PHOENIX-4732 Improve config defaults for async secondary index MR jobs
Project: http://git-wip-us.apache.org/repos/asf/phoenix/repo Commit: http://git-wip-us.apache.org/repos/asf/phoenix/commit/976ffd56 Tree: http://git-wip-us.apache.org/repos/asf/phoenix/tree/976ffd56 Diff: http://git-wip-us.apache.org/repos/asf/phoenix/diff/976ffd56 Branch: refs/heads/4.x-HBase-1.2 Commit: 976ffd56ce589595e80e9244df8cc442d17d8a43 Parents: f186c30 Author: James Taylor <jtay...@salesforce.com> Authored: Tue May 8 15:26:44 2018 -0700 Committer: James Taylor <jtay...@salesforce.com> Committed: Wed May 9 13:25:29 2018 -0700 ---------------------------------------------------------------------- .../index/automation/PhoenixMRJobSubmitter.java | 48 ++++++++++++++++++-- .../apache/phoenix/util/PhoenixMRJobUtil.java | 18 +++----- 2 files changed, 52 insertions(+), 14 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/phoenix/blob/976ffd56/phoenix-core/src/main/java/org/apache/phoenix/mapreduce/index/automation/PhoenixMRJobSubmitter.java ---------------------------------------------------------------------- diff --git a/phoenix-core/src/main/java/org/apache/phoenix/mapreduce/index/automation/PhoenixMRJobSubmitter.java b/phoenix-core/src/main/java/org/apache/phoenix/mapreduce/index/automation/PhoenixMRJobSubmitter.java index 9c447e8..d86802a 100644 --- a/phoenix-core/src/main/java/org/apache/phoenix/mapreduce/index/automation/PhoenixMRJobSubmitter.java +++ b/phoenix-core/src/main/java/org/apache/phoenix/mapreduce/index/automation/PhoenixMRJobSubmitter.java @@ -45,6 +45,7 @@ import org.apache.hadoop.conf.Configuration; import org.apache.hadoop.hbase.HBaseConfiguration; import org.apache.hadoop.hbase.HConstants; import org.apache.hadoop.hbase.zookeeper.ZooKeeperWatcher; +import org.apache.hadoop.mapreduce.MRJobConfig; import org.apache.hadoop.security.UserGroupInformation; import org.apache.phoenix.jdbc.PhoenixDatabaseMetaData; import org.apache.phoenix.mapreduce.index.IndexTool; @@ -53,8 +54,8 @@ import org.apache.phoenix.schema.PTable.IndexType; import org.apache.phoenix.schema.PTableType; import org.apache.phoenix.schema.types.PDate; import org.apache.phoenix.util.PhoenixMRJobUtil; -import org.apache.phoenix.util.UpgradeUtil; import org.apache.phoenix.util.PhoenixMRJobUtil.MR_SCHEDULER_TYPE; +import org.apache.phoenix.util.UpgradeUtil; import org.apache.phoenix.util.ZKBasedMasterElectionUtil; import org.codehaus.jettison.json.JSONArray; import org.codehaus.jettison.json.JSONObject; @@ -69,8 +70,28 @@ public class PhoenixMRJobSubmitter { // Lock to elect a master node that submits the Phoenix Secondary Index MR Jobs private static final String PHOENIX_LOCKS_PARENT = "/phoenix/automated-mr-index-build-leader-election"; + public static final String PHOENIX_MR_CLIENT_SCANNER_TIMEOUT_PERIOD = + "phoenix.mr.client.scanner.timeout.period"; + public static final String PHOENIX_MR_RPC_TIMEOUT = + "phoenix.mr.rpc.timeout"; + public static final String PHOENIX_MR_TASK_TIMEOUT = + "phoenix.mr.task.timeout"; + public static final String PHOENIX_MR_CLIENT_RETRIES_NUMBER = + "phoenix.mr.client.retries.number"; + public static final String PHOENIX_MR_CLIENT_PAUSE = + "phoenix.mr.client.retries.number"; + public static final String PHOENIX_MR_ZK_RECOVERY_RETRY = + "phoenix.mr.zk.recovery.retry"; private static final String AUTO_INDEX_BUILD_LOCK_NAME = "ActiveStandbyElectorLock"; - + private static final int DEFAULT_TIMEOUT_IN_MILLIS = 600000; + public static final int DEFAULT_MR_CLIENT_SCANNER_TIMEOUT_PERIOD = DEFAULT_TIMEOUT_IN_MILLIS; + public static final int DEFAULT_MR_RPC_TIMEOUT = DEFAULT_TIMEOUT_IN_MILLIS; + public static final int DEFAULT_MR_TASK_TIMEOUT = DEFAULT_TIMEOUT_IN_MILLIS; + // Reduced HBase/Zookeeper Client Retries + public static final int DEFAULT_MR_CLIENT_RETRIES_NUMBER = 10; + public static final int DEFAULT_MR_CLIENT_PAUSE = 1000; + public static final int DEFAULT_MR_ZK_RECOVERY_RETRY = 1; + public static final String CANDIDATE_INDEX_INFO_QUERY = "SELECT " + PhoenixDatabaseMetaData.INDEX_TYPE + "," + PhoenixDatabaseMetaData.DATA_TABLE_NAME + ", " @@ -103,7 +124,28 @@ public class PhoenixMRJobSubmitter { } this.conf = conf; - PhoenixMRJobUtil.updateTimeoutsToFailFast(conf); + // Have Phoenix specific properties for defaults to enable potential override + conf.setLong(HConstants.HBASE_CLIENT_SCANNER_TIMEOUT_PERIOD, + conf.getLong(PHOENIX_MR_CLIENT_SCANNER_TIMEOUT_PERIOD, + DEFAULT_MR_CLIENT_SCANNER_TIMEOUT_PERIOD)); + conf.setLong(HConstants.HBASE_RPC_TIMEOUT_KEY, + conf.getLong(PHOENIX_MR_RPC_TIMEOUT, + DEFAULT_MR_RPC_TIMEOUT)); + conf.setLong(MRJobConfig.TASK_TIMEOUT, + conf.getLong(PHOENIX_MR_TASK_TIMEOUT, + DEFAULT_MR_TASK_TIMEOUT)); + + // Reduced HBase Client Retries + conf.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, + conf.getInt(PHOENIX_MR_CLIENT_RETRIES_NUMBER, + DEFAULT_MR_CLIENT_RETRIES_NUMBER)); + conf.setInt(HConstants.HBASE_CLIENT_PAUSE, + conf.getInt(PHOENIX_MR_CLIENT_PAUSE, + DEFAULT_MR_CLIENT_PAUSE)); + conf.setInt("zookeeper.recovery.retry", + conf.getInt(PHOENIX_MR_ZK_RECOVERY_RETRY, + DEFAULT_MR_ZK_RECOVERY_RETRY)); + String schedulerType = conf.get(PhoenixMRJobUtil.PHOENIX_MR_SCHEDULER_TYPE_NAME, MR_SCHEDULER_TYPE.NONE.toString()); http://git-wip-us.apache.org/repos/asf/phoenix/blob/976ffd56/phoenix-core/src/main/java/org/apache/phoenix/util/PhoenixMRJobUtil.java ---------------------------------------------------------------------- diff --git a/phoenix-core/src/main/java/org/apache/phoenix/util/PhoenixMRJobUtil.java b/phoenix-core/src/main/java/org/apache/phoenix/util/PhoenixMRJobUtil.java index f12d49d..13eae98 100644 --- a/phoenix-core/src/main/java/org/apache/phoenix/util/PhoenixMRJobUtil.java +++ b/phoenix-core/src/main/java/org/apache/phoenix/util/PhoenixMRJobUtil.java @@ -52,19 +52,17 @@ public class PhoenixMRJobUtil { private static final String ACTIVE_STANDBY_ELECTOR_LOCK = "ActiveStandbyElectorLock"; private static final String RM_APPS_GET_ENDPOINT = "/ws/v1/cluster/apps"; - // Reduced HBase Client Retries - private static final int CLIENT_RETRIES_NUMBER = 2; - private static final long CLIENT_PAUSE_TIME = 1000; - private static final int ZOOKEEPER_RECOVERY_RETRY_COUNT = 1; - public static final String PHOENIX_INDEX_MR_QUEUE_NAME_PROPERTY = "phoenix.index.mr.scheduler.capacity.queuename"; public static final String PHOENIX_INDEX_MR_MAP_MEMORY_PROPERTY = "phoenix.index.mr.scheduler.capacity.mapMemoryMB"; + public static final String PHOENIX_MR_CONCURRENT_MAP_LIMIT_PROPERTY = + "phoenix.mr.concurrent.map.limit"; // Default MR Capacity Scheduler Configurations for Phoenix MR Index Build // Jobs public static final String DEFAULT_QUEUE_NAME = "default"; + public static final int DEFAULT_MR_CONCURRENT_MAP_LIMIT = 20; public static final int DEFAULT_MAP_MEMROY_MB = 5120; public static final String XMX_OPT = "-Xmx"; @@ -208,12 +206,6 @@ public class PhoenixMRJobUtil { return rmPort; } - public static void updateTimeoutsToFailFast(Configuration conf) { - conf.set("hbase.client.retries.number", String.valueOf(CLIENT_RETRIES_NUMBER)); - conf.set("zookeeper.recovery.retry", String.valueOf(ZOOKEEPER_RECOVERY_RETRY_COUNT)); - conf.set("hbase.client.pause", String.valueOf(CLIENT_PAUSE_TIME)); - } - /** * This method set the configuration values for Capacity scheduler. * @param conf - Configuration to which Capacity Queue information to be added @@ -221,6 +213,10 @@ public class PhoenixMRJobUtil { public static void updateCapacityQueueInfo(Configuration conf) { conf.set(MRJobConfig.QUEUE_NAME, conf.get(PHOENIX_INDEX_MR_QUEUE_NAME_PROPERTY, DEFAULT_QUEUE_NAME)); + + conf.setInt(MRJobConfig.JOB_RUNNING_MAP_LIMIT, + conf.getInt(PHOENIX_MR_CONCURRENT_MAP_LIMIT_PROPERTY, DEFAULT_MR_CONCURRENT_MAP_LIMIT)); + int mapMemoryMB = conf.getInt(PHOENIX_INDEX_MR_MAP_MEMORY_PROPERTY, DEFAULT_MAP_MEMROY_MB); conf.setInt(MRJobConfig.MAP_MEMORY_MB, mapMemoryMB);