lib/libzfs/libzfs_status.c

   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-EY",
  68         "ZFS-8000-EY",
  69         "ZFS-8000-HC",
  70         "ZFS-8000-JQ",
  71         "ZFS-8000-K4",
  72         "ZFS-8000-ER",
  73 };
  74
  75 #define NMSGID  (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0]))
  76
  77 /* ARGSUSED */
  78 static int
  79 vdev_missing(uint64_t state, uint64_t aux, uint64_t errs)
  80 {
  81         return (state == VDEV_STATE_CANT_OPEN &&
  82             aux == VDEV_AUX_OPEN_FAILED);
  83 }
  84
  85 /* ARGSUSED */
  86 static int
  87 vdev_faulted(uint64_t state, uint64_t aux, uint64_t errs)
  88 {
  89         return (state == VDEV_STATE_FAULTED);
  90 }
  91
  92 /* ARGSUSED */
  93 static int
  94 vdev_errors(uint64_t state, uint64_t aux, uint64_t errs)
  95 {
  96         return (state == VDEV_STATE_DEGRADED || errs != 0);
  97 }
  98
  99 /* ARGSUSED */
 100 static int
 101 vdev_broken(uint64_t state, uint64_t aux, uint64_t errs)
 102 {
 103         return (state == VDEV_STATE_CANT_OPEN);
 104 }
 105
 106 /* ARGSUSED */
 107 static int
 108 vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs)
 109 {
 110         return (state == VDEV_STATE_OFFLINE);
 111 }
 112
 113 /* ARGSUSED */
 114 static int
 115 vdev_removed(uint64_t state, uint64_t aux, uint64_t errs)
 116 {
 117         return (state == VDEV_STATE_REMOVED);
 118 }
 119
 120 /*
 121  * Detect if any leaf devices that have seen errors or could not be opened.
 122  */
 123 static boolean_t
 124 find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t))
 125 {
 126         nvlist_t **child;
 127         vdev_stat_t *vs;
 128         uint_t c, children;
 129         char *type;
 130
 131         /*
 132          * Ignore problems within a 'replacing' vdev, since we're presumably in
 133          * the process of repairing any such errors, and don't want to call them
 134          * out again.  We'll pick up the fact that a resilver is happening
 135          * later.
 136          */
 137         verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0);
 138         if (strcmp(type, VDEV_TYPE_REPLACING) == 0)
 139                 return (B_FALSE);
 140
 141         if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child,
 142             &children) == 0) {
 143                 for (c = 0; c < children; c++)
 144                         if (find_vdev_problem(child[c], func))
 145                                 return (B_TRUE);
 146         } else {
 147                 verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS,
 148                     (uint64_t **)&vs, &c) == 0);
 149
 150                 if (func(vs->vs_state, vs->vs_aux,
 151                     vs->vs_read_errors +
 152                     vs->vs_write_errors +
 153                     vs->vs_checksum_errors))
 154                         return (B_TRUE);
 155         }
 156
 157         /*
 158          * Check any L2 cache devs
 159          */
 160         if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_L2CACHE, &child,
 161             &children) == 0) {
 162                 for (c = 0; c < children; c++)
 163                         if (find_vdev_problem(child[c], func))
 164                                 return (B_TRUE);
 165         }
 166
 167         return (B_FALSE);
 168 }
 169
 170 /*
 171  * Active pool health status.
 172  *
 173  * To determine the status for a pool, we make several passes over the config,
 174  * picking the most egregious error we find.  In order of importance, we do the
 175  * following:
 176  *
 177  *      - Check for a complete and valid configuration
 178  *      - Look for any faulted or missing devices in a non-replicated config
 179  *      - Check for any data errors
 180  *      - Check for any faulted or missing devices in a replicated config
 181  *      - Look for any devices showing errors
 182  *      - Check for any resilvering devices
 183  *
 184  * There can obviously be multiple errors within a single pool, so this routine
 185  * only picks the most damaging of all the current errors to report.
 186  */
 187 static zpool_status_t
 188 check_status(nvlist_t *config, boolean_t isimport, zpool_errata_t *erratap)
 189 {
 190         nvlist_t *nvroot;
 191         vdev_stat_t *vs;
 192         pool_scan_stat_t *ps = NULL;
 193         uint_t vsc, psc;
 194         uint64_t nerr;
 195         uint64_t version;
 196         uint64_t stateval;
 197         uint64_t suspended;
 198         uint64_t hostid = 0;
 199         uint64_t errata = 0;
 200         unsigned long system_hostid = get_system_hostid();
 201
 202         verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
 203             &version) == 0);
 204         verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
 205             &nvroot) == 0);
 206         verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
 207             (uint64_t **)&vs, &vsc) == 0);
 208         verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
 209             &stateval) == 0);
 210
 211         /*
 212          * Currently resilvering a vdev
 213          */
 214         (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
 215             (uint64_t **)&ps, &psc);
 216         if (ps && ps->pss_func == POOL_SCAN_RESILVER &&
 217             ps->pss_state == DSS_SCANNING)
 218                 return (ZPOOL_STATUS_RESILVERING);
 219
 220         /*
 221          * The multihost property is set and the pool may be active.
 222          */
 223         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 224             vs->vs_aux == VDEV_AUX_ACTIVE) {
 225                 mmp_state_t mmp_state;
 226                 nvlist_t *nvinfo;
 227
 228                 nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
 229                 mmp_state = fnvlist_lookup_uint64(nvinfo,
 230                     ZPOOL_CONFIG_MMP_STATE);
 231
 232                 if (mmp_state == MMP_STATE_ACTIVE)
 233                         return (ZPOOL_STATUS_HOSTID_ACTIVE);
 234                 else if (mmp_state == MMP_STATE_NO_HOSTID)
 235                         return (ZPOOL_STATUS_HOSTID_REQUIRED);
 236                 else
 237                         return (ZPOOL_STATUS_HOSTID_MISMATCH);
 238         }
 239
 240         /*
 241          * Pool last accessed by another system.
 242          */
 243         (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
 244         if (hostid != 0 && (unsigned long)hostid != system_hostid &&
 245             stateval == POOL_STATE_ACTIVE)
 246                 return (ZPOOL_STATUS_HOSTID_MISMATCH);
 247
 248         /*
 249          * Newer on-disk version.
 250          */
 251         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 252             vs->vs_aux == VDEV_AUX_VERSION_NEWER)
 253                 return (ZPOOL_STATUS_VERSION_NEWER);
 254
 255         /*
 256          * Unsupported feature(s).
 257          */
 258         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 259             vs->vs_aux == VDEV_AUX_UNSUP_FEAT) {
 260                 nvlist_t *nvinfo;
 261
 262                 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO,
 263                     &nvinfo) == 0);
 264                 if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY))
 265                         return (ZPOOL_STATUS_UNSUP_FEAT_WRITE);
 266                 return (ZPOOL_STATUS_UNSUP_FEAT_READ);
 267         }
 268
 269         /*
 270          * Check that the config is complete.
 271          */
 272         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 273             vs->vs_aux == VDEV_AUX_BAD_GUID_SUM)
 274                 return (ZPOOL_STATUS_BAD_GUID_SUM);
 275
 276         /*
 277          * Check whether the pool has suspended.
 278          */
 279         if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED,
 280             &suspended) == 0) {
 281                 uint64_t reason;
 282
 283                 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED_REASON,
 284                     &reason) == 0 && reason == ZIO_SUSPEND_MMP)
 285                         return (ZPOOL_STATUS_IO_FAILURE_MMP);
 286
 287                 if (suspended == ZIO_FAILURE_MODE_CONTINUE)
 288                         return (ZPOOL_STATUS_IO_FAILURE_CONTINUE);
 289                 return (ZPOOL_STATUS_IO_FAILURE_WAIT);
 290         }
 291
 292         /*
 293          * Could not read a log.
 294          */
 295         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 296             vs->vs_aux == VDEV_AUX_BAD_LOG) {
 297                 return (ZPOOL_STATUS_BAD_LOG);
 298         }
 299
 300         /*
 301          * Bad devices in non-replicated config.
 302          */
 303         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 304             find_vdev_problem(nvroot, vdev_faulted))
 305                 return (ZPOOL_STATUS_FAULTED_DEV_NR);
 306
 307         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 308             find_vdev_problem(nvroot, vdev_missing))
 309                 return (ZPOOL_STATUS_MISSING_DEV_NR);
 310
 311         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 312             find_vdev_problem(nvroot, vdev_broken))
 313                 return (ZPOOL_STATUS_CORRUPT_LABEL_NR);
 314
 315         /*
 316          * Corrupted pool metadata
 317          */
 318         if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
 319             vs->vs_aux == VDEV_AUX_CORRUPT_DATA)
 320                 return (ZPOOL_STATUS_CORRUPT_POOL);
 321
 322         /*
 323          * Persistent data errors.
 324          */
 325         if (!isimport) {
 326                 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
 327                     &nerr) == 0 && nerr != 0)
 328                         return (ZPOOL_STATUS_CORRUPT_DATA);
 329         }
 330
 331         /*
 332          * Missing devices in a replicated config.
 333          */
 334         if (find_vdev_problem(nvroot, vdev_faulted))
 335                 return (ZPOOL_STATUS_FAULTED_DEV_R);
 336         if (find_vdev_problem(nvroot, vdev_missing))
 337                 return (ZPOOL_STATUS_MISSING_DEV_R);
 338         if (find_vdev_problem(nvroot, vdev_broken))
 339                 return (ZPOOL_STATUS_CORRUPT_LABEL_R);
 340
 341         /*
 342          * Devices with errors
 343          */
 344         if (!isimport && find_vdev_problem(nvroot, vdev_errors))
 345                 return (ZPOOL_STATUS_FAILING_DEV);
 346
 347         /*
 348          * Offlined devices
 349          */
 350         if (find_vdev_problem(nvroot, vdev_offlined))
 351                 return (ZPOOL_STATUS_OFFLINE_DEV);
 352
 353         /*
 354          * Removed device
 355          */
 356         if (find_vdev_problem(nvroot, vdev_removed))
 357                 return (ZPOOL_STATUS_REMOVED_DEV);
 358
 359         /*
 360          * Outdated, but usable, version
 361          */
 362         if (SPA_VERSION_IS_SUPPORTED(version) && version != SPA_VERSION)
 363                 return (ZPOOL_STATUS_VERSION_OLDER);
 364
 365         /*
 366          * Usable pool with disabled features
 367          */
 368         if (version >= SPA_VERSION_FEATURES) {
 369                 int i;
 370                 nvlist_t *feat;
 371
 372                 if (isimport) {
 373                         feat = fnvlist_lookup_nvlist(config,
 374                             ZPOOL_CONFIG_LOAD_INFO);
 375                         if (nvlist_exists(feat, ZPOOL_CONFIG_ENABLED_FEAT))
 376                                 feat = fnvlist_lookup_nvlist(feat,
 377                                     ZPOOL_CONFIG_ENABLED_FEAT);
 378                 } else {
 379                         feat = fnvlist_lookup_nvlist(config,
 380                             ZPOOL_CONFIG_FEATURE_STATS);
 381                 }
 382
 383                 for (i = 0; i < SPA_FEATURES; i++) {
 384                         zfeature_info_t *fi = &spa_feature_table[i];
 385                         if (!nvlist_exists(feat, fi->fi_guid))
 386                                 return (ZPOOL_STATUS_FEAT_DISABLED);
 387                 }
 388         }
 389
 390         /*
 391          * Informational errata available.
 392          */
 393         (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRATA, &errata);
 394         if (errata) {
 395                 *erratap = errata;
 396                 return (ZPOOL_STATUS_ERRATA);
 397         }
 398
 399         return (ZPOOL_STATUS_OK);
 400 }
 401
 402 zpool_status_t
 403 zpool_get_status(zpool_handle_t *zhp, char **msgid, zpool_errata_t *errata)
 404 {
 405         zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE, errata);
 406
 407         if (ret >= NMSGID)
 408                 *msgid = NULL;
 409         else
 410                 *msgid = zfs_msgid_table[ret];
 411
 412         return (ret);
 413 }
 414
 415 zpool_status_t
 416 zpool_import_status(nvlist_t *config, char **msgid, zpool_errata_t *errata)
 417 {
 418         zpool_status_t ret = check_status(config, B_TRUE, errata);
 419
 420         if (ret >= NMSGID)
 421                 *msgid = NULL;
 422         else
 423                 *msgid = zfs_msgid_table[ret];
 424
 425         return (ret);
 426 }
 427
 428 static void
 429 dump_ddt_stat(const ddt_stat_t *dds, int h)
 430 {
 431         char refcnt[6];
 432         char blocks[6], lsize[6], psize[6], dsize[6];
 433         char ref_blocks[6], ref_lsize[6], ref_psize[6], ref_dsize[6];
 434
 435         if (dds == NULL || dds->dds_blocks == 0)
 436                 return;
 437
 438         if (h == -1)
 439                 (void) strcpy(refcnt, "Total");
 440         else
 441                 zfs_nicenum(1ULL << h, refcnt, sizeof (refcnt));
 442
 443         zfs_nicenum(dds->dds_blocks, blocks, sizeof (blocks));
 444         zfs_nicebytes(dds->dds_lsize, lsize, sizeof (lsize));
 445         zfs_nicebytes(dds->dds_psize, psize, sizeof (psize));
 446         zfs_nicebytes(dds->dds_dsize, dsize, sizeof (dsize));
 447         zfs_nicenum(dds->dds_ref_blocks, ref_blocks, sizeof (ref_blocks));
 448         zfs_nicebytes(dds->dds_ref_lsize, ref_lsize, sizeof (ref_lsize));
 449         zfs_nicebytes(dds->dds_ref_psize, ref_psize, sizeof (ref_psize));
 450         zfs_nicebytes(dds->dds_ref_dsize, ref_dsize, sizeof (ref_dsize));
 451
 452         (void) printf("%6s   %6s   %5s   %5s   %5s   %6s   %5s   %5s   %5s\n",
 453             refcnt,
 454             blocks, lsize, psize, dsize,
 455             ref_blocks, ref_lsize, ref_psize, ref_dsize);
 456 }
 457
 458 /*
 459  * Print the DDT histogram and the column totals.
 460  */
 461 void
 462 zpool_dump_ddt(const ddt_stat_t *dds_total, const ddt_histogram_t *ddh)
 463 {
 464         int h;
 465
 466         (void) printf("\n");
 467
 468         (void) printf("bucket   "
 469             "           allocated             "
 470             "          referenced          \n");
 471         (void) printf("______   "
 472             "______________________________   "
 473             "______________________________\n");
 474
 475         (void) printf("%6s   %6s   %5s   %5s   %5s   %6s   %5s   %5s   %5s\n",
 476             "refcnt",
 477             "blocks", "LSIZE", "PSIZE", "DSIZE",
 478             "blocks", "LSIZE", "PSIZE", "DSIZE");
 479
 480         (void) printf("%6s   %6s   %5s   %5s   %5s   %6s   %5s   %5s   %5s\n",
 481             "------",
 482             "------", "-----", "-----", "-----",
 483             "------", "-----", "-----", "-----");
 484
 485         for (h = 0; h < 64; h++)
 486                 dump_ddt_stat(&ddh->ddh_stat[h], h);
 487
 488         dump_ddt_stat(dds_total, -1);
 489
 490         (void) printf("\n");
 491 }