Loknath Priyatham Teja Singamsetty created PHOENIX-3083:
------------------------------------------------------------
Summary: Avoid additional handling of concurrency control inside
timertask of BuildIndexScheduleTask when ScheduledThreadPoolExecutor
executeAtFixedRate does it internally
Key: PHOENIX-3083
URL: https://issues.apache.org/jira/browse/PHOENIX-3083
Project: Phoenix
Issue Type: Improvement
Affects Versions: 4.8.0, 4.8.1
Reporter: Loknath Priyatham Teja Singamsetty
Assignee: Loknath Priyatham Teja Singamsetty
Priority: Trivial
MetaDataRegionObserver codesnippet where inprogress AtomicInteger is used for
concurrency control:
-----------------------------------------------------------------
/**
* Task runs periodically to build indexes whose
INDEX_NEED_PARTIALLY_REBUILD is set true
*
*/
public static class BuildIndexScheduleTask extends TimerTask {
// inProgress is to prevent timer from invoking a new task while
previous one is still
// running
private final static AtomicInteger inProgress = new AtomicInteger(0);
RegionCoprocessorEnvironment env;
public BuildIndexScheduleTask(RegionCoprocessorEnvironment env) {
this.env = env;
}
@Override
public void run() {
// FIXME: we should replay the data table Put, as doing a partial
index build would only add
// the new rows and not delete the previous index value. Also, we
should restrict the scan
// to only data within this region (as otherwise *every* region
will be running this code
// separately, all updating the same data.
RegionScanner scanner = null;
PhoenixConnection conn = null;
if (inProgress.get() > 0) {
LOG.debug("New ScheduledBuildIndexTask skipped as there is
already one running");
return;
}
try {
inProgress.incrementAndGet();
----------------------------------------------------
JAVADOC:
--------------
public ScheduledFuture<?> scheduleAtFixedRate(Runnable command,
long initialDelay,
long period,
TimeUnit unit)
Description copied from interface: ScheduledExecutorService
Creates and executes a periodic action that becomes enabled first after the
given initial delay, and subsequently with the given period; that is executions
will commence after initialDelay then initialDelay+period, then initialDelay +
2 * period, and so on. If any execution of the task encounters an exception,
subsequent executions are suppressed. Otherwise, the task will only terminate
via cancellation or termination of the executor. If any execution of this task
takes longer than its period, then subsequent executions may start late, but
will not concurrently execute.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
