Gurobi Error 10030: Overage for too long, X active sessions and over the baseline for Y minutes
Hi,
I am solving a model repeatedly with different parameters using Gurobi via Julia. The program is running parallel across 100 workers and 1 node, where each worker takes the same model setup in terms of variables and constraints but different parameters in the objective function. I also set the environment for Gurobi by
using JuMP
import Gurobi
const ENV = Gurobi.Env()
ufl = Model(() -> Gurobi.Optimizer(ENV))
The program is running on a server using an academic WLS license with a limit of 2 active sessions and 100 distributed workers.
However, I keep getting the error of "Gurobi Error 10030: Overage for too long, X active sessions and over the baseline for Y minutes". Also, I found that the error eventually goes aways if I start the job a few times, and wait sufficiently long before starting a new job.
Is there any way to solve this issue?
Many thanks in advance!
Best,
Chenying
-
Hi Chenying,
Are you stopping the julia kernel after some time? It may that the Gurobi environments are not being disposed properly (freed) and therefore the tokens are held unnecessarily.
Also if you are trying to run 100 jobs in parallel, you need more than 2 tokens.
Cheers,
David0 -
Hi David,
Thanks for your reply and sorry for my late response. May I know how to stop julia kernel when running on cluster? I did use rmprocs(workers()) if that's what you mean.
Also, how should I obtain more than 2 tokens? Do you mean multiple licenses?
Best,
Chenying
0 -
I've the same problem.
Error 10030: Too many sessions, 1 active sessions for a baseline of 1
when I launch my python API customized in my second server ubuntu. not working. but if i stop python api in first server then second server is ok. I wait for 10 minutes between switch.
0
サインインしてコメントを残してください。
コメント
3件のコメント