HBASE-20548 Master fails to startup on large clusters, refreshing block distribution
Signed-off-by: Andrew Purtell <apurtell@apache.org>
This commit is contained in:
parent
554d513f50
commit
1fbce10ff4
|
@ -466,4 +466,9 @@ public class RSGroupBasedLoadBalancer implements RSGroupableBalancer {
|
|||
public void setRsGroupInfoManager(RSGroupInfoManager rsGroupInfoManager) {
|
||||
this.rsGroupInfoManager = rsGroupInfoManager;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void postMasterStartupInitialize() {
|
||||
this.internalBalancer.postMasterStartupInitialize();
|
||||
}
|
||||
}
|
||||
|
|
|
@ -1003,6 +1003,17 @@ public class HMaster extends HRegionServer implements MasterServices {
|
|||
}
|
||||
|
||||
zombieDetector.interrupt();
|
||||
|
||||
/*
|
||||
* After master has started up, lets do balancer post startup initialization. Since this runs
|
||||
* in activeMasterManager thread, it should be fine.
|
||||
*/
|
||||
long start = System.currentTimeMillis();
|
||||
this.balancer.postMasterStartupInitialize();
|
||||
if (LOG.isDebugEnabled()) {
|
||||
LOG.debug("Balancer post startup initialization complete, took " + (
|
||||
(System.currentTimeMillis() - start) / 1000) + " seconds");
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
|
@ -159,6 +159,11 @@ public interface LoadBalancer extends Configurable, Stoppable, ConfigurationObse
|
|||
@Override
|
||||
void onConfigurationChange(Configuration conf);
|
||||
|
||||
/**
|
||||
* If balancer needs to do initialization after Master has started up, lets do that here.
|
||||
*/
|
||||
void postMasterStartupInitialize();
|
||||
|
||||
/**
|
||||
* @return true if Master carries regions
|
||||
*/
|
||||
|
|
|
@ -1151,6 +1151,19 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public void postMasterStartupInitialize() {
|
||||
if (services != null && regionFinder != null) {
|
||||
try {
|
||||
Set<RegionInfo> regions =
|
||||
services.getAssignmentManager().getRegionStates().getRegionAssignments().keySet();
|
||||
regionFinder.refreshAndWait(regions);
|
||||
} catch (Exception e) {
|
||||
LOG.warn("Refreshing region HDFS Block dist failed with exception, ignoring", e);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
public void setRackManager(RackManager rackManager) {
|
||||
this.rackManager = rackManager;
|
||||
}
|
||||
|
@ -1249,7 +1262,7 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
return assignments;
|
||||
}
|
||||
|
||||
Cluster cluster = createCluster(servers, regions, false);
|
||||
Cluster cluster = createCluster(servers, regions);
|
||||
List<RegionInfo> unassignedRegions = new ArrayList<>();
|
||||
|
||||
roundRobinAssignment(cluster, regions, unassignedRegions,
|
||||
|
@ -1288,11 +1301,7 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
return assignments;
|
||||
}
|
||||
|
||||
protected Cluster createCluster(List<ServerName> servers,
|
||||
Collection<RegionInfo> regions, boolean forceRefresh) {
|
||||
if (forceRefresh && useRegionFinder) {
|
||||
regionFinder.refreshAndWait(regions);
|
||||
}
|
||||
protected Cluster createCluster(List<ServerName> servers, Collection<RegionInfo> regions) {
|
||||
// Get the snapshot of the current assignments for the regions in question, and then create
|
||||
// a cluster out of it. Note that we might have replicas already assigned to some servers
|
||||
// earlier. So we want to get the snapshot to see those assignments, but this will only contain
|
||||
|
@ -1346,7 +1355,7 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
final List<ServerName> finalServers = idleServers.isEmpty() ?
|
||||
servers : idleServers;
|
||||
List<RegionInfo> regions = Lists.newArrayList(regionInfo);
|
||||
Cluster cluster = createCluster(finalServers, regions, false);
|
||||
Cluster cluster = createCluster(finalServers, regions);
|
||||
return randomAssignment(cluster, regionInfo, finalServers);
|
||||
}
|
||||
|
||||
|
@ -1419,8 +1428,6 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
int numRandomAssignments = 0;
|
||||
int numRetainedAssigments = 0;
|
||||
|
||||
Cluster cluster = createCluster(servers, regions.keySet(), true);
|
||||
|
||||
for (Map.Entry<RegionInfo, ServerName> entry : regions.entrySet()) {
|
||||
RegionInfo region = entry.getKey();
|
||||
ServerName oldServerName = entry.getValue();
|
||||
|
@ -1463,6 +1470,7 @@ public abstract class BaseLoadBalancer implements LoadBalancer {
|
|||
|
||||
// If servers from prior assignment aren't present, then lets do randomAssignment on regions.
|
||||
if (randomAssignRegions.size() > 0) {
|
||||
Cluster cluster = createCluster(servers, regions.keySet());
|
||||
for (Map.Entry<ServerName, List<RegionInfo>> entry : assignments.entrySet()) {
|
||||
ServerName sn = entry.getKey();
|
||||
for (RegionInfo region : entry.getValue()) {
|
||||
|
|
Loading…
Reference in New Issue