perfSONAR User Q&A and Other Discussion

[Mark Feit <>]

(Sorry for the late reply; I’ve been on vacation.)

 

Siri Chandana writes:

 

I have a question regarding the pscheduler command. I want to run the pscheduler with source and destination ip’s . Is it possible  to run pscheduler with a different source ip apart from the ip of my remote server. For example, I want the source IP to be the IP of my mac on which pscheduler is installed.

 

I tried to run pscheduler using ip of another desktop and a bwctl server from my remote server. The output is as follows:  Unable to determine the lead participant:  Neither the source nor destination is running pScheduler.

 

The “pscheduler” command is a program that talks to a pScheduler server through its REST API.  For architectural reasons, it doesn’t understand anything about specific tests or tools and uses the local pScheduler server (or a remote one that you specify with the --assist switch) for assistance in converting your test’s command-line arguments to the JSON format used by the API (which succeeded) and figuring out what server plays the lead role in the test (which did not).    See the Terminology section of this page for some definitions:  http://docs.perfsonar.net/pscheduler_intro.html.

 

The command you describe asks for a throughput test between 68.181.191.6 and 150.100.209.131.  If the source address you specify is bound to an OS X interface, there is no pScheduler running there.  As far as your VM is concerned, that address could be on another machine half the world away.

 

The bottom line is that you can’t specify that your test originate on any arbitrary IP address; there has to be a pScheduler server there and, if it’s a multi-participant test like throughput, at the destination end as well.  You can determine whether or not there’s a reachable pScheduler on a host using this command:

 

$ pscheduler ping HOSTNAME

 

 

Hope that helps.

 

--Mark