Jan 22, 2011

Errors in Prod@at Work

1.ORA-Value is too long
Trying to insert a 40,978 value in a column where data type is small int.
Fix: Change the data type

2. ora-26002 table has index defined upon it
Reason: At session the load is BULK mode and the target table has index on it.
Fix: Change the Load mode to Normal.

2/a). 'ORA-01555: snapshot too old: rollback segment number 2 with name "XXXXX" too small'
Ask the DBA to increase the undo_retention.

b).ORA-24806: LOB form mismatch
When sourc data changes while laoing to target this eror will come.
Run in different time interval where source won't change the data while the job is running.

3.ORA-03106:fatal two-task communication protocol error
rerun the job again. Solved.

4. Error: Cache file  needed by named Lookup
Transformation  in this mapping
appears to be created by an unnamed cache lookup transformation.

delete the cache files and run again. solve the problem.

5. CMN_1252 Error in looup creation

This error generally comes when Informatica Server is unable to provide sufficient space to the lookup for lookup cache creation.

The way We used is if there are multiple sessions are running then wait till there execution then run this failed session separately.
And if multiple sessions failed then also run each session one by one. So server can get time to accumulate the sufficient space which is required and is released from other sessions as on there completion.

6. error code ORA-03135 i.e the connection is lost or timeout
solution:
The keepalive parameter in the oracle configuration in the Informatica server is not set properly. we have set this value in such a way that the TCP/IP channel between Informatica Server and Oracle Database will never be in idle situation. This keepalive parameter contains a time (in ms) which is basically the duration between two PINGs between these two servers. This parameter can be found at the following locations.

/usr/sbin/ndd /dev/tcp tcp_keepalive_interval

Apart from that we have also included one entry "ENABLE = BROKEN" in the entry for that particular database in the tnsnames.ora file exists in the Informatica server.
Other possibilty is if the size of the table is very high.

After applying these two scenarios, the long run query is running fine.

7. CMN_1769 Inconsistent Recovery Cache
Solution: The task'sRecovery Strategy attribute (in the Properties tab of Edit Tasks)
needs to be "Fail task and continue workflow".


8. ORA-12514:TNS: listener does not currently know of service requested in connect descriptor
Sol: Restart the Job again.

9. ORA-12569: TNS:packet checksum failure
Sol: By removing the unwanted columns with CLOB and BLOB data types solved the issue.

10. sqlstate = 40003 
Database driver error... 
Function Name : Fetch 
SQL Stmt : select 

This error is mainly due to the size of the data when you fetch.
The reasons are if you have any CLOB or BLOB fileds present hten the size will be huge
the sol is use dbms.lob function for those kind of columns or else change the session properties like DTM Buffersize and Maximum Memory Allowed For Auto Memory Attributes
















No comments:

Post a Comment