1 /* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21 22 /* 23 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. 24 * Copyright (c) 2012 by Delphix. All rights reserved. 25 * Copyright (c) 2013 Steven Hartland. All rights reserved. 26 */ 27 28 /* 29 * This file contains the functions which analyze the status of a pool. This 30 * include both the status of an active pool, as well as the status exported 31 * pools. Returns one of the ZPOOL_STATUS_* defines describing the status of 32 * the pool. This status is independent (to a certain degree) from the state of 33 * the pool. A pool's state describes only whether or not it is capable of 34 * providing the necessary fault tolerance for data. The status describes the 35 * overall status of devices. A pool that is online can still have a device 36 * that is experiencing errors. 37 * 38 * Only a subset of the possible faults can be detected using 'zpool status', 39 * and not all possible errors correspond to a FMA message ID. The explanation 40 * is left up to the caller, depending on whether it is a live pool or an 41 * import. 42 */ 43 44 #include <libzfs.h> 45 #include <string.h> 46 #include <unistd.h> 47 #include "libzfs_impl.h" 48 #include "zfeature_common.h" 49 50 /* 51 * Message ID table. This must be kept in sync with the ZPOOL_STATUS_* defines 52 * in libzfs.h. Note that there are some status results which go past the end 53 * of this table, and hence have no associated message ID. 54 */ 55 static char *zfs_msgid_table[] = { 56 "ZFS-8000-14", /* ZPOOL_STATUS_CORRUPT_CACHE */ 57 "ZFS-8000-2Q", /* ZPOOL_STATUS_MISSING_DEV_R */ 58 "ZFS-8000-3C", /* ZPOOL_STATUS_MISSING_DEV_NR */ 59 "ZFS-8000-4J", /* ZPOOL_STATUS_CORRUPT_LABEL_R */ 60 "ZFS-8000-5E", /* ZPOOL_STATUS_CORRUPT_LABEL_NR */ 61 "ZFS-8000-6X", /* ZPOOL_STATUS_BAD_GUID_SUM */ 62 "ZFS-8000-72", /* ZPOOL_STATUS_CORRUPT_POOL */ 63 "ZFS-8000-8A", /* ZPOOL_STATUS_CORRUPT_DATA */ 64 "ZFS-8000-9P", /* ZPOOL_STATUS_FAILING_DEV */ 65 "ZFS-8000-A5", /* ZPOOL_STATUS_VERSION_NEWER */ 66 "ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_MISMATCH */ 67 "ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_ACTIVE */ 68 "ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_REQUIRED */ 69 "ZFS-8000-HC", /* ZPOOL_STATUS_IO_FAILURE_WAIT */ 70 "ZFS-8000-JQ", /* ZPOOL_STATUS_IO_FAILURE_CONTINUE */ 71 "ZFS-8000-MM", /* ZPOOL_STATUS_IO_FAILURE_MMP */ 72 "ZFS-8000-K4", /* ZPOOL_STATUS_BAD_LOG */ 73 /* 74 * The following results have no message ID. 75 * ZPOOL_STATUS_UNSUP_FEAT_READ 76 * ZPOOL_STATUS_UNSUP_FEAT_WRITE 77 * ZPOOL_STATUS_FAULTED_DEV_R 78 * ZPOOL_STATUS_FAULTED_DEV_NR 79 * ZPOOL_STATUS_VERSION_OLDER 80 * ZPOOL_STATUS_FEAT_DISABLED 81 * ZPOOL_STATUS_RESILVERING 82 * ZPOOL_STATUS_OFFLINE_DEV 83 * ZPOOL_STATUS_REMOVED_DEV 84 * ZPOOL_STATUS_OK 85 */ 86 }; 87 88 #define NMSGID (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0])) 89 90 /* ARGSUSED */ 91 static int 92 vdev_missing(uint64_t state, uint64_t aux, uint64_t errs) 93 { 94 return (state == VDEV_STATE_CANT_OPEN && 95 aux == VDEV_AUX_OPEN_FAILED); 96 } 97 98 /* ARGSUSED */ 99 static int 100 vdev_faulted(uint64_t state, uint64_t aux, uint64_t errs) 101 { 102 return (state == VDEV_STATE_FAULTED); 103 } 104 105 /* ARGSUSED */ 106 static int 107 vdev_errors(uint64_t state, uint64_t aux, uint64_t errs) 108 { 109 return (state == VDEV_STATE_DEGRADED || errs != 0); 110 } 111 112 /* ARGSUSED */ 113 static int 114 vdev_broken(uint64_t state, uint64_t aux, uint64_t errs) 115 { 116 return (state == VDEV_STATE_CANT_OPEN); 117 } 118 119 /* ARGSUSED */ 120 static int 121 vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs) 122 { 123 return (state == VDEV_STATE_OFFLINE); 124 } 125 126 /* ARGSUSED */ 127 static int 128 vdev_removed(uint64_t state, uint64_t aux, uint64_t errs) 129 { 130 return (state == VDEV_STATE_REMOVED); 131 } 132 133 /* 134 * Detect if any leaf devices that have seen errors or could not be opened. 135 */ 136 static boolean_t 137 find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t)) 138 { 139 nvlist_t **child; 140 vdev_stat_t *vs; 141 uint_t c, children; 142 char *type; 143 144 /* 145 * Ignore problems within a 'replacing' vdev, since we're presumably in 146 * the process of repairing any such errors, and don't want to call them 147 * out again. We'll pick up the fact that a resilver is happening 148 * later. 149 */ 150 verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0); 151 if (strcmp(type, VDEV_TYPE_REPLACING) == 0) 152 return (B_FALSE); 153 154 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child, 155 &children) == 0) { 156 for (c = 0; c < children; c++) 157 if (find_vdev_problem(child[c], func)) 158 return (B_TRUE); 159 } else { 160 verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS, 161 (uint64_t **)&vs, &c) == 0); 162 163 if (func(vs->vs_state, vs->vs_aux, 164 vs->vs_read_errors + 165 vs->vs_write_errors + 166 vs->vs_checksum_errors)) 167 return (B_TRUE); 168 } 169 170 /* 171 * Check any L2 cache devs 172 */ 173 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_L2CACHE, &child, 174 &children) == 0) { 175 for (c = 0; c < children; c++) 176 if (find_vdev_problem(child[c], func)) 177 return (B_TRUE); 178 } 179 180 return (B_FALSE); 181 } 182 183 /* 184 * Active pool health status. 185 * 186 * To determine the status for a pool, we make several passes over the config, 187 * picking the most egregious error we find. In order of importance, we do the 188 * following: 189 * 190 * - Check for a complete and valid configuration 191 * - Look for any faulted or missing devices in a non-replicated config 192 * - Check for any data errors 193 * - Check for any faulted or missing devices in a replicated config 194 * - Look for any devices showing errors 195 * - Check for any resilvering devices 196 * 197 * There can obviously be multiple errors within a single pool, so this routine 198 * only picks the most damaging of all the current errors to report. 199 */ 200 static zpool_status_t 201 check_status(nvlist_t *config, boolean_t isimport, zpool_errata_t *erratap) 202 { 203 nvlist_t *nvroot; 204 vdev_stat_t *vs; 205 pool_scan_stat_t *ps = NULL; 206 uint_t vsc, psc; 207 uint64_t nerr; 208 uint64_t version; 209 uint64_t stateval; 210 uint64_t suspended; 211 uint64_t hostid = 0; 212 uint64_t errata = 0; 213 unsigned long system_hostid = get_system_hostid(); 214 215 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, 216 &version) == 0); 217 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, 218 &nvroot) == 0); 219 verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, 220 (uint64_t **)&vs, &vsc) == 0); 221 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, 222 &stateval) == 0); 223 224 /* 225 * Currently resilvering a vdev 226 */ 227 (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, 228 (uint64_t **)&ps, &psc); 229 if (ps != NULL && ps->pss_func == POOL_SCAN_RESILVER && 230 ps->pss_state == DSS_SCANNING) 231 return (ZPOOL_STATUS_RESILVERING); 232 233 /* 234 * The multihost property is set and the pool may be active. 235 */ 236 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 237 vs->vs_aux == VDEV_AUX_ACTIVE) { 238 mmp_state_t mmp_state; 239 nvlist_t *nvinfo; 240 241 nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); 242 mmp_state = fnvlist_lookup_uint64(nvinfo, 243 ZPOOL_CONFIG_MMP_STATE); 244 245 if (mmp_state == MMP_STATE_ACTIVE) 246 return (ZPOOL_STATUS_HOSTID_ACTIVE); 247 else if (mmp_state == MMP_STATE_NO_HOSTID) 248 return (ZPOOL_STATUS_HOSTID_REQUIRED); 249 else 250 return (ZPOOL_STATUS_HOSTID_MISMATCH); 251 } 252 253 /* 254 * Pool last accessed by another system. 255 */ 256 (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid); 257 if (hostid != 0 && (unsigned long)hostid != system_hostid && 258 stateval == POOL_STATE_ACTIVE) 259 return (ZPOOL_STATUS_HOSTID_MISMATCH); 260 261 /* 262 * Newer on-disk version. 263 */ 264 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 265 vs->vs_aux == VDEV_AUX_VERSION_NEWER) 266 return (ZPOOL_STATUS_VERSION_NEWER); 267 268 /* 269 * Unsupported feature(s). 270 */ 271 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 272 vs->vs_aux == VDEV_AUX_UNSUP_FEAT) { 273 nvlist_t *nvinfo; 274 275 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, 276 &nvinfo) == 0); 277 if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY)) 278 return (ZPOOL_STATUS_UNSUP_FEAT_WRITE); 279 return (ZPOOL_STATUS_UNSUP_FEAT_READ); 280 } 281 282 /* 283 * Check that the config is complete. 284 */ 285 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 286 vs->vs_aux == VDEV_AUX_BAD_GUID_SUM) 287 return (ZPOOL_STATUS_BAD_GUID_SUM); 288 289 /* 290 * Check whether the pool has suspended. 291 */ 292 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED, 293 &suspended) == 0) { 294 uint64_t reason; 295 296 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED_REASON, 297 &reason) == 0 && reason == ZIO_SUSPEND_MMP) 298 return (ZPOOL_STATUS_IO_FAILURE_MMP); 299 300 if (suspended == ZIO_FAILURE_MODE_CONTINUE) 301 return (ZPOOL_STATUS_IO_FAILURE_CONTINUE); 302 return (ZPOOL_STATUS_IO_FAILURE_WAIT); 303 } 304 305 /* 306 * Could not read a log. 307 */ 308 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 309 vs->vs_aux == VDEV_AUX_BAD_LOG) { 310 return (ZPOOL_STATUS_BAD_LOG); 311 } 312 313 /* 314 * Bad devices in non-replicated config. 315 */ 316 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 317 find_vdev_problem(nvroot, vdev_faulted)) 318 return (ZPOOL_STATUS_FAULTED_DEV_NR); 319 320 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 321 find_vdev_problem(nvroot, vdev_missing)) 322 return (ZPOOL_STATUS_MISSING_DEV_NR); 323 324 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 325 find_vdev_problem(nvroot, vdev_broken)) 326 return (ZPOOL_STATUS_CORRUPT_LABEL_NR); 327 328 /* 329 * Corrupted pool metadata 330 */ 331 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 332 vs->vs_aux == VDEV_AUX_CORRUPT_DATA) 333 return (ZPOOL_STATUS_CORRUPT_POOL); 334 335 /* 336 * Persistent data errors. 337 */ 338 if (!isimport) { 339 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT, 340 &nerr) == 0 && nerr != 0) 341 return (ZPOOL_STATUS_CORRUPT_DATA); 342 } 343 344 /* 345 * Missing devices in a replicated config. 346 */ 347 if (find_vdev_problem(nvroot, vdev_faulted)) 348 return (ZPOOL_STATUS_FAULTED_DEV_R); 349 if (find_vdev_problem(nvroot, vdev_missing)) 350 return (ZPOOL_STATUS_MISSING_DEV_R); 351 if (find_vdev_problem(nvroot, vdev_broken)) 352 return (ZPOOL_STATUS_CORRUPT_LABEL_R); 353 354 /* 355 * Devices with errors 356 */ 357 if (!isimport && find_vdev_problem(nvroot, vdev_errors)) 358 return (ZPOOL_STATUS_FAILING_DEV); 359 360 /* 361 * Offlined devices 362 */ 363 if (find_vdev_problem(nvroot, vdev_offlined)) 364 return (ZPOOL_STATUS_OFFLINE_DEV); 365 366 /* 367 * Removed device 368 */ 369 if (find_vdev_problem(nvroot, vdev_removed)) 370 return (ZPOOL_STATUS_REMOVED_DEV); 371 372 /* 373 * Informational errata available. 374 */ 375 (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRATA, &errata); 376 if (errata) { 377 *erratap = errata; 378 return (ZPOOL_STATUS_ERRATA); 379 } 380 381 /* 382 * Outdated, but usable, version 383 */ 384 if (SPA_VERSION_IS_SUPPORTED(version) && version != SPA_VERSION) 385 return (ZPOOL_STATUS_VERSION_OLDER); 386 387 /* 388 * Usable pool with disabled features 389 */ 390 if (version >= SPA_VERSION_FEATURES) { 391 int i; 392 nvlist_t *feat; 393 394 if (isimport) { 395 feat = fnvlist_lookup_nvlist(config, 396 ZPOOL_CONFIG_LOAD_INFO); 397 if (nvlist_exists(feat, ZPOOL_CONFIG_ENABLED_FEAT)) 398 feat = fnvlist_lookup_nvlist(feat, 399 ZPOOL_CONFIG_ENABLED_FEAT); 400 } else { 401 feat = fnvlist_lookup_nvlist(config, 402 ZPOOL_CONFIG_FEATURE_STATS); 403 } 404 405 for (i = 0; i < SPA_FEATURES; i++) { 406 zfeature_info_t *fi = &spa_feature_table[i]; 407 if (!nvlist_exists(feat, fi->fi_guid)) 408 return (ZPOOL_STATUS_FEAT_DISABLED); 409 } 410 } 411 412 return (ZPOOL_STATUS_OK); 413 } 414 415 zpool_status_t 416 zpool_get_status(zpool_handle_t *zhp, char **msgid, zpool_errata_t *errata) 417 { 418 zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE, errata); 419 420 if (ret >= NMSGID) 421 *msgid = NULL; 422 else 423 *msgid = zfs_msgid_table[ret]; 424 425 return (ret); 426 } 427 428 zpool_status_t 429 zpool_import_status(nvlist_t *config, char **msgid, zpool_errata_t *errata) 430 { 431 zpool_status_t ret = check_status(config, B_TRUE, errata); 432 433 if (ret >= NMSGID) 434 *msgid = NULL; 435 else 436 *msgid = zfs_msgid_table[ret]; 437 438 return (ret); 439 } 440 441 static void 442 dump_ddt_stat(const ddt_stat_t *dds, int h) 443 { 444 char refcnt[6]; 445 char blocks[6], lsize[6], psize[6], dsize[6]; 446 char ref_blocks[6], ref_lsize[6], ref_psize[6], ref_dsize[6]; 447 448 if (dds == NULL || dds->dds_blocks == 0) 449 return; 450 451 if (h == -1) 452 (void) strcpy(refcnt, "Total"); 453 else 454 zfs_nicenum(1ULL << h, refcnt, sizeof (refcnt)); 455 456 zfs_nicenum(dds->dds_blocks, blocks, sizeof (blocks)); 457 zfs_nicenum(dds->dds_lsize, lsize, sizeof (lsize)); 458 zfs_nicenum(dds->dds_psize, psize, sizeof (psize)); 459 zfs_nicenum(dds->dds_dsize, dsize, sizeof (dsize)); 460 zfs_nicenum(dds->dds_ref_blocks, ref_blocks, sizeof (ref_blocks)); 461 zfs_nicenum(dds->dds_ref_lsize, ref_lsize, sizeof (ref_lsize)); 462 zfs_nicenum(dds->dds_ref_psize, ref_psize, sizeof (ref_psize)); 463 zfs_nicenum(dds->dds_ref_dsize, ref_dsize, sizeof (ref_dsize)); 464 465 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 466 refcnt, 467 blocks, lsize, psize, dsize, 468 ref_blocks, ref_lsize, ref_psize, ref_dsize); 469 } 470 471 /* 472 * Print the DDT histogram and the column totals. 473 */ 474 void 475 zpool_dump_ddt(const ddt_stat_t *dds_total, const ddt_histogram_t *ddh) 476 { 477 int h; 478 479 (void) printf("\n"); 480 481 (void) printf("bucket " 482 " allocated " 483 " referenced \n"); 484 (void) printf("______ " 485 "______________________________ " 486 "______________________________\n"); 487 488 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 489 "refcnt", 490 "blocks", "LSIZE", "PSIZE", "DSIZE", 491 "blocks", "LSIZE", "PSIZE", "DSIZE"); 492 493 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 494 "------", 495 "------", "-----", "-----", "-----", 496 "------", "-----", "-----", "-----"); 497 498 for (h = 0; h < 64; h++) 499 dump_ddt_stat(&ddh->ddh_stat[h], h); 500 501 dump_ddt_stat(dds_total, -1); 502 503 (void) printf("\n"); 504 } 505