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

SOLR-10279: The autoAddReplica feature can result in SolrCores being assigned new shards when using legacyCloud=false and will also fail on a state check when taking over a core registration with a new core.

This commit is contained in:
markrmiller 2017-03-14 06:01:06 -04:00
parent 120274b451
commit bac3424936
4 changed files with 38 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
solr/CHANGES.txt
View File

@ -243,6 +243,10 @@ Bug Fixes
* SOLR-10269: MetricsHandler JSON output incorrect. (ab)
* SOLR-10279: The autoAddReplica feature can result in SolrCores being assigned new shards when using
legacyCloud=false and will also fail on a state check when taking over a core registration with a new
core. (Mark Miller, Hrishikesh Gadre, Patrick Dvorack)
Optimizations
----------------------

6
solr/core/src/java/org/apache/solr/cloud/OverseerAutoReplicaFailoverThread.java
View File

@ -243,13 +243,14 @@ public class OverseerAutoReplicaFailoverThread implements Runnable, Closeable {
final String dataDir = badReplica.replica.getStr("dataDir");
final String ulogDir = badReplica.replica.getStr("ulogDir");
final String coreNodeName = badReplica.replica.getName();
final String shardId = badReplica.slice.getName();
if (dataDir != null) {
// need an async request - full shard goes down leader election
final String coreName = badReplica.replica.getStr(ZkStateReader.CORE_NAME_PROP);
log.debug("submit call to {}", createUrl);
MDC.put("OverseerAutoReplicaFailoverThread.createUrl", createUrl);
try {
updateExecutor.submit(() -> createSolrCore(collection, createUrl, dataDir, ulogDir, coreNodeName, coreName));
updateExecutor.submit(() -> createSolrCore(collection, createUrl, dataDir, ulogDir, coreNodeName, coreName, shardId));
} finally {
MDC.remove("OverseerAutoReplicaFailoverThread.createUrl");
}
@ -440,7 +441,7 @@ public class OverseerAutoReplicaFailoverThread implements Runnable, Closeable {
private boolean createSolrCore(final String collection,
final String createUrl, final String dataDir, final String ulogDir,
final String coreNodeName, final String coreName) {
final String coreNodeName, final String coreName, final String shardId) {
try (HttpSolrClient client = new HttpSolrClient.Builder(createUrl).build()) {
log.debug("create url={}", createUrl);
@ -451,6 +452,7 @@ public class OverseerAutoReplicaFailoverThread implements Runnable, Closeable {
createCmd.setCoreNodeName(coreNodeName);
// TODO: how do we ensure unique coreName
// for now, the collections API will use unique names
createCmd.setShardId(shardId);
createCmd.setCoreName(coreName);
createCmd.setDataDir(dataDir);
createCmd.setUlogDir(ulogDir.substring(0, ulogDir.length() - "/tlog".length()));

8
solr/core/src/java/org/apache/solr/cloud/ZkController.java
View File

@ -1449,13 +1449,7 @@ public class ZkController {
errorMessage.set("coreNodeName " + coreNodeName + " does not exist in shard " + cloudDesc.getShardId());
return false;
}
String baseUrl = replica.getStr(BASE_URL_PROP);
String coreName = replica.getStr(CORE_NAME_PROP);
if (baseUrl.equals(this.baseURL) && coreName.equals(cd.getName())) {
return true;
}
errorMessage.set("coreNodeName " + coreNodeName + " exists, but does not match expected node or core name");
return false;
return true;
});
} catch (TimeoutException e) {
String error = errorMessage.get();

29
solr/core/src/test/org/apache/solr/cloud/SharedFSAutoReplicaFailoverTest.java
View File

@ -38,6 +38,7 @@ import org.apache.solr.SolrTestCaseJ4.SuppressSSL;
import org.apache.solr.client.solrj.SolrQuery;
import org.apache.solr.client.solrj.SolrRequest;
import org.apache.solr.client.solrj.SolrServerException;
import org.apache.solr.client.solrj.request.CollectionAdminRequest;
import org.apache.solr.client.solrj.request.CollectionAdminRequest.Create;
import org.apache.solr.client.solrj.request.QueryRequest;
import org.apache.solr.client.solrj.response.CollectionAdminResponse;
@ -103,6 +104,11 @@ public class SharedFSAutoReplicaFailoverTest extends AbstractFullDistribZkTestBa
public void setUp() throws Exception {
super.setUp();
collectionUlogDirMap.clear();
if (random().nextBoolean()) {
CollectionAdminRequest.setClusterProperty("legacyCloud", "false").process(cloudClient);
} else {
CollectionAdminRequest.setClusterProperty("legacyCloud", "true").process(cloudClient);
}
}
@Override
@ -313,6 +319,29 @@ public class SharedFSAutoReplicaFailoverTest extends AbstractFullDistribZkTestBa
assertSliceAndReplicaCount(collection1);
assertUlogDir(collections);
// restart all to test core saved state
ChaosMonkey.stop(jettys);
ChaosMonkey.stop(controlJetty);
assertTrue("Timeout waiting for all not live", ClusterStateUtil.waitForAllReplicasNotLive(cloudClient.getZkStateReader(), 45000));
ChaosMonkey.start(jettys);
ChaosMonkey.start(controlJetty);
assertTrue("Timeout waiting for all live and active", ClusterStateUtil.waitForAllActiveAndLiveReplicas(cloudClient.getZkStateReader(), collection1, 120000));
assertSliceAndReplicaCount(collection1);
assertUlogDir(collections);
assertSliceAndReplicaCount(collection1);
assertSingleReplicationAndShardSize(collection3, 5);
// all docs should be queried
assertSingleReplicationAndShardSize(collection4, 5);
queryAndAssertResultSize(collection4, numDocs, 10000);
}
private void queryAndAssertResultSize(String collection, int expectedResultSize, int timeoutMS)