[ https://issues.apache.org/jira/browse/HBASE-24585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17142590#comment-17142590 ]
Michael Stack commented on HBASE-24585: --------------------------------------- HBASE-24616 has suggested fix for BoundedRecoveredHFilesOutputSink not being able to get TableDescriptor at all times. > Failed start recovering crash in standalone mode if procedure-based > distributed WAL split & hbase.wal.split.to.hfile=true > ------------------------------------------------------------------------------------------------------------------------- > > Key: HBASE-24585 > URL: https://issues.apache.org/jira/browse/HBASE-24585 > Project: HBase > Issue Type: Bug > Reporter: Michael Stack > Priority: Major > > (This description got redone after I figured out what was going on. > Previously it was just a litany of me banging around trying to learn > procedure-based WAL splitting and hbase.wal.split.to.hfile; no one needs to > read that; hence the refactor). > HBASE-24574 procedure-based distributed WAL splitting is enabled and > split-to-hflie too. A force crash requires recovery with ServerCrashProcedure > splitting old WALs on restart. The recovery fails because we get stuck. The > Master can't assign meta because it is being recovered. The recovery can't > make progress because it is asking for a table descriptor for meta -- needed > by the hbase.wal.split.to.hfile feature -- and the master is not yet > initialized. After the default timeout, Master shuts down because it can't > initialize. > {code} > 2020-06-18 19:53:54,175 ERROR [main] master.HMasterCommandLine: Master > exiting > java.lang.RuntimeException: Master not initialized after 200000ms > at > org.apache.hadoop.hbase.util.JVMClusterUtil.waitForEvent(JVMClusterUtil.java:232) > at > org.apache.hadoop.hbase.util.JVMClusterUtil.startup(JVMClusterUtil.java:200) > at > org.apache.hadoop.hbase.LocalHBaseCluster.startup(LocalHBaseCluster.java:430) > at > org.apache.hadoop.hbase.master.HMasterCommandLine.startMaster(HMasterCommandLine.java:232) > at > org.apache.hadoop.hbase.master.HMasterCommandLine.run(HMasterCommandLine.java:140) > at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76) > at > org.apache.hadoop.hbase.util.ServerCommandLine.doMain(ServerCommandLine.java:149) > at org.apache.hadoop.hbase.master.HMaster.main(HMaster.java:3059) > {code} > The abort of Master interrupts other ongoing actions so later in the log > we'll see the WAL split show as interrupted > {code} > 2020-06-17 21:20:37,472 ERROR > [RS_LOG_REPLAY_OPS-regionserver/localhost:16020-0] > handler.RSProcedureHandler: Error when call RSProcedureCallable: > java.io.IOException: Failed WAL split, status=RESIGNED, > wal=file:/Users/stack/checkouts/hbase.apache.git/tmp/hbase/WALs/localhost,16020,1592440848604-splitting/localhost%2C16020%2C1592440848604.meta.1592440852959.meta > at > org.apache.hadoop.hbase.regionserver.SplitWALCallable.splitWal(SplitWALCallable.java:106) > at > org.apache.hadoop.hbase.regionserver.SplitWALCallable.call(SplitWALCallable.java:86) > at > org.apache.hadoop.hbase.regionserver.SplitWALCallable.call(SplitWALCallable.java:49) > at > org.apache.hadoop.hbase.regionserver.handler.RSProcedureHandler.process(RSProcedureHandler.java:49) > at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:104) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) > at java.lang.Thread.run(Thread.java:748) > {code} > This issue becomes how to make hbase.wal.split.to.hfile work in standalone > mode. -- This message was sent by Atlassian Jira (v8.3.4#803005)