[ 
https://issues.apache.org/jira/browse/HADOOP-17267?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Karen Coppage updated HADOOP-17267:
-----------------------------------
    Description: 
HDFS reuses the same cached FileSystem object across the file system. If the 
client calls FileSystem.close(), closeAllForUgi(), or closeAll() (if it applies 
to the instance) anywhere in the system it purges the cache of that FS 
instance, and trying to use the instance results in an IOException: FileSystem 
closed.

It would be a great help to clients to see where and when a given FS instance 
was closed. I.e. in close(), closeAllForUgi(), or closeAll(), it would be great 
to see a DEBUG-level log of
 * calling method name, class, file name/line number
 * FileSystem object's identity hash (FileSystem.close() only)

For the full calling stack, turn on TRACE logging.

  was:
HDFS reuses the same cached FileSystem object across the file system. If the 
client calls FileSystem.close(), closeAllForUgi(), or closeAll() (if it applies 
to the instance) anywhere in the system it purges the cache of that FS 
instance, and trying to use the instance results in an IOException: FileSystem 
closed.

It would be a great help to clients to see where and when a given FS instance 
was closed. I.e. in close(), closeAllForUgi(), or closeAll(), it would be great 
to see a DEBUG-level log of
 * calling method name, class, file name/line number
 * FileSystem object's identity hash (FileSystem.close() only)


> Add debug-level logs in Filesystem#close
> ----------------------------------------
>
>                 Key: HADOOP-17267
>                 URL: https://issues.apache.org/jira/browse/HADOOP-17267
>             Project: Hadoop Common
>          Issue Type: Improvement
>          Components: fs
>    Affects Versions: 3.3.0
>            Reporter: Karen Coppage
>            Assignee: Karen Coppage
>            Priority: Minor
>              Labels: pull-request-available
>          Time Spent: 4h
>  Remaining Estimate: 0h
>
> HDFS reuses the same cached FileSystem object across the file system. If the 
> client calls FileSystem.close(), closeAllForUgi(), or closeAll() (if it 
> applies to the instance) anywhere in the system it purges the cache of that 
> FS instance, and trying to use the instance results in an IOException: 
> FileSystem closed.
> It would be a great help to clients to see where and when a given FS instance 
> was closed. I.e. in close(), closeAllForUgi(), or closeAll(), it would be 
> great to see a DEBUG-level log of
>  * calling method name, class, file name/line number
>  * FileSystem object's identity hash (FileSystem.close() only)
> For the full calling stack, turn on TRACE logging.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to