[
https://issues.apache.org/jira/browse/HBASE-22739?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16951599#comment-16951599
]
Michael Stack edited comment on HBASE-22739 at 10/15/19 4:50 AM:
-----------------------------------------------------------------
HBASE-23173 is duplicate of this. Has a PR that just addresses the AIOBE.
Code has been same for long time so perturbed by some recent commit? Why it is
happening would necessitate spending time in balancer. The removeRegion is used
in a good few places. Meantime my cluster won't balance. I don't have RSGroup
enabled. Does have replicas enabled.
Here is PR w/ change and test. Needs work because some balancer tests are
failing. Closed it out while we figure what to do here.
https://github.com/apache/hbase/pull/723
was (Author: stack):
HBASE-23173 is duplicate of this. Has a PR that just addresses the AIOBE.
Code has been same for long time so perturbed by some recent commit? Why it is
happening would necessitate spending time in balancer. The removeRegion is used
in a good few places. Meantime my cluster won't balance. I don't have RSGroup
enabled. Does have replicas enabled..
> ArrayIndexOutOfBoundsException when balance
> -------------------------------------------
>
> Key: HBASE-22739
> URL: https://issues.apache.org/jira/browse/HBASE-22739
> Project: HBase
> Issue Type: Bug
> Components: Balancer
> Reporter: casuallc
> Priority: Major
> Fix For: 2.1.1
>
>
>
> {code:java}
> 2019-07-25 15:19:59,828 ERROR [master/nna:16000.Chore.1]
> hbase.ScheduledChore: Caught error
> java.lang.ArrayIndexOutOfBoundsException: 3171
> at
> org.apache.hadoop.hbase.master.balancer.BaseLoadBalancer$Cluster.removeRegion(BaseLoadBalancer.java:873)
> at
> org.apache.hadoop.hbase.master.balancer.BaseLoadBalancer$Cluster.doAction(BaseLoadBalancer.java:716)
> at
> org.apache.hadoop.hbase.master.balancer.StochasticLoadBalancer.balanceCluster(StochasticLoadBalancer.java:407)
> at
> org.apache.hadoop.hbase.master.balancer.StochasticLoadBalancer.balanceCluster(StochasticLoadBalancer.java:318)
> at org.apache.hadoop.hbase.master.HMaster.balance(HMaster.java:1650)
> at org.apache.hadoop.hbase.master.HMaster.balance(HMaster.java:1567)
> at
> org.apache.hadoop.hbase.master.balancer.BalancerChore.chore(BalancerChore.java:49)
> at org.apache.hadoop.hbase.ScheduledChore.run(ScheduledChore.java:186)
> at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
> at
> org.apache.hadoop.hbase.JitterScheduledThreadPoolExecutorImpl$JitteredRunnableScheduledFuture.run(JitterScheduledThreadPoolExecutorImpl.java:111)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> {code}
> should check if the regionIndex is valid when removeRegion,
> java:
> hbase-server/src/main/java/org/apache/hadoop/hbase/master/balancer/BaseLoadBalancer.java
> {code:java}
> int[] removeRegion(int[] regions, int regionIndex) {
> //TODO: this maybe costly. Consider using linked lists
> int[] newRegions = new int[regions.length - 1];
> int i = 0;
> for (i = 0; i < regions.length; i++) {
> if (regions[i] == regionIndex) {
> break;
> }
> if (i == regions.length - 1) {
> return Arrays.copyOf(regions, regions.length);
> }
> newRegions[i] = regions[i];
> }
> System.arraycopy(regions, i+1, newRegions, i, newRegions.length - i);
> return newRegions;
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
