Parallel processing cluster not used when running dredge on a MCMCglmm model with the parallel package

I am using MCMCglmm to run a PGLMM model. Since the aim is not to make predictions, I'm using dredge (from MuMIn) to calculate model-weighted parameter values and confidence intervals. Due to the large number of fixed effects, I thought it would be a good idea to implement a parallel processing workflow so that it doesn't take a day or two. I am using fully-updated Win10 and R Studio 2021.09.2.

The model and model selection code for reference:

 fullmod<-upd.MCMCglmm(occ~den*month+year+diet+dssi+eu_trend+hssi*migration+mass+latitude,
                random=~phylo+spp,family="gaussian",ginverse=list(phylo=inv.mat),prior=prior,
               data=merged_full,nitt=nitt,burnin=burnin,thin=thin, verbose=F)
all_mods <- dredge(fullmod, trace=2)

This model works perfectly, and using dredge without a cluser also runs with no issues, so I know the issue lies solely in my implementation of parallel processing. According to the relevant documentation, pdredge is depreciated and so dredge can be used directly; I tried pdredge regardless based on the code in this question and I get the exact same error.

Clustering code:

    library(parallel)
    library(snow)
    
    # Detect number of cores and create cluster (leave one out to not overwhelm pc)
    nCores <- detectCores() - 1
    cl <- makeCluster(nCores, type = "SOCK")
    
    
    # Export all objects to be used to all the cores in the cluster
    clusterExport(cl, list("merged_full","inv.mat","prior","nitt","burnin","thin","fullmod",
"upd.MCMCglmm"))
    
    # Load packages to be used
    clusterEvalQ(cl,library(MuMIn,logical.return =T))
    clusterEvalQ(cl,library(MCMCglmm,logical.return =T))
    
    fullmod<-upd.MCMCglmm(occ~den*month+year+diet+dssi+eu_trend+hssi*migration+mass+latitude,random=~phylo+spp,
                                                 family="gaussian",ginverse=list(phylo=inv.mat),prior=prior,
                                                 data=merged_full,nitt=nitt,burnin=burnin,thin=thin, verbose=F)
    all_mods <- MuMIn::dredge(fullmod, trace=2, cluster=cl)

And the output:

> all_mods <- MuMIn::dredge(fullmod, trace=2, cluster=cl)
Not using cluster.
Fixed term is "(Intercept)"

And of course, the code runs fine, just without clustering. Checking the Task Manager, I can see that the cluster is up and running, it's just that the other cores are not being used.

I did not provide a reproducible example as the model runs with no issues , but I can do so if required. Anyone knows what could be causing this?

I think this is a bug in the MuMIn package, and the fix is as simple as just renaming your cluster to anything else but cl. I spent an extremely frustrated day trying to figure out this issue and wound up digging around in the code for dredge.R on Github and found this:

When I ran isTRUE(evaluate) && inherits(cl, "cluster") on my cl object, it kept returning TRUE! So obviously my cluster was successfully created and passing the tests in dredge. However, I thought that perhaps cl <- match.call() (line 31) was potentially messing up the dredge function using my own cl object. Renaming my cluster to clu fixed the problem.