job was taking 1 hour to execute but now taking 15 hours

[former_member214007]

What I have to do when some job was taking 1 hour to execute but now taking 15 hours. How can I solve this problem?

[former_member211387]

As mentioned in this chain, the reasons can be numerous.

Some usual causes are

1. If there is a join in the dataflow check the join condition as there might be cartisian issue in cases which might cause the issue

2. Check the BODS Job server infrastructure setup. If the volume of data processed a lot higher than the original volume when it was taking 1 hour to execute, then you might need to upgrade the infrastructure. RAM/CPU or the combination of both can be upgraded

3. Check the indexing set up on the source and target tables. Specifically if the target table has a clustered index introduced, then it will take a lot of time to load as it will cluster after each commit.

The causes and remedies are endless.......

Raghu

[Former Member]

Hi,

Is there any calculation using the absolute time and elapsed time to check the time taken to read 1000 rows in a particular query.

[Former Member]

Hi,

     I had a similar issue where job executes for longer time.Could you please suggest in the monitor long how can we know that a particular query in the data flow is taking longer time to process.What makes the difference of the execution time and elapsed time.

Many thanks in advance.

[mageshwaran_subramanian]

Log on to Management Console  and choose operational dashboard to see the DF(s) which take longer time to execute than usual , then narrow down to the issue which causes the long execution time

[former_member198401]

Hi Sabat,

As Dirk mentioned, check your BODS Logs and try to find which data flow is taking longer to complete. There might be a possibility of index fragmentation of your fact tables(if there are any). There might be other reasons also so I am not asking to check the indexes directly.

Regards

Arun Sasi

[former_member187605]

There might be thousands of reasons. Try to narrrow down your problem using the monitor log output. Then proceed from there.