ERROR: _seqN > m_maxSequenceNo in v3.7
GeekerHua opened this issue · 9 comments
Hello @GeekerHua , do you see such issue just for under-load
scenario? Did you try ping-pong
and throughput
?
could you provide output for pp
case too.
could you provide output for
pp
case too.
pp output nothing. It's on the picture.
there is no output for pp case on your screen because it is comment out ##
.
@GeekerHua do you have any update?
Hi,
It also happens to me on Arm with pp test when using --mps parameter.
Setup
- Amazon Linux 2 (based on RHEL) on AWS with kernel (4.14.248-189.473.amzn2.aarch64).
- Arm64 (64 CPUs)
- sockperf, version 3.7-11.gita57ffb579002
It works fine with:
./sockperf server --ip x.y.z.w --tcp --port 5001 --msg-size 400
./sockperf ping-pong --ip x.y.z.w --tcp --port 5001 --time 30 --msg-size 400 --range 100 --full-rtt
It get to error with appending --mps 1000 to the client
[ec2-user@ip-172-31-16-82 sockperf]$ ./sockperf ping-pong --ip x.y.z.w --tcp --port 5001 --time 30 --msg-size 400 --range 100 --full-rtt --mps 1000
sockperf: == version #3.7-11.gita57ffb579002 ==
sockperf[CLIENT] send on:sockperf: using recvfrom() to block on socket(s)
[ 0] IP = X.Y.Z.W PORT = 5001 # TCP
sockperf: Warmup stage (sending a few dummy messages)...
sockperf: Starting test...
sockperf: ERROR: _seqN > m_maxSequenceNo (errno=0 Success)
From very fast look, maybe the calc of _maxSequenceNo at bringup() function is related.
We will look at possible reason.
Fix #187 was included into master