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, Version 1.0 only 6 * (the "License"). You may not use this file except in compliance 7 * with the License. 8 * 9 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 10 * or http://www.opensolaris.org/os/licensing. 11 * See the License for the specific language governing permissions 12 * and limitations under the License. 13 * 14 * When distributing Covered Code, include this CDDL HEADER in each 15 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 16 * If applicable, add the following below this CDDL HEADER, with the 17 * fields enclosed by brackets "[]" replaced with your own identifying 18 * information: Portions Copyright [yyyy] [name of copyright owner] 19 * 20 * CDDL HEADER END 21 */ 22 /* 23 * Copyright 2005 Sun Microsystems, Inc. All rights reserved. 24 * Use is subject to license terms. 25 */ 26 27 #pragma ident "%Z%%M% %I% %E% SMI" 28 29 /* 30 * This file contains the functions which analyze the status of a pool. This 31 * include both the status of an active pool, as well as the status exported 32 * pools. Returns one of the ZPOOL_STATUS_* defines describing the status of 33 * the pool. This status is independent (to a certain degree) from the state of 34 * the pool. A pool's state descsribes only whether or not it is capable of 35 * providing the necessary fault tolerance for data. The status describes the 36 * overall status of devices. A pool that is online can still have a device 37 * that is experiencing errors. 38 * 39 * Only a subset of the possible faults can be detected using 'zpool status', 40 * and not all possible errors correspond to a FMA message ID. The explanation 41 * is left up to the caller, depending on whether it is a live pool or an 42 * import. 43 */ 44 45 #include <libzfs.h> 46 #include <string.h> 47 #include "libzfs_impl.h" 48 49 /* 50 * Message ID table. This must be kep in sync with the ZPOOL_STATUS_* defines 51 * in libzfs.h. Note that there are some status results which go past the end 52 * of this table, and hence have no associated message ID. 53 */ 54 static char *msgid_table[] = { 55 "ZFS-8000-14", 56 "ZFS-8000-2Q", 57 "ZFS-8000-3C", 58 "ZFS-8000-4J", 59 "ZFS-8000-5E", 60 "ZFS-8000-6X", 61 "ZFS-8000-72", 62 "ZFS-8000-8A", 63 "ZFS-8000-9P", 64 "ZFS-8000-A5" 65 }; 66 67 #define NMSGID (sizeof (msgid_table) / sizeof (msgid_table[0])) 68 69 /* ARGSUSED */ 70 static int 71 vdev_missing(uint64_t state, uint64_t aux, uint64_t errs) 72 { 73 return (state == VDEV_STATE_CANT_OPEN && 74 aux == VDEV_AUX_OPEN_FAILED); 75 } 76 77 /* ARGSUSED */ 78 static int 79 vdev_errors(uint64_t state, uint64_t aux, uint64_t errs) 80 { 81 return (errs != 0); 82 } 83 84 /* ARGSUSED */ 85 static int 86 vdev_broken(uint64_t state, uint64_t aux, uint64_t errs) 87 { 88 return (state == VDEV_STATE_CANT_OPEN); 89 } 90 91 /* ARGSUSED */ 92 static int 93 vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs) 94 { 95 return (state == VDEV_STATE_OFFLINE); 96 } 97 98 /* 99 * Detect if any leaf devices that have seen errors or could not be opened. 100 */ 101 static int 102 find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t)) 103 { 104 nvlist_t **child; 105 vdev_stat_t *vs; 106 uint_t c, children; 107 char *type; 108 109 /* 110 * Ignore problems within a 'replacing' vdev, since we're presumably in 111 * the process of repairing any such errors, and don't want to call them 112 * out again. We'll pick up the fact that a resilver is happening 113 * later. 114 */ 115 verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0); 116 if (strcmp(type, VDEV_TYPE_REPLACING) == 0) 117 return (FALSE); 118 119 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child, 120 &children) == 0) { 121 for (c = 0; c < children; c++) 122 if (find_vdev_problem(child[c], func)) 123 return (TRUE); 124 } else { 125 verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_STATS, 126 (uint64_t **)&vs, &c) == 0); 127 128 if (func(vs->vs_state, vs->vs_aux, 129 vs->vs_read_errors + 130 vs->vs_write_errors + 131 vs->vs_checksum_errors)) 132 return (TRUE); 133 } 134 135 return (FALSE); 136 } 137 138 /* 139 * Active pool health status. 140 * 141 * To determine the status for a pool, we make several passes over the config, 142 * picking the most egregious error we find. In order of importance, we do the 143 * following: 144 * 145 * - Check for a complete and valid configuration 146 * - Look for any missing devices 147 * - Look for any devices showing errors 148 * - Check for any data errors 149 * - Check for any resilvering devices 150 * 151 * There can obviously be multiple errors within a single pool, so this routine 152 * only picks the most damaging of all the current errors to report. 153 */ 154 static zpool_status_t 155 check_status(nvlist_t *config, int isimport) 156 { 157 nvlist_t *nvroot; 158 vdev_stat_t *vs; 159 uint_t vsc; 160 161 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, 162 &nvroot) == 0); 163 verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, 164 (uint64_t **)&vs, &vsc) == 0); 165 166 /* 167 * Check that the config is complete. 168 */ 169 if (vs->vs_state == VDEV_STATE_CANT_OPEN && 170 vs->vs_aux == VDEV_AUX_BAD_GUID_SUM) { 171 return (ZPOOL_STATUS_BAD_GUID_SUM); 172 } 173 174 /* 175 * Missing devices 176 */ 177 if (find_vdev_problem(nvroot, vdev_missing)) { 178 if (vs->vs_state == VDEV_STATE_CANT_OPEN) 179 return (ZPOOL_STATUS_MISSING_DEV_NR); 180 else 181 return (ZPOOL_STATUS_MISSING_DEV_R); 182 } 183 184 /* 185 * Devices with corrupted labels. 186 */ 187 if (find_vdev_problem(nvroot, vdev_broken)) { 188 if (vs->vs_state == VDEV_STATE_CANT_OPEN) 189 return (ZPOOL_STATUS_CORRUPT_LABEL_NR); 190 else 191 return (ZPOOL_STATUS_CORRUPT_LABEL_R); 192 } 193 194 /* 195 * Devices with errors 196 */ 197 if (!isimport && find_vdev_problem(nvroot, vdev_errors)) 198 return (ZPOOL_STATUS_FAILING_DEV); 199 200 /* 201 * Offlined devices 202 */ 203 if (find_vdev_problem(nvroot, vdev_offlined)) 204 return (ZPOOL_STATUS_OFFLINE_DEV); 205 206 /* 207 * Currently resilvering 208 */ 209 if (!vs->vs_scrub_complete && vs->vs_scrub_type == POOL_SCRUB_RESILVER) 210 return (ZPOOL_STATUS_RESILVERING); 211 212 /* 213 * We currently have no way to detect the following errors: 214 * 215 * CORRUPT_CACHE 216 * VERSION_MISMATCH 217 * CORRUPT_POOL 218 * CORRUPT_DATA 219 */ 220 221 return (ZPOOL_STATUS_OK); 222 } 223 224 zpool_status_t 225 zpool_get_status(zpool_handle_t *zhp, char **msgid) 226 { 227 zpool_status_t ret = check_status(zhp->zpool_config, FALSE); 228 229 if (ret >= NMSGID) 230 *msgid = NULL; 231 else 232 *msgid = msgid_table[ret]; 233 234 return (ret); 235 } 236 237 zpool_status_t 238 zpool_import_status(nvlist_t *config, char **msgid) 239 { 240 zpool_status_t ret = check_status(config, TRUE); 241 242 if (ret >= NMSGID) 243 *msgid = NULL; 244 else 245 *msgid = msgid_table[ret]; 246 247 return (ret); 248 } 249