Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] Node not reporting to Maddash

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] Node not reporting to Maddash


Chronological Thread 
  • From: Mark Feit <>
  • To: Raul Lopes <>, "" <>
  • Subject: Re: [perfsonar-user] Node not reporting to Maddash
  • Date: Tue, 6 Jul 2021 13:25:51 +0000

Raul Lopes writes:

 

 I have a couple of nodes in a 20 nodes mesh that run the tests, but do not report to Maddash.
 
  The logs are full of lines like
 
  Jul  5 19:19:25 ps03-em1 journal: runner INFO     489461: Posted result to
https://ps03-lat.rl.ac.uk/pscheduler/tasks/0203bb86-10af-4d77-a607-a51de117f9cf/runs/defa4e86-a17
  7-41a5-aebc-e1256683bd07
 
  I can use pscheduler on the command line and get latency and bandwidth run with success. However, they do not send anything to Maddash. There is nothing going from that host to the Maddash server.

 

The usual first step would be to grab a few of the run URLs for the posted results and ask pScheduler how it thinks the archvings went:

 

pscheduler result --archivings RUN-URL

 

At the bottom, you’ll see something that looks like this that should provide diagnostic information like this contrived example:

 

Archivings:

 

  To failer, Unfinished

    2021-07-06T13:05:51Z      Possibly-permanent archiver error: Program failed to start after 3 tries: Program exited 1: Forced to fail badly.

 

 

However also shows

 

2021-07-06 09:58:14,944 [INFO] /usr/lib/python3.6/site-packages/pycassa/logging/pool_logger.py: Connection 139755773060824 (localhost:9160) in pool 139755772886656 failed:
TSocket read 0 bytes

 

esmond/django.log has an error

 

2021-07-06 10:50:25,567 [ERROR] /usr/lib/esmond/lib/python3.6/site-packages/django/core/handlers/exception.py: Internal Server Error: /esmond/perfsonar/archive/4dbdcd5297fa41eaa01b5474ff5b186f/

Traceback (most recent call last):

  File "/usr/lib64/python3.6/site-packages/thrift/transport/TSocket.py", line 143, in read

    message='TSocket read 0 bytes')

thrift.transport.TTransport.TTransportException: TSocket read 0 bytes

Has anyone seen a situation like that?

 

I haven’t seen that specifically, but zero-byte reads are often a sign that the server closed the connection while the client was expecting it to send something.  Andy might have something more specific since he’s our Esmond maven.  He’s just back from vacation and should get to this while digging out from under a week’s worth of mail.

 

--Mark

 




Archive powered by MHonArc 2.6.24.

Top of Page