[netperf-talk] TCP_RR not returning results

[Rick Jones]

shwetabh.goel at calsoftinc.com wrote:
>>>Shwetabh wrote:
>>>
>>>>
>>>>
>>>>Rick Jones wrote:
>>>>
>>>>
>>>>>When you say did not stop - you mean it was still generating traffic,
>>>>>or simply didn't emit a result?
>>>>>
>>>>>
>>>>>>After this, no output was given. I hope this helps with the
>>>>>>diagnosis.
>>>>>>Please let me know if any more information is needed.
>>>>>
>>>>>
>>>>>A system call trace after the elapsed time could help
>>>>>
>>>>>
>>>>>thanks,
>>>>>
>>>>>rick jones
>>>>>
>>>>
>>>>Hi,
>>>>
>>>>I started the TCP_RR test for 12 hours. After 12 hours, the test still
>>>>did not stop.
>>>>Then I ran strace for the process. Here's the output:
>>>>
>>>>[root at localhost ~]# strace -p 5709
>>>>Process 5709 attached - interrupt to quit
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>recvfrom(4, "", 1, 0, NULL, NULL)       = 0
>>>>
>>>>I hope this information helps. Please let me know if some other info is
>>>>needed.
>>>
>>>It should - this on netperf or netserver?
>>>
>>>Interesting - looks like the code is posting a single-byte recv as
>>>expected, getting a zero byte return, and ignoring it.  That keeps it in
>>>the loop trying to complete a transaction.
>>>
>>>rick
>>>
>>
>>This was on netperf. I didn't collect netserver's logs. I will get them if
>>they are needed.
>>
>>Regards,
>>Shwetabh
>>
> 
> 
> strace from netserver:
> 
> [root at sunc1 ~]# strace -p 14662
> Process 14662 attached - interrupt to quit
> accept(4,

That would be the remote parent netserver.  Assuming there is no other 
netserver process on the system it implies that the remote side has 
terminated.

rick