infra(docker)+test(siteeventlog): close register tripwires 79 + 93 — rig data to named volumes; pin site_events schema to the cap-model calibration

Row 79 executed its own prescribed remedy without waiting for a second
corruption: all ten rig data dirs (8 docker/ nodes + 2 docker-env2/ site
nodes) move from virtiofs bind mounts to named docker volumes
(scadabridge-<node>-data), seeded in place with both rigs stopped and
live-verified (site-a localDbReplicationConnected=true / backlog 0 over
the migrated files; env2 stable, 0 restarts). The feared inspectability
loss is moot — OrbStack exposes volumes at ~/OrbStack/docker/volumes/.
The rebuild surfaced latent env2 drift (pre-Secrets-0.6.2 image): env2
now carries the x-secrets-store-env posture + central data volumes.

Row 93's silent maintenance trap is now loud: SiteEventsSchemaCalibrationTests
pins the site_events columns/indexes the EventLogPurgeService size model is
calibrated against, failing with a pointer to the memo §4 recalibration on
any schema drift. Rows 92 (accepted by design) and 94 (standing tripwire)
annotated accordingly.
This commit is contained in:
Joseph Doherty
2026-08-18 08:43:52 -04:00
parent 5936573e71
commit 324ba369d9
8 changed files with 284 additions and 83 deletions
+6 -5
View File
@@ -3,13 +3,14 @@ set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
# The compose file's bind mounts are RELATIVE (./site-a-node-a/data etc.), so running
# this script from a git worktree silently rehomes every node's data directory to the
# worktree's copy — the rig keeps running, but its SQLite stores split from the main
# checkout's and all state written there is lost on the next deploy from main.
# The compose file's remaining bind mounts (appsettings, logs) are RELATIVE, so running
# this script from a git worktree silently rehomes every node's config to the worktree's
# copy — the rig keeps running, but on stale/divergent appsettings. (The data dirs moved
# to named volumes — register row 79 — so rig state itself no longer splits, but a
# worktree deploy is still wrong by default.)
if [[ "$SCRIPT_DIR" == *"/.claude/worktrees/"* && "${SCADABRIDGE_ALLOW_WORKTREE_DEPLOY:-}" != "1" ]]; then
echo "ERROR: refusing to deploy from a git worktree ($SCRIPT_DIR)." >&2
echo "Bind-mount data dirs would rehome to the worktree copy and split rig state." >&2
echo "Bind-mounted appsettings would rehome to the worktree copy." >&2
echo "Deploy from the main checkout, or set SCADABRIDGE_ALLOW_WORKTREE_DEPLOY=1 to override." >&2
exit 1
fi