]> git.proxmox.com Git - mirror_zfs-debian.git/blob - lib/libzfs/libzfs_status.c
Align parition end on 1 MiB boundary
[mirror_zfs-debian.git] / 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 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
23 */
24
25 /*
26 * This file contains the functions which analyze the status of a pool. This
27 * include both the status of an active pool, as well as the status exported
28 * pools. Returns one of the ZPOOL_STATUS_* defines describing the status of
29 * the pool. This status is independent (to a certain degree) from the state of
30 * the pool. A pool's state describes only whether or not it is capable of
31 * providing the necessary fault tolerance for data. The status describes the
32 * overall status of devices. A pool that is online can still have a device
33 * that is experiencing errors.
34 *
35 * Only a subset of the possible faults can be detected using 'zpool status',
36 * and not all possible errors correspond to a FMA message ID. The explanation
37 * is left up to the caller, depending on whether it is a live pool or an
38 * import.
39 */
40
41 #include <libzfs.h>
42 #include <string.h>
43 #include <unistd.h>
44 #include "libzfs_impl.h"
45
46 /*
47 * Message ID table. This must be kept in sync with the ZPOOL_STATUS_* defines
48 * in libzfs.h. Note that there are some status results which go past the end
49 * of this table, and hence have no associated message ID.
50 */
51 static char *zfs_msgid_table[] = {
52 "ZFS-8000-14",
53 "ZFS-8000-2Q",
54 "ZFS-8000-3C",
55 "ZFS-8000-4J",
56 "ZFS-8000-5E",
57 "ZFS-8000-6X",
58 "ZFS-8000-72",
59 "ZFS-8000-8A",
60 "ZFS-8000-9P",
61 "ZFS-8000-A5",
62 "ZFS-8000-EY",
63 "ZFS-8000-HC",
64 "ZFS-8000-JQ",
65 "ZFS-8000-K4",
66 };
67
68 #define NMSGID (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0]))
69
70 /* ARGSUSED */
71 static int
72 vdev_missing(uint64_t state, uint64_t aux, uint64_t errs)
73 {
74 return (state == VDEV_STATE_CANT_OPEN &&
75 aux == VDEV_AUX_OPEN_FAILED);
76 }
77
78 /* ARGSUSED */
79 static int
80 vdev_faulted(uint64_t state, uint64_t aux, uint64_t errs)
81 {
82 return (state == VDEV_STATE_FAULTED);
83 }
84
85 /* ARGSUSED */
86 static int
87 vdev_errors(uint64_t state, uint64_t aux, uint64_t errs)
88 {
89 return (state == VDEV_STATE_DEGRADED || errs != 0);
90 }
91
92 /* ARGSUSED */
93 static int
94 vdev_broken(uint64_t state, uint64_t aux, uint64_t errs)
95 {
96 return (state == VDEV_STATE_CANT_OPEN);
97 }
98
99 /* ARGSUSED */
100 static int
101 vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs)
102 {
103 return (state == VDEV_STATE_OFFLINE);
104 }
105
106 /* ARGSUSED */
107 static int
108 vdev_removed(uint64_t state, uint64_t aux, uint64_t errs)
109 {
110 return (state == VDEV_STATE_REMOVED);
111 }
112
113 /*
114 * Detect if any leaf devices that have seen errors or could not be opened.
115 */
116 static boolean_t
117 find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t))
118 {
119 nvlist_t **child;
120 vdev_stat_t *vs;
121 uint_t c, children;
122 char *type;
123
124 /*
125 * Ignore problems within a 'replacing' vdev, since we're presumably in
126 * the process of repairing any such errors, and don't want to call them
127 * out again. We'll pick up the fact that a resilver is happening
128 * later.
129 */
130 verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0);
131 if (strcmp(type, VDEV_TYPE_REPLACING) == 0)
132 return (B_FALSE);
133
134 if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child,
135 &children) == 0) {
136 for (c = 0; c < children; c++)
137 if (find_vdev_problem(child[c], func))
138 return (B_TRUE);
139 } else {
140 verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS,
141 (uint64_t **)&vs, &c) == 0);
142
143 if (func(vs->vs_state, vs->vs_aux,
144 vs->vs_read_errors +
145 vs->vs_write_errors +
146 vs->vs_checksum_errors))
147 return (B_TRUE);
148 }
149
150 return (B_FALSE);
151 }
152
153 /*
154 * Active pool health status.
155 *
156 * To determine the status for a pool, we make several passes over the config,
157 * picking the most egregious error we find. In order of importance, we do the
158 * following:
159 *
160 * - Check for a complete and valid configuration
161 * - Look for any faulted or missing devices in a non-replicated config
162 * - Check for any data errors
163 * - Check for any faulted or missing devices in a replicated config
164 * - Look for any devices showing errors
165 * - Check for any resilvering devices
166 *
167 * There can obviously be multiple errors within a single pool, so this routine
168 * only picks the most damaging of all the current errors to report.
169 */
170 static zpool_status_t
171 check_status(nvlist_t *config, boolean_t isimport)
172 {
173 nvlist_t *nvroot;
174 vdev_stat_t *vs;
175 pool_scan_stat_t *ps = NULL;
176 uint_t vsc, psc;
177 uint64_t nerr;
178 uint64_t version;
179 uint64_t stateval;
180 uint64_t suspended;
181 uint64_t hostid = 0;
182 unsigned long system_hostid = gethostid() & 0xffffffff;
183
184 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
185 &version) == 0);
186 verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
187 &nvroot) == 0);
188 verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
189 (uint64_t **)&vs, &vsc) == 0);
190 verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
191 &stateval) == 0);
192
193 /*
194 * Currently resilvering a vdev
195 */
196 (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
197 (uint64_t **)&ps, &psc);
198 if (ps && ps->pss_func == POOL_SCAN_RESILVER &&
199 ps->pss_state == DSS_SCANNING)
200 return (ZPOOL_STATUS_RESILVERING);
201
202 /*
203 * Pool last accessed by another system.
204 */
205 (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
206 if (hostid != 0 && (unsigned long)hostid != system_hostid &&
207 stateval == POOL_STATE_ACTIVE)
208 return (ZPOOL_STATUS_HOSTID_MISMATCH);
209
210 /*
211 * Newer on-disk version.
212 */
213 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
214 vs->vs_aux == VDEV_AUX_VERSION_NEWER)
215 return (ZPOOL_STATUS_VERSION_NEWER);
216
217 /*
218 * Check that the config is complete.
219 */
220 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
221 vs->vs_aux == VDEV_AUX_BAD_GUID_SUM)
222 return (ZPOOL_STATUS_BAD_GUID_SUM);
223
224 /*
225 * Check whether the pool has suspended due to failed I/O.
226 */
227 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED,
228 &suspended) == 0) {
229 if (suspended == ZIO_FAILURE_MODE_CONTINUE)
230 return (ZPOOL_STATUS_IO_FAILURE_CONTINUE);
231 return (ZPOOL_STATUS_IO_FAILURE_WAIT);
232 }
233
234 /*
235 * Could not read a log.
236 */
237 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
238 vs->vs_aux == VDEV_AUX_BAD_LOG) {
239 return (ZPOOL_STATUS_BAD_LOG);
240 }
241
242 /*
243 * Bad devices in non-replicated config.
244 */
245 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
246 find_vdev_problem(nvroot, vdev_faulted))
247 return (ZPOOL_STATUS_FAULTED_DEV_NR);
248
249 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
250 find_vdev_problem(nvroot, vdev_missing))
251 return (ZPOOL_STATUS_MISSING_DEV_NR);
252
253 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
254 find_vdev_problem(nvroot, vdev_broken))
255 return (ZPOOL_STATUS_CORRUPT_LABEL_NR);
256
257 /*
258 * Corrupted pool metadata
259 */
260 if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
261 vs->vs_aux == VDEV_AUX_CORRUPT_DATA)
262 return (ZPOOL_STATUS_CORRUPT_POOL);
263
264 /*
265 * Persistent data errors.
266 */
267 if (!isimport) {
268 if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
269 &nerr) == 0 && nerr != 0)
270 return (ZPOOL_STATUS_CORRUPT_DATA);
271 }
272
273 /*
274 * Missing devices in a replicated config.
275 */
276 if (find_vdev_problem(nvroot, vdev_faulted))
277 return (ZPOOL_STATUS_FAULTED_DEV_R);
278 if (find_vdev_problem(nvroot, vdev_missing))
279 return (ZPOOL_STATUS_MISSING_DEV_R);
280 if (find_vdev_problem(nvroot, vdev_broken))
281 return (ZPOOL_STATUS_CORRUPT_LABEL_R);
282
283 /*
284 * Devices with errors
285 */
286 if (!isimport && find_vdev_problem(nvroot, vdev_errors))
287 return (ZPOOL_STATUS_FAILING_DEV);
288
289 /*
290 * Offlined devices
291 */
292 if (find_vdev_problem(nvroot, vdev_offlined))
293 return (ZPOOL_STATUS_OFFLINE_DEV);
294
295 /*
296 * Removed device
297 */
298 if (find_vdev_problem(nvroot, vdev_removed))
299 return (ZPOOL_STATUS_REMOVED_DEV);
300
301 /*
302 * Outdated, but usable, version
303 */
304 if (version < SPA_VERSION)
305 return (ZPOOL_STATUS_VERSION_OLDER);
306
307 return (ZPOOL_STATUS_OK);
308 }
309
310 zpool_status_t
311 zpool_get_status(zpool_handle_t *zhp, char **msgid)
312 {
313 zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE);
314
315 if (ret >= NMSGID)
316 *msgid = NULL;
317 else
318 *msgid = zfs_msgid_table[ret];
319
320 return (ret);
321 }
322
323 zpool_status_t
324 zpool_import_status(nvlist_t *config, char **msgid)
325 {
326 zpool_status_t ret = check_status(config, B_TRUE);
327
328 if (ret >= NMSGID)
329 *msgid = NULL;
330 else
331 *msgid = zfs_msgid_table[ret];
332
333 return (ret);
334 }
335
336 static void
337 dump_ddt_stat(const ddt_stat_t *dds, int h)
338 {
339 char refcnt[6];
340 char blocks[6], lsize[6], psize[6], dsize[6];
341 char ref_blocks[6], ref_lsize[6], ref_psize[6], ref_dsize[6];
342
343 if (dds == NULL || dds->dds_blocks == 0)
344 return;
345
346 if (h == -1)
347 (void) strcpy(refcnt, "Total");
348 else
349 zfs_nicenum(1ULL << h, refcnt, sizeof (refcnt));
350
351 zfs_nicenum(dds->dds_blocks, blocks, sizeof (blocks));
352 zfs_nicenum(dds->dds_lsize, lsize, sizeof (lsize));
353 zfs_nicenum(dds->dds_psize, psize, sizeof (psize));
354 zfs_nicenum(dds->dds_dsize, dsize, sizeof (dsize));
355 zfs_nicenum(dds->dds_ref_blocks, ref_blocks, sizeof (ref_blocks));
356 zfs_nicenum(dds->dds_ref_lsize, ref_lsize, sizeof (ref_lsize));
357 zfs_nicenum(dds->dds_ref_psize, ref_psize, sizeof (ref_psize));
358 zfs_nicenum(dds->dds_ref_dsize, ref_dsize, sizeof (ref_dsize));
359
360 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n",
361 refcnt,
362 blocks, lsize, psize, dsize,
363 ref_blocks, ref_lsize, ref_psize, ref_dsize);
364 }
365
366 /*
367 * Print the DDT histogram and the column totals.
368 */
369 void
370 zpool_dump_ddt(const ddt_stat_t *dds_total, const ddt_histogram_t *ddh)
371 {
372 int h;
373
374 (void) printf("\n");
375
376 (void) printf("bucket "
377 " allocated "
378 " referenced \n");
379 (void) printf("______ "
380 "______________________________ "
381 "______________________________\n");
382
383 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n",
384 "refcnt",
385 "blocks", "LSIZE", "PSIZE", "DSIZE",
386 "blocks", "LSIZE", "PSIZE", "DSIZE");
387
388 (void) printf("%6s %6s %5s %5s %5s %6s %5s %5s %5s\n",
389 "------",
390 "------", "-----", "-----", "-----",
391 "------", "-----", "-----", "-----");
392
393 for (h = 0; h < 64; h++)
394 dump_ddt_stat(&ddh->ddh_stat[h], h);
395
396 dump_ddt_stat(dds_total, -1);
397
398 (void) printf("\n");
399 }