[mirror_zfs.git] / lib / libzfs / libzfs_status.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or https://opensource.org/licenses/CDDL-1.0.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */

/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2012 by Delphix. All rights reserved.
 * Copyright (c) 2013 Steven Hartland. All rights reserved.
 * Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
 */

/*
 * This file contains the functions which analyze the status of a pool.  This
 * include both the status of an active pool, as well as the status exported
 * pools.  Returns one of the ZPOOL_STATUS_* defines describing the status of
 * the pool.  This status is independent (to a certain degree) from the state of
 * the pool.  A pool's state describes only whether or not it is capable of
 * providing the necessary fault tolerance for data.  The status describes the
 * overall status of devices.  A pool that is online can still have a device
 * that is experiencing errors.
 *
 * Only a subset of the possible faults can be detected using 'zpool status',
 * and not all possible errors correspond to a FMA message ID.  The explanation
 * is left up to the caller, depending on whether it is a live pool or an
 * import.
 */

#include <libzfs.h>
#include <libzutil.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/systeminfo.h>
#include "libzfs_impl.h"
#include "zfeature_common.h"

/*
 * Message ID table.  This must be kept in sync with the ZPOOL_STATUS_* defines
 * in include/libzfs.h.  Note that there are some status results which go past
 * the end of this table, and hence have no associated message ID.
 */
static const char *const zfs_msgid_table[] = {
	"ZFS-8000-14", /* ZPOOL_STATUS_CORRUPT_CACHE */
	"ZFS-8000-2Q", /* ZPOOL_STATUS_MISSING_DEV_R */
	"ZFS-8000-3C", /* ZPOOL_STATUS_MISSING_DEV_NR */
	"ZFS-8000-4J", /* ZPOOL_STATUS_CORRUPT_LABEL_R */
	"ZFS-8000-5E", /* ZPOOL_STATUS_CORRUPT_LABEL_NR */
	"ZFS-8000-6X", /* ZPOOL_STATUS_BAD_GUID_SUM */
	"ZFS-8000-72", /* ZPOOL_STATUS_CORRUPT_POOL */
	"ZFS-8000-8A", /* ZPOOL_STATUS_CORRUPT_DATA */
	"ZFS-8000-9P", /* ZPOOL_STATUS_FAILING_DEV */
	"ZFS-8000-A5", /* ZPOOL_STATUS_VERSION_NEWER */
	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_MISMATCH */
	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_ACTIVE */
	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_REQUIRED */
	"ZFS-8000-HC", /* ZPOOL_STATUS_IO_FAILURE_WAIT */
	"ZFS-8000-JQ", /* ZPOOL_STATUS_IO_FAILURE_CONTINUE */
	"ZFS-8000-MM", /* ZPOOL_STATUS_IO_FAILURE_MMP */
	"ZFS-8000-K4", /* ZPOOL_STATUS_BAD_LOG */
	"ZFS-8000-ER", /* ZPOOL_STATUS_ERRATA */
	/*
	 * The following results have no message ID.
	 *	ZPOOL_STATUS_UNSUP_FEAT_READ
	 *	ZPOOL_STATUS_UNSUP_FEAT_WRITE
	 *	ZPOOL_STATUS_FAULTED_DEV_R
	 *	ZPOOL_STATUS_FAULTED_DEV_NR
	 *	ZPOOL_STATUS_VERSION_OLDER
	 *	ZPOOL_STATUS_FEAT_DISABLED
	 *	ZPOOL_STATUS_RESILVERING
	 *	ZPOOL_STATUS_OFFLINE_DEV
	 *	ZPOOL_STATUS_REMOVED_DEV
	 *	ZPOOL_STATUS_REBUILDING
	 *	ZPOOL_STATUS_REBUILD_SCRUB
	 *	ZPOOL_STATUS_COMPATIBILITY_ERR
	 *	ZPOOL_STATUS_INCOMPATIBLE_FEAT
	 *	ZPOOL_STATUS_OK
	 */
};

#define	NMSGID	(sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0]))

static int
vdev_missing(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_OPEN_FAILED);
}

static int
vdev_faulted(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_FAULTED);
}

static int
vdev_errors(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_DEGRADED ||
	    vs->vs_read_errors != 0 || vs->vs_write_errors != 0 ||
	    vs->vs_checksum_errors != 0);
}

static int
vdev_broken(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_CANT_OPEN);
}

static int
vdev_offlined(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_OFFLINE);
}

static int
vdev_removed(vdev_stat_t *vs, uint_t vsc)
{
	(void) vsc;
	return (vs->vs_state == VDEV_STATE_REMOVED);
}

static int
vdev_non_native_ashift(vdev_stat_t *vs, uint_t vsc)
{
	if (getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") != NULL)
		return (0);

	return (VDEV_STAT_VALID(vs_physical_ashift, vsc) &&
	    vs->vs_configured_ashift < vs->vs_physical_ashift);
}

/*
 * Detect if any leaf devices that have seen errors or could not be opened.
 */
static boolean_t
find_vdev_problem(nvlist_t *vdev, int (*func)(vdev_stat_t *, uint_t),
    boolean_t ignore_replacing)
{
	nvlist_t **child;
	uint_t c, children;

	/*
	 * Ignore problems within a 'replacing' vdev, since we're presumably in
	 * the process of repairing any such errors, and don't want to call them
	 * out again.  We'll pick up the fact that a resilver is happening
	 * later.
	 */
	if (ignore_replacing == B_TRUE) {
		const char *type = fnvlist_lookup_string(vdev,
		    ZPOOL_CONFIG_TYPE);
		if (strcmp(type, VDEV_TYPE_REPLACING) == 0)
			return (B_FALSE);
	}

	if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child,
	    &children) == 0) {
		for (c = 0; c < children; c++)
			if (find_vdev_problem(child[c], func, ignore_replacing))
				return (B_TRUE);
	} else {
		uint_t vsc;
		vdev_stat_t *vs = (vdev_stat_t *)fnvlist_lookup_uint64_array(
		    vdev, ZPOOL_CONFIG_VDEV_STATS, &vsc);
		if (func(vs, vsc) != 0)
			return (B_TRUE);
	}

	/*
	 * Check any L2 cache devs
	 */
	if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_L2CACHE, &child,
	    &children) == 0) {
		for (c = 0; c < children; c++)
			if (find_vdev_problem(child[c], func, ignore_replacing))
				return (B_TRUE);
	}

	return (B_FALSE);
}

/*
 * Active pool health status.
 *
 * To determine the status for a pool, we make several passes over the config,
 * picking the most egregious error we find.  In order of importance, we do the
 * following:
 *
 *	- Check for a complete and valid configuration
 *	- Look for any faulted or missing devices in a non-replicated config
 *	- Check for any data errors
 *	- Check for any faulted or missing devices in a replicated config
 *	- Look for any devices showing errors
 *	- Check for any resilvering or rebuilding devices
 *
 * There can obviously be multiple errors within a single pool, so this routine
 * only picks the most damaging of all the current errors to report.
 */
static zpool_status_t
check_status(nvlist_t *config, boolean_t isimport,
    zpool_errata_t *erratap, const char *compat)
{
	pool_scan_stat_t *ps = NULL;
	uint_t vsc, psc;
	uint64_t suspended;
	uint64_t hostid = 0;
	uint64_t errata = 0;
	unsigned long system_hostid = get_system_hostid();

	uint64_t version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
	nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
	    ZPOOL_CONFIG_VDEV_TREE);
	vdev_stat_t *vs = (vdev_stat_t *)fnvlist_lookup_uint64_array(nvroot,
	    ZPOOL_CONFIG_VDEV_STATS, &vsc);
	uint64_t stateval = fnvlist_lookup_uint64(config,
	    ZPOOL_CONFIG_POOL_STATE);

	/*
	 * Currently resilvering a vdev
	 */
	(void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
	    (uint64_t **)&ps, &psc);
	if (ps != NULL && ps->pss_func == POOL_SCAN_RESILVER &&
	    ps->pss_state == DSS_SCANNING)
		return (ZPOOL_STATUS_RESILVERING);

	/*
	 * Currently rebuilding a vdev, check top-level vdevs.
	 */
	vdev_rebuild_stat_t *vrs = NULL;
	nvlist_t **child;
	uint_t c, i, children;
	uint64_t rebuild_end_time = 0;
	if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
	    &child, &children) == 0) {
		for (c = 0; c < children; c++) {
			if ((nvlist_lookup_uint64_array(child[c],
			    ZPOOL_CONFIG_REBUILD_STATS,
			    (uint64_t **)&vrs, &i) == 0) && (vrs != NULL)) {
				uint64_t state = vrs->vrs_state;

				if (state == VDEV_REBUILD_ACTIVE) {
					return (ZPOOL_STATUS_REBUILDING);
				} else if (state == VDEV_REBUILD_COMPLETE &&
				    vrs->vrs_end_time > rebuild_end_time) {
					rebuild_end_time = vrs->vrs_end_time;
				}
			}
		}

		/*
		 * If we can determine when the last scrub was run, and it
		 * was before the last rebuild completed, then recommend
		 * that the pool be scrubbed to verify all checksums.  When
		 * ps is NULL we can infer the pool has never been scrubbed.
		 */
		if (rebuild_end_time > 0) {
			if (ps != NULL) {
				if ((ps->pss_state == DSS_FINISHED &&
				    ps->pss_func == POOL_SCAN_SCRUB &&
				    rebuild_end_time > ps->pss_end_time) ||
				    ps->pss_state == DSS_NONE)
					return (ZPOOL_STATUS_REBUILD_SCRUB);
			} else {
				return (ZPOOL_STATUS_REBUILD_SCRUB);
			}
		}
	}

	/*
	 * The multihost property is set and the pool may be active.
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_ACTIVE) {
		mmp_state_t mmp_state;
		nvlist_t *nvinfo;

		nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
		mmp_state = fnvlist_lookup_uint64(nvinfo,
		    ZPOOL_CONFIG_MMP_STATE);

		if (mmp_state == MMP_STATE_ACTIVE)
			return (ZPOOL_STATUS_HOSTID_ACTIVE);
		else if (mmp_state == MMP_STATE_NO_HOSTID)
			return (ZPOOL_STATUS_HOSTID_REQUIRED);
		else
			return (ZPOOL_STATUS_HOSTID_MISMATCH);
	}

	/*
	 * Pool last accessed by another system.
	 */
	(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
	if (hostid != 0 && (unsigned long)hostid != system_hostid &&
	    stateval == POOL_STATE_ACTIVE)
		return (ZPOOL_STATUS_HOSTID_MISMATCH);

	/*
	 * Newer on-disk version.
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_VERSION_NEWER)
		return (ZPOOL_STATUS_VERSION_NEWER);

	/*
	 * Unsupported feature(s).
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_UNSUP_FEAT) {
		nvlist_t *nvinfo = fnvlist_lookup_nvlist(config,
		    ZPOOL_CONFIG_LOAD_INFO);
		if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY))
			return (ZPOOL_STATUS_UNSUP_FEAT_WRITE);
		return (ZPOOL_STATUS_UNSUP_FEAT_READ);
	}

	/*
	 * Check that the config is complete.
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_BAD_GUID_SUM)
		return (ZPOOL_STATUS_BAD_GUID_SUM);

	/*
	 * Check whether the pool has suspended.
	 */
	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED,
	    &suspended) == 0) {
		uint64_t reason;

		if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED_REASON,
		    &reason) == 0 && reason == ZIO_SUSPEND_MMP)
			return (ZPOOL_STATUS_IO_FAILURE_MMP);

		if (suspended == ZIO_FAILURE_MODE_CONTINUE)
			return (ZPOOL_STATUS_IO_FAILURE_CONTINUE);
		return (ZPOOL_STATUS_IO_FAILURE_WAIT);
	}

	/*
	 * Could not read a log.
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_BAD_LOG) {
		return (ZPOOL_STATUS_BAD_LOG);
	}

	/*
	 * Bad devices in non-replicated config.
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    find_vdev_problem(nvroot, vdev_faulted, B_TRUE))
		return (ZPOOL_STATUS_FAULTED_DEV_NR);

	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    find_vdev_problem(nvroot, vdev_missing, B_TRUE))
		return (ZPOOL_STATUS_MISSING_DEV_NR);

	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    find_vdev_problem(nvroot, vdev_broken, B_TRUE))
		return (ZPOOL_STATUS_CORRUPT_LABEL_NR);

	/*
	 * Corrupted pool metadata
	 */
	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	    vs->vs_aux == VDEV_AUX_CORRUPT_DATA)
		return (ZPOOL_STATUS_CORRUPT_POOL);

	/*
	 * Persistent data errors.
	 */
	if (!isimport) {
		uint64_t nerr;
		if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
		    &nerr) == 0 && nerr != 0)
			return (ZPOOL_STATUS_CORRUPT_DATA);
	}

	/*
	 * Missing devices in a replicated config.
	 */
	if (find_vdev_problem(nvroot, vdev_faulted, B_TRUE))
		return (ZPOOL_STATUS_FAULTED_DEV_R);
	if (find_vdev_problem(nvroot, vdev_missing, B_TRUE))
		return (ZPOOL_STATUS_MISSING_DEV_R);
	if (find_vdev_problem(nvroot, vdev_broken, B_TRUE))
		return (ZPOOL_STATUS_CORRUPT_LABEL_R);

	/*
	 * Devices with errors
	 */
	if (!isimport && find_vdev_problem(nvroot, vdev_errors, B_TRUE))
		return (ZPOOL_STATUS_FAILING_DEV);

	/*
	 * Offlined devices
	 */
	if (find_vdev_problem(nvroot, vdev_offlined, B_TRUE))
		return (ZPOOL_STATUS_OFFLINE_DEV);

	/*
	 * Removed device
	 */
	if (find_vdev_problem(nvroot, vdev_removed, B_TRUE))
		return (ZPOOL_STATUS_REMOVED_DEV);

	/*
	 * Suboptimal, but usable, ashift configuration.
	 */
	if (find_vdev_problem(nvroot, vdev_non_native_ashift, B_FALSE))
		return (ZPOOL_STATUS_NON_NATIVE_ASHIFT);

	/*
	 * Informational errata available.
	 */
	(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRATA, &errata);
	if (errata) {
		*erratap = errata;
		return (ZPOOL_STATUS_ERRATA);
	}

	/*
	 * Outdated, but usable, version
	 */
	if (SPA_VERSION_IS_SUPPORTED(version) && version != SPA_VERSION) {
		/* "legacy" compatibility disables old version reporting */
		if (compat != NULL && strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0)
			return (ZPOOL_STATUS_OK);
		else
			return (ZPOOL_STATUS_VERSION_OLDER);
	}

	/*
	 * Usable pool with disabled or superfluous features
	 * (superfluous = beyond what's requested by 'compatibility')
	 */
	if (version >= SPA_VERSION_FEATURES) {
		int i;
		nvlist_t *feat;

		if (isimport) {
			feat = fnvlist_lookup_nvlist(config,
			    ZPOOL_CONFIG_LOAD_INFO);
			if (nvlist_exists(feat, ZPOOL_CONFIG_ENABLED_FEAT))
				feat = fnvlist_lookup_nvlist(feat,
				    ZPOOL_CONFIG_ENABLED_FEAT);
		} else {
			feat = fnvlist_lookup_nvlist(config,
			    ZPOOL_CONFIG_FEATURE_STATS);
		}

		/* check against all features, or limited set? */
		boolean_t c_features[SPA_FEATURES];

		switch (zpool_load_compat(compat, c_features, NULL, 0)) {
		case ZPOOL_COMPATIBILITY_OK:
		case ZPOOL_COMPATIBILITY_WARNTOKEN:
			break;
		default:
			return (ZPOOL_STATUS_COMPATIBILITY_ERR);
		}
		for (i = 0; i < SPA_FEATURES; i++) {
			zfeature_info_t *fi = &spa_feature_table[i];
			if (!fi->fi_zfs_mod_supported)
				continue;
			if (c_features[i] && !nvlist_exists(feat, fi->fi_guid))
				return (ZPOOL_STATUS_FEAT_DISABLED);
			if (!c_features[i] && nvlist_exists(feat, fi->fi_guid))
				return (ZPOOL_STATUS_INCOMPATIBLE_FEAT);
		}
	}

	return (ZPOOL_STATUS_OK);
}

zpool_status_t
zpool_get_status(zpool_handle_t *zhp, const char **msgid,
    zpool_errata_t *errata)
{
	/*
	 * pass in the desired feature set, as
	 * it affects check for disabled features
	 */
	char compatibility[ZFS_MAXPROPLEN];
	if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compatibility,
	    ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
		compatibility[0] = '\0';

	zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE, errata,
	    compatibility);

	if (msgid != NULL) {
		if (ret >= NMSGID)
			*msgid = NULL;
		else
			*msgid = zfs_msgid_table[ret];
	}
	return (ret);
}

zpool_status_t
zpool_import_status(nvlist_t *config, const char **msgid,
    zpool_errata_t *errata)
{
	zpool_status_t ret = check_status(config, B_TRUE, errata, NULL);

	if (ret >= NMSGID)
		*msgid = NULL;
	else
		*msgid = zfs_msgid_table[ret];

	return (ret);
}
Commit	Line	Data
34dc7c2f BB	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
1d3ba0bf	9	* or https://opensource.org/licenses/CDDL-1.0.
34dc7c2f BB	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	*/
9ae529ec	21
34dc7c2f	22	/*
428870ff	23	* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
9ae529ec	24	* Copyright (c) 2012 by Delphix. All rights reserved.
34ffbed8	25	* Copyright (c) 2013 Steven Hartland. All rights reserved.
658fb802	26	* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
34dc7c2f BB	27	*/
34dc7c2f BB	28
34dc7c2f BB	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 describes 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>
e89f1295	46	#include <libzutil.h>
6fe3498c	47	#include <stdlib.h>
34dc7c2f BB	48	#include <string.h>
34dc7c2f BB	49	#include <unistd.h>
46364cb2	50	#include <sys/systeminfo.h>
34dc7c2f	51	#include "libzfs_impl.h"
b9b24bb4	52	#include "zfeature_common.h"
34dc7c2f BB	53
	54	/*
	55	* Message ID table. This must be kept in sync with the ZPOOL_STATUS_* defines
5365b074	56	* in include/libzfs.h. Note that there are some status results which go past
5365b074	57	* the end of this table, and hence have no associated message ID.
34dc7c2f	58	*/
a926aab9	59	static const char *const zfs_msgid_table[] = {
5365b074	60	"ZFS-8000-14", /* ZPOOL_STATUS_CORRUPT_CACHE */
	61	"ZFS-8000-2Q", /* ZPOOL_STATUS_MISSING_DEV_R */
	62	"ZFS-8000-3C", /* ZPOOL_STATUS_MISSING_DEV_NR */
	63	"ZFS-8000-4J", /* ZPOOL_STATUS_CORRUPT_LABEL_R */
	64	"ZFS-8000-5E", /* ZPOOL_STATUS_CORRUPT_LABEL_NR */
	65	"ZFS-8000-6X", /* ZPOOL_STATUS_BAD_GUID_SUM */
	66	"ZFS-8000-72", /* ZPOOL_STATUS_CORRUPT_POOL */
	67	"ZFS-8000-8A", /* ZPOOL_STATUS_CORRUPT_DATA */
	68	"ZFS-8000-9P", /* ZPOOL_STATUS_FAILING_DEV */
	69	"ZFS-8000-A5", /* ZPOOL_STATUS_VERSION_NEWER */
	70	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_MISMATCH */
	71	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_ACTIVE */
	72	"ZFS-8000-EY", /* ZPOOL_STATUS_HOSTID_REQUIRED */
	73	"ZFS-8000-HC", /* ZPOOL_STATUS_IO_FAILURE_WAIT */
	74	"ZFS-8000-JQ", /* ZPOOL_STATUS_IO_FAILURE_CONTINUE */
	75	"ZFS-8000-MM", /* ZPOOL_STATUS_IO_FAILURE_MMP */
	76	"ZFS-8000-K4", /* ZPOOL_STATUS_BAD_LOG */
	77	"ZFS-8000-ER", /* ZPOOL_STATUS_ERRATA */
	78	/*
	79	* The following results have no message ID.
	80	* ZPOOL_STATUS_UNSUP_FEAT_READ
	81	* ZPOOL_STATUS_UNSUP_FEAT_WRITE
	82	* ZPOOL_STATUS_FAULTED_DEV_R
	83	* ZPOOL_STATUS_FAULTED_DEV_NR
	84	* ZPOOL_STATUS_VERSION_OLDER
	85	* ZPOOL_STATUS_FEAT_DISABLED
	86	* ZPOOL_STATUS_RESILVERING
	87	* ZPOOL_STATUS_OFFLINE_DEV
	88	* ZPOOL_STATUS_REMOVED_DEV
9a49d3f3 BB	89	* ZPOOL_STATUS_REBUILDING
9a49d3f3 BB	90	* ZPOOL_STATUS_REBUILD_SCRUB
658fb802	91	* ZPOOL_STATUS_COMPATIBILITY_ERR
e086db16	92	* ZPOOL_STATUS_INCOMPATIBLE_FEAT
5365b074	93	* ZPOOL_STATUS_OK
5365b074	94	*/
34dc7c2f BB	95	};
	96
	97	#define NMSGID (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0]))
	98
34dc7c2f	99	static int
6fe3498c	100	vdev_missing(vdev_stat_t *vs, uint_t vsc)
34dc7c2f	101	{
1f182103	102	(void) vsc;
6fe3498c RM	103	return (vs->vs_state == VDEV_STATE_CANT_OPEN &&
6fe3498c RM	104	vs->vs_aux == VDEV_AUX_OPEN_FAILED);
34dc7c2f BB	105	}
34dc7c2f BB	106
34dc7c2f	107	static int
6fe3498c	108	vdev_faulted(vdev_stat_t *vs, uint_t vsc)
34dc7c2f	109	{
1f182103	110	(void) vsc;
6fe3498c	111	return (vs->vs_state == VDEV_STATE_FAULTED);
34dc7c2f BB	112	}
34dc7c2f BB	113
34dc7c2f	114	static int
6fe3498c	115	vdev_errors(vdev_stat_t *vs, uint_t vsc)
34dc7c2f	116	{
1f182103	117	(void) vsc;
6fe3498c RM	118	return (vs->vs_state == VDEV_STATE_DEGRADED \|\|
	119	vs->vs_read_errors != 0 \|\| vs->vs_write_errors != 0 \|\|
	120	vs->vs_checksum_errors != 0);
34dc7c2f BB	121	}
34dc7c2f BB	122
34dc7c2f	123	static int
6fe3498c	124	vdev_broken(vdev_stat_t *vs, uint_t vsc)
34dc7c2f	125	{
1f182103	126	(void) vsc;
6fe3498c	127	return (vs->vs_state == VDEV_STATE_CANT_OPEN);
34dc7c2f BB	128	}
34dc7c2f BB	129
34dc7c2f	130	static int
6fe3498c	131	vdev_offlined(vdev_stat_t *vs, uint_t vsc)
34dc7c2f	132	{
1f182103	133	(void) vsc;
6fe3498c	134	return (vs->vs_state == VDEV_STATE_OFFLINE);
34dc7c2f BB	135	}
34dc7c2f BB	136
45d1cae3	137	static int
6fe3498c	138	vdev_removed(vdev_stat_t *vs, uint_t vsc)
45d1cae3	139	{
1f182103	140	(void) vsc;
6fe3498c RM	141	return (vs->vs_state == VDEV_STATE_REMOVED);
	142	}
	143
	144	static int
	145	vdev_non_native_ashift(vdev_stat_t *vs, uint_t vsc)
	146	{
	147	if (getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") != NULL)
	148	return (0);
	149
	150	return (VDEV_STAT_VALID(vs_physical_ashift, vsc) &&
	151	vs->vs_configured_ashift < vs->vs_physical_ashift);
45d1cae3 BB	152	}
45d1cae3 BB	153
34dc7c2f BB	154	/*
	155	* Detect if any leaf devices that have seen errors or could not be opened.
	156	*/
	157	static boolean_t
6fe3498c RM	158	find_vdev_problem(nvlist_t vdev, int (func)(vdev_stat_t *, uint_t),
6fe3498c RM	159	boolean_t ignore_replacing)
34dc7c2f BB	160	{
34dc7c2f BB	161	nvlist_t **child;
8bb9ecf4	162	uint_t c, children;
34dc7c2f BB	163
	164	/*
	165	* Ignore problems within a 'replacing' vdev, since we're presumably in
	166	* the process of repairing any such errors, and don't want to call them
	167	* out again. We'll pick up the fact that a resilver is happening
	168	* later.
	169	*/
6fe3498c	170	if (ignore_replacing == B_TRUE) {
d1807f16 RY	171	const char *type = fnvlist_lookup_string(vdev,
d1807f16 RY	172	ZPOOL_CONFIG_TYPE);
6fe3498c RM	173	if (strcmp(type, VDEV_TYPE_REPLACING) == 0)
	174	return (B_FALSE);
	175	}
34dc7c2f BB	176
	177	if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child,
	178	&children) == 0) {
	179	for (c = 0; c < children; c++)
6fe3498c	180	if (find_vdev_problem(child[c], func, ignore_replacing))
34dc7c2f BB	181	return (B_TRUE);
34dc7c2f BB	182	} else {
8bb9ecf4 RM	183	uint_t vsc;
	184	vdev_stat_t vs = (vdev_stat_t )fnvlist_lookup_uint64_array(
	185	vdev, ZPOOL_CONFIG_VDEV_STATS, &vsc);
6fe3498c	186	if (func(vs, vsc) != 0)
34dc7c2f BB	187	return (B_TRUE);
	188	}
	189
34ffbed8 SH	190	/*
	191	* Check any L2 cache devs
	192	*/
	193	if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_L2CACHE, &child,
	194	&children) == 0) {
	195	for (c = 0; c < children; c++)
6fe3498c	196	if (find_vdev_problem(child[c], func, ignore_replacing))
34ffbed8 SH	197	return (B_TRUE);
	198	}
	199
34dc7c2f BB	200	return (B_FALSE);
	201	}
	202
	203	/*
	204	* Active pool health status.
	205	*
	206	* To determine the status for a pool, we make several passes over the config,
	207	* picking the most egregious error we find. In order of importance, we do the
	208	* following:
	209	*
	210	* - Check for a complete and valid configuration
	211	* - Look for any faulted or missing devices in a non-replicated config
	212	* - Check for any data errors
	213	* - Check for any faulted or missing devices in a replicated config
	214	* - Look for any devices showing errors
9a49d3f3	215	* - Check for any resilvering or rebuilding devices
34dc7c2f BB	216	*
	217	* There can obviously be multiple errors within a single pool, so this routine
	218	* only picks the most damaging of all the current errors to report.
	219	*/
	220	static zpool_status_t
658fb802 CB	221	check_status(nvlist_t *config, boolean_t isimport,
658fb802 CB	222	zpool_errata_t erratap, const char compat)
34dc7c2f	223	{
428870ff BB	224	pool_scan_stat_t *ps = NULL;
428870ff BB	225	uint_t vsc, psc;
b128c09f	226	uint64_t suspended;
34dc7c2f	227	uint64_t hostid = 0;
ffe9d382	228	uint64_t errata = 0;
53698a45	229	unsigned long system_hostid = get_system_hostid();
34dc7c2f	230
8bb9ecf4 RM	231	uint64_t version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
	232	nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
	233	ZPOOL_CONFIG_VDEV_TREE);
	234	vdev_stat_t vs = (vdev_stat_t )fnvlist_lookup_uint64_array(nvroot,
	235	ZPOOL_CONFIG_VDEV_STATS, &vsc);
	236	uint64_t stateval = fnvlist_lookup_uint64(config,
	237	ZPOOL_CONFIG_POOL_STATE);
428870ff BB	238
	239	/*
	240	* Currently resilvering a vdev
	241	*/
	242	(void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
	243	(uint64_t **)&ps, &psc);
d4a72f23	244	if (ps != NULL && ps->pss_func == POOL_SCAN_RESILVER &&
428870ff BB	245	ps->pss_state == DSS_SCANNING)
428870ff BB	246	return (ZPOOL_STATUS_RESILVERING);
34dc7c2f	247
9a49d3f3 BB	248	/*
	249	* Currently rebuilding a vdev, check top-level vdevs.
	250	*/
	251	vdev_rebuild_stat_t *vrs = NULL;
	252	nvlist_t **child;
	253	uint_t c, i, children;
	254	uint64_t rebuild_end_time = 0;
	255	if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
	256	&child, &children) == 0) {
	257	for (c = 0; c < children; c++) {
	258	if ((nvlist_lookup_uint64_array(child[c],
	259	ZPOOL_CONFIG_REBUILD_STATS,
	260	(uint64_t **)&vrs, &i) == 0) && (vrs != NULL)) {
	261	uint64_t state = vrs->vrs_state;
	262
	263	if (state == VDEV_REBUILD_ACTIVE) {
	264	return (ZPOOL_STATUS_REBUILDING);
	265	} else if (state == VDEV_REBUILD_COMPLETE &&
	266	vrs->vrs_end_time > rebuild_end_time) {
	267	rebuild_end_time = vrs->vrs_end_time;
	268	}
	269	}
	270	}
	271
	272	/*
	273	* If we can determine when the last scrub was run, and it
	274	* was before the last rebuild completed, then recommend
	275	* that the pool be scrubbed to verify all checksums. When
	276	* ps is NULL we can infer the pool has never been scrubbed.
	277	*/
	278	if (rebuild_end_time > 0) {
	279	if (ps != NULL) {
	280	if ((ps->pss_state == DSS_FINISHED &&
	281	ps->pss_func == POOL_SCAN_SCRUB &&
	282	rebuild_end_time > ps->pss_end_time) \|\|
	283	ps->pss_state == DSS_NONE)
	284	return (ZPOOL_STATUS_REBUILD_SCRUB);
	285	} else {
	286	return (ZPOOL_STATUS_REBUILD_SCRUB);
	287	}
	288	}
	289	}
	290
379ca9cf OF	291	/*
	292	* The multihost property is set and the pool may be active.
	293	*/
	294	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	295	vs->vs_aux == VDEV_AUX_ACTIVE) {
	296	mmp_state_t mmp_state;
	297	nvlist_t *nvinfo;
	298
	299	nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
	300	mmp_state = fnvlist_lookup_uint64(nvinfo,
	301	ZPOOL_CONFIG_MMP_STATE);
	302
	303	if (mmp_state == MMP_STATE_ACTIVE)
	304	return (ZPOOL_STATUS_HOSTID_ACTIVE);
	305	else if (mmp_state == MMP_STATE_NO_HOSTID)
	306	return (ZPOOL_STATUS_HOSTID_REQUIRED);
	307	else
	308	return (ZPOOL_STATUS_HOSTID_MISMATCH);
	309	}
	310
34dc7c2f BB	311	/*
	312	* Pool last accessed by another system.
	313	*/
428870ff	314	(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
492b8e9e	315	if (hostid != 0 && (unsigned long)hostid != system_hostid &&
34dc7c2f BB	316	stateval == POOL_STATE_ACTIVE)
	317	return (ZPOOL_STATUS_HOSTID_MISMATCH);
	318
	319	/*
	320	* Newer on-disk version.
	321	*/
	322	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	323	vs->vs_aux == VDEV_AUX_VERSION_NEWER)
	324	return (ZPOOL_STATUS_VERSION_NEWER);
	325
9ae529ec CS	326	/*
	327	* Unsupported feature(s).
	328	*/
	329	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	330	vs->vs_aux == VDEV_AUX_UNSUP_FEAT) {
8bb9ecf4 RM	331	nvlist_t *nvinfo = fnvlist_lookup_nvlist(config,
8bb9ecf4 RM	332	ZPOOL_CONFIG_LOAD_INFO);
9ae529ec CS	333	if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY))
	334	return (ZPOOL_STATUS_UNSUP_FEAT_WRITE);
	335	return (ZPOOL_STATUS_UNSUP_FEAT_READ);
	336	}
	337
34dc7c2f BB	338	/*
	339	* Check that the config is complete.
	340	*/
	341	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	342	vs->vs_aux == VDEV_AUX_BAD_GUID_SUM)
	343	return (ZPOOL_STATUS_BAD_GUID_SUM);
	344
b128c09f	345	/*
cec3a0a1	346	* Check whether the pool has suspended.
b128c09f BB	347	*/
	348	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED,
	349	&suspended) == 0) {
cec3a0a1 OF	350	uint64_t reason;
	351
	352	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED_REASON,
	353	&reason) == 0 && reason == ZIO_SUSPEND_MMP)
	354	return (ZPOOL_STATUS_IO_FAILURE_MMP);
	355
b128c09f BB	356	if (suspended == ZIO_FAILURE_MODE_CONTINUE)
	357	return (ZPOOL_STATUS_IO_FAILURE_CONTINUE);
	358	return (ZPOOL_STATUS_IO_FAILURE_WAIT);
	359	}
	360
	361	/*
	362	* Could not read a log.
	363	*/
	364	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	365	vs->vs_aux == VDEV_AUX_BAD_LOG) {
	366	return (ZPOOL_STATUS_BAD_LOG);
	367	}
	368
34dc7c2f BB	369	/*
	370	* Bad devices in non-replicated config.
	371	*/
	372	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
6fe3498c	373	find_vdev_problem(nvroot, vdev_faulted, B_TRUE))
34dc7c2f BB	374	return (ZPOOL_STATUS_FAULTED_DEV_NR);
	375
	376	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
6fe3498c	377	find_vdev_problem(nvroot, vdev_missing, B_TRUE))
34dc7c2f BB	378	return (ZPOOL_STATUS_MISSING_DEV_NR);
	379
	380	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
6fe3498c	381	find_vdev_problem(nvroot, vdev_broken, B_TRUE))
34dc7c2f BB	382	return (ZPOOL_STATUS_CORRUPT_LABEL_NR);
	383
	384	/*
	385	* Corrupted pool metadata
	386	*/
	387	if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
	388	vs->vs_aux == VDEV_AUX_CORRUPT_DATA)
	389	return (ZPOOL_STATUS_CORRUPT_POOL);
	390
	391	/*
	392	* Persistent data errors.
	393	*/
	394	if (!isimport) {
018f2604	395	uint64_t nerr;
34dc7c2f BB	396	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
	397	&nerr) == 0 && nerr != 0)
	398	return (ZPOOL_STATUS_CORRUPT_DATA);
	399	}
	400
	401	/*
	402	* Missing devices in a replicated config.
	403	*/
6fe3498c	404	if (find_vdev_problem(nvroot, vdev_faulted, B_TRUE))
34dc7c2f	405	return (ZPOOL_STATUS_FAULTED_DEV_R);
6fe3498c	406	if (find_vdev_problem(nvroot, vdev_missing, B_TRUE))
34dc7c2f	407	return (ZPOOL_STATUS_MISSING_DEV_R);
6fe3498c	408	if (find_vdev_problem(nvroot, vdev_broken, B_TRUE))
34dc7c2f BB	409	return (ZPOOL_STATUS_CORRUPT_LABEL_R);
	410
	411	/*
	412	* Devices with errors
	413	*/
6fe3498c	414	if (!isimport && find_vdev_problem(nvroot, vdev_errors, B_TRUE))
34dc7c2f BB	415	return (ZPOOL_STATUS_FAILING_DEV);
	416
	417	/*
	418	* Offlined devices
	419	*/
6fe3498c	420	if (find_vdev_problem(nvroot, vdev_offlined, B_TRUE))
34dc7c2f BB	421	return (ZPOOL_STATUS_OFFLINE_DEV);
34dc7c2f BB	422
45d1cae3 BB	423	/*
	424	* Removed device
	425	*/
6fe3498c	426	if (find_vdev_problem(nvroot, vdev_removed, B_TRUE))
45d1cae3 BB	427	return (ZPOOL_STATUS_REMOVED_DEV);
45d1cae3 BB	428
6fe3498c RM	429	/*
	430	* Suboptimal, but usable, ashift configuration.
	431	*/
	432	if (find_vdev_problem(nvroot, vdev_non_native_ashift, B_FALSE))
	433	return (ZPOOL_STATUS_NON_NATIVE_ASHIFT);
	434
ae76f45c TC	435	/*
	436	* Informational errata available.
	437	*/
	438	(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRATA, &errata);
	439	if (errata) {
	440	*erratap = errata;
	441	return (ZPOOL_STATUS_ERRATA);
	442	}
	443
34dc7c2f BB	444	/*
	445	* Outdated, but usable, version
	446	*/
e086db16 CB	447	if (SPA_VERSION_IS_SUPPORTED(version) && version != SPA_VERSION) {
	448	/* "legacy" compatibility disables old version reporting */
	449	if (compat != NULL && strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0)
	450	return (ZPOOL_STATUS_OK);
	451	else
	452	return (ZPOOL_STATUS_VERSION_OLDER);
	453	}
34dc7c2f	454
b9b24bb4	455	/*
e086db16 CB	456	* Usable pool with disabled or superfluous features
e086db16 CB	457	* (superfluous = beyond what's requested by 'compatibility')
b9b24bb4 CS	458	*/
	459	if (version >= SPA_VERSION_FEATURES) {
	460	int i;
	461	nvlist_t *feat;
	462
	463	if (isimport) {
	464	feat = fnvlist_lookup_nvlist(config,
	465	ZPOOL_CONFIG_LOAD_INFO);
379ca9cf OF	466	if (nvlist_exists(feat, ZPOOL_CONFIG_ENABLED_FEAT))
	467	feat = fnvlist_lookup_nvlist(feat,
	468	ZPOOL_CONFIG_ENABLED_FEAT);
b9b24bb4 CS	469	} else {
	470	feat = fnvlist_lookup_nvlist(config,
	471	ZPOOL_CONFIG_FEATURE_STATS);
	472	}
	473
658fb802	474	/* check against all features, or limited set? */
e086db16	475	boolean_t c_features[SPA_FEATURES];
658fb802	476
e086db16 CB	477	switch (zpool_load_compat(compat, c_features, NULL, 0)) {
	478	case ZPOOL_COMPATIBILITY_OK:
	479	case ZPOOL_COMPATIBILITY_WARNTOKEN:
	480	break;
	481	default:
658fb802	482	return (ZPOOL_STATUS_COMPATIBILITY_ERR);
e086db16	483	}
b9b24bb4 CS	484	for (i = 0; i < SPA_FEATURES; i++) {
b9b24bb4 CS	485	zfeature_info_t *fi = &spa_feature_table[i];
03ef8f09 MM	486	if (!fi->fi_zfs_mod_supported)
03ef8f09 MM	487	continue;
e086db16	488	if (c_features[i] && !nvlist_exists(feat, fi->fi_guid))
b9b24bb4	489	return (ZPOOL_STATUS_FEAT_DISABLED);
e086db16 CB	490	if (!c_features[i] && nvlist_exists(feat, fi->fi_guid))
e086db16 CB	491	return (ZPOOL_STATUS_INCOMPATIBLE_FEAT);
b9b24bb4 CS	492	}
	493	}
	494
34dc7c2f BB	495	return (ZPOOL_STATUS_OK);
	496	}
	497
	498	zpool_status_t
a926aab9 AZ	499	zpool_get_status(zpool_handle_t zhp, const char *msgid,
a926aab9 AZ	500	zpool_errata_t *errata)
34dc7c2f	501	{
658fb802 CB	502	/*
	503	* pass in the desired feature set, as
	504	* it affects check for disabled features
	505	*/
	506	char compatibility[ZFS_MAXPROPLEN];
	507	if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compatibility,
	508	ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
	509	compatibility[0] = '\0';
	510
	511	zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE, errata,
	512	compatibility);
	513
f0ed6c74 TH	514	if (msgid != NULL) {
	515	if (ret >= NMSGID)
	516	*msgid = NULL;
	517	else
	518	*msgid = zfs_msgid_table[ret];
	519	}
34dc7c2f BB	520	return (ret);
	521	}
	522
	523	zpool_status_t
a926aab9 AZ	524	zpool_import_status(nvlist_t config, const char *msgid,
a926aab9 AZ	525	zpool_errata_t *errata)
34dc7c2f	526	{
658fb802	527	zpool_status_t ret = check_status(config, B_TRUE, errata, NULL);
34dc7c2f BB	528
	529	if (ret >= NMSGID)
	530	*msgid = NULL;
	531	else
	532	*msgid = zfs_msgid_table[ret];
	533
	534	return (ret);
	535	}