Files
stalwart-migrator/cmd/stalwart-migrate/run.go
T
jcoffey-dev 7e04351b0f Add cutover; drop rollback in favour of operator-provided recovery
Two changes that arrived together: the cutover phase (ARCHITECTURE.md 4.5)
is implemented, and the rollback phase is deleted. Recovery from a failed
migration is now explicitly the operator's own snapshot or backup, and out
of scope for this tool.

internal/cutover implements 4.5 as seven checkpointed steps: verify the
staged binary's version, install it, preserve and rewrite the service
definition, reload, start, wait for a healthy JMAP session, recalculate
quotas.

The unit is rewritten in place rather than generated from a template. An
operator's unit carries hardening options, limits and dependencies this
tool has no business having an opinion about, and regenerating it would
silently drop them. It repoints ExecStart (preserving systemd's -@:+!
prefix characters and every argument after the executable), updates
--config, and strips recovery-mode Environment lines - leaving
STALWART_RECOVERY_MODE=1 set would recovery-boot the service on every
restart, forever. It refuses on a unit with no ExecStart, and on an
Environment line mixing a recovery variable with others: a line it only
partly understands is one it must not edit.

Quota recalculation is the one step allowed to fail without failing the
phase. Its wire format is grounded in Stalwart's x:Task schema reference -
Task/set creating one AccountMaintenance per account with maintenanceType
recalculateQuota - but the upgrade guide only documents the WebUI path, so
two details remain inferred and are called out in stalwartapi/task.go:
whether the schema's "read-only" annotation on accountId/maintenanceType
means "immutable after creation", and whether a finished task simply leaves
the queue (TaskStatus documents Pending/Retry/Failed with no success
state). Warning rather than failing is the honest response to that
uncertainty, and stale counters are an accounting problem next to calling
for a restore of a machine that is otherwise migrated and serving mail.

Docker deployments are refused outright: cutting a container over means
pulling an image and recreating it, not swapping a binary.

On removing rollback. The implementation worked and was tested, and it was
removed because restoring bytes correctly is not the hard part. It copied
file contents and permissions and verified every restored file against a
manifest - and did not preserve ownership. Run as root, as this tool
requires, it would have produced a byte-perfect, checksum-verified,
root-owned data directory that Stalwart, running as its own user, could not
open, and it would have reported success. The PostgreSQL path was worse:
pg_dump without --clean emits CREATE TABLE + COPY, which fails replaying
into a database whose tables still exist, and the ON_ERROR_STOP=1 added so
a half-applied restore couldn't be reported as success turned that into a
hard failure. None of it had ever run against a real server. A filesystem
snapshot has none of these failure modes, because it never lost the
metadata to begin with.

So cutover's gate is no longer rollback.CanRollBack but an explicit
RecoveryPointConfirmed acknowledgement. That is an assertion, not a check -
this tool cannot verify someone else's snapshot - and its only value is
that nobody migrates a production mail server having never been asked the
question. Two consequences are accepted deliberately: restoring any
pre-migration recovery point discards mail delivered since, and a failed
migration now stops and reports rather than undoing itself.

What the tool still does to make a manual restore easier: the old binary is
preserved and never deleted, the original service definition is preserved
before the rewrite, the settings and principals dumps stay on disk, and
every artifact path and checksum stays in the checkpoint where `status
<run-id>` can print it.

Also removed: the `confirm` command stub and RollbackWindowClosed, whose
only purpose was closing a rollback window that no longer exists, and
checkpoint.PhaseRollback. Old state.json files still load - JSON ignores
the now-unknown field.

Still open, and recorded in 8: cutover ignores systemd drop-ins, so an
ExecStart or Environment override in stalwart.service.d/*.conf is invisible
to the rewrite - including the recovery variable it exists to strip;
nothing prevents concurrent runs on the same run-id; and nothing in this
repo has ever run against a real Stalwart, real systemd, or a real store.
2026-08-23 17:52:47 -07:00

249 lines
12 KiB
Go

package main
import (
"context"
"flag"
"fmt"
"net/http"
"os"
"path/filepath"
"github.com/LINUXexpert-org/stalwart-migrator/internal/backup"
"github.com/LINUXexpert-org/stalwart-migrator/internal/checkpoint"
"github.com/LINUXexpert-org/stalwart-migrator/internal/plan"
"github.com/LINUXexpert-org/stalwart-migrator/internal/preflight"
"github.com/LINUXexpert-org/stalwart-migrator/internal/recovery"
"github.com/LINUXexpert-org/stalwart-migrator/internal/validate"
)
// runRun implements `stalwart-migrate run`. Only --dry-run is available
// today. internal/cutover exists, so the gap is no longer a phase but the
// pipeline around it: §4.3 staging, and the wiring that would drive
// preflight -> backup -> stage -> recovery-mode -> cutover -> validate
// against real paths rather than a sandbox. See ARCHITECTURE.md §8. `run`
// without --dry-run refuses rather than doing a migration partway.
//
// --dry-run runs preflight and a real backup (see the caveat printed below
// about why backup still touches the live data directory), then - if the
// plan crosses the 0.15/0.16 boundary - clones the verified backup into a
// disposable sandbox, converts the settings snapshot to point at that
// sandbox (via migrate_v016.py's own documented --patch-paths mechanism,
// not by this tool guessing at config.json's schema), runs the real
// recovery-mode migration against the sandbox, and boots the result
// normally to confirm it comes up. Nothing at the real binary path or the
// real service is ever touched.
//
// Every byte a dry run writes - the fs-backup copy, the settings/principals
// dumps, the downloaded migrate_v016.py, the sandbox clone and its
// config/export files - lives under one per-run directory
// (work-dir/<run-id>) that a deferred cleanup at the bottom of this
// function removes on every exit path: success, a failed check partway
// through, or an early refusal. The only thing left behind afterward is the
// checkpoint's state.json under --state-dir, which is exactly the
// success/failure log a rerun's `status <run-id>` reads - not bulk data.
// --keep-artifacts opts out, for when a failure needs inspecting.
func runRun(args []string) (err error) {
fs := flag.NewFlagSet("run", flag.ExitOnError)
binaryPath := fs.String("binary", "/usr/local/bin/stalwart", "path to the currently-installed stalwart binary")
targetBinaryPath := fs.String("target-binary", "", "path to an already-downloaded target-version stalwart binary (required to simulate a major-boundary migration)")
configPath := fs.String("config", "/etc/stalwart/config.toml", "path to stalwart's current config file")
dataDir := fs.String("data-dir", "/var/lib/stalwart", "stalwart data directory")
containerName := fs.String("container", "stalwart", "docker container name, if applicable")
adminURL := fs.String("admin-url", "", "base URL for the live instance's admin/JMAP API (required)")
adminUser := fs.String("admin-user", "", "admin username")
adminPassword := fs.String("admin-password", os.Getenv("STALWART_MIGRATE_ADMIN_PASSWORD"),
"admin password (or set STALWART_MIGRATE_ADMIN_PASSWORD)")
targetVersion := fs.String("target", "latest", `target Stalwart version, or "latest"`)
stateDir := fs.String("state-dir", checkpoint.DefaultBaseDir, "directory to store run checkpoints in")
workDir := fs.String("work-dir", "/var/lib/stalwart-migrator/work", "scratch directory for backups, dumps, and the dry-run sandbox (cleaned up afterward - see --keep-artifacts)")
stalwartCLI := fs.String("stalwart-cli", "stalwart-cli", "path to the stalwart-cli binary")
pythonPath := fs.String("python", "python3", "path to python3")
migrationScriptSHA256 := fs.String("migration-script-sha256", "", "pinned sha256 of migrate_v016.py (recommended; see preflight/backup output for the hash to pin after a first unpinned run)")
recoveryPort := fs.Int("recovery-port", 8080, "port recovery mode's HTTP listener binds, per UPGRADING/v0_16.md's own examples")
minFree := fs.Float64("min-free-multiple", 2.0, "required free disk space as a multiple of the data directory size")
dryRun := fs.Bool("dry-run", false, "simulate and validate the migration against a disposable sandbox, without touching production")
keepArtifacts := fs.Bool("keep-artifacts", false, "don't delete work-dir/<run-id> afterward (the fs-backup copy, dumps, and sandbox) - useful for inspecting a failure")
if err := fs.Parse(args); err != nil {
return err
}
if !*dryRun {
return fmt.Errorf(
"real (non-dry-run) migrations aren't available yet: cutover is implemented (ARCHITECTURE.md §4.5), but nothing wires it " +
"into a production run - the staging phase (§4.3) and the real pipeline don't exist yet, so this command has no path " +
"that touches production. Run with --dry-run to validate the migration mechanics against a disposable sandbox copy " +
"of your data. Note that when a real run does land, recovery from a failed migration will be your own snapshot or " +
"backup - this tool does not undo a migration (§4.8)",
)
}
if *adminURL == "" {
return fmt.Errorf("--admin-url is required")
}
ctx := context.Background()
httpClient := &http.Client{}
if err := os.MkdirAll(*workDir, 0o750); err != nil {
return fmt.Errorf("create work dir %s: %w", *workDir, err)
}
store := checkpoint.NewStore(*stateDir)
rs, err := store.Create("", *targetVersion)
if err != nil {
return fmt.Errorf("create run: %w", err)
}
fmt.Printf("run id: %s\n\n", rs.RunID)
runWorkDir := filepath.Join(*workDir, rs.RunID)
logPath := filepath.Join(*stateDir, rs.RunID, "state.json")
defer func() {
if _, statErr := os.Stat(runWorkDir); os.IsNotExist(statErr) {
return // nothing was ever written (e.g. refused before backup ran)
}
if *keepArtifacts {
fmt.Printf("\nartifacts kept at %s (--keep-artifacts) - remove manually when done inspecting\n", runWorkDir)
return
}
outcome := "succeeded"
if err != nil {
outcome = "failed"
}
if rmErr := os.RemoveAll(runWorkDir); rmErr != nil {
fmt.Fprintf(os.Stderr, "\nwarning: dry run %s, but failed to clean up %s: %v (remove it manually)\n", outcome, runWorkDir, rmErr)
return
}
fmt.Printf("\ndry run %s - cleaned up %s; the run log is at %s\n", outcome, runWorkDir, logPath)
}()
fmt.Println("--- preflight ---")
checker := preflight.New(preflight.Options{
BinaryPath: *binaryPath, ConfigPath: *configPath, DataDir: *dataDir, ContainerName: *containerName,
AdminURL: *adminURL, AdminUser: *adminUser, AdminPassword: *adminPassword,
TargetVersion: *targetVersion, MinFreeMultiple: *minFree, HTTPClient: httpClient,
})
pfReport, err := checker.Run(ctx, store, rs)
fmt.Print(pfReport.String())
if err != nil {
return fmt.Errorf("preflight failed to complete: %w", err)
}
if pfReport.Blocking() {
return fmt.Errorf("preflight found blocking issues - see FAIL lines above")
}
p, err := plan.Decide(rs.SourceVersion, rs.TargetVersion)
if err != nil {
return fmt.Errorf("plan: %w", err)
}
fmt.Printf("\nplan: %s\n", p.Reason)
fmt.Println("\n--- backup ---")
fmt.Println("(dry-run does not stop the live Stalwart service itself - internal/service can now do that, but dry-run " +
"isn't wired to offer it. For a guaranteed-consistent snapshot, stop stalwart before running this; otherwise the " +
"filesystem copy may reflect a live, in-use store. This is unrelated to whether production gets touched - it never does.)")
backupDir := filepath.Join(runWorkDir, "backup")
scriptDest := filepath.Join(runWorkDir, "migrate_v016.py")
settingsPath := filepath.Join(runWorkDir, "settings.json")
principalsPath := filepath.Join(runWorkDir, "principals.json")
backupOpts := backup.Options{
BinaryPath: *binaryPath,
SkipBinaryPreservation: true, // dry-run: never touch the production binary
DataDir: *dataDir,
BackupDir: backupDir,
MigrationScriptSHA256: *migrationScriptSHA256,
ScriptDestPath: scriptDest,
AdminURL: *adminURL,
AdminUser: *adminUser,
AdminPassword: *adminPassword,
SettingsDumpPath: settingsPath,
PrincipalsDumpPath: principalsPath,
PythonPath: *pythonPath,
HTTPClient: httpClient,
}
bkReport, err := backup.Run(ctx, store, rs, backupOpts)
fmt.Print(bkReport.String())
if err != nil {
return fmt.Errorf("backup failed: %w", err)
}
if !p.CrossesMajorBoundary {
fmt.Println("\nthis is a same-boundary patch upgrade: there's no recovery-mode phase to simulate. " +
"preflight and backup above are as far as a dry-run goes for this path - a real run would be a binary swap and restart.")
return nil
}
if *targetBinaryPath == "" {
return fmt.Errorf("--target-binary is required to simulate a major-boundary migration (0.15 -> 0.16 crosses one here)")
}
fmt.Println("\n--- convert (settings -> sandbox config) ---")
sandboxDataDir := filepath.Join(runWorkDir, "sandbox-data")
sandboxConfigPath := filepath.Join(runWorkDir, "sandbox-config.json")
sandboxExportPath := filepath.Join(runWorkDir, "sandbox-export.json")
if _, err := store.RunStep(rs, checkpoint.PhaseStage, "clone-sandbox-data", func() (checkpoint.StepOutcome, error) {
manifest, err := backup.CopyDataDir(backupDir, sandboxDataDir)
if err != nil {
return checkpoint.StepOutcome{}, err
}
return checkpoint.StepOutcome{Detail: fmt.Sprintf("cloned the verified backup (%d files) into the sandbox at %s", len(manifest.Files), sandboxDataDir)}, nil
}); err != nil {
return fmt.Errorf("clone sandbox data: %w", err)
}
fmt.Printf("cloned verified backup into sandbox: %s\n", sandboxDataDir)
if _, err := store.RunStep(rs, checkpoint.PhaseStage, "convert-settings", func() (checkpoint.StepOutcome, error) {
if err := backup.RunSettingsConvert(ctx, backup.SettingsConvertOptions{
PythonPath: *pythonPath, ScriptPath: scriptDest,
SettingsPath: settingsPath, PrincipalsPath: principalsPath,
ConfigPath: sandboxConfigPath, OutputPath: sandboxExportPath,
PatchPaths: map[string]string{*dataDir: sandboxDataDir},
}); err != nil {
return checkpoint.StepOutcome{}, err
}
return checkpoint.StepOutcome{Detail: fmt.Sprintf("generated %s and %s, patched to point at the sandbox", sandboxConfigPath, sandboxExportPath)}, nil
}); err != nil {
return fmt.Errorf("convert settings: %w", err)
}
fmt.Println("generated sandbox config.json and export.json")
fmt.Println("\n--- recovery-mode migration (against the sandbox) ---")
listenURL := fmt.Sprintf("http://127.0.0.1:%d/", *recoveryPort)
recReport, err := recovery.Run(ctx, store, rs, recovery.Options{
BinaryPath: *targetBinaryPath, ConfigPath: sandboxConfigPath, ListenURL: listenURL,
AdminUser: "admin", ApplyFiles: []string{sandboxExportPath}, CLIBinaryPath: *stalwartCLI,
HTTPClient: httpClient,
})
fmt.Print(recReport.String())
if err != nil {
return fmt.Errorf("recovery-mode migration against the sandbox failed: %w", err)
}
fmt.Println("\n--- boot check (normal boot of the migrated sandbox) ---")
if rs.PreflightSnapshot != nil {
fmt.Println("(comparing against the pre-migration account/mailbox snapshot preflight captured - " +
"this is the actual no-data-loss check, not just a reachability probe)")
} else {
fmt.Println("(no pre-migration snapshot to compare against - preflight couldn't capture one, most likely " +
"because --admin-url wasn't set; only reachability is checked)")
}
valReport, err := validate.Run(ctx, store, rs, validate.BootCheckOptions{
BinaryPath: *targetBinaryPath, ConfigPath: sandboxConfigPath, ListenURL: listenURL, HTTPClient: httpClient,
ContentIntegrityBefore: rs.PreflightSnapshot,
AdminUser: *adminUser,
AdminPassword: *adminPassword,
})
fmt.Print(valReport.String())
if err != nil {
return fmt.Errorf("post-migration validation failed: %w", err)
}
verified := "the migration mechanics succeeded"
if rs.PreflightSnapshot != nil {
verified = "the migration mechanics succeeded AND every account/mailbox message count matched before vs. after"
}
fmt.Printf("\nDRY RUN COMPLETE for run %s: %s, against a disposable sandbox copy of your data. Nothing in production was touched.\n", rs.RunID, verified)
return nil
}