Skip to content

HBASE-26163 Better logging in RSGroupInfoManagerImpl #3610

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Apache9 merged 11 commits into from
Sep 7, 2021
Merged

HBASE-26163 Better logging in RSGroupInfoManagerImpl #3610

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
61450e6
HBASE-26163 Better logging in RSGroupInfoManagerImpl
caroliney14 Aug 20, 2021
c8145e0
move long messages to debug level
caroliney14 Aug 23, 2021
80e7bf7
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Aug 23, 2021
e7aeb13
pr rev
caroliney14 Aug 24, 2021
4ecd3ed
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Aug 24, 2021
608893b
remove tostring
caroliney14 Aug 26, 2021
fb2282e
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Aug 26, 2021
1c7711a
move lists to debug
caroliney14 Aug 27, 2021
6847cce
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Aug 31, 2021
87d6d71
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Aug 31, 2021
a8720f8
Merge branch 'master' of https://github.com/caroliney14/hbase into HB…
caroliney14 Sep 2, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions hbase-server/src/main/java/org/apache/hadoop/hbase/rsgroup/RSGroupBasedLoadBalancer.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,8 @@
import java.util.Map;
import java.util.Set;
import java.util.TreeMap;
import java.util.stream.Collectors;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.ClusterMetrics;
import org.apache.hadoop.hbase.HBaseIOException;
Expand Down Expand Up @@ -129,6 +131,7 @@ public synchronized List<RegionPlan> balanceCluster(
try {
// For each rsgroup
for (RSGroupInfo rsgroup : rsGroupInfoManager.listRSGroups()) {
LOG.debug("Balancing RSGroup={}", rsgroup.getName());
Map<TableName, Map<ServerName, List<RegionInfo>>> loadOfTablesInGroup = new HashMap<>();
for (Map.Entry<TableName, Map<ServerName, List<RegionInfo>>> entry : correctedLoadOfAllTable
.entrySet()) {
Expand Down Expand Up @@ -234,6 +237,11 @@ private List<Pair<List<RegionInfo>, List<ServerName>>> generateGroupAssignments(
if (!fallbackRegions.isEmpty()) {
List<ServerName> candidates = null;
if (isFallbackEnabled()) {
if (LOG.isDebugEnabled()) {
LOG.debug("Falling back {} regions to servers outside their RSGroup. Regions: {}",
fallbackRegions.size(), fallbackRegions.stream()
.map(RegionInfo::getRegionNameAsString).collect(Collectors.toSet()));
}
candidates = getFallBackCandidates(servers);
}
candidates = (candidates == null || candidates.isEmpty()) ?
Expand Down
31 changes: 25 additions & 6 deletions hbase-server/src/main/java/org/apache/hadoop/hbase/rsgroup/RSGroupInfoManagerImpl.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -257,6 +257,9 @@ private synchronized void updateDefaultServers() {
// do not need to persist, as we do not persist default group.
resetRSGroupMap(newGroupMap);
LOG.info("Updated default servers, {} servers", newDefaultGroupInfo.getServers().size());
if (LOG.isDebugEnabled()) {
LOG.debug("New default servers list: {}", newDefaultGroupInfo.getServers());
}
}

private synchronized void init() throws IOException {
Expand Down Expand Up @@ -298,6 +301,7 @@ public synchronized void addRSGroup(RSGroupInfo rsGroupInfo) throws IOException
Map<String, RSGroupInfo> newGroupMap = Maps.newHashMap(rsGroupMap);
newGroupMap.put(rsGroupInfo.getName(), rsGroupInfo);
flushConfig(newGroupMap);
LOG.info("Add group {} done.", rsGroupInfo.getName());
}

private RSGroupInfo getRSGroupInfo(final String groupName) throws ConstraintException {
Expand Down Expand Up @@ -331,7 +335,7 @@ public synchronized Set<Address> moveServers(Set<Address> servers, String srcGro
if (onlineServers != null) {
if (!onlineServers.contains(el)) {
if (LOG.isDebugEnabled()) {
LOG.debug("Dropping " + el + " during move-to-default rsgroup because not online");
LOG.debug("Dropping " + el + " during move-to-default RSGroup because not online");
}
continue;
}
Expand Down Expand Up @@ -373,8 +377,8 @@ public synchronized void removeRSGroup(String groupName) throws IOException {
for (TableDescriptor td : masterServices.getTableDescriptors().getAll().values()) {
if (td.getRegionServerGroup().map(groupName::equals).orElse(false)) {
throw new ConstraintException("RSGroup " + groupName + " is already referenced by " +
td.getTableName() + "; you must remove all the tables from the rsgroup before " +
"the rsgroup can be removed.");
td.getTableName() + "; you must remove all the tables from the RSGroup before " +
"the RSGroup can be removed.");
}
}
for (NamespaceDescriptor ns : masterServices.getClusterSchema().getNamespaces()) {
Expand All @@ -392,6 +396,7 @@ public synchronized void removeRSGroup(String groupName) throws IOException {
Map<String, RSGroupInfo> newGroupMap = Maps.newHashMap(rsGroupMap);
newGroupMap.remove(groupName);
flushConfig(newGroupMap);
LOG.info("Remove group {} done", groupName);
}

@Override
Expand Down Expand Up @@ -639,6 +644,8 @@ private synchronized void flushConfig(Map<String, RSGroupInfo> newGroupMap) thro
return;
}

LOG.debug("Offline mode, cannot persist to {}", RSGROUP_TABLE_NAME);

Map<String, RSGroupInfo> oldGroupMap = Maps.newHashMap(holder.groupName2Group);
RSGroupInfo oldDefaultGroup = oldGroupMap.remove(RSGroupInfo.DEFAULT_GROUP);
RSGroupInfo newDefaultGroup = newGroupMap.remove(RSGroupInfo.DEFAULT_GROUP);
Expand All @@ -655,22 +662,27 @@ private synchronized void flushConfig(Map<String, RSGroupInfo> newGroupMap) thro
// according to the inputted newGroupMap (an updated copy of rsGroupMap)
this.holder = new RSGroupInfoHolder(newGroupMap);

LOG.debug("New RSGroup map: {}", newGroupMap);

// Do not need to update tableMap
// because only the update on servers in default group is allowed above,
// or IOException will be thrown
return;
}

/* For online mode, persist to hbase:rsgroup and Zookeeper */
LOG.debug("Online mode, persisting to {} and ZK", RSGROUP_TABLE_NAME);
flushConfigTable(newGroupMap);

// Make changes visible after having been persisted to the source of truth
resetRSGroupMap(newGroupMap);
saveRSGroupMapToZK(newGroupMap);
updateCacheOfRSGroups(newGroupMap.keySet());
LOG.info("Flush config done, new RSGroup map: {}", newGroupMap);
}

private void saveRSGroupMapToZK(Map<String, RSGroupInfo> newGroupMap) throws IOException {
LOG.debug("Saving RSGroup info to ZK");
try {
String groupBasePath =
ZNodePaths.joinZNode(watcher.getZNodePaths().baseZNode, RS_GROUP_ZNODE);
Expand Down Expand Up @@ -800,6 +812,7 @@ private void createRSGroupTable() throws IOException {
if (optProcId.isPresent()) {
procId = optProcId.getAsLong();
} else {
LOG.debug("Creating group table {}", RSGROUP_TABLE_NAME);
procId = masterServices.createSystemTable(RSGROUP_TABLE_DESC);
}
// wait for region to be online
Expand Down Expand Up @@ -849,9 +862,11 @@ private void multiMutate(List<Mutation> mutations) throws IOException {
}
MutateRowsRequest request = builder.build();
AsyncTable<?> table = conn.getTable(RSGROUP_TABLE_NAME);
LOG.debug("Multimutating {} with {} mutations", RSGROUP_TABLE_NAME, mutations.size());
FutureUtils.get(table.<MultiRowMutationService, MutateRowsResponse> coprocessorService(
MultiRowMutationService::newStub,
(stub, controller, done) -> stub.mutateRows(controller, request, done), ROW_KEY));
LOG.info("Multimutating {} with {} mutations done", RSGROUP_TABLE_NAME, mutations.size());
}

private void checkGroupName(String groupName) throws ConstraintException {
Expand All @@ -871,7 +886,7 @@ public RSGroupInfo getRSGroupForTable(TableName tableName) throws IOException {
* @param servers servers to remove
*/
private void checkForDeadOrOnlineServers(Set<Address> servers) throws IOException {
// This uglyness is because we only have Address, not ServerName.
// This ugliness is because we only have Address, not ServerName.
Set<Address> onlineServers = new HashSet<>();
List<ServerName> drainingServers = masterServices.getServerManager().getDrainingServersList();
for (ServerName server : masterServices.getServerManager().getOnlineServers().keySet()) {
Expand Down Expand Up @@ -1001,7 +1016,7 @@ private <T> void moveRegionsBetweenGroups(Set<T> regionsOwners, Set<Address> new
// Get regions that are associated with this server and filter regions by group tables.
for (RegionInfo region : getRegionsInfo.apply((T) owner.getAddress())) {
if (!validation.apply(region)) {
LOG.info("Moving region {}, which do not belong to RSGroup {}",
LOG.info("Moving region {}, which does not belong to RSGroup {}",
region.getShortNameToLog(), targetGroupName);
// Move region back to source RSGroup servers
ServerName dest =
Expand Down Expand Up @@ -1186,6 +1201,7 @@ public boolean balanceRSGroup(String groupName) throws IOException {
}

private void moveTablesAndWait(Set<TableName> tables, String targetGroup) throws IOException {
LOG.debug("Moving {} tables to target group {}", tables.size(), targetGroup);
List<Long> procIds = new ArrayList<Long>();
for (TableName tableName : tables) {
TableDescriptor oldTd = masterServices.getTableDescriptors().get(tableName);
Expand All @@ -1205,6 +1221,7 @@ private void moveTablesAndWait(Set<TableName> tables, String targetGroup) throws
ProcedureSyncWait.waitForProcedureToCompleteIOE(masterServices.getMasterProcedureExecutor(),
proc, Long.MAX_VALUE);
}
LOG.info("Move tables done. Moved {} tables to {}: {}", tables.size(), targetGroup, tables);
Copy link
Contributor

@Apache9 Apache9 Aug 26, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We will log too much if the level is info here? Can we just info the size and the target group, and add another debug log to log all the tables?

Copy link
Contributor Author

@caroliney14 caroliney14 Aug 27, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

}

@Override
Expand Down Expand Up @@ -1270,7 +1287,8 @@ public void moveServers(Set<Address> servers, String targetGroupName) throws IOE
Set<Address> movedServers = moveServers(servers, srcGrp.getName(),
targetGroupName);
moveServerRegionsFromGroup(movedServers, srcGrp.getServers(), targetGroupName, srcGrp.getName());
LOG.info("Move servers done: {} => {}", srcGrp.getName(), targetGroupName);
LOG.info("Move servers done. Moved {} servers from {} => {}: {}", movedServers.size(),
Copy link
Contributor

@Apache9 Apache9 Aug 26, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ditto, I'm afraid it will too much to also log all the servers name, but log the size in info and then all the servers with debug.

Copy link
Contributor Author

@caroliney14 caroliney14 Aug 27, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

srcGrp.getName(), targetGroupName, movedServers);
}
}

Expand Down Expand Up @@ -1304,6 +1322,7 @@ public synchronized void renameRSGroup(String oldName, String newName) throws IO
.map(TableDescriptor::getTableName)
.collect(Collectors.toSet());
setRSGroup(updateTables, newName);
LOG.info("Rename RSGroup done: {} => {}", oldName, newName);
}

@Override
Expand Down