[
https://issues.apache.org/jira/browse/HIVE-25112?focusedWorklogId=596229&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-596229
]
ASF GitHub Bot logged work on HIVE-25112:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 13/May/21 16:22
Start Date: 13/May/21 16:22
Worklog Time Spent: 10m
Work Description: belugabehr commented on a change in pull request #2270:
URL: https://github.com/apache/hive/pull/2270#discussion_r631935634
##########
File path: ql/src/java/org/apache/hadoop/hive/ql/txn/compactor/Worker.java
##########
@@ -250,49 +251,48 @@ void gatherStats() {
static final class CompactionHeartbeater extends Thread {
static final private Logger LOG =
LoggerFactory.getLogger(CompactionHeartbeater.class);
- private final AtomicBoolean stop = new AtomicBoolean();
private final CompactionTxn compactionTxn;
private final String tableName;
private final HiveConf conf;
- private final long interval;
+ private final long txnTimeout;
+
public CompactionHeartbeater(CompactionTxn compactionTxn, String
tableName, HiveConf conf) {
- this.tableName = tableName;
- this.compactionTxn = compactionTxn;
- this.conf = conf;
+ this.tableName = Objects.requireNonNull(tableName);
+ this.compactionTxn = Objects.requireNonNull(compactionTxn);
+ this.conf = Objects.requireNonNull(conf);
+
+ this.txnTimeout = MetastoreConf.getTimeVar(conf,
MetastoreConf.ConfVars.TXN_TIMEOUT, TimeUnit.MILLISECONDS);
- this.interval =
- MetastoreConf.getTimeVar(conf, MetastoreConf.ConfVars.TXN_TIMEOUT,
TimeUnit.MILLISECONDS) / 2;
setDaemon(true);
setPriority(MIN_PRIORITY);
setName("CompactionHeartbeater-" + compactionTxn.getTxnId());
}
+
@Override
public void run() {
+ LOG.debug("Heartbeating compaction transaction id {} for table: {}",
compactionTxn, tableName);
+
IMetaStoreClient msc = null;
try {
- // We need to create our own metastore client since the thrifts clients
- // are not thread safe.
+ // Create a metastore client for each thread since it is not thread
safe
msc = HiveMetaStoreUtils.getHiveMetastoreClient(conf);
- LOG.debug("Heartbeating compaction transaction id {} for table: {}",
compactionTxn, tableName);
- while(!stop.get()) {
+ while (true) {
msc.heartbeat(compactionTxn.getTxnId(), 0);
- Thread.sleep(interval);
+
+ // Send a heart beat before a timeout occurs. Scale the interval
based
+ // on the server's transaction timeout allowance
+ Thread.sleep(txnTimeout / 2);
}
+ } catch (InterruptedException ie) {
+ LOG.debug("Successfully stop the heartbeating the transaction {}",
this.compactionTxn);
} catch (Exception e) {
- LOG.error("Error while heartbeating txn {} in {}, error: ",
compactionTxn, Thread.currentThread().getName(), e.getMessage());
Review comment:
The thread name will be captured by the logging framework, and the
message too. Just need to pass in the entire Exception object.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 596229)
Time Spent: 20m (was: 10m)
> Simplify TXN Compactor Heartbeat Thread
> ---------------------------------------
>
> Key: HIVE-25112
> URL: https://issues.apache.org/jira/browse/HIVE-25112
> Project: Hive
> Issue Type: Improvement
> Reporter: David Mollitor
> Assignee: David Mollitor
> Priority: Minor
> Labels: pull-request-available
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Simplify the Thread structure. Threads do not need a "start"/"stop" state,
> they already have it. It is running/interrupted and it is designed to work
> this way with thread pools and forced exits.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
