compass check leaves the service fleet running, so the next check cannot rebuild the database
Table of Contents
This page is a capture in the inbox bucket of the product backlog — a pre-sprint idea, not yet pulled into a sprint as a story.
1. What
On brave_hopper, the first compass check passed its db step. Its
ticks-smoke step provisions a tenant on the running fleet. After it, all 25
services were running. The second compass check then failed db with
"there are 2 active connection(s) to ores_dev_brave_hopper", and its ctest ran
against a database it had not rebuilt. The cause is inferred from that order;
the check source was not read. Stop the fleet at the end of ticks-smoke, or
stop it before db.
2. Why
A second check run in one environment reports false failures, and a reviewer cannot trust its ctest result.