Oozie Workflow | Hadoop

In this video we are discussing OZ workflow. OZ workflow is nothing but sequence of actions to be carried out and represented in the form of directed a cyclic graph

Each and every action is dependent on the previous action. If the previous action works well and has its completion successfully, then its output will be going to the next action

So let us go for more detailing about this OZ workflow. So, the workflow techniques in OZ

So the OZ workflow is a sequence of actions arranged in a DAG, that is a directed acyclic graph

and these actions are dependent on the previous tasks or actions. And one action can run properly if the previous one returns a valid output to it

So, workflow actions can be different kind of tasks, like our hype task, we can have the big task

We can have the Java action, shell action, etc. So, different kinds of tasks, it can do the scheduling

In a flow of different sequential tasks, some tasks can be performed in parallel, some

tasks can be performed only in sequential. So to execute some tasks in parallel, we can use the fork option in OZ

So if I want to execute some tasks in parallel, then we'll be using the fork option so

that these two tasks can execute in parallel and later we can go for the join option so that again

the tasks will get the sequence will get merged so the join option is used to merge two parallel tasks into one So let us discuss this one with one proper diagram Just find this one

This is the different components of OZ workflow. Just see. Here we are having this start

Then we are having this MR job that is a map reduce job. We are having this peak job. So two jobs

are there, then OZ will do the fork. So, as a fork, so these two jobs will be executed in parallel, then join will be there

then it will have the proper completion and end. Nodes in OZ workflow

So there are mainly three control nodes in the OZ workflow and they are start, end, and

the kill nodes. You can find here we're having the start, we're having the end and the kill node

So, there are mainly three control flows in our OZ control flow nodes in our OZ workflow

So, a workflow can start from the start node and ends at the end node, but when an error

occurs, it moves to the kill node to kill the process because the process has got some interruption

So here you can find here we're having this start. So now this map-produce word count is the process to be executed

So if it faces any kind of error, then the kill node. will be reached, otherwise it will go to the end if the process has got okay with the operation

and processing. So, this is the basic overview of this OZ workflow and we have explained this one with proper

diagram and explanation. Thanks for watching this video