[ 
https://issues.apache.org/jira/browse/HDFS-15196?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17070193#comment-17070193
 ] 

Fengnan Li commented on HDFS-15196:
-----------------------------------

Thanks [~elgoiri] 

I fixed point one, but it is hard to simplify the logic of point 2 here since 
what we are doing is 
 * insert mount points into one batch of listing where they fit by order
 * append mount points when remaining is 0 and these points are bigger to make 
sure previously inserted batch not inserted again.

I tried but couldn't find a better way to express this logic and other ways 
either lose the order or have duplicates.

> RBF: RouterRpcServer getListing cannot list large dirs correctly
> ----------------------------------------------------------------
>
>                 Key: HDFS-15196
>                 URL: https://issues.apache.org/jira/browse/HDFS-15196
>             Project: Hadoop HDFS
>          Issue Type: Bug
>            Reporter: Fengnan Li
>            Assignee: Fengnan Li
>            Priority: Critical
>         Attachments: HDFS-15196.001.patch, HDFS-15196.002.patch, 
> HDFS-15196.003.patch, HDFS-15196.003.patch, HDFS-15196.004.patch, 
> HDFS-15196.005.patch, HDFS-15196.006.patch, HDFS-15196.007.patch, 
> HDFS-15196.008.patch, HDFS-15196.009.patch, HDFS-15196.010.patch, 
> HDFS-15196.011.patch, HDFS-15196.012.patch
>
>
> In RouterRpcServer, getListing function is handled as two parts:
>  # Union all partial listings from destination ns + paths
>  # Append mount points for the dir to be listed
> In the case of large dir which is bigger than DFSConfigKeys.DFS_LIST_LIMIT 
> (with default value 1k), the batch listing will be used and the startAfter 
> will be used to define the boundary of each batch listing. However, step 2 
> here will add existing mount points, which will mess up with the boundary of 
> the batch, thus making the next batch startAfter wrong.
> The initial fix is just to append the mount points when there is no more 
> batch query necessary, but this will break the order of returned entries. 
> Therefore more complex logic is added to make sure the order is kept. At the 
> same time the remainingEntries variable inside DirectoryListing is also 
> updated to include the remaining mount points.



--
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