Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] missing maddash host data

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] missing maddash host data


Chronological Thread 
  • From: "Andrew Lake" <>
  • To: "SCHAER Frederic" <>
  • Cc: , "Shawn McKee" <>
  • Subject: Re: [perfsonar-user] missing maddash host data
  • Date: Mon, 23 Nov 2015 08:59:33 -0800 (PST)

Hi,

In both cases it is trying to get data from the Central MA run by OSG at http://psds.grid.iu.edu/esmond/perfsonar/archive/. That is populated by periodically polling the MAs running locally on each of your machines and copying the results to psds.grid.iu.edu. It might be worth checking that the software pulling the results is able to retrieve them. That is something developed within OSG so I have copied Shawn McKee who can help point you in the right direction. Since your MA is only open to certain locations it would seem, it’s possible its just a firewall issue or similar and the software pulling the results is just getting blocked. That’s just a guess, but may be a good place to start if you have already confirmed you have the data locally.

Hope that helps,
Andy






On Mon, Nov 23, 2015 at 9:16 AM, SCHAER Frederic <> wrote:

Hi,

 

My perfonsar hosts appear in maddash with missing data for all lines/columns. All squares are half orange…

The issue can be seen on this dashboard (GRIF perfsonar02 host) : http://psmad.grid.iu.edu/maddash-webui/index.cgi?dashboard=Dual-Stack%20Mesh%20Config

 

The errors displayed is :

Unable to find any tests with data in the given time range where source is ccperfsonar1.in2p3.fr and destination is perfsonar02.datagrid.cea.fr”

 

Thing is.. I don’t see many things in the error logs.

It is as if tests were running, but data did not appear in the results/measurement archives in the right place. Or at least in maddash.

Because when I look here : http://ccperfsonar1.in2p3.fr/toolkit/ , I see bidirectional results.

When I look here : https://perfsonar02.datagrid.cea.fr/toolkit/, I also see bidirectional results

 

Side question would be : where does maddash look for those test results ? Locally on the perfsonar host ? using http (filtered out except for the perfsonar maddash network, i.e 129.79.53.0/24 and 192.41.231.110) or https ?

 

The only thing that I see repeated many times a day in the logs are peer errors :  “Peer cancelled test before expected”

 

Example :

Nov 23 11:27:09 perfsonar02 bwctld[39868]: FILE=sapi.c, LINE=353, Connection to (perfsonar02.datagrid.cea.fr:16104) from (perfsonar-bw.cern.ch:45395)

Nov 23 11:27:09 perfsonar02 bwctld[39868]: FILE=sapi.c, LINE=510, ControlSession([perfsonar02.datagrid.cea.fr]:16104) accepted from userid(nil):([perfsonar-bw.cern.ch]:45395)

Nov 23 11:27:15 perfsonar02 bwtraceroute: Local server has exited

Nov 23 11:27:15 perfsonar02 bwtraceroute: 1090 seconds until next testing period

Nov 23 11:27:15 perfsonar02 bwtraceroute: Spawning endpoint to handle remote side

Nov 23 11:27:15 perfsonar02 bwtraceroute: ParisTracerouteAvailable(): Unable to verify that 'paris-traceroute' is working. It may not be installed. exit status: 1: output: Cannot create a raw socket (are you root?): Operation not permitted#012E: Cannot create libparistraceroute loopUnknown error: Invalid argument

Nov 23 11:27:15 perfsonar02 bwtraceroute: Couldn't initialize tool "paris-traceroute". Disabling it.

Nov 23 11:27:15 perfsonar02 bwtraceroute: Connection to (unixsock:unnamed) from (unixsock:unnamed)

Nov 23 11:27:15 perfsonar02 bwtraceroute: ControlSession([unixsock]:unnamed) accepted from userid(nil):([unixsock]:unnamed)

Nov 23 11:27:15 perfsonar02 bwctld[39900]: FILE=sapi.c, LINE=353, Connection to (perfsonar02.datagrid.cea.fr:4823) from (perfsonar02.datagrid.cea.fr:58635)

Nov 23 11:27:15 perfsonar02 bwctld[39900]: FILE=sapi.c, LINE=510, ControlSession([perfsonar02.datagrid.cea.fr]:4823) accepted from userid(nil):([perfsonar02.datagrid.cea.fr]:58635)

Nov 23 11:27:15 perfsonar02 bwtraceroute: Using perfsonar02.datagrid.cea.fr as the address for remote sender

Nov 23 11:27:15 perfsonar02 bwtraceroute: Using perfsonar-bw.tier2.hep.manchester.ac.uk as the address for remote receiver

Nov 23 11:27:15 perfsonar02 bwtraceroute: Available in-common: traceroute tracepath

Nov 23 11:27:15 perfsonar02 bwtraceroute: Using tool: tracepath

Nov 23 11:27:15 perfsonar02 bwtraceroute: Server 'perfsonar-bw.tier2.hep.manchester.ac.uk' accepted test request at time 1448274446.807215

Nov 23 11:27:15 perfsonar02 bwtraceroute: Client 'perfsonar02.datagrid.cea.fr' accepted test request at time 1448274446.807215

Nov 23 11:27:15 perfsonar02 bwtraceroute: 21 seconds until test results available

Nov 23 11:27:16 perfsonar02 bwctld[39908]: FILE=sapi.c, LINE=353, Connection to (perfsonar02.datagrid.cea.fr:4823) from (perfsonar2.mi.infn.it:36227)

Nov 23 11:27:16 perfsonar02 bwctld[39908]: FILE=sapi.c, LINE=510, ControlSession([perfsonar02.datagrid.cea.fr]:4823) accepted from userid(nil):([perfsonar2.mi.infn.it]:36227)

Nov 23 11:27:17 perfsonar02 bwtraceroute: 1116 seconds until next testing period

Nov 23 11:27:21 perfsonar02 bwctld[39868]: FILE=endpoint.c, LINE=1310, PeerAgent: Peer cancelled test before expected

 

I am assuming there is some port filtered out somewere or some misconfiguration, but I’d appreciate some help hunting down the issue as my logs do not help much…

 

Thanks && regards

Frederic

 





Archive powered by MHonArc 2.6.16.

Top of Page