APT_BadAlloc: Heap allocation failed
Posted: Mon Dec 15, 2008 11:56 am
Hi ,
I have a simple job that extracts data from db2 database and loads in to sql server. The job design is source drs stage ---> copy stage ---> target drs stage (SQLserver) and from copy stage ---> aggregator --> audit table.(In aggregator iam taking count of source records).
The job is processing with around 6 million records. The source stage is DRS stage and the array size we have put is 5000 and the transaction isolation level is Read Uncommitted.The target DRS stage has properties array size 1000 and transaction size of 10000. Isolation level is Read committed. The job is aborting after processing around 4 million records and the fatal error is
sql_F4108,0: Caught exception from runLocally(): APT_BadAlloc: Heap allocation failed..
So i changed the target DRS stage properties as array size 1 and transaction size 10 .. the job finished with out any warning .. but my concern here is the throughput is just 350 rows/sec. so the job is taking longtime to finish..
I have searched the forum by key word heap allocation failed and many posts says that might be due to virtual memory.. Why the job is aborting when the array size and transaction size are 1000 and 10000 and why not with arraysize 1 and transaction size 10000. what are the recommended settings for the array size and transaction size to put in the target DRS stage not to abort and also to improve the performance of the job.
Please help me to solve this issue .. i have three more jobs failing with the same issue.
Thanks in Advance
I have a simple job that extracts data from db2 database and loads in to sql server. The job design is source drs stage ---> copy stage ---> target drs stage (SQLserver) and from copy stage ---> aggregator --> audit table.(In aggregator iam taking count of source records).
The job is processing with around 6 million records. The source stage is DRS stage and the array size we have put is 5000 and the transaction isolation level is Read Uncommitted.The target DRS stage has properties array size 1000 and transaction size of 10000. Isolation level is Read committed. The job is aborting after processing around 4 million records and the fatal error is
sql_F4108,0: Caught exception from runLocally(): APT_BadAlloc: Heap allocation failed..
So i changed the target DRS stage properties as array size 1 and transaction size 10 .. the job finished with out any warning .. but my concern here is the throughput is just 350 rows/sec. so the job is taking longtime to finish..
I have searched the forum by key word heap allocation failed and many posts says that might be due to virtual memory.. Why the job is aborting when the array size and transaction size are 1000 and 10000 and why not with arraysize 1 and transaction size 10000. what are the recommended settings for the array size and transaction size to put in the target DRS stage not to abort and also to improve the performance of the job.
Please help me to solve this issue .. i have three more jobs failing with the same issue.
Thanks in Advance