perfSONAR User Q&A and Other Discussion

[Brian Tehan <>]

Small update on this, I found an errant JSON file hanging out in /etc/perfsonar/psconfig/pscheduler.d and I had put repeat-cron in there so it was throwing errors to /var/log/maddash/psconfig-maddash-agent.log.  I’ve deleted that file and that exposed some problems with my published psconfig file.  I’ve attached my current psconfig JSON file that is published to all my nodes and is successfully getting them to generate latencybg tests.  I put a throughput test and task in there just to get a mesh to display on MaDDash, and then setup repeating tasks via the CLI with the repeat-cron flag.  I’ve adjusted the MaDDash yaml file so that the check interval and retry interval are both set to 60 seconds and bandwidth tests are running at roughly once per hour.

 

Three squares are showing results now:

 

 

However, the tests I’m running today are 50M across the entire mesh:

 

 

 

So MaDDash still isn’t picking up the tests I’m doing via CLI.  I feel like I’ve got something wrong in my psconfig file or my CLI commands.  Anyone have any ideas?

 

Thanks!

 

 

Brian Tehan

Regional Network Engineer

University of Nebraska System

EAB009F 6001 Dodge St

Omaha, NE 68182-0051

Office: (402) 554-6428

 

 

From: <> On Behalf Of Brian Tehan
Sent: Tuesday, November 16, 2021 4:00 PM
To:
Subject: [perfsonar-user] Test results not showing up in MaDDash mesh

 

Non-NU Email

---

Hello,

 

I’ve been working my way through getting bandwidth tests to run at specific times each day and have been successful in doing so with the --repeat-cron flag on CLI commands (can’t do that via PWA or psconfig).  However, I’m struggling to get CLI tests to show up in a throughput mesh in MaDDash.  I’ve tried archiving to Esmond directly via /usr/share/pscheduler/psc-archiver-esmond.json and using http, with commands like the following:

 

pscheduler task --archive @/usr/share/pscheduler/psc-archiver-esmond.json --slip PT1H --sliprand --repeat-cron "0 0-5 * * 0-6" throughput --source 162.127.1.5 --dest 162.127.1.1 --duration PT30S --bandwidth 41943040 --ip-version 4 --omit PT5S

 

 

pscheduler task --archive '{"archiver": "esmond","data":{"url":http://162.127.1.5/esmond/perfsonar/archive/,"_auth-token": "<<redacted>>"}}' throughput --source 162.127.1.5 --dest 162.127.1.1 --duration PT30S --bandwidth 62914560 --ip-version 4 --omit PT5S

 

(token pulled from /usr/share/pscheduler/ psc-archiver-esmond.json)

 

 

Result data will show up on the test results graph (either by clicking on the maddash mesh square or via the main toolkit page) for each of these tests, but the mesh will still show that it’s unable to find test data, even after the next check interval has passed.  I’ve attached my pSConfig JSON that I use to make latencybg tests.  Those tests run successfully, and the results show up in the loss mesh.  Anyone have any suggestions on what I should be doing differently?

 

 

Thanks!

 

 

Brian Tehan

Regional Network Engineer

University of Nebraska System

EAB009F 6001 Dodge St

Omaha, NE 68182-0051

Office: (402) 554-6428

 

 

Attachment: latency.json
Description: latency.json