| 1 | package host |
| 2 | |
| 3 | import ( |
| 4 | "context" |
| 5 | "encoding/json" |
| 6 | "errors" |
| 7 | "os" |
| 8 | "os/exec" |
| 9 | "path/filepath" |
| 10 | "testing" |
| 11 | "time" |
| 12 | |
| 13 | "github.com/carlosframework/platform/internal/hostbox" |
| 14 | "github.com/carlosframework/platform/internal/hostclaim" |
| 15 | "github.com/carlosframework/platform/internal/hostguard" |
| 16 | "github.com/carlosframework/platform/internal/hostwrite" |
| 17 | "github.com/carlosframework/platform/internal/registry" |
| 18 | "github.com/carlosframework/platform/internal/release" |
| 19 | "github.com/carlosframework/platform/internal/rootparticipation" |
| 20 | ) |
| 21 | |
| 22 | type restartRequestProcessFixture struct { |
| 23 | Unit rollbackProcessFixture |
| 24 | Claim *hostclaim.Claim |
| 25 | Installations, Journal string |
| 26 | } |
| 27 | |
| 28 | func TestPlan14RootRestartProductionCrash(t *testing.T) { |
| 29 | t.Parallel() |
| 30 | if boundary := os.Getenv("PLAN14_REQUEST_CRASH"); boundary != "" { |
| 31 | dir := os.Getenv("PLAN14_REQUEST_DIR") |
| 32 | f := newRestartConsumptionFixture(t) |
| 33 | if err := f.run(t); err != nil { |
| 34 | t.Fatal(err) |
| 35 | } |
| 36 | f.request(t, release.RestartScopeHost, observedHost) |
| 37 | storeDir := filepath.Join(dir, "deployment") |
| 38 | store := release.NewDirStore(storeDir) |
| 39 | keys, err := f.e.store.List(t.Context(), "") |
| 40 | if err != nil { |
| 41 | t.Fatal(err) |
| 42 | } |
| 43 | for _, key := range keys { |
| 44 | raw, _, err := f.e.store.Get(t.Context(), key) |
| 45 | if err != nil { |
| 46 | t.Fatal(err) |
| 47 | } |
| 48 | if _, err := store.Put(t.Context(), key, raw, ""); err != nil { |
| 49 | t.Fatal(err) |
| 50 | } |
| 51 | } |
| 52 | crash := func() { |
| 53 | state := restartRequestProcessFixture{Unit: rollbackProcessFixture{Props: f.u.props, Exe: f.u.exe, Opt: f.u.opt, Bake: f.e.bake, Store: storeDir, Effects: f.u.effects}, Claim: f.claim, |
| 54 | Installations: f.p.source.state.(*rollbackDisk).dir, Journal: f.p.journal.Store.(*rollbackDisk).dir} |
| 55 | raw, err := json.Marshal(state) |
| 56 | if err != nil { |
| 57 | t.Fatal(err) |
| 58 | } |
| 59 | if err := os.WriteFile(filepath.Join(dir, "fixture.json"), raw, 0600); err != nil { |
| 60 | t.Fatal(err) |
| 61 | } |
| 62 | os.Exit(74) |
| 63 | } |
| 64 | disk := f.p.journal.Store.(*rollbackDisk) |
| 65 | classify := func(raw []byte) string { |
| 66 | var r struct { |
| 67 | Pending *release.RestartIntent |
| 68 | ScopeID string `json:"scope_id"` |
| 69 | } |
| 70 | if json.Unmarshal(raw, &r) != nil { |
| 71 | return "" |
| 72 | } |
| 73 | if r.Pending == nil { |
| 74 | if r.ScopeID != release.RestartScopeNone { |
| 75 | return "consumed" |
| 76 | } |
| 77 | return "baseline" |
| 78 | } |
| 79 | if r.Pending.Targets[0].After != "" { |
| 80 | return "completed" |
| 81 | } |
| 82 | if r.Pending.Targets[0].Before != "" { |
| 83 | return "intent" |
| 84 | } |
| 85 | return "pending" |
| 86 | } |
| 87 | disk.afterWrite = func(_ string, raw []byte) { |
| 88 | if classify(raw) == boundary { |
| 89 | crash() |
| 90 | } |
| 91 | } |
| 92 | if boundary == "intent-visible" { |
| 93 | disk.beforeWrite = func(name string, raw []byte) { |
| 94 | if classify(raw) != "intent" { |
| 95 | return |
| 96 | } |
| 97 | file, err := os.CreateTemp(disk.dir, "visible-") |
| 98 | if err != nil { |
| 99 | t.Fatal(err) |
| 100 | } |
| 101 | if _, err = file.Write(raw); err != nil { |
| 102 | t.Fatal(err) |
| 103 | } |
| 104 | if err = file.Sync(); err != nil { |
| 105 | t.Fatal(err) |
| 106 | } |
| 107 | if err = file.Close(); err != nil { |
| 108 | t.Fatal(err) |
| 109 | } |
| 110 | if err = os.Rename(file.Name(), filepath.Join(disk.dir, name)); err != nil { |
| 111 | t.Fatal(err) |
| 112 | } |
| 113 | // Exit before syncing the parent of the visible rename. |
| 114 | crash() |
| 115 | } |
| 116 | } |
| 117 | f.u.afterEffect = func() { |
| 118 | if boundary == "effect" { |
| 119 | crash() |
| 120 | } |
| 121 | } |
| 122 | _ = f.run(t) |
| 123 | t.Fatal("crash boundary not reached") |
| 124 | } |
| 125 | for _, boundary := range []string{"pending", "intent", "intent-visible", "effect", "completed", "consumed"} { |
| 126 | t.Run(boundary, func(t *testing.T) { |
| 127 | dir := t.TempDir() |
| 128 | ctx, cancel := context.WithTimeout(t.Context(), 30*time.Second) |
| 129 | defer cancel() |
| 130 | cmd := exec.CommandContext(ctx, os.Args[0], "-test.run=^TestPlan14RootRestartProductionCrash$", "-test.count=1") |
| 131 | cmd.Dir = dir |
| 132 | cmd.Env = append(os.Environ(), "TMPDIR="+dir, "PLAN14_REQUEST_DIR="+dir, "PLAN14_REQUEST_CRASH="+boundary, "CARLOS_HOST_CLAIM=") |
| 133 | out, err := cmd.CombinedOutput() |
| 134 | var exit *exec.ExitError |
| 135 | if !errors.As(err, &exit) || exit.ExitCode() != 74 { |
| 136 | t.Fatalf("child: %v\n%s", err, out) |
| 137 | } |
| 138 | raw, err := os.ReadFile(filepath.Join(dir, "fixture.json")) |
| 139 | if err != nil { |
| 140 | t.Fatal(err) |
| 141 | } |
| 142 | var saved restartRequestProcessFixture |
| 143 | if err = json.Unmarshal(raw, &saved); err != nil { |
| 144 | t.Fatal(err) |
| 145 | } |
| 146 | u := &plan14RestartUnits{props: saved.Unit.Props, exe: saved.Unit.Exe, opt: saved.Unit.Opt, effects: saved.Unit.Effects} |
| 147 | store := release.NewDirStore(saved.Unit.Store) |
| 148 | state := &rollbackDisk{dir: saved.Installations} |
| 149 | guard := hostguard.Guard{DataDir: t.TempDir(), LockDir: t.TempDir(), Root: hostguard.NewRootAuthority(state), Claims: rootViewClaimReader{obs: hostclaim.Observed{Host: observedHost, ETag: "claim", Claim: saved.Claim}}} |
| 150 | source := &proofBindingSource{store: store, sys: u, optDir: u.opt, openFile: os.Open, unitDirs: []string{filepath.Dir(u.props.FragmentPath)}, healthy: func(acct, app, version, digest string) *release.HealthyBuild { |
| 151 | for _, b := range release.ReadHealthy(saved.Unit.Bake, acct, app) { |
| 152 | if b.Version == version && b.SHA256 == digest { |
| 153 | return &b |
| 154 | } |
| 155 | } |
| 156 | return nil |
| 157 | }} |
| 158 | now, sleep := settleClock() |
| 159 | p := rootRestartRequests{journal: &release.RootRestarts{Store: &rollbackDisk{dir: saved.Journal}}, source: rootRestartRequestSource{state: state, store: store, census: func() (rootparticipation.Census, error) { |
| 160 | return rootparticipation.Census{Box: "box", Enrollment: "en", Migration: 1, Complete: true}, nil |
| 161 | }}, |
| 162 | sys: u, optDir: u.opt, getenv: func(string) string { return "5s" }, checks: rootRestartChecks{host: source.forHost, ready: func(registry.Route) func() error { return func() error { return nil } }, now: now, sleep: sleep}, |
| 163 | writes: hostwrite.Config{Enabled: true, Selector: hostbox.WorkSource{Guard: guard, Box: "box", Region: "ie", Now: time.Now, Policy: func(_ context.Context, c *hostclaim.Claim, s hostbox.Selection, _ hostwrite.Identity) error { |
| 164 | return hostbox.Serving(c, s) |
| 165 | }, Local: func(context.Context, *hostclaim.Claim, hostbox.Selection) error { return nil }}}} |
| 166 | err = p.run(t.Context()) |
| 167 | if boundary == "intent" || boundary == "intent-visible" { |
| 168 | if err == nil || len(u.effects) != 0 { |
| 169 | t.Fatalf("ambiguous replay: %v %v", err, u.effects) |
| 170 | } |
| 171 | u.props.InvocationID += "reconciled" |
| 172 | u.props.MainPID++ |
| 173 | if err = p.run(t.Context()); err != nil || len(u.effects) != 0 { |
| 174 | t.Fatalf("reconcile: %v %v", err, u.effects) |
| 175 | } |
| 176 | } else if err != nil || len(u.effects) != 1 { |
| 177 | t.Fatalf("resume: %v %v", err, u.effects) |
| 178 | } |
| 179 | count := len(u.effects) |
| 180 | if err = p.run(t.Context()); err != nil || len(u.effects) != count { |
| 181 | t.Fatalf("replay: %v %v", err, u.effects) |
| 182 | } |
| 183 | }) |
| 184 | } |
| 185 | } |
| 186 | |
amadan