Hiran Chaudhuri created NUTCH-3082:
--------------------------------------
Summary: Improve logging in case of Nutch job failure
Key: NUTCH-3082
URL: https://issues.apache.org/jira/browse/NUTCH-3082
Project: Nutch
Issue Type: Improvement
Affects Versions: 1.21
Reporter: Hiran Chaudhuri
Please improve logging in case a Nutch job fails.
In NUTCH-3075 we found a situation where the injector job dies but the error
message is just a 'failed, reason: not available'. This does not help users at
all.
So the reason of job failure needs to be printed. Or, if the concise reason is
not available some hint needs to be printed where to look for the real reason.
This is especially true for distributed job execution.
One speciality that may have been overlooked previously:
In NUTCH-3075 the job failed during setup, not at runtime. So with this ticket
make sure either a reason or a hint pointing to the reason is printed for
* local execution failing during setup
* local execution failing during execution
* distributed execution failing during setup
* distributed execution failing during execution
These tests need to be applied to Injector but potentially to all Nutch hadoop
jobs.
I opened this issue based on
https://issues.apache.org/jira/browse/NUTCH-3075?focusedCommentId=17891812&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-17891812
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
