Return code from Pig is 0 even if the job fails when using -M flag
------------------------------------------------------------------
Key: PIG-1615
URL: https://issues.apache.org/jira/browse/PIG-1615
Project: Pig
Issue Type: Bug
Affects Versions: 0.7.0, 0.6.0
Reporter: Viraj Bhat
Fix For: 0.8.0
I have a Pig script of this form, which I used inside a workflow system such as
Oozie.
{code}
A = load '$INPUT' using PigStorage();
store A into '$OUTPUT';
{code}
I run this as with Multi-query optimization turned off :
{quote}
$java -cp ~/pig-svn/trunk/pig.jar:$HADOOP_CONF_DIR org.apache.pig.Main -p
INPUT=/user/viraj/junk1 -M -p OUTPUT=/user/viraj/junk2 loadpigstorage.pig
{quote}
The directory "/user/viraj/junk1" is not present
I get the following results:
{quote}
Input(s):
Failed to read data from "/user/viraj/junk1"
Output(s):
Failed to produce result in "/user/viraj/junk2"
{quote}
This is expected, but the return code is still 0
{code}
$ echo $?
0
{code}
If I run this script with Multi-query optimization turned on, it gives, a
return code of 2, which is correct.
{code}
$ java -cp ~/pig-svn/trunk/pig.jar:$HADOOP_CONF_DIR org.apache.pig.Main -p
INPUT=/user/viraj/junk1 -p OUTPUT=/user/viraj/junk2 loadpigstorage.pig
...
$ echo $?
2
{code}
I believe a wrong return code from Pig, is causing Oozie to believe that Pig
script succeeded.
Viraj
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
