Throughout my seven years working with Cloudera/Hortonworks, I'm always learning new things.  One thing I've learned from the Cloudera/Hortonoworks merger was how amazing CDSW/CML is as a product.  CML isn't only for Data Scientists, it's for anyone that needs an IDE (Integrated Development Environment).  Coming from a development background working with Eclipse and IntelliJ, you become dependent on a solid IDE.   In the past, I've used IDEs to develop applications, and then a build process would eventually deploy to a run environment.  This is where CML shines, you're able to run your applications within CDP at scale with enormous amounts of data.  Anyone using CML is over the moon running their applications/projects within CML because of the simplicity, as I'll demonstrate below.  The tagline of CML is "BYOL" (Bring your own libraries), meaning ALL libraries are welcome (outside of the Cloudera ecosystem).  This differentiates Cloudera from others such as native Azure, where native Azure is HIGHLY dependent on all things Microsoft (unless the application owner created something native Azure-specific).  I'll demonstrate how easy deploying a third party such as Django "The web framework for perfectionists with deadlines." where the installation/run feels like you're running on your local laptop, instead of a highly scalable IDE that runs anywhere.     Remember that CML runs ANYWHERE, within the cloud providers such as Azure, AWS, or GCP, and on-premise.  Cloudera abstracts out the complexities.  Using CML, we'll go from installing Django and then running Django in a matter of minutes.     Step 1: Find the read-only URL to run your embedded application (Django)   import os url=os.environ["CDSW_ENGINE_ID"]+"."+os.environ["CDSW_DOMAIN"] print("http://read-only-%s"%url)     Step 2: Install Django   !pip install django     Step 3: Create a Django project as instructed here   !django-admin startproject mysite cd mysite     Step 4: Modify the settings.py file adding the 'read-only-%s' value from step 1 and localhost   ALLOWED_HOSTS = ['localhost','read-only-yourhostnamefromstep1']     Step 5: Run Django   !python manage.py runserver localhost:$CDSW_READONLY_PORT     That's IT!  If you'd like to access your Django page, navigate to the URL in step 1!  Nothing specialized for CML, as we say BYOL!   ... View more

In order to debug pairing DLM, you'll need the following pre-req: 1) Root access to the DPS VM Problem statement - have you received an error when pairing a cluster? Follow these step-by-step instructions to access the DLM log, to gain granular log information that will help you debug: 1) Run the "sudo docker ps" command to gain the container id for "dlm-app": In the image above, the container id for "dlm-app" is "83d879e9a45e". 2) Once you receive the container id, you can run the following command "sudo docker exec -it 83d879e9a45e /bin/tailf /usr/dlm-app/logs/application.log" This will give you insight into the DPS-DLM application, in the example above you'll see "ERROR". The error log will post once you click "pair" in the DLM UI. Using the information from the log, you'll be able to troubleshoot your issue. ... View more

Reading and writing files to a MapR cluster (version 6) is simple, using the standard PutFile or GetFile, utilizing the MapR NFS. If you've searched high and low on how to do this, you've likely read articles and GitHub projects specifying steps. I've tried these steps without success, meaning whats out there is too complicated or out-dated to solve NiFi reading/writing to MapR. You don't need to re-compile the HDFS processors with the MapR dependencies, just follow the steps below: 1) Install the MapR client on each NiFi node #Install syslinux (for rpm install) sudo yum install syslinux #Download the RPM for your OS http://package.mapr.com/releases/v6.0.0/redhat/ rpm -Uvh mapr-client-6.0.0.20171109191718.GA-1.x86_64.rpm #Configure the mapr client connecting with the cldb /opt/mapr/server/configure.sh -c -N ryancicak.com -C cicakmapr0.field.hortonworks.com:7222 -genkeys -secure #Once you have the same users/groups on your OS (as MapR), you will be able to use maprlogin password (allowing you to login with a Kerberos ticket) #Prove that you can access the MapR FS hadoop fs -ls / 2) Mount the MaprR FS on each NiFi node sudo mount -o hard,nolock cicakmapr0.field.hortonworks.com:/mapr /mapr *This will allow you to access the MapRFS on the mount point /mapr/yourclustername.com/location 3) Use the PutFile and GetFile processor referencing the /mapr directory on your NiFi nodes *Following 1-3 allows you to quickly read/write to MapR, using NiFi. ... View more

Step-by-step instructions on how-to install a CRAN package from a local repo - without an internet connection. I'll be installing the package called "tidyr". In order to fully install the package, I need to first download tidyr and all dependencies. To do this, I use the CRAN PACKAGE: https://cran.r-project.org/src/contrib/PACKAGES 1) Finding "Package: tidyr", I can see the dependencies (imports): Imports: dplyr (>= 0.7.0), glue, magrittr, purrr, Rcpp, rlang, stringi, tibble, tidyselect and from them, I can build a list of all the imported packages I need: assertthat_0.2.0.tar.gz BH_1.66.0-1.tar.gz bindr_0.1.1.tar.gz bindrcpp_0.2.tar.gz cli_1.0.0.tar.gz crayon_1.3.4.tar.gz dplyr_0.7.4.tar.gz glue_1.2.0.tar.gz magrittr_1.5.tar.gz pillar_1.2.1.tar.gz plogr_0.1-1.tar.gz purrr_0.2.4.tar.gz Rcpp_0.12.16.tar.gz rlang_0.2.0.tar.gz stringi_1.1.7.tar.gz tibble_1.4.2.tar.gz tidyr_0.8.0.tar.gz tidyselect_0.2.4.tar.gz utf8_1.1.3.tar.gz R6_2.2.2.tar.gz pkgconfig_2.0.1.tar.gz 2) Download the packages from https://cran.r-project.org/src/contrib/ 3) Create a directory to emulate the CRAN repo, in my example I created /tmp/ryantester/src/contrib 4) Create a PACKAGES file within /tmp/ryantester/src/contrib - since this tutorial covers tidyr, I'll include the necessary packages Package: R6 Version: 2.2.2 Depends: R (>= 3.0) Suggests: knitr, microbenchmark, pryr, testthat, ggplot2, scales License: MIT + file LICENSE NeedsCompilation: no Package: pkgconfig Version: 2.0.1 Imports: utils Suggests: covr, testthat, disposables (>= 1.0.3) License: MIT + file LICENSE NeedsCompilation: no Package: bindr Version: 0.1.1 Suggests: testthat License: MIT + file LICENSE NeedsCompilation: no Package: bindrcpp Version: 0.2 Imports: Rcpp, bindr LinkingTo: Rcpp, plogr Suggests: testthat License: MIT + file LICENSE NeedsCompilation: yes Package: plogr Version: 0.1-1 Suggests: Rcpp License: MIT + file LICENSE NeedsCompilation: no Package: BH Version: 1.66.0-1 License: BSL-1.0 NeedsCompilation: no Package: plogr Version: 0.1-1 Suggests: Rcpp License: MIT + file LICENSE NeedsCompilation: no Package: dplyr Version: 0.7.4 Depends: R (>= 3.1.2) Imports: assertthat, bindrcpp (>= 0.2), glue (>= 1.1.1), magrittr, methods, pkgconfig, rlang (>= 0.1.2), R6, Rcpp (>= 0.12.7), tibble (>= 1.3.1), utils LinkingTo: Rcpp (>= 0.12.0), BH (>= 1.58.0-1), bindrcpp, plogr Suggests: bit64, covr, dbplyr, dtplyr, DBI, ggplot2, hms, knitr, Lahman (>= 3.0-1), mgcv, microbenchmark, nycflights13, rmarkdown, RMySQL, RPostgreSQL, RSQLite, testthat, withr License: MIT + file LICENSE NeedsCompilation: yes Package: utf8 Version: 1.1.3 Depends: R (>= 2.10) Suggests: corpus, knitr, rmarkdown, testthat License: Apache License (== 2.0) | file LICENSE NeedsCompilation: yes Package: assertthat Version: 0.2.0 Imports: tools Suggests: testthat License: GPL-3 NeedsCompilation: no Package: cli Version: 1.0.0 Depends: R (>= 2.10) Imports: assertthat, crayon, methods Suggests: covr, mockery, testthat, withr License: MIT + file LICENSE NeedsCompilation: no Package: crayon Version: 1.3.4 Imports: grDevices, methods, utils Suggests: mockery, rstudioapi, testthat, withr License: MIT + file LICENSE NeedsCompilation: no Package: pillar Version: 1.2.1 Imports: cli (>= 1.0.0), crayon (>= 1.3.4), methods, rlang (>= 0.2.0), utf8 (>= 1.1.3) Suggests: knitr (>= 1.19), lubridate, testthat (>= 2.0.0) License: GPL-3 NeedsCompilation: no Package: tidyselect Version: 0.2.4 Depends: R (>= 3.1) Imports: glue, purrr, rlang (>= 0.2.0), Rcpp (>= 0.12.0) LinkingTo: Rcpp (>= 0.12.0), Suggests: covr, dplyr, testthat License: GPL-3 NeedsCompilation: yes Package: tibble Version: 1.4.2 Depends: R (>= 3.1.0) Imports: cli, crayon, methods, pillar (>= 1.1.0), rlang, utils Suggests: covr, dplyr, import, knitr (>= 1.5.32), microbenchmark, mockr, nycflights13, rmarkdown, testthat, withr License: MIT + file LICENSE NeedsCompilation: yes Package: stringi Version: 1.1.7 Depends: R (>= 2.14) Imports: tools, utils, stats License: file LICENSE License_is_FOSS: yes NeedsCompilation: yes Package: rlang Version: 0.2.0 Depends: R (>= 3.1.0) Suggests: crayon, knitr, methods, pillar, rmarkdown (>= 0.2.65), testthat, covr License: GPL-3 NeedsCompilation: yes Package: Rcpp Version: 0.12.16 Depends: R (>= 3.0.0) Imports: methods, utils Suggests: RUnit, inline, rbenchmark, knitr, rmarkdown, pinp, pkgKitten (>= 0.1.2) License: GPL (>= 2) NeedsCompilation: yes Package: purrr Version: 0.2.4 Depends: R (>= 3.1) Imports: magrittr (>= 1.5), rlang (>= 0.1), tibble Suggests: covr, dplyr (>= 0.4.3), knitr, rmarkdown, testthat License: GPL-3 | file LICENSE NeedsCompilation: yes Package: magrittr Version: 1.5 Suggests: testthat, knitr License: MIT + file LICENSE NeedsCompilation: no Package: glue Version: 1.2.0 Depends: R (>= 3.1) Imports: methods Suggests: testthat, covr, magrittr, crayon, knitr, rmarkdown, DBI, RSQLite, R.utils, forcats, microbenchmark, rprintf, stringr, ggplot2 License: MIT + file LICENSE NeedsCompilation: yes Package: tidyr Version: 0.8.0 Depends: R (>= 3.2) Imports: dplyr (>= 0.7.0), glue, magrittr, purrr, Rcpp, rlang, stringi, tibble, tidyselect LinkingTo: Rcpp Suggests: covr, gapminder, knitr, rmarkdown, testthat License: MIT + file LICENSE NeedsCompilation: yes 5) Move the downloaded packages from #1 and #2 to the /tmp/ryantester/src/contrib 6) The final step is to install, pointing to your local repo (in our case, /tmp/ryantester) install.packages('tidyr', repos = "file:///tmp/ryantester") ... View more

Installing the Alarm Fatigue Demo via Cloudbreak: There are multiple ways to deploy the Alarm Fatigue Demo via Cloudbreak. Below are four options: 1) Deploy via the Cloudbreak UI a) Login to https://cbdtest.field.hortonworks.com b) Select your credentials – if you credentials don’t exist, create them under “Manage Credentials” c) Once your credentials are selected, click “Create Cluster” d) Make-up a cluster name and choose the Availability Zone (SE) and then click “Setup Network and Security" e) “fieldcloud-openstack-network” should be selected and click “Choose Blueprint” f) Select the Blueprint called “alarm_fatigue_v2” Host Group 1 (Select Ambari Server, alarm-fatigue-demo and pre-install-java8) Host Group 2 (select pre-install-java8) Host Group 3 (select pre-install-java8) g) Click on “Review and Launch” e) Click on “Create and start cluster” (After clicking, the deployment via Cloudbreak will likely take 30-50 minutes, go get a coffee) 2) Deploy via Bash Script (specifying configuration file) Create file .deploy.config with the following Version=0.5 CloudBreakServer= https://cbdtest.field.hortonworks.com CloudBreakIdentityServer= http://cbdtest.field.hortonworks.com:8089 CloudBreakUser=admin@example.com CloudBreakPassword=yourpassword CloudBreakCredentials= CloudBreakClusterName=alarmfatigue-auto CloudBreakTemplate=openstack-m3-xlarge CloudBreakRegion=RegionOne CloudBreakSecurityGroup=openstack-connected-platform-demo-all-services-port-v3 CloudBreakNetwork=fieldcloud-openstack-network CloudBreakAvailabilityZone=SE Change the highlighted Then execute the following: wget -O - https://raw.githubusercontent.com/ryancicak/northcentral_hackathon/master/CloudBreakArtifacts/cloudbreak-cmd/deployer.sh| bash 3) Deploy via Bash Script inputting configurations (while prompted) Just execute wget -O -https://raw.githubusercontent.com/ryancicak/northcentral_hackathon/master/CloudBreakArtifacts/cloudbreak-cmd/deployer.sh| bash and fill out the information as prompted 4) Deploy via Jenkins All four options will deploy install, configure and run all necessary services including "Alarm Fatigue Demo Control" ... View more

Many technologists ask "What is the easiest way to import data into Hive?" It's true, you can write a Sqoop job (making sure YARN containers are tuned properly in regards to # of mappers, the size of the mappers, and the sort allocated memory). And you'll deal with a two step process of moving the RDBMS tables into plain text and then creating an ORC table with an INSERT INTO statement.... I don't call this easy. What if there were a tool where you could simply drag-and-drop processes (processors) on a canvas and connect them together creating a workflow? There is a tool for this called NiFi! Before we go any further, I'll assume you know and understand NiFi - what is a processor, scheduling a processor, what is a connection (relationship), what are different relationship types, etc:. Small-Medium RDBMS tables This article will cover small to medium sized RDBMS tables. This means I can run a single select * from myrdbmstable; without returning millions of rows (where we're taxing the RDBMS), I'll write a second article on how-to use the processor GenerateTableFetch that generates select queries that fetch "pages" of rows from a table. Using pages of rows from a table will distribute the select queries amongst multiple NiFi nodes, similar to what Sqoop does with the #mappers where each mapper pages through results. Prerequisites for Hive In order to stream data into Hive, we'll utilize Hive's transactional capabilities which require the following: 1) Enable ACID in Hive 2) Use bucketing on your Hive table 3) Store the Hive table as ORC 4) Set the following property on your Hive table TBLPROPERTIES ("transactional"="true") 1-4 will be followed below in Step 2 Step 1 - Query an RDBMS table using the QueryDatabaseTable processor As described above, choose a small-medium sized RDBMS table (we'll tackle large database tables in another article). a) Add the QueryDatabaseTable processor to your canvas You'll see you we need to choose a Database Connection Pooling Service (which we'll define below in step 2), add a table name and finally create a successful relationship b) Create an RDBMS Connection Pooling Service - Right click on the processor and go to Configure c) Under the "Properties" tab, click the property "Database Connection Pooling Service" and click "Create new service..." on the drop-down d) Choose the "DBCPConnectionPool" and click on Create e) Click on the arrow to go to configure the DBCPConnectionPool f) Click on the pencil to Edit the DBCPConnectionPool g) Change the name in Settings - to something that is easily identifiable for your database connection pool h) Finally go to Properties and define your connection. I'm creating a connection for MySQL, but the general rule of thumb is if a JDBC driver exists, you'll be able to connect (for example I wrote this article to connect to Teradata from within NiFi) Database Connection URL: jdbc:mysql://localhost:3306/hortonworks Database Driver Class Name: com.mysql.jdbc.Driver Database Driver Location(s): /Users/rcicak/Desktop/mysql-connector-java-5.0.8-bin.jar Click Apply - and you've created your RDBMS connection pool i) Enable your connection pool j) Define a table name - in our case we'll choose "people3", where the table people3 is described in MySQL as the following: *Note: Don't forget the Maximum-value columns, QueryDatabaseTable will keep track of the last row that was fetched mysql> DESCRIBE people3; +-------+--------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-------+--------------+------+-----+---------+-------+ | id | datetime | NO | PRI | NULL | | | name | varchar(255) | YES | | NULL | | | age | int(11) | YES | | NULL | | +-------+--------------+------+-----+---------+-------+ 3 rows in set (0.00 sec) k) Change the schedule of the QueryDatabaseProcessor to run every 10 minutes (or whatever your requirement calls for), therefore the select statement is not being execute multiple times per second Step 2 - Stream the RDBMS rows into Hive using the PutHiveStreaming processor a) Add the PutHiveStreamingProcessor to your canvas b) Create a successful relationship between QueryDatabaseTable (in step 1) and PutHiveStreaming (step 2) c) The caution went away on QueryDatabaseTable after we added the relationship, now we'll need to resolve the caution on the PutHiveStreaming processor d) Right click on the PutHiveStreaming processor and choose Configure, going to the "Properties" tab - Adding the following Hive configurations: Hive Metastore URI: thrift://cregion-hdpmaster2.field.hortonworks.com:9083 Database Name: default Table Name: people3 *Adjust the "Transactions per Batch" accordingly to your SLA requirements e) Verify your Hive table "people3" exists - and as explained in the Hive prerequisites above, you'll need ACID enabled, store the table as ORC, table properties set to transactional = true and also bucketing in order for PutHiveStreaming to work create table people3 (id timestamp, name varchar(255), age int) CLUSTERED BY(id) INTO 3 BUCKETS STORED AS ORC tblproperties("transactional"="true"); You've successfully imported your RDBMS table into Hive with two processors - wasn't that easy? ... View more

Prerequisites 1) Service Ambari Infra installed -> Ranger will use Ambari Infra's SolrCloud for Ranger Audit 2) MySQL installed and running (I'll use Hive's Metastore MySQL instance * MySQL is one of the many DB options) Installing Apache Ranger using Ambari Infra (SolrCloud) for Ranger Audit 1) Find the location of mysql-connector-java.jar (assume /usr/share/java/mysql-connector-java.jar) run the following command on the Ambari server sudo ambari-server setup --jdbc-db=mysql --jdbc-driver=/usr/share/java/mysql-connector-java.jar 2) In Ambari, click Add Service 3) Choose Ranger and click Next 4) Choose "I have met all the requirements above." and click Proceed (this was done in #1 above) 5) Assign master(s) for "Ranger Usersync" and "Ranger Admin" and click Next 6) Assign Slaves and Clients - since we did not install Apache Atlas, Ranger TagSync is not required and click Next 7) Customize Services -> Ranger Audit, click on "OFF" to enable SolrCloud Before clicking: After clicking: 😎 Customize Services -> Ranger Admin, enter "Ranger DB host" the DB you chose (in my case, I chose MySQL) and a password "Ranger DB password" for the user rangeradmin *Ranger will automatically add the user "rangeradmin" Add the proper credentials for a DB user that has administrator credentials (this administrator will create the user rangeradmin and Ranger tables) MySQL create an administrator user *Note: rcicak2.field.hortonworks.com is the server where Ranger is being installed CREATE USER 'ryan'@'rcicak2.field.hortonworks.com' IDENTIFIED BY 'lebronjamesisawesome'; GRANT ALL PRIVILEGES ON *.* TO 'ryan'@'rcicak2.field.hortonworks.com' WITH GRANT OPTION; Click Next 9) Review -> Click Deploy * Install, Start and Test will show you the progess of Ranger installing 10) Choose Ranger in Ambari 11) Choose "Configs" and "Ranger Plugin" and select the services you'd like Ranger to authorize (You'll need to restart the service after saving changes) ... View more

The goal of this article is to ingest log data from multiple servers (running MiNiFi) that push log data to a NiFi cluster. The NiFi cluster will listen for the log data on an input port and route to an HDFS directory (determined by the host name). This article will assume you are using Ambari for NiFi installation/administration. 1) Set the nifi.remote.input.host and nifi.remote.input.socket.port of NiFi cluster Reason: Allows NiFi cluster to use an Input Port (MiNiFi will push to a Remote Processing Group) * NiFi will listen using the Input Port a) In Ambari, go to NiFi b) Choose the Configs tab c) Choose the Advanced nifi-properties d) Set the nifi.remote.input.host to your NiFi hostname and nifi.remote.input.socket.port to 10000 e) Restart NiFi using Ambari 2) In NiFi – Create a flow for incoming log data (listening [input port] for MiNiFi data) Reason: Listen for incoming log data and route the log data to an HDFS directory a) On the NiFi Flow Canvas, drag-and-drop an Input port (Name your Input Port – I named mine “listen_for_minifi_logs”) b) Drag-and-drop processor RouteOnAttribute c) Create a connection between Input Port and RouteOnAttribute d) Configure RouteOnAttribute - Properties (removing the red caution), adding two properties (one property per server you’re installing MiNiFi on) I have two servers – rcicak0.field.hortonworks.com and rcicak1.field.hortonworks.com, the incoming log data (flowfile) will contain an attribute called “host_name” where we’ll properly route the flowfile depending on the host_name property e) Drag-and-drop three putHDFS processors & create a connection using hostname_rcicak0, hostname_rcicak1 and unmatched f) Each putHDFS processor will have a different HDFS directory – configure properties for each putHDFS processor /tmp/rcicak0/, /tmp/rcicak1/ and /tmp/unmatched g) Configure the HDFS directory – properties (adding a core-site.xml and the directory – depending on the connection) h) Play the processors – at this point, the NiFi flow is ready to receive Log data from MiNiFi 3) Setup MiNiFi on at least one server Reason: MiNiFi needs to push the log data to a remote processing group and delete the log file a) Download MiNiFi (from http://hortonworks.com/downloads/ ) on each of the servers (that contain log data) b) Unzip minifi-0.0.1-bin.zip to a directory c) Complete step 4 below before continuing to d d) Running with an account that has the read/write permission to the log data directory (to read the file and delete the file) run *location/minifi-0.0.1/bin/minifi.sh start 4) Using a processor group in NiFi, create a MiNiFi Flow (pushing log data to a remote processing group) Reason: Push the log data to a remote processing group and delete the log file a) Create a process group (call the group “minifi_flow”) b) Go into the process group “minifi_flow" c) Drag-and-drop the processor GetFile d) Configure the processor GetFile – Properties (IMPORTANT: Any file matching the file filter’s regular expression under input directory [and the recursive subdirectories when set to true], the file will be deleted once the file is stored in MiNiFi’s content repository In the example above, the file filter looks for hdfs-audit.log.Archive -> in this case is a date e) Drag-and-drop the UpdateAttribute processor and create a successful connection between GetFile and UpdateAttribute f) Configure the UpdateAttribute processor – Properties, setting the host_name attribute adding the nifi expression language getting the hostname g) Drag-and-drop a remote process group – use the nifi.remote.input.host from above for the URL Wait for the connection to establish before continuing to h h) Add a connection between UpdateAttribute and the Remote Process Group – under To Input choose listen_for_minifi_logs i) Select all processors and relationships to create a template (Download the template’s xml file) j) Use the minifi-toolkit (https://www.apache.org/dyn/closer.lua?path=/nifi/minifi/0.0.1/minifi-toolkit-0.0.1-bin.zip ) and run the config.sh tool “config.sh transform theminifi_flow_template.xml config.yml” -> which will convert the XML to a YML file that will be read by MiNiFi k) Copy the config.yml file into the minifi-0.0.1/conf directory on each of the MiNiFi servers (if you already have your MiNiFi agent started, restart the agent) ... View more

If you've received the error exitCode=7 after enabling Kerberos, you are hitting this Jira bug. Notice the bug outlines the issue but does not outline a solution. The good news is the solution is simple, as I'll document below. Problem: If you've enabled Kerberos through Ambari, you'll get through around 90-95% of the last step "Start and Test Services" and then receive the error: 16/09/26 23:42:49 INFO mapreduce.Job: Running job: job_1474928865338_0022 16/09/26 23:42:55 INFO mapreduce.Job: Job job_1474928865338_0022 running in uber mode : false 16/09/26 23:42:55 INFO mapreduce.Job: map 0% reduce 0% 16/09/26 23:42:55 INFO mapreduce.Job: Job job_1474928865338_0022 failed with state FAILED due to: Application application_1474928865338_0022 failed 2 times due to AM Container for appattempt_1474928865338_0022_000002 exited with exitCode: 7 For more detailed output, check application tracking page: http://master2.fqdn.com:8088/cluster/app/application_1474928865338_0022 Then, click on links to logs of each attempt.Diagnostics: Exception from container-launch. Container id: container_e05_1474928865338_0022_02_000001 Exit code: 7 Stack trace: ExitCodeException exitCode=7: at org.apache.hadoop.util.Shell.runCommand(Shell.java:576) at org.apache.hadoop.util.Shell.run(Shell.java:487) at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:753) at org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor.launchContainer(LinuxContainerExecutor.java:371) at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:303) at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:82) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745) Shell output: main : command provided 1 main : run as user is ambari-qa main : requested yarn user is ambari-qa Container exited with a non-zero exit code 7 Failing this attempt. Failing the application. You'll notice running "Service Checks" for Tez, MapReduce2, YARN, Pig (any service that involves creating a YARN container) will fail with the exitCode=7. This is because in YARN, the local-dirs likely has the "noexec" flag specified meaning the binaries that are added to these directories cannot be executed. Solution: Open /etc/fstab (with the proper permissions) and remove the noexec flag under all mounted drives specified under "local-dirs" in YARN. Then either remount or reboot your machine - problem solved. ... View more