[
https://issues.apache.org/jira/browse/HDDS-699?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16791051#comment-16791051
]
Tsz Wo Nicholas Sze commented on HDDS-699:
------------------------------------------
Thanks [~Sammi] for the 06 patch. The example in the javadoc of getLeaf is
very useful!
{code}
* / --- root
* / \
* / \
* / \
* / \
* dc1 dc2
* / \ / \
* / \ / \
* / \ / \
* rack1 rack2 rack1 rack2
* / \ / \ / \ / \
* n1 n2 n3 n4 n5 n6 n7 n8
{code}
Consider the following two sets of input:
# leafIndex = 2, excludedScope = /dc1/rack1, excludedNodes = \{/dc1/rack1/n1},
ancestorGen = 0
# leafIndex = 2, excludedScope = /dc1/rack1, excludedNodes = \{/dc1/rack1/n1},
ancestorGen = 2
In #1, the entire /dc1/rack1 is excluded so that the output is n4.
In #2, the entire /dc1 is excluded so that the output is n6.
Therefore, we should calculate the overlap and remove it, if there is any.
{code}
public Node getLeaf(int leafIndex, String excludedScope,
Collection<Node> excludedNodes, int ancestorGen) {
...
// build an ancestor(children) to exclude node count map
Map<Node, Integer> countMap =
getAncestorCountMap(excludedNodes, ancestorGen, currentGen);
// check overlap between excludedScope and countMap
if (excludedScope != null) {
for(Iterator<Map.Entry<Node, Integer>> i =
countMap.entrySet().iterator(); i.hasNext(); ) {
final Map.Entry<Node, Integer> entry = i.next();
final String path = entry.getKey().getNetworkFullPath();
if (path.startsWith(excludedScope)) {
// this node is a part of the excludedScope
i.remove();
} else if (excludedScope.startsWith(path)) {
// the excludedScope is already excluded by the this node
excludedScope = null;
}
}
}
// nodes covered by excluded scope
int excludedNodeCount = getExcludedScopeNodeCount(excludedScope);
...
}
{code}
> Detect Ozone Network topology
> -----------------------------
>
> Key: HDDS-699
> URL: https://issues.apache.org/jira/browse/HDDS-699
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Reporter: Xiaoyu Yao
> Assignee: Sammi Chen
> Priority: Major
> Attachments: HDDS-699.00.patch, HDDS-699.01.patch, HDDS-699.02.patch,
> HDDS-699.03.patch, HDDS-699.04.patch, HDDS-699.05.patch, HDDS-699.06.patch
>
>
> Traditionally this has been implemented in Hadoop via script or customizable
> java class. One thing we want to add here is the flexible multi-level support
> instead of fixed levels like DC/Rack/NG/Node.
--
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
