The schema applies itself, and the setup script stops hiding failures

Migrations were run by hand and nothing recorded which had run, so
re-running the setup script against an existing database failed on the
first CREATE TABLE, and shipping a new migration gave an operator no way
to know whether an estate had it. A missed migration is not a startup
error - it is a query referencing a column that is not there, surfacing
later on whichever endpoint touches it first.

server/internal/migrate applies pending migrations at boot and refuses to
start against a schema it does not match. One transaction per file
holding both the DDL and the row that records it; an advisory lock so two
servers starting at once cannot both apply 008; checksums so an edited
migration is refused by name rather than silently skipped; numeric
ordering so 010 does not run before 009. `migrate -baseline N` adopts a
database built before any of this existed, because "the clients table
exists" does not say whether 007's index does.

Verified on the live database: adopted 001-007, applied 008.

008 adds two indexes on `purchases`, found by asking the database which
foreign keys had nothing behind them and then checking what queries the
table. The conversion report filters client_id + occurred_at, which is
exactly the estate-wide case with no site to narrow it.

run-local.sh had two bugs, both found by running it rather than reading
it: it reused a broker container whose bind mount pointed at a directory
that no longer existed, and it discarded stderr on the mosquitto_passwd
call, so under `set -e` it exited at step 5 with no output at all.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HViLj9gYNRtSr7YVZmW5sn
This commit is contained in:
2026-09-04 12:06:52 +05:30
parent dad04e8cda
commit 5453c26e4c
11 changed files with 941 additions and 8 deletions

View File

@@ -26,9 +26,11 @@ import (
"github.com/loyaly/behavision-server/internal/assistant"
"github.com/loyaly/behavision-server/internal/blob"
"github.com/loyaly/behavision-server/internal/ingest"
"github.com/loyaly/behavision-server/internal/migrate"
"github.com/loyaly/behavision-server/internal/web"
"github.com/loyaly/behavision-server/internal/secret"
"github.com/loyaly/behavision-server/internal/store"
"github.com/loyaly/behavision-server/migrations"
)
var version = "dev"
@@ -44,6 +46,13 @@ func main() {
}
return
}
if len(os.Args) > 1 && os.Args[1] == "migrate" {
if err := runMigrate(os.Args[2:]); err != nil {
fmt.Fprintln(os.Stderr, err)
os.Exit(1)
}
return
}
if err := run(); err != nil {
log.Fatalf("behavision-server: %v", err)
}
@@ -79,6 +88,29 @@ func run() error {
st.UseLogger(logger)
logger.Print("database connected")
// Applied here, not by hand, because an upgrade of this product is "copy
// the new binary and restart it". A migration an operator has to remember
// to run is a migration that does not get run, and its failure is not a
// startup error - it is a query referencing a column that is not there,
// surfacing later on whichever endpoint touches it first.
//
// Refusing to start on failure is deliberate: a server running against a
// schema it does not match writes wrong data, and wrong data outlives the
// outage that stopping causes.
if os.Getenv("BEHAVISION_SKIP_MIGRATE") == "1" {
logger.Print("WARN BEHAVISION_SKIP_MIGRATE=1 - schema not checked")
} else {
applied, err := migrate.Apply(ctx, st.Pool(), migrations.FS)
if err != nil {
return fmt.Errorf("schema: %w", err)
}
if len(applied) == 0 {
logger.Print("schema up to date")
} else {
logger.Printf("schema: applied %s", strings.Join(applied, ", "))
}
}
// Without the key the server still ingests events and serves reports; only
// enrolment fails, and it fails with a message naming the missing variable.
// Refusing to start would take a working estate down over a feature that