My goal is to provide Hadoop jobs statistics web UI for administrative users.
I use HortonWorks Hadoop2 cluster and jobs run on YARN.
From the architecture perspective , I am planning to collect jobs related information ( such as start time, end time, mappers, etc ) from YARN Resource Manager REST API as scheduled cron job >> index them in to elastic search >> show them in Kibana.
I wonder if there is better way to do this.
Have you looked into Ambari? It provides metrics, dashboards, and alerting without having to create the framework from scratch.
Apache Ambari
Ambari provides statistics on an infrastructure level not on job level. So, you need to write a custom code to use yarn-rest API which provides you a JSON response. Based on which you can use the JSON parser and get the exact details. I have written one on python, you can refer to this link-https://dzone.com/articles/customized-alerts-for-hadoop-jobs-using-yarn-rest
http://thelearnguru.com/customized-alerts-for-hadoop-jobs-using-yarn-rest-api
Related
I have a Hadoop/Spark cluster setup via Ambari (HDP -2.6.2.0). Now that I have my cluster running, I want to feed some data into it. We have an Elasticsearch cluster on premise (version 5.6). I want to setup the ES-Hadoop Connector (https://www.elastic.co/guide/en/elasticsearch/hadoop/current/doc-sections.html) that Elastic provides so I can dump some data from Elastic to HDFS.
I grabbed the ZIP file with the JARS and followed the directions on a blog post at CERN:
https://db-blog.web.cern.ch/blog/prasanth-kothuri/2016-05-integrating-hadoop-and-elasticsearch-%E2%80%93-part-2-%E2%80%93-writing-and-querying
So far, this seems reasonable, but I have some questions:
We have SSL/TLS setup on our Elasticsearch cluster, so when I perform a query, I obviously get an error using the example on the blog. What do I need to do on my Hadoop/Spark side and on the Elastic side to make this communication work?
I read that I need to add those JARS to the Spark classpath - is there a rule of thumb as to where i should put those on my cluster? I assume on of my Spark Client nodes, but I am not sure. Also, once i put them there, is there a way to add them to the classpath so that all of my nodes / client nodes have the same classpath? Maybe something in Ambari provides that?
Basically what I am looking for is to be able to preform a query to ES from Spark that triggers a job that tells ES to push "X" amount of data to my HDFS. Based on what I can read on the Elastic site, this is how I think it should work, but I am really confused by the documentation. It's lacking and has confused both me and my Elastic team. Can someone provide some clear directions or some clarity around what I need to do to set this up?
For the project setup part of the question you can take a look at
https://github.com/zouzias/elasticsearch-spark-example
which a project template integrating elasticsearch with spark.
I have a three node nifi cluster. We just installed nifi packages on linux machines and cluster with separate zookeeper cluster. I am planning to monitor nifi performance via nagios but we saw hortonworks ambari provides fetures for management and monitoring also.
What is the best practice for nifi deployment on prod
how should we scale up
how can we monitor nifi
Should we monitor queue/process performance
Should use something like ambari
regards..
Edit-1:
#James actually I am collecting user event logs from several sources within company.
All events are first written to Kafka. Nifi consumes kafka, does simple transformations like getting a field from payload to attribute.
After transformations data is written to both elasticsearch and hdfs. Before writing to hdfs we are merging flowfiles so writing to hdfs is in batches.
I have around 50k/s event.
What I would like to do is make workflow and job metadata such as start date, end date and status available in a hive table to be consumed by a BI tool for visualization purposes. I would like to be able to monitor for example if a certain workflow fails on certain hours, success rate, ...
For this purpose I need access to the same data Hue is able to show in the job browser and Oozie dashboard. What I am looking for specifically for workflows for example is the name, submitter, status, start and end time. The reason that I want this is that in my opinion this tool lacks a general overview and good search.
The idea is that once I locate this data I will directly -or trough some processing steps- load it into Hive.
Questions that I would like to see answered:
Is this data stored in HDFS or is it scattered in local data nodes?
If it is stored in HDFS. Where can I find it? If it is stored in local data nodes, how does Hue find and show this?
Assuming I can access the data. In what format would I expect this data. Is this stored in general log files or can I expect somewhat structured data?
I am using CDH 5.8
If jobs are submitted through other ways than Oozie , my approach won't be helpful.
We have collected all the logs from the oozie server through the Oozie Java API and iterated over the coordinator information to get the required info.
You need to think, what kind of information you need to retrieve.
If you have all jobs submitted through Bundle then come from bundle to coordinator then to workflow to find out the info.
If you want to get all the coordinator info then simply call the api with the number of coordinator to bring and fetch required info.
And then we have loaded the fetched result into a hive table and there one can filter results for failed or time out coordinators & various other parameters.
You can start looking into the example given from Oozie site:-
https://oozie.apache.org/docs/3.2.0-incubating/DG_Examples.html#Java_API_Example]
If you want to track the status of your jobs scheduled in oozie, you should use oozie RESTful API or JavaAPI. I didn't work with Hue version for operation Oozie, but I guess it still uses rest api behind the scene. It provides you with all necessary information and you can create some service which would consume this data and push it into Hive table.
Another option is to access Oozie database. As you probably know Oozie keeps all the data about the scheduled jobs within some RDBMS like MqSql or Postgres. You can consume this information through some JDBC connector. An interesting way would actually be to try to link this information directly into Hive as a set of external tables though JDBCStorageHandler. Not sure if it work, but it worth to try.
I am trying to update HDP architecture so data residing in Hive tables can be accessed by REST APIs. What are the best approaches how to expose data from HDP to other services?
This is my initial idea:
I am storing data in Hive tables and I want to expose some of the information through REST API therefore I thought that using HCatalog/WebHCat would be the best solution. However, I found out that it allows only to query metadata.
What are the options that I have here?
Thank you
You can very well use WebHDFS which is basically a REST Service over Hadoop.
Please see documentation below:
https://hadoop.apache.org/docs/r1.0.4/webhdfs.html
The REST API gateway for the Apache Hadoop Ecosystem is called KNOX
I would check it before explore any other options. In other words, Do you have any reason to avoid using KNOX?
What version of HDP are you running?
The Knox component has been available for quite a while and manageable via Ambari.
Can you get an instance of HiveServer2 running in HTTP mode?
This would give you SQL access through J/ODBC drivers without requiring Hadoop config and binaries (other than those required for the drivers) on the client machines.
I have few Hive jobs and Mapreduce programs running in my cluster. I am able to check in Ambari about general resource utilization. But I want to see the resources utilized by individual applications. Is it possible through Ambari API? Can you provide some clues?
To my knowledge metrics that are provided by Ambari are for whole cluster.
But you can check MapReduce2 Job History UI, it seems like you are looking for this stuff. Check this link out, there is more detailed description there
http://hortonworks.com/blog/elephants-can-remember-mapreduce-job-history-in-hdp-2-0/