perfSONAR User Q&A and Other Discussion

[Aaron Brown <>]

Hey Roderick,

That was part of a bugfix in the bwctl 1.5.4 release. The "-O 5” flag wasn’t 
being counted against the total test time even though the test is actually 
running (its results are just being ignored). Sorry for that screwing up the 
tests.

Cheers,
Aaron

> On Apr 9, 2015, at 5:41 AM, Roderick Mooi 
> <>
>  wrote:
> 
> Dear Soh,
> 
> Thanks for the assistance - that explains it. Our bwctl commands do use -O 
> 5 as they are generated from the mesh. We liked it that way but didn’t 
> realise that it actually adds 5s to the test time :-/
> 
> Considering that increasing the limit to 35s or reducing the -t to 25 
> worked this seems to be it.
> 
> Ivan, do you still require the other information from me?
> 
> Something that bothers me a bit is that this used to work fine until 27 
> March and then stopped working (without any intervention from our side). 
> Did something change in iperf or bwctl around that time related to the 
> handling of -O and -t? (our nodes do perform automatic updates)
> 
> Thanks!
> 
>> On 09 Apr 2015, at 11:24 AM, 
>> 
>>  wrote:
>> 
>> Dear all,
>> 
>> Possibly it is caused by "-O 5" option passed to bwctl.
>> Bwctl will pass "-O 5" to iperf3.
>> 
>> This option is used to ignore data during initial 5 seconds
>> to reduce the effect of TCP slow start.
>> Then additional 5 seconds are needed to the time duration specified by 
>> "-t".
>> 
>> You can check whether "-O" is specified or not by ps command.
>> I have same problem, but the "-O" option appears only for the test
>> generated from mesh.
>> 
>> If 3.4.2 is cleanly installed, the bwctl.limits file permits 60 seconds.
>> So some pair will meet this problem, and other pare will not.
>> 
>> regards,
>> Soh Y. Suzuki
>> 
>> On 2015/04/09, at 17:34, "Garnizov, Ivan (RRZE)" 
>> <>
>>  wrote:
>> 
>>> Dear Roderick,
>>> 
>>> Unfortunately, I am unable to reproduce your problem.
>>> Please first specify versions of the tools you use: toolkit and iperf!
>>> 
>>> Then please explain, where do you set the bwctld.limits file with 35s. 
>>> Both sides or just destination...or source, so that we do not have to 
>>> guess.
>>> 
>>> It would be helpful for us to make the same test from our systems, if you 
>>> are willing to provide us with the information about the hosts themselves 
>>> as well.
>>> 
>>> The subject of the message is also signifying that something was working 
>>> and then stopped, but we have no idea, based on what changes the error 
>>> was introduced.
>>> 
>>> Best regards,
>>> Ivan
>>> 
>>> -----Original Message-----
>>> From: 
>>> 
>>>  
>>> [mailto:]
>>>  On Behalf Of Roderick Mooi
>>> Sent: Thursday, April 09, 2015 9:19 AM
>>> To: perfsonar-user
>>> Cc: Kevin Draai
>>> Subject: [perfsonar-user] bwctl scheduled 30 second tests stopped working
>>> 
>>> Greetings
>>> 
>>> Recently our scheduled tests stopped working. We noticed lines like these 
>>> in owamp_bwctl.log:
>>> 
>>> bwctld[19119]: FILE=endpoint.c, LINE=1310, PeerAgent: Peer cancelled test 
>>> before expected
>>> 
>>> Further investigation: Our tests are scheduled for 30s each and although 
>>> bwctld.limits has duration=30, testing manually revealed that the tests 
>>> are being rejected.
>>> 
>>> Changing either the limits file to duration=35 or our test duration to 
>>> 25s resolved the problem. 
>>> 
>>> Any explanations?
>>> 
>>> Thanks,
>>> 
>>> Roderick
>>> --
>>> This message is subject to the CSIR's copyright terms and conditions, 
>>> e-mail legal notice, and implemented Open Document Format (ODF) standard. 
>>> The full disclaimer details can be found at 
>>> http://www.csir.co.za/disclaimer.html.
>>> 
>>> This message has been scanned for viruses and dangerous content by 
>>> MailScanner, and is believed to be clean.
>>> 
>>> Please consider the environment before printing this email.
>>> 
>>> 
>> 
>> -- 
>> This message is subject to the CSIR's copyright terms and conditions, 
>> e-mail legal notice, and implemented Open Document Format (ODF) standard. 
>> The full disclaimer details can be found at 
>> http://www.csir.co.za/disclaimer.html.
>> 
>> This message has been scanned for viruses and dangerous content by 
>> MailScanner, 
>> and is believed to be clean.
>> 
>> Please consider the environment before printing this email.
>> 
> 
> 
> -- 
> This message is subject to the CSIR's copyright terms and conditions, 
> e-mail legal notice, and implemented Open Document Format (ODF) standard. 
> The full disclaimer details can be found at 
> http://www.csir.co.za/disclaimer.html.
> 
> This message has been scanned for viruses and dangerous content by 
> MailScanner, 
> and is believed to be clean.
> 
> Please consider the environment before printing this email.
>