[
https://issues.apache.org/jira/browse/HBASE-28150?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Duo Zhang reopened HBASE-28150:
-------------------------------
> Too many master region WAL cause HDFS disk space full
> -----------------------------------------------------
>
> Key: HBASE-28150
> URL: https://issues.apache.org/jira/browse/HBASE-28150
> Project: HBase
> Issue Type: Bug
> Affects Versions: 2.4.14
> Reporter: chaijunjie
> Assignee: chaijunjie
> Priority: Critical
>
> create a table, but it failed when execute CREATE_TABLE_WRITE_FS_LAYOUT, then
> will try again and again, will write too many proc record to master:store, we
> find num of the master WAL in oldWALs more than 13000..
>
> Q: should add a suspend time logic for create table proc retry? i see
> TransitRegionStateProcedure has the logic..
>
> ---------------------------------------------------------------
> sorry, i upload screenshot failed, just copy to here
> {code:java}
> // 2023-10-12 12:34:35,360 | INFO | RegionOpenAndInit-themis:a-pool-0 |
> Closing region themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. |
> org.apache.hadoop.hbase.regionserver.HRegion.doClose(HRegion.java:1688)
> 2023-10-12 12:34:35,360 | INFO | RegionOpenAndInit-themis:a-pool-0 | Closed
> themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. |
> org.apache.hadoop.hbase.regionserver.HRegion.doClose(HRegion.java:1900)
> 2023-10-12 12:34:35,360 | INFO | PEWorker-1 | Region directories are created
> at hdfs://hacluster/hbase/.tmp for table themis:a |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:346)
> 2023-10-12 12:34:35,362 | WARN | PEWorker-1 | Retriable error trying to
> create table=themis:a state=CREATE_TABLE_WRITE_FS_LAYOUT |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:159)
> java.io.IOException: Unable to move table from
> temp=hdfs://hacluster/hbase/.tmp/data/themis/a to hbase
> root=hdfs://hacluster/hbase/data/themis/a
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.moveTempDirectoryToHBaseRoot(CreateTableProcedure.java:391)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:350)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:318)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:121)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:75)
> at
> org.apache.hadoop.hbase.procedure2.StateMachineProcedure.execute(StateMachineProcedure.java:188)
> at
> org.apache.hadoop.hbase.procedure2.Procedure.doExecute(Procedure.java:922)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.execProcedure(ProcedureExecutor.java:1650)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.executeProcedure(ProcedureExecutor.java:1396)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.access$1000(ProcedureExecutor.java:75)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor$WorkerThread.runProcedure(ProcedureExecutor.java:1962)
> at org.apache.hadoop.hbase.trace.TraceUtil.trace(TraceUtil.java:221)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor$WorkerThread.run(ProcedureExecutor.java:1988)
> 2023-10-12 12:34:35,387 | INFO | PEWorker-1 | pid=917,
> state=RUNNABLE:CREATE_TABLE_WRITE_FS_LAYOUT, locked=true;
> CreateTableProcedure table=themis:a execute
> state=CREATE_TABLE_WRITE_FS_LAYOUT |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:102)
> 2023-10-12 12:34:35,414 | INFO | RegionOpenAndInit-themis:a-pool-0 |
> creating {ENCODED => 513d3d5b4d3ad5c8f13bacea4a888d69, NAME =>
> 'themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69.', STARTKEY => '',
> ENDKEY => ''}, tableDescriptor='themis:a', {NAME => 'f1', BLOOMFILTER =>
> 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE',
> DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER',
> MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536',
> REPLICATION_SCOPE => '0'}, regionDir=hdfs://hacluster/hbase/.tmp |
> org.apache.hadoop.hbase.regionserver.HRegion.createHRegion(HRegion.java:7906)
> 2023-10-12 12:34:35,432 | INFO | RegionOpenAndInit-themis:a-pool-0 | Waiting
> for flushes and compactions to finish for the region
> themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. |
> org.apache.hadoop.hbase.regionserver.HRegion.waitForFlushesAndCompactions(HRegion.java:1911)
> 2023-10-12 12:34:35,432 | INFO | RegionOpenAndInit-themis:a-pool-0 | Total
> wait time for flushes and compaction for the region
> themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. is: 0ms |
> org.apache.hadoop.hbase.regionserver.HRegion.waitForFlushesAndCompactions(HRegion.java:1946)
> 2023-10-12 12:34:35,432 | INFO | RegionOpenAndInit-themis:a-pool-0 | Closing
> region themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. |
> org.apache.hadoop.hbase.regionserver.HRegion.doClose(HRegion.java:1688)
> 2023-10-12 12:34:35,432 | INFO | RegionOpenAndInit-themis:a-pool-0 | Closed
> themis:a,,1697025107991.513d3d5b4d3ad5c8f13bacea4a888d69. |
> org.apache.hadoop.hbase.regionserver.HRegion.doClose(HRegion.java:1900)
> 2023-10-12 12:34:35,432 | INFO | PEWorker-1 | Region directories are created
> at hdfs://hacluster/hbase/.tmp for table themis:a |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:346)
> 2023-10-12 12:34:35,434 | WARN | PEWorker-1 | Retriable error trying to
> create table=themis:a state=CREATE_TABLE_WRITE_FS_LAYOUT |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:159)
> java.io.IOException: Unable to move table from
> temp=hdfs://hacluster/hbase/.tmp/data/themis/a to hbase
> root=hdfs://hacluster/hbase/data/themis/a
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.moveTempDirectoryToHBaseRoot(CreateTableProcedure.java:391)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:350)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.createFsLayout(CreateTableProcedure.java:318)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:121)
> at
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:75)
> at
> org.apache.hadoop.hbase.procedure2.StateMachineProcedure.execute(StateMachineProcedure.java:188)
> at
> org.apache.hadoop.hbase.procedure2.Procedure.doExecute(Procedure.java:922)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.execProcedure(ProcedureExecutor.java:1650)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.executeProcedure(ProcedureExecutor.java:1396)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor.access$1000(ProcedureExecutor.java:75)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor$WorkerThread.runProcedure(ProcedureExecutor.java:1962)
> at org.apache.hadoop.hbase.trace.TraceUtil.trace(TraceUtil.java:221)
> at
> org.apache.hadoop.hbase.procedure2.ProcedureExecutor$WorkerThread.run(ProcedureExecutor.java:1988)
> 2023-10-12 12:34:35,469 | INFO | PEWorker-1 | pid=917,
> state=RUNNABLE:CREATE_TABLE_WRITE_FS_LAYOUT, locked=true;
> CreateTableProcedure table=themis:a execute
> state=CREATE_TABLE_WRITE_FS_LAYOUT |
> org.apache.hadoop.hbase.master.procedure.CreateTableProcedure.executeFromState(CreateTableProcedure.java:102)
> {code}
> {code:java}
> //hdfs dfs -ls /hbase/oldWALs | grep 'masterlocal' |wc -l
> 13398
> {code}
>
> analysis:
> this was beacuse {color:#ff0000}*i delete namespace dir in HDFS
> directly*{color}...but {color:#ff0000}*did not delete from
> hbase:namespce*{color}, so when i want to create a table in this namespce
> will hang....
> it's a operation error...
> but if some logic failed in CreateTableProcedure, i think will cause this
> issue again...
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
