Teradata connector stage - immediate mode
Moderators: chulett, rschirm, roy
Teradata connector stage - immediate mode
Hi,
I'm using teradata connector to read data from teradata db. I have gone through documentation and came to know that to achieve parallel reads parallel synchronization property has to be set. I'm using a 2 node configuration. When i tried reading in immediate mode i'm not able to acheive parallelism even after setting parallel synchronization property. Please suggest.
I'm using teradata connector to read data from teradata db. I have gone through documentation and came to know that to achieve parallel reads parallel synchronization property has to be set. I'm using a 2 node configuration. When i tried reading in immediate mode i'm not able to acheive parallelism even after setting parallel synchronization property. Please suggest.
Hi,
No the job did not abort. Its completing normally. But the teradata stage is running in sequential mode.
I've monitored the job in Director. For other parallel stages apart from teradata, it showed operator name X 2.
example: peek x 2. It means it should be running on two nodes.
Also in the logs, under event detail operator name, node no. will be there. For other parallel stages logs messages came for both nodes. For teradata messages are from only node 0. Based on this i'm assuming that teradata reading is happening sequentially.
No the job did not abort. Its completing normally. But the teradata stage is running in sequential mode.
I've monitored the job in Director. For other parallel stages apart from teradata, it showed operator name X 2.
example: peek x 2. It means it should be running on two nodes.
Also in the logs, under event detail operator name, node no. will be there. For other parallel stages logs messages came for both nodes. For teradata messages are from only node 0. Based on this i'm assuming that teradata reading is happening sequentially.
The immediate mode runs on one node. Where as the bulk option can run on both multiple and single node depending on if the sync property is set to yes or no. The only reason which comes in my mind which may sound daft is the fact that immediate mode is prefered for small amount of data, hence should not make any difference if it runs in parallel or sequential mode where as bulk menthod is meant for large chunks of data.
maybe there is a patch for it.........
maybe there is a patch for it.........
Teradata Certified Master V2R5
Hi,
Yes. And i've read in a blog yesterday that immediate read works in sequential mode only.
I've tried a sample job using bulk mode to read data. I'm getting an error.
Error info -
Teradata_Connector_0,0: [IIS-CONN-TERA-005004] RDBMS code 21012: Could Not Find Message Catalog opermsgs.cat SQL statement: SELECT IRX.UPC_ID,1 AS ISPREPRICED FROM DW_DSS_DEV.CIC_UPC_ROG IRX (CC_TeraParallelTransporter::initiate, file CC_TeraParallelTransporter.cpp, line 1185)
I've googled on this error. could not get much help. Can you help me in resolving this.
Yes. And i've read in a blog yesterday that immediate read works in sequential mode only.
I've tried a sample job using bulk mode to read data. I'm getting an error.
Error info -
Teradata_Connector_0,0: [IIS-CONN-TERA-005004] RDBMS code 21012: Could Not Find Message Catalog opermsgs.cat SQL statement: SELECT IRX.UPC_ID,1 AS ISPREPRICED FROM DW_DSS_DEV.CIC_UPC_ROG IRX (CC_TeraParallelTransporter::initiate, file CC_TeraParallelTransporter.cpp, line 1185)
I've googled on this error. could not get much help. Can you help me in resolving this.
The bulk mode in with parallel sync might sometimes be slower than immediate mode itself. So just keep an eye on that.
Make sure you have the column ISPREPRICED in your DS job metadata and UPC_ID and ISPREPRICED are in the same order in which your selecting them..
Code: Select all
SELECT IRX.UPC_ID,1 AS ISPREPRICED FROM DW_DSS_DEV.CIC_UPC_ROG IRX
Teradata Certified Master V2R5