[uuliver]

WFLYJCA0019: Stopped Driver service with driver-name = oracle

the vdb state is always loading ,but one hour later,the cmd console show

WFLYJCA0019: Stopped Driver service with driver-name = oracle

[org.jboss.as] (MSC service thread 1-5)  WFLYSRV0050: WildFly Full 9.0.2.Final (WildFly Core 1.0.2.Final) stopped in 554153ms

[uuliver]

ERROR [org.jboss.as.controller.management-operation] (XNIO-1 task-6)  WFLYCTL0013: Operation ("get-schema") failed - address: ([("subsystem" => "teiid")]) - failure description: "TEIID50096 VDB ln628.1 does not exist or is not ACTIVE. "

[shawkins]

> TEIID60011 No column found on table Epbank.MDSYS.SYS_NTj4Pd6KReXlTgQOUKi4pslA== with name NESTED_TABLE_ID

It is a system table that the oracle driver is providing incomplete metadata for.  It should be excluded from import using a schemaPattern and/or excludeTables settings.

> Try "importer.importApproximateIndexes" to false as another property?

importApproximateIndexes should be left as true - especially for oracle.  When set to false any attempt to get index info, which includes counts, will do full row counts against the tables rather than consulting statistics.

Let's see we can get the import properties correct to pull in only the subset of metadata you are interested in, and then see if additional issues occur.  Do you now have some combination of tableTypes, schemaPattern, and excludeTables you know to try, or do you need additional help in limiting the metadata import?

[uuliver]

Thank you very much for your answer，It helps me a lot,i want to know about the import properties,Dose the tableNamePattern  allow set sevral table,and how to set.

Second questions is ,why i deploy vdb ,the souce model loading long,What caused it,and how to set the properties can improves speed

[shawkins]

> It helps me a lot,i want to know about the import properties,Dose the tableNamePattern  allow set sevral table,and how to set.

See JDBC Translators · Teiid Documentation the tableNamePattern is expected to be a SQL like pattern that sets what tables you are interested in importing.  Generally tableTypes and schemaPattern will be more useful than tableNamePattern.

> Second questions is ,why i deploy vdb ,the souce model loading long,What caused it,and how to set the properties can improves speed

Especially for oracle having no import properties to limit the metadata import is problematic.  Oracle will allow the crawling of the entire database metadata and several of their JDBC DatabaseMetaData interrogation methods are not efficient.  So we suspect that you are staying in the loading state due to fetching the metadata.  If the logging level were increased to debug, you should be able to confirm what the importer / vdb load process is doing.

[uuliver]

thank you for your answer ,and how to set the logging level to debug.

and what means the vdb's property Dynamic what different between true and false

set true ,when restart the server the vdb loading again

[rareddy]

Add the following to standalone-teiid.xml file in the "logging" subsystem element, then restart the server.

            <logger category="org.teiid.CONNECTOR">

                <level name="DEBUG"/>

            </logger>

You can also use Web Console to add the same.

[uuliver]

thank you  and what means the vdb's Dynamic property .what different between true and false.

[uuliver]

i set the logger to debug,and  the log of debug information is below

2017-07-10 09:44:17,942 INFO  [org.teiid.RUNTIME] (Worker0_async-teiid-threads0)  TEIID50030 VDB ln630epdm.1 model "Epbank" metadata loaded. End Time: 17-7-10  9:44

2017-07-10 10:13:33,598 INFO  [org.teiid.RUNTIME] (Worker1_async-teiid-threads1)  TEIID50030 VDB ln630epdm.1 model "Geol" metadata loaded. End Time: 17-7-10 10:13

the first souce model have more tables than the second ,but why ,the second loaded More slowly

[rareddy]

That depends on the type of database, JDBC driver, load the database or it could be latency in the network. If you want to figure out you need to a JAVA profiler to see where the time delay is. As per Teiid, it gives fair priority to both sources in terms of reading metadata.