zpool_vdev.c revision 99653d4ee642c6528e88224f12409a5f23060994
fa9e4066f08beec538e775443c5be79dd423fcabahrens * CDDL HEADER START
fa9e4066f08beec538e775443c5be79dd423fcabahrens * The contents of this file are subject to the terms of the
ea8dc4b6d2251b437950c0056bc626b311c73c27eschrock * Common Development and Distribution License (the "License").
ea8dc4b6d2251b437950c0056bc626b311c73c27eschrock * You may not use this file except in compliance with the License.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
fa9e4066f08beec538e775443c5be79dd423fcabahrens * See the License for the specific language governing permissions
fa9e4066f08beec538e775443c5be79dd423fcabahrens * and limitations under the License.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * When distributing Covered Code, include this CDDL HEADER in each
fa9e4066f08beec538e775443c5be79dd423fcabahrens * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If applicable, add the following below this CDDL HEADER, with the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * fields enclosed by brackets "[]" replaced with your own identifying
fa9e4066f08beec538e775443c5be79dd423fcabahrens * information: Portions Copyright [yyyy] [name of copyright owner]
fa9e4066f08beec538e775443c5be79dd423fcabahrens * CDDL HEADER END
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock * Copyright 2006 Sun Microsystems, Inc. All rights reserved.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Use is subject to license terms.
fa9e4066f08beec538e775443c5be79dd423fcabahrens#pragma ident "%Z%%M% %I% %E% SMI"
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Functions to convert between a list of vdevs and an nvlist representing the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * configuration. Each entry in the list can be one of:
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Device vdevs
fa9e4066f08beec538e775443c5be79dd423fcabahrens * disk=(path=..., devid=...)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * file=(path=...)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Group vdevs
99653d4ee642c6528e88224f12409a5f23060994eschrock * raidz[1|2]=(...)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * mirror=(...)
99653d4ee642c6528e88224f12409a5f23060994eschrock * Hot spares
fa9e4066f08beec538e775443c5be79dd423fcabahrens * While the underlying implementation supports it, group vdevs cannot contain
fa9e4066f08beec538e775443c5be79dd423fcabahrens * other group vdevs. All userland verification of devices is contained within
fa9e4066f08beec538e775443c5be79dd423fcabahrens * this file. If successful, the nvlist returned can be passed directly to the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * kernel; we've done as much verification as possible in userland.
99653d4ee642c6528e88224f12409a5f23060994eschrock * Hot spares are a special case, and passed down as an array of disk vdevs, at
99653d4ee642c6528e88224f12409a5f23060994eschrock * the same level as the root of the vdev tree.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * The only function exported by this file is 'get_vdev_spec'. The function
fa9e4066f08beec538e775443c5be79dd423fcabahrens * performs several passes:
fa9e4066f08beec538e775443c5be79dd423fcabahrens * 1. Construct the vdev specification. Performs syntax validation and
fa9e4066f08beec538e775443c5be79dd423fcabahrens * makes sure each device is valid.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * 2. Check for devices in use. Using libdiskmgt, makes sure that no
fa9e4066f08beec538e775443c5be79dd423fcabahrens * devices are also in use. Some can be overridden using the 'force'
fa9e4066f08beec538e775443c5be79dd423fcabahrens * flag, others cannot.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * 3. Check for replication errors if the 'force' flag is not specified.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * validates that the replication level is consistent across the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * entire pool.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * 4. Label any whole disks with an EFI label.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * For any given vdev specification, we can have multiple errors. The
fa9e4066f08beec538e775443c5be79dd423fcabahrens * vdev_error() function keeps track of whether we have seen an error yet, and
fa9e4066f08beec538e775443c5be79dd423fcabahrens * prints out a header if its the first error we've seen.
99653d4ee642c6528e88224f12409a5f23060994eschrock/*PRINTFLIKE1*/
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("invalid vdev specification\n"));
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("use '-f' to override "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "the following errors:\n"));
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("the following errors "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "must be manually repaired:\n"));
99653d4ee642c6528e88224f12409a5f23060994eschrock * ENXIO/ENODEV is a valid error message if the device doesn't live in
ea8dc4b6d2251b437950c0056bc626b311c73c27eschrock * /dev/dsk. Don't bother printing an error message in this case.
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock (void) fprintf(stderr, gettext("warning: device in use checking "
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock * Validate a device, passing the bulk of the work off to libdiskmgt.
99653d4ee642c6528e88224f12409a5f23060994eschrockcheck_slice(const char *path, int force, boolean_t wholedisk, boolean_t isspare)
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock force ? DM_WHO_ZPOOL_FORCE : DM_WHO_ZPOOL, &error) || error) {
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock return (0);
99653d4ee642c6528e88224f12409a5f23060994eschrock } else if (!isspare ||
99653d4ee642c6528e88224f12409a5f23060994eschrock * The above check is a rather severe hack. It would
99653d4ee642c6528e88224f12409a5f23060994eschrock * probably make more sense to have DM_WHO_ZPOOL_SPARE
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock * If we're given a whole disk, ignore overlapping slices since we're
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock * about to label it anyway.
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock (dm_isoverlapping((char *)path, &msg, &error) || error)) {
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Validate a whole disk. Iterate over all slices on the disk and make sure
fa9e4066f08beec538e775443c5be79dd423fcabahrens * that none is in use by calling check_slice().
fa9e4066f08beec538e775443c5be79dd423fcabahrens/* ARGSUSED */
99653d4ee642c6528e88224f12409a5f23060994eschrockcheck_disk(const char *name, dm_descriptor_t disk, int force, int isspare)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Get the drive associated with this disk. This should never fail,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * because we already have an alias handle open for the device.
fa9e4066f08beec538e775443c5be79dd423fcabahrens if ((drive = dm_get_associated_descriptors(disk, DM_DRIVE,
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens if ((media = dm_get_associated_descriptors(*drive, DM_MEDIA,
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * It is possible that the user has specified a removable media drive,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * and the media is not present.
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock vdev_error(gettext("'%s' has no media in drive\n"), name);
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens if ((slice = dm_get_associated_descriptors(*media, DM_SLICE,
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Iterate over all slices and report any errors. We don't care about
fa9e4066f08beec538e775443c5be79dd423fcabahrens * overlapping slices because we are using the whole disk.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (check_slice(name, force, B_TRUE, isspare) != 0)
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock * Validate a device.
99653d4ee642c6528e88224f12409a5f23060994eschrockcheck_device(const char *path, boolean_t force, boolean_t isspare)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * For whole disks, libdiskmgt does not include the leading dev path.
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock if ((desc = dm_get_descriptor_by_name(DM_ALIAS, dev, &err)) != NULL) {
99653d4ee642c6528e88224f12409a5f23060994eschrock return (check_slice(path, force, B_FALSE, isspare));
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Check that a file is valid. All we can do in this case is check that it's
fa9e4066f08beec538e775443c5be79dd423fcabahrens * not in use by another pool.
99653d4ee642c6528e88224f12409a5f23060994eschrockcheck_file(const char *file, boolean_t force, boolean_t isspare)
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
99653d4ee642c6528e88224f12409a5f23060994eschrock if (zpool_in_use(g_zfs, fd, &state, &name, &inuse) == 0 && inuse) {
46a2abf27af40eda17a3f97e79eda1aef4e3c3c8eschrock const char *desc;
99653d4ee642c6528e88224f12409a5f23060994eschrock * Allow hot spares to be shared between pools.
99653d4ee642c6528e88224f12409a5f23060994eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrensis_whole_disk(const char *arg, struct stat64 *statbuf)
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) snprintf(path, sizeof (path), "%s%s", arg, BACKUP_SLICE);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Create a leaf vdev. Determine if this is a file or a device. If it's a
fa9e4066f08beec538e775443c5be79dd423fcabahrens * device, fill in the device id to make a complete nvlist. Valid forms for a
fa9e4066f08beec538e775443c5be79dd423fcabahrens * leaf vdev are:
fa9e4066f08beec538e775443c5be79dd423fcabahrens * /dev/dsk/xxx Complete disk path
fa9e4066f08beec538e775443c5be79dd423fcabahrens * /xxx Full path to file
fa9e4066f08beec538e775443c5be79dd423fcabahrens * xxx Shorthand for /dev/dsk/xxx
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Determine what type of vdev this is, and put the full path into
fa9e4066f08beec538e775443c5be79dd423fcabahrens * 'path'. We detect whether this is a device of file afterwards by
fa9e4066f08beec538e775443c5be79dd423fcabahrens * checking the st_mode of the file.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Complete device or file path. Exact type is determined by
fa9e4066f08beec538e775443c5be79dd423fcabahrens * examining the file descriptor afterwards.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * This may be a short path for a device, or it could be total
fa9e4066f08beec538e775443c5be79dd423fcabahrens * gibberish. Check to see if it's a known device in
fa9e4066f08beec538e775443c5be79dd423fcabahrens * /dev/dsk/. As part of this check, see if we've been given a
fa9e4066f08beec538e775443c5be79dd423fcabahrens * an entire disk (minus the slice number).
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) snprintf(path, sizeof (path), "%s/%s", DISK_ROOT,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If we got ENOENT, then the user gave us
fa9e4066f08beec538e775443c5be79dd423fcabahrens * gibberish, so try to direct them with a
fa9e4066f08beec538e775443c5be79dd423fcabahrens * reasonable error message. Otherwise,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * regurgitate strerror() since it's the best we
fa9e4066f08beec538e775443c5be79dd423fcabahrens "shorthand device name\n"));
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Determine whether this is a device or a file.
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("cannot use '%s': must be a "
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Finally, we have the complete device or file, and we know that it is
fa9e4066f08beec538e775443c5be79dd423fcabahrens * acceptable to use. Construct the nvlist to describe this vdev. All
fa9e4066f08beec538e775443c5be79dd423fcabahrens * vdevs have a 'path' element, and devices also have a 'devid' element.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_add_string(vdev, ZPOOL_CONFIG_PATH, path) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE, type) == 0);
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_WHOLE_DISK,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * For a whole disk, defer getting its devid until after labeling it.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Get the devid for the device.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Go through and verify the replication level of the pool is consistent.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Performs the following checks:
fa9e4066f08beec538e775443c5be79dd423fcabahrens * For the new spec, verifies that devices in mirrors and raidz are the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * same size.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If the current configuration already has inconsistent replication
fa9e4066f08beec538e775443c5be79dd423fcabahrens * levels, ignore any other potential problems in the new spec.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Otherwise, make sure that the current spec (if there is one) and the new
fa9e4066f08beec538e775443c5be79dd423fcabahrens * spec have consistent replication levels.
fa9e4066f08beec538e775443c5be79dd423fcabahrenstypedef struct replication_level {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Given a list of toplevel vdevs, return the current replication level. If
fa9e4066f08beec538e775443c5be79dd423fcabahrens * the config is inconsistent, then NULL is returned. If 'fatal' is set, then
fa9e4066f08beec538e775443c5be79dd423fcabahrens * an error message will be displayed for each self-inconsistent vdev.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
fa9e4066f08beec538e775443c5be79dd423fcabahrens for (t = 0; t < toplevels; t++) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * This is a 'file' or 'disk' vdev.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * This is a mirror or RAID-Z vdev. Go through and make
fa9e4066f08beec538e775443c5be79dd423fcabahrens * sure the contents are all the same (files vs. disks),
fa9e4066f08beec538e775443c5be79dd423fcabahrens * keeping track of the number of elements in the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * We also check that the size of each vdev (if it can
fa9e4066f08beec538e775443c5be79dd423fcabahrens * be determined) is the same.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * The 'dontreport' variable indicatest that we've
fa9e4066f08beec538e775443c5be79dd423fcabahrens * already reported an error for this spec, so don't
fa9e4066f08beec538e775443c5be79dd423fcabahrens * bother doing it again.
fa9e4066f08beec538e775443c5be79dd423fcabahrens for (c = 0; c < children; c++) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If we have a raidz/mirror that combines disks
fa9e4066f08beec538e775443c5be79dd423fcabahrens * with files, report it as an error.
fa9e4066f08beec538e775443c5be79dd423fcabahrens "mismatched replication "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "level: %s contains both "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "files and devices\n"),
fa9e4066f08beec538e775443c5be79dd423fcabahrens * According to stat(2), the value of 'st_size'
fa9e4066f08beec538e775443c5be79dd423fcabahrens * is undefined for block devices and character
fa9e4066f08beec538e775443c5be79dd423fcabahrens * devices. But there is no effective way to
fa9e4066f08beec538e775443c5be79dd423fcabahrens * determine the real size in userland.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Instead, we'll take advantage of an
fa9e4066f08beec538e775443c5be79dd423fcabahrens * implementation detail of spec_size(). If the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * device is currently open, then we (should)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * return a valid size.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If we still don't get a valid size (indicated
fa9e4066f08beec538e775443c5be79dd423fcabahrens * by a size of 0 or MAXOFFSET_T), then ignore
fa9e4066f08beec538e775443c5be79dd423fcabahrens * this device altogether.
fa9e4066f08beec538e775443c5be79dd423fcabahrens if (err != 0 ||
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Also check the size of each device. If they
fa9e4066f08beec538e775443c5be79dd423fcabahrens * differ, then report an error.
fa9e4066f08beec538e775443c5be79dd423fcabahrens "%s contains devices of "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "different sizes\n"),
fa9e4066f08beec538e775443c5be79dd423fcabahrens * At this point, we have the replication of the last toplevel
fa9e4066f08beec538e775443c5be79dd423fcabahrens * vdev in 'rep'. Compare it to 'lastrep' to see if its
fa9e4066f08beec538e775443c5be79dd423fcabahrens * different.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (strcmp(lastrep.zprl_type, rep.zprl_type) != 0) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: "
99653d4ee642c6528e88224f12409a5f23060994eschrock "both %s and %s vdevs are "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "present\n"),
99653d4ee642c6528e88224f12409a5f23060994eschrock } else if (lastrep.zprl_parity != rep.zprl_parity) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: "
99653d4ee642c6528e88224f12409a5f23060994eschrock "both %llu and %llu device parity "
99653d4ee642c6528e88224f12409a5f23060994eschrock "%s vdevs are present\n"),
99653d4ee642c6528e88224f12409a5f23060994eschrock } else if (lastrep.zprl_children != rep.zprl_children) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: "
99653d4ee642c6528e88224f12409a5f23060994eschrock "both %llu-way and %llu-way %s "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "vdevs are present\n"),
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Check the replication level of the vdev spec against the current pool. Calls
fa9e4066f08beec538e775443c5be79dd423fcabahrens * get_replication() to make sure the new spec is self-consistent. If the pool
fa9e4066f08beec538e775443c5be79dd423fcabahrens * has a consistent replication level, then we ignore any errors. Otherwise,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * report any difference between the two.
fa9e4066f08beec538e775443c5be79dd423fcabahrenscheck_replication(nvlist_t *config, nvlist_t *newroot)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If we have a current pool configuration, check to see if it's
fa9e4066f08beec538e775443c5be79dd423fcabahrens * self-consistent. If not, simply return success.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
99653d4ee642c6528e88224f12409a5f23060994eschrock if ((current = get_replication(nvroot, B_FALSE)) == NULL)
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Get the replication level of the new vdev spec, reporting any
fa9e4066f08beec538e775443c5be79dd423fcabahrens * inconsistencies found.
99653d4ee642c6528e88224f12409a5f23060994eschrock if ((new = get_replication(newroot, B_TRUE)) == NULL) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Check to see if the new vdev spec matches the replication level of
fa9e4066f08beec538e775443c5be79dd423fcabahrens * the current pool.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (strcmp(current->zprl_type, new->zprl_type) != 0) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: pool uses %s "
99653d4ee642c6528e88224f12409a5f23060994eschrock "and new vdev is %s\n"),
99653d4ee642c6528e88224f12409a5f23060994eschrock } else if (current->zprl_parity != new->zprl_parity) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: pool uses %llu "
99653d4ee642c6528e88224f12409a5f23060994eschrock "device parity and new vdev uses %llu\n"),
99653d4ee642c6528e88224f12409a5f23060994eschrock } else if (current->zprl_children != new->zprl_children) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "mismatched replication level: pool uses %llu-way "
99653d4ee642c6528e88224f12409a5f23060994eschrock "%s and new vdev uses %llu-way %s\n"),
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Label an individual disk. The name provided is the short name, stripped of
fa9e4066f08beec538e775443c5be79dd423fcabahrens * any leading /dev path.
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) snprintf(path, sizeof (path), "%s/%s%s", RDISK_ROOT, name,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * This shouldn't happen. We've long since verified that this
fa9e4066f08beec538e775443c5be79dd423fcabahrens * is a valid device.
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("cannot open '%s': %s\n"),
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * The only way this can fail is if we run out of memory, or we
fa9e4066f08beec538e775443c5be79dd423fcabahrens * were unable to read the disk geometry.
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("cannot label '%s': unable to "
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens vtoc->efi_parts[0].p_size = vtoc->efi_last_u_lba + 1 -
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Why we use V_USR: V_BACKUP confuses users, and is considered
fa9e4066f08beec538e775443c5be79dd423fcabahrens * disposable by some EFI utilities (since EFI doesn't have a backup
fa9e4066f08beec538e775443c5be79dd423fcabahrens * slice). V_UNASSIGNED is supposed to be used only for zero size
fa9e4066f08beec538e775443c5be79dd423fcabahrens * partitions, and efi_write() will fail if we use it. V_ROOT, V_BOOT,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * etc. were all pretty specific. V_USR is as close to reality as we
fa9e4066f08beec538e775443c5be79dd423fcabahrens * can get, in the absence of V_OTHER.
fa9e4066f08beec538e775443c5be79dd423fcabahrens vtoc->efi_parts[8].p_start = vtoc->efi_last_u_lba + 1 - resv;
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Currently, EFI labels are not supported for IDE disks, and it
fa9e4066f08beec538e775443c5be79dd423fcabahrens * is likely that they will not be supported on other drives for
fa9e4066f08beec538e775443c5be79dd423fcabahrens * some time. Print out a helpful error message directing the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * user to manually label the disk and give a specific slice.
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("cannot label '%s': failed to "
fa9e4066f08beec538e775443c5be79dd423fcabahrens (void) fprintf(stderr, gettext("use fdisk(1M) to partition "
fa9e4066f08beec538e775443c5be79dd423fcabahrens "the disk, and provide a specific slice\n"));
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Go through and find any whole disks in the vdev specification, labelling them
fa9e4066f08beec538e775443c5be79dd423fcabahrens * as appropriate. When constructing the vdev spec, we were unable to open this
fa9e4066f08beec538e775443c5be79dd423fcabahrens * device in order to provide a devid. Now that we have labelled the disk and
fa9e4066f08beec538e775443c5be79dd423fcabahrens * know that slice 0 is valid, we can construct the devid now.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If the disk was already labelled with an EFI label, we will have gotten the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * devid already (because we were able to open the whole disk). Otherwise, we
fa9e4066f08beec538e775443c5be79dd423fcabahrens * need to get the devid after we label the disk.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * We have a disk device. Get the path to the device
fa9e4066f08beec538e775443c5be79dd423fcabahrens * and see if its a whole disk by appending the backup
fa9e4066f08beec538e775443c5be79dd423fcabahrens * slice and stat()ing the device.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0);
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Fill in the devid, now that we've labeled the disk.
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (-1);
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock * Update the path to refer to the 's0' slice. The presence of
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock * the 'whole_disk' field indicates to the CLI that we should
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock * chop off the slice number when displaying the device in
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock * future output.
afefbcddfd8caf5f3b2da510d9439471ab225040eschrock verify(nvlist_add_string(nv, ZPOOL_CONFIG_PATH, buf) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens for (c = 0; c < children; c++)
99653d4ee642c6528e88224f12409a5f23060994eschrock if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
99653d4ee642c6528e88224f12409a5f23060994eschrock for (c = 0; c < children; c++)
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
99653d4ee642c6528e88224f12409a5f23060994eschrock * Determine if the given path is a hot spare within the given configuration.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (zpool_in_use(g_zfs, fd, &state, &name, &inuse) != 0 ||
99653d4ee642c6528e88224f12409a5f23060994eschrock verify(nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) == 0);
99653d4ee642c6528e88224f12409a5f23060994eschrock verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
99653d4ee642c6528e88224f12409a5f23060994eschrock if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
99653d4ee642c6528e88224f12409a5f23060994eschrock for (i = 0; i < nspares; i++) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Go through and find any devices that are in use. We rely on libdiskmgt for
fa9e4066f08beec538e775443c5be79dd423fcabahrens * the majority of this task.
99653d4ee642c6528e88224f12409a5f23060994eschrockcheck_in_use(nvlist_t *config, nvlist_t *nv, int force, int isreplacing,
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0);
99653d4ee642c6528e88224f12409a5f23060994eschrock * As a generic check, we look to see if this is a replace of a
99653d4ee642c6528e88224f12409a5f23060994eschrock * hot spare within the same pool. If so, we allow it
99653d4ee642c6528e88224f12409a5f23060994eschrock * regardless of what libdiskmgt or zpool_in_use() says.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
99653d4ee642c6528e88224f12409a5f23060994eschrock return (0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens for (c = 0; c < children; c++)
99653d4ee642c6528e88224f12409a5f23060994eschrock if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
99653d4ee642c6528e88224f12409a5f23060994eschrock for (c = 0; c < children; c++)
fa9e4066f08beec538e775443c5be79dd423fcabahrens return (0);
99653d4ee642c6528e88224f12409a5f23060994eschrockconst char *
99653d4ee642c6528e88224f12409a5f23060994eschrock if (strcmp(type, "raidz") == 0 || strcmp(type, "raidz1") == 0) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Construct a syntactically valid vdev specification,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * and ensure that all devices and files exist and can be opened.
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Note: we don't bother freeing anything in the error paths
fa9e4066f08beec538e775443c5be79dd423fcabahrens * because the program is just going to exit anyway.
99653d4ee642c6528e88224f12409a5f23060994eschrock const char *type;
fa9e4066f08beec538e775443c5be79dd423fcabahrens while (argc > 0) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * If it's a mirror or raidz, the subsequent arguments are
fa9e4066f08beec538e775443c5be79dd423fcabahrens * its leaves -- until we encounter the next mirror or raidz.
99653d4ee642c6528e88224f12409a5f23060994eschrock if ((type = is_grouping(argv[0], &mindev)) != NULL) {
99653d4ee642c6528e88224f12409a5f23060994eschrock "specification: 'spare' can be "
99653d4ee642c6528e88224f12409a5f23060994eschrock "specified only once\n"));
99653d4ee642c6528e88224f12409a5f23060994eschrock "specification: %s requires at least %d "
99653d4ee642c6528e88224f12409a5f23060994eschrock for (c = 0; c < children; c++)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * We have a device. Pass off to make_leaf_vdev() to
fa9e4066f08beec538e775443c5be79dd423fcabahrens * construct the appropriate nvlist describing the vdev.
99653d4ee642c6528e88224f12409a5f23060994eschrock "specification: at least one toplevel vdev must be "
99653d4ee642c6528e88224f12409a5f23060994eschrock "specified\n"));
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Finally, create nvroot and add all top-level vdevs to it.
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_alloc(&nvroot, NV_UNIQUE_NAME, 0) == 0);
fa9e4066f08beec538e775443c5be79dd423fcabahrens verify(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
99653d4ee642c6528e88224f12409a5f23060994eschrock verify(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
fa9e4066f08beec538e775443c5be79dd423fcabahrens for (t = 0; t < toplevels; t++)
99653d4ee642c6528e88224f12409a5f23060994eschrock for (t = 0; t < nspares; t++)
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Get and validate the contents of the given vdev specification. This ensures
fa9e4066f08beec538e775443c5be79dd423fcabahrens * that the nvlist returned is well-formed, that all the devices exist, and that
fa9e4066f08beec538e775443c5be79dd423fcabahrens * they are not currently in use by any other known consumer. The 'poolconfig'
fa9e4066f08beec538e775443c5be79dd423fcabahrens * parameter is the current configuration of the pool when adding devices
fa9e4066f08beec538e775443c5be79dd423fcabahrens * existing pool, and is used to perform additional checks, such as changing the
fa9e4066f08beec538e775443c5be79dd423fcabahrens * replication level of the pool. It can be 'NULL' to indicate that this is a
fa9e4066f08beec538e775443c5be79dd423fcabahrens * new pool. The 'force' flag controls whether devices should be forcefully
fa9e4066f08beec538e775443c5be79dd423fcabahrens * added, even if they appear in use.
fa9e4066f08beec538e775443c5be79dd423fcabahrensmake_root_vdev(nvlist_t *poolconfig, int force, int check_rep,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Construct the vdev specification. If this is successful, we know
fa9e4066f08beec538e775443c5be79dd423fcabahrens * that we have a valid specification, and that all devices can be
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Validate each device to make sure that its not shared with another
fa9e4066f08beec538e775443c5be79dd423fcabahrens * subsystem. We do this even if 'force' is set, because there are some
fa9e4066f08beec538e775443c5be79dd423fcabahrens * uses (such as a dedicated dump device) that even '-f' cannot
fa9e4066f08beec538e775443c5be79dd423fcabahrens * override.
99653d4ee642c6528e88224f12409a5f23060994eschrock if (check_in_use(poolconfig, newroot, force, isreplacing,
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Check the replication level of the given vdevs and report any errors
fa9e4066f08beec538e775443c5be79dd423fcabahrens * found. We include the existing pool spec, if any, as we need to
fa9e4066f08beec538e775443c5be79dd423fcabahrens * catch changes against the existing replication level.
fa9e4066f08beec538e775443c5be79dd423fcabahrens if (check_rep && check_replication(poolconfig, newroot) != 0) {
fa9e4066f08beec538e775443c5be79dd423fcabahrens * Run through the vdev specification and label any whole disks found.