尝试将数据插入HBase时无限期暂停

时间：2015-11-17 08:57:32

标签： java hadoop garbage-collection hbase

在尝试将数据插入HBase时，我看到在执行多次写入（~100000000）后，插入操作只会挂起（UI中写入0次/秒）。主服务器和区域服务器保持运行状态，java HBase客户端进程似乎仍然活跃。

我只能看到以下日志条目。可能是什么原因造成的？如何解决？

hbase-master.log 

2015-11-17 02:16:52,007 DEBUG [htable-pool279-t1] ipc.AbstractRpcClient: Use SIMPLE authentication for service ClientService, sasl=false
2015-11-17 02:16:52,007 DEBUG [htable-pool279-t1] ipc.AbstractRpcClient: Connecting to server2.example.com/<IP>:16020
2015-11-17 02:17:17,966 DEBUG [ProcedureExecutorTimeout] procedure2.ProcedureExecutor$CompletedProcedureCleaner: No completed procedures to cleanup.
2015-11-17 02:17:47,966 DEBUG [ProcedureExecutorTimeout] procedure2.ProcedureExecutor$CompletedProcedureCleaner: No completed procedures to cleanup.
2015-11-17 02:18:17,966 DEBUG [ProcedureExecutorTimeout] procedure2.ProcedureExecutor$CompletedProcedureCleaner: No completed procedures to cleanup.
2015-11-17 02:18:47,966 DEBUG [ProcedureExecutorTimeout] procedure2.ProcedureExecutor$CompletedProcedureCleaner: No completed procedures to cleanup.

HBase的-master.out 

2015-11-16T03:26:46.519-0600: [CMS-concurrent-sweep-start]
2015-11-16T03:26:46.519-0600: [CMS-concurrent-sweep: 0.000/0.000 secs] [Times: user=0.00 sys=0.00, real=0.00 secs]
2015-11-16T03:26:46.519-0600: [CMS-concurrent-reset-start]
2015-11-16T03:26:46.580-0600: [CMS-concurrent-reset: 0.061/0.062 secs] [Times: user=0.22 sys=0.05, real=0.06 secs]
2015-11-16T03:36:46.351-0600: [GC (Allocation Failure) 2015-11-16T03:36:46.351-0600: [ParNew: 297460K->14971K(309696K), 0.0321644 secs] 297460K->20626K(997824K), 0.0323361 secs] [Times: user=0.18 sys=0.02, real=0.03 secs]
2015-11-16T07:30:46.984-0600: [GC (Allocation Failure) 2015-11-16T07:30:46.984-0600: [ParNew: 290299K->6737K(309696K), 0.0220497 secs] 295954K->12392K(997824K), 0.0222000 secs] [Times: user=0.09 sys=0.01, real=0.02 secs]
2015-11-16T07:56:52.016-0600: [GC (Allocation Failure) 2015-11-16T07:56:52.017-0600: [ParNew: 282065K->7829K(309696K), 0.0071377 secs] 287720K->13484K(997824K), 0.0072568 secs] [Times: user=0.11 sys=0.00, real=0.01 secs]
2015-11-16T08:42:46.946-0600: [GC (Allocation Failure) 2015-11-16T08:42:46.946-0600: [ParNew: 283157K->6516K(309696K), 0.0056529 secs] 288812K->12171K(997824K), 0.0057721 secs] [Times: user=0.08 sys=0.00, real=0.00 secs]

区域的server.log 

2015-11-17 01:24:08,833 INFO  [MemStoreFlusher.1] regionserver.HStore: Added hdfs://server0.example.com:9000/hbase/data/default/MYTABLE/323da4642dc634afdbbc1fef03568a04/node_info/49b6b7c247c64cf49b75414f30cc1646, entries=302884, sequenceid=91538227, filesize=31.6 M
2015-11-17 01:24:08,833 INFO  [MemStoreFlusher.1] regionserver.HRegion: Finished memstore flush of ~67.84 MB/71133312, currentsize=0 B/0 for region MYTABLE,,1447717440847.323da4642dc634afdbbc1fef03568a04. in 579ms, sequenceid=91538227, compaction requested=true
2015-11-17 01:24:53,525 INFO  [regionserver/server1.example.com/<IP>:16020.logRoller] wal.FSHLog: Rolled WAL /hbase/WALs/server1.example.com,16020,1447666006438/server1.example.com%2C16020%2C1447666006438.default.1447741493458 with entries=126566, filesize=26.02 MB; new WAL /hbase/WALs/server1.example.com,16020,1447666006438/server1.example.com%2C16020%2C1447666006438.default.1447745093512
2015-11-17 01:24:53,526 INFO  [regionserver/server1.example.com/<IP>:16020.logRoller] wal.FSHLog: Archiving hdfs://server0.example.com:9000/hbase/WALs/server1.example.com,16020,1447666006438/server1.example.com%2C16020%2C1447666006438.default.1447741081730 to hdfs://server0.example.com:9000/hbase/oldWALs/server1.example.com%2C16020%2C1447666006438.default.144
7741081730
2015-11-17 01:24:53,527 INFO  [regionserver/server1.example.com/9.12.235.213:16020.logRoller] wal.FSHLog: Archiving hdfs://server0.example.com:9000/hbase/WALs/server1.example.com,16020,1447666006438/server1.example.com%2C16020%2C1447666006438.def
ault.1447741493458 to hdfs://server0.example.com:9000/hbase/oldWALs/server1.example.com%2C16020%2C1447666006438.default.1447741493458
2015-11-17 01:26:47,759 INFO  [LruBlockCacheStatsExecutor] hfile.LruBlockCache: totalSize=7.78 MB, freeSize=6.21 GB, max=6.21 GB, blockCount=10, accesses=1953643, hits=30700, hitRatio=1.57%, , cachingAccesses=30779, cachingHits=30171, cachingHitsRatio=98.02%, evictions=7919, evicted=598, evictedPerRun=0.07551458477973938
2015-11-17 01:31:47,759 INFO  [LruBlockCacheStatsExecutor] hfile.LruBlockCache: totalSize=7.78 MB, freeSize=6.21 GB, max=6.21 GB, blockCount=10, accesses=1953643, hits=30700, hitRatio=1.57%, , cachingAccesses=30779, cachingHits=30171, cachingHitsRatio=98.02%, evictions=7949, evicted=598, evictedPerRun=0.07522958517074585
2015-11-17 01:36:47,759 INFO  [LruBlockCacheStatsExecutor] hfile.LruBlockCache: totalSize=7.78 MB, freeSize=6.21 GB, max=6.21 GB, blockCount=10, accesses=1953643, hits=30700, hitRatio=1.57%, , cachingAccesses=30779, cachingHits=30171, cachingHitsRatio=98.02%, evictions=7979, evicted=598, evictedPerRun=0.07494673877954483

区域-server.out 

2015-11-17T00:25:48.005-0600: [GC (Allocation Failure) 2015-11-17T00:25:48.005-0600: [ParNew: 307577K->1134K(309696K), 0.0038813 secs] 1265221K->962084K(1735184K), 0.0039946 secs] [Times: user=0.06 sys=0.00, real=0.00 secs]
2015-11-17T00:25:49.891-0600: [GC (Allocation Failure) 2015-11-17T00:25:49.891-0600: [ParNew: 276462K->1354K(309696K), 0.0025507 secs] 1237412K->962304K(1735184K), 0.0026612 secs] [Times: user=0.04 sys=0.00, real=0.00 secs]
2015-11-17T00:25:51.727-0600: [GC (Allocation Failure) 2015-11-17T00:25:51.727-0600: [ParNew: 276682K->34368K(309696K), 0.0050290 secs] 1237632K->997653K(1735184K), 0.0051420 secs] [Times: user=0.10 sys=0.00, real=0.01 secs]
2015-11-17T00:25:53.553-0600: [GC (Allocation Failure) 2015-11-17T00:25:53.553-0600: [ParNew: 309696K->1598K(309696K), 0.0040175 secs] 1272981K->966446K(1735184K), 0.0041283 secs] [Times: user=0.07 sys=0.00, real=0.00 secs]
2015-11-17T00:25:55.334-0600: [GC (Allocation Failure) 2015-11-17T00:25:55.334-0600: [ParNew: 276926K->34368K(309696K), 0.0048517 secs]1241774K->1001404K(1735184K), 0.0049688 secs] [Times: user=0.10 sys=0.00, real=0.00 secs]

1 个答案:

答案 0 :(得分：0)

似乎这个问题与老旧的HBase无关。由于我使用 opencsv CSVReader 来读取插入代码中大小为＆gt; 5GB的csv文件，因此导致插入数据时出现了可怕且无法解释的暂停。（问题类似于opencsv-not-reading-my-entire-file）。我刚刚更改了插入代码，使用BufferedReader读取文件，延迟消失了！

相关问题
最新问题