[
https://issues.apache.org/jira/browse/HBASE-23186?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Xiaolin Ha updated HBASE-23186:
-------------------------------
Summary: Fsck threads block master's exiting (was: Set Fsck thread be
daemon and close its OutputStream when master abort)
> Fsck threads block master's exiting
> -----------------------------------
>
> Key: HBASE-23186
> URL: https://issues.apache.org/jira/browse/HBASE-23186
> Project: HBase
> Issue Type: Bug
> Components: master
> Reporter: Xiaolin Ha
> Assignee: Xiaolin Ha
> Priority: Major
>
> HBASE-21072 imported to use HBaseFsck as default in hbase2.
> {code:java}
> if (this.conf.getBoolean("hbase.write.hbck1.lock.file", true)) {
> HBaseFsck.checkAndMarkRunningHbck(this.conf,
> HBaseFsck.createLockRetryCounterFactory(this.conf).create());
> }{code}
> But the fsck thread is not daemon，
> {code:java}
> public static Pair<Path, FSDataOutputStream>
> checkAndMarkRunningHbck(Configuration conf,
> RetryCounter retryCounter) throws IOException {
> FileLockCallable callable = new FileLockCallable(conf, retryCounter);
> ExecutorService executor = Executors.newFixedThreadPool(1);
> ...{code}
> This will make JVM not exit.
> We should set it be daemon and close the dfs output stream when master
> abort/stop.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
