[Hadoop-common-user] Hadoop mapred - Out of swap space

What's the RAM on each node?

7GB

Sent from my mobile

On 06/11/2010, at 11:00, "Hari Sreekumar" wrote:

What's the RAM on each node?

On Sat, Nov 6, 2010 at 11:03 AM, Shavit Netzer wrote:

Hello,

I have a question regarding MapRed jobs.

I have 24 nodes, each node have 4 disks (mnt – mnt3), 500GB each mnt.

All balanced ( I used the balancer, except mnt, which have 97% used).

My question is:
I got the following error and I relate it to the disk space (maybe I'm
wrong).

Maybe there is a configuration that I can add, change in order to have few
more retries on separate disk:


10/10/27 21:59:01 INFO mapred.JobClient: map 100% reduce 26%

10/10/27 21:59:02 INFO mapred.JobClient: Task Id :
attempt_201010201240_4059_r_000023_0, Status : FAILED

java.io.IOException: Task process exit with nonzero status of 134.

at
org.apache.hadoop.mapred.TaskRunner.runChild(TaskRunner.java:462)

at
org.apache.hadoop.mapred.TaskRunner.run(TaskRunner.java:403)



attempt_201010201240_4059_r_000023_0: #

attempt_201010201240_4059_r_000023_0: # A fatal error has been detected by
the Java Runtime Environment:

attempt_201010201240_4059_r_000023_0: #

attempt_201010201240_4059_r_000023_0: # java.lang.OutOfMemoryError:
requested 32744 bytes for ChunkPool::allocate. Out of swap space?

attempt_201010201240_4059_r_000023_0: #

attempt_201010201240_4059_r_000023_0: # Internal Error
(allocation.cpp:117), pid=15974, tid=1089702224

attempt_201010201240_4059_r_000023_0: # Error: ChunkPool::allocate

attempt_201010201240_4059_r_000023_0: #

attempt_201010201240_4059_r_000023_0: # JRE version: 6.0_14-b08

attempt_201010201240_4059_r_000023_0: # Java VM: Java HotSpot(TM) 64-Bit
Server VM (14.0-b16 mixed mode linux-amd64 )

attempt_201010201240_4059_r_000023_0: # An error report file with more
information is saved as:

attempt_201010201240_4059_r_000023_0: #
/mnt2/hadoop/mapred/local/taskTracker/jobcache/job_201010201240_4059/attempt_201010201240_4059_r_000023_0/work/hs_err_pid15974.log

attempt_201010201240_4059_r_000023_0: #

attempt_201010201240_4059_r_000023_0: # If you would like to submit a bug
report, please visit:

attempt_201010201240_4059_r_000023_0: #
http://java.sun.com/webapps/bugreport/crash.jsp

attempt_201010201240_4059_r_000023_0: #

Regards,
Shavit

It's an out of mem error, so I feel it has to do with ram rather tham disk
space. Did you check if it's swapping? (top/htop)... Is your reduce phase
very mem-intensive? Seems to be a memory leak somewhere.. What does htop
say? What processes are you running on each node? What does the log file
that it is showing say?

Hi,

I will get the information and will reply to you.

Sent from my mobile