iperf test server needs to be changed

This is a test result example for a working server available:

Running 'iperf -y c -c 10.246.72.3 -t 300 -P 5 '
Linux 4.4.0-130-generic (s2lp3) 07/09/2018 _s390x_ (4 CPU)

04:48:05 AM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
04:49:05 AM all 0.05 0.00 8.44 0.00 0.22 8.46 0.04 0.00 0.00 82.79
04:49:05 AM 0 0.07 0.00 9.45 0.00 0.22 7.87 0.03 0.00 0.00 82.37
04:49:05 AM 1 0.03 0.00 8.65 0.00 0.17 4.05 0.02 0.00 0.00 87.08
04:49:05 AM 2 0.05 0.00 5.32 0.00 0.32 19.46 0.07 0.00 0.00 74.79
04:49:05 AM 3 0.05 0.00 10.36 0.00 0.22 2.43 0.03 0.00 0.00 86.90

04:49:05 AM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
04:50:05 AM all 0.03 0.00 7.21 0.00 0.22 8.21 0.03 0.00 0.00 84.30
04:50:05 AM 0 0.03 0.00 6.78 0.00 0.18 7.20 0.03 0.00 0.00 85.76
04:50:05 AM 1 0.03 0.00 7.70 0.00 0.17 2.92 0.03 0.00 0.00 89.15
04:50:05 AM 2 0.05 0.00 3.75 0.00 0.32 20.83 0.05 0.00 0.00 75.00
04:50:05 AM 3 0.02 0.00 10.62 0.00 0.20 1.90 0.02 0.00 0.00 87.25

04:50:05 AM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
04:51:05 AM all 0.05 0.00 6.73 0.00 0.24 9.26 0.03 0.00 0.00 83.69
04:51:05 AM 0 0.05 0.00 6.48 0.00 0.20 6.90 0.02 0.00 0.00 86.35
04:51:05 AM 1 0.02 0.00 6.05 0.00 0.15 3.45 0.02 0.00 0.00 90.32
04:51:05 AM 2 0.05 0.00 4.68 0.00 0.35 24.15 0.05 0.00 0.00 70.71
04:51:05 AM 3 0.05 0.00 9.72 0.00 0.25 2.53 0.02 0.00 0.00 87.43

04:51:05 AM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
04:52:05 AM all 0.03 0.00 7.77 0.00 0.22 8.02 0.04 0.00 0.00 83.92
04:52:05 AM 0 0.05 0.00 8.98 0.00 0.20 6.03 0.05 0.00 0.00 84.69
04:52:05 AM 1 0.03 0.00 7.90 0.00 0.17 4.03 0.02 0.00 0.00 87.85
04:52:05 AM 2 0.05 0.00 5.00 0.00 0.32 19.80 0.07 0.00 0.00 74.77
04:52:05 AM 3 0.02 0.00 9.15 0.00 0.20 2.23 0.02 0.00 0.00 88.38

04:52:05 AM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
04:53:05 AM all 0.07 0.00 7.84 0.00 0.24 8.65 0.05 0.00 0.00 83.15
04:53:05 AM 0 0.07 0.00 10.25 0.00 0.20 5.35 0.03 0.00 0.00 84.10
04:53:05 AM 1 0.07 0.00 6.85 0.00 0.17 4.05 0.03 0.00 0.00 88.84
04:53:05 AM 2 0.08 0.00 4.13 0.00 0.37 23.11 0.12 0.00 0.00 72.19
04:53:05 AM 3 0.05 0.00 10.14 0.00 0.22 2.10 0.02 0.00 0.00 87.48

Average: CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
Average: a...

The iperf test issue fixed with http://kernel.ubuntu.com/git/ubuntu/autotest-client-tests.git/commit/?id=481edaa9bb6a0e2964e99cd824b9aa26cb8055f7

But this server issue is still valid.

do you have anymore info on this? I can run the tests manually and find out if not.

Hi Sean,

the whole story for this test is that we used to have the iperf server installed on one of our internal server, with IP 10.245.80.28

Now that the system has gone, I have setup the new iperf server on the new MaaS server (10.246.72.3) and it works well.

The major problem is that this server only works for bare-metal system, but not for our cloud instances (VLAN issue)

Brad agreed to skip them on clouds for now.

So I think the next step is:
1. Fix the ip address in ACT
2. Blacklist this test for clouds
3. Find a solution so it can be tested on cloud as well.

I have forward the discussion and unassigned myself, please feel free to take this.
Thanks

The Iperf test has been passing on all metals in the past SRU cycles, yet the data in the test is rather stale. The iperf server is still set for 10.245.80.28, Which is from years ago.

We should probably find a new system (Not the MAAS server) to host iperf server as this will generate random traffic. The system should be bare metal and not a VM or Container.

balboa probably works moving forward and I'll look into getting that done. The test needs to be updated to properly exit, in the event the host is unreachable. I can also own that.

And we can probably elaborate on the iperf tests as well since those have sure evolved a bit since this was last run.