[
https://issues.apache.org/jira/browse/HBASE-11276?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14630623#comment-14630623
]
Enis Soztutar commented on HBASE-11276:
---------------------------------------
This will eat up the CPU. Can you do smt like while(true) {
Threads.sleep(10000) }
{code}
+ while (true)
+ ;// loop here until got killed
{code}
These should not be needed. We can always assume a distributed cluster.
{code}
util.initializeCluster(isDistributed ? 1 : this.NUM_SLAVES_BASE);
+ if (!isDistributed) {
+ util.startMiniMapReduceCluster();
+ }
{code}
Also this:
{code}
+ protected int NUM_SLAVES_BASE = 3; // number of slaves when running test
with mini cluster
{code}
Did you test this with and without table / cfs ?
> Add back support for running ChaosMonkey as standalone tool
> -----------------------------------------------------------
>
> Key: HBASE-11276
> URL: https://issues.apache.org/jira/browse/HBASE-11276
> Project: HBase
> Issue Type: Task
> Affects Versions: 0.98.0, 0.96.0, 0.99.0
> Reporter: Dima Spivak
> Assignee: Yu Li
> Priority: Minor
> Attachments: HBASE-11276.patch, HBASE-11276_v2.patch
>
>
> [According to the ref
> guide|http://hbase.apache.org/book/hbase.tests.html#integration.tests], it
> was once possible to run ChaosMonkey as a standalone tool against a deployed
> cluster. After 0.94, this is no longer possible.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
