Files
stalwart-migrator/internal/preflight/checks_test.go
T
jcoffey-dev 3e155fa42c Fix three defects a full VM migration exposed
Ran a complete 0.15.5 -> 0.16.14 migration of the smoke VM, driving the
phases in the order the real pipeline will. It worked - all mail intact and
readable afterwards, all ten listeners up, cutover executed for the first
time ever and checkpoint resume exercised - and it exposed three defects.

1. The converted config was installed root-owned while the service runs as
   its own user. Stalwart crash-looped 28 times on "Failed to read data
   store settings: Permission denied", minutes after the mistake and
   nowhere near it. This is the same ownership trap that retired the
   rollback implementation, in a new place: writing files as root is the
   natural thing for a tool running as root to do, and it is wrong every
   time the service is not root.

   Cutover now installs the config itself, copying ownership and mode from
   the config being replaced.

2. v0.16.14 does not serve /api - the endpoint stalwartapi assumed.
   Confirmed against a fully migrated, fully configured, serving instance
   rather than a sandbox: /api, /api/principal and /jmap/ all 404. The JMAP
   endpoint is the one the session document advertises, which is what RFC
   8620 discovery is for.

   The client now discovers it, re-basing the advertised path onto the
   operator's host: a real instance advertises its canonical public URL
   ("https://mail.smoke.test/jmap/") which frequently isn't reachable from
   where this tool runs. The session is authoritative about the path; the
   operator is authoritative about the host.

3. Dispatching on the urn:stalwart:jmap capability was wrong, because
   NEITHER version advertises it - not 0.15.5, and not a fully migrated
   0.16.14. That sent 0.16 instances down the 0.15 REST path where every
   call 404s. The client probes what the instance actually serves instead.
   Less elegant than a declared capability, with the advantage of being
   true.

Also: a JMAP "forbidden" now explains itself. An account holding the admin
role before the migration was refused x:Account/query afterwards, and a
bare "forbidden" gives an operator nowhere to start. Whether the role
failed to carry or v0.16 wants different permissions was not isolated, and
that question is recorded as open - it gates quota recalculation and any
post-migration validation.

Verified against both live instances: the 0.15.5 reports 3 accounts and its
domain over REST, and the migrated 0.16.14 routes to JMAP, finds the right
endpoint, and returns the explained refusal.
2026-08-23 21:32:30 -07:00

317 lines
10 KiB
Go

// SPDX-FileCopyrightText: 2026 LINUXexpert-org
// SPDX-License-Identifier: GPL-3.0-or-later
package preflight
import (
"context"
"encoding/json"
"fmt"
"net/http"
"net/http/httptest"
"os"
"path/filepath"
"strings"
"testing"
"github.com/LINUXexpert-org/stalwart-migrator/internal/checkpoint"
)
// writeFakeBinary creates a shell script that behaves like `stalwart
// --version` and records each invocation to counterPath, so tests can
// assert a checkpointed step was (or wasn't) re-executed on resume.
func writeFakeBinary(t *testing.T, version, counterPath string) string {
t.Helper()
dir := t.TempDir()
scriptPath := filepath.Join(dir, "fake-stalwart.sh")
script := fmt.Sprintf("#!/bin/sh\necho invoked >> %q\necho 'stalwart %s'\n", counterPath, version)
if err := os.WriteFile(scriptPath, []byte(script), 0o755); err != nil {
t.Fatal(err)
}
return scriptPath
}
func countLines(t *testing.T, path string) int {
t.Helper()
data, err := os.ReadFile(path)
if os.IsNotExist(err) {
return 0
}
if err != nil {
t.Fatal(err)
}
return len(strings.Split(strings.TrimSpace(string(data)), "\n"))
}
func TestCheckerRunEndToEndAndResume(t *testing.T) {
// -- fixtures --------------------------------------------------------
counterPath := filepath.Join(t.TempDir(), "invocations")
binaryPath := writeFakeBinary(t, "0.15.5", counterPath)
configPath := filepath.Join(t.TempDir(), "config.toml")
tomlCfg := "[store.\"rocksdb\"]\ntype = \"rocksdb\"\npath = \"/var/lib/stalwart/data\"\n"
if err := os.WriteFile(configPath, []byte(tomlCfg), 0o644); err != nil {
t.Fatal(err)
}
dataDir := t.TempDir()
if err := os.WriteFile(filepath.Join(dataDir, "db"), make([]byte, 1024), 0o644); err != nil {
t.Fatal(err)
}
withFakeGithub(t, func(w http.ResponseWriter, r *http.Request) {
json.NewEncoder(w).Encode(Release{
TagName: "v0.16.14",
Assets: []ReleaseAsset{{Name: "checksums.txt"}},
})
})
stateDir := t.TempDir()
store := checkpoint.NewStore(stateDir)
rs, err := store.Create("", "latest")
if err != nil {
t.Fatalf("store.Create: %v", err)
}
checker := New(Options{
BinaryPath: binaryPath,
ConfigPath: configPath,
DataDir: dataDir,
TargetVersion: "latest",
})
// -- first run: everything should execute -----------------------------
report, err := checker.Run(context.Background(), store, rs)
if err != nil {
t.Fatalf("Run #1: %v", err)
}
if report.Blocking() {
t.Fatalf("Run #1: unexpected blocking report:\n%s", report.String())
}
if got := countLines(t, counterPath); got != 1 {
t.Fatalf("binary invocations after Run #1 = %d, want 1", got)
}
if rs.SourceVersion != "0.15.5" {
t.Errorf("rs.SourceVersion = %q, want 0.15.5", rs.SourceVersion)
}
if rs.TargetVersion != "0.16.14" {
t.Errorf("rs.TargetVersion = %q, want 0.16.14 (resolved from \"latest\")", rs.TargetVersion)
}
if rs.Topology.StoreBackend != "rocksdb" {
t.Errorf("rs.Topology.StoreBackend = %q, want rocksdb", rs.Topology.StoreBackend)
}
foundDirection := false
for _, res := range report.Results {
if res.Name == "upgrade-direction" {
foundDirection = true
if !strings.Contains(res.Detail, "major boundary") {
t.Errorf("upgrade-direction detail = %q, want mention of the major boundary", res.Detail)
}
}
}
if !foundDirection {
t.Error("report missing upgrade-direction check")
}
// -- simulate a crash and resume: reload state from disk fresh --------
resumed, err := store.Load(rs.RunID)
if err != nil {
t.Fatalf("store.Load (resume): %v", err)
}
report2, err := checker.Run(context.Background(), store, resumed)
if err != nil {
t.Fatalf("Run #2 (resume): %v", err)
}
if got := countLines(t, counterPath); got != 1 {
t.Errorf("binary invocations after resumed Run = %d, want 1 (already-done steps must not re-execute)", got)
}
if len(report2.Results) != len(report.Results) {
t.Errorf("resumed report has %d results, want %d (same as first run)", len(report2.Results), len(report.Results))
}
}
func TestCheckerRunFlagsTooOldSource(t *testing.T) {
counterPath := filepath.Join(t.TempDir(), "invocations")
binaryPath := writeFakeBinary(t, "0.14.2", counterPath)
configPath := filepath.Join(t.TempDir(), "config.toml")
if err := os.WriteFile(configPath, []byte("[server]\nhostname = \"mail.example.com\"\n"), 0o644); err != nil {
t.Fatal(err)
}
dataDir := t.TempDir()
withFakeGithub(t, func(w http.ResponseWriter, r *http.Request) {
json.NewEncoder(w).Encode(Release{TagName: "v0.16.14"})
})
store := checkpoint.NewStore(t.TempDir())
rs, err := store.Create("", "latest")
if err != nil {
t.Fatal(err)
}
checker := New(Options{BinaryPath: binaryPath, ConfigPath: configPath, DataDir: dataDir, TargetVersion: "latest"})
report, err := checker.Run(context.Background(), store, rs)
if err != nil {
t.Fatalf("Run: %v", err)
}
if !report.Blocking() {
t.Fatalf("expected a blocking report for a too-old source version, got:\n%s", report.String())
}
}
func TestCheckerRunFlagsInsufficientDiskSpace(t *testing.T) {
counterPath := filepath.Join(t.TempDir(), "invocations")
binaryPath := writeFakeBinary(t, "0.15.5", counterPath)
configPath := filepath.Join(t.TempDir(), "config.toml")
if err := os.WriteFile(configPath, []byte("[store.\"rocksdb\"]\ntype = \"rocksdb\"\n"), 0o644); err != nil {
t.Fatal(err)
}
dataDir := t.TempDir()
if err := os.WriteFile(filepath.Join(dataDir, "db"), make([]byte, 1024), 0o644); err != nil {
t.Fatal(err)
}
withFakeGithub(t, func(w http.ResponseWriter, r *http.Request) {
json.NewEncoder(w).Encode(Release{TagName: "v0.16.14"})
})
store := checkpoint.NewStore(t.TempDir())
rs, err := store.Create("", "latest")
if err != nil {
t.Fatal(err)
}
// An absurd multiple guarantees the free-space check fails regardless
// of how much space the test host actually has free.
checker := New(Options{
BinaryPath: binaryPath, ConfigPath: configPath, DataDir: dataDir,
TargetVersion: "latest", MinFreeMultiple: 1e12,
})
report, err := checker.Run(context.Background(), store, rs)
if err != nil {
t.Fatalf("Run: %v", err)
}
if !report.Blocking() {
t.Fatalf("expected a blocking report for insufficient disk space, got:\n%s", report.String())
}
}
func TestCheckerRunCapturesAccountSnapshotWhenAdminURLSet(t *testing.T) {
counterPath := filepath.Join(t.TempDir(), "invocations")
binaryPath := writeFakeBinary(t, "0.15.5", counterPath)
configPath := filepath.Join(t.TempDir(), "config.toml")
if err := os.WriteFile(configPath, []byte("[store.\"rocksdb\"]\ntype = \"rocksdb\"\n"), 0o644); err != nil {
t.Fatal(err)
}
dataDir := t.TempDir()
if err := os.WriteFile(filepath.Join(dataDir, "db"), make([]byte, 1024), 0o644); err != nil {
t.Fatal(err)
}
withFakeGithub(t, func(w http.ResponseWriter, r *http.Request) {
json.NewEncoder(w).Encode(Release{TagName: "v0.16.14"})
})
// A fake admin server: answers Ping's session-discovery GET, the
// impersonated session-discovery GET MailboxSnapshot makes for
// [email protected], and the POST /api calls for x:Account/query,
// x:Account/get, and Mailbox/get.
var apiURL string
adminSrv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/api/principal" {
w.WriteHeader(http.StatusNotFound) // v0.16 shape: no REST management API
return
}
switch {
case r.Method == http.MethodGet && r.URL.Path == "/.well-known/jmap":
user, _, _ := r.BasicAuth()
if strings.Contains(user, "%") {
json.NewEncoder(w).Encode(map[string]any{
"apiUrl": apiURL,
"primaryAccounts": map[string]string{"urn:ietf:params:jmap:mail": "mail-alice"},
})
return
}
// A 0.16-era instance: the urn:stalwart:jmap capability is what
// says its management API is the JMAP one.
json.NewEncoder(w).Encode(map[string]any{
"apiUrl": apiURL,
"capabilities": map[string]any{"urn:ietf:params:jmap:core": map[string]any{}, "urn:stalwart:jmap": map[string]any{}},
})
case r.Method == http.MethodPost && r.URL.Path == "/api":
var body map[string]any
json.NewDecoder(r.Body).Decode(&body)
methodCalls := body["methodCalls"].([]any)
name := methodCalls[0].([]any)[0].(string)
switch name {
case "x:Account/query":
json.NewEncoder(w).Encode(map[string]any{"methodResponses": []any{
[]any{"x:Account/query", map[string]any{"ids": []string{"a1"}}, "q"},
}})
case "x:Account/get":
json.NewEncoder(w).Encode(map[string]any{"methodResponses": []any{
[]any{"x:Account/get", map[string]any{"list": []map[string]any{
{"id": "a1", "name": "[email protected]", "domainId": "example.com"},
}}, "g"},
}})
case "Mailbox/get":
json.NewEncoder(w).Encode(map[string]any{"methodResponses": []any{
[]any{"Mailbox/get", map[string]any{"list": []map[string]any{
{"name": "Inbox", "totalEmails": 5},
}}, "m"},
}})
}
default:
w.WriteHeader(http.StatusNotFound)
}
}))
apiURL = adminSrv.URL + "/api"
defer adminSrv.Close()
store := checkpoint.NewStore(t.TempDir())
rs, err := store.Create("", "latest")
if err != nil {
t.Fatal(err)
}
checker := New(Options{
BinaryPath: binaryPath, ConfigPath: configPath, DataDir: dataDir, TargetVersion: "latest",
AdminURL: adminSrv.URL, AdminUser: "admin", AdminPassword: "hunter2",
})
report, err := checker.Run(context.Background(), store, rs)
if err != nil {
t.Fatalf("Run: %v", err)
}
if report.Blocking() {
t.Fatalf("unexpected blocking report:\n%s", report.String())
}
if rs.PreflightSnapshot == nil {
t.Fatal("PreflightSnapshot should be populated when AdminURL is set")
}
if rs.PreflightSnapshot.AccountCount != 1 {
t.Errorf("AccountCount = %d, want 1", rs.PreflightSnapshot.AccountCount)
}
if len(rs.PreflightSnapshot.Domains) != 1 || rs.PreflightSnapshot.Domains[0] != "example.com" {
t.Errorf("Domains = %v, want [example.com]", rs.PreflightSnapshot.Domains)
}
aliceMailboxes := rs.PreflightSnapshot.MailboxCounts["[email protected]"]
if len(aliceMailboxes) != 1 || aliceMailboxes[0].Mailbox != "Inbox" || aliceMailboxes[0].Messages != 5 {
t.Errorf("alice's mailbox counts = %+v, want [{Inbox 5}]", aliceMailboxes)
}
// Resume: the snapshot must survive a reload from disk without
// re-running the check (the admin server would still work, but this
// confirms the persisted value is what's actually being relied on).
resumed, err := store.Load(rs.RunID)
if err != nil {
t.Fatalf("store.Load: %v", err)
}
if resumed.PreflightSnapshot == nil || resumed.PreflightSnapshot.AccountCount != 1 {
t.Errorf("resumed PreflightSnapshot = %+v, want AccountCount 1", resumed.PreflightSnapshot)
}
}