I am trying to load a large amount of data into an Oracle target. The sequential input file is about 620,000 records, and each record is about 2500 bytes wide. The job ran for about 5.5 hours, and then failed with many fatal errors. Actually, the status was set to 'has been reset', not 'failure'. Some of the errors are:
APT_CombinedOperatorController,0: Failure during execution of operator logic.
APT_CombinedOperatorController,0: Fatal Error: Tsort merger aborting: Scratch space full
cs}}},0: Failure during execution of operator logic.
APT_CombinedOperatorController,0: Fatal Error: Pipe read failed: short read
node_node1: combination of 9 operators [SEQL_Src_Customer_Ship; SEQL_Src_Customer_Ship_1;
ORCL_Tgt_Customer_Ship,0: Fatal Error: waitForWriteSignal(): Premature EOF on node ucas002 Socket operation on non-socket
When I divide the input file into 3 smaller ones, and load one at a time, each run is successful.
Does anyone know what is causing the scratch space error?
Thanks - cbloomfi
NOTICE: developerWorks Community will be offline May 29-30, 2015 while we upgrade to the latest version of IBM Connections. For more information, read our upgrade FAQ.
This topic has been locked.
1 reply Latest Post - 2009-12-27T16:55:37Z by Bromium
Pinned topic What would cause a scratch space error?
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2009-12-27T16:55:37Z at 2009-12-27T16:55:37Z by Bromium
Bromium 2700026WU23 PostsACCEPTED ANSWER
Re: What would cause a scratch space error?2009-12-27T16:55:37Z in response to SystemAdminWell, you'll have to look at your APT file to see where your scratch space is sitting. Your scratch space needs are dependent upon your input file and your job complexity. But basically... your scratch mount filled up.
With the cost of SAN / DISK / NAS these days... you can afford to beef up your scratch pool. If you can't, how did you afford DataStage to begin with? :)
FYI: Don't put your scratch pool on NAS.