[
https://issues.apache.org/jira/browse/HADOOP-12173?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611186#comment-14611186
]
Inigo Goiri commented on HADOOP-12173:
--------------------------------------
Currently, the code does:
{code}
public void add(Node node) {
if (node==null) return;
int newDepth = NodeBase.locationToDepth(node.getNetworkLocation()) + 1;
netlock.writeLock().lock();
try {
String oldTopoStr = this.toString();
if( node instanceof InnerNode ) {
throw new IllegalArgumentException(
"Not allow to add an inner node: "+NodeBase.getPath(node));
}
if ((depthOfAllLeaves != -1) && (depthOfAllLeaves != newDepth)) {
LOG.error("Error: can't add leaf node " + NodeBase.getPath(node) +
" at depth " + newDepth + " to topology:\n" + oldTopoStr);
throw new InvalidTopologyException("Failed to add " +
NodeBase.getPath(node) +
": You cannot have a rack and a non-rack node at the same " +
"level of the network topology.");
}
{code}
This implies doing a toString for every single addition.
> NetworkTopology#add calls NetworkTopology#toString always
> ---------------------------------------------------------
>
> Key: HADOOP-12173
> URL: https://issues.apache.org/jira/browse/HADOOP-12173
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 2.7.0
> Reporter: Inigo Goiri
> Fix For: 2.7.1
>
>
> It always does a toString of the whole topology but this is not required when
> there are no errors. This is adding a very big overhead to large clusters as
> it's walking the whole tree every time we add a node to the cluster.
> HADOOP-10953 did some fix in that area but the errors is still there.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
