Hi, Peter.
Thanks for your input.
It is completely tester's failure.
The calucuration is done only by back calculation from the unicast NS.
So, your situation must be happen.
Usually, the problem is hidden by repeating the test
because ReachableTime has the randomness.
I will change the judgement.
The tester must also care about actual time when NUT reply the Echo.
Regards,
On Wed, 5 Mar 2008 17:57:00 -0800
"Hunt Peter (Nokia-S&S/MtView)" <Peter.Hunt@nokia.com> wrote:
>
> I'm running the Self Test suite in the following configuration:
>
> * Environment: FreeBSD 6.3
> * Test Tool version: v6eval 3.0.12
> * Self Test package version 1-4-11
>
> I have a question about the test V6LC_2_2_14_C in the v2 Core suite, which tests the reachable time of a router.
>
> The test determines the reachable time of the router by resolving an ND entry, sending echo requests each second until it receives an NS from the NUT, and then subtracting 5 seconds (i.e. the DELAY_FIRST_PROBE_TIME) from the time it took to receive the NS. It then verifies the reachable time is between 5 and 15 seconds (i.e. the valid range for the randomized reachable time derived from a 10-second configured value).
>
> My concern is that the test might result in a false failure if the randomized reachable time value is close to 15 seconds, and there is a delay between when the reachable time for the entry is reached, and when the NUT receives the next echo request from the TN.
>
> For example, if the randomized reachable time is 14.8 seconds:
>
> * At 0 seconds, the entry on the NUT is resolved
> * At 14.8 seconds, the reachable time is reached and the entry
> enters STALE state
> * At 15.3 seconds (say) the NUT receives the next echo request
> from the TN, sends an echo reply, and the entry enters DELAY state.
> * At 20.3 seconds, the entry enters PROBE state and the NUT sends
> the next NS.
>
> In this case, the test would calculate a reachable time of 15.3 seconds and fail. Yet the NUT derived the randomized value correctly.
>
> Does the test account for the time that the ND entry on the NUT is in STALE state? That is, does it account for the time between when the reachable time expires and the next echo reply is sent?
>
> Thanks in advance,
>
> Peter Hunt
> Software Engineer
> Nokia S&S
>
----
Yukiyo Akisada <akisada@tahi.org>