perfSONAR User Q&A and Other Discussion

[Shawn McKee <>]

Hi Winnie,

Yes, since a perfSONAR BW test involves both ends (and the network in the middle) there can be many things that can cause a test to fail.   For WLCG/OSG, we have the check_mk monitoring setup to test basic services on each host to help us understand where problems might be if measurements are not working.

I think your site should be good now that we have fixed the firewall port ranges.   We will have to watch closely since the hosts only have 2GB of memory and testing at this scale has caused issues in hosts <4GB.   

Thanks,

Shawn

On Wed, May 20, 2015 at 10:08 AM, Winnie Lacesso <> wrote:

Forgot to add: another point of confusion of "is the site's Perfsonar BW box
borken or not?" was that the monitoring pages all show
lcgnetmon02.phy.bris.ac.uk as Fine.

https://maddash.aglt2.org/WLCGperfSONAR/check_mk/index.py?start_url=%2FWLCGperfSONAR%2Fcheck_mk%2Fview.py%3Fview_name%3Dhosts%26host%3Dbris.ac.uk
https://pfomd.grid.iu.edu/WLCGperfSONAR/check_mk/index.py?start_url=%2FWLCGperfSONAR%2Fcheck_mk%2Fview.py%3Fview_name%3Dhosts%26host%3Dbris.ac.uk
https://perfsonar-itb.grid.iu.edu/WLCGperfSONAR/check_mk/index.py?start_url=%2FWLCGperfSONAR%2Fcheck_mk%2Fview.py%3Fview_name%3Dhosts%26host%3Dbris.ac.uk


State Service         Status detail         Age                 Checked Icons Perf-O-Meter
OK    perfSONAR BWCTL TCP OK - 0.136 second 2015-05-15 21:53:42 6 min         135.507 ms
      Bandwidth Test  response time on
      Controller      137.222.171.39 port 4823

It was only noticing that graphs to sister sites in nearby UK stopped
showing stuff around 27 April that made one wonder "is it borken? Partly?"

But then the graphs are 100% fine to RightNow to a few sites. So not all
borken...