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", 57 "ZFS-8000-2Q", 58 "ZFS-8000-3C", 59 "ZFS-8000-4J", 60 "ZFS-8000-5E", 61 "ZFS-8000-6X", 62 "ZFS-8000-72", 63 "ZFS-8000-8A", 64 "ZFS-8000-9P", 65 "ZFS-8000-A5", 66 "ZFS-8000-EY", 67 "ZFS-8000-HC", 68 "ZFS-8000-JQ", 69 "ZFS-8000-K4", 70 }; 71 72 #define NMSGID (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0])) 73 74 /* ARGSUSED */ 75 static int 76 vdev_missing(uint64_t state, uint64_t aux, uint64_t errs) 77 { 78 return (state == VDEV_STATE_CANT_OPEN && 79 aux == VDEV_AUX_OPEN_FAILED); 80 } 81 82 /* ARGSUSED */ 83 static int 84 vdev_faulted(uint64_t state, uint64_t aux, uint64_t errs) 85 { 86 return (state == VDEV_STATE_FAULTED); 87 } 88 89 /* ARGSUSED */ 90 static int 91 vdev_errors(uint64_t state, uint64_t aux, uint64_t errs) 92 { 93 return (state == VDEV_STATE_DEGRADED || errs != 0); 94 } 95 96 /* ARGSUSED */ 97 static int 98 vdev_broken(uint64_t state, uint64_t aux, uint64_t errs) 99 { 100 return (state == VDEV_STATE_CANT_OPEN); 101 } 102 103 /* ARGSUSED */ 104 static int 105 vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs) 106 { 107 return (state == VDEV_STATE_OFFLINE); 108 } 109 110 /* ARGSUSED */ 111 static int 112 vdev_removed(uint64_t state, uint64_t aux, uint64_t errs) 113 { 114 return (state == VDEV_STATE_REMOVED); 115 } 116 117 /* 118 * Detect if any leaf devices that have seen errors or could not be opened. 119 */ 120 static boolean_t 121 find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t)) 122 { 123 nvlist_t **child; 124 vdev_stat_t *vs; 125 uint_t c, children; 126 char *type; 127 128 /* 129 * Ignore problems within a 'replacing' vdev, since we're presumably in 130 * the process of repairing any such errors, and don't want to call them 131 * out again. We'll pick up the fact that a resilver is happening 132 * later. 133 */ 134 verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0); 135 if (strcmp(type, VDEV_TYPE_REPLACING) == 0) 136 return (B_FALSE); 137 138 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child, 139 &children) == 0) { 140 for (c = 0; c < children; c++) 141 if (find_vdev_problem(child[c], func)) 142 return (B_TRUE); 143 } else { 144 verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS, 145 (uint64_t **)&vs, &c) == 0); 146 147 if (func(vs->vs_state, vs->vs_aux, 148 vs->vs_read_errors + 149 vs->vs_write_errors + 150 vs->vs_checksum_errors)) 151 return (B_TRUE); 152 } 153 154 /* 155 * Check any L2 cache devs 156 */ 157 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_L2CACHE, &child, 158 &children) == 0) { 159 for (c = 0; c < children; c++) 160 if (find_vdev_problem(child[c], func)) 161 return (B_TRUE); 162 } 163 164 return (B_FALSE); 165 } 166 167 /* 168 * Active pool health status. 169 * 170 * To determine the status for a pool, we make several passes over the config, 171 * picking the most egregious error we find. In order of importance, we do the 172 * following: 173 * 174 * - Check for a complete and valid configuration 175 * - Look for any faulted or missing devices in a non-replicated config 176 * - Check for any data errors 177 * - Check for any faulted or missing devices in a replicated config 178 * - Look for any devices showing errors 179 * - Check for any resilvering devices 180 * 181 * There can obviously be multiple errors within a single pool, so this routine 182 * only picks the most damaging of all the current errors to report. 183 */ 184 static zpool_status_t 185 check_status(nvlist_t *config, boolean_t isimport) 186 { 187 nvlist_t *nvroot; 188 vdev_stat_t *vs; 189 pool_scan_stat_t *ps = NULL; 190 uint_t vsc, psc; 191 uint64_t nerr; 192 uint64_t version; 193 uint64_t stateval; 194 uint64_t suspended; 195 uint64_t hostid = 0; 196 197 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, 198 &version) == 0); 199 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, 200 &nvroot) == 0); 201 verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, 202 (uint64_t **)&vs, &vsc) == 0); 203 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, 204 &stateval) == 0); 205 206 /* 207 * Currently resilvering a vdev 208 */ 209 (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, 210 (uint64_t **)&ps, &psc); 211 if (ps && ps->pss_func == POOL_SCAN_RESILVER && 212 ps->pss_state == DSS_SCANNING) 213 return (ZPOOL_STATUS_RESILVERING); 214 215 /* 216 * Pool last accessed by another system. 217 */ 218 (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid); 219 if (hostid != 0 && (unsigned long)hostid != gethostid() && 220 stateval == POOL_STATE_ACTIVE) 221 return (ZPOOL_STATUS_HOSTID_MISMATCH); 222 223 /* 224 * Newer on-disk version. 225 */ 226 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 227 vs->vs_aux == VDEV_AUX_VERSION_NEWER) 228 return (ZPOOL_STATUS_VERSION_NEWER); 229 230 /* 231 * Unsupported feature(s). 232 */ 233 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 234 vs->vs_aux == VDEV_AUX_UNSUP_FEAT) { 235 nvlist_t *nvinfo; 236 237 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, 238 &nvinfo) == 0); 239 if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY)) 240 return (ZPOOL_STATUS_UNSUP_FEAT_WRITE); 241 return (ZPOOL_STATUS_UNSUP_FEAT_READ); 242 } 243 244 /* 245 * Check that the config is complete. 246 */ 247 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 248 vs->vs_aux == VDEV_AUX_BAD_GUID_SUM) 249 return (ZPOOL_STATUS_BAD_GUID_SUM); 250 251 /* 252 * Check whether the pool has suspended due to failed I/O. 253 */ 254 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED, 255 &suspended) == 0) { 256 if (suspended == ZIO_FAILURE_MODE_CONTINUE) 257 return (ZPOOL_STATUS_IO_FAILURE_CONTINUE); 258 return (ZPOOL_STATUS_IO_FAILURE_WAIT); 259 } 260 261 /* 262 * Could not read a log. 263 */ 264 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 265 vs->vs_aux == VDEV_AUX_BAD_LOG) { 266 return (ZPOOL_STATUS_BAD_LOG); 267 } 268 269 /* 270 * Bad devices in non-replicated config. 271 */ 272 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 273 find_vdev_problem(nvroot, vdev_faulted)) 274 return (ZPOOL_STATUS_FAULTED_DEV_NR); 275 276 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 277 find_vdev_problem(nvroot, vdev_missing)) 278 return (ZPOOL_STATUS_MISSING_DEV_NR); 279 280 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 281 find_vdev_problem(nvroot, vdev_broken)) 282 return (ZPOOL_STATUS_CORRUPT_LABEL_NR); 283 284 /* 285 * Corrupted pool metadata 286 */ 287 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 288 vs->vs_aux == VDEV_AUX_CORRUPT_DATA) 289 return (ZPOOL_STATUS_CORRUPT_POOL); 290 291 /* 292 * Persistent data errors. 293 */ 294 if (!isimport) { 295 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT, 296 &nerr) == 0 && nerr != 0) 297 return (ZPOOL_STATUS_CORRUPT_DATA); 298 } 299 300 /* 301 * Missing devices in a replicated config. 302 */ 303 if (find_vdev_problem(nvroot, vdev_faulted)) 304 return (ZPOOL_STATUS_FAULTED_DEV_R); 305 if (find_vdev_problem(nvroot, vdev_missing)) 306 return (ZPOOL_STATUS_MISSING_DEV_R); 307 if (find_vdev_problem(nvroot, vdev_broken)) 308 return (ZPOOL_STATUS_CORRUPT_LABEL_R); 309 310 /* 311 * Devices with errors 312 */ 313 if (!isimport && find_vdev_problem(nvroot, vdev_errors)) 314 return (ZPOOL_STATUS_FAILING_DEV); 315 316 /* 317 * Offlined devices 318 */ 319 if (find_vdev_problem(nvroot, vdev_offlined)) 320 return (ZPOOL_STATUS_OFFLINE_DEV); 321 322 /* 323 * Removed device 324 */ 325 if (find_vdev_problem(nvroot, vdev_removed)) 326 return (ZPOOL_STATUS_REMOVED_DEV); 327 328 /* 329 * Outdated, but usable, version 330 */ 331 if (SPA_VERSION_IS_SUPPORTED(version) && version != SPA_VERSION) 332 return (ZPOOL_STATUS_VERSION_OLDER); 333 334 /* 335 * Usable pool with disabled features 336 */ 337 if (version >= SPA_VERSION_FEATURES) { 338 int i; 339 nvlist_t *feat; 340 341 if (isimport) { 342 feat = fnvlist_lookup_nvlist(config, 343 ZPOOL_CONFIG_LOAD_INFO); 344 feat = fnvlist_lookup_nvlist(feat, 345 ZPOOL_CONFIG_ENABLED_FEAT); 346 } else { 347 feat = fnvlist_lookup_nvlist(config, 348 ZPOOL_CONFIG_FEATURE_STATS); 349 } 350 351 for (i = 0; i < SPA_FEATURES; i++) { 352 zfeature_info_t *fi = &spa_feature_table[i]; 353 if (!nvlist_exists(feat, fi->fi_guid)) 354 return (ZPOOL_STATUS_FEAT_DISABLED); 355 } 356 } 357 358 return (ZPOOL_STATUS_OK); 359 } 360 361 zpool_status_t 362 zpool_get_status(zpool_handle_t *zhp, char **msgid) 363 { 364 zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE); 365 366 if (ret >= NMSGID) 367 *msgid = NULL; 368 else 369 *msgid = zfs_msgid_table[ret]; 370 371 return (ret); 372 } 373 374 zpool_status_t 375 zpool_import_status(nvlist_t *config, char **msgid) 376 { 377 zpool_status_t ret = check_status(config, B_TRUE); 378 379 if (ret >= NMSGID) 380 *msgid = NULL; 381 else 382 *msgid = zfs_msgid_table[ret]; 383 384 return (ret); 385 } 386 387 static void 388 dump_ddt_stat(const ddt_stat_t *dds, int h) 389 { 390 char refcnt[6]; 391 char blocks[6], lsize[6], psize[6], dsize[6]; 392 char ref_blocks[6], ref_lsize[6], ref_psize[6], ref_dsize[6]; 393 394 if (dds == NULL || dds->dds_blocks == 0) 395 return; 396 397 if (h == -1) 398 (void) strcpy(refcnt, "Total"); 399 else 400 zfs_nicenum(1ULL << h, refcnt, sizeof (refcnt)); 401 402 zfs_nicenum(dds->dds_blocks, blocks, sizeof (blocks)); 403 zfs_nicenum(dds->dds_lsize, lsize, sizeof (lsize)); 404 zfs_nicenum(dds->dds_psize, psize, sizeof (psize)); 405 zfs_nicenum(dds->dds_dsize, dsize, sizeof (dsize)); 406 zfs_nicenum(dds->dds_ref_blocks, ref_blocks, sizeof (ref_blocks)); 407 zfs_nicenum(dds->dds_ref_lsize, ref_lsize, sizeof (ref_lsize)); 408 zfs_nicenum(dds->dds_ref_psize, ref_psize, sizeof (ref_psize)); 409 zfs_nicenum(dds->dds_ref_dsize, ref_dsize, sizeof (ref_dsize)); 410 411 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 412 refcnt, 413 blocks, lsize, psize, dsize, 414 ref_blocks, ref_lsize, ref_psize, ref_dsize); 415 } 416 417 /* 418 * Print the DDT histogram and the column totals. 419 */ 420 void 421 zpool_dump_ddt(const ddt_stat_t *dds_total, const ddt_histogram_t *ddh) 422 { 423 int h; 424 425 (void) printf("\n"); 426 427 (void) printf("bucket " 428 " allocated " 429 " referenced \n"); 430 (void) printf("______ " 431 "______________________________ " 432 "______________________________\n"); 433 434 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 435 "refcnt", 436 "blocks", "LSIZE", "PSIZE", "DSIZE", 437 "blocks", "LSIZE", "PSIZE", "DSIZE"); 438 439 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n", 440 "------", 441 "------", "-----", "-----", "-----", 442 "------", "-----", "-----", "-----"); 443 444 for (h = 0; h < 64; h++) 445 dump_ddt_stat(&ddh->ddh_stat[h], h); 446 447 dump_ddt_stat(dds_total, -1); 448 449 (void) printf("\n"); 450 } 451