflow is running continuously.
All of the change records are stored within a condense file and ready for
consumption by PowerCenter when the request is received.
PowerCenter has access to the change data records even though the Return to Top
6 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
PowerExchange condense file has not been closed.

The following is a sample of a pwxccl.cfg parameters file that controls the


Oracle CDC condense process:
/* ----------------------------------------------------------------*/
/* PowerExchange Condense Configuration File
/*
/* See PowerExchange CDC Guide for Linux, UNIX and Windows
/* Chapter 3 PowerExchange Logger for Linux, UNIX and Windows
/* ----------------------------------------------------------------*/
/* The value for the DBID parameter must match the Collection-ID
/* contained in the ORACLE-ID statement in the dbmover.cfg file.
/* ----------------------------------------------------------------*/
/*
DBID=ORACDC
DB_TYPE=ORA
EXT_CAPT_MASK=/home/pwx/v861/condense/condense
CHKPT_BASENAME=/home/pwx/v861/condense/condense.CHKPT
CHKPT_NUM=3
COND_CDCT_RET_P=5
NO_DATA_WAIT=1
NO_DATA_WAIT2=60
/*
CONDENSE_SHUTDOWN_TIMEOUT=60
/*
/* ----------------------------------------------------------------*/
/* COLL_END_LOG equal to 1 means BATCH MODE
/* COLL_END_LOG equal to 0 means CONTINUOUS MODE
/* ----------------------------------------------------------------*/
/*
COLL_END_LOG=0
/*
/* ----------------------------------------------------------------*/
/* FILE_SWITCH_CRIT of M means minutes
/* FILE_SWITCH_CRIT of R means records
/* ----------------------------------------------------------------*/
/*
FILE_SWITCH_CRIT=M
FILE_SWITCH_VAL=20
FILE_SWITCH_MIN=(1,2)
FILE_FLUSH_VAL=10
/*
/* ----------------------------------------------------------------*/
/* CAPT_IMAGE of AI means AFTER IMAGE
/* CAPT_IMAGE of BA means BEFORE and AFTER IMAGE
/* ----------------------------------------------------------------*/
/*
CAPT_IMAGE=BA
SIGNALLING=Y
/*
/* ----------------------------------------------------------------*/
/* Oracle User id and Password for PowerExchange CDC
/* ----------------------------------------------------------------*/ Return to Top
7 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
/*
UID=database userid
PWD=database password
/*
/********************************************************************/
/* The following parameters are only used during a cold start and forces
/* the cold start to use the most recent catalog copy. Without these
/* parameter, if the v_$transaction and v_$archive_log views are out of
/* sync and there is a very good chance that the most recent
/* catalog copy will not be used for the cold start.
/********************************************************************/
/*
SEQUENCE_TOKEN=0
RESTART_TOKEN=0
These parameters have the following description and syntax requirements.

Configuration Parameter Description


DBID=collection id Specifies the PowerExchange
change registration collection
identifier, also called instance
name. The value used must be
identical to the first parameter of
the ORACLEID statement in the
dbmover.cfg.
DB_TYPE=ORA Mandatory parameter and must
contain ORA for Oracle.
EXT_CAPT_MASK=/directory/mask Unique mask for the data files
name created by the Condense. A suffix
will be added to this mask name
which contains a date/time stamp
of when the file was created. In
addition, the directory must already
exist.
CHKPT_BASENAME=/directory/mask Unique mask for the checkpoint
name files created by the Condense. A
suffix is added to this mask name
which contains a Vnn. In addition,
the directory must already exist.
CHKPT_NUM=number checkpoint files Specifies the number of checkpoint
files. The default is 3.
COND_CDCT_RET_P=number of days CDCT and Condensed files
retention period in days. Files older
than this period and their
corresponding CDCT records are
deleted during start-up, file switch,
or shut down processing.

Return to Top
8 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
NO_DATA_WAIT=number of minutes When running in Continuous mode,
it defines the number of minutes to
wait on commands manually
entered through the Command
Handler before starting the next
Condense. The default is 60.
NO_DATA_WAIT2=number of seconds Defines the number of seconds
before the Condenser stops. The
default is 600 seconds.
CONDENSE_SHUTDOWN_TIMEOUT= Specifies the maximum time period
number of seconds for the PowerExchange
Condenser, DTLCACON, to shut
down normally after a shutdown
command. The default is 600.
COLL_END_LOG=0 or 1 Specifies whether to use Batch or
Continuous.
0 - Continuous mode. After each
condense run, the system waits for
the number of minutes defined in
the NO_DATA_WAIT parameter,
then performs another Condense.
1 - Batch mode. The system shuts
down after a single condense run.
For example, a single condense
run might be scheduled following a
particular batch update job.
FILE_SWITCH_CRIT=Records or Defines the criteria to use when
Minutes deciding when to do an automatic
file switch. The default is M.
R Records
M Minutes
FILE_SWITCH_VAL=file switch units Defines the number of
FILE_SWITCH_CRIT units at
which to perform a file switch
automatically.
Default is 30.

Return to Top
9 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
FILE_SWITCH_MIN=(number of units, Specifies file switch criteria for
number of units ignored) Condense when changes for new
sources are encountered. Use this
to reduce latency for continuous
extraction mode.
Number of Unitsspecifies the
minimum number of
FILE_SWITCH_CRIT units that
must pass before a file switch is
done when encountering a change
for a source with no entry in the
CDCT.
Number of Units Ignored specifies
the number of
FILE_SWITCH_CRIT units which
must occur during cold start
processing before Condense uses
the Number of Units value.
The default is (-1,0).
FILE_FLUSH_VAL=number of seconds Specifies the file flush interval in
seconds. The file flush interval is
the number of seconds that elapse
before a flush is performed on the
current partial condense file.
When the Condense task flushes,
the data is written to the disk
condense files allowing it to be
read by continuous extraction
mode extractions.
CAPT_IMAGE=before or after image Specifies whether before images or
after images should be captured.
BA Before and after images
AI Just after images
SIGNALLING=Y or N Specifies how the system should
come down when an abnormal
condition occurs.
Y The system tries to shutdown
normally
N The system abends with a
dump
Default is N.
UID=user id Oracle database user id
PWD=password Oracle database password
SEQUENCE_TOKEN=0 Sequence portion of the restart
token to be used when doing a
cold start.
Return to Top
10 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
RESTART_TOKEN=0 Restart token to be used when
doing a cold start.

Condense Option Considerations


The condense option for Oracle CDC provides only the required data by
reducing the collected data based on the Unit of Work information. This can
prevent the transfer of unnecessary data and save CPU and memory
resources. In order to properly allocate space for the files created by the
condense process it is necessary to perform capacity planning.
In determining the space required for the CDC data files it is important to
know whether before and after images (or just after images) are required.
Also, the retention period for these files must be considered. The retention
period is defined in the COND_CDCT_RET_P parameter in the dtlca.cfg
file. The value that appears for this parameter specifies the retention period
in days. The general algorithms for calculating this space are outlined
below.
After Image Only
Estimated condense file disk space for Table A =
((The width of Table A in bytes * Estimated number of data
changes for Table A per
24 hour period) + 700 bytes for the six fields added to each CDC
record) *
the value on theCOND_CDCT_RET_P parameter

Before/After Image
Estimated condense file disk space for Table A =
((The width of Table A in bytes * Estimated number of data
changes for Table A per
24 hour period) * 2) + 700 bytes for the six fields added to each
CDC record) *
the value on theCOND_CDCT_RET_P
parameter

Accurate capacity planning can be accomplished by running sample


condense jobs for a given number of source changes to determine the
storage required. The size of files created by the condense process can be
used for projecting the actual storage required in a production environment.

PowerExchange CDC Restart


Performance
The amount of time required to restart the PowerExchange CDC process
should be considered when determining performance. The PowerExchange
CDC process will need to be restarted whenever any of the following events
occur:
A schema change is made to a table.
Return to Top
11 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
An existing Change Registration is amended.
The PowerExchange service pack is applied or a configuration file is
changed.
An Oracle patch or bug fix is applied.
An Operating System patch or upgrade is applied.
A copy of the Oracle catalog must be placed on the archive log in order for
LogMiner to function correctly. The frequency of these copies is very site
specific and it can impact the amount of time that it takes the CDC process
to restart. As a best practice, Informatica recommends avoiding major
changes to the Oracle production environment. Significant changes can
negatively impact the performance and re-start time of the Oracle CDC
environment.
There are several parameters that appear in the dbmover.cfg configuration
file that can assist in optimizing restart performance. These parameters
are:
RSTRADV: The RSTRADV parameter specifies the number of seconds to
wait after receiving a Unit of Work (UOW) for a source table before
advancing the restart tokens by returning an empty UOW. This parameter
is very beneficial in cases where the frequency of updates on some tables
is low in comparison to other tables.
CATINT: The CATINT parameter specifies the frequency in which the
Oracle catalog is copied to the archive logs. Since LogMiner needs a copy
of the catalog on the archive log to become operational, this is an important
parameter as it will have an impact on which archive log is used to restart
the CDC process. When Oracle places a catalog copy on the archive log, it
will first flush all of the online redo logs to the archive logs prior to writing
out the catalog.
CATBEGIN: The CATBEGIN parameter specifies the time of day that the
Oracle catalog copy process should begin. The time of day that is specified
in this parameter is based on a 24 hour clock.
CATEND: The CATEND parameter specifies the time of day that the Oracle
catalog copy process should end. The time of day that is specified in this
parameter is based on a 24 hour clock.
Refer to the dbmover.cfg file below for a description of these parameters. It
is important to carefully code these parameters as they can impact the
amount of time it takes to restart the PowerExchange CDC process.
/* ----------------------------------------------------------------*/
/* Trace Parameter
/* ----------------------------------------------------------------*/
TRACE=(ORAC,2,99)
/*
/* ----------------------------------------------------------------*/
/* Capture Registrations and Extraction Maps
/* ----------------------------------------------------------------*/
/*
CAPT_XTRA=/home/pwx/v861/capture
CAPT_PATH=/home/pwx/v861/capture/camaps
/*
/* ----------------------------------------------------------------*/
/* Oracle Change Data Capture Parms
/* see PowerExchange CDC Guide for Linux, UNIX, and Windows Return to Top
12 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
/* Chapter 6 - Oracle Change Data Capture
/* see PowerExchange Reference Guide
/* Chapter 2 DBMOVER Configuration File Parameters
/* see Readme_ORACAPT.txt
/* ----------------------------------------------------------------*/
/*
/*ORACLEID=
(collection_id,oracle_sid,connect_string,capture_connect_string)
/*
/*
/* ----------------------------------------------------------------*/
/* Change Data Capture Parameters
/* ----------------------------------------------------------------*/
/*
ORACLEID=(ORACDC,orcl,orcl,orcl)
/*
CAPI_CONNECTION=(NAME=CAPCAPX,TYPE=
(CAPX,DFLTINST=ORACDC))
CAPI_CONNECTION=(NAME=CAPIUOW,TYPE=
(UOWC,CAPINAME=CAPIORA,
RSTRADV=60,MEMCACHE=4096))
CAPI_CONNECTION=(NAME=CAPIORA,DLLTRACE=ORAC,TYPE=
(ORCL,CATINT=30,
CATBEGIN=00:01,CATEND=23:59,,ARRAYSIZE=1000,
COMMITINT=5,
REPNODE=local,BYPASSUF=Y,ORACOLL=ORACDC))
These parameters have the following description and syntax requirements.

Configuration Parameter Description


TRACE=(ORAC,2,99) Trace severely impacts performance
and should only be done under the
direction of Informatica Global
Customer Support. In an Oracle
CDC environment, ALWAYS
turn-on this trace ahead of time to
see what is going on even if there
are no problems.

CAPT_XTRA=qualified directory Path to the local directory where the


pathname change registration extraction maps
are to be stored. This is a required
parameter for CDC and the directory
path specified must exist or an error
message will be generated.

Return to Top
13 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
CAPT_PATH=qualified directory Path to the local directory where the
pathname following CDC files are to be stored.
CCT file contains the change
registrations
CDEP file contains the
application names for PowerCenter
extractions that use ODBC
connections.
CDCT file contains information
about the PowerExchange condense.
This is a required parameter for CDC
and the directory path specified must
exist or an error message will be
generated.

ORACLEID=(collection id, Specifies the Oracle instance name


and connection information.
Oracle SID,
PowerExchange requires an
connect string, ORACLEID statement for each
Oracle instance which is being used
capture connect string)
with CDC. The maximum number of
ORACLEID statements contained in
the dbmover.cfg is 20.
- collection id An Oracle instance identifier that
matches the collection id specified in
the change registration for the Oracle
source table. The collection id is
required and there is no default
value.
- Oracle SID Name of the Oracle database that
contains the tables that have been
registered for change data capture.
The Oracle SID is required and there
is no default value.
- connect string The database service name and
could be the same as the Oracle SID.
The connect string is not a required
field; however, the best practice is to
enter it since it can be required under
certain conditions.
- capture connect string The database service name (could
be the same as the Oracle SID). The
connect string is not a required field;
however, the best practice is to enter
it since it can be required under
certain conditions.

Return to Top
14 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
CAPI_CONNECTION= The CAPX CAPI_CONNECTION
(NAME=name, statement specifies the parameters
used for continuous extraction from
TYPE=(CAPX,
condense files.
DFLTINST=collection id))
- NAME A unique name for the
CAPI_CONNECTION statement. The
maximum length is eight characters
and must be unique within all of the
CAPI-CONNECTION statements.
- CAPX Mandatory parameter.
- DFLTINST Specifies the PowerExchange
instance to process and must be
identical to the first parameter in the
ORACLEID statement.

CAPI_CONNECTION= The UOWC CAPI_CONNECTION


(NAME=name, statement is used to specify
parameters for the Unit of Work
TYPE= (UOWC,
(UOW) Cleanser. In the change
CAPINAME=capi name, stream, the changes from multiple
units of work are inter-mingled. The
RSTRADV=no of seconds,
UOW Cleanser reconstructs each
MEMCACHE=cache size)) interleaved unit of work (UOW) from
the change stream into complete
units of work in chronological order
based on change end time.
- NAME A unique name for the
CAPI_CONNECTION statement. The
maximum length is eight characters
and must be unique within all of the
CAPI-CONNECTION statements.
- UOWC Mandatory parameter
- CAPI NAME A unique name with a maximum
length of eight characters.
- RSTADV Specifies the number of seconds
PowerExchange waits before
advancing the restart tokens by
returning an empty Unit of Work
(UOW). Empty UOWs contain no
data, only restart tokens. This
parameter is very beneficial in cases
where the frequency of updates on
some tables is low in comparison to
other tables.

Return to Top
15 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
- MEMCACHE Specifies the memory cache in
Kilobytes allocated to the UOW
Cleanser to reconstruct complete
UOWs. The UOW Cleanser keeps all
changes in each UOW in cache until
the end-UOW (commit record) is
read. The default value is 1024.

CAPI_CONNECTION= The ORCL CAPI_CONNECTION


(NAME=name, statement is used to specify
parameters for Oracle CDC real-time
DLLTRACE=trace name,
extraction mode and for Oracle
TYPE=(ORCL, Condense.
CATINT=interval,
CATBEGIN=beginning time,
CATEND=ending time,
ARRAYSIZE=array size,
COMMITINT=commit interval,
REPNODE=node name,
BYPASSUF=yes or no,
ORACOLL=collection id))
- NAME The name specified MUST be
identical to the CAPINAME used in
the UOW CAPI statement.
- DLLTRACE The trace name MUST be identical to
the name specified in the TRACE
statement.
- ORCL Mandatory parameter
- CATINT The number of minutes between
attempts to write the Oracle catalog
to the Oracle redo log. The default
value is 1440, once a day.
- CATBEGIN The earliest time of day at which
Oracle can attempt to write the
Oracle catalog to the Oracle redo
log. The default value is 00:00.
- CATEND The latest time of day at which
Oracle can attempt to write the
Oracle catalog to the Oracle redo
log. The default value is 24:00
- ARRAYSIZE This parameter is used to control the
size of the prefetch array Oracle
capture uses to read the Oracle
archive logs. The default value is
100.
Return to Top
16 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
- COMMITINT Specifies the number of minutes
between Oracle Capture commit
points. The default value is 5.
- REPNODE Specifies the name of the NODE
statement in dbmover.cfg that points
to the Capture repository. The NODE
statement used should be for either
local or node1 where the IP address
is set to 127.0.0.1 which is a loop
back address.
- BYPASSUF If the Oracle instance has tables that
contain LOB columns that can be
included in the table row (this is the
default), then specify BYPASSUF=Y.
The default is N.
- ORACOLL Specifies the PowerExchange
instance to process and must be
identical to the first parameter in the
ORACLEID statement.

Tags: Data Synchronization (/index.php/component/customproperties


/tag/Project%20Types-Data-Synchronization)Data Warehousing (/index.php
/component/customproperties/tag/Project%20Types-Data-Warehousing)

Next (/index.php/best-practices-all/158-powerexchange-configuration
/443-m1bp524)

Latest 5 Articles Top 5 Articles About Us


RASCI Matrix Template Project Sponsor (/index.php Contact Us (/index.php
(/index.php /roles-all/92-m1r112) /contact-us)
/velocitytemplates/814-rasci- Business Analyst Informatica
matrix-template) (/index.php/roles-all/83- (http://www.informatica.com)
BOST Organization Model m1r101) Customer Community
Using RASCI (/index.php Technical Project Manager (https://network.informatica.com/)
/best-practices-all/177- (/index.php/roles-all/108- Marketplace
ent-arch-bpcat/813- m1r469) (https://marketplace.informatica.com/
bost-org-rasci) Data Steward/Data Quality Informatica Professional
Informatica v10 - Steward (/index.php/roles- Services
Administrator Tool Guide all/101-m1r379) (http://www.informatica.com
(/index.php/best-practices- Data Integration Developer /us/services-and-training
all/155-operations/812-v10- (/index.php/roles-all/82- /professional-services
admin-tool-guide) m1r102) /#fbid=kqutOHK30wS)
Data Governance Policy Service Offerings
Document Example (http://www.informatica.com
(/index.php/sample- /us/services-and-training
deliverables-all/810- /professional-services
dg-policy-example) /service-offerings
Informatica v10 Deciding /#fbid=kqutOHK30wS)
where to Develop and Return to Top
17 of 18 7/3/2017 5:59 PM
Velocity - PowerExchange for Oracle CDC https://velocity.informatica.com/index.php/best-practices-all/158-power...
Deploy (/index.php/best-
practices-all/149-
development-techniques
/809-v10-dev-deploy)

2017 Informatica All rights reserved

Return to Top
18 of 18 7/3/2017 5:59 PM

Anda mungkin juga menyukai