[
https://issues.apache.org/jira/browse/HDFS-13233?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16387202#comment-16387202
]
Íñigo Goiri commented on HDFS-13233:
------------------------------------
Thanks [~striver.wang] for the report and the fix.
I would probably make the full condition a private function for readability.
Please double check as I haven't gone through it carefully but I would do
something like:
{code}
while (entry != null && isParentEntry(path, entry.getKey())) {
...
private boolean isParentEntry(final String path, final String parent) {
if (!path.startsWith(parent)) {
return false;
}
if (path.equals(parent)) {
return true;
}
return path.charAt(entry.getKey().length()) == '/';
}
{code}
I would also add checks to avoid having the catch for
{{StringIndexOutOfBoundsException}}.
> RBF：getMountPoint doesn't return the correct mount point of the mount table
> ---------------------------------------------------------------------------
>
> Key: HDFS-13233
> URL: https://issues.apache.org/jira/browse/HDFS-13233
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs
> Affects Versions: 3.2.0
> Reporter: wangzhiyuan
> Priority: Major
> Attachments: HDFS-13233.001.patch
>
>
> Method MountTableResolver->getMountPoint will traverse mount table and return
> the first mount point which the input path starts with, but the condition is
> not sufficient.
> Suppose the mount point table contains: "/user" "/user/test" "/user/test1",
> if the input path is "/user/test111", the return mount point is
> "/user/test1", but the correct one should be "/user".
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
