Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hadoop >> mail # general >> problem starting cdh3b2 jobtracker


Copy link to this message
-
Re: problem starting cdh3b2 jobtracker
The IOException in MRAsyncDiskService is being logged with severity level
WARN; I believe that system operation continues normally despite being
unable to clean some of the directories.

Are you experiencing problems where a partial mis-match of mapred.local.dir
configuration and available disks caused a daemon to fail to start? If so,
can you post the configuration and log?

I'm inclined to believe that the fact that the above warning appears in
Ted's log immediately before the fatal problem (trying to bind to a port
already in use) is a red herring. It would be quite surprising to me if the
disk cleanup service affected the socket binding component of the
jobtracker.
- Aaron

On Fri, Aug 6, 2010 at 3:43 AM, Harsh J <[EMAIL PROTECTED]> wrote:

> java.io.IOException: Cannot create toBeDeleted in /data1/mapred/local
>
> This line points at the solution actually. In earlier versions of CDH
> if the list of local mapred directories had false ones (like say the
> jobtracker machine not having 2 disks like all the tasktracking
> machines and it not being in the slaves list either), it used to
> ignore it. Now it doesn't seem to and instead tries to operate things
> upon it? Looks like a major bug Cloudera folks! Encountered this using
> CDH3 +320. Not using my jobtracker machine to perform tasks as well.
>
> It gets resolved after you validate the mapred local directory list on
> the job tracker machine's config alone. However, this would lead to
> issues with conf-syncing between nodes if it acts this way forever.
>
> On Fri, Jul 2, 2010 at 8:32 AM, Ted Yu <[EMAIL PROTECTED]> wrote:
> > We installed cdh3b2 0.20.2+320 and saw some strange error in jobtracker
> log:
> >
> > 2010-07-02 01:49:31,977 INFO org.apache.hadoop.mapred.JobTracker:
> JobTracker
> > up at: 9001
> > 2010-07-02 01:49:31,977 INFO org.apache.hadoop.mapred.JobTracker:
> JobTracker
> > webserver: 50030
> > 2010-07-02 01:49:31,988 WARN org.apache.hadoop.mapred.JobTracker: Error
> > starting tracker: java.io.IOException: Cannot create toBeDeleted in
> > /data1/mapred/local
> >    at
> >
> org.apache.hadoop.util.MRAsyncDiskService.<init>(MRAsyncDiskService.java:85)
> >    at org.apache.hadoop.mapred.JobTracker.<init>(JobTracker.java:1688)
> >    at
> org.apache.hadoop.mapred.JobTracker.startTracker(JobTracker.java:199)
> >    at
> org.apache.hadoop.mapred.JobTracker.startTracker(JobTracker.java:191)
> >    at org.apache.hadoop.mapred.JobTracker.main(JobTracker.java:3765)
> >
> > 2010-07-02 01:49:32,990 INFO org.apache.hadoop.mapred.JobTracker:
> Scheduler
> > configured with (memSizeForMapSlotOnJT, memSizeForReduceSlotOnJT,
> > limitMaxMemForMapTasks, limitMaxMemForReduceTasks) (-1, -1, -1, -1)
> > 2010-07-02 01:49:32,991 FATAL org.apache.hadoop.mapred.JobTracker:
> > java.net.BindException: Problem binding to
> > sjc1-hadoop0.sjc1.ciq.com/10.201.8.204:9001<
> http://sjc1-hadoop0.sjc1.carrieriq.com/10.201.8.204:9001>:
> > Address already in use
> >    at org.apache.hadoop.ipc.Server.bind(Server.java:198)
> >    at org.apache.hadoop.ipc.Server$Listener.<init>(Server.java:261)
> >    at org.apache.hadoop.ipc.Server.<init>(Server.java:1043)
> >    at org.apache.hadoop.ipc.RPC$Server.<init>(RPC.java:492)
> >    at org.apache.hadoop.ipc.RPC.getServer(RPC.java:454)
> >    at org.apache.hadoop.mapred.JobTracker.<init>(JobTracker.java:1628)
> >    at
> org.apache.hadoop.mapred.JobTracker.startTracker(JobTracker.java:199)
> >    at
> org.apache.hadoop.mapred.JobTracker.startTracker(JobTracker.java:191)
> >    at org.apache.hadoop.mapred.JobTracker.main(JobTracker.java:3765)
> > Caused by: java.net.BindException: Address already in use
> >    at sun.nio.ch.Net.bind(Native Method)
> >    at
> > sun.nio.ch.ServerSocketChannelImpl.bind(ServerSocketChannelImpl.java:119)
> >    at sun.nio.ch.ServerSocketAdaptor.bind(ServerSocketAdaptor.java:59)
> >    at org.apache.hadoop.ipc.Server.bind(Server.java:196)
> >    ... 8 more
> >
> > 2010-07-02 01:49:32,992 INFO org.apache.hadoop.mapred.JobTracker:
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB