Hadoop is an open source volunteer project under the Apache Software Foundation. We do not pretend it's perfect (yet!) and do recognise that there might be warts. This page describes how one can get help from the hadoop community.


Problem!

When you run into issues using hadoop some options are:

  • Frequently Asked Questions
    Please browse through the FAQ to check if there is a fix described there.
  • Jira
    Please browse through our bug-tracker to see if someone else has already reported a similar problem. If so, please consider filing a patch (if you can) or just vote for that issue to be fixed.
  • IRC Channel
    There is an IRC Channel dedicated to hadoop at irc.freenode.org. The name of the channel is #hadoop. Typically some users/developers hang around here, so this might be a good avenue for a quick chat, but please understand that they might be busy and may not be able to help you right-away!
  • User Mailing List
    core-user@hadoop.apache.org is the primary Hadoop User mailing list, and is the main avenue for users to interact with each other and Hadoop developers. This is typically where users run-to for help, suggestions etc. If you are hadoop-user, you (maybe) are already there. More information about hadoop mailing lists here. There are also searchable MailingListArchives.
    Please direct most of your concerns and questions to this mailing list.


Asking Questions

Here are a few tips on how to go about it:

  • First, and foremost, please mention the hadoop version you are using.
  • Explicitly provide the following information:
    • Any exceptions/errors encountered by your application.
    • Any exceptions/errors found while browsing through the hadoop-daemons' logs.
      • The hadoop daemons are: namenode & datanode (hdfs) and jobtracker & tasktracker (map-reduce).
      • The daemons' logs are found in the ${HADOOP_LOG_DIR} directory.
  • Don't cross post to the developer lists -they will ignore you.
  • Don't file a bug report if its just you not being able to get Hadoop working. They'll be closed as invalid.
  • Do try and debug it yourself.
  • Don't start resending the question every 30 minutes if you don't get an answer at first. Wait 24 hours (for everyone round the world to see it)
  • More generic tips on asking questions here.

Do realise that everyone on the user list is doing it co-operatively and not paid for it. You can't expect timely replies -many of the people will be in different time zones, and they are busy. Because they are busy you do have to do your homework -especially as they can't debug your server configuration for you. All they can do is give suggestions.


I got a response!

Once you have a response to your question(s), here is how you share your knowledge and contribute back:

  • I found a new bug
    Create a new bug, and consider fixing it yourself. We definitely hope you can learn about the hadoop project and contribute your expertise.
  • Problem Solved
    If you think others can benefit from your what you've just learnt please create a new documentation request and/or help us out by adding a section to our FAQ.

Thanks!

  • No labels