accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Keith Turner (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-2050) log recovery fails when using two namenodes and viewfs://
Date Thu, 19 Dec 2013 02:00:09 GMT


Keith Turner commented on ACCUMULO-2050:

I looked into this.  HadoopLogCloser currently calls recoverLease() on DistributedFileSystem.
 In the case of this exception instead of a DistributedFileSystem we have a ViewFileSystem
(there error message is misleading ACCUMULO-2063)  AFAICT  ViewFileSystem does not provide
a way to translate a viewfs:// path to a hdfs:// path in its API.  This would be needed to
get the corresponding DistributedFileSystem.  So it seems like the options are to not support
viewfs or to re-implement the viewfs translation logic in Accumulo.   I opened HDFS-5684 in
the course of investigating this.   I would really like to support viewfs, so I am going to
keep poking at this.   If anyone has suggestions I would love to hear them.

> log recovery fails when using two namenodes and viewfs://
> ---------------------------------------------------------
>                 Key: ACCUMULO-2050
>                 URL:
>             Project: Accumulo
>          Issue Type: Sub-task
>            Reporter: Keith Turner
>             Fix For: 1.6.0
> See parent ticket for env info.  Saw the following error in log recovery.
> {noformat}
> Failed to initiate log sort viewfs:/nn2/accumulo/wal/ip-10-1-3-25+9997/777b3e10-da00-47d2-819b-24464997d784
> 	java.lang.IllegalStateException: Don't know how to recover a lease for org.apache.accumulo.server.fs.VolumeManagerImpl
> 		at org.apache.accumulo.server.master.recovery.HadoopLogCloser.close(
> 		at org.apache.accumulo.master.recovery.RecoveryManager$
> 		at java.util.concurrent.Executors$
> 		at
> 		at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(
> 		at java.util.concurrent.ScheduledThreadPoolExecutor$
> 		at java.util.concurrent.ThreadPoolExecutor.runWorker(
> 		at java.util.concurrent.ThreadPoolExecutor$
> 		at
> 		at
> 		at
> {noformat}

This message was sent by Atlassian JIRA

View raw message