[sles 12 sp5] iperf3 error: iperf3: error - unable to send control message - port may not be available, the other side may have stopped running, etc.: Broken pipe

[LiliDeng]

Build iperf3 from master branch.
Run iperf3 -s -D on machine A.
Run iperf3 -t 30 -c 10.0.0.5 -P 128 on machine B.

It is related with the parameter P
When the parallel value is 50, it works, when it is 55, it starts to show this failure.
Expected to see no error code but see below errors.
A

[252]   0.00-30.01  sec   158 MBytes  44.2 Mbits/sec                  receiver
[254]   0.00-30.01  sec   157 MBytes  43.8 Mbits/sec                  receiver
[256]   0.00-30.01  sec   173 MBytes  48.5 Mbits/sec                  receiver
[258]   0.00-30.01  sec   202 MBytes  56.4 Mbits/sec                  receiver
[260]   0.00-30.01  sec   140 MBytes  39.1 Mbits/sec                  receiver
[SUM]   0.00-30.01  sec  20.1 GBytes  5.76 Gbits/sec                  receiver
warning: JSON data length overflow
iperf3: error - unable to receive results: Bad file descriptor
-----------------------------------------------------------
Server listening on 5201 (test #2)
-----------------------------------------------------------

B

[255]  28.00-29.00  sec  5.62 MBytes  47.2 Mbits/sec    8   36.9 KBytes
[257]  28.00-29.00  sec  5.62 MBytes  47.2 Mbits/sec    8   21.8 KBytes
[259]  28.00-29.00  sec  5.50 MBytes  46.1 Mbits/sec    5   23.2 KBytes
[SUM]  28.00-29.00  sec   684 MBytes  5.74 Gbits/sec  869
warning: Failed to read JSON data size
iperf3: error - unable to send control message - port may not be available, the other side may have stopped running, etc.: Broken pipe

[davidBar-On]

At least A is a known issue that is supposed to be fixed by PR #1779 (that was not merged yet into the master branch). Try using the code from this PR to see if it resolves both issues.

[LiliDeng]

@davidBar-On with above change, I encounter a different crash
iperf client output segment

[SUM]  28.00-29.00  sec   682 MBytes  5.73 Gbits/sec    0
warning: Failed to read JSON data size
Segmentation fault (core dumped)

dmesg segment

[ 1251.041615] iperf3[4388]: segfault at 7f4583cff9d0 ip 00007f4584fb52f1 sp 00007fff8c631830 error 4 in libpthread-2.22.so[7f4584fa6000+18000]
[ 1344.265343] iperf3[4883]: segfault at 7fef7bfe79d0 ip 00007fef7d29d2f1 sp 00007ffe6dfdf320 error 4 in libpthread-2.22.so[7fef7d28e000+18000]

[davidBar-On]

The segmentation fault is a different issue related to duplicate canceling of a thread. Fortunately, it was already fixed by PR #1761 (see this comment by @answer9030 that explains the issue).

@LiliDeng, I have rebased PR #1779 so it now includes PR #1761 fix. Can you retest to make sure the segmentation fault is resolved?

[LiliDeng]

@davidBar-On the issue has been resolved with the latest fix.

[davidBar-On]

@LiliDeng, great. Note that you can close the issue and go back to use iperf3 master branch, as the change was merged into iperf3.

[LiliDeng]

@davidBar-On thanks.