[
https://issues.apache.org/jira/browse/HBASE-18192?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ashu Pachauri updated HBASE-18192:
----------------------------------
Attachment: HBASE-18192.master.001.patch
HBASE-18192.branch-1.001.patch
HBASE-18192.branch-1.3.003.patch
Uploading patches for master and branch-1.
Also, uploading a new patch for branch-1.3 which has an extra check for proper
cleanup of recovered queues. Please use HBASE-18192.branch-1.3.003.patch for
branch-1.3.
> Replication drops recovered queues on region server shutdown
> ------------------------------------------------------------
>
> Key: HBASE-18192
> URL: https://issues.apache.org/jira/browse/HBASE-18192
> Project: HBase
> Issue Type: Bug
> Components: Replication
> Affects Versions: 1.3.1, 1.2.6
> Reporter: Ashu Pachauri
> Assignee: Ashu Pachauri
> Priority: Blocker
> Fix For: 2.0.0, 1.4.0, 1.3.2, 1.2.7
>
> Attachments: HBASE-18192.branch-1.001.patch,
> HBASE-18192.branch-1.3.001.patch, HBASE-18192.branch-1.3.002.patch,
> HBASE-18192.branch-1.3.003.patch, HBASE-18192.master.001.patch
>
>
> When a recovered queue has only one active ReplicationWorkerThread, the
> recovered queue is completely dropped on a region server shutdown. This will
> happen in situation when
> 1. DefaultWALProvider is used.
> 2. RegionGroupingProvider provider is used but replication is stuck on one
> WAL group for some reason (for example HBASE-18137)
> 3. All other replication workers have died due to unhandled exception, and
> the only one finishes. This will cause the recovered queue to get deleted
> without a regionserver shutdown. This can happen on deployments without fix
> for HBASE-17381.
> The problematic piece of code is:
> {Code}
> while (isWorkerActive()){
> // The worker thread run loop...
> }
> if (replicationQueueInfo.isQueueRecovered()) {
> // use synchronize to make sure one last thread will clean the queue
> synchronized (workerThreads) {
> Threads.sleep(100);// wait a short while for other worker thread to
> fully exit
> boolean allOtherTaskDone = true;
> for (ReplicationSourceWorkerThread worker : workerThreads.values())
> {
> if (!worker.equals(this) && worker.isAlive()) {
> allOtherTaskDone = false;
> break;
> }
> }
> if (allOtherTaskDone) {
> manager.closeRecoveredQueue(this.source);
> LOG.info("Finished recovering queue " + peerClusterZnode
> + " with the following stats: " + getStats());
> }
> }
> {Code}
> The conceptual issue is that isWorkerActive() tells whether a worker is
> currently running or not and it's being used as a proxy for whether a worker
> has finished it's work. But, in fact, "Should a worker should exit?" and "Has
> completed it's work?" are two different questions.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
