perfSONAR User Q&A and Other Discussion

[Casey Russell <>]

Thank you Andy and Larry,

     I must have conflated some concepts.  It's been quite a while now since I watched the videos, seminars, and presentations on 4.0 and it's clear I mis-remembered.  I'm getting ready to re-install several of our nodes and it looks like I'll stick with the dual-nic setup.

Thanks again.

Sincerely,

Casey Russell

Network Engineer

785-856-9809

2029 Becker Drive, Suite 282
Lawrence, Kansas 66047

On Thu, Mar 29, 2018 at 11:31 AM, Andrew Lake <> wrote:

Hi,

Your OWAMP tests by default at a constant rate of N packets per second. Things are done this way because loss events are often short-lived and if you aren't doing that you will miss important stuff..and sometimes we still do miss stuff if N is not big enough even at the sub-second level. That being the case, there’s not much pScheduler or any software can do to schedule a throughput test around an OWAMP test running at that rate. pScheduler is aware that both tests are running, but the OWAMP tests (called ‘latencybg’ tests in pscheduler terminology) are a classification of test that can run at the same time as other tests. As has always been the case, there’s nothing stopping you from running both throughput and these background OWAMP tests on the same NIC, you just need to be aware that you may see some noise in your OWAMP data related to the throughput tests. 

It is possible to use pScheduler to run OWAMP tests that run for a fixed duration, take a break for awhile and run again. This are classified in a way that will not allow them to run at the same time as a throughput test. These are called just “latency” tests in pscheduler terminology. You could do something similar with the old bwctl bwping tool, so that’s not really new. The cost of doing things this way is that you will likely miss important events, and as such it is not a widely used feature. 

Hope that helps,

Andy

On March 29, 2018 at 11:42:57 AM, Casey Russell () wrote:

I seem to remember that pscheduler was going to have a more holistic view of the testing going on on a machine, and that doing latency and bandwidth testing on a single machine/NIC would no longer be the taboo that it once was in 3.x.  Did I dream that, or is that now safe to do?

I ask because I came across the note on this page that still warns against it.  

http://docs.perfsonar.net/install_hardware_details.html   

Sincerely,

Casey Russell

Network Engineer

785-856-9809

2029 Becker Drive, Suite 282
Lawrence, Kansas 66047