Cluster with Python and Dispy: cannot connect node

Question

I'm trying to create a raspberry-pi cluster using dispy for a school project.

Just to try it, I'm using only one RPi as a node and my pc as master that request jobs.

Below the devices and their ip(s):

master (pc)  -->  192.168.1.106
node (rpi)   -->  192.168.1.105

Node configuration:

I execute dispynode in my node with:

sudo dispynode.py -i 192.168.1.105

Where:

• sudo to make sure there won't problem about permission
• -i 192.168.1.105 to tell the node which address set up for communication (documentation)

The console output:

2015...12:21 - dispynode - dispynode version 4.5
2015...12:21 - dispynode - serving 1 cpus at 192.168.1.105:51348

I said: "Seems good!"

Master configuration:

I've written a small python script to test the cluster:

import dispy

cluster = dispy.JobCluster(
    "/bin/ls",
    nodes=["192.168.1.105"],
    ip_addr="192.168.1.106"
)

cluster.stats()

Where:

• cluster will be a new JobCluster that have the following specs
• "/bin/ls" is the path of an executable, the job for the node
• nodes contains the addresses of nodes that will receive the job (in this case just one, my rpi)
• ip_addr is the address used by master for communication
• cluster.stats() print out the stats about the cluster created

I expect just one node connected, but when I run this script I got this stats:

2015...12:22 - dispy - Storing fault recovery information in "_dispy_2015..."

Node |  CPUs |    Jobs |    Sec/Job | Node Time Sec
-----------------------------------------------------------------------------

Total job time: 0.000 sec

As you can see there isn't rpi node connected to master...

Also the fault recovery file created is empty.

How can I connect node to master?

I'm using in the wrong way dispynode or master script?

NOTES & USEFUL:

• Obiuvsly I've installed dispy in both devices.
• I can reach the node with ping from master and vice versa.
• ufw is disabled.
• Documentation here.

Answer 1

Did you try submitting a job? Dispy might only initiate a connection when there is an actual job.