| 1 | //go:build linux |
| 2 | |
| 3 | package layout |
| 4 | |
| 5 | import ( |
| 6 | "context" |
| 7 | "errors" |
| 8 | "fmt" |
| 9 | "os" |
| 10 | "path/filepath" |
| 11 | "strings" |
| 12 | "testing" |
| 13 | |
| 14 | "github.com/carlosframework/platform/internal/registry" |
| 15 | ) |
| 16 | |
| 17 | // seedAppEntries writes what an app makes for itself beside its flat |
| 18 | // database (discussion #256): a directory named for the host holding its |
| 19 | // key, a `-blobs` directory, and a dot-suffixed file. |
| 20 | func seedAppEntries(t *testing.T, dd, h string) { |
| 21 | t.Helper() |
| 22 | for path, body := range map[string]string{ |
| 23 | filepath.Join(dd, h, "instance.key"): "key " + h, |
| 24 | filepath.Join(dd, h, "report.token"): "token " + h, |
| 25 | filepath.Join(dd, h+"-blobs", "ab", "cd"): "blob " + h, |
| 26 | filepath.Join(dd, h+".cache"): "cache " + h, |
| 27 | } { |
| 28 | if err := os.MkdirAll(filepath.Dir(path), 0o700); err != nil { |
| 29 | t.Fatal(err) |
| 30 | } |
| 31 | if err := os.WriteFile(path, []byte(body), 0o600); err != nil { |
| 32 | t.Fatal(err) |
| 33 | } |
| 34 | } |
| 35 | } |
| 36 | |
| 37 | // Discussion #256: the first migration moved the nine platform names and |
| 38 | // left behind /data/<host>/instance.key and /data/<host>-blobs; the app |
| 39 | // started nested without them and minted a new key. Every entry the flat |
| 40 | // naming rules make the host's now moves into the instance directory at |
| 41 | // the same relative name, and comes back on revert. An entry whose name a |
| 42 | // longer host owns (`<host>-x` is the host `<host>-x` itself) stays, and so |
| 43 | // does a shared directory no host owns, which is reported. |
| 44 | func TestConvergeMovesAppCreatedEntriesWithTheHost(t *testing.T) { |
| 45 | p, f := box(t) |
| 46 | var logged []string |
| 47 | p.Logf = func(format string, a ...any) { logged = append(logged, fmt.Sprintf(format, a...)) } |
| 48 | seedAppEntries(t, p.DataDir, host) |
| 49 | // A route whose name is host plus "-x": an externally supervised one |
| 50 | // (not activator-managed), so its own entries must be left alone. |
| 51 | decoy := host + "-x" |
| 52 | if err := p.Registry.Upsert(registry.Route{Host: decoy, Kind: "instance", Socket: filepath.Join(p.SocketDir, decoy+".sock")}); err != nil { |
| 53 | t.Fatal(err) |
| 54 | } |
| 55 | seedAppEntries(t, p.DataDir, decoy) |
| 56 | // What a flat app with cwd / wrote as ./blobs. |
| 57 | if err := os.MkdirAll(filepath.Join(p.DataDir, "blobs", "aa"), 0o700); err != nil { |
| 58 | t.Fatal(err) |
| 59 | } |
| 60 | |
| 61 | res := byHost(t, p.Converge(context.Background(), ModeNested)) |
| 62 | if res[host].Outcome != "migrated" { |
| 63 | t.Fatalf("%s: %+v", host, res[host]) |
| 64 | } |
| 65 | raw := rawInstanceDir(p.DataDir, host) |
| 66 | for rel, want := range map[string]string{ |
| 67 | filepath.Join(host, "instance.key"): "key " + host, |
| 68 | filepath.Join(host, "report.token"): "token " + host, |
| 69 | filepath.Join(host+"-blobs", "ab", "cd"): "blob " + host, |
| 70 | host + ".cache": "cache " + host, |
| 71 | } { |
| 72 | if b, err := os.ReadFile(filepath.Join(raw, rel)); err != nil || string(b) != want { |
| 73 | t.Errorf("%s not in the instance directory: %q %v", rel, b, err) |
| 74 | } |
| 75 | if _, err := os.Lstat(filepath.Join(p.DataDir, rel)); !errors.Is(err, os.ErrNotExist) { |
| 76 | t.Errorf("%s left behind in the data directory: %v", rel, err) |
| 77 | } |
| 78 | } |
| 79 | for _, rel := range []string{filepath.Join(decoy, "instance.key"), filepath.Join(decoy+"-blobs", "ab", "cd"), decoy + ".cache", filepath.Join("blobs", "aa")} { |
| 80 | if _, err := os.Lstat(filepath.Join(p.DataDir, rel)); err != nil { |
| 81 | t.Errorf("%s, not this host's, was moved: %v", rel, err) |
| 82 | } |
| 83 | if _, err := os.Lstat(filepath.Join(raw, rel)); !errors.Is(err, os.ErrNotExist) { |
| 84 | t.Errorf("%s, not this host's, is in the instance directory: %v", rel, err) |
| 85 | } |
| 86 | } |
| 87 | if JournalPresent(p.DataDir, host) || f.fences[host] != 0 { |
| 88 | t.Fatal("journal or fence left after the move") |
| 89 | } |
| 90 | if n := p.Note(host); !strings.Contains(n, "moved with it: "+host+", "+host+"-blobs, "+host+".cache") { |
| 91 | t.Errorf("note = %q", n) |
| 92 | } |
| 93 | // The aside directory holds nothing: these were moved, not quarantined. |
| 94 | if q, err := os.ReadDir(asideDir(p.DataDir, host)); err == nil && len(q) > 0 { |
| 95 | t.Errorf("quarantined instead of moved: %v", q) |
| 96 | } |
| 97 | joined := strings.Join(logged, "\n") |
| 98 | if !strings.Contains(joined, "owned by no host on this box") || !strings.Contains(joined, filepath.Join(p.DataDir, "blobs")) { |
| 99 | t.Errorf("the shared directory was not reported:\n%s", joined) |
| 100 | } |
| 101 | if strings.Contains(joined, filepath.Join(p.DataDir, decoy)) { |
| 102 | t.Errorf("the decoy host's own directory was reported as unowned:\n%s", joined) |
| 103 | } |
| 104 | if !strings.Contains(joined, "named for another host: "+decoy+" ("+decoy+"'s), "+decoy+"-blobs ("+decoy+"'s)") { |
| 105 | t.Errorf("the decoy host's entries were not reported as left for it:\n%s", joined) |
| 106 | } |
| 107 | |
| 108 | // Revert brings the same names back, and only those. |
| 109 | res = byHost(t, p.Converge(context.Background(), ModeFlat)) |
| 110 | if res[host].Outcome != "reverted" { |
| 111 | t.Fatalf("revert %s: %+v", host, res[host]) |
| 112 | } |
| 113 | assertFlatUntouched(t, p, host) |
| 114 | for rel, want := range map[string]string{ |
| 115 | filepath.Join(host, "instance.key"): "key " + host, |
| 116 | filepath.Join(host+"-blobs", "ab", "cd"): "blob " + host, |
| 117 | host + ".cache": "cache " + host, |
| 118 | } { |
| 119 | if b, err := os.ReadFile(filepath.Join(p.DataDir, rel)); err != nil || string(b) != want { |
| 120 | t.Errorf("%s not back in the data directory: %q %v", rel, b, err) |
| 121 | } |
| 122 | } |
| 123 | if entries, _ := os.ReadDir(raw); len(entries) != 0 { |
| 124 | t.Errorf("instance directory not emptied by the revert: %v", entries) |
| 125 | } |
| 126 | } |
| 127 | |
| 128 | // The same protections the platform names get, on the way in: a file |
| 129 | // inside an app's directory that has another name OUTSIDE it would be |
| 130 | // bound read-write into the namespace while something else still reaches |
| 131 | // it, so the move is refused with the inode and the find, and the host is |
| 132 | // held (a tenant can plant exactly this). A hard link whose every name is |
| 133 | // inside the directory is the app's own deduplication and moves. |
| 134 | func TestConvergeRefusesAnAppEntryWithANameOutsideIt(t *testing.T) { |
| 135 | p, f := box(t) |
| 136 | seedAppEntries(t, p.DataDir, host) |
| 137 | secret := filepath.Join(p.DataDir, "litestream.env") |
| 138 | if err := os.WriteFile(secret, []byte("AWS_SECRET_ACCESS_KEY=x\n"), 0o600); err != nil { |
| 139 | t.Fatal(err) |
| 140 | } |
| 141 | if err := os.Link(secret, filepath.Join(p.DataDir, host, "creds")); err != nil { |
| 142 | t.Skipf("hard links unavailable here: %v", err) |
| 143 | } |
| 144 | // And a legitimate pair, both names inside the blob store. |
| 145 | if err := os.Link(filepath.Join(p.DataDir, host+"-blobs", "ab", "cd"), filepath.Join(p.DataDir, host+"-blobs", "dup")); err != nil { |
| 146 | t.Fatal(err) |
| 147 | } |
| 148 | res := byHost(t, p.Converge(context.Background(), ModeNested)) |
| 149 | r := res[host] |
| 150 | if r.Outcome != "held" || !strings.Contains(r.Reason, "creds") || !strings.Contains(r.Reason, "2 names") || !strings.Contains(r.Reason, "only 1 of them are inside") || !strings.Contains(r.Reason, "-samefile") { |
| 151 | t.Fatalf("linked outside: %+v", r) |
| 152 | } |
| 153 | assertFlatUntouched(t, p, host) |
| 154 | if f.fences[host] == 0 { |
| 155 | t.Fatal("the fence was lifted; the host would wake flat with the link in place") |
| 156 | } |
| 157 | if _, err := os.Stat(filepath.Join(p.DataDir, host, "instance.key")); err != nil { |
| 158 | t.Fatalf("a refused move touched the app's directory: %v", err) |
| 159 | } |
| 160 | // The operator removes the outside name; the inside pair is fine. |
| 161 | os.Remove(filepath.Join(p.DataDir, host, "creds")) |
| 162 | res = byHost(t, p.Converge(context.Background(), ModeNested)) |
| 163 | if res[host].Outcome != "migrated" || f.fences[host] != 0 { |
| 164 | t.Fatalf("after the link is gone: %+v fence=%d", res[host], f.fences[host]) |
| 165 | } |
| 166 | raw := rawInstanceDir(p.DataDir, host) |
| 167 | a, err1 := os.Stat(filepath.Join(raw, host+"-blobs", "ab", "cd")) |
| 168 | b, err2 := os.Stat(filepath.Join(raw, host+"-blobs", "dup")) |
| 169 | if err1 != nil || err2 != nil || !os.SameFile(a, b) { |
| 170 | t.Fatalf("the app's own hard-link pair did not move together: %v %v", err1, err2) |
| 171 | } |
| 172 | } |
| 173 | |
| 174 | // A symlink inside an app's directory is refused like one anywhere else |
| 175 | // the pass moves, and the host is held. |
| 176 | func TestConvergeRefusesASymlinkInsideAnAppEntry(t *testing.T) { |
| 177 | p, f := box(t) |
| 178 | seedAppEntries(t, p.DataDir, host) |
| 179 | if err := os.Symlink("/etc/passwd", filepath.Join(p.DataDir, host, "link")); err != nil { |
| 180 | t.Fatal(err) |
| 181 | } |
| 182 | res := byHost(t, p.Converge(context.Background(), ModeNested)) |
| 183 | if r := res[host]; r.Outcome != "held" || !strings.Contains(r.Reason, "link") || !strings.Contains(r.Reason, "not a regular file") { |
| 184 | t.Fatalf("symlink: %+v", r) |
| 185 | } |
| 186 | if f.fences[host] == 0 { |
| 187 | t.Fatal("the fence was lifted") |
| 188 | } |
| 189 | assertFlatUntouched(t, p, host) |
| 190 | } |
| 191 | |
| 192 | // An extra whose destination is already taken is refused before the |
| 193 | // journal, naming both, never a choice. On the way to nested the |
| 194 | // destination check has just cleared the instance directory, so a taken |
| 195 | // destination was planted since (here: inside the serialized section, |
| 196 | // after the check and before the scan), and the host is held. |
| 197 | func TestConvergeRefusesAnAppEntryWhoseDestinationIsTaken(t *testing.T) { |
| 198 | p, f := box(t) |
| 199 | seedAppEntries(t, p.DataDir, host) |
| 200 | p.Serialize = func(fn func()) { |
| 201 | if !JournalPresent(p.DataDir, host) { |
| 202 | os.MkdirAll(filepath.Join(rawInstanceDir(p.DataDir, host), host), 0o700) |
| 203 | } |
| 204 | fn() |
| 205 | } |
| 206 | res := byHost(t, p.Converge(context.Background(), ModeNested)) |
| 207 | r := res[host] |
| 208 | if r.Outcome != "held" || !strings.Contains(r.Reason, "exists at both") || !strings.Contains(r.Reason, "refusing to choose") { |
| 209 | t.Fatalf("taken destination: %+v", r) |
| 210 | } |
| 211 | if f.fences[host] == 0 { |
| 212 | t.Fatal("the fence was lifted") |
| 213 | } |
| 214 | assertFlatUntouched(t, p, host) |
| 215 | if _, err := os.Stat(filepath.Join(p.DataDir, host, "instance.key")); err != nil { |
| 216 | t.Fatalf("a refused move touched the app's directory: %v", err) |
| 217 | } |
| 218 | } |
| 219 | |
| 220 | // Recovery completes a move with extras from its journal: the names are |
| 221 | // recorded, re-derived on recovery, and moved with the rest. |
| 222 | func TestRecoverCompletesAMoveWithAppEntries(t *testing.T) { |
| 223 | for _, step := range []string{"after-journal", "mid-files", "after-files"} { |
| 224 | t.Run(step, func(t *testing.T) { |
| 225 | p, _ := box(t) |
| 226 | seedAppEntries(t, p.DataDir, host) |
| 227 | p.abortAt = step |
| 228 | p.Converge(context.Background(), ModeNested) |
| 229 | p.abortAt = "" |
| 230 | j, found, err := ReadJournal(p.DataDir, host) |
| 231 | if err != nil || !found || len(j.Extra) != 3 { |
| 232 | t.Fatalf("journal = %+v %v %v, want three extras recorded", j, found, err) |
| 233 | } |
| 234 | res := byHost(t, p.Recover(context.Background(), ModeNested)) |
| 235 | if res[host].Outcome != "recovered" { |
| 236 | t.Fatalf("recover = %+v", res[host]) |
| 237 | } |
| 238 | raw := rawInstanceDir(p.DataDir, host) |
| 239 | for _, rel := range []string{filepath.Join(host, "instance.key"), filepath.Join(host+"-blobs", "ab", "cd"), host + ".cache"} { |
| 240 | if _, err := os.Stat(filepath.Join(raw, rel)); err != nil { |
| 241 | t.Errorf("%s not at destination after recovery from %s: %v", rel, step, err) |
| 242 | } |
| 243 | if _, err := os.Lstat(filepath.Join(p.DataDir, rel)); !errors.Is(err, os.ErrNotExist) { |
| 244 | t.Errorf("%s left behind after recovery from %s", rel, step) |
| 245 | } |
| 246 | } |
| 247 | if JournalPresent(p.DataDir, host) { |
| 248 | t.Error("journal remains") |
| 249 | } |
| 250 | }) |
| 251 | } |
| 252 | } |
| 253 | |
| 254 | // A forged journal for host H that names H-x, the directory of the host |
| 255 | // H-x itself, is refused on recovery: the name passes the shape rule, but |
| 256 | // ownership is judged against the box's rows as they stand, and it is not |
| 257 | // H's. The host stays fenced, nothing moves. |
| 258 | func TestRecoverRefusesAJournalNamingAnotherHostsEntry(t *testing.T) { |
| 259 | p, _ := box(t) |
| 260 | decoy := host + "-x" |
| 261 | if err := p.Registry.Upsert(registry.Route{Host: decoy, Kind: "instance", Socket: filepath.Join(p.SocketDir, decoy+".sock")}); err != nil { |
| 262 | t.Fatal(err) |
| 263 | } |
| 264 | seedAppEntries(t, p.DataDir, decoy) |
| 265 | plan, err := PlanFor(p.DataDir, p.SocketDir, host, ModeNested) |
| 266 | if err != nil { |
| 267 | t.Fatal(err) |
| 268 | } |
| 269 | present := make([]bool, len(plan.Artifacts)) |
| 270 | present[0] = true |
| 271 | j := plan.Journal(present) |
| 272 | j.Extra = []JournalArtifact{{Name: decoy, Present: true}} |
| 273 | if err := WriteJournal(p.DataDir, j); err != nil { |
| 274 | t.Fatal(err) |
| 275 | } |
| 276 | res := byHost(t, p.Recover(context.Background(), ModeNested)) |
| 277 | if r := res[host]; r.Outcome != "refused" || !strings.Contains(r.Reason, "belongs to "+decoy) || !strings.HasSuffix(r.Reason, "not acting") { |
| 278 | t.Fatalf("forged extra: %+v", r) |
| 279 | } |
| 280 | if _, err := os.Stat(filepath.Join(p.DataDir, decoy, "instance.key")); err != nil { |
| 281 | t.Fatalf("the other host's directory was moved: %v", err) |
| 282 | } |
| 283 | if !JournalPresent(p.DataDir, host) { |
| 284 | t.Fatal("the refused journal was removed") |
| 285 | } |
| 286 | } |
| 287 | |
| 288 | // Review (#256 B): the hosts migrated before the pass carried the app's |
| 289 | // entries along (the 104 on IE) still have them in the data directory, |
| 290 | // where the namespace cannot see them. The pass never moves a nested |
| 291 | // host's files, but it says so on every pass, in the log and in the |
| 292 | // host's status note, until an operator has moved them. |
| 293 | func TestConvergeReportsFlatLeftoversOfANestedHost(t *testing.T) { |
| 294 | p, _ := box(t) |
| 295 | if res := byHost(t, p.Converge(context.Background(), ModeNested)); res[host].Outcome != "migrated" { |
| 296 | t.Fatalf("setup: %+v", res[host]) |
| 297 | } |
| 298 | // What an earlier release left behind: the app's directory, and a |
| 299 | // platform-named file minted at the flat path since. |
| 300 | seedAppEntries(t, p.DataDir, host) |
| 301 | if err := os.WriteFile(filepath.Join(p.DataDir, host+".env"), []byte("CARLOS_ADMIN_TOKEN=old\n"), 0o600); err != nil { |
| 302 | t.Fatal(err) |
| 303 | } |
| 304 | var logged []string |
| 305 | p.Logf = func(f string, a ...any) { logged = append(logged, fmt.Sprintf(f, a...)) } |
| 306 | res := byHost(t, p.Converge(context.Background(), ModeNested)) |
| 307 | if res[host].Outcome != "unchanged" { |
| 308 | t.Fatalf("%+v", res[host]) |
| 309 | } |
| 310 | want := "still in " + p.DataDir + " from before its move, where the instance cannot see them: " + |
| 311 | filepath.Join(p.DataDir, host) + ", " + filepath.Join(p.DataDir, host+"-blobs") + ", " + filepath.Join(p.DataDir, host+".cache") + ", " + filepath.Join(p.DataDir, host+".env") + |
| 312 | ". move what the app needs into " + rawInstanceDir(p.DataDir, host) + " by hand" |
| 313 | if j := strings.Join(logged, "\n"); !strings.Contains(j, "layout: "+host+": "+want) { |
| 314 | t.Errorf("log:\n%s\nwant a line with:\n%s", j, want) |
| 315 | } |
| 316 | if n := p.Note(host); n != want { |
| 317 | t.Errorf("note = %q", n) |
| 318 | } |
| 319 | for _, rel := range []string{filepath.Join(host, "instance.key"), host + ".env"} { |
| 320 | if _, err := os.Stat(filepath.Join(p.DataDir, rel)); err != nil { |
| 321 | t.Errorf("a report moved %s: %v", rel, err) |
| 322 | } |
| 323 | } |
| 324 | if n := p.Note(replica); strings.Contains(n, "still in") { |
| 325 | t.Errorf("the replica, with nothing left behind, got the note: %q", n) |
| 326 | } |
| 327 | } |
| 328 | |
amadan