You are viewing a plain text version of this content. The canonical link for it is here.

Posted to hdfs-issues@hadoop.apache.org by "Andrew Wang (JIRA)" <ji...@apache.org> on 2014/09/18 00:12:34 UTC

[jira] [Updated] (HDFS-7078) Fix getEZForPath to work correctly with snapshots

     [ https://issues.apache.org/jira/browse/HDFS-7078?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Andrew Wang updated HDFS-7078:
------------------------------
    Attachment: hdfs-7078.001.patch

It turns out that my understanding of the EZ map was off. It's actually cleaned when the inode is removed, which happens when no snapshots reference it. However, this means we have an issue in listEZs, since it'll list snapshot EZs as well.

So, I added filtering in listEZ to filter out EZs that aren't present in the current FS state. This involves re-resolving the path to check that the inode IDs are the same.

> Fix getEZForPath to work correctly with snapshots
> -------------------------------------------------
>
>                 Key: HDFS-7078
>                 URL: https://issues.apache.org/jira/browse/HDFS-7078
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: encryption
>            Reporter: Andrew Wang
>            Assignee: Andrew Wang
>         Attachments: hdfs-7078.001.patch
>
>
> getEZForPath is checking if something is in an EZ based on the current filesystem state, rather than the snapshot filesystem state.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)