jochen
6c98e7ea51
mesh/merge-gate pass: builds build-agent, mesh-controller, route-proxy → ace, g14, novox, shanks; no bus step; every machine composes with the change as it…
mesh/repo-check pass: its merge-check.sh passed
mesh/delivery delivered
mesh/delivery-group group fix/315-a-failed-unit-is-a-condition delivered: every member is delivered
A module's failed unit now arrives among its resources and raises the module's own condition, named by the unit. The machine's own failed units, which no module places, are one warning for the machine listing them, cleared when none is listed. Nothing a send moved is among them, so the gate never holds a send on that finding. The statement's units are kept with the machine's health (migration 0080) and node show says them.
10 lines
728 B
SQL
10 lines
728 B
SQL
-- A failed unit is never silent (novox/hq issue 315).
|
|
--
|
|
-- Beside the state of every long-running resource and its networking, each machine's node-engine states
|
|
-- its service managers: running, degraded or unknown, and every failed unit no module places — a mount of
|
|
-- the machine's own table, a unit a removed package left behind. A module's failed unit is among the
|
|
-- resources, as that module's. Kept with the machine's newest statement, replaced with it, so `node show`
|
|
-- and a controller started again read the same word. Null for a machine whose node-engine is older than
|
|
-- this reading: its units are not known — never healthy, never a reason to raise anything.
|
|
alter table node_health add column units jsonb;
|