[
https://issues.apache.org/jira/browse/HBASE-13965?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14623634#comment-14623634
]
Ted Yu commented on HBASE-13965:
--------------------------------
{code}
+ /** Config for balancing the cluster by each table */
{code}
'by each table' -> 'by table'
{code}
+ /** The name of the combined table when bytable is off*/
{code}
'combined table' -> 'combined tables'
{code}
+ // Use Most Recent Used(MRU) cache
{code}
'Most Recent' -> 'Most Recently'
{code}
+ synchronized (stochasticCosts) {
+ if (stochasticCosts != null) {
{code}
Looks like the null check is not needed.
{code}
+ private void updateStochasticCosts(TableName tableName, Double overall,
Double[] subCosts) {
{code}
Should we do early return if the table is system table ?
{code}
+ //TODO: should hbase:mespace be passed in by HMaster?
{code}
Can the above TODO be dropped ?
{code}
+ ((MetricsStochasticBalancer)
metricsBalancer).updateStochasticCost(tableName.getNameAsString(), "Overall",
+ "Overall cost", overall);
{code}
Since metricsBalancer would be cast in the loop that follows, you can use local
variable of type MetricsStochasticBalancer to save the casts in the loop.
{code}
+ if (null == tableName)
+ return;
{code}
nit: move return on the same line as if statement to save a pair of curly
braces.
> Stochastic Load Balancer JMX Metrics
> ------------------------------------
>
> Key: HBASE-13965
> URL: https://issues.apache.org/jira/browse/HBASE-13965
> Project: HBase
> Issue Type: Improvement
> Components: Balancer, metrics
> Reporter: Lei Chen
> Assignee: Lei Chen
> Attachments: HBASE-13965-v3.patch, HBASE-13965-v4.patch,
> HBASE-13965-v5.patch, HBASE-13965-v6.patch, HBASE-13965-v7.patch,
> HBASE-13965_v2.patch, HBase-13965-v1.patch,
> stochasticloadbalancerclasses_v2.png
>
>
> Today’s default HBase load balancer (the Stochastic load balancer) is cost
> function based. The cost function weights are tunable but no visibility into
> those cost function results is directly provided.
> A driving example is a cluster we have been tuning which has skewed rack size
> (one rack has half the nodes of the other few racks). We are tuning the
> cluster for uniform response time from all region servers with the ability to
> tolerate a rack failure. Balancing LocalityCost, RegionReplicaRack Cost and
> RegionCountSkew Cost is difficult without a way to attribute each cost
> function’s contribution to overall cost.
> What this jira proposes is to provide visibility via JMX into each cost
> function of the stochastic load balancer, as well as the overall cost of the
> balancing plan.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
