You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Lee Faris (JIRA)" <ji...@apache.org> on 2007/04/25 19:50:15 UTC
[jira] Created: (HADOOP-1295) hadoop-config.sh resolving symlinks
leads to errors
hadoop-config.sh resolving symlinks leads to errors
----------------------------------------------------
Key: HADOOP-1295
URL: https://issues.apache.org/jira/browse/HADOOP-1295
Project: Hadoop
Issue Type: Improvement
Components: scripts
Environment: RHEL3
Reporter: Lee Faris
My company uses a versioned deployment system where the final results are symlinked.
For example:The hadoop package would be located at this location on all boxes
/apollo/env/Hadoop/
This is a symlink generated by the system. The hardlink can look like this:
box1: /apollo/env/Hadoop -> /apollo/version/Hadoop-11114456
box2: /apollo/env/Hadoop -> /apollo/version/Hadoop-10039445
This piece of script from hadoop-config.sh resolves symlinks into hard links:
while [ -h "$this" ]; do
ls=`ls -ld "$this"`
link=`expr "$ls" : '.*-> \(.*\)$'`
if expr "$link" : '.*/.*' > /dev/null; then
this="$link"
else
this=`dirname "$this"`/"$link"
fi
done
I am not sure why this is done. Commenting out the code makes things work for our system.
I assume that was put in for a reason. Is there a solution for the original need for this code to that can work with our use case?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1295) hadoop-config.sh resolving symlinks
leads to errors
Posted by "Doug Cutting (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1295?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491810 ]
Doug Cutting commented on HADOOP-1295:
--------------------------------------
bq. Can it only run this code if HADOOP_HOME is not already set?
That should work. It looks like all of the stuff there to get an absolute path, etc. is only needed to determine HADOOP_HOME, so if that's already defined it could be skipped.
> hadoop-config.sh resolving symlinks leads to errors
> ----------------------------------------------------
>
> Key: HADOOP-1295
> URL: https://issues.apache.org/jira/browse/HADOOP-1295
> Project: Hadoop
> Issue Type: Improvement
> Components: scripts
> Environment: RHEL3
> Reporter: Lee Faris
> Priority: Minor
>
> My company uses a versioned deployment system where the final results are symlinked.
> For example:The hadoop package would be located at this location on all boxes
> /apollo/env/Hadoop/
> This is a symlink generated by the system. The hardlink can look like this:
> box1: /apollo/env/Hadoop -> /apollo/version/Hadoop-11114456
> box2: /apollo/env/Hadoop -> /apollo/version/Hadoop-10039445
> This piece of script from hadoop-config.sh resolves symlinks into hard links:
> while [ -h "$this" ]; do
> ls=`ls -ld "$this"`
> link=`expr "$ls" : '.*-> \(.*\)$'`
> if expr "$link" : '.*/.*' > /dev/null; then
> this="$link"
> else
> this=`dirname "$this"`/"$link"
> fi
> done
> I am not sure why this is done. Commenting out the code makes things work for our system.
> I assume that was put in for a reason. Is there a solution for the original need for this code to that can work with our use case?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1295) hadoop-config.sh resolving symlinks
leads to errors
Posted by "Lee Faris (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1295?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Lee Faris updated HADOOP-1295:
------------------------------
Priority: Minor (was: Major)
> hadoop-config.sh resolving symlinks leads to errors
> ----------------------------------------------------
>
> Key: HADOOP-1295
> URL: https://issues.apache.org/jira/browse/HADOOP-1295
> Project: Hadoop
> Issue Type: Improvement
> Components: scripts
> Environment: RHEL3
> Reporter: Lee Faris
> Priority: Minor
>
> My company uses a versioned deployment system where the final results are symlinked.
> For example:The hadoop package would be located at this location on all boxes
> /apollo/env/Hadoop/
> This is a symlink generated by the system. The hardlink can look like this:
> box1: /apollo/env/Hadoop -> /apollo/version/Hadoop-11114456
> box2: /apollo/env/Hadoop -> /apollo/version/Hadoop-10039445
> This piece of script from hadoop-config.sh resolves symlinks into hard links:
> while [ -h "$this" ]; do
> ls=`ls -ld "$this"`
> link=`expr "$ls" : '.*-> \(.*\)$'`
> if expr "$link" : '.*/.*' > /dev/null; then
> this="$link"
> else
> this=`dirname "$this"`/"$link"
> fi
> done
> I am not sure why this is done. Commenting out the code makes things work for our system.
> I assume that was put in for a reason. Is there a solution for the original need for this code to that can work with our use case?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1295) hadoop-config.sh resolving symlinks
leads to errors
Posted by "Lee Faris (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1295?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491804 ]
Lee Faris commented on HADOOP-1295:
-----------------------------------
Yes I get it. We need to find the hadoop root. Is there a way to just pass in the hadoop root? Can it only run this code if HADOOP_HOME is not already set?
Btw, it was nice to meet you yesterday at Amazon. Thanks for stopping by!
> hadoop-config.sh resolving symlinks leads to errors
> ----------------------------------------------------
>
> Key: HADOOP-1295
> URL: https://issues.apache.org/jira/browse/HADOOP-1295
> Project: Hadoop
> Issue Type: Improvement
> Components: scripts
> Environment: RHEL3
> Reporter: Lee Faris
> Priority: Minor
>
> My company uses a versioned deployment system where the final results are symlinked.
> For example:The hadoop package would be located at this location on all boxes
> /apollo/env/Hadoop/
> This is a symlink generated by the system. The hardlink can look like this:
> box1: /apollo/env/Hadoop -> /apollo/version/Hadoop-11114456
> box2: /apollo/env/Hadoop -> /apollo/version/Hadoop-10039445
> This piece of script from hadoop-config.sh resolves symlinks into hard links:
> while [ -h "$this" ]; do
> ls=`ls -ld "$this"`
> link=`expr "$ls" : '.*-> \(.*\)$'`
> if expr "$link" : '.*/.*' > /dev/null; then
> this="$link"
> else
> this=`dirname "$this"`/"$link"
> fi
> done
> I am not sure why this is done. Commenting out the code makes things work for our system.
> I assume that was put in for a reason. Is there a solution for the original need for this code to that can work with our use case?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1295) hadoop-config.sh resolving symlinks
leads to errors
Posted by "Doug Cutting (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1295?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491769 ]
Doug Cutting commented on HADOOP-1295:
--------------------------------------
I think this is for when someone makes a link to bin/hadoop in, e.g., ~/bin. When this is done, $this/.. is not the hadoop root but rather ~/bin, and the symlink must thus be resolved. Does that make sense?
> hadoop-config.sh resolving symlinks leads to errors
> ----------------------------------------------------
>
> Key: HADOOP-1295
> URL: https://issues.apache.org/jira/browse/HADOOP-1295
> Project: Hadoop
> Issue Type: Improvement
> Components: scripts
> Environment: RHEL3
> Reporter: Lee Faris
> Priority: Minor
>
> My company uses a versioned deployment system where the final results are symlinked.
> For example:The hadoop package would be located at this location on all boxes
> /apollo/env/Hadoop/
> This is a symlink generated by the system. The hardlink can look like this:
> box1: /apollo/env/Hadoop -> /apollo/version/Hadoop-11114456
> box2: /apollo/env/Hadoop -> /apollo/version/Hadoop-10039445
> This piece of script from hadoop-config.sh resolves symlinks into hard links:
> while [ -h "$this" ]; do
> ls=`ls -ld "$this"`
> link=`expr "$ls" : '.*-> \(.*\)$'`
> if expr "$link" : '.*/.*' > /dev/null; then
> this="$link"
> else
> this=`dirname "$this"`/"$link"
> fi
> done
> I am not sure why this is done. Commenting out the code makes things work for our system.
> I assume that was put in for a reason. Is there a solution for the original need for this code to that can work with our use case?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.