You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hadoop.apache.org by Onder SEZGIN <on...@gmail.com> on 2015/09/28 10:47:03 UTC
yarn nodemanager local-dirs are bad error & yarn resourcemanager
maximum-am-resource-percent is insufficient warning
Hi,
after i started hdfs and yarn deamons
and
run the example as below;
%HADOOP_PREFIX%\bin\yarn jar
%HADOOP_PREFIX%\share\hadoop\mapreduce\hadoop-mapreduce-examples-2.6.1.jar
wordcount /myfile.txt /out
i can see such error on yarn nodemanager logs;
15/09/28 11:23:17 ERROR nodemanager.LocalDirsHandlerService: Most of the
disks f
ailed. 1/1 local-dirs are bad: /tmp/hadoop-<MYUSER>/nm-local-dir; 1/1
log-dirs
are bad: /dep/logs/userlogs
and also i can see such warnings in resource manager logs
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is
insuf
icient to start a single application in queue, it is likely set too low.
skippi
g enforcement to allow at least one application to start
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is
insuf
icient to start a single application in queue for user, it is likely set
too lo
. skipping enforcement to allow at least one application to start
Afterwards, i can see the logs in resource manager and task is not finished
and stays hanged.
15/09/28 11:24:14 INFO capacity.LeafQueue: Application
application_1443428598680
_0001 from user: <MYUSER> activated in queue: default
15/09/28 11:24:14 INFO capacity.LeafQueue: Application added - appId:
applicatio
n_1443428598680_0001 user:
org.apache.hadoop.yarn.server.resourcemanager.schedul
er.capacity.LeafQueue$User@62ac4540, leaf-queue: default
#user-pending-applicati
ons: 0 #user-active-applications: 1 #queue-pending-applications: 0
#queue-active
-applications: 1
15/09/28 11:24:14 INFO capacity.CapacityScheduler: Added Application
Attempt app
attempt_1443428598680_0001_000001 to scheduler from user <MYUSER> in queue
defa
ult
15/09/28 11:24:14 INFO attempt.RMAppAttemptImpl:
appattempt_1443428598680_0001_0
00001 State change from SUBMITTED to SCHEDULED
15/09/28 11:33:18 INFO scheduler.AbstractYarnScheduler: Release request
cache is
cleaned up
Could anybody help me to run this simple wordcount example on my windows
box?
Best Regards
--
Regards
Onder
Re: yarn nodemanager local-dirs are bad error & yarn resourcemanager
maximum-am-resource-percent is insufficient warning
Posted by Varun Vasudev <vv...@apache.org>.
Hi Onder,
What are the permissions for /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs? Can you search for messages containing /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs in the node manager logs? There should be a message about why the dirs were marked as bad.
-Varun
From: Onder SEZGIN
Reply-To: <us...@hadoop.apache.org>
Date: Monday, September 28, 2015 at 2:17 PM
To: <us...@hadoop.apache.org>
Subject: yarn nodemanager local-dirs are bad error & yarn resourcemanager maximum-am-resource-percent is insufficient warning
Hi,
after i started hdfs and yarn deamons
and
run the example as below;
%HADOOP_PREFIX%\bin\yarn jar %HADOOP_PREFIX%\share\hadoop\mapreduce\hadoop-mapreduce-examples-2.6.1.jar wordcount /myfile.txt /out
i can see such error on yarn nodemanager logs;
15/09/28 11:23:17 ERROR nodemanager.LocalDirsHandlerService: Most of the disks f
ailed. 1/1 local-dirs are bad: /tmp/hadoop-<MYUSER>/nm-local-dir; 1/1 log-dirs
are bad: /dep/logs/userlogs
and also i can see such warnings in resource manager logs
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue, it is likely set too low. skippi
g enforcement to allow at least one application to start
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue for user, it is likely set too lo
. skipping enforcement to allow at least one application to start
Afterwards, i can see the logs in resource manager and task is not finished and stays hanged.
15/09/28 11:24:14 INFO capacity.LeafQueue: Application application_1443428598680
_0001 from user: <MYUSER> activated in queue: default
15/09/28 11:24:14 INFO capacity.LeafQueue: Application added - appId: applicatio
n_1443428598680_0001 user: org.apache.hadoop.yarn.server.resourcemanager.schedul
er.capacity.LeafQueue$User@62ac4540, leaf-queue: default #user-pending-applicati
ons: 0 #user-active-applications: 1 #queue-pending-applications: 0 #queue-active
-applications: 1
15/09/28 11:24:14 INFO capacity.CapacityScheduler: Added Application Attempt app
attempt_1443428598680_0001_000001 to scheduler from user <MYUSER> in queue defa
ult
15/09/28 11:24:14 INFO attempt.RMAppAttemptImpl: appattempt_1443428598680_0001_0
00001 State change from SUBMITTED to SCHEDULED
15/09/28 11:33:18 INFO scheduler.AbstractYarnScheduler: Release request cache is
cleaned up
Could anybody help me to run this simple wordcount example on my windows box?
Best Regards
--
Regards
Onder
Re: yarn nodemanager local-dirs are bad error & yarn resourcemanager
maximum-am-resource-percent is insufficient warning
Posted by Varun Vasudev <vv...@apache.org>.
Hi Onder,
What are the permissions for /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs? Can you search for messages containing /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs in the node manager logs? There should be a message about why the dirs were marked as bad.
-Varun
From: Onder SEZGIN
Reply-To: <us...@hadoop.apache.org>
Date: Monday, September 28, 2015 at 2:17 PM
To: <us...@hadoop.apache.org>
Subject: yarn nodemanager local-dirs are bad error & yarn resourcemanager maximum-am-resource-percent is insufficient warning
Hi,
after i started hdfs and yarn deamons
and
run the example as below;
%HADOOP_PREFIX%\bin\yarn jar %HADOOP_PREFIX%\share\hadoop\mapreduce\hadoop-mapreduce-examples-2.6.1.jar wordcount /myfile.txt /out
i can see such error on yarn nodemanager logs;
15/09/28 11:23:17 ERROR nodemanager.LocalDirsHandlerService: Most of the disks f
ailed. 1/1 local-dirs are bad: /tmp/hadoop-<MYUSER>/nm-local-dir; 1/1 log-dirs
are bad: /dep/logs/userlogs
and also i can see such warnings in resource manager logs
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue, it is likely set too low. skippi
g enforcement to allow at least one application to start
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue for user, it is likely set too lo
. skipping enforcement to allow at least one application to start
Afterwards, i can see the logs in resource manager and task is not finished and stays hanged.
15/09/28 11:24:14 INFO capacity.LeafQueue: Application application_1443428598680
_0001 from user: <MYUSER> activated in queue: default
15/09/28 11:24:14 INFO capacity.LeafQueue: Application added - appId: applicatio
n_1443428598680_0001 user: org.apache.hadoop.yarn.server.resourcemanager.schedul
er.capacity.LeafQueue$User@62ac4540, leaf-queue: default #user-pending-applicati
ons: 0 #user-active-applications: 1 #queue-pending-applications: 0 #queue-active
-applications: 1
15/09/28 11:24:14 INFO capacity.CapacityScheduler: Added Application Attempt app
attempt_1443428598680_0001_000001 to scheduler from user <MYUSER> in queue defa
ult
15/09/28 11:24:14 INFO attempt.RMAppAttemptImpl: appattempt_1443428598680_0001_0
00001 State change from SUBMITTED to SCHEDULED
15/09/28 11:33:18 INFO scheduler.AbstractYarnScheduler: Release request cache is
cleaned up
Could anybody help me to run this simple wordcount example on my windows box?
Best Regards
--
Regards
Onder
Re: yarn nodemanager local-dirs are bad error & yarn resourcemanager
maximum-am-resource-percent is insufficient warning
Posted by Varun Vasudev <vv...@apache.org>.
Hi Onder,
What are the permissions for /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs? Can you search for messages containing /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs in the node manager logs? There should be a message about why the dirs were marked as bad.
-Varun
From: Onder SEZGIN
Reply-To: <us...@hadoop.apache.org>
Date: Monday, September 28, 2015 at 2:17 PM
To: <us...@hadoop.apache.org>
Subject: yarn nodemanager local-dirs are bad error & yarn resourcemanager maximum-am-resource-percent is insufficient warning
Hi,
after i started hdfs and yarn deamons
and
run the example as below;
%HADOOP_PREFIX%\bin\yarn jar %HADOOP_PREFIX%\share\hadoop\mapreduce\hadoop-mapreduce-examples-2.6.1.jar wordcount /myfile.txt /out
i can see such error on yarn nodemanager logs;
15/09/28 11:23:17 ERROR nodemanager.LocalDirsHandlerService: Most of the disks f
ailed. 1/1 local-dirs are bad: /tmp/hadoop-<MYUSER>/nm-local-dir; 1/1 log-dirs
are bad: /dep/logs/userlogs
and also i can see such warnings in resource manager logs
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue, it is likely set too low. skippi
g enforcement to allow at least one application to start
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue for user, it is likely set too lo
. skipping enforcement to allow at least one application to start
Afterwards, i can see the logs in resource manager and task is not finished and stays hanged.
15/09/28 11:24:14 INFO capacity.LeafQueue: Application application_1443428598680
_0001 from user: <MYUSER> activated in queue: default
15/09/28 11:24:14 INFO capacity.LeafQueue: Application added - appId: applicatio
n_1443428598680_0001 user: org.apache.hadoop.yarn.server.resourcemanager.schedul
er.capacity.LeafQueue$User@62ac4540, leaf-queue: default #user-pending-applicati
ons: 0 #user-active-applications: 1 #queue-pending-applications: 0 #queue-active
-applications: 1
15/09/28 11:24:14 INFO capacity.CapacityScheduler: Added Application Attempt app
attempt_1443428598680_0001_000001 to scheduler from user <MYUSER> in queue defa
ult
15/09/28 11:24:14 INFO attempt.RMAppAttemptImpl: appattempt_1443428598680_0001_0
00001 State change from SUBMITTED to SCHEDULED
15/09/28 11:33:18 INFO scheduler.AbstractYarnScheduler: Release request cache is
cleaned up
Could anybody help me to run this simple wordcount example on my windows box?
Best Regards
--
Regards
Onder
Re: yarn nodemanager local-dirs are bad error & yarn resourcemanager
maximum-am-resource-percent is insufficient warning
Posted by Varun Vasudev <vv...@apache.org>.
Hi Onder,
What are the permissions for /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs? Can you search for messages containing /tmp/hadoop-<MYUSER>/nm-local-dir and /dep/logs/userlogs in the node manager logs? There should be a message about why the dirs were marked as bad.
-Varun
From: Onder SEZGIN
Reply-To: <us...@hadoop.apache.org>
Date: Monday, September 28, 2015 at 2:17 PM
To: <us...@hadoop.apache.org>
Subject: yarn nodemanager local-dirs are bad error & yarn resourcemanager maximum-am-resource-percent is insufficient warning
Hi,
after i started hdfs and yarn deamons
and
run the example as below;
%HADOOP_PREFIX%\bin\yarn jar %HADOOP_PREFIX%\share\hadoop\mapreduce\hadoop-mapreduce-examples-2.6.1.jar wordcount /myfile.txt /out
i can see such error on yarn nodemanager logs;
15/09/28 11:23:17 ERROR nodemanager.LocalDirsHandlerService: Most of the disks f
ailed. 1/1 local-dirs are bad: /tmp/hadoop-<MYUSER>/nm-local-dir; 1/1 log-dirs
are bad: /dep/logs/userlogs
and also i can see such warnings in resource manager logs
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue, it is likely set too low. skippi
g enforcement to allow at least one application to start
15/09/28 11:24:14 WARN capacity.LeafQueue: maximum-am-resource-percent is insuf
icient to start a single application in queue for user, it is likely set too lo
. skipping enforcement to allow at least one application to start
Afterwards, i can see the logs in resource manager and task is not finished and stays hanged.
15/09/28 11:24:14 INFO capacity.LeafQueue: Application application_1443428598680
_0001 from user: <MYUSER> activated in queue: default
15/09/28 11:24:14 INFO capacity.LeafQueue: Application added - appId: applicatio
n_1443428598680_0001 user: org.apache.hadoop.yarn.server.resourcemanager.schedul
er.capacity.LeafQueue$User@62ac4540, leaf-queue: default #user-pending-applicati
ons: 0 #user-active-applications: 1 #queue-pending-applications: 0 #queue-active
-applications: 1
15/09/28 11:24:14 INFO capacity.CapacityScheduler: Added Application Attempt app
attempt_1443428598680_0001_000001 to scheduler from user <MYUSER> in queue defa
ult
15/09/28 11:24:14 INFO attempt.RMAppAttemptImpl: appattempt_1443428598680_0001_0
00001 State change from SUBMITTED to SCHEDULED
15/09/28 11:33:18 INFO scheduler.AbstractYarnScheduler: Release request cache is
cleaned up
Could anybody help me to run this simple wordcount example on my windows box?
Best Regards
--
Regards
Onder