[
https://issues.apache.org/jira/browse/HDFS-15009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16980106#comment-16980106
]
Ayush Saxena commented on HDFS-15009:
-------------------------------------
Logically, I too think it shouldn’t check for dir10. It should check dir1 and
its descendent
The only doubt i got by the words in the javadoc which said starting from so I
got a doubt that, doesn’t that mean /dir1* when we specify /dir1
Maybe hemanth if no further opinions for a day, you may clear the explanation
in the javadoc too while fixing this that it means the defendants.
> FSCK "-list-corruptfileblocks" return Invalid Entries
> -----------------------------------------------------
>
> Key: HDFS-15009
> URL: https://issues.apache.org/jira/browse/HDFS-15009
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: hemanthboyina
> Assignee: hemanthboyina
> Priority: Major
>
> Scenario : if we have two directories dir1, dir10 and only dir10 have
> corrupt files
> Now if we run -list-corruptfileblocks for dir1, corrupt files count for dir1
> showing is of dir10
> {code:java}
> while (blkIterator.hasNext()) {
> BlockInfo blk = blkIterator.next();
> final INodeFile inode = getBlockCollection(blk);
> skip++;
> if (inode != null) {
> String src = inode.getFullPathName();
> if (src.startsWith(path)){
> corruptFiles.add(new CorruptFileBlockInfo(src, blk));
> count++;
> if (count >= DEFAULT_MAX_CORRUPT_FILEBLOCKS_RETURNED)
> break;
> }
> }
> } {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
