[HIVE-21206] Bootstrap replication is slow as it opens lot of metastore connections.

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 4.0.0
• Fix Version/s: 4.0.0-alpha-1
• Component/s: repl
• Labels: DR pull-request-available replication
• Target Version/s: 4.0.0
• Assignee: Sankar Hariappan
• Reporter: Sankar Hariappan
• Created: 04/Feb/19 06:01
• Updated: 17/Nov/22 08:54
• Resolved: 12/Feb/19 05:05

Description

Hive bootstrap replication of 1TB data onprem to onprem in Hive3 is running slower compared to Hive2.

Time taken for bootstrap replication of table with 1000 partitions are as below:

Hive2- Hive2	Hive3 - Hive3
Bootstrap: 7m	BootStrap: 17m

Every MoveTask is closing and opening new metastore connection which is causing slow down.

2019-02-08T12:28:30,174 INFO  [HiveServer2-Background-Pool: Thread-1134]: ql.Driver (:()) - Starting task [Stage-5:MOVE] in serial mode
2019-02-08T12:28:30,177 INFO  [HiveServer2-Background-Pool: Thread-1134]: exec.Task (:()) - Loading data to table nondefault.nondefault_table1 from hdfs://mycluster1/warehouse/tablespace/managed/hive/nondefault.db/nondefault_table1/.hive-staging_hive_2019-02-08_12-28-23_584_1482331698286040936-3/-ext-10001
2019-02-08T12:28:30,189 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Trying to connect to metastore with URI thrift://ctr-e139-1542663976389-62755-01-000014.hwx.site:9083
2019-02-08T12:28:30,189 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - HMSC::open(): Could not find delegation token. Creating KERBEROS-based thrift connection.
2019-02-08T12:28:30,206 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Opened a connection to metastore, current connections: 4
2019-02-08T12:28:30,206 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Connected to metastore.
2019-02-08T12:28:30,206 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.RetryingMetaStoreClient (:()) - RetryingMetaStoreClient proxy=class org.apache.hadoop.hive.ql.metadata.SessionHiveMetaStoreClient ugi=hive/[e-mail] (auth:KERBEROS) retries=24 delay=5 lifetime=0
2019-02-08T12:28:30,325 INFO  [org.apache.ranger.audit.queue.AuditBatchQueue1]: provider.BaseAuditHandler (:()) - Audit Status Log: name=hiveServer2.async.multi_dest.batch, finalDestination=hiveServer2.async.multi_dest.batch.solr, interval=01:00.002 minutes, events=2, succcessCount=1, totalEvents=56, totalSuccessCount=25
2019-02-08T12:28:30,520 INFO  [HiveServer2-Background-Pool: Thread-1134]: common.FileUtils (FileUtils.java:mkdir(580)) - Creating directory if it doesn't exist: hdfs://mycluster1/warehouse/tablespace/managed/hive/nondefault.db/nondefault_table1/base_0000001
2019-02-08T12:28:31,245 INFO  [HiveServer2-Background-Pool: Thread-1134]: ql.Driver (:()) - Starting task [Stage-11:MOVE] in serial mode
2019-02-08T12:28:31,245 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Closed a connection to metastore, current connections: 3
2019-02-08T12:28:31,246 INFO  [HiveServer2-Background-Pool: Thread-1134]: exec.Task (:()) - Loading data to table nondefault.nondefault_table2 from hdfs://mycluster1/warehouse/tablespace/managed/hive/nondefault.db/nondefault_table2/.hive-staging_hive_2019-02-08_12-28-23_810_7457138692783022870-3/-ext-10002
2019-02-08T12:28:31,327 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Trying to connect to metastore with URI thrift://ctr-e139-1542663976389-62755-01-000014.hwx.site:9083
2019-02-08T12:28:31,327 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - HMSC::open(): Could not find delegation token. Creating KERBEROS-based thrift connection.
2019-02-08T12:28:31,336 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Opened a connection to metastore, current connections: 4
2019-02-08T12:28:31,337 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.HiveMetaStoreClient (:()) - Connected to metastore.
2019-02-08T12:28:31,337 INFO  [HiveServer2-Background-Pool: Thread-1134]: metastore.RetryingMetaStoreClient (:()) - RetryingMetaStoreClient proxy=class org.apache.hadoop.hive.ql.metadata.SessionHiveMetaStoreClient ugi=hive/[e-mail] (auth:KERBEROS) retries=24 delay=5 lifetime=0
2019-02-08T12:28:31,642 INFO  [HiveServer2-Background-Pool: Thread-1134]: common.FileUtils (FileUtils.java:mkdir(580)) - Creating directory if it doesn't exist: hdfs://mycluster1/warehouse/tablespace/managed/hive/nondefault.db/nondefault_table2/base_0000001

Issue Links

• links to: GitHub Pull Request #529