TCP, the shortest partition window found for one exchange

What I ran

A partition window that ended a TCP connection between the worlds of a pair, on a round switch, was shrunk from 6,000 ms to 3,409 ms, opening at 1,310 ms of virtual time, in 30 runs, each an exchange between the worlds of the pair.

The shortest window ended the connection in 3 of 3 re-runs, so in all four of its runs. The two one-step neighbours, each a window of 3,408 ms, let the connection survive. The bundle was replayed two times from its own schedule and ended the connection each time.

What it shows

MIGHT. The connection ended, seen once in the window of 3409 ms opening at 1310 ms on a round switch and then in 3 of 3 re-runs.

What it does not show

That the number holds on another host, kernel or hypervisor build, that the exchange gives the same result in general, or anything below the step of 1 ms. Exact replay of the exchange is not shown: the saved output lists it under what the gate does not claim.

According to the repository's source (src/campaign/ComposeShrink.fs), the product's shrinker, molina campaign --compose, takes an isolate and a stop and keeps the chance dials as they were drawn. This shrink was made by the gate's own code. No command reduces a lying-flush file or a window of virtual time on a customer's file, so a customer receives a smallest set of failures only where the shrinker can reduce the failure, and what it returns is the smallest it found against the runs made.

That the worlds ran on cores reserved for them. The ledger row of this gate says isolation was not applicable, and the machine's load average when the run began, a figure the runner recorded, was 12.26, 11.92 and 13.75 over one, five and fifteen minutes.

Conditions

Each trial booted its worlds from nothing and no world was restored. The run is in screening mode.