Why tOracleOutput using existing connection (tOracleConnection) is getting poor performance in comparison to the tOracleOutput which is standalone(it contains the db details in it , hostname ,service name,username ,portnumber and password)?
What is the logic behind this slowness?
What is the correct situation to these differrent options.
How did you measure the performance difference between two scenarios? We will have to make sure that the system resources are aligned same for both cases. If there are any underling queries running in Oracle database, that also can change the statistics.
tOracleconnection just helps to use an existing open connection to do subsequent jobs. Could you check the details DB statistics with the help of an Oracle DBA and let us know whether issue persists?
Even in a simple scenario , I see the difference
I checked with simple one to one flow,
tFileInputDelimited--->tMap--->tOracleOuput(using Use an existing connection) its **bleep** slow.
When I update the job like below
tFileInputDelimited--->tMap--->tOracleOuput(in-built connection details), the throughout is high .
Whats happening internally , I'm not sure ?
Please check the batch size of the tOracleOutput component whether it is too low. Could you please quantify the time difference between both ways of data loading?
If you are absolutely sure that all the other parameters remain same (CPU utilization, number of connections, commit size memory usage in Talend and DB servers etc.) , culd you please log a JIRA ticket with details so that R&D team can have a look into it. The link for JIRA is as shown below.
Yes even I faced performance issues with use an existing conn enabled for tOracleOutput during the data loading.If it leads to performance issues then whats the use of tOracleConnection when I would like to connec to same db multiple times and ending up with performance issues.