You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Thomas Koch (JIRA)" <ji...@apache.org> on 2011/01/10 21:46:48 UTC
[jira] Created: (HADOOP-7092) Consolidate startup shell scripts
Consolidate startup shell scripts
---------------------------------
Key: HADOOP-7092
URL: https://issues.apache.org/jira/browse/HADOOP-7092
Project: Hadoop Common
Issue Type: Task
Components: scripts
Reporter: Thomas Koch
Priority: Minor
- Today we struggled ~1-2 hours to find out that the Cloudera init scripts have a bug in the shell code (bug filled by my colleague)
- Many projects that started at hadoop copied and adapted the shell startup code. So there's a lot of code duplication between hbase, zookeeper, hadoop and maybe others
- There already are some open issues regarding to shell code
- The shell code isn't properly tested (automaticaly) and will first probably fail at customer side
Would it make sense to build a shell library of most often used functionality for startup scripts? Is there already such a library somewhere?
This issue should collect thoughts in this area.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-7092) Consolidate startup shell scripts
Posted by "Thomas Koch (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12979749#action_12979749 ]
Thomas Koch commented on HADOOP-7092:
-------------------------------------
http://stackoverflow.com/questions/78497/design-patterns-or-best-practices-for-shell-scripts
> Consolidate startup shell scripts
> ---------------------------------
>
> Key: HADOOP-7092
> URL: https://issues.apache.org/jira/browse/HADOOP-7092
> Project: Hadoop Common
> Issue Type: Task
> Components: scripts
> Reporter: Thomas Koch
> Priority: Minor
>
> - Today we struggled ~1-2 hours to find out that the Cloudera init scripts have a bug in the shell code (bug filled by my colleague)
> - Many projects that started at hadoop copied and adapted the shell startup code. So there's a lot of code duplication between hbase, zookeeper, hadoop and maybe others
> - There already are some open issues regarding to shell code
> - The shell code isn't properly tested (automaticaly) and will first probably fail at customer side
> Would it make sense to build a shell library of most often used functionality for startup scripts? Is there already such a library somewhere?
> This issue should collect thoughts in this area.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] [Commented] (HADOOP-7092) Consolidate startup shell scripts
Posted by "Steve Loughran (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-7092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13185625#comment-13185625 ]
Steve Loughran commented on HADOOP-7092:
----------------------------------------
I'm going to ask something more subversive: why are we using Bash?
Could we say "use python" for the scripts. It's a better language from a maintenance perspective, because it is closer to the java language that most developers in the Hadoop stack will have encountered. On a windows system, it will add one place that Cygwin is not needed.
It would also offer a more powerful base set of functions in the python runtime -things that won't need team maintenance (and the delays to trickle down to other project/cost of maintaining forks).
I'd be tempted to try writing some .py versions of the basic script if anyone else was interested.
> Consolidate startup shell scripts
> ---------------------------------
>
> Key: HADOOP-7092
> URL: https://issues.apache.org/jira/browse/HADOOP-7092
> Project: Hadoop Common
> Issue Type: Task
> Components: scripts
> Reporter: Thomas Koch
> Priority: Minor
>
> - Today we struggled ~1-2 hours to find out that the Cloudera init scripts have a bug in the shell code (bug filled by my colleague)
> - Many projects that started at hadoop copied and adapted the shell startup code. So there's a lot of code duplication between hbase, zookeeper, hadoop and maybe others
> - There already are some open issues regarding to shell code
> - The shell code isn't properly tested (automaticaly) and will first probably fail at customer side
> Would it make sense to build a shell library of most often used functionality for startup scripts? Is there already such a library somewhere?
> This issue should collect thoughts in this area.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira