[HBASE-22292] PreemptiveFastFailInterceptor clean repeatedFailuresMap issue

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 3.0.0-alpha-1, 2.2.0, 2.3.0, 2.1.5
• Release Note: Adds new configuration hbase.client.failure.map.cleanup.interval which defaults to ten minutes.
• Assignee: zou
• Reporter: zou
• Created: 23/Apr/19 06:07
• Updated: 28/Apr/19 00:49
• Resolved: 23/Apr/19 15:01

Description

PreemptiveFastFailInterceptor do not set fastFailClearingTimeMilliSec, so in 

occasionallyCleanupFailureInformation function this branch

else if (now > entry.getValue().timeOfFirstFailureMilliSec
    + this.fastFailClearingTimeMilliSec) {

will be always be true,then the repeatedFailuresMap will be clean.

and in the constructor function 

public PreemptiveFastFailInterceptor(Configuration conf) {
  this.fastFailThresholdMilliSec = conf.getLong(
      HConstants.HBASE_CLIENT_FAST_FAIL_THREASHOLD_MS,
      HConstants.HBASE_CLIENT_FAST_FAIL_THREASHOLD_MS_DEFAULT);
  this.failureMapCleanupIntervalMilliSec = conf.getLong(
// this constant seem to set fastFailClearingTimeMilliSec, it may be a mistake.
      HConstants.HBASE_CLIENT_FAST_FAIL_CLEANUP_MS_DURATION_MS,
      HConstants.HBASE_CLIENT_FAST_FAIL_CLEANUP_DURATION_MS_DEFAULT);
  lastFailureMapCleanupTimeMilliSec = EnvironmentEdgeManager.currentTime();
}