In the following documentation:
- recommended: designates an environment recommended by Talend based on our experiences and customer usage;
- supported: designates a supported environment for use with the listed component or service;
- supported with limitations: designates an environment that is supported by Talend but with certain conditions explained in notes.
|Support type||Database (64-bit)||Note|
|Recommended||Oracle 10g R2 to 12c, 18c and 19c||-|
|Supported||Microsoft SQL Server 2008 R2 to 20171||-|
|PostgreSQL 10 or newer1||-|
1 Amazon RDS is supported.
If you reuse an existing database on Oracle
- The character set of the database must be AL32UTF8 (UTF8).
- To find out what exact Oracle edition/version is actually
sqlplus SYS@<DB-NAME> as SYSDBA select banner from v$version where BANNER like '%Edition%';
- To find out how much memory is actually available to your Oracle database, it is important to first understand How Oracle's memory is configured and used.
- To find out how much processing CPU/Cores is actually available to the Oracle database,
query the table v$parameter for the value of cpu_count, or query the table v$license as
sqlplus SYS@<DB-NAME> as SYSDBA select * from v$license;
If you reuse an existing database on Microsoft SQL Server
- The database must be configured to interpret SQL in a case insensitive manner. The case insensitive collation must be Latin1_General_CI_AS.
- You have applied the current Microsoft patches.
If you reuse an existing database on PostgreSQL
- libxml is needed, you might need to rebuild PostgreSQL with it.
Database administrator privileges are required to install, setup and uninstall the database.
In general, you must always install the latest service packs for a given database version before creating Talend Data Catalog database.
Oracle 12c is recommended for large scale deployments where indexes must be kept in memory to avoid any disk I/O, and therefore obtain fast search response time. In such case, the Oracle 12c Server must have enough memory (for KEEP buffer pool), a fast-enough processor speed (for a single search), and enough cores (for multiple user searches in parallel).