Could Hadoop framework (or runtime) prevent (or constrain) an application MapReduce program from accessing local resource like local filesystem?
I guess the answer should be true, especially when the MapReduce program is running a cluster.
A secure (Kerberized) cluster will run containers under the user that submitted the job. Ordinary access control can then isolate this user access to local resources.
Non secure clusters run containers as the NM (I'm talking a modern Yarn cluster, not a 1.x version).
Most recent Hadoop version (2.6, very soon to be released) contains YARN-1964 which allows Docker based containers. They are fully isolated (Docker) but this was committed in 2.6 on 2014-11-12, so is about 2 week mature. You'll be living on the edge.
Ofcourse, MapReduce will use local resources in Map/Reduce phase.
The output of Map will be stored in local file system and then suffle and sort will happen.
Next the data will be input to the Reduce phase.
You can specify the path for local path to store the intermediate results of Map by property in Hadoop V1 mapred.local.dir
Hadoop V2,
From the Docs,
Property : mapreduce.cluster.local.dir
Value : ${hadoop.tmp.dir}/mapred/local
Description : The local directory where MapReduce stores intermediate data files. May be a comma-separated list of directories on different devices in order to spread disk i/o. Directories that do not exist are ignored.
Hope it helps!
Related
What is the recommended DefaultFS (File system) for Hadoop on Dataproc. Are there any benchmarks, considerations available around using GCS vs HDFS as the default file system?
I was also trying to test things out and discovered that when I set the DefaultFS to a gs:// path, the Hive scratch files are getting created - both on HDFS as well as the GCS paths. Is this happening synchronously and adding to latency or does the write to GCS happen after the fact?
Would appreciate any guidance, reference around this.
Thank you
PS: These are ephemeral Dataproc clusters that are going to be using GCS for all persistent data.
HDFS is faster. There should already be public benchmarks for that, or just taken as a fact because GCS is networked storage where HDFS is directly mounted in the Dataproc VMs.
"Recommended" would be persistent storage, though, so GCS, but maybe only after finalizing the data in the applications. For example, you might not want Hive scratch files in GCS since they'll never be used outside of the current query session, but you would want Spark checkpoints if you're running periodic batch jobs that scale down the HDFS cluster in between executions
I would say the default (HDFS) is the recommended. Typically, the input and output data of Dataproc jobs are persisted outside of the cluster in GCS or BigQuery, the cluster is used for compute and intermediate data. These intermediate data are stored on local disks directly or through HDFS which eventually also goes to local disks. After the job is done, you can safely delete the cluster, only pay for the storage of input and output data to save cost.
Also HDFS usually has lower latency for intermediate data, especially for lots of small files and metadata operations, e.g. dir rename. GCS is better at throughput for large files.
But when using HDFS, you need to provision sufficient disk space (at least 1TB each node) and consider using local SSDs. See https://cloud.google.com/dataproc/docs/support/spark-job-tuning#optimize_disk_size for more details.
I want to store flink store in HDFS so that after crash I can recover the flink state from HDFS. I am planning to write state to HDFS every 60 second. How Can I achieve this ?
Is this the config I need to follow ?
https://ci.apache.org/projects/flink/flink-docs-release-1.4/ops/state/state_backends.html#setting-default-state-backend
And where do I specify the check point interval ? Any link or sample code would be helpful
Choosing where checkpoints are stored (e.g., HDFS) is separate from deciding which state backend to use for managing your working state (which can be on-heap, or in local files managed by the RocksDB library).
These two concepts were cleanly separated in Flink 1.12. In early versions of Flink, the two appeared to be more strongly related than they actually are because the filesystem and rocksdb state backend constructors took a file URI as a parameter, specifying where the checkpoints should be stored.
The best way to manage all of this is to leave this out of your code, and to specify the configuration you want in flink-conf.yaml, e.g.,
state.backend: filesystem
state.checkpoints.dir: hdfs://namenode-host:port/flink-checkpoints
execution.checkpointing.interval: 10s
Information about checkpointing and savepointing can be found at https://nightlies.apache.org/flink/flink-docs-master/docs/dev/datastream/fault-tolerance/checkpointing/
On how to configure HDFS as a filesystem, you should check https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/filesystems/overview/
Are there any dependencies between Spark and Hadoop?
If not, are there any features I'll miss when I run Spark without Hadoop?
Spark is an in-memory distributed computing engine.
Hadoop is a framework for distributed storage (HDFS) and distributed processing (YARN).
Spark can run with or without Hadoop components (HDFS/YARN)
Distributed Storage:
Since Spark does not have its own distributed storage system, it has to depend on one of these storage systems for distributed computing.
S3 – Non-urgent batch jobs. S3 fits very specific use cases when data locality isn’t critical.
Cassandra – Perfect for streaming data analysis and an overkill for batch jobs.
HDFS – Great fit for batch jobs without compromising on data locality.
Distributed processing:
You can run Spark in three different modes: Standalone, YARN and Mesos
Have a look at the below SE question for a detailed explanation about both distributed storage and distributed processing.
Which cluster type should I choose for Spark?
Spark can run without Hadoop but some of its functionality relies on Hadoop's code (e.g. handling of Parquet files). We're running Spark on Mesos and S3 which was a little tricky to set up but works really well once done (you can read a summary of what needed to properly set it here).
(Edit) Note: since version 2.3.0 Spark also added native support for Kubernetes
By default , Spark does not have storage mechanism.
To store data, it needs fast and scalable file system. You can use S3 or HDFS or any other file system. Hadoop is economical option due to low cost.
Additionally if you use Tachyon, it will boost performance with Hadoop. It's highly recommended Hadoop for apache spark processing.
As per Spark documentation, Spark can run without Hadoop.
You may run it as a Standalone mode without any resource manager.
But if you want to run in multi-node setup, you need a resource manager like YARN or Mesos and a distributed file system like HDFS,S3 etc.
Yes, spark can run without hadoop. All core spark features will continue to work, but you'll miss things like easily distributing all your files (code as well as data) to all the nodes in the cluster via hdfs, etc.
Yes, you can install the Spark without the Hadoop.
That would be little tricky
You can refer arnon link to use parquet to configure on S3 as data storage.
http://arnon.me/2015/08/spark-parquet-s3/
Spark is only do processing and it uses dynamic memory to perform the task, but to store the data you need some data storage system. Here hadoop comes in role with Spark, it provide the storage for Spark.
One more reason for using Hadoop with Spark is they are open source and both can integrate with each other easily as compare to other data storage system. For other storage like S3, you should be tricky to configure it like mention in above link.
But Hadoop also have its processing unit called Mapreduce.
Want to know difference in Both?
Check this article: https://www.dezyre.com/article/hadoop-mapreduce-vs-apache-spark-who-wins-the-battle/83
I think this article will help you understand
what to use,
when to use and
how to use !!!
Yes, of course. Spark is an independent computation framework. Hadoop is a distribution storage system(HDFS) with MapReduce computation framework. Spark can get data from HDFS, as well as any other data source such as traditional database(JDBC), kafka or even local disk.
Yes, Spark can run with or without Hadoop installation for more details you can visit -https://spark.apache.org/docs/latest/
Yes spark can run without Hadoop. You can install spark in your local machine with out Hadoop. But Spark lib comes with pre Haddop libraries i.e. are used while installing on your local machine.
You can run spark without hadoop but spark has dependency on hadoop win-utils. so some features may not work, also if you want to read hive tables from spark then you need hadoop.
Not good at english,Forgive me!
TL;DR
Use local(single node) or standalone(cluster) to run spark without Hadoop,but stills need hadoop dependencies for logging and some file process.
Windows is strongly NOT recommend to run spark!
Local mode
There are so many running mode with spark,one of it is called local will running without hadoop dependencies.
So,here is the first question:how to tell spark we want to run on local mode?
After read this official doc,i just give it a try on my linux os:
Must install java and scala,not the core content so skip it.
Download spark package
There are "without hadoop" and "hadoop integrated" 2 type of package
The most important thing is "without hadoop" do NOT mean run without hadoop but just not bundle with hadoop so you can bundle it with your custom hadoop!
Spark can run without hadoop(HDFS and YARN) but need hadoop dependency jar such as parquet/avro etc SerDe class,so strongly recommend to use "integrated" package(and you will found missing some log dependencies like log4j and slfj and other common utils class if chose "without hadoop" package but all this bundled with hadoop integrated pacakge)!
Run on local mode
Most simple way is just run shell，and you will see the welcome log
# as same as ./bin/spark-shell --master local[*]
./bin/spark-shell
Standalone mode
As same as blew,but different with step 3.
# Starup cluster
# if you want run on frontend
# export SPARK_NO_DAEMONIZE=true
./sbin/start-master.sh
# run this on your every worker
./sbin/start-worker.sh spark://VMS110109:7077
# Submit job or just shell
./bin/spark-shell spark://VMS110109:7077
On windows?
I kown so many people run spark on windown just for study,but here is so different on windows and really strongly NOT recommend to use windows.
The most important things is download winutils.exe from here and configure system variable HADOOP_HOME to point where winutils located.
At this moment 3.2.1 is the most latest release version of spark,but a bug is exist.You will got a exception like Illegal character in path at index 32: spark://xxxxxx:63293/D:\classe when run ./bin/spark-shell.cmd,only startup a standalone cluster then use ./bin/sparkshell.cmd or use lower version can temporary fix this.
For more detail and solution you can refer for here
No. It requires full blown Hadoop installation to start working - https://issues.apache.org/jira/browse/SPARK-10944
I am currently "playing around" with Hadoop in a VM (CDH4.1.3 image from cloudera). What I am wondering about is the following (and the documentation did not really help me in that regard).
Following the tutorial, I would format a NameNode first - OK, that is already done if one uses the cloudera image. Likewise the HDFS file structure is already present. In the hdfs-site.xml the datanode data dir is set to:
/var/lib/hadoop-hdfs/cache/${user.name}/dfs/data
which is obviously where the blocks are supposed to be copied to in a real distributed setting. In the cloudera tutorial, one is told to create hdfs "home directories" for each user (/users/<username>), which I do not understand what they are for. Are they just for local test-runs in a single-node setup?
Say I really had petabytes of data on type not fitting into my local storage. This data would have to be distributed straight away, rendering a local "home directory" entirely useless.
Could someone tell me, just to give me an intuition, how a real Hadoop workflow with massive data would look like? What kind of distinct nodes would I have running for a start?
There's the master (JobTracker) with its slave file (where would I put that) allowing the master to resolve all the DataNodes. Then there is my NameNode that keeps track of where the block IDs are stored. The DataNodes are also carry TaskTracker responsibility. In the config files, the NameNode's URI is included -- am I correct so far? Then there is still the ${user.name} variable in the configuration which apparently, if I understood it right, has something to do with WebHDFS, which would also be great if someone could explain to me. In the running examples, the directions tend to be hardcoded to
/var/lib/hadoop-hdfs/cache/1/dfs/data, /var/lib/hadoop-hdfs/cache/2/dfs/data and so on.
So, back to the example: Say, I have my tape and want to import data into my HDFS (and I am required to stream data into the filesystem because I lack the local storage to save it locally on a single machine). Where would I start with the migration process? On an arbitrary DataNode? On the NameNode that distributes the chunks? After all, I cannot assume the data just to "be there", because the name node has to be aware of the block IDs.
It would be great if someone could shortly elaborate on these topics:
What is the home directory really for?
Do I migrate data to the home directory first and to the real distributed system afterwards?
How does WebHDFS work and what role does it play with regards to the user.name variable
How would I migrate "big data" into my HDFS on the fly - or even if it's not big data, how do I populate my file system in a proper way (meaning, that the chunks get randomly distributed across the cluster?
What is the home directory really for?
You have a small confusion here. Just like /home exists for local filesystems on Linux, where users are given their own storage space, /users is a home mount ON the HDFS (Distributed FS). The tutorial needs you to administratively create a home directory for the user you wish to later be running data loads and queries as, such that they get adequate permissions and storage access onto the HDFS. The tutorial is not asking you to create these directories locally.
Do I migrate data to the home directory first and to the real distributed system afterwards?
I believe my above answer should clarify this for you. You should create your home directory on the HDFS, and then load all your data inside of that directory.
How does WebHDFS work and what role does it play with regards to the user.name variable
WebHDFS is one of the various ways to access HDFS. Regular clients to talk to HDFS require use of Java APIs. WebHDFS (and also HttpFs) techniques were added to HDFS to let other languages have their own set of APIs by providing a REST front-end to HDFS. WebHDFS allows user-authentication, to help persist the permission and security models.
How would I migrate "big data" into my HDFS on the fly - or even if it's not big data, how do I populate my file system in a proper way (meaning, that the chunks get randomly distributed across the cluster?
The large part of problem HDFS solves for you is that of managing distribution of data. When loading files or data streams to HDFS (via CLI tools, sinks from Apache Flume, etc.), the blocks are spread in an ideal distribution by HDFS itself, and the chunking is managed by it as well. All you need to do is use the user-side regular FileSystem style APIs and forget about what goes where underneath - its all managed for you.
I'm able to run a local mapper and reducer built using ruby with an input file.
I'm unclear about the behavior of the distributed system though.
For the production system, I have a HDFS set up across two machines. I know that if I store a large file on the HDFS, it will have some blocks on both machines to allow for parallelization. Do I also need to store the actual mappers and reducer files (my ruby files in this case) on the HDFS as well?
Also, how would I then go about actually running the streaming job so that it runs in a parallel manner on both systems?
If you were to use mapper/reducers written in ruby (or anything other than Java), you would have to use hadoop-streaming. Hadoop streaming has an option to package your mapper/reducer files when sending your job to the cluster. The following link should have what you are looking for.
http://hadoop.apache.org/common/docs/r0.15.2/streaming.html