DSXchange

HI ,

My job design
Load Type:Insert/Load

I am loading the data in production around 2 lks records ,but its not loading throughing below error>after running long time the below error .

[IIS-CONN-TERA-005023] Sync main_program: Startup time, 0:53; production run time, 5:47:39.timeout of 14,400 seconds expired while waiting for the other instances to reach state 1,004 (CC_TeraAdapter::waitForState, file CC_TeraAdapter.cpp, line 5,816)

Info:main_program: Startup time, 0:53; production run time, 5:47:39.

Please suggest for the same .

Share any ideas on the same PLease

No clue. Have you run this by your Teradata DBA? Official support provider?

Manu,

Need more details to be able to help you.
Are you using any load utility? (Access method bulk??)
Are you using parallel sync? if yes, what are the parameters used?
What config are you running this on? (Grid or cluster)?

The solution would vary depending on the details for above qns..
It may be just increasing sync parameters or changing config to have all the parallel processes to go in same phase.

-Chandra.

HI ,

Please find my teradata connector settings while loading the data
Total source count 17 Laksrecords in the sequntial file

Write Mode: Insert
Access method:Bulk
Table action :Append
Bulk access:Load
Error table 1:Yes,as parameter
Error table 2:Yes
,as parameter
Log table:yes as parameter
Work table: as parameter
start mode:clean
clean up mode:drop

Sync table :Yes
Sync table action:Replace
Sync table clen up:Keep

Example:
pTD_SYNC_USER=
pTD_SYNC_PASSWORD=
pTD_SYNC_DATABASE=
pTD_SYNC_ID=
pTD_SYNC_POLL=5
pTD_SYNC_TIMEOUT=1800
pTD_CHECK_TIMEOUT=0
pTD_END_TIMEOUT=0
pTD_PROGRESS=0
pTD_SESSMAX=8
pTD_SESSMIN=0
pTD_SLEEP=0
pTD_TENACITY=0
pTD_BULK_LOADTYPE=Load
pGDW_MACRO_DB=
pTD_BEFOREAFTER_SQL=No
pTD_ARRAYSIZE=1000
pTD_RECORDCOUNT=2000
pNUMROWS=50000000
pTD_ERROR_LIMIT=0

PLease find above params and suggest for the same

Is your input 17million records? If yes, It's not too big to get any issues with load.
We were running into issues when we load billions of records in one streach, so the approach we used to split the data to load in smaller chunks (~800 million in one execution) and loop thru to process all data.
How many nodes are you using to do the load?

1. Could you check if the head node can be removed from the Load to teradata (by using a config/ node pool contraints?)
2. I noticed that you have two target teradata connector stages. Are you using load utility on both? One all records and other the aggregated one?
I think you need to change your design to separate aggreagted summary data load into a separate process.

Thanks,
Chandra.