ORA-03113: end-of-file on communication channel

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
Anishn
Participant
Posts: 8
Joined: Tue Feb 01, 2011 11:24 am

ORA-03113: end-of-file on communication channel

Post by Anishn »

Hi ,

One of the job is failing in Production by giving the below error..


ORA-03113: end-of-file on communication channel

If i re-run then it was running fine.

I am getting the data from one table and loading it in to table using Oracle connector.Any one have any idea what was causing this issue.Please help me..
MarkB
Premium Member
Premium Member
Posts: 95
Joined: Fri Oct 27, 2006 9:13 am

Post by MarkB »

That's a generic Oracle error stating that the connection to the server timed out or was lost. It could have been an Oracle server issue or a network issue.
Anishn
Participant
Posts: 8
Joined: Tue Feb 01, 2011 11:24 am

Post by Anishn »

Do you have any idea what are the things i need to check with DBA's please let me know..
GJ_Stage
Participant
Posts: 131
Joined: Mon Oct 27, 2008 6:59 am

Post by GJ_Stage »

Make sure Database (Oracle) should be up and running while DS jobs are running.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

There really isn't much to check, as stated all it says is for some unknown reason it lost connection to the database while the job was running. One time, I wouldn't worry about it. If it persists, take the message to your DBAs and probably even Network guys, see if they can help.
Last edited by chulett on Tue Jul 30, 2013 7:42 am, edited 1 time in total.
-craig

"You can never have too many knives" -- Logan Nine Fingers
Anishn
Participant
Posts: 8
Joined: Tue Feb 01, 2011 11:24 am

Post by Anishn »

Hi,

Thank you for all your support.Still My issue is not resolved:


> SELECT SP.*
> FROM TABLE1 SP
> INNER JOIN (
> SELECT i.EMPID
> , i.YEAR,
> , MAX(i.EFF_END_DT) EFF_END_DT
> FROM TABLE1 i
> WHERE SRC='A'
> GROUP BY i.EMPID, i.YEAR
> ) C
> ON SP.EMPID=C.EMPID
> AND SP.YEAR=C.YEAR
> AND SP.SRC=C.SRC
> AND SP.EFF_END_DT=C.EFF_END_DT

The above is the Source query,Still we are trying to resolve. I showed to DBA. He is telliing the inner query was reading first time and getting Database Server timeout (connection was broken). Storing this data in Disk. When you run the second time it was working fine,because it was pulling the Stored data from memory. So i am trying to change the job design..Any one has any more opinions on this..please help me.
Thanks...
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Has this error occurred again? Your sql didn't cause anything and (IMHO) redesigning the job is a waste of time... unless your error is reproducible.
-craig

"You can never have too many knives" -- Logan Nine Fingers
Anishn
Participant
Posts: 8
Joined: Tue Feb 01, 2011 11:24 am

Post by Anishn »

Hi,

This error is coming repeatedly lots of other jobs started failed in the Production, and DBA's are trying to find out the issue. And more strange part is before any job fails first time if we restart the job they will work fine,but the new failed jobs are not running second also.It was spreading more in DataStage 8.5 Production. Can anyone help me..
Thanks..
kondeti
Premium Member
Premium Member
Posts: 67
Joined: Sat Mar 04, 2006 11:38 am

Post by kondeti »

Hello, Is it possible to replace OCI with DRS Stage? If So, try once.
Thanks
Kondeti
Anishn
Participant
Posts: 8
Joined: Tue Feb 01, 2011 11:24 am

Post by Anishn »

kondeti wrote:Hello, Is it possible to replace OCI with DRS Stage? If So, try once.
I was using the Oracle Connector stage and the DBA's concluded that they don;t have any solution for that. I can replace some other stage too..but it that works means i need to change lot of job designs...
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

I don't see any need to replace stages. If you can't get any joy from your DBAs and your network people (you have involved them as well, yes?) then get your official support provider involved.
-craig

"You can never have too many knives" -- Logan Nine Fingers
PaulS
Premium Member
Premium Member
Posts: 45
Joined: Fri Nov 05, 2010 4:38 am

Post by PaulS »

Did you find a cause for this?
rajendra123
Participant
Posts: 2
Joined: Tue Jul 13, 2010 11:37 am
Location: Chennai

Post by rajendra123 »

Try doing this.
1)Hold this job in the production environment
2)Ensure all other jobs complete. Let the dependent jobs on this job wait
3)Hold all the dependent jobs
4)Release this job alone and see if it works fine in the first run itself(without a rerun).

If it is a success, then obviously it is a network/traffic problem at the time of execution of this job, which you would have to resolve with the network team
Rajendra Tadavarthy
rajendra.ece@gmail.com
Post Reply