[ 
https://issues.apache.org/jira/browse/HADOOP-15423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16522431#comment-16522431
 ] 

Sean Mackrory commented on HADOOP-15423:
----------------------------------------

Fair enough - I thought we had seen some issues in which a file was created and 
then used as a directory and thought this would increase the risk of related 
issues, but I found the issue I was thinking of and it's actually not S3Guard 
related. Will commit shortly.

> Merge fileCache and dirCache into one single cache in LocalMetadataStore
> ------------------------------------------------------------------------
>
>                 Key: HADOOP-15423
>                 URL: https://issues.apache.org/jira/browse/HADOOP-15423
>             Project: Hadoop Common
>          Issue Type: Sub-task
>            Reporter: Gabor Bota
>            Assignee: Gabor Bota
>            Priority: Minor
>         Attachments: HADOOP-15423.001.patch
>
>
> Right now the s3guard.LocalMetadataStore uses two HashMap in the 
> implementation - one for the file and one for the dir hash.
> {code:java}
>   /** Contains directories and files. */
>   private Cache<Path, PathMetadata> fileCache;
>   /** Contains directory listings. */
>   private Cache<Path, DirListingMetadata> dirCache;
> {code}
> It would be nice to have only one hash instead of these two for storing the 
> values. An idea for the implementation would be to have a class with nullable 
> fields:
> {code:java}
>   static class LocalMetaEntry {
>     @Nullable
>     public PathMetadata pathMetadata;
>     @Nullable
>     public DirListingMetadata dirListingMetadata;
>   }
> {code}
> or a Pair (tuple):
> {code:java}
> Pair<PathMetadata, DirListingMetadata> metaEntry;
> {code}
> And only one hash/cache for these elements.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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

Reply via email to