Data Flow Error - Named pipe error occurred: <The pipe has been ended.

[Former Member]

Hi Friends,

We get the below Error intermittently in a few of our Data Flows - " Named pipe error occurred: <The pipe has been ended"

We just upgraded from BODS 4.0 to BODS 4.2 and this Error has started appearing and the Data Flow fails with the Error. This happens only with our long running data flows that run from 2-3 hours. The shorter jobs run fine. Also this was rare in V 4.0 but after coming onto 4.2 this is happening more often. I did refer the KBA 1305751. We are using the correct Oracle Client. 

Error Snapshot:

(14.2.5) 10-20-15 11:33:14 (W) (000:000) ABT-200201: Job server killed job with pid <11260>, execName <10_20_2015_11_09_20_116__b6cdf656_b004_4ee3_b334_19940cbd88ed> due to job server shutdown.

(14.2) 10-20-15 11:33:14 (E) (8036:6340) FIL-080134: |Dataflow Adhoc_DF_IML_Previous_Day_Shipment|Pipe Listener for Adhoc_DF_IML_Previous_Day_Shipment_1_3

                                                     Named pipe error occurred: <The pipe has been ended.

                                                     >

Monitor Log Snapshot:

@FIL-080134, ERROR, 0, 0.000, 1432.382, 0.000, 0.000, 0:0, 0.000, 0, FIL-080134: |Dataflow Adhoc_DF_IML_Previous_Day_Shipment|Pipe Listener for Adhoc_DF_IML_Previous_Day_Shipment_1_3: Named pipe error occurred: <The pipe has been ended.

>

Any inputs will be very much appreciated.

Thanks,

SA

[former_member198401 · Accepted solution]

Stephen,

Greetings for the day!!

Below are the basic reasons Named Pipe errors :

1. The dataflow is having Degree of Parallelism, set as 4 and also the dataflow has got one or more 'Run as separate process' set in it.

2. There might be lookups set with 'Run as separate process'.

3. When the target table is partitioned and 'Enable partition' is set.

4. Loss of communication between Job server and Oracle database(In your case this might be the issue as per the error log which you have attached)

As a workaround you can do the below changes

1) Change the Data Flow property in the job to ‘Pageable’ from ‘In-Memory’

2) Add the below line to the DSConfig file on the job server under the [AL_ENGINE] section and see if that helps

NamedPipeWaitTime = 500

*****CAUTION*****

Always back up dsconfig.txt file

Regards

Arun Sasi

[Former Member · Accepted solution]

Thank You Arun.

Removing the "Run as separate process" from the Query Tranforms and Lookups has fixed the Pipe Error. we also set the Data Flow  property to Pageable.

Thanks you for your inputs and helping us with the resolution.

Thanks,

SA

[former_member198401 · Accepted solution]

Cool!!

Regards

Arun Sasi

[Former Member]

The trace messages are usually a bit more helpful in this error case, but when we get this error, its usually when we try to put a null into a non-nullable field, a primary key constraint, or a check constraint on the data base side.

edit: that being said, your snapshot shows that "job server shutdown" meaning that it looks like whatever server you run your jobs on got shut off