[
https://issues.apache.org/jira/browse/HBASE-6280?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13402074#comment-13402074
]
Jieshan Bean commented on HBASE-6280:
-------------------------------------
Good finding, qiujaiwei. It seems a concurrency issue. We can see from the
method of HConnectionManager#processExecs:
{noformat}
Future<R> future = pool.submit(
new Callable<R>() {
public R call() throws Exception {
T instance = (T)Proxy.newProxyInstance(conf.getClassLoader(),
new Class[]{protocol},
invoker);
R result = callable.call(instance);
byte[] region = invoker.getRegionName();
if (callback != null) {
callback.update(region, r, result);
}
return result;
}
});
{noformat}
I think we should use a concurrent map instead of using TreeMap.
> why using treeMap at default implement with class Batch.Callback
> -----------------------------------------------------------------
>
> Key: HBASE-6280
> URL: https://issues.apache.org/jira/browse/HBASE-6280
> Project: HBase
> Issue Type: Bug
> Components: coprocessors
> Affects Versions: 0.92.0
> Environment: centos 6.2 hbase 0.92.0
> Reporter: qiujaiwei
>
> public <T extends CoprocessorProtocol, R> Map<byte[],R> coprocessorExec(
> Class<T> protocol, byte[] startKey, byte[] endKey,
> Batch.Call<T,R> callable)
> throws IOException, Throwable {
> final Map<byte[],R> results = new TreeMap<byte[],R>(
> Bytes.BYTES_COMPARATOR);
> coprocessorExec(protocol, startKey, endKey, callable,
> new Batch.Callback<R>(){
> public void update(byte[] region, byte[] row, R value) {
> results.put(region, value);
> }
> });
> return results;
> }
> when mulit region call the Batch.Callback ,the treemap should lockup.
> we meet this situation after we run 3 month.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
