perfSONAR User Q&A and Other Discussion

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

Hi m0nk3ys are dangerous ???

 

Could you please add the –vv switch to the bwctl tests?

 

Regards,

Ivan Garnizov

 

GEANT SA1T2: pS deployments GN Operations

GEANT SA2T3: pS development team

GEANT SA3T5: eduPERT team

 

 

 

From: [mailto:] On Behalf Of m0nk3ys are dangerous ???
Sent: Dienstag, 20. September 2016 14:33
To: Michael Johnson
Cc:
Subject: Re: [perfsonar-user] PerfSonar Toolkit - Bandwidth tests on 2 sites are reporting incorrect speeds

 

Hi Michael,

 

Sorry for my late response, but I've spend most of my time in the hospital over the last weeks..

All the hosts are internal so I'm afraid I won't be able to provide access to them. But more than happy to supply you with any data you might require to analyze the issue.

 

Please see below my mesh config, which is as you can see setup with iperf3 TCP. Not running any UDP ones at the moment.

Tested with the identical commands the mesh config is using. The results are within the values I would expect during work hours.

But the graphs are not showing these values.

 

###############################################################################

###############################################################################

 

/usr/bin/bwctl -s SiteA -c SiteB -T iperf3 -B SiteB -I 3600 -R 10 –M 1360

[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd

[ 15]   0.00-1.00   sec  16.2 MBytes   136 Mbits/sec  1029   2.36 MBytes       

[ 15]   1.00-2.00   sec  10.0 MBytes  83.9 Mbits/sec  1247    609 KBytes    

[ 15]   2.00-3.00   sec  10.0 MBytes  83.9 Mbits/sec    0    631 KBytes       

[ 15]   3.00-4.00   sec  10.0 MBytes  83.9 Mbits/sec    0    702 KBytes       

[ 15]   4.00-5.00   sec  7.50 MBytes  62.9 Mbits/sec    2    382 KBytes       

[ 15]   5.00-6.00   sec  7.50 MBytes  62.9 Mbits/sec    0    423 KBytes       

[ 15]   6.00-7.00   sec  8.75 MBytes  73.4 Mbits/sec    0    575 KBytes       

[ 15]   7.00-8.00   sec  11.2 MBytes  94.4 Mbits/sec    0    810 KBytes       

[ 15]   8.00-9.00   sec  8.75 MBytes  73.4 Mbits/sec   15    448 KBytes       

[ 15]   9.00-10.00  sec  7.50 MBytes  62.9 Mbits/sec    0    484 KBytes       

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval           Transfer     Bandwidth       Retr

[ 15]   0.00-10.00  sec  97.4 MBytes  81.7 Mbits/sec  2293             sender

[ 15]   0.00-10.00  sec  87.2 MBytes  73.1 Mbits/sec                  receiver

 

/usr/bin/bwctl -s SiteB -c SiteA -T iperf3 -B SiteA -I 3600 -R 10 –M 1360

[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd

[ 15]   0.00-1.00   sec  2.80 MBytes  23.5 Mbits/sec  129    141 KBytes       

[ 15]   1.00-2.00   sec  2.75 MBytes  23.0 Mbits/sec    0    151 KBytes       

[ 15]   2.00-3.00   sec  3.13 MBytes  26.2 Mbits/sec    0    197 KBytes       

[ 15]   3.00-4.00   sec  4.55 MBytes  38.2 Mbits/sec    0    296 KBytes       

[ 15]   4.00-5.00   sec  7.05 MBytes  59.2 Mbits/sec    0    449 KBytes       

[ 15]   5.00-6.00   sec  8.34 MBytes  70.0 Mbits/sec    4    296 KBytes       

[ 15]   6.00-7.00   sec  5.70 MBytes  47.8 Mbits/sec    0    321 KBytes       

[ 15]   7.00-8.00   sec  6.66 MBytes  55.9 Mbits/sec    0    417 KBytes       

[ 15]   8.00-9.00   sec  8.11 MBytes  68.0 Mbits/sec    3    291 KBytes       

[ 15]   9.00-10.00  sec  5.61 MBytes  47.0 Mbits/sec    0    316 KBytes       

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval           Transfer     Bandwidth       Retr

[ 15]   0.00-10.00  sec  54.7 MBytes  45.9 Mbits/sec  136             sender

[ 15]   0.00-10.00  sec  53.4 MBytes  44.8 Mbits/sec                  receiver

 

 

###############################################################################

###############################################################################

 

<test_spec bwctl_1h_tcp_test>

  type              perfsonarbuoy/bwctl  # Perform a bwctl test (i.e. achievable bandwidth)

  tool              bwctl/iperf3         # Use 'iperf' to do the bandwidh test

  protocol          tcp                  # Run a TCP bandwidth test

  window_size       1360                 # TCP window size (bytes)

  interval          3600                 # Run the test every 1 hours

  duration          20                   # Perform a 20 second test

  random_start_percentage 10             # randomize start time

  force_bidirectional 1                  # do bidirectional test

</test_spec>

 

<test_spec owamp_test>

  type              perfsonarbuoy/owamp  # Perform a constant low-bandwidth OWAMP test

  packet_interval   0.1                  # Send 10 packets every second (i.e. pause 0.1 seconds between each packet)

  loss_threshold    10                   # Wait no more than 10 seconds for a response

  session_count     10800                # Refresh the test every half hour (once every 18000 packets)

  sample_count      600                  # Send results back every 60 seconds (once every 600 packets)

  packet_padding    0                    # The size of the packets (not including the IP/UDP headers)

  bucket_width      0.0001               # The granularity of the measurements

  force_bidirectional 1                  # do bidirectional test

</test_spec>

 

<test_spec ping_test>

  type              pinger               # Perform PingER tests

  test_interval     300                  # Perform a test every 5 minutes

  packet_count      10                   # Send 10 packets for each test

  packet_interval   1                    # Pause 1 second between each packet sent

  packet_size       1000                 # The size of the packets (not include the IP headers)

</test_spec>

 

<test_spec traceroute_test>

  type              traceroute           # Perform a traceroute test

  test_interval     600                  # Run the test every 5 minutes

  packet_size       40                   # The size of the packets (not include IP/UDP headers)

</test_spec>

 

<group emea_cross_connect>

  type             mesh

 

  member           a

  member           b

  member           c

  member           d

  member           e

  member           f

  member           g

  member           h

  member           i

</group>

 

<group ukbh_hub_to_emea_sites>

  type             disjoint

 

  a_member         a

  b_member         b

  b_member         c

  b_member         d

  b_member         e

  b_member         f

  b_member         g

  b_member         h

  b_member         i

</group>

 

<test>

  description       Throughput TCP - Bidirectional UK Hub and EMEA Sites

  group             ukbh_hub_to_emea_sites

  test_spec         bwctl_1h_tcp_test

</test>

 

<test>

  description       Traceroute - UK Hub to EMEA Sites

  group             ukbh_hub_to_emea_sites

  test_spec         traceroute_test

</test>

 

<test>

  description       Packetloss - EMEA Sites Cross Connect

  group             emea_cross_connect

  test_spec         owamp_test

</test>

 

<test>

  description       Ping - EMEA Sites Cross Connect

  group             emea_cross_connect

  test_spec         ping_test

</test>

 

On Mon, Aug 1, 2016 at 10:51 PM, Michael Johnson <> wrote:

Hi,

 

That's very odd, I haven't seen this before. Your CLI iperf tests may be using a different protocol (TCP vs UDP), or maybe the mesh is iperf2 and your manual tests use iperf3 (or vice versa)?

 

If your toolkit host is public, can you share a link to it? I would be curious to see whether these results are wrong in the database, or if they're being charted incorrectly for some reason.

 

Thanks,

Michael Johnson

GlobalNOC Software Engineering

Indiana University

812-856-2771

 

On Jul 30, 2016, at 8:29 AM, m0nk3ys are dangerous ??? <> wrote:

 

Hi,

 

I've got a toolkit setup with a mesh configuration to keep an eye on the network performance of our worldwide WAN.

<image.png>

Most of the test are running as they should and reporting the correct values.

Although two of my sites are reporting over the 100mbps speeds, while the local links are 100mbps.

Site1:

<image.png>

<image.png>

 

Site2

<image.png>

|<image.png>

 

As you can see they are reporting over the 100mbps for these sites which isn't correct.

These are all European sites, going to my HQ toolkit server. But the rest of the 15 sites connected to the same server are working fine.

When I run a manual iPerf test the bandwidth values are within the correct values so there must be something wrong with the config of these two toolkits.

 

If anyone has an idea where I should start looking or what to do, it would be appreciated!

 

Thanks