Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

SOLR-12051: Adding error log in case of data loss

This commit is contained in:
Cao Manh Dat 2018-03-10 07:20:05 +07:00
parent 78097d2098
commit 05d4a9320c
1 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
solr/core/src/java/org/apache/solr/cloud/ElectionContext.java
View File

@ -347,6 +347,7 @@ final class ShardLeaderElectionContext extends ShardLeaderElectionContextBase {
Replica.Type replicaType; Replica.Type replicaType;
String coreNodeName; String coreNodeName;
boolean setTermToMax = false;
try (SolrCore core = cc.getCore(coreName)) { try (SolrCore core = cc.getCore(coreName)) {
if (core == null) { if (core == null) {
@ -365,9 +366,11 @@ final class ShardLeaderElectionContext extends ShardLeaderElectionContextBase {
ZkShardTerms zkShardTerms = zkController.getShardTerms(collection, shardId); ZkShardTerms zkShardTerms = zkController.getShardTerms(collection, shardId);
if (zkShardTerms.registered(coreNodeName) && !zkShardTerms.canBecomeLeader(coreNodeName)) { if (zkShardTerms.registered(coreNodeName) && !zkShardTerms.canBecomeLeader(coreNodeName)) {
if (!waitForEligibleBecomeLeaderAfterTimeout(zkShardTerms, coreNodeName, leaderVoteWait)) { if (!waitForEligibleBecomeLeaderAfterTimeout(zkShardTerms, coreNodeName, leaderVoteWait)) {
log.info("Can't become leader, term of replica {} less than leader", coreNodeName);
rejoinLeaderElection(core); rejoinLeaderElection(core);
return; return;
} else {
// only log an error if this replica win the election
setTermToMax = true;
} }
} }
@ -470,7 +473,9 @@ final class ShardLeaderElectionContext extends ShardLeaderElectionContextBase {
} }
} }
// in case of leaderVoteWait timeout, a replica with lower term can win the election // in case of leaderVoteWait timeout, a replica with lower term can win the election
if (zkController.getShardTerms(collection, shardId).registered(coreNodeName)) { if (setTermToMax) {
log.error("WARNING: Potential data loss -- Replica {} became leader after timeout (leaderVoteWait) " +
"without being up-to-date with the previous leader", coreNodeName);
zkController.getShardTerms(collection, shardId).setTermEqualsToLeader(coreNodeName); zkController.getShardTerms(collection, shardId).setTermEqualsToLeader(coreNodeName);
} }
super.runLeaderProcess(weAreReplacement, 0); super.runLeaderProcess(weAreReplacement, 0);