[
https://issues.apache.org/jira/browse/HDFS-7933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14485837#comment-14485837
]
Chris Nauroth commented on HDFS-7933:
-------------------------------------
Hi Xiaoyu. This patch looks good. Here are a few comments:
# In {{BlockManager#chooseSourceDatanode}}, should the {{decommissioning}}
counter be incremented by {{countableReplica}}, like for the other counters?
{code}
if ((nodesCorrupt != null) && (nodesCorrupt.contains(node)))
corrupt += countableReplica;
else if (node.isDecommissionInProgress()) {
decommissioning++;
} else if (node.isDecommissioned()) {
decommissioned += countableReplica;
{code}
# Could the deprecated {{NumberReplicas#decommissionedReplicas}} be implemented
to forward the call to {{decommissionedAndDecommissioning}}? If so, then we
can eliminate the {{decommissionedReplicas}} member variable.
# The new test in {{TestFsck}} appears to create a {{StringBuilder}}, append
data to it, and then never use it for anything. Can it be removed?
> fsck should also report decommissioning replicas.
> --------------------------------------------------
>
> Key: HDFS-7933
> URL: https://issues.apache.org/jira/browse/HDFS-7933
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: namenode
> Reporter: Jitendra Nath Pandey
> Assignee: Xiaoyu Yao
> Attachments: HDFS-7933.00.patch, HDFS-7933.01.patch,
> HDFS-7933.02.patch
>
>
> Fsck doesn't count replicas that are on decommissioning nodes. If a block has
> all replicas on the decommissioning nodes, it will be marked as missing,
> which is alarming for the admins, although the system will replicate them
> before nodes are decommissioned.
> Fsck output should also show decommissioning replicas along with the live
> replicas.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
