perfSONAR User Q&A and Other Discussion

[Brian Candler <>]

On 18/12/2014 22:42, Michael Johnson wrote:

Here's an example:
http://stats.es.net/serviceTest/graphWidget.cgi?url="http://anl-pt1.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.22&source=198.124.252.117&url=http://anl-owamp.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.22&source=198.124.252.97&url=http://anl-pt1.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.34&source=198.124.252.117&url=http://anl-owamp.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.34&source=198.124.252.97

There are some limitations to this approach -- it doesn't like it more than one host has data for a given type. For example, if more than one host has Throughput data it will not handle it nicely, as far as I recall.

Thank you, I think I'm slowly getting there :-)

That long URL contains four (url/dest/source) tuples, but the graph shows only three data pairs: (throughput/reverse throughput), (latency/reverse latency), (loss/reverse loss).

Now, if I trim the URL to contain only one pair at a time, I get these:

http://stats.es.net/serviceTest/graphWidget.cgi?url="http://anl-pt1.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.22&source=198.124.252.117
=> shows only throughput

http://stats.es.net/serviceTest/graphWidget.cgi?url="http://anl-owamp.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.22&source=198.124.252.97
=> "ERROR: No data to plot for the hosts and time range selected."

http://stats.es.net/serviceTest/graphWidget.cgi?url="http://anl-pt1.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.34&source=198.124.252.117
=> "ERROR: No data to plot for the hosts and time range selected."

http://stats.es.net/serviceTest/graphWidget.cgi?url="http://anl-owamp.es.net:8085/esmond/perfsonar/archive&dest=198.129.254.34&source=198.124.252.97
=> shows only latency and loss

So I guess what's happening is that the Dashboard knows about two endpoints at each node (one for throughput and one for latency), and graphs all the combinations including the useless ones like throughput A <-> latency B and vice versa.

Regards,

Brian.