I am facing problem in INFORMATICA
I have deployed informatica mapping in DEVELOPMENT server which is running fine and 5 million records processed in 2 or 3 minutes ,
however the same mapping transfer into PRODUCTION server which takes too much time and processed 3 million records in 45 minutes or more.
I really dont know why it is impact so much difference between DEVELOPMENT and PRODUCTION server.
Problem might be because in live enviornment there are few parametres like :
1. Target table might be accessed by more processes.
2. CPU utilization might be more as too many processes running at he same tym.
Check with your INFA Admin and also check the batch logs to get the proesses running at the same time.
1. Check your session config and see if its not running in "verbose" mode
2. Check where you are writing your logs. whether on the local drive/NAS
3. Check the Reader/Writer/Process threads and find out the Utilization %
4. Analyze your map and findout where the cache happens for lookup,aggreg, so on..
5. Check Integration service - Ascii mode is faster
6. You have not mentioned about the source/target - Next steps will depend on their types
But the problem is that we are still not sure about the sources and targets Habib is dealing with.
Your tip holds good if he is dealing with relational environment
DEV config: Client DB (System A) --> Informatica Server (System B) --> Target Database (System C)
Client DB (System A) --> Informatica Server (System D) --> Target Database (System E)
All machines are physical.
System B - 8GB, Xeon.
System C - 16GB, Xeon.
PROD servers are Blade servers:
System D - 32GB, Xeon.
System E - 128GB, Xeon.
The ETL process simply fetches data from Source, performs aggregations and validity checks, and writes to Target Database. There are about 300,000 rows imported on average daily.
Informatica config for both DEV and PROD is exactly the same. Indexes on target table on both DEV and PROD are the same.
Verbose logging is disabled on all machines. There are no other processes running on PROD as we have not gone live and no users have access to the system.
Hi Zafar, not sure if this issue is still open. Did you check the busy percentage for reader, writer and transformations. It will indicate the one thats taking most time. Also, the session log will provide information on when the sql was sent to the source database and when the first record was retrieved.
This issue is still persist.
I have monitoring worflow run time
In my mapping there is 2 sorter transformation like
Router----> 1 SORT----> Joiner
2 SORT Transformation
I have a router and 2 sort transformation from router and then joiner . the 2nd transformation
have all columns and sorting into branch and date whereas the 1st sorter transformation only have aggregated data and join branch and date too. then join through joiner transformation. I have monitoring sorter 2 transformation takes too much time to load data.
On QA server it runs perfectly but the problem in production server. plz help.