I have two VM setup for the Hadoop cluster, as below.
VM-MASTER, 4GB Memory
VM-SLAVE, 4GB Memory
I have the following config for yarn-site.xml. When I goto http://VM-MASTEr:8088/cluster. I see Memory Total is 0, and VCores Total is 0.
Am I missing something here?
I think this problem caused the job I submitted always in ACCEPTED state, and never move into RUNNING state. I'm using Hadoop 2.8.0.
<configuration>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>500</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>2048</value>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>1</value>
</property>
</configuration>
Related
I have setup Hadoop cluster with 20GB RAM and 6 cores. I have around 8 GB data in 3 csv files and I have to join them. For this purpose, I have used Apache Hive for this. Hadoop, Hive are 3.x version.
Here is the Hive query
SELECT distinct rm.UID,rm.Num_Period , rpd.C_Mon-rpd.Non_Cred_Inputs as Claimed_Mon, rpd.Splr_UID, rpd.Doc_Type,rpd.Doc_No_Num ,rpd.Doc_Date, rpd.Purchased_Type,rpd.Rate_ID, rpd.C_Withheld, rpd.Non_Creditable_Inputs, rsd.G_UID , rsd.G_Type,rsd.Doc_Type as G_doc_type, rsd.Doc_No_Num as G_doc_no_num, rsd.Doc_Date as G_doc_date, rsd.Sale_Type as G_sale_type, rsd.Rate_ID as G_rate_id, rsd.Rate_Value as G_rate_value,rsd.hscode as G_hscode
from ZUniq rm inner join Zpurchasedetails rpd
on rm.UniqID = rpd.UniqID
inner join Zsaledetails rsd on rpd.UniqID = rsd.UniqID
where rpd.Non_Cred_Inputs < rpd.C_Mon;
Now,there is around 300 GB disk free on one node and 400 GB on other. When I run above query, all disks are used and then job goes to pending with a message that no healthy node exits.
Here is the Hadoop configuration
yarn-site.xml
<configuration>
<property>
<name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
<property>
<name>yarn.resourcemanager.scheduler.class</name>
<value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler</value>
<!-- <value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.fifo.FifoScheduler</value> -->
<!-- <value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler</value> -->
</property>
<!-- Site specific YARN configuration properties -->
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.env-whitelist</name>
<value>JAVA_HOME,HADOOP_COMMON_HOME,HADOOP_HDFS_HOME,HADOOP_CONF_DIR,CLASSPATH_PREPEND_DISTCACHE,HADOOP_YARN_HOME,HADOOP_MAPRED_HOME</value>
</property>
<property>
<name>yarn.nodemanager.local-dirs</name>
<value>/mnt/disk1/.hdfs/tmp</value>
</property>
<property>
<name>yarn.resourcemanager.hostname</name>
<value>hms-master</value>
</property>
<property>
<name>yarn.log-aggregation.retain-seconds</name>
<value>5184000</value>
<description>Delete the logs after 60 days </description>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>3</value>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-vcores</name>
<value>1</value>
</property>
<property>
<name>yarn.nodemanager.vmem-check-enabled</name>
<value>false</value>
<description>Whether virtual memory limits will be enforced for containers</description>
</property>
<property>
<name>yarn.nodemanager.vmem-pmem-ratio</name>
<value>4</value>
<description>Ratio between virtual memory to physical memory when setting memory limits for containers</description>
</property>
<!-- Logging related option -->
<property>
<name>yarn.log-aggregation-enable</name>
<value>true</value>
</property>
<property>
<name>yarn.log.server.url</name>
<value>http://hms-master:19888/jobhistory/logs</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>10240</value>
<description>Total RAM that can be used in single system by all containers.</description>
</property>
<property>
<name>yarn.scheduler.maximum-allocation-mb</name>
<value>10240</value>
<description>Maximum RAM that one continer can get </description>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>1024</value>
<description>Minimum RAM that one continer (e.g. map or reduce) can get. It should be less or equal to yarn.nodemanager.resource.memory-mb value </description>
</property>
</configuration>
I tried to run simple word count as MapReduce job. Everything works fine when run locally (all work done on Name Node). But, when I try to run it on a cluster using YARN (adding mapreduce.framework.name=yarn to mapred-site.conf) job hangs.
I came across a similar problem here:
MapReduce jobs get stuck in Accepted state
Output from job:
*** START ***
15/12/25 17:52:50 INFO client.RMProxy: Connecting to ResourceManager at /0.0.0.0:8032
15/12/25 17:52:51 WARN mapreduce.JobResourceUploader: Hadoop command-line option parsing not performed. Implement the Tool interface and execute your application with ToolRunner to remedy this.
15/12/25 17:52:51 INFO input.FileInputFormat: Total input paths to process : 5
15/12/25 17:52:52 INFO mapreduce.JobSubmitter: number of splits:5
15/12/25 17:52:52 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1451083949804_0001
15/12/25 17:52:53 INFO impl.YarnClientImpl: Submitted application application_1451083949804_0001
15/12/25 17:52:53 INFO mapreduce.Job: The url to track the job: http://hadoop-droplet:8088/proxy/application_1451083949804_0001/
15/12/25 17:52:53 INFO mapreduce.Job: Running job: job_1451083949804_0001
mapred-site.xml:
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<name>mapreduce.job.tracker</name>
<value>localhost:54311</value>
</property>
<!--
<property>
<name>mapreduce.job.tracker.reserved.physicalmemory.mb</name>
<value></value>
</property>
<property>
<name>mapreduce.map.memory.mb</name>
<value>1024</value>
</property>
<property>
<name>mapreduce.reduce.memory.mb</name>
<value>2048</value>
</property>
<property>
<name>yarn.app.mapreduce.am.resource.mb</name>
<value>3000</value>
<source>mapred-site.xml</source>
</property> -->
</configuration>
yarn-site.xml
<configuration>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
<!--
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>3000</value>
<source>yarn-site.xml</source>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>500</value>
</property>
<property>
<name>yarn.scheduler.capacity.maximum-am-resource-percent</name>
<value>3000</value>
</property>
-->
</configuration>
//I the left commented options - they were not solving the problem
YarnApplicationState: ACCEPTED: waiting for AM container to be allocated, launched and register with RM.
What can be the problem?
EDIT:
I tried this configuration (commented) on machines: NameNode(8GB RAM) + 2x DataNode (4GB RAM). I get the same effect: Job hangs on ACCEPTED state.
EDIT2:
changed configuration (thanks #Manjunath Ballur) to:
yarn-site.xml:
<configuration>
<property>
<name>yarn.resourcemanager.hostname</name>
<value>hadoop-droplet</value>
</property>
<property>
<name>yarn.resourcemanager.resource-tracker.address</name>
<value>hadoop-droplet:8031</value>
</property>
<property>
<name>yarn.resourcemanager.address</name>
<value>hadoop-droplet:8032</value>
</property>
<property>
<name>yarn.resourcemanager.scheduler.address</name>
<value>hadoop-droplet:8030</value>
</property>
<property>
<name>yarn.resourcemanager.admin.address</name>
<value>hadoop-droplet:8033</value>
</property>
<property>
<name>yarn.resourcemanager.webapp.address</name>
<value>hadoop-droplet:8088</value>
</property>
<property>
<description>Classpath for typical applications.</description>
<name>yarn.application.classpath</name>
<value>
$HADOOP_CONF_DIR,
$HADOOP_COMMON_HOME/*,$HADOOP_COMMON_HOME/lib/*,
$HADOOP_HDFS_HOME/*,$HADOOP_HDFS_HOME/lib/*,
$HADOOP_MAPRED_HOME/*,$HADOOP_MAPRED_HOME/lib/*,
$YARN_HOME/*,$YARN_HOME/lib/*
</value>
</property>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce.shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
<property>
<name>yarn.nodemanager.local-dirs</name>
<value>/data/1/yarn/local,/data/2/yarn/local,/data/3/yarn/local</value>
</property>
<property>
<name>yarn.nodemanager.log-dirs</name>
<value>/data/1/yarn/logs,/data/2/yarn/logs,/data/3/yarn/logs</value>
</property>
<property>
<description>Where to aggregate logs</description>
<name>yarn.nodemanager.remote-app-log-dir</name>
<value>/var/log/hadoop-yarn/apps</value>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>50</value>
</property>
<property>
<name>yarn.scheduler.maximum-allocation-mb</name>
<value>390</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>390</value>
</property>
</configuration>
mapred-site.xml:
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<name>yarn.app.mapreduce.am.resource.mb</name>
<value>50</value>
</property>
<property>
<name>yarn.app.mapreduce.am.command-opts</name>
<value>-Xmx40m</value>
</property>
<property>
<name>mapreduce.map.memory.mb</name>
<value>50</value>
</property>
<property>
<name>mapreduce.reduce.memory.mb</name>
<value>50</value>
</property>
<property>
<name>mapreduce.map.java.opts</name>
<value>-Xmx40m</value>
</property>
<property>
<name>mapreduce.reduce.java.opts</name>
<value>-Xmx40m</value>
</property>
</configuration>
Still not working.
Additional info: I can see no nodes on cluster preview (similar problem here: Slave nodes not in Yarn ResourceManager )
You should check the status of Node managers in your cluster. If the NM nodes are short on disk space then RM will mark them "unhealthy" and those NMs can't allocate new containers.
1) Check the Unhealthy nodes: http://<active_RM>:8088/cluster/nodes/unhealthy
If the "health report" tab says "local-dirs are bad" then it means you need to cleanup some disk space from these nodes.
2) Check the DFS dfs.data.dir property in hdfs-site.xml. It points the location on local file system where hdfs data is stored.
3) Login to those machines and use df -h & hadoop fs - du -h commands to measure the space occupied.
4) Verify hadoop trash and delete it if it's blocking you.
hadoop fs -du -h /user/user_name/.Trash and hadoop fs -rm -r /user/user_name/.Trash/*
I feel, you are getting your memory settings wrong.
To understand the tuning of YARN configuration, I found this to be a very good source: http://www.cloudera.com/content/www/en-us/documentation/enterprise/latest/topics/cdh_ig_yarn_tuning.html
I followed the instructions given in this blog and was able to get my jobs running. You should alter your settings proportional to the physical memory you have on your nodes.
Key things to remember is:
Values of mapreduce.map.memory.mb and mapreduce.reduce.memory.mb should be at least yarn.scheduler.minimum-allocation-mb
Values of mapreduce.map.java.opts and mapreduce.reduce.java.opts should be around "0.8 times the value of" corresponding mapreduce.map.memory.mb and mapreduce.reduce.memory.mb configurations. (In my case it is 983 MB ~ (0.8 * 1228 MB))
Similarly, value of yarn.app.mapreduce.am.command-opts should be "0.8 times the value of" yarn.app.mapreduce.am.resource.mb
Following are the settings I use and they work perfectly for me:
yarn-site.xml:
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>1228</value>
</property>
<property>
<name>yarn.scheduler.maximum-allocation-mb</name>
<value>9830</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>9830</value>
</property>
mapred-site.xml
<property>
<name>yarn.app.mapreduce.am.resource.mb</name>
<value>1228</value>
</property>
<property>
<name>yarn.app.mapreduce.am.command-opts</name>
<value>-Xmx983m</value>
</property>
<property>
<name>mapreduce.map.memory.mb</name>
<value>1228</value>
</property>
<property>
<name>mapreduce.reduce.memory.mb</name>
<value>1228</value>
</property>
<property>
<name>mapreduce.map.java.opts</name>
<value>-Xmx983m</value>
</property>
<property>
<name>mapreduce.reduce.java.opts</name>
<value>-Xmx983m</value>
</property>
You can also refer to the answer here: Yarn container understanding and tuning
You can add vCore settings, if you want your container allocation to take into account CPU also. But, for this to work, you need to use CapacityScheduler with DominantResourceCalculator. See the discussion about this here: How are containers created based on vcores and memory in MapReduce2?
This has solved my case for this error:
<property>
<name>yarn.scheduler.capacity.maximum-am-resource-percent</name>
<value>100</value>
</property>
Check your hosts file on master and slave nodes. I had exactly this problem. My hosts file looked like this on master node for example
127.0.0.0 localhost
127.0.1.1 master-virtualbox
192.168.15.101 master
I changed it like below
192.168.15.101 master master-virtualbox localhost
So it worked.
These lines
<property>
<name>yarn.nodemanager.disk-health-checker.max-disk-utilization-per-disk-percentage</name>
<value>100</value>
</property>
in the yarn-site.xml solved my problem since the node will be marked as unhealthy when disk usage is >=95%. Solution mainly suitable for pseudodistributed mode.
You have 512 MB RAM on each of the instance and all your memory configurations in yarn-site.xml and mapred-site.xml are 500 MB to 3 GB. You will not be able to run any thing on the cluster. Change every thing to ~256 MB.
Also your mapred-site.xml is using framework to by yarn and you have job tracker address which is not correct. You need to have resource manager related parameters in yarn-site.xml on a multinode cluster (including resourcemanager web address). With out that, the cluster does not know where your cluster is.
You need to revisit both your xml files.
anyway that's work for me .thank you a lot! #KaP
that's my yarn-site.xml
<property>
<name>yarn.resourcemanager.hostname</name>
<value>MacdeMacBook-Pro.local</value>
</property>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.resourcemanager.webapp.address</name>
<value>${yarn.resourcemanager.hostname}:8088</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>4096</value>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>2048</value>
</property>
<property>
<name>yarn.nodemanager.vmem-pmem-ratio</name>
<value>2.1</value>
that's my mapred-site.xml
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
The first thing is to check yarn resource manager logs. I had searched the Internet about this problem for a very long time, but nobody told me how to find out what is really happening. It's so straightforward and simple to check yarn resource manager logs. I am confused why people ignore logs.
For me, there was a error in log
Caused by: org.apache.hadoop.net.ConnectTimeoutException: 20000 millis timeout while waiting for channel to be ready for connect. ch : java.nio.channels.SocketChannel[connection-pending remote=172.16.0.167/172.16.0.167:55622]
That's because I switched wifi network in my work place, so my computer IP changed.
Old question, but I got on the same issue recently and in my case it was due to manually setting the master to local in the code.
Please, search for conf.setMaster("local[*]") and remove it.
Hope it helps.
I finally have got my Hadoop 2.6 running on rasbian on a RB Pi 2. I'm now going to tune it before deploying another node.
My current config is the pretty much default from a tutorial I've found on :
http://nextgenhadoop.blogspot.pt/2013/10/steps-to-install-hadoop-220-stable.html
And some troubleshooting tips from stackoverflow.
Due to the RAM available and CPU the Pi2 has, I'm sure there will be a 'ideal' config to it, I've been messing with yarn-site.xml adding and removing memory in a way that seems logical to me, but the Pi doesnt agree.
Right now I have 1 node, with '8 vcores' and '8GB' of ram. This can't of course be true
Can anyone share their config with me ?
These two propeties from yarn-site.xml file are relevant to nodemanager resources:
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>4096</value>
<description>Physical memory, in MB, to be made available to running containers</description>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>4</value>
<description>Number of CPU cores that can be allocated for containers.</description>
</property>
You can find more in Hadoop documentation
I've managed to trim it down, here is my congig, please feel free to comment on them. thanks
my yarn-site.xml
<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>128</value>
<description>Minimum limit of memory to allocate to each container request at the Resource Manager.</description>
</property>
<property>
<name>yarn.scheduler.maximum-allocation-mb</name>
<value>768</value>
<description>Maximum limit of memory to allocate to each container request at the Resource Manager.</description>
</property>
<property>
<name>yarn.scheduler.minimum-allocation-vcores</name>
<value>1</value>
<description>The minimum allocation for every container request at the RM, in terms of virtual CPU cores. Requests lower than this won't take effect, and the specified value will get allocated the minimum.</description>
</property>
<property>
<name>yarn.scheduler.maximum-allocation-vcores</name>
<value>2</value>
<description>The maximum allocation for every container request at the RM, in terms of virtual CPU cores. Requests higher than this won't take effect, and will get capped to this value.</description>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>768</value>
<description>Physical memory, in MB, to be made available to running containers</description>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>2</value>
<description>Number of CPU cores that can be allocated for containers.</description>
</property>
and my mapred-site.xml
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<name>mapred.job.tracker</name>
<value>hadoopi00:54311</value>
</property>
<property>
<name>yarn.app.mapreduce.am.resource.mb</name>
<value>768</value>
</property>
<property>
<name>mapreduce.map.memory.mb</name>
<value>512</value>
</property>
<property>
<name>mapreduce.reduce.memory.mb</name>
<value>512</value>
</property>
<property>
<name>mapreduce.map.java.opts</name>
<value>-Xmx384m</value>
</property>
<property>
<name>mapreduce.reduce.java.opts</name>
<value>-Xmx384m</value>
</property>
</configuration>
I know there is another question about this, but there are no answers yet, so I'm going to try and ask it in a more detail.
I am running a map-reduce job using Hadoop 2.2.0 on a 2 node cluster that I have setup on Amazon 2 EC2 instances; the master node is a medium instance and the slave node is also a medium instance. It runs extremely slowly, it takes over 17 minutes, but when I run the same exact job on the same cluster without yarn it runs in under 1 minute. Here is what my mapred-site.xml looks like:
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
</configuration>
If I change the mapreduce.framework property to 'local, so that the file simply reads:
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>local</value>
</property>
</configuration>
I can then run the same map-reduce job in less than a minute. However, I would like to use YARN, so that I can track the map-reduce job through the webapp. When I run the job with the mapreduce.framework property set to yarn it takes 17+ minutes to run the same exact job. I cannot imagine that YARN would slow down a map-reduce job to such an extreme level.
I am also using "top" to track my CPU usage, and it seems that when I run it with yarn, the CPU usage is split between the different nodes, however, when I change run it with "local" all of the processing is done on the master node. I'm not sure how this makes sense, because it seems to me, that when the CPU processing is split between the different nodes, it should run faster, not slower. Is there something I forgot to configure in Hadoop to make running on a cluster faster?
Here are the rest of my configuration files:
core-site.xml
<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://namenode:8020</value>
</property>
</configuration>
hdfs-site.xml
<configuration>
<property>
<name>dfs.replication</name>
<value>2</value>
</property>
<property>
<name>dfs.permissions</name>
<value>false</value>
</property>
<property>
<name>fs.checkpoint.dir</name>
<value>file:/home/ubuntu/hadoop/hdfs/snn</value>
</property>
<property>
<name>fs.checkpoint.edits.dir</name>
<value>file:/home/ubuntu/hadoop/hdfs/snn</value>
</property>
<property>
<name>dfs.namenode.name.dir</name>
<value>file:/home/ubuntu/hadoop/hdfs/nn</value>
</property>
<property>
<name>dfs.namenode.name.dir</name>
<value>file:/home/ubuntu/hadoop/hdfs/nn</value>
</property>
</configuration>
yarn-site.xml
<configuration>
<property>
<name>yarn.resourcemanager.resource-tracker.address</name>
<value>namenode:8031</value>
</property>
<property>
<name>yarn.resourcemanager.address</name>
<value>namenode:8032</value>
</property>
<property>
<name>yarn.resourcemanager.scheduler.address</name>
<value>namenode:8030</value>
</property>
<property>
<name>yarn.resourcemanager.admin.address</name>
<value>namenode:8033</value>
</property>
<property>
<name>yarn.resourcemanager.webapp.address</name>
<value>namenode:8088</value>
</property>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
</configuration>
Is there something wrong with the way I set this up? Has anyone else ran into this problem? Any help will be greatly appreciated, Thanks!
I wish i still remember where I read this so I can give you a reference. You won't benefit with yarn unless you have large size cluster.
A node in my YARN cluster has 64GB memory and 24 cores. I set the following properties in the yarn-site.xml:
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>32768</value>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>16</value>
</property>
But I found still the nodemanager in the node has only 7 containers. What are other properties I need to set?
You need to tell YARN how to break down the memory to containers so for instance if you set the memory per container to 2GB will give you 16 containers
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>2048</value>
try something like:
<property>
<name>mapreduce.map.memory.mb</name>
<value>1024</value>
</property>
<property>
<name>mapreduce.reduce.memory.mb</name>
<value>2048</value>
</property>