1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-21 13:34:40 +03:00
lvm2/tools/lvconvert.c

4802 lines
135 KiB
C
Raw Normal View History

/*
2016-02-25 17:01:12 +03:00
* Copyright (C) 2005-2016 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
* This copyrighted material is made available to anyone wishing to use,
* modify, copy, or redistribute it subject to the terms and conditions
* of the GNU Lesser General Public License v.2.1.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "tools.h"
2007-12-22 15:13:29 +03:00
#include "polldaemon.h"
2005-11-29 21:20:23 +03:00
#include "lv_alloc.h"
#include "lvconvert_poll.h"
#define MAX_PDATA_ARGS 10 /* Max number of accepted args for d-m-p-d tools */
typedef enum {
/* Split:
* For a mirrored or raid LV, split mirror into two mirrors, optionally tracking
* future changes to the main mirror to allow future recombination.
*/
CONV_SPLIT_MIRRORS = 2,
/* Every other segment type or mirror log conversion we haven't separated out */
CONV_OTHER = 3,
} conversion_type_t;
struct lvconvert_params {
2016-08-02 18:40:33 +03:00
/* Exactly one of these 12 command categories is determined */
2016-12-21 00:17:48 +03:00
int keep_mimages; /* 2 */ /* --splitmirrors */
/* other */ /* 3 */
/* FIXME Eliminate all cases where more than one of the above are set then use conv_type instead */
conversion_type_t conv_type;
int track_changes; /* CONV_SPLIT_MIRRORS is set */
2016-08-02 18:40:33 +03:00
int corelog; /* Equivalent to --mirrorlog core */
int mirrorlog; /* Only one of corelog and mirrorlog may be set */
int mirrors_supplied; /* When type_str is not set, this may be set with keep_mimages for --splitmirrors */
const char *type_str; /* When this is set, mirrors_supplied may optionally also be set */
/* Holds what you asked for based on --type or other arguments, else "" */
const struct segment_type *segtype; /* Holds what segment type you will get */
int force;
int yes;
2006-04-06 00:43:23 +04:00
int zero;
const char *lv_name;
2010-01-09 01:00:31 +03:00
const char *lv_split_name;
2007-12-22 15:13:29 +03:00
const char *lv_name_full;
2006-04-06 00:43:23 +04:00
const char *vg_name;
int wait_completion;
int need_polling;
2006-04-06 00:43:23 +04:00
uint32_t region_size;
unsigned region_size_supplied;
2006-04-06 00:43:23 +04:00
uint32_t mirrors;
2005-11-29 21:20:23 +03:00
sign_t mirrors_sign;
2010-04-13 05:54:32 +04:00
uint32_t stripes;
uint32_t stripe_size;
unsigned stripes_supplied;
unsigned stripe_size_supplied;
uint32_t read_ahead;
2006-04-06 00:43:23 +04:00
unsigned target_attr;
alloc_policy_t alloc;
int pv_count;
char **pvs;
struct dm_list *pvh;
struct logical_volume *lv_to_poll;
struct dm_list idls;
const char *origin_name;
};
struct convert_poll_id_list {
struct dm_list list;
struct poll_operation_id *id;
unsigned is_merging_origin:1;
unsigned is_merging_origin_thin:1;
};
/* FIXME Temporary function until the enum replaces the separate variables */
static void _set_conv_type(struct lvconvert_params *lp, int conv_type)
{
if (lp->conv_type != CONV_OTHER)
log_error(INTERNAL_ERROR "Changing conv_type from %d to %d.", lp->conv_type, conv_type);
lp->conv_type = conv_type;
}
static int _raid0_type_requested(const char *type_str)
{
return (!strcmp(type_str, SEG_TYPE_NAME_RAID0) || !strcmp(type_str, SEG_TYPE_NAME_RAID0_META));
}
/* mirror/raid* (1,10,4,5,6 and their variants) reshape */
static int _mirror_or_raid_type_requested(struct cmd_context *cmd, const char *type_str)
{
return (arg_is_set(cmd, mirrors_ARG) || !strcmp(type_str, SEG_TYPE_NAME_MIRROR) ||
(!strncmp(type_str, SEG_TYPE_NAME_RAID, 4) && !_raid0_type_requested(type_str)));
}
static int _linear_type_requested(const char *type_str)
{
return (!strcmp(type_str, SEG_TYPE_NAME_LINEAR));
2016-07-02 00:20:54 +03:00
}
static int _striped_type_requested(const char *type_str)
2016-07-02 00:20:54 +03:00
{
return (!strcmp(type_str, SEG_TYPE_NAME_STRIPED) || _linear_type_requested(type_str));
}
static int _read_conversion_type(struct cmd_context *cmd,
struct lvconvert_params *lp)
{
const char *type_str = arg_str_value(cmd, type_ARG, "");
lp->type_str = type_str;
if (!lp->type_str[0])
return 1;
/* FIXME: Check thin-pool and thin more thoroughly! */
if (!strcmp(type_str, SEG_TYPE_NAME_SNAPSHOT) || _striped_type_requested(type_str) ||
!strncmp(type_str, SEG_TYPE_NAME_RAID, 4) || !strcmp(type_str, SEG_TYPE_NAME_MIRROR) ||
!strcmp(type_str, SEG_TYPE_NAME_CACHE_POOL) || !strcmp(type_str, SEG_TYPE_NAME_CACHE) ||
!strcmp(type_str, SEG_TYPE_NAME_THIN_POOL) || !strcmp(type_str, SEG_TYPE_NAME_THIN))
return 1;
log_error("Conversion using --type %s is not supported.", type_str);
2016-10-11 12:55:19 +03:00
return 0;
}
static int _read_params(struct cmd_context *cmd, struct lvconvert_params *lp)
{
const char *vg_name = NULL;
2006-04-28 19:01:39 +04:00
if (!_read_conversion_type(cmd, lp))
return_0;
if (!arg_is_set(cmd, background_ARG))
lp->wait_completion = 1;
if (arg_is_set(cmd, corelog_ARG))
lp->corelog = 1;
if (arg_is_set(cmd, mirrorlog_ARG)) {
if (lp->corelog) {
log_error("--mirrorlog and --corelog are incompatible.");
return 0;
}
lp->mirrorlog = 1;
}
if (arg_is_set(cmd, trackchanges_ARG))
lp->track_changes = 1;
2010-01-09 01:00:31 +03:00
/*
* The '--splitmirrors n' argument is equivalent to '--mirrors -n'
* (note the minus sign), except that it signifies the additional
* intent to keep the mimage that is detached, rather than
* discarding it.
*/
if (arg_is_set(cmd, splitmirrors_ARG)) {
if ((lp->lv_split_name = arg_str_value(cmd, name_ARG, NULL))) {
if (!validate_restricted_lvname_param(cmd, &vg_name, &lp->lv_split_name))
return_0;
}
if (_mirror_or_raid_type_requested(cmd, lp->type_str)) {
log_error("--mirrors/--type mirror/--type raid* and --splitmirrors are "
"mutually exclusive.");
return 0;
}
if (!arg_is_set(cmd, name_ARG) && !lp->track_changes) {
2010-01-09 01:00:31 +03:00
log_error("Please name the new logical volume using '--name'");
return 0;
}
if ((lp->lv_split_name = arg_str_value(cmd, name_ARG, NULL))) {
if (!validate_restricted_lvname_param(cmd, &vg_name, &lp->lv_split_name))
return_0;
}
2010-01-09 01:00:31 +03:00
lp->keep_mimages = 1;
_set_conv_type(lp, CONV_SPLIT_MIRRORS);
2010-01-09 01:00:31 +03:00
lp->mirrors = arg_uint_value(cmd, splitmirrors_ARG, 0);
lp->mirrors_sign = SIGN_MINUS;
}
/* If no other case was identified, then use of --stripes means --type striped */
if (!arg_is_set(cmd, type_ARG) && !*lp->type_str &&
!lp->mirrorlog && !lp->corelog &&
(arg_is_set(cmd, stripes_long_ARG) || arg_is_set(cmd, stripesize_ARG)))
lp->type_str = SEG_TYPE_NAME_STRIPED;
if ((arg_is_set(cmd, stripes_long_ARG) || arg_is_set(cmd, stripesize_ARG)) &&
!(_mirror_or_raid_type_requested(cmd, lp->type_str) || _striped_type_requested(lp->type_str) ||
_raid0_type_requested(lp->type_str) || arg_is_set(cmd, thinpool_ARG))) {
log_error("--stripes or --stripesize argument is only valid "
2016-07-02 00:20:54 +03:00
"with --mirrors/--type mirror/--type raid*/--type striped/--type linear, --repair and --thinpool");
return 0;
}
if (arg_is_set(cmd, mirrors_ARG)) {
2016-10-11 12:55:19 +03:00
/* --splitmirrors is the mechanism for detaching and keeping a mimage */
lp->mirrors_supplied = 1;
2006-04-06 00:43:23 +04:00
lp->mirrors = arg_uint_value(cmd, mirrors_ARG, 0);
lp->mirrors_sign = arg_sign_value(cmd, mirrors_ARG, SIGN_NONE);
2006-04-06 00:43:23 +04:00
}
lp->alloc = (alloc_policy_t) arg_uint_value(cmd, alloc_ARG, ALLOC_INHERIT);
/*
* Final checking of each case:
* lp->keep_mimages
* --type mirror|raid lp->mirrorlog lp->corelog
* --type raid0|striped
*/
switch(lp->conv_type) {
case CONV_SPLIT_MIRRORS:
break;
2006-04-06 00:43:23 +04:00
case CONV_OTHER:
if (_mirror_or_raid_type_requested(cmd, lp->type_str) ||
lp->mirrorlog || lp->corelog) { /* Mirrors (and some RAID functions) */
if (arg_is_set(cmd, chunksize_ARG)) {
log_error("--chunksize is only available with snapshots or pools.");
2006-04-06 00:43:23 +04:00
return 0;
}
if (arg_is_set(cmd, zero_ARG)) {
log_error("--zero is only available with snapshots or thin pools.");
2006-04-28 19:01:39 +04:00
return 0;
}
if (arg_is_set(cmd, regionsize_ARG)) {
lp->region_size = arg_uint_value(cmd, regionsize_ARG, 0);
lp->region_size_supplied = 1;
} else {
lp->region_size = get_default_region_size(cmd);
lp->region_size_supplied = 0;
}
/* FIXME man page says in one place that --type and --mirrors can't be mixed */
if (lp->mirrors_supplied && !lp->mirrors)
/* down-converting to linear/stripe? */
lp->type_str = SEG_TYPE_NAME_STRIPED;
} else if (_raid0_type_requested(lp->type_str) || _striped_type_requested(lp->type_str)) { /* striped or linear or raid0 */
if (arg_from_list_is_set(cmd, "cannot be used with --type raid0 or --type striped or --type linear",
chunksize_ARG, corelog_ARG, mirrors_ARG, mirrorlog_ARG, regionsize_ARG, zero_ARG,
-1))
return_0;
} /* else segtype will default to current type */
}
lp->force = arg_count(cmd, force_ARG);
lp->yes = arg_count(cmd, yes_ARG);
return 1;
}
2007-12-22 15:13:29 +03:00
static struct poll_functions _lvconvert_mirror_fns = {
.poll_progress = poll_mirror_progress,
.finish_copy = lvconvert_mirror_finish,
2007-12-22 15:13:29 +03:00
};
static struct poll_functions _lvconvert_merge_fns = {
.poll_progress = poll_merge_progress,
.finish_copy = lvconvert_merge_finish,
};
2013-11-30 00:28:18 +04:00
static struct poll_functions _lvconvert_thin_merge_fns = {
.poll_progress = poll_thin_merge_progress,
.finish_copy = lvconvert_merge_finish,
2013-11-30 00:28:18 +04:00
};
static struct poll_operation_id *_create_id(struct cmd_context *cmd,
const char *vg_name,
const char *lv_name,
const char *uuid)
{
struct poll_operation_id *id;
char lv_full_name[NAME_LEN];
if (!vg_name || !lv_name || !uuid) {
log_error(INTERNAL_ERROR "Wrong params for lvconvert _create_id.");
return NULL;
}
if (dm_snprintf(lv_full_name, sizeof(lv_full_name), "%s/%s", vg_name, lv_name) < 0) {
log_error(INTERNAL_ERROR "Name \"%s/%s\" is too long.", vg_name, lv_name);
return NULL;
}
if (!(id = dm_pool_alloc(cmd->mem, sizeof(*id)))) {
log_error("Poll operation ID allocation failed.");
return NULL;
}
if (!(id->display_name = dm_pool_strdup(cmd->mem, lv_full_name)) ||
!(id->lv_name = strchr(id->display_name, '/')) ||
!(id->vg_name = dm_pool_strdup(cmd->mem, vg_name)) ||
!(id->uuid = dm_pool_strdup(cmd->mem, uuid))) {
log_error("Failed to copy one or more poll operation ID members.");
dm_pool_free(cmd->mem, id);
return NULL;
}
id->lv_name++; /* skip over '/' */
return id;
}
static int _lvconvert_poll_by_id(struct cmd_context *cmd, struct poll_operation_id *id,
unsigned background,
int is_merging_origin,
int is_merging_origin_thin)
{
2015-09-02 17:53:23 +03:00
if (test_mode())
return ECMD_PROCESSED;
if (is_merging_origin)
return poll_daemon(cmd, background,
(MERGING | (is_merging_origin_thin ? THIN_VOLUME : SNAPSHOT)),
is_merging_origin_thin ? &_lvconvert_thin_merge_fns : &_lvconvert_merge_fns,
"Merged", id);
return poll_daemon(cmd, background, CONVERTING,
&_lvconvert_mirror_fns, "Converted", id);
}
int lvconvert_poll(struct cmd_context *cmd, struct logical_volume *lv,
unsigned background)
2007-12-22 15:13:29 +03:00
{
int r;
struct poll_operation_id *id = _create_id(cmd, lv->vg->name, lv->name, lv->lvid.s);
int is_merging_origin = 0;
int is_merging_origin_thin = 0;
if (!id) {
log_error("Failed to allocate poll identifier for lvconvert.");
return ECMD_FAILED;
}
/* FIXME: check this in polling instead */
if (lv_is_merging_origin(lv)) {
is_merging_origin = 1;
is_merging_origin_thin = seg_is_thin_volume(find_snapshot(lv));
}
r = _lvconvert_poll_by_id(cmd, id, background, is_merging_origin, is_merging_origin_thin);
return r;
2007-12-22 15:13:29 +03:00
}
static int _insert_lvconvert_layer(struct cmd_context *cmd,
struct logical_volume *lv)
{
char format[NAME_LEN], layer_name[NAME_LEN];
int i;
/*
* We would like to give the same number for this layer
* and the newly added mimage.
* However, LV name of newly added mimage is determined *after*
* the LV name of this layer is determined.
*
* So, use generate_lv_name() to generate mimage name first
* and take the number from it.
*/
if (dm_snprintf(format, sizeof(format), "%s_mimage_%%d", lv->name) < 0) {
log_error("lvconvert: layer name creation failed.");
return 0;
}
if (!generate_lv_name(lv->vg, format, layer_name, sizeof(layer_name)) ||
sscanf(layer_name, format, &i) != 1) {
log_error("lvconvert: layer name generation failed.");
return 0;
}
if (dm_snprintf(layer_name, sizeof(layer_name), MIRROR_SYNC_LAYER "_%d", i) < 0) {
log_error("layer name creation failed.");
return 0;
}
if (!insert_layer_for_lv(cmd, lv, 0, layer_name)) {
log_error("Failed to insert resync layer");
return 0;
}
return 1;
}
2009-05-21 07:04:52 +04:00
static int _failed_mirrors_count(struct logical_volume *lv)
{
struct lv_segment *lvseg;
int ret = 0;
unsigned s;
2009-05-21 07:04:52 +04:00
dm_list_iterate_items(lvseg, &lv->segments) {
if (!seg_is_mirrored(lvseg))
return -1;
for (s = 0; s < lvseg->area_count; s++) {
if (seg_type(lvseg, s) == AREA_LV) {
if (is_temporary_mirror_layer(seg_lv(lvseg, s)))
ret += _failed_mirrors_count(seg_lv(lvseg, s));
else if (lv_is_partial(seg_lv(lvseg, s)))
++ ret;
}
2015-05-25 14:16:57 +03:00
else if (seg_type(lvseg, s) == AREA_PV &&
is_missing_pv(seg_pv(lvseg, s)))
++ret;
}
}
2009-05-21 07:04:52 +04:00
return ret;
}
static int _failed_logs_count(struct logical_volume *lv)
{
2012-02-24 02:36:56 +04:00
int ret = 0;
unsigned s;
struct logical_volume *log_lv = first_seg(lv)->log_lv;
if (log_lv && lv_is_partial(log_lv)) {
if (lv_is_mirrored(log_lv))
ret += _failed_mirrors_count(log_lv);
else
ret += 1;
}
for (s = 0; s < first_seg(lv)->area_count; s++) {
if (seg_type(first_seg(lv), s) == AREA_LV &&
is_temporary_mirror_layer(seg_lv(first_seg(lv), s)))
ret += _failed_logs_count(seg_lv(first_seg(lv), s));
}
return ret;
}
static struct dm_list *_failed_pv_list(struct volume_group *vg)
{
2009-05-21 07:04:52 +04:00
struct dm_list *failed_pvs;
struct pv_list *pvl, *new_pvl;
2009-05-21 07:04:52 +04:00
if (!(failed_pvs = dm_pool_alloc(vg->vgmem, sizeof(*failed_pvs)))) {
log_error("Allocation of list of failed_pvs failed.");
2014-05-20 14:53:51 +04:00
return NULL;
}
2009-05-21 07:04:52 +04:00
dm_list_init(failed_pvs);
dm_list_iterate_items(pvl, &vg->pvs) {
2010-03-16 17:37:38 +03:00
if (!is_missing_pv(pvl->pv))
continue;
2014-07-11 14:25:18 +04:00
/*
* Finally, --repair will remove empty PVs.
* But we only want remove these which are output of repair,
* Do not count these which are already empty here.
* FIXME: code should traverse PV in LV not in whole VG.
* FIXME: layer violation? should it depend on vgreduce --removemising?
*/
if (pvl->pv->pe_alloc_count == 0)
continue;
if (!(new_pvl = dm_pool_alloc(vg->vgmem, sizeof(*new_pvl)))) {
2009-05-21 07:04:52 +04:00
log_error("Allocation of failed_pvs list entry failed.");
2014-05-20 14:53:51 +04:00
return NULL;
}
new_pvl->pv = pvl->pv;
2009-05-21 07:04:52 +04:00
dm_list_add(failed_pvs, &new_pvl->list);
}
2009-05-21 07:04:52 +04:00
return failed_pvs;
}
static int _is_partial_lv(struct logical_volume *lv,
void *baton __attribute__((unused)))
{
return lv_is_partial(lv);
}
2009-05-21 07:04:52 +04:00
/*
* Walk down the stacked mirror LV to the original mirror LV.
*/
static struct logical_volume *_original_lv(struct logical_volume *lv)
{
struct logical_volume *next_lv = lv, *tmp_lv;
while ((tmp_lv = find_temporary_mirror(next_lv)))
next_lv = tmp_lv;
return next_lv;
}
static void _lvconvert_mirrors_repair_ask(struct cmd_context *cmd,
int failed_log, int failed_mirrors,
int *replace_log, int *replace_mirrors)
{
const char *leg_policy, *log_policy;
int force = arg_count(cmd, force_ARG);
int yes = arg_count(cmd, yes_ARG);
if (arg_is_set(cmd, usepolicies_ARG)) {
leg_policy = find_config_tree_str(cmd, activation_mirror_image_fault_policy_CFG, NULL);
log_policy = find_config_tree_str(cmd, activation_mirror_log_fault_policy_CFG, NULL);
*replace_mirrors = strcmp(leg_policy, "remove");
*replace_log = strcmp(log_policy, "remove");
return;
}
if (force != PROMPT) {
*replace_log = *replace_mirrors = 0;
return;
}
*replace_log = *replace_mirrors = 1;
if (yes)
return;
if (failed_log &&
2014-05-20 22:10:55 +04:00
yes_no_prompt("Attempt to replace failed mirror log? [y/n]: ") == 'n')
*replace_log = 0;
if (failed_mirrors &&
yes_no_prompt("Attempt to replace failed mirror images "
2014-05-20 22:10:55 +04:00
"(requires full device resync)? [y/n]: ") == 'n')
*replace_mirrors = 0;
}
/*
* _get_log_count
* @lv: the mirror LV
*
* Get the number of on-disk copies of the log.
* 0 = 'core'
* 1 = 'disk'
* 2+ = 'mirrored'
*/
2013-11-25 16:42:30 +04:00
static uint32_t _get_log_count(struct logical_volume *lv)
{
struct logical_volume *log_lv;
log_lv = first_seg(_original_lv(lv))->log_lv;
if (log_lv)
return lv_mirror_count(log_lv);
return 0;
}
static int _lv_update_mirrored_log(struct logical_volume *lv,
struct dm_list *operable_pvs,
int log_count)
{
int old_log_count;
struct logical_volume *log_lv;
Taka's fix for handling failure of all mirrored log devices and all but one mirror leg. <patch header> To handle a double failure of a mirrored log, Jon's two patches are commited, however, lvconvert command can't still handle an error when mirror leg and mirrored log got failure at the same time. [Patch]: Handle both devices of a mirrored log failing (bug 607347) posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00009.html commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00027.html [Patch]: Handle both devices of a mirrored log failing (bug 607347) - additional fix posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00093.html commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00101.html In the second patch, the target type of mirrored log is replaced with error target when remove_log is set to 1, but this procedure should be also used in other cases such as the number of mirror leg is 1. This patch relocates the procedure to the main path. In addition, I added following three changes. - Removed tmp_orphan_lvs handling procedure It seems that _delete_lv() can handle detached_log_lv properly without adding mirror legs in mirrored log to tmp_orphan_lvs. Therefore, I removed the procedure. - Removed vg_write()/vg_commit() Metadata is saved by vg_write()/vg_commit() just after detached_log_lv is handled. Therefore, I removed vg_write()/vg_commit(). - With Jon's second patch, we think that we don't have to call remove_mirror_log() in _lv_update_mirrored_log() because will be handled remove_mirror_images() in _lvconvert_mirrors_repaire(). </patch header> Signed-off-by: Takahiro Yasui <takahiro.yasui@hds.com> Reviewed-by: Petr Rockai <prockai@redhat.com> Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
2010-08-03 01:07:40 +04:00
/*
* When log_count is 0, mirrored log doesn't need to be
* updated here but it will be removed later.
*/
if (!log_count)
return 1;
log_lv = first_seg(_original_lv(lv))->log_lv;
if (!log_lv || !lv_is_mirrored(log_lv))
return 1;
old_log_count = _get_log_count(lv);
if (old_log_count == log_count)
return 1;
/* Reducing redundancy of the log */
Taka's fix for handling failure of all mirrored log devices and all but one mirror leg. <patch header> To handle a double failure of a mirrored log, Jon's two patches are commited, however, lvconvert command can't still handle an error when mirror leg and mirrored log got failure at the same time. [Patch]: Handle both devices of a mirrored log failing (bug 607347) posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00009.html commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00027.html [Patch]: Handle both devices of a mirrored log failing (bug 607347) - additional fix posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00093.html commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00101.html In the second patch, the target type of mirrored log is replaced with error target when remove_log is set to 1, but this procedure should be also used in other cases such as the number of mirror leg is 1. This patch relocates the procedure to the main path. In addition, I added following three changes. - Removed tmp_orphan_lvs handling procedure It seems that _delete_lv() can handle detached_log_lv properly without adding mirror legs in mirrored log to tmp_orphan_lvs. Therefore, I removed the procedure. - Removed vg_write()/vg_commit() Metadata is saved by vg_write()/vg_commit() just after detached_log_lv is handled. Therefore, I removed vg_write()/vg_commit(). - With Jon's second patch, we think that we don't have to call remove_mirror_log() in _lv_update_mirrored_log() because will be handled remove_mirror_images() in _lvconvert_mirrors_repaire(). </patch header> Signed-off-by: Takahiro Yasui <takahiro.yasui@hds.com> Reviewed-by: Petr Rockai <prockai@redhat.com> Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
2010-08-03 01:07:40 +04:00
return remove_mirror_images(log_lv, log_count,
is_mirror_image_removable,
operable_pvs, 0U);
}
static int _lv_update_log_type(struct cmd_context *cmd,
struct lvconvert_params *lp,
struct logical_volume *lv,
struct dm_list *operable_pvs,
int log_count)
{
int old_log_count;
uint32_t region_size = (lp) ? lp->region_size :
first_seg(lv)->region_size;
alloc_policy_t alloc = (lp) ? lp->alloc : lv->alloc;
struct logical_volume *original_lv;
struct logical_volume *log_lv;
old_log_count = _get_log_count(lv);
if (old_log_count == log_count)
return 1;
original_lv = _original_lv(lv);
/* Remove an existing log completely */
if (!log_count) {
if (!remove_mirror_log(cmd, original_lv, operable_pvs,
arg_count(cmd, yes_ARG) ||
arg_count(cmd, force_ARG)))
return_0;
return 1;
}
log_lv = first_seg(original_lv)->log_lv;
/* Adding redundancy to the log */
if (old_log_count < log_count) {
if (!(region_size = adjusted_mirror_region_size(cmd, lv->vg->extent_size,
lv->le_count,
region_size, 0,
vg_is_clustered(lv->vg))))
return_0;
if (!add_mirror_log(cmd, original_lv, log_count,
region_size, operable_pvs, alloc))
return_0;
/*
* FIXME: This simple approach won't work in cluster mirrors,
* but it doesn't matter because we don't support
* mirrored logs in cluster mirrors.
*/
if (old_log_count &&
!lv_update_and_reload(log_lv))
return_0;
return 1;
}
/* Reducing redundancy of the log */
return remove_mirror_images(log_lv, log_count,
is_mirror_image_removable, operable_pvs, 1U);
}
/*
* Reomove missing and empty PVs from VG, if are also in provided list
*/
static void _remove_missing_empty_pv(struct volume_group *vg, struct dm_list *remove_pvs)
{
struct pv_list *pvl, *pvl_vg, *pvlt;
int removed = 0;
if (!remove_pvs)
return;
dm_list_iterate_items(pvl, remove_pvs) {
dm_list_iterate_items_safe(pvl_vg, pvlt, &vg->pvs) {
if (!id_equal(&pvl->pv->id, &pvl_vg->pv->id) ||
2010-03-16 17:37:38 +03:00
!is_missing_pv(pvl_vg->pv) ||
pvl_vg->pv->pe_alloc_count != 0)
continue;
/* FIXME: duplication of vgreduce code, move this to library */
vg->free_count -= pvl_vg->pv->pe_count;
vg->extent_count -= pvl_vg->pv->pe_count;
del_pvl_from_vgs(vg, pvl_vg);
free_pv_fid(pvl_vg->pv);
removed++;
}
}
if (removed) {
if (!vg_write(vg) || !vg_commit(vg)) {
stack;
return;
}
log_warn("WARNING: %d missing and now unallocated Physical Volumes removed from VG.", removed);
}
}
/*
* _lvconvert_mirrors_parse_params
*
* This function performs the following:
* 1) Gets the old values of mimage and log counts
* 2) Parses the CLI args to find the new desired values
* 3) Adjusts 'lp->mirrors' to the appropriate absolute value.
* (Remember, 'lp->mirrors' is specified in terms of the number of "copies"
* vs. the number of mimages. It can also be a relative value.)
* 4) Sets 'lp->need_polling' if collapsing
* 5) Validates other mirror params
*
* Returns: 1 on success, 0 on error
*/
static int _lvconvert_mirrors_parse_params(struct cmd_context *cmd,
struct logical_volume *lv,
struct lvconvert_params *lp,
uint32_t *old_mimage_count,
uint32_t *old_log_count,
uint32_t *new_mimage_count,
uint32_t *new_log_count)
{
*old_mimage_count = lv_mirror_count(lv);
*old_log_count = _get_log_count(lv);
2007-12-21 04:08:18 +03:00
if (is_lockd_type(lv->vg->lock_type) && lp->keep_mimages) {
/* FIXME: we need to create a lock for the new LV. */
log_error("Unable to split mirrors in VG with lock_type %s", lv->vg->lock_type);
return 0;
}
/*
* Adjusting mimage count?
*/
if (!lp->mirrors_supplied && !lp->keep_mimages)
lp->mirrors = *old_mimage_count;
else if (lp->mirrors_sign == SIGN_PLUS)
lp->mirrors = *old_mimage_count + lp->mirrors;
else if (lp->mirrors_sign == SIGN_MINUS)
lp->mirrors = (*old_mimage_count > lp->mirrors) ?
*old_mimage_count - lp->mirrors: 0;
else
2005-11-29 21:20:23 +03:00
lp->mirrors += 1;
*new_mimage_count = lp->mirrors;
/* Too many mimages? */
if (lp->mirrors > DEFAULT_MIRROR_MAX_IMAGES) {
log_error("Only up to %d images in mirror supported currently.",
DEFAULT_MIRROR_MAX_IMAGES);
return 0;
}
/* Did the user try to subtract more legs than available? */
if (lp->mirrors < 1) {
log_error("Unable to reduce images by specified amount - only %d in %s",
*old_mimage_count, lv->name);
return 0;
}
/*
* FIXME: It would be nice to say what we are adjusting to, but
* I really don't know whether to specify the # of copies or mimages.
*/
if (*old_mimage_count != *new_mimage_count)
log_verbose("Adjusting mirror image count of %s", lv->name);
2009-04-30 00:11:46 +04:00
/* If region size is not given by user - use value from mirror */
if (lv_is_mirrored(lv) && !lp->region_size_supplied) {
lp->region_size = first_seg(lv)->region_size;
log_debug("Copying region size %s from existing mirror.",
display_size(lv->vg->cmd, lp->region_size));
}
/*
* Adjust log type
*
* If we are converting from a mirror to another mirror or simply
* changing the log type, we start by assuming they want the log
* type the same and then parse the given args. OTOH, If we are
* converting from linear to mirror, then we start from the default
* position that the user would like a 'disk' log.
*/
*new_log_count = (*old_mimage_count > 1) ? *old_log_count : 1;
if (!lp->corelog && !lp->mirrorlog)
return 1;
*new_log_count = arg_int_value(cmd, mirrorlog_ARG, lp->corelog ? MIRROR_LOG_CORE : DEFAULT_MIRRORLOG);
/*
* No mirrored logs for cluster mirrors until
* log daemon is multi-threaded.
*/
if ((*new_log_count == MIRROR_LOG_MIRRORED) && vg_is_clustered(lv->vg)) {
log_error("Log type, \"mirrored\", is unavailable to cluster mirrors.");
return 0;
}
log_verbose("Setting logging type to %s.", get_mirror_log_name(*new_log_count));
/*
* Region size must not change on existing mirrors
*/
if (arg_is_set(cmd, regionsize_ARG) && lv_is_mirrored(lv) &&
(lp->region_size != first_seg(lv)->region_size)) {
2005-11-29 21:20:23 +03:00
log_error("Mirror log region size cannot be changed on "
"an existing mirror.");
return 0;
}
/*
* For the most part, we cannot handle multi-segment mirrors. Bail out
* early if we have encountered one.
*/
if (lv_is_mirrored(lv) && dm_list_size(&lv->segments) != 1) {
log_error("Logical volume %s has multiple "
"mirror segments.", display_lvname(lv));
return 0;
}
2009-04-30 00:11:46 +04:00
return 1;
}
/*
* _lvconvert_mirrors_aux
*
* Add/remove mirror images and adjust log type. 'operable_pvs'
* are the set of PVs open to removal or allocation - depending
* on the operation being performed.
*/
static int _lvconvert_mirrors_aux(struct cmd_context *cmd,
struct logical_volume *lv,
struct lvconvert_params *lp,
struct dm_list *operable_pvs,
uint32_t new_mimage_count,
uint32_t new_log_count,
struct dm_list *pvh)
{
uint32_t region_size;
struct lv_segment *seg = first_seg(lv);
struct logical_volume *layer_lv;
uint32_t old_mimage_count = lv_mirror_count(lv);
uint32_t old_log_count = _get_log_count(lv);
if ((lp->mirrors == 1) && !lv_is_mirrored(lv)) {
log_warn("WARNING: Logical volume %s is already not mirrored.",
display_lvname(lv));
return 1;
}
if (!(region_size = adjusted_mirror_region_size(cmd, lv->vg->extent_size,
lv->le_count,
lp->region_size ? : seg->region_size, 0,
vg_is_clustered(lv->vg))))
return_0;
if (lv_component_is_active(lv)) {
log_error("Cannot convert logical volume %s with active component LV(s).",
display_lvname(lv));
return 0;
}
if (!operable_pvs)
operable_pvs = pvh;
2010-01-09 01:00:31 +03:00
/*
* Up-convert from linear to mirror
*/
if (!lv_is_mirrored(lv)) {
/* FIXME Share code with lvcreate */
/*
* FIXME should we give not only pvh, but also all PVs
* currently taken by the mirror? Would make more sense from
* user perspective.
*/
2010-04-13 05:54:32 +04:00
if (!lv_add_mirrors(cmd, lv, new_mimage_count - 1, lp->stripes,
lp->stripe_size, region_size, new_log_count, operable_pvs,
lp->alloc, MIRROR_BY_LV))
return_0;
if (!arg_is_set(cmd, background_ARG))
lp->need_polling = 1;
goto out;
}
/*
* Up-convert m-way mirror to n-way mirror
*/
if (new_mimage_count > old_mimage_count) {
2016-07-14 16:21:01 +03:00
if (lv_is_not_synced(lv)) {
log_error("Can't add mirror to out-of-sync mirrored "
"LV: use lvchange --resync first.");
return 0;
}
2009-10-26 13:01:56 +03:00
/*
* We allow snapshots of mirrors, but for now, we
* do not allow up converting mirrors that are under
* snapshots. The layering logic is somewhat complex,
* and preliminary test show that the conversion can't
* seem to get the correct %'age of completion.
*/
if (lv_is_origin(lv)) {
log_error("Can't add additional mirror images to "
"mirror %s which is under snapshots.",
display_lvname(lv));
return 0;
2009-10-26 13:01:56 +03:00
}
/*
* Is there already a convert in progress? We do not
* currently allow more than one.
*/
if (find_temporary_mirror(lv) || lv_is_converting(lv)) {
log_error("%s is already being converted. Unable to start another conversion.",
display_lvname(lv));
return 0;
}
/*
* Log addition/removal should be done before the layer
* insertion to make the end result consistent with
* linear-to-mirror conversion.
*/
if (!_lv_update_log_type(cmd, lp, lv,
operable_pvs, new_log_count))
return_0;
/* Insert a temporary layer for syncing,
* only if the original lv is using disk log. */
if (seg->log_lv && !_insert_lvconvert_layer(cmd, lv)) {
log_error("Failed to insert resync layer.");
2007-12-20 21:55:46 +03:00
return 0;
}
/* FIXME: can't have multiple mlogs. force corelog. */
if (!lv_add_mirrors(cmd, lv,
new_mimage_count - old_mimage_count,
lp->stripes, lp->stripe_size,
region_size, 0U, operable_pvs, lp->alloc,
MIRROR_BY_LV)) {
/* FIXME: failure inside library -> move error path processing into library. */
layer_lv = seg_lv(first_seg(lv), 0);
if (!remove_layer_from_lv(lv, layer_lv) ||
(lv_is_active(lv) && !deactivate_lv(cmd, layer_lv)) ||
!lv_remove(layer_lv) ||
!vg_write(lv->vg) || !vg_commit(lv->vg)) {
log_error("ABORTING: Failed to remove "
"temporary mirror layer %s.",
display_lvname(layer_lv));
log_error("Manual cleanup with vgcfgrestore "
"and dmsetup may be required.");
return 0;
}
return_0;
}
if (seg->log_lv)
lv->status |= CONVERTING;
lp->need_polling = 1;
goto out_skip_log_convert;
}
/*
* Down-convert (reduce # of mimages).
*/
if (new_mimage_count < old_mimage_count) {
uint32_t nmc = old_mimage_count - new_mimage_count;
uint32_t nlc = (!new_log_count || lp->mirrors == 1) ? 1U : 0U;
2010-04-20 16:18:31 +04:00
/* FIXME: Why did nlc used to be calculated that way? */
/* Reduce number of mirrors */
if (lp->keep_mimages) {
if (lp->track_changes) {
Add the ability to split an image from the mirror and track changes. ~> lvconvert --splitmirrors 1 --trackchanges vg/lv The '--trackchanges' option allows a user the ability to use an image of a RAID1 array for the purposes of temporary read-only access. The image can be merged back into the array at a later time and only the blocks that have changed in the array since the split will be resync'ed. This operation can be thought of as a partial split. The image is never completely extracted from the array, in that the array reserves the position the device occupied and tracks the differences between the array and the split image via a bitmap. The image itself is rendered read-only and the name (<LV>_rimage_*) cannot be changed. The user can complete the split (permanently splitting the image from the array) by re-issuing the 'lvconvert' command without the '--trackchanges' argument and specifying the '--name' argument. ~> lvconvert --splitmirrors 1 --name my_split vg/lv Merging the tracked image back into the array is done with the '--merge' option (included in a follow-on patch). ~> lvconvert --merge vg/lv_rimage_<n> The internal mechanics of this are relatively simple. The 'raid' device- mapper target allows for the specification of an empty slot in an array via '- -'. This is what will be used if a partial activation of an array is ever required. (It would also be possible to use 'error' targets in place of the '- -'.) If a RAID image is found to be both read-only and visible, then it is considered separate from the array and '- -' is used to hold it's position in the array. So, all that needs to be done to temporarily split an image from the array /and/ cause the kernel target's bitmap to track (aka "mark") changes made is to make the specified image visible and read-only. To merge the device back into the array, the image needs to be returned to the read/write state of the top-level LV and made invisible.
2011-08-18 23:38:26 +04:00
log_error("--trackchanges is not available "
"to 'mirror' segment type.");
Add the ability to split an image from the mirror and track changes. ~> lvconvert --splitmirrors 1 --trackchanges vg/lv The '--trackchanges' option allows a user the ability to use an image of a RAID1 array for the purposes of temporary read-only access. The image can be merged back into the array at a later time and only the blocks that have changed in the array since the split will be resync'ed. This operation can be thought of as a partial split. The image is never completely extracted from the array, in that the array reserves the position the device occupied and tracks the differences between the array and the split image via a bitmap. The image itself is rendered read-only and the name (<LV>_rimage_*) cannot be changed. The user can complete the split (permanently splitting the image from the array) by re-issuing the 'lvconvert' command without the '--trackchanges' argument and specifying the '--name' argument. ~> lvconvert --splitmirrors 1 --name my_split vg/lv Merging the tracked image back into the array is done with the '--merge' option (included in a follow-on patch). ~> lvconvert --merge vg/lv_rimage_<n> The internal mechanics of this are relatively simple. The 'raid' device- mapper target allows for the specification of an empty slot in an array via '- -'. This is what will be used if a partial activation of an array is ever required. (It would also be possible to use 'error' targets in place of the '- -'.) If a RAID image is found to be both read-only and visible, then it is considered separate from the array and '- -' is used to hold it's position in the array. So, all that needs to be done to temporarily split an image from the array /and/ cause the kernel target's bitmap to track (aka "mark") changes made is to make the specified image visible and read-only. To merge the device back into the array, the image needs to be returned to the read/write state of the top-level LV and made invisible.
2011-08-18 23:38:26 +04:00
return 0;
}
if (!lv_split_mirror_images(lv, lp->lv_split_name,
nmc, operable_pvs))
2014-05-20 14:53:51 +04:00
return_0;
} else if (!lv_remove_mirrors(cmd, lv, nmc, nlc,
is_mirror_image_removable, operable_pvs, 0))
return_0;
goto out; /* Just in case someone puts code between */
}
out:
/*
* Converting the log type
*/
if (lv_is_mirrored(lv) && (old_log_count != new_log_count)) {
if (!_lv_update_log_type(cmd, lp, lv,
operable_pvs, new_log_count))
return_0;
}
out_skip_log_convert:
if (!lv_update_and_reload(lv))
return_0;
return 1;
}
int mirror_remove_missing(struct cmd_context *cmd,
struct logical_volume *lv, int force)
{
struct dm_list *failed_pvs;
int log_count = _get_log_count(lv) - _failed_logs_count(lv);
if (!(failed_pvs = _failed_pv_list(lv->vg)))
return_0;
2013-11-25 16:42:30 +04:00
if (force && _failed_mirrors_count(lv) == (int)lv_mirror_count(lv)) {
2016-02-25 17:01:12 +03:00
log_error("No usable images left in %s.", display_lvname(lv));
return lv_remove_with_dependencies(cmd, lv, DONT_PROMPT, 0);
}
/*
* We must adjust the log first, or the entire mirror
* will get stuck during a suspend.
*/
if (!_lv_update_mirrored_log(lv, failed_pvs, log_count))
2014-05-20 14:53:51 +04:00
return_0;
if (_failed_mirrors_count(lv) > 0 &&
!lv_remove_mirrors(cmd, lv, _failed_mirrors_count(lv),
log_count ? 0U : 1U,
_is_partial_lv, NULL, 0))
2014-05-20 14:53:51 +04:00
return_0;
mirror: Mirrored log should be fixed before mirror when double fault occurs This patch is intended to fix bug 825323 - FS turns read-only during a double fault of a mirror leg and mirrored log's leg at the same time. It only affects a 2-way mirror with a mirrored log. 3+-way mirrors and mirrors without a mirrored log are not affected. The problem resulted from the fact that the top level mirror was not using 'noflush' when suspending before its "down-convert". When a mirror image fails, the bios are queue until a suspend is recieved. If it is a 'noflush' suspend, the bios can be safely requeued in the DM core. If 'noflush' is not used, the bios must be pushed through the target and if a device is failed for a mirror, that means issuing an error. When an error is received by a file system, it results in it turning read-only (depending on the FS). Part of the problem was is due to the nature of the stacking involved in using a mirror as a mirror's log. When an image in each fail, the top level mirror stalls because it is waiting for a log flush. The other stalls waiting for corrective action. When the repair command is issued, the entire stacked arrangement is collapsed to a linear LV. The log flush then fails (somewhat uncleanly) and the top-level mirror is suspended without 'noflush' because it is a linear device. This patch allows the log to be repaired first, which in turn allows the top-level mirror's log flush to complete cleanly. The top-level mirror is then secondarily reduced to a linear device - at which time this mirror is suspended properly with 'noflush'.
2012-11-15 00:58:47 +04:00
if (lv_is_mirrored(lv) &&
!_lv_update_log_type(cmd, NULL, lv, failed_pvs, log_count))
2014-05-20 14:53:51 +04:00
return_0;
if (!lv_update_and_reload(lv))
return_0;
return 1;
}
/*
* _lvconvert_mirrors_repair
*
* This function operates in two phases. First, all of the bad
* devices are removed from the mirror. Then, if desired by the
* user, the devices are replaced.
*
* 'old_mimage_count' and 'old_log_count' are there so we know
* what to convert to after the removal of devices.
*/
static int _lvconvert_mirrors_repair(struct cmd_context *cmd,
struct logical_volume *lv,
struct lvconvert_params *lp,
struct dm_list *pvh)
{
int failed_logs;
int failed_mimages;
int replace_logs = 0;
int replace_mimages = 0;
uint32_t log_count;
uint32_t original_mimages = lv_mirror_count(lv);
uint32_t original_logs = _get_log_count(lv);
cmd->partial_activation = 1;
lp->need_polling = 0;
lv_check_transient(lv); /* TODO check this in lib for all commands? */
if (!lv_is_partial(lv)) {
2016-02-25 17:01:12 +03:00
log_print_unless_silent("Volume %s is consistent. Nothing to repair.",
display_lvname(lv));
return 1;
}
failed_mimages = _failed_mirrors_count(lv);
failed_logs = _failed_logs_count(lv);
/* Retain existing region size in case we need it later */
if (!lp->region_size)
lp->region_size = first_seg(lv)->region_size;
if (!mirror_remove_missing(cmd, lv, 0))
return_0;
if (failed_mimages)
log_print_unless_silent("Mirror status: %d of %d images failed.",
failed_mimages, original_mimages);
/*
* Count the failed log devices
*/
if (failed_logs)
log_print_unless_silent("Mirror log status: %d of %d images failed.",
failed_logs, original_logs);
/*
* Find out our policies
*/
_lvconvert_mirrors_repair_ask(cmd, failed_logs, failed_mimages,
&replace_logs, &replace_mimages);
/*
* Second phase - replace faulty devices
*/
lp->mirrors = replace_mimages ? original_mimages : (original_mimages - failed_mimages);
/*
* It does not make sense to replace the log if the volume is no longer
* a mirror.
*/
if (lp->mirrors == 1)
replace_logs = 0;
log_count = replace_logs ? original_logs : (original_logs - failed_logs);
while (replace_mimages || replace_logs) {
log_warn("WARNING: Trying to up-convert to %d images, %d logs.", lp->mirrors, log_count);
if (_lvconvert_mirrors_aux(cmd, lv, lp, NULL,
lp->mirrors, log_count, pvh))
break;
2014-05-20 22:10:55 +04:00
if (lp->mirrors > 2)
--lp->mirrors;
else if (log_count > 0)
--log_count;
else
break; /* nowhere to go, anymore... */
}
if (replace_mimages && lv_mirror_count(lv) != original_mimages)
2016-02-25 17:01:12 +03:00
log_warn("WARNING: Failed to replace %d of %d images in volume %s.",
original_mimages - lv_mirror_count(lv), original_mimages,
display_lvname(lv));
if (replace_logs && _get_log_count(lv) != original_logs)
2016-02-25 17:01:12 +03:00
log_warn("WARNING: Failed to replace %d of %d logs in volume %s.",
original_logs - _get_log_count(lv), original_logs,
display_lvname(lv));
/* if (!arg_is_set(cmd, use_policies_ARG) && (lp->mirrors != old_mimage_count
|| log_count != old_log_count))
return 0; */
return 1;
}
static int _lvconvert_validate_thin(struct logical_volume *lv,
struct lvconvert_params *lp)
{
if (!lv_is_thin_pool(lv) && !lv_is_thin_volume(lv))
return 1;
2016-02-25 17:01:12 +03:00
log_error("Converting thin%s segment type for %s to %s is not supported.",
lv_is_thin_pool(lv) ? " pool" : "",
2016-02-25 17:01:12 +03:00
display_lvname(lv), lp->segtype->name);
if (lv_is_thin_volume(lv))
return 0;
/* Give advice for thin pool conversion */
2016-02-25 17:01:12 +03:00
log_error("For pool data volume conversion use %s.",
display_lvname(seg_lv(first_seg(lv), 0)));
log_error("For pool metadata volume conversion use %s.",
display_lvname(first_seg(lv)->metadata_lv));
return 0;
}
/*
* _lvconvert_mirrors
*
* Determine what is being done. Are we doing a conversion, repair, or
* collapsing a stack? Once determined, call helper functions.
*/
static int _lvconvert_mirrors(struct cmd_context *cmd,
struct logical_volume *lv,
struct lvconvert_params *lp)
{
uint32_t old_mimage_count = 0;
uint32_t old_log_count = 0;
uint32_t new_mimage_count = 0;
uint32_t new_log_count = 0;
if ((lp->corelog || lp->mirrorlog) && *lp->type_str && strcmp(lp->type_str, SEG_TYPE_NAME_MIRROR)) {
log_error("--corelog and --mirrorlog are only compatible with mirror devices.");
return 0;
}
if (!_lvconvert_validate_thin(lv, lp))
return_0;
Mirror/Thin: Disallow thinpools on mirror logical volumes The same corner cases that exist for snapshots on mirrors exist for any logical volume layered on top of mirror. (One example is when a mirror image fails and a non-repair LVM command is the first to detect it via label reading. In this case, the LVM command will hang and prevent the necessary LVM repair command from running.) When a better alternative exists, it makes no sense to allow a new target to stack on mirrors as a new feature. Since, RAID is now capable of running EX in a cluster and thin is not active-active aware, it makes sense to pair these two rather than mirror+thinpool. As further background, here are some additional comments that I made when addressing a bug related to mirror+thinpool: (https://bugzilla.redhat.com/show_bug.cgi?id=919604#c9) I am going to disallow thin* on top of mirror logical volumes. Users will have to use the "raid1" segment type if they want this. This bug has come down to a choice between: 1) Disallowing thin-LVs from being used as PVs. 2) Disallowing thinpools on top of mirrors. The problem is that the code in dev_manager.c:device_is_usable() is unable to tell whether there is a mirror device lower in the stack from the device being checked. Pretty much anything layered on top of a mirror will suffer from this problem. (Snapshots are a good example of this; and option #1 above has been chosen to deal with them. This can also be seen in dev_manager.c:device_is_usable().) When a mirror failure occurs, the kernel blocks all I/O to it. If there is an LVM command that comes along to do the repair (or a different operation that requires label reading), it would normally avoid the mirror when it sees that it is blocked. However, if there is a snapshot or a thin-LV that is on a mirror, the above code will not detect the mirror underneath and will issue label reading I/O. This causes the command to hang. Choosing #1 would mean that thin-LVs could never be used as PVs - even if they are stacked on something other than mirrors. Choosing #2 means that thinpools can never be placed on mirrors. This is probably better than we think, since it is preferred that people use the "raid1" segment type in the first place. However, RAID* cannot currently be used in a cluster volume group - even in EX-only mode. Thus, a complete solution for option #2 must include the ability to activate RAID logical volumes (and perform RAID operations) in a cluster volume group. I've already begun working on this.
2013-09-12 00:58:44 +04:00
if (lv_is_thin_type(lv)) {
log_error("Mirror segment type cannot be used for thinpool%s.\n"
"Try \"%s\" segment type instead.",
lv_is_thin_pool_data(lv) ? "s" : " metadata",
SEG_TYPE_NAME_RAID1);
Mirror/Thin: Disallow thinpools on mirror logical volumes The same corner cases that exist for snapshots on mirrors exist for any logical volume layered on top of mirror. (One example is when a mirror image fails and a non-repair LVM command is the first to detect it via label reading. In this case, the LVM command will hang and prevent the necessary LVM repair command from running.) When a better alternative exists, it makes no sense to allow a new target to stack on mirrors as a new feature. Since, RAID is now capable of running EX in a cluster and thin is not active-active aware, it makes sense to pair these two rather than mirror+thinpool. As further background, here are some additional comments that I made when addressing a bug related to mirror+thinpool: (https://bugzilla.redhat.com/show_bug.cgi?id=919604#c9) I am going to disallow thin* on top of mirror logical volumes. Users will have to use the "raid1" segment type if they want this. This bug has come down to a choice between: 1) Disallowing thin-LVs from being used as PVs. 2) Disallowing thinpools on top of mirrors. The problem is that the code in dev_manager.c:device_is_usable() is unable to tell whether there is a mirror device lower in the stack from the device being checked. Pretty much anything layered on top of a mirror will suffer from this problem. (Snapshots are a good example of this; and option #1 above has been chosen to deal with them. This can also be seen in dev_manager.c:device_is_usable().) When a mirror failure occurs, the kernel blocks all I/O to it. If there is an LVM command that comes along to do the repair (or a different operation that requires label reading), it would normally avoid the mirror when it sees that it is blocked. However, if there is a snapshot or a thin-LV that is on a mirror, the above code will not detect the mirror underneath and will issue label reading I/O. This causes the command to hang. Choosing #1 would mean that thin-LVs could never be used as PVs - even if they are stacked on something other than mirrors. Choosing #2 means that thinpools can never be placed on mirrors. This is probably better than we think, since it is preferred that people use the "raid1" segment type in the first place. However, RAID* cannot currently be used in a cluster volume group - even in EX-only mode. Thus, a complete solution for option #2 must include the ability to activate RAID logical volumes (and perform RAID operations) in a cluster volume group. I've already begun working on this.
2013-09-12 00:58:44 +04:00
return 0;
}
if (lv_is_cache_type(lv)) {
log_error("Mirrors are not yet supported on cache LVs %s.",
display_lvname(lv));
return 0;
}
if (_linear_type_requested(lp->type_str)) {
if (arg_is_set(cmd, mirrors_ARG) && (arg_uint_value(cmd, mirrors_ARG, 0) != 0)) {
log_error("Cannot specify mirrors with linear type.");
return 0;
}
lp->mirrors_supplied = 1;
lp->mirrors = 0;
}
/* Adjust mimage and/or log count */
if (!_lvconvert_mirrors_parse_params(cmd, lv, lp,
&old_mimage_count, &old_log_count,
&new_mimage_count, &new_log_count))
2016-11-23 12:30:52 +03:00
return_0;
if (((old_mimage_count < new_mimage_count && old_log_count > new_log_count) ||
(old_mimage_count > new_mimage_count && old_log_count < new_log_count)) &&
lp->pv_count) {
log_error("Cannot both allocate and free extents when "
"specifying physical volumes to use.");
log_error("Please specify the operation in two steps.");
return 0;
}
/* Nothing to do? (Probably finishing collapse.) */
if ((old_mimage_count == new_mimage_count) &&
(old_log_count == new_log_count))
return 1;
if ((old_log_count != new_log_count) &&
(new_log_count == MIRROR_LOG_MIRRORED)) {
log_warn("WARNING: Log type \"mirrored\" is DEPRECATED and will be removed in the future. Use RAID1 LV or disk log instead.");
}
if (!_lvconvert_mirrors_aux(cmd, lv, lp, NULL,
new_mimage_count, new_log_count, lp->pvh))
2016-11-23 12:30:52 +03:00
return_0;
backup(lv->vg);
if (!lp->need_polling)
log_print_unless_silent("Logical volume %s converted.",
display_lvname(lv));
else
log_print_unless_silent("Logical volume %s being converted.",
display_lvname(lv));
return 1;
}
2013-12-04 06:09:37 +04:00
static int _is_valid_raid_conversion(const struct segment_type *from_segtype,
const struct segment_type *to_segtype)
{
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!from_segtype)
return 1;
if (from_segtype == to_segtype)
return 1;
/* Support raid0 <-> striped conversions */
if (segtype_is_striped(from_segtype) && segtype_is_striped(to_segtype))
return 1;
if (!segtype_is_raid(from_segtype) && !segtype_is_raid(to_segtype))
return_0; /* Not converting to or from RAID? */
return 1;
}
/* Check for dm-raid target supporting raid4 conversion properly. */
static int _raid4_conversion_supported(struct logical_volume *lv, struct lvconvert_params *lp)
{
int ret = 1;
struct lv_segment *seg = first_seg(lv);
if (seg_is_raid4(seg))
ret = raid4_is_supported(lv->vg->cmd, seg->segtype);
else if (segtype_is_raid4(lp->segtype))
ret = raid4_is_supported(lv->vg->cmd, lp->segtype);
if (ret)
return 1;
log_error("Cannot convert %s LV %s to %s.",
lvseg_name(seg), display_lvname(lv), lp->segtype->name);
return 0;
}
2013-12-04 06:09:37 +04:00
static int _lvconvert_raid(struct logical_volume *lv, struct lvconvert_params *lp)
{
int image_count = 0;
struct cmd_context *cmd = lv->vg->cmd;
struct lv_segment *seg = first_seg(lv);
if (_linear_type_requested(lp->type_str)) {
if (arg_is_set(cmd, mirrors_ARG) && (arg_uint_value(cmd, mirrors_ARG, 0) != 0)) {
log_error("Cannot specify mirrors with linear type.");
return 0;
}
lp->mirrors_supplied = 1;
lp->mirrors = 0;
}
/* Can only change image count for raid1 and linear */
if (lp->mirrors_supplied) {
if (_raid0_type_requested(lp->type_str)) {
log_error("--mirrors/-m is not compatible with conversion to %s.",
lp->type_str);
return 0;
}
if (!seg_is_mirrored(seg) && !seg_is_linear(seg)) {
log_error("--mirrors/-m is not compatible with %s.",
lvseg_name(seg));
return 0;
}
if (seg_is_raid10(seg)) {
log_error("--mirrors/-m cannot be changed with %s.",
lvseg_name(seg));
return 0;
}
}
if (!_lvconvert_validate_thin(lv, lp))
return_0;
if (!_is_valid_raid_conversion(seg->segtype, lp->segtype))
goto try_new_takeover_or_reshape;
2016-12-21 00:17:48 +03:00
if (seg_is_linear(seg) && !lp->mirrors_supplied) {
if (_raid0_type_requested(lp->type_str)) {
log_error("Linear LV %s cannot be converted to %s.",
display_lvname(lv), lp->type_str);
return 0;
}
if (!strcmp(lp->type_str, SEG_TYPE_NAME_RAID1)) {
log_error("Raid conversions of LV %s require -m/--mirrors.",
display_lvname(lv));
return 0;
}
goto try_new_takeover_or_reshape;
}
/* Change number of RAID1 images */
if (lp->mirrors_supplied || lp->keep_mimages) {
image_count = lv_raid_image_count(lv);
if (lp->mirrors_sign == SIGN_PLUS)
image_count += lp->mirrors;
else if (lp->mirrors_sign == SIGN_MINUS)
image_count -= lp->mirrors;
else
image_count = lp->mirrors + 1;
if (image_count < 1) {
log_error("Unable to %s images by specified amount.",
lp->keep_mimages ? "split" : "reduce");
return 0;
}
/* --trackchanges requires --splitmirrors which always has SIGN_MINUS */
if (lp->track_changes && lp->mirrors != 1) {
log_error("Exactly one image must be split off from %s when tracking changes.",
display_lvname(lv));
return 0;
}
}
if ((lp->corelog || lp->mirrorlog) && strcmp(lp->type_str, SEG_TYPE_NAME_MIRROR)) {
log_error("--corelog and --mirrorlog are only compatible with mirror devices");
return 0;
}
if (lp->track_changes)
return lv_raid_split_and_track(lv, lp->yes, lp->pvh);
if (lp->keep_mimages)
return lv_raid_split(lv, lp->yes, lp->lv_split_name, image_count, lp->pvh);
if (lp->mirrors_supplied) {
if (!*lp->type_str || !strcmp(lp->type_str, SEG_TYPE_NAME_RAID1) || !strcmp(lp->type_str, SEG_TYPE_NAME_LINEAR) ||
(!strcmp(lp->type_str, SEG_TYPE_NAME_STRIPED) && image_count == 1)) {
if (image_count > DEFAULT_RAID1_MAX_IMAGES) {
log_error("Only up to %u mirrors in %s LV %s supported currently.",
DEFAULT_RAID1_MAX_IMAGES, lp->segtype->name, display_lvname(lv));
return 0;
}
if (!lv_raid_change_image_count(lv, lp->yes, image_count,
(lp->region_size_supplied || !seg->region_size) ?
lp->region_size : seg->region_size , lp->pvh))
return_0;
log_print_unless_silent("Logical volume %s successfully converted.",
display_lvname(lv));
return 1;
}
goto try_new_takeover_or_reshape;
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
}
if ((seg_is_linear(seg) || seg_is_striped(seg) || seg_is_mirrored(seg) || lv_is_raid(lv)) &&
(lp->type_str && lp->type_str[0])) {
/* Activation is required later which precludes existing supported raid0 segment */
if ((seg_is_any_raid0(seg) || segtype_is_any_raid0(lp->segtype)) &&
!(lp->target_attr & RAID_FEATURE_RAID0)) {
log_error("RAID module does not support RAID0.");
return 0;
}
/* Activation is required later which precludes existing supported raid4 segment */
if (!_raid4_conversion_supported(lv, lp))
2016-11-23 12:30:52 +03:00
return_0;
/* Activation is required later which precludes existing supported raid10 segment */
if ((seg_is_raid10(seg) || segtype_is_raid10(lp->segtype)) &&
!(lp->target_attr & RAID_FEATURE_RAID10)) {
log_error("RAID module does not support RAID10.");
return 0;
}
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
/* FIXME This needs changing globally. */
if (!arg_is_set(cmd, stripes_long_ARG))
lp->stripes = 0;
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!arg_is_set(cmd, type_ARG))
lp->segtype = NULL;
if (!arg_is_set(cmd, regionsize_ARG))
lp->region_size = 0;
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!lv_raid_convert(lv, lp->segtype,
lp->yes, lp->force, lp->stripes, lp->stripe_size_supplied, lp->stripe_size,
lp->region_size, lp->pvh))
return_0;
log_print_unless_silent("Logical volume %s successfully converted.",
display_lvname(lv));
return 1;
}
try_new_takeover_or_reshape:
if (!_raid4_conversion_supported(lv, lp))
return 0;
/* FIXME This needs changing globally. */
if (!arg_is_set(cmd, stripes_long_ARG))
lp->stripes = 0;
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!arg_is_set(cmd, type_ARG))
lp->segtype = NULL;
/* Only let raid4 through for now. */
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!lp->segtype ||
(lp->type_str && lp->type_str[0] && lp->segtype != seg->segtype &&
((seg_is_raid4(seg) && seg_is_striped(lp) && lp->stripes > 1) ||
(seg_is_striped(seg) && seg->area_count > 1 && seg_is_raid4(lp))))) {
if (!lv_raid_convert(lv, lp->segtype,
lp->yes, lp->force, lp->stripes, lp->stripe_size_supplied, lp->stripe_size,
(lp->region_size_supplied || !seg->region_size) ?
lp->region_size : seg->region_size , lp->pvh))
return_0;
log_print_unless_silent("Logical volume %s successfully converted.",
display_lvname(lv));
return 1;
}
log_error("Conversion operation not yet supported.");
return 0;
}
2016-12-21 00:17:48 +03:00
/*
* Functions called to perform a specific operation on a specific LV type.
*
* _convert_<lvtype>_<operation>
*
* For cases where an operation does not apply to the LV itself, but
* is implicitly redirected to a sub-LV, these functions locate the
* correct sub-LV and call the operation on that sub-LV. If a sub-LV
* of the proper type is not found, these functions report the error.
*
* FIXME: the _lvconvert_foo() functions can be cleaned up since they
* are now only called for valid combinations of LV type and operation.
* After that happens, the code remaining in those functions can be
* moved into the _convert_lvtype_operation() functions below.
*/
/*
* Change the number of images in a mirror LV.
* lvconvert --mirrors Number LV
*/
static int _convert_mirror_number(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
2013-12-04 06:09:37 +04:00
{
2016-12-21 00:17:48 +03:00
return _lvconvert_mirrors(cmd, lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Split images from a mirror LV and use them to create a new LV.
* lvconvert --splitmirrors Number LV
*
* Required options:
* --name Name
*/
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
static int _convert_mirror_splitmirrors(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_mirrors(cmd, lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Change the type of log used by a mirror LV.
* lvconvert --mirrorlog Type LV
*/
static int _convert_mirror_log(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_mirrors(cmd, lv, lp);
}
2016-12-21 00:17:48 +03:00
/*
* Convert mirror LV to linear LV.
* lvconvert --type linear LV
*
* Alternate syntax:
* lvconvert --mirrors 0 LV
*/
static int _convert_mirror_linear(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_mirrors(cmd, lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Convert mirror LV to raid1 LV.
* lvconvert --type raid1 LV
*/
static int _convert_mirror_raid(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Change the number of images in a raid1 LV.
* lvconvert --mirrors Number LV
*/
static int _convert_raid_number(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Split images from a raid1 LV and use them to create a new LV.
* lvconvert --splitmirrors Number LV
*
* Required options:
* --trackchanges | --name Name
*/
static int _convert_raid_splitmirrors(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
/* FIXME: split the splitmirrors section out of _lvconvert_raid and call it here. */
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Convert a raid* LV to use a different raid level.
* lvconvert --type raid* LV
*/
static int _convert_raid_raid(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Convert a raid* LV to a mirror LV.
* lvconvert --type mirror LV
*/
static int _convert_raid_mirror(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Convert a raid* LV to a striped LV.
* lvconvert --type striped LV
*/
static int _convert_raid_striped(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2013-12-04 06:09:37 +04:00
2016-12-21 00:17:48 +03:00
/*
* Convert a raid* LV to a linear LV.
* lvconvert --type linear LV
*/
static int _convert_raid_linear(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
2013-12-04 06:09:37 +04:00
}
2016-12-21 00:17:48 +03:00
/*
* Convert a striped/linear LV to a mirror LV.
* lvconvert --type mirror LV
*
* Required options:
* --mirrors Number
*
* Alternate syntax:
* This is equivalent to above when global/mirror_segtype_default="mirror".
* lvconvert --mirrors Number LV
*/
static int _convert_striped_mirror(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
2016-12-21 00:17:48 +03:00
return _lvconvert_mirrors(cmd, lv, lp);
}
2016-12-21 00:17:48 +03:00
/*
* Convert a striped/linear LV to a raid* LV.
* lvconvert --type raid* LV
*
* Required options:
* --mirrors Number
*
* Alternate syntax:
* This is equivalent to above when global/mirror_segtype_default="raid1".
* lvconvert --mirrors Number LV
*/
static int _convert_striped_raid(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
return _lvconvert_raid(lv, lp);
}
2016-12-21 00:17:48 +03:00
static int _convert_mirror(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
if (arg_is_set(cmd, mirrors_ARG))
return _convert_mirror_number(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (arg_is_set(cmd, splitmirrors_ARG))
return _convert_mirror_splitmirrors(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (arg_is_set(cmd, mirrorlog_ARG) || arg_is_set(cmd, corelog_ARG))
return _convert_mirror_log(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (_linear_type_requested(lp->type_str))
return _convert_mirror_linear(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (segtype_is_raid(lp->segtype))
return _convert_mirror_raid(cmd, lv, lp);
log_error("Unknown operation on mirror LV %s.", display_lvname(lv));
return 0;
}
2016-12-21 00:17:48 +03:00
static int _convert_raid(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
2016-12-21 00:17:48 +03:00
if (arg_is_set(cmd, mirrors_ARG))
return _convert_raid_number(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (arg_is_set(cmd, splitmirrors_ARG))
return _convert_raid_splitmirrors(cmd, lv, lp);
2016-12-21 00:17:48 +03:00
if (segtype_is_raid(lp->segtype))
return _convert_raid_raid(cmd, lv, lp);
if (segtype_is_mirror(lp->segtype))
return _convert_raid_mirror(cmd, lv, lp);
if (!strcmp(lp->type_str, SEG_TYPE_NAME_STRIPED))
return _convert_raid_striped(cmd, lv, lp);
if (_linear_type_requested(lp->type_str))
return _convert_raid_linear(cmd, lv, lp);
log_error("Unknown operation on raid LV %s.", display_lvname(lv));
return 0;
}
static int _convert_striped(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
const char *mirrors_type = find_config_tree_str(cmd, global_mirror_segtype_default_CFG, NULL);
if (!strcmp(lp->type_str, SEG_TYPE_NAME_MIRROR))
return _convert_striped_mirror(cmd, lv, lp);
if (segtype_is_raid(lp->segtype))
return _convert_striped_raid(cmd, lv, lp);
/* --mirrors can mean --type mirror or --type raid1 depending on config setting. */
if (arg_is_set(cmd, mirrors_ARG) && mirrors_type && !strcmp(mirrors_type, SEG_TYPE_NAME_MIRROR))
return _convert_striped_mirror(cmd, lv, lp);
if (arg_is_set(cmd, mirrors_ARG) && mirrors_type && !strcmp(mirrors_type, SEG_TYPE_NAME_RAID1))
return _convert_striped_raid(cmd, lv, lp);
log_error("Unknown operation on striped or linear LV %s.", display_lvname(lv));
return 0;
}
static int _lvconvert_raid_types(struct cmd_context *cmd, struct logical_volume *lv,
struct lvconvert_params *lp)
{
struct lv_segment *seg = first_seg(lv);
int ret = 0;
/* If LV is inactive here, ensure it's not active elsewhere. */
if (!lockd_lv(cmd, lv, "ex", 0))
return_ECMD_FAILED;
2016-12-21 00:17:48 +03:00
/* Set up segtype either from type_str or else to match the existing one. */
if (!*lp->type_str)
lp->segtype = seg->segtype;
else if (!(lp->segtype = get_segtype_from_string(cmd, lp->type_str)))
goto_out;
if (!strcmp(lp->type_str, SEG_TYPE_NAME_MIRROR)) {
if (!lp->mirrors_supplied && !seg_is_raid1(seg)) {
log_error("Conversions to --type mirror require -m/--mirrors");
goto out;
}
}
/* lv->segtype can't be NULL */
if (activation() && lp->segtype->ops->target_present &&
!lp->segtype->ops->target_present(cmd, NULL, &lp->target_attr)) {
log_error("%s: Required device-mapper target(s) not "
"detected in your kernel.", lp->segtype->name);
goto out;
}
/* Process striping parameters */
/* FIXME This is incomplete */
if (_mirror_or_raid_type_requested(cmd, lp->type_str) || _raid0_type_requested(lp->type_str) ||
_striped_type_requested(lp->type_str) || lp->mirrorlog || lp->corelog) {
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!arg_is_set(cmd, type_ARG))
lp->segtype = first_seg(lv)->segtype;
2016-12-21 00:17:48 +03:00
/* FIXME Handle +/- adjustments too? */
if (!get_stripe_params(cmd, lp->segtype, &lp->stripes, &lp->stripe_size, &lp->stripes_supplied, &lp->stripe_size_supplied))
goto_out;
if (_raid0_type_requested(lp->type_str) || _striped_type_requested(lp->type_str))
/* FIXME Shouldn't need to override get_stripe_params which defaults to 1 stripe (i.e. linear)! */
/* The default keeps existing number of stripes, handled inside the library code */
if (!arg_is_set(cmd, stripes_long_ARG))
lp->stripes = 0;
}
if (lv_is_cache(lv))
lv = seg_lv(first_seg(lv), 0);
if (lv_is_mirror(lv)) {
ret = _convert_mirror(cmd, lv, lp);
goto out;
}
if (lv_is_raid(lv)) {
ret = _convert_raid(cmd, lv, lp);
goto out;
}
/*
* FIXME: add lv_is_striped() and lv_is_linear()?
* This does not include raid0 which is caught by the test above.
* If operations differ between striped and linear, split this case.
*/
if (segtype_is_striped(seg->segtype) || segtype_is_linear(seg->segtype)) {
ret = _convert_striped(cmd, lv, lp);
goto out;
}
/*
* The intention is to explicitly check all cases above and never
* reach here, but this covers anything that was missed.
*/
log_error("Cannot convert LV %s.", display_lvname(lv));
out:
return ret ? ECMD_PROCESSED : ECMD_FAILED;
}
static int _lvconvert_splitsnapshot(struct cmd_context *cmd, struct logical_volume *cow)
{
struct volume_group *vg = cow->vg;
const char *cow_name = display_lvname(cow);
if (lv_is_virtual_origin(origin_from_cow(cow))) {
log_error("Unable to split off snapshot %s with virtual origin.", cow_name);
return 0;
}
if (is_lockd_type(vg->lock_type)) {
/* FIXME: we need to create a lock for the new LV. */
log_error("Unable to split snapshots in VG with lock_type %s.", vg->lock_type);
return 0;
}
if (lv_is_active_locally(cow)) {
if (!lv_check_not_in_use(cow, 1))
return_0;
if ((arg_count(cmd, force_ARG) == PROMPT) &&
!arg_count(cmd, yes_ARG) &&
lv_is_visible(cow) &&
lv_is_active(cow)) {
if (yes_no_prompt("Do you really want to split off active "
"logical volume %s? [y/n]: ", display_lvname(cow)) == 'n') {
log_error("Logical volume %s not split.", display_lvname(cow));
return 0;
}
}
}
if (!archive(vg))
return_0;
log_verbose("Splitting snapshot %s from its origin.", display_lvname(cow));
if (!vg_remove_snapshot(cow))
return_0;
backup(vg);
log_print_unless_silent("Logical Volume %s split from its origin.", display_lvname(cow));
return 1;
}
static int _lvconvert_split_and_keep_cachepool(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume *cachepool_lv)
{
log_debug("Detaching cache pool %s from cache LV %s.",
display_lvname(cachepool_lv), display_lvname(lv));
if (!archive(lv->vg))
return_0;
if (!lv_cache_remove(lv))
return_0;
if (!vg_write(lv->vg) || !vg_commit(lv->vg))
return_0;
backup(lv->vg);
log_print_unless_silent("Logical volume %s is not cached and cache pool %s is unused.",
display_lvname(lv), display_lvname(cachepool_lv));
return 1;
}
static int _lvconvert_split_and_remove_cachepool(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume *cachepool_lv)
{
struct lv_segment *seg;
struct logical_volume *remove_lv;
seg = first_seg(lv);
if (lv_is_partial(seg_lv(seg, 0))) {
log_warn("WARNING: Cache origin logical volume %s is missing.",
display_lvname(seg_lv(seg, 0)));
remove_lv = lv; /* When origin is missing, drop everything */
} else
remove_lv = seg->pool_lv;
if (lv_is_partial(seg_lv(first_seg(seg->pool_lv), 0)))
log_warn("WARNING: Cache pool data logical volume %s is missing.",
display_lvname(seg_lv(first_seg(seg->pool_lv), 0)));
if (lv_is_partial(first_seg(seg->pool_lv)->metadata_lv))
log_warn("WARNING: Cache pool metadata logical volume %s is missing.",
display_lvname(first_seg(seg->pool_lv)->metadata_lv));
/* TODO: Check for failed cache as well to get prompting? */
if (lv_is_partial(lv)) {
if (first_seg(seg->pool_lv)->cache_mode != CACHE_MODE_WRITETHROUGH) {
if (!arg_count(cmd, force_ARG)) {
log_error("Conversion aborted.");
log_error("Cannot uncache writethrough cache volume %s without --force.",
display_lvname(lv));
return 0;
}
log_warn("WARNING: Uncaching of partially missing writethrough cache volume %s might destroy your data.",
display_lvname(lv));
}
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Do you really want to uncache %s with missing LVs? [y/n]: ",
display_lvname(lv)) == 'n') {
log_error("Conversion aborted.");
return 0;
}
}
if (lvremove_single(cmd, remove_lv, NULL) != ECMD_PROCESSED)
return_0;
if (remove_lv != lv)
log_print_unless_silent("Logical volume %s is not cached.", display_lvname(lv));
return 1;
}
2013-12-04 06:09:37 +04:00
static int _lvconvert_snapshot(struct cmd_context *cmd,
struct logical_volume *lv,
const char *origin_name)
2006-04-06 00:43:23 +04:00
{
struct logical_volume *org;
2016-03-07 12:45:50 +03:00
const char *snap_name = display_lvname(lv);
uint32_t chunk_size;
int zero;
2006-04-06 00:43:23 +04:00
if (strcmp(lv->name, origin_name) == 0) {
log_error("Unable to use %s as both snapshot and origin.", snap_name);
return 0;
}
chunk_size = arg_uint_value(cmd, chunksize_ARG, 8);
if (chunk_size < 8 || chunk_size > 1024 || !is_power_of_2(chunk_size)) {
log_error("Chunk size must be a power of 2 in the range 4K to 512K.");
return 0;
}
if (!cow_has_min_chunks(lv->vg, lv->le_count, chunk_size))
return_0;
2006-04-06 00:43:23 +04:00
log_verbose("Setting chunk size to %s.", display_size(cmd, chunk_size));
if (!(org = find_lv(lv->vg, origin_name))) {
log_error("Couldn't find origin volume %s in Volume group %s.",
origin_name, lv->vg->name);
return 0;
}
/*
* check_lv_rules() checks cannot be done via command definition
* rules because this LV is not processed by process_each_lv.
*/
/*
* check_lv_types() checks cannot be done via command definition
* LV_foo specification because this LV is not processed by process_each_lv.
*/
if (!validate_snapshot_origin(org))
2018-02-26 16:44:51 +03:00
return_0;
2006-04-06 00:43:23 +04:00
if (lv_component_is_active(org)) {
log_error("Cannot use logical volume %s with active component LVs for snapshot origin.",
display_lvname(org));
return 0;
}
log_warn("WARNING: Converting logical volume %s to snapshot exception store.",
2016-03-07 12:45:50 +03:00
snap_name);
log_warn("THIS WILL DESTROY CONTENT OF LOGICAL VOLUME (filesystem etc.)");
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Do you really want to convert %s? [y/n]: ",
2016-03-07 12:45:50 +03:00
snap_name) == 'n') {
log_error("Conversion aborted.");
return 0;
2006-04-06 00:43:23 +04:00
}
if (!deactivate_lv(cmd, lv)) {
2016-03-07 12:45:50 +03:00
log_error("Couldn't deactivate logical volume %s.", snap_name);
2006-04-06 00:43:23 +04:00
return 0;
}
if (first_seg(lv)->segtype->flags & SEG_CANNOT_BE_ZEROED)
zero = 0;
else
zero = arg_int_value(cmd, zero_ARG, 1);
if (!zero || !(lv->status & LVM_WRITE))
log_warn("WARNING: %s not zeroed.", snap_name);
else {
lv->status |= LV_TEMPORARY;
if (!activate_lv_excl_local(cmd, lv) ||
!wipe_lv(lv, (struct wipe_params) { .do_zero = 1 })) {
log_error("Aborting. Failed to wipe snapshot exception store.");
return 0;
}
lv->status &= ~LV_TEMPORARY;
/* Deactivates cleared metadata LV */
if (!deactivate_lv(lv->vg->cmd, lv)) {
log_error("Failed to deactivate zeroed snapshot exception store.");
return 0;
}
}
if (!archive(lv->vg))
return_0;
if (!vg_add_snapshot(org, lv, NULL, org->le_count, chunk_size)) {
2006-04-06 00:43:23 +04:00
log_error("Couldn't create snapshot.");
return 0;
}
/* store vg on disk(s) */
if (!lv_update_and_reload(org))
2006-04-06 00:43:23 +04:00
return_0;
2016-03-07 12:45:50 +03:00
log_print_unless_silent("Logical volume %s converted to snapshot.", snap_name);
return 1;
2006-04-06 00:43:23 +04:00
}
static int _lvconvert_merge_old_snapshot(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume **lv_to_poll)
{
int merge_on_activate = 0;
struct logical_volume *origin = origin_from_cow(lv);
struct lv_segment *snap_seg = find_snapshot(lv);
struct lvinfo info;
dm_percent_t snap_percent;
/* Check if merge is possible */
if (lv_is_merging_origin(origin)) {
log_error("Cannot merge snapshot %s into the origin %s "
"with merging snapshot %s.",
display_lvname(lv), display_lvname(origin),
display_lvname(find_snapshot(origin)->lv));
return 0;
}
if (lv_is_external_origin(origin_from_cow(lv))) {
log_error("Cannot merge snapshot %s into "
"the read-only external origin %s.",
display_lvname(lv),
display_lvname(origin_from_cow(lv)));
return 0;
}
2014-09-23 18:49:46 +04:00
/* FIXME: test when snapshot is remotely active */
if (lv_info(cmd, lv, 0, &info, 1, 0)
&& info.exists && info.live_table &&
(!lv_snapshot_percent(lv, &snap_percent) ||
snap_percent == DM_PERCENT_INVALID)) {
log_error("Unable to merge invalidated snapshot LV %s.",
display_lvname(lv));
return 0;
}
if (snap_seg->segtype->ops->target_present &&
!snap_seg->segtype->ops->target_present(cmd, snap_seg, NULL)) {
log_error("Can't initialize snapshot merge. "
"Missing support in kernel?");
return 0;
}
if (!archive(lv->vg))
return_0;
/*
* Prevent merge with open device(s) as it would likely lead
* to application/filesystem failure. Merge on origin's next
* activation if either the origin or snapshot LV are currently
* open.
*
* FIXME testing open_count is racey; snapshot-merge target's
* constructor and DM should prevent appropriate devices from
* being open.
*/
if (lv_is_active_locally(origin)) {
if (!lv_check_not_in_use(origin, 0)) {
log_print_unless_silent("Delaying merge since origin is open.");
merge_on_activate = 1;
} else if (!lv_check_not_in_use(lv, 0)) {
log_print_unless_silent("Delaying merge since snapshot is open.");
merge_on_activate = 1;
}
} else if (vg_is_clustered(origin->vg) && lv_is_active(origin)) {
/* When it's active somewhere else */
log_print_unless_silent("Delaying merge since origin is remotely active.");
merge_on_activate = 1;
}
init_snapshot_merge(snap_seg, origin);
if (merge_on_activate) {
/* Store and commit vg but skip starting the merge */
if (!vg_write(lv->vg) || !vg_commit(lv->vg))
return_0;
backup(lv->vg);
} else {
/* Perform merge */
if (!lv_update_and_reload(origin))
return_0;
if (!lv_has_target_type(origin->vg->vgmem, origin, NULL,
TARGET_NAME_SNAPSHOT_MERGE)) {
/* Race during table reload prevented merging */
merge_on_activate = 1;
} else if (!lv_info(cmd, origin, 0, &info, 0, 0) || !info.exists) {
log_print_unless_silent("Conversion starts after activation.");
merge_on_activate = 1;
} else {
*lv_to_poll = origin;
}
}
if (merge_on_activate)
log_print_unless_silent("Merging of snapshot %s will occur on "
"next activation of %s.",
display_lvname(lv), display_lvname(origin));
else
log_print_unless_silent("Merging of volume %s started.",
display_lvname(lv));
return 1;
}
2013-11-30 00:28:18 +04:00
static int _lvconvert_merge_thin_snapshot(struct cmd_context *cmd,
struct logical_volume *lv)
2013-11-30 00:28:18 +04:00
{
int origin_is_active = 0;
2013-11-30 00:28:18 +04:00
struct lv_segment *snap_seg = first_seg(lv);
struct logical_volume *origin = snap_seg->origin;
if (!origin) {
log_error("%s is not a mergeable logical volume.",
display_lvname(lv));
2013-11-30 00:28:18 +04:00
return 0;
}
/* Check if merge is possible */
if (lv_is_merging_origin(origin)) {
log_error("Cannot merge snapshot %s into the origin %s "
"with merging snapshot %s.",
display_lvname(lv), display_lvname(origin),
display_lvname(find_snapshot(origin)->lv));
2013-11-30 00:28:18 +04:00
return 0;
}
if (lv_is_external_origin(origin)) {
if (!(origin = origin_from_cow(lv)))
log_error(INTERNAL_ERROR "%s is missing origin.",
display_lvname(lv));
else
log_error("%s is read-only external origin %s.",
display_lvname(lv), display_lvname(origin));
2013-11-30 00:28:18 +04:00
return 0;
}
if (lv_is_origin(origin)) {
log_error("Merging into the old snapshot origin %s is not supported.",
display_lvname(origin));
2013-11-30 00:28:18 +04:00
return 0;
}
if (!archive(lv->vg))
return_0;
/*
* Prevent merge with open device(s) as it would likely lead
* to application/filesystem failure. Merge on origin's next
* activation if either the origin or snapshot LV can't be
* deactivated.
*/
if (!deactivate_lv(cmd, lv))
log_print_unless_silent("Delaying merge since snapshot is open.");
else if ((origin_is_active = lv_is_active(origin)) &&
!deactivate_lv(cmd, origin))
log_print_unless_silent("Delaying merge since origin volume is open.");
else {
/*
* Both thin snapshot and origin are inactive,
* replace the origin LV with its snapshot LV.
*/
if (!thin_merge_finish(cmd, origin, lv))
return_0;
log_print_unless_silent("Volume %s replaced origin %s.",
display_lvname(origin), display_lvname(lv));
2013-11-30 00:28:18 +04:00
if (origin_is_active && !activate_lv_excl_local(cmd, lv)) {
log_error("Failed to reactivate origin %s.",
display_lvname(lv));
return 0;
2013-11-30 00:28:18 +04:00
}
return 1;
2013-11-30 00:28:18 +04:00
}
init_snapshot_merge(snap_seg, origin);
/* Commit vg, merge will start with next activation */
if (!vg_write(lv->vg) || !vg_commit(lv->vg))
return_0;
log_print_unless_silent("Merging of thin snapshot %s will occur on "
"next activation of %s.",
display_lvname(lv), display_lvname(origin));
2013-11-30 00:28:18 +04:00
backup(lv->vg);
return 1;
2013-11-30 00:28:18 +04:00
}
static int _lvconvert_thin_pool_repair(struct cmd_context *cmd,
struct logical_volume *pool_lv,
struct dm_list *pvh, int poolmetadataspare)
{
const char *dmdir = dm_dir();
const char *thin_dump =
find_config_tree_str_allow_empty(cmd, global_thin_dump_executable_CFG, NULL);
const char *thin_repair =
find_config_tree_str_allow_empty(cmd, global_thin_repair_executable_CFG, NULL);
const struct dm_config_node *cn;
const struct dm_config_value *cv;
int ret = 0, status;
int args = 0;
const char *argv[MAX_PDATA_ARGS + 7]; /* Max supported args */
char *dm_name, *trans_id_str;
char meta_path[PATH_MAX];
char pms_path[PATH_MAX];
uint64_t trans_id;
struct logical_volume *pmslv;
struct logical_volume *mlv = first_seg(pool_lv)->metadata_lv;
struct pipe_data pdata;
FILE *f;
if (!thin_repair || !thin_repair[0]) {
log_error("Thin repair commnand is not configured. Repair is disabled.");
return 0; /* Checking disabled */
}
pmslv = pool_lv->vg->pool_metadata_spare_lv;
/* Check we have pool metadata spare LV */
if (!handle_pool_metadata_spare(pool_lv->vg, 0, pvh, 1))
return_0;
if (pmslv != pool_lv->vg->pool_metadata_spare_lv) {
if (!vg_write(pool_lv->vg) || !vg_commit(pool_lv->vg))
return_0;
pmslv = pool_lv->vg->pool_metadata_spare_lv;
}
if (!(dm_name = dm_build_dm_name(cmd->mem, mlv->vg->name,
mlv->name, NULL)) ||
(dm_snprintf(meta_path, sizeof(meta_path), "%s/%s", dmdir, dm_name) < 0)) {
log_error("Failed to build thin metadata path.");
return 0;
}
if (!(dm_name = dm_build_dm_name(cmd->mem, pmslv->vg->name,
pmslv->name, NULL)) ||
(dm_snprintf(pms_path, sizeof(pms_path), "%s/%s", dmdir, dm_name) < 0)) {
log_error("Failed to build pool metadata spare path.");
return 0;
}
if (!(cn = find_config_tree_array(cmd, global_thin_repair_options_CFG, NULL))) {
log_error(INTERNAL_ERROR "Unable to find configuration for global/thin_repair_options");
return 0;
}
for (cv = cn->v; cv && args < MAX_PDATA_ARGS; cv = cv->next) {
if (cv->type != DM_CFG_STRING) {
log_error("Invalid string in config file: "
"global/thin_repair_options");
return 0;
}
argv[++args] = cv->v.str;
}
if (args >= MAX_PDATA_ARGS) {
log_error("Too many options for thin repair command.");
return 0;
}
argv[0] = thin_repair;
argv[++args] = "-i";
argv[++args] = meta_path;
argv[++args] = "-o";
argv[++args] = pms_path;
argv[++args] = NULL;
if (pool_is_active(pool_lv)) {
log_error("Active pools cannot be repaired. Use lvchange -an first.");
return 0;
}
if (!activate_lv_excl_local(cmd, pmslv)) {
log_error("Cannot activate pool metadata spare volume %s.",
pmslv->name);
return 0;
}
if (!activate_lv_excl_local(cmd, mlv)) {
log_error("Cannot activate thin pool metadata volume %s.",
mlv->name);
goto deactivate_pmslv;
}
if (!(ret = exec_cmd(cmd, (const char * const *)argv, &status, 1))) {
log_error("Repair of thin metadata volume of thin pool %s failed (status:%d). "
"Manual repair required!",
display_lvname(pool_lv), status);
goto deactivate_mlv;
}
if (thin_dump[0]) {
argv[0] = thin_dump;
argv[1] = pms_path;
argv[2] = NULL;
if (!(f = pipe_open(cmd, argv, 0, &pdata)))
log_warn("WARNING: Cannot read output from %s %s.", thin_dump, pms_path);
else {
/*
* Scan only the 1st. line for transation id.
* Watch out, if the thin_dump format changes
*/
if (fgets(meta_path, sizeof(meta_path), f) &&
(trans_id_str = strstr(meta_path, "transaction=\"")) &&
(sscanf(trans_id_str + 13, FMTu64, &trans_id) == 1) &&
(trans_id != first_seg(pool_lv)->transaction_id) &&
((trans_id - 1) != first_seg(pool_lv)->transaction_id))
log_error("Transaction id " FMTu64 " from pool \"%s/%s\" "
"does not match repaired transaction id "
FMTu64 " from %s.",
first_seg(pool_lv)->transaction_id,
pool_lv->vg->name, pool_lv->name, trans_id,
pms_path);
(void) pipe_close(&pdata); /* killing pipe */
}
}
deactivate_mlv:
if (!deactivate_lv(cmd, mlv)) {
log_error("Cannot deactivate thin pool metadata volume %s.",
display_lvname(mlv));
ret = 0;
}
deactivate_pmslv:
if (!deactivate_lv(cmd, pmslv)) {
log_error("Cannot deactivate pool metadata spare volume %s.",
display_lvname(pmslv));
ret = 0;
}
if (!ret)
return 0;
if (pmslv == pool_lv->vg->pool_metadata_spare_lv) {
pool_lv->vg->pool_metadata_spare_lv = NULL;
pmslv->status &= ~POOL_METADATA_SPARE;
lv_set_visible(pmslv);
}
/* Try to allocate new pool metadata spare LV */
if (!handle_pool_metadata_spare(pool_lv->vg, 0, pvh, poolmetadataspare))
stack;
if (dm_snprintf(meta_path, sizeof(meta_path), "%s_meta%%d", pool_lv->name) < 0) {
log_error("Can't prepare new metadata name for %s.", pool_lv->name);
return 0;
}
if (!generate_lv_name(pool_lv->vg, meta_path, pms_path, sizeof(pms_path))) {
log_error("Can't generate new name for %s.", meta_path);
return 0;
}
if (!detach_pool_metadata_lv(first_seg(pool_lv), &mlv))
return_0;
/* Swap _pmspare and _tmeta name */
if (!swap_lv_identifiers(cmd, mlv, pmslv))
return_0;
if (!attach_pool_metadata_lv(first_seg(pool_lv), pmslv))
return_0;
/* Used _tmeta (now _pmspare) becomes _meta%d */
if (!lv_rename_update(cmd, mlv, pms_path, 0))
return_0;
if (!vg_write(pool_lv->vg) || !vg_commit(pool_lv->vg))
return_0;
log_warn("WARNING: LV %s holds a backup of the unrepaired metadata. Use lvremove when no longer required.",
display_lvname(mlv));
if (dm_list_size(&pool_lv->vg->pvs) > 1)
log_warn("WARNING: New metadata LV %s might use different PVs. Move it with pvmove if required.",
display_lvname(first_seg(pool_lv)->metadata_lv));
return 1;
}
/* TODO: lots of similar code with thinpool repair
* investigate possible better code sharing...
*/
static int _lvconvert_cache_repair(struct cmd_context *cmd,
struct logical_volume *cache_lv,
struct dm_list *pvh, int poolmetadataspare)
{
const char *dmdir = dm_dir();
const char *cache_repair =
find_config_tree_str_allow_empty(cmd, global_cache_repair_executable_CFG, NULL);
const struct dm_config_node *cn;
const struct dm_config_value *cv;
int ret = 0, status;
int args = 0;
const char *argv[MAX_PDATA_ARGS + 7]; /* Max supported args */
char *dm_name;
char meta_path[PATH_MAX];
char pms_path[PATH_MAX];
struct logical_volume *pool_lv;
struct logical_volume *pmslv;
struct logical_volume *mlv;
pool_lv = lv_is_cache_pool(cache_lv) ? cache_lv : first_seg(cache_lv)->pool_lv;
mlv = first_seg(pool_lv)->metadata_lv;
if (!cache_repair || !cache_repair[0]) {
log_error("Cache repair commnand is not configured. Repair is disabled.");
return 0; /* Checking disabled */
}
pmslv = cache_lv->vg->pool_metadata_spare_lv;
/* Check we have pool metadata spare LV */
if (!handle_pool_metadata_spare(cache_lv->vg, 0, pvh, 1))
return_0;
if (pmslv != cache_lv->vg->pool_metadata_spare_lv) {
if (!vg_write(cache_lv->vg) || !vg_commit(cache_lv->vg))
return_0;
pmslv = cache_lv->vg->pool_metadata_spare_lv;
}
if (!(dm_name = dm_build_dm_name(cmd->mem, mlv->vg->name,
mlv->name, NULL)) ||
(dm_snprintf(meta_path, sizeof(meta_path), "%s/%s", dmdir, dm_name) < 0)) {
log_error("Failed to build cache metadata path.");
return 0;
}
if (!(dm_name = dm_build_dm_name(cmd->mem, pmslv->vg->name,
pmslv->name, NULL)) ||
(dm_snprintf(pms_path, sizeof(pms_path), "%s/%s", dmdir, dm_name) < 0)) {
log_error("Failed to build pool metadata spare path.");
return 0;
}
if (!(cn = find_config_tree_array(cmd, global_cache_repair_options_CFG, NULL))) {
log_error(INTERNAL_ERROR "Unable to find configuration for global/cache_repair_options");
return 0;
}
for (cv = cn->v; cv && args < MAX_PDATA_ARGS; cv = cv->next) {
if (cv->type != DM_CFG_STRING) {
log_error("Invalid string in config file: "
"global/cache_repair_options");
return 0;
}
argv[++args] = cv->v.str;
}
if (args >= MAX_PDATA_ARGS) {
log_error("Too many options for cache repair command.");
return 0;
}
argv[0] = cache_repair;
argv[++args] = "-i";
argv[++args] = meta_path;
argv[++args] = "-o";
argv[++args] = pms_path;
argv[++args] = NULL;
if (lv_is_active(cache_lv)) {
log_error("Only inactive cache can be repaired.");
return 0;
}
if (!activate_lv_excl_local(cmd, pmslv)) {
log_error("Cannot activate pool metadata spare volume %s.",
pmslv->name);
return 0;
}
if (!activate_lv_excl_local(cmd, mlv)) {
log_error("Cannot activate cache pool metadata volume %s.",
mlv->name);
goto deactivate_pmslv;
}
if (!(ret = exec_cmd(cmd, (const char * const *)argv, &status, 1))) {
log_error("Repair of cache metadata volume of cache %s failed (status:%d). "
"Manual repair required!",
display_lvname(cache_lv), status);
goto deactivate_mlv;
}
/* TODO: any active validation of cache-pool metadata? */
deactivate_mlv:
if (!deactivate_lv(cmd, mlv)) {
log_error("Cannot deactivate pool metadata volume %s.",
display_lvname(mlv));
ret = 0;
}
deactivate_pmslv:
if (!deactivate_lv(cmd, pmslv)) {
log_error("Cannot deactivate pool metadata spare volume %s.",
display_lvname(pmslv));
ret = 0;
}
if (!ret)
return 0;
if (pmslv == cache_lv->vg->pool_metadata_spare_lv) {
cache_lv->vg->pool_metadata_spare_lv = NULL;
pmslv->status &= ~POOL_METADATA_SPARE;
lv_set_visible(pmslv);
}
/* Try to allocate new pool metadata spare LV */
if (!handle_pool_metadata_spare(cache_lv->vg, 0, pvh, poolmetadataspare))
stack;
if (dm_snprintf(meta_path, sizeof(meta_path), "%s_meta%%d", cache_lv->name) < 0) {
log_error("Can't prepare new metadata name for %s.", cache_lv->name);
return 0;
}
if (!generate_lv_name(cache_lv->vg, meta_path, pms_path, sizeof(pms_path))) {
log_error("Can't generate new name for %s.", meta_path);
return 0;
}
if (!detach_pool_metadata_lv(first_seg(pool_lv), &mlv))
return_0;
/* Swap _pmspare and _cmeta name */
if (!swap_lv_identifiers(cmd, mlv, pmslv))
return_0;
if (!attach_pool_metadata_lv(first_seg(pool_lv), pmslv))
return_0;
/* Used _cmeta (now _pmspare) becomes _meta%d */
if (!lv_rename_update(cmd, mlv, pms_path, 0))
return_0;
if (!vg_write(cache_lv->vg) || !vg_commit(cache_lv->vg))
return_0;
/* FIXME: just as with thinpool repair - fix the warning
* where moving doesn't make any sense (same disk storage)
*/
log_warn("WARNING: If everything works, remove %s volume.",
display_lvname(mlv));
log_warn("WARNING: Use pvmove command to move %s on the best fitting PV.",
display_lvname(first_seg(pool_lv)->metadata_lv));
return 1;
}
2016-12-21 00:17:48 +03:00
static int _lvconvert_to_thin_with_external(struct cmd_context *cmd,
struct logical_volume *lv,
2016-12-21 00:17:48 +03:00
struct logical_volume *thinpool_lv)
{
struct volume_group *vg = lv->vg;
2016-12-21 00:17:48 +03:00
struct logical_volume *thin_lv;
const char *origin_name;
struct lvcreate_params lvc = {
.activate = CHANGE_AEY,
.alloc = ALLOC_INHERIT,
.major = -1,
.minor = -1,
.suppress_zero_warn = 1, /* Suppress warning for this thin */
.permission = LVM_READ,
2016-12-21 00:17:48 +03:00
.pool_name = thinpool_lv->name,
.pvh = &vg->pvs,
.read_ahead = DM_READ_AHEAD_AUTO,
.stripes = 1,
.virtual_extents = lv->le_count,
};
2016-12-21 00:17:48 +03:00
if (lv == thinpool_lv) {
log_error("Can't use same LV %s for thin pool and thin volume.",
2016-12-21 00:17:48 +03:00
display_lvname(thinpool_lv));
return 0;
}
2016-12-21 00:17:48 +03:00
if ((origin_name = arg_str_value(cmd, originname_ARG, NULL)))
if (!validate_restricted_lvname_param(cmd, &vg->name, &origin_name))
return_0;
/*
* If NULL, an auto-generated 'lvol' name is used.
* If set, the lv create code checks the name isn't used.
*/
lvc.lv_name = origin_name;
2016-12-21 00:17:48 +03:00
if (is_lockd_type(vg->lock_type)) {
/*
* FIXME: external origins don't work in lockd VGs.
* Prior to the lvconvert, there's a lock associated with
* the uuid of the external origin LV. After the convert,
* that uuid belongs to the new thin LV, and a new LV with
* a new uuid exists as the non-thin, readonly external LV.
* We'd need to remove the lock for the previous uuid
* (the new thin LV will have no lock), and create a new
* lock for the new LV uuid used by the external LV.
*/
2015-03-05 23:00:44 +03:00
log_error("Can't use lock_type %s LV as external origin.",
2016-12-21 00:17:48 +03:00
vg->lock_type);
2015-03-05 23:00:44 +03:00
return 0;
}
dm_list_init(&lvc.tags);
2016-12-21 00:17:48 +03:00
if (!pool_supports_external_origin(first_seg(thinpool_lv), lv))
return_0;
if (!(lvc.segtype = get_segtype_from_string(cmd, SEG_TYPE_NAME_THIN)))
return_0;
if (!archive(vg))
return_0;
2016-12-21 00:17:48 +03:00
/*
* New thin LV needs to be created (all messages sent to pool) In this
* case thin volume is created READ-ONLY and also warn about not
* zeroing is suppressed.
*
* The new thin LV is created with the origin_name, or an autogenerated
* 'lvol' name. Then the names and ids are swapped between the thin LV
* and the original/external LV. So, the thin LV gets the name and id
* of the original LV arg, and the original LV arg gets the origin_name
* or the autogenerated name.
*/
if (!(thin_lv = lv_create_single(vg, &lvc)))
return_0;
2016-12-21 00:17:48 +03:00
if (!deactivate_lv(cmd, thin_lv)) {
log_error("Aborting. Unable to deactivate new LV. "
"Manual intervention required.");
return 0;
}
/*
* Crashing till this point will leave plain thin volume
* which could be easily removed by the user after i.e. power-off
*/
2016-12-21 00:17:48 +03:00
if (!swap_lv_identifiers(cmd, thin_lv, lv)) {
stack;
goto revert_new_lv;
}
/* Preserve read-write status of original LV here */
2016-12-21 00:17:48 +03:00
thin_lv->status |= (lv->status & LVM_WRITE);
2016-12-21 00:17:48 +03:00
if (!attach_thin_external_origin(first_seg(thin_lv), lv)) {
stack;
goto revert_new_lv;
}
2016-12-21 00:17:48 +03:00
if (!lv_update_and_reload(thin_lv)) {
stack;
goto deactivate_and_revert_new_lv;
}
2016-12-21 00:17:48 +03:00
log_print_unless_silent("Converted %s to thin volume with external origin %s.",
display_lvname(thin_lv), display_lvname(lv));
return 1;
deactivate_and_revert_new_lv:
2016-12-21 00:17:48 +03:00
if (!swap_lv_identifiers(cmd, thin_lv, lv))
stack;
2016-12-21 00:17:48 +03:00
if (!deactivate_lv(cmd, thin_lv)) {
log_error("Unable to deactivate failed new LV. "
"Manual intervention required.");
return 0;
}
2016-12-21 00:17:48 +03:00
if (!detach_thin_external_origin(first_seg(thin_lv)))
return_0;
revert_new_lv:
/* FIXME Better to revert to backup of metadata? */
2016-12-21 00:17:48 +03:00
if (!lv_remove(thin_lv) || !vg_write(vg) || !vg_commit(vg))
log_error("Manual intervention may be required to remove "
"abandoned LV(s) before retrying.");
else
backup(vg);
return 0;
}
2016-12-21 00:17:48 +03:00
static int _lvconvert_swap_pool_metadata(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume *metadata_lv)
{
struct volume_group *vg = lv->vg;
2016-12-21 00:17:48 +03:00
struct logical_volume *prev_metadata_lv;
struct lv_segment *seg;
struct lv_type *lvtype;
2016-12-21 00:17:48 +03:00
char meta_name[NAME_LEN];
const char *swap_name;
uint32_t chunk_size;
int is_thinpool;
int is_cachepool;
int lvt_enum;
2016-12-21 00:17:48 +03:00
is_thinpool = lv_is_thin_pool(lv);
is_cachepool = lv_is_cache_pool(lv);
lvt_enum = get_lvt_enum(metadata_lv);
lvtype = get_lv_type(lvt_enum);
2016-12-21 00:17:48 +03:00
if (lvt_enum != striped_LVT && lvt_enum != linear_LVT && lvt_enum != raid_LVT) {
log_error("LV %s with type %s cannot be used as a metadata LV.",
display_lvname(metadata_lv), lvtype ? lvtype->name : "unknown");
return 0;
}
2016-12-21 00:17:48 +03:00
if (!lv_is_visible(metadata_lv)) {
log_error("Can't convert internal LV %s.",
display_lvname(metadata_lv));
return 0;
}
2016-12-21 00:17:48 +03:00
if (lv_is_locked(metadata_lv)) {
log_error("Can't convert locked LV %s.",
display_lvname(metadata_lv));
return 0;
}
2016-12-21 00:17:48 +03:00
if (lv_is_origin(metadata_lv) ||
lv_is_merging_origin(metadata_lv) ||
lv_is_external_origin(metadata_lv) ||
lv_is_virtual(metadata_lv)) {
log_error("Pool metadata LV %s is of an unsupported type.",
display_lvname(metadata_lv));
return 0;
}
2016-12-21 00:17:48 +03:00
/* FIXME cache pool */
if (is_thinpool && pool_is_active(lv)) {
/* If any volume referencing pool active - abort here */
log_error("Cannot convert pool %s with active volumes.",
display_lvname(lv));
return 0;
}
2016-12-21 00:17:48 +03:00
if ((dm_snprintf(meta_name, sizeof(meta_name), "%s%s", lv->name, is_cachepool ? "_cmeta" : "_tmeta") < 0)) {
log_error("Failed to create internal lv names, pool name is too long.");
return 0;
}
2016-12-21 00:17:48 +03:00
seg = first_seg(lv);
2016-12-21 00:17:48 +03:00
/* Normally do NOT change chunk size when swapping */
2016-12-21 00:17:48 +03:00
if (arg_is_set(cmd, chunksize_ARG)) {
chunk_size = arg_uint_value(cmd, chunksize_ARG, 0);
2016-12-21 00:17:48 +03:00
if ((chunk_size != seg->chunk_size) && !dm_list_empty(&lv->segs_using_this_lv)) {
if (arg_count(cmd, force_ARG) == PROMPT) {
log_error("Chunk size can be only changed with --force. Conversion aborted.");
return 0;
}
2016-12-21 00:17:48 +03:00
if (!validate_pool_chunk_size(cmd, seg->segtype, chunk_size))
return_0;
2016-12-21 00:17:48 +03:00
log_warn("WARNING: Changing chunk size %s to %s for %s pool volume.",
display_size(cmd, seg->chunk_size),
display_size(cmd, chunk_size),
display_lvname(lv));
2016-12-21 00:17:48 +03:00
/* Ok, user has likely some serious reason for this */
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Do you really want to change chunk size for %s pool volume? [y/n]: ",
display_lvname(lv)) == 'n') {
log_error("Conversion aborted.");
return 0;
}
}
2016-12-21 00:17:48 +03:00
seg->chunk_size = chunk_size;
}
2016-12-21 00:17:48 +03:00
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Do you want to swap metadata of %s pool with metadata volume %s? [y/n]: ",
display_lvname(lv),
display_lvname(metadata_lv)) == 'n') {
log_error("Conversion aborted.");
return 0;
}
2016-12-21 00:17:48 +03:00
if (!deactivate_lv(cmd, metadata_lv)) {
log_error("Aborting. Failed to deactivate %s.",
display_lvname(metadata_lv));
return 0;
}
2016-12-21 00:17:48 +03:00
if (!archive(vg))
return_0;
2016-12-21 00:17:48 +03:00
/* Swap names between old and new metadata LV */
2016-12-21 00:17:48 +03:00
if (!detach_pool_metadata_lv(seg, &prev_metadata_lv))
return_0;
2016-12-21 00:17:48 +03:00
swap_name = metadata_lv->name;
2016-12-21 00:17:48 +03:00
if (!lv_rename_update(cmd, metadata_lv, "pvmove_tmeta", 0))
return_0;
2016-12-21 00:17:48 +03:00
/* Give the previous metadata LV the name of the LV replacing it. */
2016-12-21 00:17:48 +03:00
if (!lv_rename_update(cmd, prev_metadata_lv, swap_name, 0))
return_0;
2016-12-21 00:17:48 +03:00
/* Rename deactivated metadata LV to have _tmeta suffix */
if (!lv_rename_update(cmd, metadata_lv, meta_name, 0))
return_0;
if (!attach_pool_metadata_lv(seg, metadata_lv))
return_0;
if (!vg_write(vg) || !vg_commit(vg))
return_0;
2016-12-21 00:17:48 +03:00
backup(vg);
return 1;
}
/*
2016-12-21 00:17:48 +03:00
* Create a new pool LV, using the lv arg as the data sub LV.
* The metadata sub LV is either a new LV created here, or an
* existing LV specified by --poolmetadata.
*
* process_single_lv is the LV currently being processed by
* process_each_lv(). It will sometimes be the same as the
* lv arg, and sometimes not.
*/
2016-12-21 00:17:48 +03:00
static int _lvconvert_to_pool(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume *process_single_lv,
2016-12-21 00:17:48 +03:00
int to_thinpool,
int to_cachepool,
struct dm_list *use_pvh)
{
2016-12-21 00:17:48 +03:00
struct volume_group *vg = lv->vg;
struct logical_volume *metadata_lv = NULL; /* existing or created */
struct logical_volume *data_lv; /* lv arg renamed */
struct logical_volume *pool_lv; /* new lv created here */
const char *pool_metadata_name; /* user-specified lv name */
const char *pool_name; /* name of original lv arg */
char meta_name[NAME_LEN]; /* generated sub lv name */
char data_name[NAME_LEN]; /* generated sub lv name */
char converted_names[3*NAME_LEN]; /* preserve names of converted lv */
2016-12-21 00:17:48 +03:00
struct segment_type *pool_segtype; /* thinpool or cachepool */
struct lv_segment *seg;
2016-12-21 00:17:48 +03:00
unsigned int target_attr = ~0;
unsigned int activate_pool;
unsigned int zero_metadata;
uint64_t meta_size;
uint32_t meta_extents;
uint32_t chunk_size;
int chunk_calc;
cache_metadata_format_t cache_metadata_format;
cache_mode_t cache_mode;
const char *policy_name;
struct dm_config_tree *policy_settings = NULL;
int pool_metadata_spare;
thin_discards_t discards;
thin_zero_t zero_new_blocks;
2016-12-21 00:17:48 +03:00
int r = 0;
/* for handling lvmlockd cases */
2015-03-05 23:00:44 +03:00
char *lockd_data_args = NULL;
char *lockd_meta_args = NULL;
char *lockd_data_name = NULL;
char *lockd_meta_name = NULL;
struct id lockd_data_id;
struct id lockd_meta_id;
const char *str_seg_type = to_cachepool ? SEG_TYPE_NAME_CACHE_POOL : SEG_TYPE_NAME_THIN_POOL;
2015-03-05 23:00:44 +03:00
2016-12-21 00:17:48 +03:00
if (lv_is_thin_pool(lv) || lv_is_cache_pool(lv)) {
log_error(INTERNAL_ERROR "LV %s is already a pool.", display_lvname(lv));
return 0;
}
pool_segtype = get_segtype_from_string(cmd, str_seg_type);
if (!pool_segtype ||
!pool_segtype->ops->target_present(cmd, NULL, &target_attr)) {
log_error("%s: Required device-mapper target(s) not detected in your kernel.",
str_seg_type);
return 0;
}
2016-12-21 00:17:48 +03:00
/* Allow to have only thinpool active and restore it's active state. */
activate_pool = to_thinpool && lv_is_active(lv);
2016-12-21 00:17:48 +03:00
/* Wipe metadata_lv by default, but allow skipping this for cache pools. */
zero_metadata = (to_cachepool) ? arg_int_value(cmd, zero_ARG, 1) : 1;
2016-12-21 00:17:48 +03:00
/* An existing LV needs to have its lock freed once it becomes a data LV. */
if (is_lockd_type(vg->lock_type) && lv->lock_args) {
lockd_data_args = dm_pool_strdup(cmd->mem, lv->lock_args);
lockd_data_name = dm_pool_strdup(cmd->mem, lv->name);
memcpy(&lockd_data_id, &lv->lvid.id[1], sizeof(struct id));
}
/*
* The internal LV names for pool data/meta LVs.
*/
if ((dm_snprintf(meta_name, sizeof(meta_name), "%s%s", lv->name, to_cachepool ? "_cmeta" : "_tmeta") < 0) ||
(dm_snprintf(data_name, sizeof(data_name), "%s%s", lv->name, to_cachepool ? "_cdata" : "_tdata") < 0)) {
log_error("Failed to create internal lv names, pool name is too long.");
return 0;
}
/* If LV is inactive here, ensure it's not active elsewhere. */
if (!lockd_lv(cmd, lv, "ex", 0))
return 0;
/*
2016-12-21 00:17:48 +03:00
* If an existing LV is to be used as the metadata LV,
* verify that it's in a usable state. These checks are
* not done by command def rules because this LV is not
* processed by process_each_lv.
*/
2016-12-21 00:17:48 +03:00
if ((pool_metadata_name = arg_str_value(cmd, poolmetadata_ARG, NULL))) {
if (!validate_lvname_param(cmd, &vg->name, &pool_metadata_name)) {
log_error("Metadata LV %s not found.", pool_metadata_name);
return 0;
}
2016-12-21 00:17:48 +03:00
if (!(metadata_lv = find_lv(vg, pool_metadata_name))) {
log_error("Unknown pool metadata LV %s.", pool_metadata_name);
return 0;
}
2015-03-05 23:00:44 +03:00
/* An existing LV needs to have its lock freed once it becomes a meta LV. */
if (is_lockd_type(vg->lock_type) && metadata_lv->lock_args) {
lockd_meta_args = dm_pool_strdup(cmd->mem, metadata_lv->lock_args);
lockd_meta_name = dm_pool_strdup(cmd->mem, metadata_lv->name);
memcpy(&lockd_meta_id, &metadata_lv->lvid.id[1], sizeof(struct id));
}
2016-12-21 00:17:48 +03:00
if (metadata_lv == lv) {
log_error("Can't use same LV for pool data and metadata LV %s.",
display_lvname(metadata_lv));
return 0;
}
if (metadata_lv == process_single_lv) {
log_error("Use a different LV for pool metadata %s.",
display_lvname(metadata_lv));
return 0;
}
if (!lv_is_visible(metadata_lv)) {
log_error("Can't convert internal LV %s.",
display_lvname(metadata_lv));
return 0;
}
if (lv_is_locked(metadata_lv)) {
log_error("Can't convert locked LV %s.",
display_lvname(metadata_lv));
return 0;
}
if (lv_is_mirror(metadata_lv)) {
log_error("Mirror logical volumes cannot be used for pool metadata.");
log_print_unless_silent("Try \"%s\" segment type instead.", SEG_TYPE_NAME_RAID1);
return 0;
}
/* FIXME Tidy up all these type restrictions. */
if (lv_is_cache_type(metadata_lv) ||
lv_is_thin_type(metadata_lv) ||
lv_is_cow(metadata_lv) || lv_is_merging_cow(metadata_lv) ||
lv_is_origin(metadata_lv) || lv_is_merging_origin(metadata_lv) ||
lv_is_external_origin(metadata_lv) ||
lv_is_virtual(metadata_lv)) {
log_error("Pool metadata LV %s is of an unsupported type.",
display_lvname(metadata_lv));
return 0;
}
/* If LV is inactive here, ensure it's not active elsewhere. */
if (!lockd_lv(cmd, metadata_lv, "ex", 0))
return 0;
}
if (!get_pool_params(cmd, pool_segtype,
&meta_size, &pool_metadata_spare,
&chunk_size, &discards, &zero_new_blocks))
goto_bad;
if (to_cachepool &&
!get_cache_params(cmd, &chunk_size, &cache_metadata_format, &cache_mode, &policy_name, &policy_settings))
goto_bad;
if (metadata_lv)
2016-12-21 00:17:48 +03:00
meta_extents = metadata_lv->le_count;
else if (meta_size)
meta_extents = extents_from_size(cmd, meta_size, vg->extent_size);
else
meta_extents = 0; /* A default will be chosen by the "update" function. */
2016-12-21 00:17:48 +03:00
/*
* Validate and/or choose defaults for meta_extents and chunk_size,
* this involves some complicated calculations.
*/
2016-12-21 00:17:48 +03:00
if (to_cachepool) {
if (!update_cache_pool_params(cmd, vg->profile, vg->extent_size,
pool_segtype, target_attr,
lv->le_count,
2016-12-21 00:17:48 +03:00
&meta_extents,
&chunk_calc,
&chunk_size))
goto_bad;
2016-12-21 00:17:48 +03:00
} else {
if (!update_thin_pool_params(cmd, vg->profile, vg->extent_size,
pool_segtype, target_attr,
lv->le_count,
2016-12-21 00:17:48 +03:00
&meta_extents,
&chunk_calc,
&chunk_size,
&discards, &zero_new_blocks))
goto_bad;
2016-12-21 00:17:48 +03:00
}
2016-12-21 00:17:48 +03:00
if (metadata_lv && (meta_extents > metadata_lv->le_count)) {
log_error("Pool metadata LV %s is too small (%u extents) for required metadata (%u extents).",
display_lvname(metadata_lv), metadata_lv->le_count, meta_extents);
goto bad;
2016-12-21 00:17:48 +03:00
}
2016-12-21 00:17:48 +03:00
log_verbose("Pool metadata extents %u chunk_size %u", meta_extents, chunk_size);
(void) dm_snprintf(converted_names, sizeof(converted_names), "%s%s%s",
display_lvname(lv),
metadata_lv ? " and " : "",
metadata_lv ? display_lvname(metadata_lv) : "");
2016-12-21 00:17:48 +03:00
/*
* Verify that user wants to use these LVs.
*/
log_warn("WARNING: Converting %s to %s pool's data%s %s metadata wiping.",
converted_names,
2016-12-21 00:17:48 +03:00
to_cachepool ? "cache" : "thin",
metadata_lv ? " and metadata volumes" : " volume",
zero_metadata ? "with" : "WITHOUT");
2016-12-21 00:17:48 +03:00
if (zero_metadata)
log_warn("THIS WILL DESTROY CONTENT OF LOGICAL VOLUME (filesystem etc.)");
else if (to_cachepool)
log_warn("WARNING: Using mismatched cache pool metadata MAY DESTROY YOUR DATA!");
2016-12-21 00:17:48 +03:00
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Do you really want to convert %s? [y/n]: ",
converted_names) == 'n') {
2016-12-21 00:17:48 +03:00
log_error("Conversion aborted.");
goto bad;
}
2016-12-21 00:17:48 +03:00
/*
* If a new metadata LV needs to be created, collect the settings for
* the new LV and create it.
*
* If an existing LV is used for metadata, deactivate/activate/wipe it.
*/
if (!metadata_lv) {
2016-12-21 00:17:48 +03:00
uint32_t meta_stripes;
uint32_t meta_stripe_size;
uint32_t meta_readahead;
alloc_policy_t meta_alloc;
unsigned meta_stripes_supplied;
unsigned meta_stripe_size_supplied;
if (!get_stripe_params(cmd, get_segtype_from_string(cmd, SEG_TYPE_NAME_STRIPED),
2016-12-21 00:17:48 +03:00
&meta_stripes,
&meta_stripe_size,
&meta_stripes_supplied,
&meta_stripe_size_supplied))
goto_bad;
2016-12-21 00:17:48 +03:00
meta_readahead = arg_uint_value(cmd, readahead_ARG, cmd->default_settings.read_ahead);
meta_alloc = (alloc_policy_t) arg_uint_value(cmd, alloc_ARG, ALLOC_INHERIT);
if (!archive(vg))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!(metadata_lv = alloc_pool_metadata(lv,
meta_name,
meta_readahead,
meta_stripes,
meta_stripe_size,
meta_extents,
meta_alloc,
use_pvh)))
goto_bad;
} else {
if (!deactivate_lv(cmd, metadata_lv)) {
log_error("Aborting. Failed to deactivate %s.",
display_lvname(metadata_lv));
goto bad;
}
if (!archive(vg))
goto_bad;
if (zero_metadata) {
metadata_lv->status |= LV_TEMPORARY;
if (!activate_lv_excl_local(cmd, metadata_lv)) {
log_error("Aborting. Failed to activate metadata lv.");
goto bad;
}
metadata_lv->status &= ~LV_TEMPORARY;
if (!wipe_lv(metadata_lv, (struct wipe_params) { .do_zero = 1 })) {
log_error("Aborting. Failed to wipe metadata lv.");
goto bad;
}
}
}
2016-12-21 00:17:48 +03:00
/*
* Deactivate the data LV and metadata LV.
* We are changing target type, so deactivate first.
*/
if (!deactivate_lv(cmd, metadata_lv)) {
log_error("Aborting. Failed to deactivate metadata lv. "
"Manual intervention required.");
goto bad;
}
2016-12-21 00:17:48 +03:00
if (!deactivate_lv(cmd, lv)) {
log_error("Aborting. Failed to deactivate logical volume %s.",
2016-12-21 00:17:48 +03:00
display_lvname(lv));
goto bad;
}
/*
2016-12-21 00:17:48 +03:00
* When the LV referenced by the original function arg "lv"
* is renamed, it is then referenced as "data_lv".
*
* pool_name pool name taken from lv arg
* data_name sub lv name, generated
* meta_name sub lv name, generated
*
* pool_lv new lv for pool object, created here
* data_lv sub lv, was lv arg, now renamed
* metadata_lv sub lv, existing or created here
*/
data_lv = lv;
pool_name = lv->name; /* Use original LV name for pool name */
/*
* Rename the original LV arg to the internal data LV naming scheme.
*
* Since we wish to have underlaying devs to match _[ct]data
* rename data LV to match pool LV subtree first,
* also checks for visible LV.
2016-12-21 00:17:48 +03:00
*
* FIXME: any more types prohibited here?
*/
2016-12-21 00:17:48 +03:00
if (!lv_rename_update(cmd, data_lv, data_name, 0))
goto_bad;
2016-12-21 00:17:48 +03:00
/*
* Create LV structures for the new pool LV object,
* and connect it to the data/meta LVs.
*/
if (!(pool_lv = lv_create_empty(pool_name, NULL,
(to_cachepool ? CACHE_POOL : THIN_POOL) | VISIBLE_LV | LVM_READ | LVM_WRITE,
ALLOC_INHERIT, vg))) {
log_error("Creation of pool LV failed.");
goto bad;
}
/* Allocate a new pool segment */
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
if (!(seg = alloc_lv_segment(pool_segtype, pool_lv, 0, data_lv->le_count, 0,
pool_lv->status, 0, NULL, 1,
lvconvert: add infrastructure for RaidLV reshaping support In order to support striped raid5/6/10 LV reshaping (change of LV type, stripesize or number of legs), this patch introduces infrastructure prerequisites to be used by raid_manip.c extensions in followup patches. This base is needed for allocation of out-of-place reshape space required by the MD raid personalities to avoid writing over data in-place when reading off the current RAID layout or number of legs and writing out the new layout or to a different number of legs (i.e. restripe) Changes: - add members reshape_len to 'struct lv_segment' to store out-of-place reshape length per component rimage - add member data_copies to struct lv_segment to support more than 2 raid10 data copies - make alloc_lv_segment() aware of both reshape_len and data_copies - adjust all alloc_lv_segment() callers to the new API - add functions to retrieve the current data offset (needed for out-of-place reshaping space allocation) and the devices count from the kernel - make libdm deptree code aware of reshape_len - add LV flags for disk add/remove reshaping - support import/export of the new 'struct lv_segment' members - enhance lv_extend/_lv_reduce to cope with reshape_len - add seg_is_*/segtype_is_* macros related to reshaping - add target version check for reshaping - grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal - enhance libdm deptree code to support data_offset (out-of-place reshaping) and delta_disk (legs add/remove reshaping) target arguments Related: rhbz834579 Related: rhbz1191935 Related: rhbz1191978
2017-02-24 02:50:00 +03:00
data_lv->le_count, 0, 0, 0, 0, NULL)))
goto_bad;
/* Add the new segment to the layer LV */
dm_list_add(&pool_lv->segments, &seg->list);
pool_lv->le_count = data_lv->le_count;
pool_lv->size = data_lv->size;
if (!attach_pool_data_lv(seg, data_lv))
goto_bad;
2015-03-05 23:00:44 +03:00
/*
* Create a new lock for a thin pool LV. A cache pool LV has no lock.
2015-03-05 23:00:44 +03:00
* Locks are removed from existing LVs that are being converted to
* data and meta LVs (they are unlocked and deleted below.)
*/
if (is_lockd_type(vg->lock_type)) {
2016-12-21 00:17:48 +03:00
if (to_cachepool) {
2015-03-05 23:00:44 +03:00
data_lv->lock_args = NULL;
metadata_lv->lock_args = NULL;
} else {
data_lv->lock_args = NULL;
metadata_lv->lock_args = NULL;
if (!strcmp(vg->lock_type, "sanlock"))
pool_lv->lock_args = "pending";
else if (!strcmp(vg->lock_type, "dlm"))
pool_lv->lock_args = "dlm";
/* The lock_args will be set in vg_write(). */
2016-12-21 00:17:48 +03:00
}
}
/* Apply settings to the new pool seg */
2016-12-21 00:17:48 +03:00
if (to_cachepool) {
if (!cache_set_params(seg, chunk_size, cache_metadata_format, cache_mode, policy_name, policy_settings))
goto_bad;
2016-12-21 00:17:48 +03:00
} else {
seg->transaction_id = 0;
seg->chunk_size = chunk_size;
seg->discards = discards;
seg->zero_new_blocks = zero_new_blocks;
2016-12-21 00:17:48 +03:00
}
2016-12-21 00:17:48 +03:00
/*
* Rename deactivated metadata LV to have _tmeta suffix.
* Implicit checks if metadata_lv is visible.
*/
if (pool_metadata_name &&
!lv_rename_update(cmd, metadata_lv, meta_name, 0))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!attach_pool_metadata_lv(seg, metadata_lv))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!handle_pool_metadata_spare(vg,
metadata_lv->le_count,
use_pvh, pool_metadata_spare))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!vg_write(vg) || !vg_commit(vg))
goto_bad;
2016-12-21 00:17:48 +03:00
if (activate_pool && !lockd_lv(cmd, pool_lv, "ex", LDLV_PERSISTENT)) {
log_error("Failed to lock pool LV %s.", display_lvname(pool_lv));
goto out;
}
2016-12-21 00:17:48 +03:00
if (activate_pool &&
!activate_lv_excl(cmd, pool_lv)) {
log_error("Failed to activate pool logical volume %s.",
display_lvname(pool_lv));
/* Deactivate subvolumes */
if (!deactivate_lv(cmd, seg_lv(seg, 0)))
log_error("Failed to deactivate pool data logical volume %s.",
display_lvname(seg_lv(seg, 0)));
if (!deactivate_lv(cmd, seg->metadata_lv))
log_error("Failed to deactivate pool metadata logical volume %s.",
display_lvname(seg->metadata_lv));
goto out;
}
2016-12-21 00:17:48 +03:00
r = 1;
2016-12-21 00:17:48 +03:00
out:
backup(vg);
2016-12-21 00:17:48 +03:00
if (r)
log_print_unless_silent("Converted %s to %s pool.",
converted_names, to_cachepool ? "cache" : "thin");
2016-12-21 00:17:48 +03:00
/*
* Unlock and free the locks from existing LVs that became pool data
* and meta LVs.
*/
if (lockd_data_name) {
if (!lockd_lv_name(cmd, vg, lockd_data_name, &lockd_data_id, lockd_data_args, "un", LDLV_PERSISTENT))
log_error("Failed to unlock pool data LV %s/%s", vg->name, lockd_data_name);
lockd_free_lv(cmd, vg, lockd_data_name, &lockd_data_id, lockd_data_args);
}
2016-12-21 00:17:48 +03:00
if (lockd_meta_name) {
if (!lockd_lv_name(cmd, vg, lockd_meta_name, &lockd_meta_id, lockd_meta_args, "un", LDLV_PERSISTENT))
log_error("Failed to unlock pool metadata LV %s/%s", vg->name, lockd_meta_name);
lockd_free_lv(cmd, vg, lockd_meta_name, &lockd_meta_id, lockd_meta_args);
}
bad:
if (policy_settings)
dm_config_destroy(policy_settings);
2016-12-21 00:17:48 +03:00
return r;
#if 0
revert_new_lv:
/* TBD */
if (!pool_metadata_lv_name) {
if (!deactivate_lv(cmd, metadata_lv)) {
log_error("Failed to deactivate metadata lv.");
return 0;
}
if (!lv_remove(metadata_lv) || !vg_write(vg) || !vg_commit(vg))
log_error("Manual intervention may be required to remove "
"abandoned LV(s) before retrying.");
else
backup(vg);
}
return 0;
2016-12-21 00:17:48 +03:00
#endif
}
2016-12-21 00:17:48 +03:00
static int _lvconvert_to_cache_vol(struct cmd_context *cmd,
struct logical_volume *lv,
struct logical_volume *cachepool_lv)
{
2016-12-21 00:17:48 +03:00
struct logical_volume *cache_lv;
2017-03-09 18:20:44 +03:00
uint32_t chunk_size = 0;
cache_metadata_format_t cache_metadata_format;
cache_mode_t cache_mode;
const char *policy_name;
2016-12-21 00:17:48 +03:00
struct dm_config_tree *policy_settings = NULL;
int r = 0;
/* If LV is inactive here, ensure it's not active elsewhere. */
if (!lockd_lv(cmd, lv, "ex", 0))
return_0;
2016-12-21 00:17:48 +03:00
if (!validate_lv_cache_create_pool(cachepool_lv))
return_0;
if (!get_cache_params(cmd, &chunk_size, &cache_metadata_format, &cache_mode, &policy_name, &policy_settings))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!archive(lv->vg))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!(cache_lv = lv_cache_create(cachepool_lv, lv)))
goto_bad;
if (!cache_set_params(first_seg(cache_lv), chunk_size, cache_metadata_format, cache_mode, policy_name, policy_settings))
goto_bad;
2016-12-21 00:17:48 +03:00
if (!lv_update_and_reload(cache_lv))
goto_bad;
2016-12-21 00:17:48 +03:00
log_print_unless_silent("Logical volume %s is now cached.",
display_lvname(cache_lv));
r = 1;
bad:
if (policy_settings)
dm_config_destroy(policy_settings);
return r;
}
static struct convert_poll_id_list* _convert_poll_id_list_create(struct cmd_context *cmd,
const struct logical_volume *lv)
{
struct convert_poll_id_list *idl = (struct convert_poll_id_list *) dm_pool_alloc(cmd->mem, sizeof(struct convert_poll_id_list));
if (!idl) {
log_error("Convert poll ID list allocation failed.");
return NULL;
}
if (!(idl->id = _create_id(cmd, lv->vg->name, lv->name, lv->lvid.s))) {
dm_pool_free(cmd->mem, idl);
return_NULL;
}
idl->is_merging_origin = lv_is_merging_origin(lv);
idl->is_merging_origin_thin = idl->is_merging_origin && seg_is_thin_volume(find_snapshot(lv));
return idl;
}
/*
* Data/results accumulated during processing.
*/
struct lvconvert_result {
int need_polling;
struct dm_list poll_idls;
};
/*
* repair-related lvconvert utilities
*/
static int _lvconvert_repair_pvs_mirror(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle,
struct dm_list *use_pvh)
{
struct lvconvert_result *lr = (struct lvconvert_result *) handle->custom_handle;
struct lvconvert_params lp = { 0 };
struct convert_poll_id_list *idl;
struct lvinfo info;
int ret;
/*
* FIXME: temporary use of lp because _lvconvert_mirrors_repair()
* and _aux() still use lp fields everywhere.
* Migrate them away from using lp (for the most part just use
* local variables, and check arg_values directly).
*/
/*
* Fill in any lp fields here that this fn expects to be set before
* it's called. It's hard to tell what the old code expects in lp
* for repair; it doesn't take the stripes option, but it seems to
* expect lp.stripes to be set to 1.
*/
lp.alloc = (alloc_policy_t) arg_uint_value(cmd, alloc_ARG, ALLOC_INHERIT);
lp.stripes = 1;
ret = _lvconvert_mirrors_repair(cmd, lv, &lp, use_pvh);
if (lp.need_polling) {
if (!lv_info(cmd, lv, 0, &info, 0, 0) || !info.exists)
log_print_unless_silent("Conversion starts after activation.");
else {
if (!(idl = _convert_poll_id_list_create(cmd, lv)))
return 0;
dm_list_add(&lr->poll_idls, &idl->list);
}
lr->need_polling = 1;
}
return ret;
}
static void _lvconvert_repair_pvs_raid_ask(struct cmd_context *cmd, int *do_it)
{
const char *dev_policy;
*do_it = 1;
if (arg_is_set(cmd, usepolicies_ARG)) {
dev_policy = find_config_tree_str(cmd, activation_raid_fault_policy_CFG, NULL);
if (!strcmp(dev_policy, "allocate") ||
!strcmp(dev_policy, "replace"))
return;
/* else if (!strcmp(dev_policy, "anything_else")) -- no replace */
*do_it = 0;
return;
}
if (!arg_count(cmd, yes_ARG) &&
yes_no_prompt("Attempt to replace failed RAID images "
"(requires full device resync)? [y/n]: ") == 'n') {
*do_it = 0;
}
}
static int _lvconvert_repair_pvs_raid(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle,
struct dm_list *use_pvh)
{
struct dm_list *failed_pvs;
int do_it;
if (!lv_is_active_exclusive_locally(lv_lock_holder(lv))) {
log_error("%s must be active %sto perform this operation.",
display_lvname(lv),
vg_is_clustered(lv->vg) ?
"exclusive locally " : "");
return 0;
}
lv_check_transient(lv); /* TODO check this in lib for all commands? */
_lvconvert_repair_pvs_raid_ask(cmd, &do_it);
if (do_it) {
if (!(failed_pvs = _failed_pv_list(lv->vg)))
return_0;
if (!lv_raid_replace(lv, arg_count(cmd, force_ARG), failed_pvs, use_pvh)) {
log_error("Failed to replace faulty devices in %s.",
display_lvname(lv));
return 0;
}
log_print_unless_silent("Faulty devices in %s successfully replaced.",
display_lvname(lv));
return 1;
}
/* "warn" if policy not set to replace */
if (arg_is_set(cmd, usepolicies_ARG))
log_warn("Use 'lvconvert --repair %s' to replace "
"failed device.", display_lvname(lv));
return 1;
}
static int _lvconvert_repair_pvs(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
struct dm_list *failed_pvs;
struct dm_list *use_pvh;
int ret;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
} else
use_pvh = &lv->vg->pvs;
if (lv_is_raid(lv))
ret = _lvconvert_repair_pvs_raid(cmd, lv, handle, use_pvh);
else if (lv_is_mirror(lv))
ret = _lvconvert_repair_pvs_mirror(cmd, lv, handle, use_pvh);
else
ret = 0;
if (ret && arg_is_set(cmd, usepolicies_ARG)) {
if ((failed_pvs = _failed_pv_list(lv->vg)))
_remove_missing_empty_pv(lv->vg, failed_pvs);
}
return ret ? ECMD_PROCESSED : ECMD_FAILED;
}
static int _lvconvert_repair_cachepool_thinpool(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
int poolmetadataspare = arg_int_value(cmd, poolmetadataspare_ARG, DEFAULT_POOL_METADATA_SPARE);
struct dm_list *use_pvh;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
} else
use_pvh = &lv->vg->pvs;
if (lv_is_thin_pool(lv)) {
if (!_lvconvert_thin_pool_repair(cmd, lv, use_pvh, poolmetadataspare))
return_ECMD_FAILED;
} else /* cache */ {
if (!_lvconvert_cache_repair(cmd, lv, use_pvh, poolmetadataspare))
return_ECMD_FAILED;
}
return ECMD_PROCESSED;
}
static int _lvconvert_repair_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
if (lv_is_thin_pool(lv) ||
lv_is_cache(lv) ||
lv_is_cache_pool(lv))
return _lvconvert_repair_cachepool_thinpool(cmd, lv, handle);
if (lv_is_raid(lv) || lv_is_mirror(lv))
return _lvconvert_repair_pvs(cmd, lv, handle);
log_error("Unsupported volume type for repair of volume %s.",
display_lvname(lv));
return ECMD_FAILED;
}
/*
* FIXME: add option --repair-pvs to call _lvconvert_repair_pvs() directly,
* and option --repair-thinpool to call _lvconvert_repair_thinpool().
* and option --repair-cache to call _lvconvert_repair_cache().
* and option --repair-cachepool to call _lvconvert_repair_cachepool().
*/
int lvconvert_repair_cmd(struct cmd_context *cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_result lr = { 0 };
struct convert_poll_id_list *idl;
int saved_ignore_suspended_devices;
int ret, poll_ret;
dm_list_init(&lr.poll_idls);
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lr;
saved_ignore_suspended_devices = ignore_suspended_devices();
init_ignore_suspended_devices(1);
cmd->handles_missing_pvs = 1;
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_repair_single);
init_ignore_suspended_devices(saved_ignore_suspended_devices);
if (lr.need_polling) {
dm_list_iterate_items(idl, &lr.poll_idls) {
poll_ret = _lvconvert_poll_by_id(cmd, idl->id,
arg_is_set(cmd, background_ARG), 0, 0);
if (poll_ret > ret)
ret = poll_ret;
}
}
destroy_processing_handle(cmd, handle);
return ret;
}
static int _lvconvert_replace_pv_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
struct arg_value_group_list *group;
const char *tmp_str;
struct dm_list *use_pvh;
struct dm_list *replace_pvh;
char **replace_pvs;
int replace_pv_count;
int i;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
} else
use_pvh = &lv->vg->pvs;
if (!(replace_pv_count = arg_count(cmd, replace_ARG)))
return_ECMD_FAILED;
if (!(replace_pvs = dm_pool_alloc(cmd->mem, sizeof(char *) * replace_pv_count)))
return_ECMD_FAILED;
i = 0;
dm_list_iterate_items(group, &cmd->arg_value_groups) {
if (!grouped_arg_is_set(group->arg_values, replace_ARG))
continue;
if (!(tmp_str = grouped_arg_str_value(group->arg_values, replace_ARG, NULL))) {
log_error("Failed to get '--replace' argument");
return_ECMD_FAILED;
}
if (!(replace_pvs[i++] = dm_pool_strdup(cmd->mem, tmp_str)))
return_ECMD_FAILED;
}
if (!(replace_pvh = create_pv_list(cmd->mem, lv->vg, replace_pv_count, replace_pvs, 0)))
return_ECMD_FAILED;
if (!lv_raid_replace(lv, arg_count(cmd, force_ARG), replace_pvh, use_pvh))
return_ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_replace_pv_cmd(struct cmd_context *cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_result lr = { 0 };
int ret;
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lr;
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_replace_pv_single);
destroy_processing_handle(cmd, handle);
return ret;
}
/*
* Merge a COW snapshot LV into its origin.
*/
static int _lvconvert_merge_snapshot_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct lvconvert_result *lr = (struct lvconvert_result *) handle->custom_handle;
struct logical_volume *lv_to_poll = NULL;
struct convert_poll_id_list *idl;
if (!_lvconvert_merge_old_snapshot(cmd, lv, &lv_to_poll))
return_ECMD_FAILED;
if (lv_to_poll) {
if (!(idl = _convert_poll_id_list_create(cmd, lv_to_poll)))
return_ECMD_FAILED;
dm_list_add(&lr->poll_idls, &idl->list);
lr->need_polling = 1;
}
return ECMD_PROCESSED;
}
int lvconvert_merge_snapshot_cmd(struct cmd_context *cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_result lr = { 0 };
struct convert_poll_id_list *idl;
int ret, poll_ret;
dm_list_init(&lr.poll_idls);
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lr;
ret = process_each_lv(cmd, cmd->position_argc, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_merge_snapshot_single);
if (lr.need_polling) {
dm_list_iterate_items(idl, &lr.poll_idls) {
poll_ret = _lvconvert_poll_by_id(cmd, idl->id,
arg_is_set(cmd, background_ARG), 1, 0);
if (poll_ret > ret)
ret = poll_ret;
}
}
destroy_processing_handle(cmd, handle);
return ret;
}
/*
* Separate a COW snapshot from its origin.
*
* lvconvert --splitsnapshot LV_snapshot
* lvconvert_split_cow_snapshot
*/
static int _lvconvert_split_snapshot_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
if (!_lvconvert_splitsnapshot(cmd, lv))
return_ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_split_snapshot_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_split_snapshot_single);
}
/*
* Combine two LVs that were once an origin/cow pair of LVs, were then
* separated with --splitsnapshot, and now with this command are combined again
* into the origin/cow pair.
*
* This is an obscure command that has little to no real uses.
*
* The command has unusual handling of position args. The first position arg
* will become the origin LV, and is not processed by process_each_lv. The
* second position arg will become the cow LV and is processed by
* process_each_lv.
*
* The single function can grab the origin LV from position_argv[0].
*
* begin with an ordinary LV foo:
* lvcreate -n foo -L 1 vg
*
* create a cow snapshot of foo named foosnap:
* lvcreate -s -L 1 -n foosnap vg/foo
*
* now, foo is an "origin LV" and foosnap is a "cow LV"
* (foosnap matches LV_snapshot aka lv_is_cow)
*
* split the two LVs apart:
* lvconvert --splitsnapshot vg/foosnap
*
* now, foo is *not* an origin LV and foosnap is *not* a cow LV
* (foosnap does not match LV_snapshot)
*
* now, combine the two LVs again:
* lvconvert --snapshot vg/foo vg/foosnap
*
* after this, foosnap will match LV_snapshot again.
*
* FIXME: when splitsnapshot is run, the previous cow LV should be
* flagged in the metadata somehow, and then that flag should be
* required here. As it is now, the first and second args
* (origin and cow) can be swapped and nothing catches it.
*/
static int _lvconvert_combine_split_snapshot_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
const char *origin_name = cmd->position_argv[0];
if (is_lockd_type(lv->vg->lock_type)) {
log_error("Unable to combine split snapshots in VG with lock_type %s", lv->vg->lock_type);
return ECMD_FAILED;
}
/* If origin_name includes VG name, the VG name is removed. */
if (!validate_lvname_param(cmd, &lv->vg->name, &origin_name))
return_ECMD_FAILED;
if (!_lvconvert_snapshot(cmd, lv, origin_name))
return_ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_combine_split_snapshot_cmd(struct cmd_context *cmd, int argc, char **argv)
{
const char *vgname = NULL;
const char *lvname1_orig;
const char *lvname2_orig;
const char *lvname1_split;
char *vglv;
int vglv_sz;
/*
* Hack to accomodate an old parsing quirk that allowed the
* the VG name to be attached to only the LV in arg pos 1,
* i.e. lvconvert -s vgname/lvname lvname
*
* The LV name in arg pos 2 is the one that is processed
* by process_each_lv(). If that LV has no VG name, but
* the first LV does, then copy the VG name from arg pos 1
* and add it to the LV name in arg pos 2 so that the
* standard arg parsing in process_each_lv will find it.
*
* This is the only instance in all commands.
*/
lvname1_orig = cmd->position_argv[0];
lvname2_orig = cmd->position_argv[1];
if (strchr(lvname1_orig, '/') && !strchr(lvname2_orig, '/') && !getenv("LVM_VG_NAME")) {
if (!(lvname1_split = dm_pool_strdup(cmd->mem, lvname1_orig)))
return_ECMD_FAILED;
if (!validate_lvname_param(cmd, &vgname, &lvname1_split))
return_ECMD_FAILED;
vglv_sz = strlen(vgname) + strlen(lvname2_orig) + 2;
if (!(vglv = dm_pool_alloc(cmd->mem, vglv_sz)) ||
dm_snprintf(vglv, vglv_sz, "%s/%s", vgname, lvname2_orig) < 0) {
log_error("vg/lv string alloc failed.");
return_ECMD_FAILED;
}
/* vglv is now vgname/lvname2 and replaces lvname2_orig */
cmd->position_argv[1] = vglv;
}
return process_each_lv(cmd, 1, cmd->position_argv + 1, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_combine_split_snapshot_single);
}
static int _lvconvert_start_poll_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct lvconvert_result *lr = (struct lvconvert_result *) handle->custom_handle;
struct convert_poll_id_list *idl;
if (!(idl = _convert_poll_id_list_create(cmd, lv)))
return_ECMD_FAILED;
dm_list_add(&lr->poll_idls, &idl->list);
lr->need_polling = 1;
return ECMD_PROCESSED;
}
int lvconvert_start_poll_cmd(struct cmd_context *cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_result lr = { 0 };
struct convert_poll_id_list *idl;
int saved_ignore_suspended_devices;
int ret, poll_ret;
dm_list_init(&lr.poll_idls);
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lr;
saved_ignore_suspended_devices = ignore_suspended_devices();
init_ignore_suspended_devices(1);
cmd->handles_missing_pvs = 1;
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_start_poll_single);
init_ignore_suspended_devices(saved_ignore_suspended_devices);
if (lr.need_polling) {
dm_list_iterate_items(idl, &lr.poll_idls) {
poll_ret = _lvconvert_poll_by_id(cmd, idl->id,
arg_is_set(cmd, background_ARG), 0, 0);
if (poll_ret > ret)
ret = poll_ret;
}
}
destroy_processing_handle(cmd, handle);
return ret;
}
static int _lvconvert_to_pool_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct dm_list *use_pvh = NULL;
int to_thinpool = 0;
int to_cachepool = 0;
switch (cmd->command->command_enum) {
case lvconvert_to_thinpool_CMD:
to_thinpool = 1;
break;
case lvconvert_to_cachepool_CMD:
to_cachepool = 1;
break;
default:
log_error(INTERNAL_ERROR "Invalid lvconvert pool command");
return 0;
};
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
} else
use_pvh = &lv->vg->pvs;
if (!_lvconvert_to_pool(cmd, lv, lv, to_thinpool, to_cachepool, use_pvh))
return_ECMD_FAILED;
return ECMD_PROCESSED;
}
/*
* The LV position arg is used as thinpool/cachepool data LV.
*/
int lvconvert_to_pool_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_to_pool_single);
}
static int _lvconvert_to_cache_vol_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct volume_group *vg = lv->vg;
struct logical_volume *cachepool_lv;
const char *cachepool_name;
if (!(cachepool_name = arg_str_value(cmd, cachepool_ARG, NULL)))
goto_out;
if (!validate_lvname_param(cmd, &vg->name, &cachepool_name))
goto_out;
if (!(cachepool_lv = find_lv(vg, cachepool_name))) {
log_error("Cache pool %s not found.", cachepool_name);
goto out;
}
/*
* If cachepool_lv is not yet a cache pool, convert it to one.
* If using an existing cache pool, wipe it.
*/
if (!lv_is_cache_pool(cachepool_lv)) {
int lvt_enum = get_lvt_enum(cachepool_lv);
struct lv_type *lvtype = get_lv_type(lvt_enum);
if (lvt_enum != striped_LVT && lvt_enum != linear_LVT && lvt_enum != raid_LVT) {
log_error("LV %s with type %s cannot be converted to a cache pool.",
display_lvname(cachepool_lv), lvtype ? lvtype->name : "unknown");
goto out;
}
if (cachepool_lv == lv) {
log_error("Use a different LV for cache pool LV and cache LV %s.",
display_lvname(cachepool_lv));
goto out;
}
if (!_lvconvert_to_pool(cmd, cachepool_lv, lv, 0, 1, &vg->pvs)) {
log_error("LV %s could not be converted to a cache pool.",
display_lvname(cachepool_lv));
goto out;
}
if (!(cachepool_lv = find_lv(vg, cachepool_name))) {
log_error("LV %s cannot be found.", display_lvname(cachepool_lv));
goto out;
}
if (!lv_is_cache_pool(cachepool_lv)) {
log_error("LV %s is not a cache pool.", display_lvname(cachepool_lv));
goto out;
}
} else {
if (!dm_list_empty(&cachepool_lv->segs_using_this_lv)) {
log_error("Cache pool %s is already in use.", cachepool_name);
goto out;
}
/* Note: requires rather deep know-how to skip zeroing */
if (!arg_is_set(cmd, zero_ARG)) {
if (!arg_is_set(cmd, yes_ARG) &&
yes_no_prompt("Do you want wipe existing metadata of cache pool %s? [y/n]: ",
display_lvname(cachepool_lv)) == 'n') {
log_error("Conversion aborted.");
log_error("To preserve cache metadata add option \"--zero n\".");
log_warn("WARNING: Reusing mismatched cache pool metadata MAY DESTROY YOUR DATA!");
goto out;
}
/* Wiping confirmed, go ahead */
if (!wipe_cache_pool(cachepool_lv))
goto_out;
} else if (arg_int_value(cmd, zero_ARG, 0)) {
if (!wipe_cache_pool(cachepool_lv))
goto_out;
} else {
log_warn("WARNING: Reusing cache pool metadata %s for volume caching.",
display_lvname(cachepool_lv));
}
}
/* When the lv arg is a thinpool, redirect command to data sub lv. */
if (lv_is_thin_pool(lv)) {
lv = seg_lv(first_seg(lv), 0);
log_verbose("Redirecting operation to data sub LV %s.", display_lvname(lv));
}
/* Convert lv to cache vol using cachepool_lv. */
if (!_lvconvert_to_cache_vol(cmd, lv, cachepool_lv))
goto_out;
return ECMD_PROCESSED;
out:
return ECMD_FAILED;
}
int lvconvert_to_cache_vol_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_to_cache_vol_single);
}
static int _lvconvert_to_thin_with_external_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct volume_group *vg = lv->vg;
struct logical_volume *thinpool_lv;
const char *thinpool_name;
if (!(thinpool_name = arg_str_value(cmd, thinpool_ARG, NULL)))
goto_out;
if (!validate_lvname_param(cmd, &vg->name, &thinpool_name))
goto_out;
if (!(thinpool_lv = find_lv(vg, thinpool_name))) {
log_error("Thin pool %s not found.", thinpool_name);
goto out;
}
/* If thinpool_lv is not yet a thin pool, convert it to one. */
if (!lv_is_thin_pool(thinpool_lv)) {
int lvt_enum = get_lvt_enum(thinpool_lv);
struct lv_type *lvtype = get_lv_type(lvt_enum);
if (lvt_enum != striped_LVT && lvt_enum != linear_LVT && lvt_enum != raid_LVT) {
log_error("LV %s with type %s cannot be converted to a thin pool.",
display_lvname(thinpool_lv), lvtype ? lvtype->name : "unknown");
goto out;
}
if (thinpool_lv == lv) {
log_error("Use a different LV for thin pool LV and thin LV %s.",
display_lvname(thinpool_lv));
goto out;
}
if (!_lvconvert_to_pool(cmd, thinpool_lv, lv, 1, 0, &vg->pvs)) {
log_error("LV %s could not be converted to a thin pool.",
display_lvname(thinpool_lv));
goto out;
}
if (!(thinpool_lv = find_lv(vg, thinpool_name))) {
log_error("LV %s cannot be found.", display_lvname(thinpool_lv));
goto out;
}
if (!lv_is_thin_pool(thinpool_lv)) {
log_error("LV %s is not a thin pool.", display_lvname(thinpool_lv));
goto out;
}
}
/* If lv is a cache volume, all data must be flushed. */
if (lv_is_cache(lv)) {
const struct lv_segment *pool_seg = first_seg(first_seg(lv)->pool_lv);
int is_clean;
if (pool_seg->cache_mode != CACHE_MODE_WRITETHROUGH) {
log_error("Cannot convert cache volume %s with %s cache mode to external origin.",
display_lvname(lv), get_cache_mode_name(pool_seg));
log_error("To proceed, run 'lvchange --cachemode writethrough %s'.",
display_lvname(lv));
goto out;
}
if (!lv_cache_wait_for_clean(lv, &is_clean))
goto_out;
if (!is_clean) {
log_error("Cache %s is not clean, refusing to convert to external origin.",
display_lvname(lv));
goto out;
}
}
/* Convert lv to thin with external origin using thinpool_lv. */
if (!_lvconvert_to_thin_with_external(cmd, lv, thinpool_lv))
goto_out;
return ECMD_PROCESSED;
out:
return ECMD_FAILED;
}
int lvconvert_to_thin_with_external_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_to_thin_with_external_single);
}
static int _lvconvert_swap_pool_metadata_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct volume_group *vg = lv->vg;
struct logical_volume *metadata_lv;
const char *metadata_name;
if (is_lockd_type(lv->vg->lock_type)) {
/* FIXME: need to swap locks betwen LVs? */
log_error("Unable to swap pool metadata in VG with lock_type %s", lv->vg->lock_type);
goto out;
}
if (!(metadata_name = arg_str_value(cmd, poolmetadata_ARG, NULL)))
goto_out;
if (!validate_lvname_param(cmd, &vg->name, &metadata_name))
goto_out;
if (!(metadata_lv = find_lv(vg, metadata_name))) {
log_error("Metadata LV %s not found.", metadata_name);
goto out;
}
if (metadata_lv == lv) {
log_error("Can't use same LV for pool data and metadata LV %s.",
display_lvname(metadata_lv));
goto out;
}
if (!_lvconvert_swap_pool_metadata(cmd, lv, metadata_lv))
goto_out;
return ECMD_PROCESSED;
out:
return ECMD_FAILED;
}
int lvconvert_swap_pool_metadata_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_swap_pool_metadata_single);
}
static int _lvconvert_to_pool_or_swap_metadata_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct dm_list *use_pvh = NULL;
int to_thinpool = 0;
int to_cachepool = 0;
switch (cmd->command->command_enum) {
case lvconvert_to_thinpool_or_swap_metadata_CMD:
to_thinpool = 1;
break;
case lvconvert_to_cachepool_or_swap_metadata_CMD:
to_cachepool = 1;
break;
default:
log_error(INTERNAL_ERROR "Invalid lvconvert pool command");
return 0;
};
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
} else
use_pvh = &lv->vg->pvs;
/*
* We can finally determine if this command is supposed to create
* a pool or swap the metadata in an existing pool.
*
* This allows the ambiguous command:
* 'lvconvert --thinpool LV1 --poolmetadata LV2' to mean either:
* 1. convert LV2 to a pool using the specified meta LV2
* 2. swap the meta lv in LV1 with LV2
*
* In case 2, the poolmetadata option is required, but in case 1
* it is optional. So, the command def is not able to validate
* the required/optional option, and we have to check here
* for missing poolmetadata in case 2.
*/
if (lv_is_pool(lv)) {
if (!arg_is_set(cmd, poolmetadata_ARG)) {
log_error("The --poolmetadata option is required to swap metadata.");
return ECMD_FAILED;
}
return _lvconvert_swap_pool_metadata_single(cmd, lv, handle);
}
if (!_lvconvert_to_pool(cmd, lv, lv, to_thinpool, to_cachepool, use_pvh))
return_ECMD_FAILED;
return ECMD_PROCESSED;
}
/*
* In the command variants with no position LV arg, the LV arg is taken from
* the --thinpool/--cachepool arg, and the position args are modified to match
* the standard command form.
*/
int lvconvert_to_pool_or_swap_metadata_cmd(struct cmd_context *cmd, int argc, char **argv)
{
char *pool_data_name;
int i, p;
switch (cmd->command->command_enum) {
case lvconvert_to_thinpool_or_swap_metadata_CMD:
pool_data_name = (char *)arg_str_value(cmd, thinpool_ARG, NULL);
break;
case lvconvert_to_cachepool_or_swap_metadata_CMD:
pool_data_name = (char *)arg_str_value(cmd, cachepool_ARG, NULL);
break;
default:
log_error(INTERNAL_ERROR "Unknown pool conversion.");
return 0;
};
/* Make the LV the first position arg. */
p = cmd->position_argc;
for (i = 0; i < cmd->position_argc; i++)
cmd->position_argv[p] = cmd->position_argv[p-1];
cmd->position_argv[0] = pool_data_name;
cmd->position_argc++;
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_to_pool_or_swap_metadata_single);
}
static int _lvconvert_merge_thin_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
if (!_lvconvert_merge_thin_snapshot(cmd, lv))
return ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_merge_thin_cmd(struct cmd_context *cmd, int argc, char **argv)
{
return process_each_lv(cmd, cmd->position_argc, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_merge_thin_single);
}
static int _lvconvert_split_cachepool_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
struct logical_volume *cache_lv = NULL;
struct logical_volume *cachepool_lv = NULL;
struct lv_segment *seg;
int ret;
if (lv_is_cache(lv)) {
cache_lv = lv;
cachepool_lv = first_seg(cache_lv)->pool_lv;
} else if (lv_is_cache_pool(lv)) {
cachepool_lv = lv;
if ((dm_list_size(&cachepool_lv->segs_using_this_lv) == 1) &&
(seg = get_only_segment_using_this_lv(cachepool_lv)) &&
seg_is_cache(seg))
cache_lv = seg->lv;
} else if (lv_is_thin_pool(lv)) {
cache_lv = seg_lv(first_seg(lv), 0); /* cached _tdata */
cachepool_lv = first_seg(cache_lv)->pool_lv;
}
if (!cache_lv) {
log_error("Cannot find cache LV from %s.", display_lvname(lv));
return ECMD_FAILED;
}
if (!cachepool_lv) {
log_error("Cannot find cache pool LV from %s.", display_lvname(lv));
return ECMD_FAILED;
}
/* If LV is inactive here, ensure it's not active elsewhere. */
if (!lockd_lv(cmd, cache_lv, "ex", 0))
return_0;
switch (cmd->command->command_enum) {
case lvconvert_split_and_keep_cachepool_CMD:
ret = _lvconvert_split_and_keep_cachepool(cmd, cache_lv, cachepool_lv);
break;
case lvconvert_split_and_remove_cachepool_CMD:
ret = _lvconvert_split_and_remove_cachepool(cmd, cache_lv, cachepool_lv);
break;
default:
log_error(INTERNAL_ERROR "Unknown cache pool split.");
ret = 0;
}
if (!ret)
return ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_split_cachepool_cmd(struct cmd_context *cmd, int argc, char **argv)
{
if (cmd->command->command_enum == lvconvert_split_and_remove_cachepool_CMD) {
cmd->handles_missing_pvs = 1;
cmd->partial_activation = 1;
}
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, NULL, &_lvconvert_split_cachepool_single);
}
static int _lvconvert_raid_types_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
struct lvconvert_params *lp = (struct lvconvert_params *) handle->custom_handle;
struct dm_list *use_pvh;
struct convert_poll_id_list *idl;
struct lvinfo info;
int ret;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
lp->pv_count = cmd->position_argc - 1;
} else
use_pvh = &lv->vg->pvs;
lp->pvh = use_pvh;
lp->lv_to_poll = lv;
2016-12-21 00:17:48 +03:00
ret = _lvconvert_raid_types(cmd, lv, lp);
if (ret != ECMD_PROCESSED)
return_ECMD_FAILED;
if (lp->need_polling) {
/* _lvconvert() call may alter the reference in lp->lv_to_poll */
if (!lv_info(cmd, lp->lv_to_poll, 0, &info, 0, 0) || !info.exists)
log_print_unless_silent("Conversion starts after activation.");
else {
if (!(idl = _convert_poll_id_list_create(cmd, lp->lv_to_poll)))
return_ECMD_FAILED;
dm_list_add(&lp->idls, &idl->list);
}
}
return ECMD_PROCESSED;
}
static int _lvconvert_raid_types_check(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle,
int lv_is_named_arg)
{
int lvt_enum = get_lvt_enum(lv);
struct lv_type *lvtype = get_lv_type(lvt_enum);
if (!lv_is_visible(lv)) {
if (!lv_is_cache_pool_metadata(lv) &&
!lv_is_cache_pool_data(lv) &&
!lv_is_thin_pool_metadata(lv) &&
!lv_is_thin_pool_data(lv) &&
!lv_is_used_cache_pool(lv) &&
!lv_is_mirrored(lv) &&
!lv_is_raid(lv))
goto fail_hidden;
}
/*
* FIXME: this validation could be done by command defs.
*
* Outside the standard linear/striped/mirror/raid LV
* types, cache is the only special LV type that is handled
* (the command is redirected to origin).
*/
switch (lvt_enum) {
case thin_LVT:
case thinpool_LVT:
case cachepool_LVT:
case snapshot_LVT:
log_error("Operation not permitted on LV %s type %s.",
display_lvname(lv), lvtype ? lvtype->name : "unknown");
return 0;
}
return 1;
fail_hidden:
log_error("Operation not permitted on hidden LV %s.", display_lvname(lv));
return 0;
}
int lvconvert_raid_types_cmd(struct cmd_context * cmd, int argc, char **argv)
{
int poll_ret, ret;
int saved_ignore_suspended_devices;
struct processing_handle *handle;
struct convert_poll_id_list *idl;
struct lvconvert_params lp = {
.conv_type = CONV_OTHER,
.target_attr = ~0,
.idls = DM_LIST_HEAD_INIT(lp.idls),
};
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lp;
if (!_read_params(cmd, &lp)) {
ret = EINVALID_CMD_LINE;
goto_out;
}
saved_ignore_suspended_devices = ignore_suspended_devices();
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, &_lvconvert_raid_types_check, &_lvconvert_raid_types_single);
init_ignore_suspended_devices(saved_ignore_suspended_devices);
dm_list_iterate_items(idl, &lp.idls) {
poll_ret = _lvconvert_poll_by_id(cmd, idl->id,
lp.wait_completion ? 0 : 1U,
idl->is_merging_origin,
idl->is_merging_origin_thin);
if (poll_ret > ret)
ret = poll_ret;
}
out:
destroy_processing_handle(cmd, handle);
return ret;
}
/*
* change mirror log
*/
static int _lvconvert_visible_check(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle,
int lv_is_named_arg)
{
if (!lv_is_visible(lv)) {
log_error("Operation not permitted on hidden LV %s.", display_lvname(lv));
return 0;
}
return 1;
}
static int _lvconvert_change_mirrorlog_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
struct lvconvert_params *lp = (struct lvconvert_params *) handle->custom_handle;
struct dm_list *use_pvh;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
lp->pv_count = cmd->position_argc - 1;
} else
use_pvh = &lv->vg->pvs;
lp->pvh = use_pvh;
2016-12-21 00:17:48 +03:00
/* FIXME: extract the mirrorlog functionality out of _lvconvert_raid_types()? */
return _lvconvert_raid_types(cmd, lv, lp);
}
int lvconvert_change_mirrorlog_cmd(struct cmd_context * cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_params lp = {
.conv_type = CONV_OTHER,
.target_attr = ~0,
.idls = DM_LIST_HEAD_INIT(lp.idls),
};
int ret;
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lp;
/* FIXME: extract the relevant bits of read_params and put here. */
if (!_read_params(cmd, &lp)) {
ret = EINVALID_CMD_LINE;
goto_out;
}
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, &_lvconvert_visible_check, &_lvconvert_change_mirrorlog_single);
out:
destroy_processing_handle(cmd, handle);
return ret;
}
static int _lvconvert_change_region_size_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
if (!lv_raid_change_region_size(lv, arg_is_set(cmd, yes_ARG), arg_count(cmd, force_ARG),
arg_int_value(cmd, regionsize_ARG, 0)))
return ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_change_region_size_cmd(struct cmd_context * cmd, int argc, char **argv)
{
return process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, &_lvconvert_visible_check, &_lvconvert_change_region_size_single);
}
/*
* split mirror images
*/
static int _lvconvert_split_mirror_images_single(struct cmd_context *cmd, struct logical_volume *lv,
struct processing_handle *handle)
{
struct lvconvert_params *lp = (struct lvconvert_params *) handle->custom_handle;
struct dm_list *use_pvh;
if (cmd->position_argc > 1) {
/* First pos arg is required LV, remaining are optional PVs. */
if (!(use_pvh = create_pv_list(cmd->mem, lv->vg, cmd->position_argc - 1, cmd->position_argv + 1, 0)))
return_ECMD_FAILED;
lp->pv_count = cmd->position_argc - 1;
} else
use_pvh = &lv->vg->pvs;
lp->pvh = use_pvh;
2016-12-21 00:17:48 +03:00
/* FIXME: extract the split functionality out of _lvconvert_raid_types()? */
return _lvconvert_raid_types(cmd, lv, lp);
}
int lvconvert_split_mirror_images_cmd(struct cmd_context * cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_params lp = {
.conv_type = CONV_OTHER,
.target_attr = ~0,
.idls = DM_LIST_HEAD_INIT(lp.idls),
};
int ret;
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lp;
/* FIXME: extract the relevant bits of read_params and put here. */
if (!_read_params(cmd, &lp)) {
ret = EINVALID_CMD_LINE;
goto_out;
}
/* FIXME: are there any hidden LVs that should be disallowed? */
ret = process_each_lv(cmd, 1, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_split_mirror_images_single);
out:
destroy_processing_handle(cmd, handle);
return ret;
}
/*
* merge mirror images
*
* Called from both lvconvert --mergemirrors and lvconvert --merge.
*/
static int _lvconvert_merge_mirror_images_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
if (!lv_raid_merge(lv))
return ECMD_FAILED;
return ECMD_PROCESSED;
}
int lvconvert_merge_mirror_images_cmd(struct cmd_context *cmd, int argc, char **argv)
{
/* arg can be a VG name, which is the standard option usage */
cmd->cname->flags &= ~GET_VGNAME_FROM_OPTIONS;
return process_each_lv(cmd, cmd->position_argc, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
NULL, &_lvconvert_visible_check, &_lvconvert_merge_mirror_images_single);
}
static int _lvconvert_merge_generic_single(struct cmd_context *cmd,
struct logical_volume *lv,
struct processing_handle *handle)
{
int ret;
if (lv_is_cow(lv))
ret = _lvconvert_merge_snapshot_single(cmd, lv, handle);
else if (lv_is_thin_volume(lv))
ret = _lvconvert_merge_thin_single(cmd, lv, handle);
else
ret = _lvconvert_merge_mirror_images_single(cmd, lv, handle);
return ret;
}
int lvconvert_merge_cmd(struct cmd_context *cmd, int argc, char **argv)
{
struct processing_handle *handle;
struct lvconvert_result lr = { 0 };
struct convert_poll_id_list *idl;
int ret, poll_ret;
dm_list_init(&lr.poll_idls);
if (!(handle = init_processing_handle(cmd, NULL))) {
log_error("Failed to initialize processing handle.");
return ECMD_FAILED;
}
handle->custom_handle = &lr;
cmd->cname->flags &= ~GET_VGNAME_FROM_OPTIONS;
ret = process_each_lv(cmd, cmd->position_argc, cmd->position_argv, NULL, NULL, READ_FOR_UPDATE,
handle, NULL, &_lvconvert_merge_generic_single);
/* polling is only used by merge_snapshot */
if (lr.need_polling) {
dm_list_iterate_items(idl, &lr.poll_idls) {
poll_ret = _lvconvert_poll_by_id(cmd, idl->id,
arg_is_set(cmd, background_ARG), 1, 0);
if (poll_ret > ret)
ret = poll_ret;
}
}
destroy_processing_handle(cmd, handle);
return ret;
}
/*
* All lvconvert command defs have their own function,
* so the generic function name is unused.
*/
int lvconvert(struct cmd_context *cmd, int argc, char **argv)
{
log_error(INTERNAL_ERROR "Missing function for command definition %d:%s.",
cmd->command->command_index, cmd->command->command_id);
return ECMD_FAILED;
}