6450bf1b88
Closes the schedule reconciliation loop end-to-end.
* New `internal/agent/scheduler` package wraps robfig/cron/v3 with
the lifecycle the agent needs:
- Apply(ScheduleSetPayload, Sender) stops the prior cron (waiting
for in-flight entries to return), rebuilds from scratch, starts,
and emits schedule.ack with the version we just applied.
- Disabled entries skipped silently; bad cron exprs (which
shouldn't reach us — the server validates — but defensive)
log a warn and skip.
- On each cron tick the entry sends a new schedule.fire envelope
to the server with {schedule_id, scheduled_at}. The scheduler
itself never builds CommandRunPayloads — server is the source
of truth for jobs.
- tx is swapped on every Apply, so reconnect is handled
naturally: cron entries that fire against a dropped tx log
"no active connection" and skip the tick.
- Stop() is idempotent and waits for the cron's in-flight
workers via cron.Stop().Done().
* New wire message api.MsgScheduleFire + api.ScheduleFirePayload
for the agent → server "I just fired locally" RPC.
* Server-side dispatch (schedule_push.go: dispatchScheduledJob):
looks up the schedule by id, validates ownership + that it's
enabled, builds args from kind (paths for backup; other kinds
are still arg-less in Phase 2 and grow as those job kinds land
in P2-05..08), persists a jobs row with actor_kind=schedule +
scheduled_id, and writes command.run back on the same conn so
the agent runs through its existing dispatch path.
* store.CreateJob now writes scheduled_id. This column was in the
schema since 0001 but never populated — the original P1 path
only had operator-driven jobs, so actor_kind was always 'user'
and scheduled_id was always nil.
* cmd/agent/main.go integration: dispatcher gains a
*scheduler.Scheduler; the MsgScheduleSet case now hands the
payload to scheduler.Apply (in a goroutine so the WS read loop
keeps draining other messages).
* WS dispatcher gains OnScheduleFire alongside OnScheduleAck.
* Tests:
- scheduler unit tests (4): ack-on-apply, cron tick fires
schedule.fire envelope, disabled entries don't fire, replace-
prior-state stops the old cron.
- Server-side end-to-end: schedule.fire → command.run with the
right job_id / kind / args, plus jobs row with actor_kind=
"schedule" and scheduled_id linking back to the schedule.
Persistence of next-fire times across agent restarts is
deliberately deferred. A missed fire window during downtime
simply fires once on reconnect — that's the desirable behaviour
(the operator wants the missed backup to run, not be silently
skipped because we lost track of when it was due).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
317 lines
9.5 KiB
Go
317 lines
9.5 KiB
Go
package main
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
"flag"
|
|
"fmt"
|
|
"log/slog"
|
|
"os"
|
|
"os/signal"
|
|
"syscall"
|
|
"time"
|
|
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/config"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/runner"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/scheduler"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/secrets"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/sysinfo"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/wsclient"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/api"
|
|
"gitea.dcglab.co.uk/steve/restic-manager/internal/restic"
|
|
)
|
|
|
|
var version = "dev"
|
|
|
|
func main() {
|
|
if err := run(); err != nil {
|
|
slog.Error("agent fatal", "err", err)
|
|
os.Exit(1)
|
|
}
|
|
}
|
|
|
|
func run() error {
|
|
configPath := flag.String("config", config.DefaultPath(), "path to agent.yaml")
|
|
enrollServer := flag.String("enroll-server", "", "server URL (used with -enroll-token to perform first-run enrollment)")
|
|
enrollToken := flag.String("enroll-token", "", "one-time enrollment token (operator copies this from the UI)")
|
|
showVersion := flag.Bool("version", false, "print version and exit")
|
|
flag.Parse()
|
|
|
|
if *showVersion {
|
|
fmt.Println("restic-manager-agent", version)
|
|
return nil
|
|
}
|
|
|
|
logger := slog.New(slog.NewJSONHandler(os.Stdout, &slog.HandlerOptions{Level: slog.LevelInfo}))
|
|
slog.SetDefault(logger)
|
|
|
|
cfg, err := config.Load(*configPath)
|
|
if err != nil {
|
|
return fmt.Errorf("config: %w", err)
|
|
}
|
|
|
|
if *enrollToken != "" {
|
|
if *enrollServer == "" {
|
|
return errors.New("enrollment: -enroll-server is required with -enroll-token")
|
|
}
|
|
return doEnroll(*enrollServer, *enrollToken, cfg, version)
|
|
}
|
|
|
|
if !cfg.Enrolled() {
|
|
return fmt.Errorf("agent is not enrolled; run with -enroll-server and -enroll-token first (config %q)", *configPath)
|
|
}
|
|
|
|
ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
|
|
defer stop()
|
|
|
|
snap, err := sysinfo.Collect(ctx, cfg.ResticPath)
|
|
if err != nil {
|
|
return fmt.Errorf("sysinfo: %w", err)
|
|
}
|
|
slog.Info("agent starting",
|
|
"version", version,
|
|
"host_id", cfg.HostID,
|
|
"server", cfg.ServerURL,
|
|
"restic_version", snap.ResticVersion,
|
|
"protocol_version", snap.ProtocolVersion,
|
|
)
|
|
|
|
resticBin, _ := restic.Locate(cfg.ResticPath) // empty is fine; commands fail with a clear error later
|
|
|
|
// Open the secrets store. If the agent is enrolled but has no
|
|
// secrets key yet (legacy YAML), mint one and migrate any
|
|
// plaintext repo fields into the encrypted blob.
|
|
sec, err := openSecretsStore(cfg)
|
|
if err != nil {
|
|
return fmt.Errorf("secrets: %w", err)
|
|
}
|
|
|
|
wsCfg := wsclient.Config{
|
|
ServerURL: cfg.ServerURL,
|
|
AgentToken: cfg.AgentToken,
|
|
HostID: cfg.HostID,
|
|
CertPinSHA256: cfg.CertPinSHA256,
|
|
HelloPayload: api.HelloPayload{
|
|
ProtocolVersion: snap.ProtocolVersion,
|
|
AgentVersion: version,
|
|
ResticVersion: snap.ResticVersion,
|
|
Hostname: snap.Hostname,
|
|
OS: snap.OS,
|
|
Arch: snap.Arch,
|
|
},
|
|
}
|
|
|
|
d := &dispatcher{
|
|
resticBin: resticBin,
|
|
secrets: sec,
|
|
scheduler: scheduler.New(),
|
|
}
|
|
if err := wsclient.Run(ctx, wsCfg, d.handle); err != nil {
|
|
return fmt.Errorf("ws run: %w", err)
|
|
}
|
|
slog.Info("agent shutting down")
|
|
return nil
|
|
}
|
|
|
|
// openSecretsStore opens (or one-time migrates) the agent's encrypted
|
|
// secrets file. Side effects:
|
|
// - mints SecretsKey if absent and persists agent.yaml.
|
|
// - if legacy plaintext repo_url/repo_password sit in agent.yaml,
|
|
// copies them into secrets.enc and clears the YAML fields on
|
|
// the next save.
|
|
func openSecretsStore(cfg *config.Config) (*secrets.Store, error) {
|
|
if err := cfg.EnsureSecretsKey(); err != nil {
|
|
return nil, err
|
|
}
|
|
keyBytes, err := cfg.SecretsKeyBytes()
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
st, err := secrets.New(cfg.ResolvedSecretsPath(), keyBytes)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
migrated := false
|
|
if cfg.LegacyRepoURL != "" || cfg.LegacyRepoPassword != "" {
|
|
cur, _ := st.Load() // empty Repo on first run is fine
|
|
if cur.URL == "" {
|
|
cur.URL = cfg.LegacyRepoURL
|
|
}
|
|
if cur.Password == "" {
|
|
cur.Password = cfg.LegacyRepoPassword
|
|
}
|
|
if err := st.Save(cur); err != nil {
|
|
return nil, fmt.Errorf("migrate legacy creds into secrets.enc: %w", err)
|
|
}
|
|
cfg.LegacyRepoURL = ""
|
|
cfg.LegacyRepoPassword = ""
|
|
migrated = true
|
|
slog.Info("agent: migrated legacy plaintext repo creds into secrets.enc")
|
|
}
|
|
|
|
// Persist key (and the cleared legacy fields) regardless of
|
|
// whether we migrated, in case we just minted SecretsKey.
|
|
if migrated || cfg.SecretsKey != "" {
|
|
if err := cfg.Save(); err != nil {
|
|
return nil, fmt.Errorf("persist agent config: %w", err)
|
|
}
|
|
}
|
|
return st, nil
|
|
}
|
|
|
|
// dispatcher closes over the long-lived agent settings (restic path
|
|
// + secrets handle) so handle() can spawn the runner without
|
|
// re-loading config every time. Repo creds are read fresh from the
|
|
// secrets store on each job — config.update writes through to disk,
|
|
// so a job dispatched in the same session sees the latest values.
|
|
type dispatcher struct {
|
|
resticBin string
|
|
secrets *secrets.Store
|
|
scheduler *scheduler.Scheduler
|
|
}
|
|
|
|
func (d *dispatcher) handle(ctx context.Context, env api.Envelope, tx wsclient.Sender) error {
|
|
switch env.Type {
|
|
case api.MsgCommandRun:
|
|
var p api.CommandRunPayload
|
|
if err := env.UnmarshalPayload(&p); err != nil {
|
|
return fmt.Errorf("command.run: %w", err)
|
|
}
|
|
return d.runJob(ctx, p, tx)
|
|
|
|
case api.MsgCommandCancel:
|
|
// TODO(P2): cancellation requires keeping a job→cancelFunc map.
|
|
slog.Info("ws agent: command.cancel received (cancellation lands in P2)", "id", env.ID)
|
|
|
|
case api.MsgScheduleSet:
|
|
var p api.ScheduleSetPayload
|
|
if err := env.UnmarshalPayload(&p); err != nil {
|
|
return fmt.Errorf("schedule.set: %w", err)
|
|
}
|
|
// scheduler.Apply rebuilds the local cron from scratch and
|
|
// emits schedule.ack via tx. Async-safe: tx may have to wait
|
|
// briefly on the connection's writeMu, but the read loop
|
|
// keeps draining other messages.
|
|
go d.scheduler.Apply(p, tx)
|
|
|
|
case api.MsgConfigUpdate:
|
|
var p api.ConfigUpdatePayload
|
|
_ = env.UnmarshalPayload(&p)
|
|
// Merge with whatever's already in secrets.enc — empty fields
|
|
// in the push mean "leave alone." Atomic write underneath.
|
|
cur, err := d.secrets.Load()
|
|
if err != nil {
|
|
slog.Error("ws agent: load secrets for merge", "err", err)
|
|
return nil
|
|
}
|
|
changed := false
|
|
if p.RepoURL != "" && p.RepoURL != cur.URL {
|
|
cur.URL = p.RepoURL
|
|
changed = true
|
|
}
|
|
if p.RepoUsername != "" && p.RepoUsername != cur.Username {
|
|
cur.Username = p.RepoUsername
|
|
changed = true
|
|
}
|
|
if p.RepoPassword != "" && p.RepoPassword != cur.Password {
|
|
cur.Password = p.RepoPassword
|
|
changed = true
|
|
}
|
|
if changed {
|
|
if err := d.secrets.Save(cur); err != nil {
|
|
slog.Error("ws agent: persist secrets", "err", err)
|
|
return nil
|
|
}
|
|
slog.Info("ws agent: repo credentials updated via config.update")
|
|
}
|
|
|
|
case api.MsgAgentUpdateAvail:
|
|
var p api.AgentUpdateAvailablePayload
|
|
_ = env.UnmarshalPayload(&p)
|
|
slog.Info("ws agent: update available", "version", p.LatestVersion, "url", p.PackageURL)
|
|
|
|
default:
|
|
slog.Debug("ws agent: ignored message", "type", env.Type)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// runJob spawns a runner for one job. We launch a goroutine so the
|
|
// WS read loop keeps draining messages while restic chugs along.
|
|
func (d *dispatcher) runJob(ctx context.Context, p api.CommandRunPayload, tx wsclient.Sender) error {
|
|
if d.resticBin == "" {
|
|
return fmt.Errorf("restic binary not located on this agent")
|
|
}
|
|
creds, err := d.secrets.Load()
|
|
if err != nil {
|
|
return fmt.Errorf("load repo credentials: %w", err)
|
|
}
|
|
if creds.Empty() {
|
|
return fmt.Errorf("repo credentials not configured (waiting for server config.update push)")
|
|
}
|
|
r := runner.New(runner.Config{
|
|
ResticBin: d.resticBin,
|
|
RepoURL: creds.URL,
|
|
RepoUsername: creds.Username,
|
|
RepoPassword: creds.Password,
|
|
}, tx, time.Second)
|
|
|
|
switch p.Kind {
|
|
case api.JobBackup:
|
|
// Agent.Args carries [paths...]. Excludes/tags are not yet
|
|
// surfaced over the wire; they come with P2 schedule support.
|
|
slog.Info("agent: accepting backup job",
|
|
"job_id", p.JobID, "paths", p.Args)
|
|
go func() {
|
|
if err := r.RunBackup(ctx, p.JobID, p.Args, nil, nil); err != nil {
|
|
slog.Warn("agent: backup job failed", "job_id", p.JobID, "err", err)
|
|
return
|
|
}
|
|
slog.Info("agent: backup job complete", "job_id", p.JobID)
|
|
}()
|
|
case api.JobInit:
|
|
slog.Info("agent: accepting init job", "job_id", p.JobID)
|
|
go func() {
|
|
if err := r.RunInit(ctx, p.JobID); err != nil {
|
|
slog.Warn("agent: init job failed", "job_id", p.JobID, "err", err)
|
|
return
|
|
}
|
|
slog.Info("agent: init job complete", "job_id", p.JobID)
|
|
}()
|
|
default:
|
|
return fmt.Errorf("kind %q not implemented yet (Phase 2 lands the rest)", p.Kind)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func doEnroll(serverURL, token string, cfg *config.Config, agentVersion string) error {
|
|
ctx, cancel := context.WithTimeout(context.Background(), 60*time.Second)
|
|
defer cancel()
|
|
|
|
snap, err := sysinfo.Collect(ctx, cfg.ResticPath)
|
|
if err != nil {
|
|
return fmt.Errorf("sysinfo: %w", err)
|
|
}
|
|
res, err := wsclient.Enroll(ctx, serverURL, wsclient.EnrollRequest{
|
|
Token: token,
|
|
HostName: snap.Hostname,
|
|
OS: snap.OS,
|
|
Arch: snap.Arch,
|
|
AgentVersion: agentVersion,
|
|
ResticVersion: snap.ResticVersion,
|
|
})
|
|
if err != nil {
|
|
return fmt.Errorf("enroll: %w", err)
|
|
}
|
|
cfg.ServerURL = serverURL
|
|
cfg.HostID = res.HostID
|
|
cfg.AgentToken = res.AgentToken
|
|
cfg.CertPinSHA256 = res.CertPinSHA256
|
|
if err := cfg.Save(); err != nil {
|
|
return fmt.Errorf("save config: %w", err)
|
|
}
|
|
fmt.Fprintf(os.Stderr, "enrolled as host %s on %s\n", res.HostID, serverURL)
|
|
return nil
|
|
}
|