[
https://issues.apache.org/jira/browse/HDFS-14760?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16912393#comment-16912393
]
Erik Krogen commented on HDFS-14760:
------------------------------------
I think at least WARN level is warranted, given that this is an issue which
could potentially cause directories to exceed their quota or be blocked even
though their quota is not yet met. But I think your point about logging and not
doing anything is valid. Perhaps in cases such as this we should be fixing the
cached version after logging -- self-heal, but also warn that something is not
quite right.
> Log INFO mode if snapshot usage and actual usage differ
> -------------------------------------------------------
>
> Key: HDFS-14760
> URL: https://issues.apache.org/jira/browse/HDFS-14760
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: CR Hota
> Assignee: CR Hota
> Priority: Major
> Attachments: HDFS-14760.001.patch
>
>
> In DirectoryWithQuotaFeature#checkStoragespace code logs in error mode
> without throwing any exceptions or action and pollutes logs. This should be
> in INFO mode.
> {code}
> private void checkStoragespace(final INodeDirectory dir, final long
> computed) {
> if (-1 != quota.getStorageSpace() && usage.getStorageSpace() != computed)
> {
> NameNode.LOG.error("BUG: Inconsistent storagespace for directory "
> + dir.getFullPathName() + ". Cached = " + usage.getStorageSpace()
> + " != Computed = " + computed);
> }
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
