perfSONAR User Q&A and Other Discussion

["Garnizov, Ivan (RRZE)" <>]

Hi Andy,

 

The problem persist….with the new configuration for MADDASH.

Please note the command stated on the web interface of the service is executed successfully on the CLI with the correct substitutions from that same page. It has a valid state, which is not UNKNOWN, but MADDASH persistently state UNKNOWN and only for BWCTL. You should be able to initiate the check from your location.

 

Please give hints on how to debug this case.

 

 

Best regards,

Ivan

 

 

From: Andrew Lake [mailto:]
Sent: Mittwoch, 28. Oktober 2015 16:35
To: Garnizov, Ivan (RRZE)
Cc:
Subject: Re: MADDASH fails with Nagios check

 

Hi Ivan,

 

Remove the lines below from gui_agent_configuration.conf and I think you will be OK. They have given a few people trouble on the list and weren’t supposed to be in the example config:

 

<ma_filter>

    ma_filter_name  bw-ignore-first-seconds

    mesh_parameter_name  omit_interval   

</ma_filter>

 

After removing those, regenerate the JSON and your maddash config. It may take some time for your dashboard to update since it will wait until the next scheduled time to run the checks. 

 

Thanks,

Andy

 

 

 

 

On Wed, Oct 28, 2015 at 11:30 AM, Garnizov, Ivan (RRZE) <> wrote:

Hi Andy,

 

Here is an instance of MADDASH, that operates well with OWAMP tests, but BWCTL is giving me a headache.

 

http://test-psmdsh-gn.geant.net/maddash-webui/details.cgi?uri=/maddash/grids/GEANT+Mesh+-+GN+IPv4+Throughput+internal+testing/psmp-gn-bw-01-ams-nl.geant.net/psmp-gn-bw-01-bud-hu.geant.net/Throughput

 

If I run the command on the CLI with the parameters  substituted I get a valid result: PS_CHECK_THROUGHPUT OK

But, as it can be seen, the result on the MADDASH states UNKNOWN

 

I have also restarted the service and have Re-scheduled the checks. I have done that once for the Amsterdam column, but also have done it for all Rows and Columns.

Nevertheless the result remains with status UNKNOWN.

 

Please give a hint on how to debug this case.

 

Best regards,

Ivan