Calling multiple times a function that uses Rmpi

I would like to repeatedly use a function that calls on clusters from Rmpi, but I am having issues where the first function call hangs indefinitely.

Here is a simple code example:

library(parallel)
library(snow)
library(Rmpi)

run_rmpi_test <- function(
    X, fun, iter, numer,filepath) {
  
  n.cores <- mpi.universe.size()-1
  cl <- makeMPIcluster(n.cores)

  varlist <- c(ls(envir = parent.frame()))

  parallel::clusterExport(cl, varlist = varlist,envir=environment())

  parallel::clusterSetRNGStream(cl, 20)
  
  # run parallel computation
  computed <- parallel::parLapply(cl=cl, X = X, fun=fun, iter=iter, numer=numer) 
 
  save(list=varlist,
       file=filepath,
       envir=environment())

  # stop the cluster
  snow::stopCluster(cl)
  mpi.finalize()
  
  # return result
  return(computed)
}


fun_function <- function(i,iter,numer){
    
    l1 <- rnorm(iter)
    l2 <- rnorm(numer)

    return(list(l1,l2))
}

funct1 <- run_rmpi_test(X=1:10,fun_function,iter=5,numer=8,filepath="test1.RData")

funct2 <- run_rmpi_test(X=1:10,fun_function,iter=3,numer=2,filepath="test2.RData")

Thus far, it seems that the function run_rmpi_test() hangs right after save(). Do you have any suggestions on how to fix the above? My OpenMPI version is 4.1.4. Ideally, I would like to have both function calls (funct1 and funct2) in the same script, and I should have both outputs test1.RData and test2.RData after the code is ran.