Page 1 of 1

Cluster Error In High Performance Computing

Posted: 16 Jan 2022, 15:13
by punyabeet
Dear Prof Hess

I get the following error while estimating a mixed ICLV model in the HPC with 250 mlhs draws in the beginning. I am running R from Conda environment.

Error in socketAccept(socket = socket, blocking = TRUE, open = "a+b", :
all connections are in use
Calls: apollo_estimate ... apollo_makeCluster -> <Anonymous> -> makePSOCKcluster -> socketAccept
Execution halted

I have attached the job submission file for reference. There are in total 5 nodes and each node is allocated 24 cores.

Please help me regarding this. I am stuck in it.

Regards
Thank You

Attachment of Job submission File

Re: Cluster Error In High Performance Computing

Posted: 16 Jan 2022, 15:36
by stephanehess
Hi

are you sure this is an issue related to Apollo as opposed to your HPC environment.

Could you show us the model code too, and can you confirm that it also doesn't run with fewer cores?

Stephane

Re: Cluster Error In High Performance Computing

Posted: 08 Feb 2022, 06:31
by punyabeet
Apologies for the late reply. It is solved. It was an HPC error.

Thanks for the reply.