we have 7 different servers and would like to balance the workload between these servers.
Do you have any advices how to manage the following example?
One workflow WORKFLOW A consists of a lot JOBS. These JOBS run on 2 of these servers. This workflow starts Thursday at night and takes at least 30 hours.
Another workflow WORKFLOW B starts one day later and its JOBS should balance its workload between all of the 7 servers.
In this case our approach is to use a HOSTG-object with all the 7 servers and divide the JOBS between these servers.
Now our problem:
The 2 servers, that are used by the JOBS of WORKFLOW A should not be used by the JOBS of WORKFLOW B until WORKFLOW A is finished.
WORKFLOW B gets more and more JOBS while it is running with the UC-Scrpit "MODIFY_TASK...ADD_TASK". These step should run on the 7 servers in the end but only on the 5 left servers until WORKFLOW A is running.
I hope someone could understand my use problem and has some advices for me.
Feel free to ask further questions.
Thanks for the help in advance!
I've never had to work on this type of problem before, but I know others here have. The product does not offer any automatic workload balancing features.
If I had to write a workload balancing solution myself, I might try to build a PowerShell utility to collect current usage data from a list of candidate servers to determine which one is currently under the least stress, populate a UC4 variable with that server name, and use that variable to set the HOST value for the next UC4 job.
what about the following (almost simple) attempt:
For Workflow A create a HOSTG containing agents on your both servers
For workflow B create a HOSTG containing agents on your seven servers with mode LOAD dependent
As long as the workload on the 2 Servers of Workflow A is higher than the average workload it works...
This does not meet your requirements to 100% but mostly :-)
Any other workload balancing logic must be scripted, as mentione by Pete.