Simple clustering for high availability and loa... Skip navigation

Simple clustering for high availability and load balancing

score 45
You have not voted. Delivered

I would like to see an agent clustering option.  It does not need to be overly slick and can run on top of the existing secure agent implementation.  I am just looking to be able to define a simple cluster that would support high availability and automate the manual load balancing that I am forced to implement.

 

Step 1: Define a cluster secure agent: CLUS01

Step 2: Assign servers to the cluster secure agent: SERV01, SERV02, SERV03

Step 3: Assign connections to run on the cluster agent

Step 4: Informatica Cloud Runtime processing

 

     4.a If a task is set to run on CLUS01, assign the tasks in order SERV01, then 02, etc.

     4.b If the server (e.g. SERV01) is not found, remove it from the pool and send an email notification

 

That's it.

 

Now, if I can get greedy, if it could keep track of load on each server and factot that into the algorithm that would be a bonus.  E.g. SERV02 might have 2 very long running tasks on it, yet the tasks that were sent to SERV03 already finished.  The algorithm would skip SERV02 and assign the next task to SERV03.

Comments

Vote history