hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "dhruba borthakur (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-1162) Reduce the time required for a checkpoint
Date Tue, 18 May 2010 22:22:43 GMT

    [ https://issues.apache.org/jira/browse/HDFS-1162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12868871#action_12868871
] 

dhruba borthakur commented on HDFS-1162:
----------------------------------------

A cluster with 70 million files. The fsimage is 11GB. 

1. The secondary namenode fetches this file from the namenode via http. This takes about 8
minutes. 
2. Then the secondary namenode uploads this file back to the namenode. This takes another
8 minutes

Thus, a total of 16+ minutes is needed to do a checkpoint. This leads to a scenario where
the checkpointing cannot keep up with the rate of growth of the transaction log. Very Bad!


> Reduce the time required for a checkpoint
> -----------------------------------------
>
>                 Key: HDFS-1162
>                 URL: https://issues.apache.org/jira/browse/HDFS-1162
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: name-node
>            Reporter: dhruba borthakur
>            Assignee: dhruba borthakur
>
> The checkpoint time increases linearly with the number of files in the cluster. This
is a problem with large clusters.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message