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

891 lines
26 KiB
C
Raw Normal View History

/*
2004-03-30 23:35:44 +04:00
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
2016-05-24 02:57:48 +03:00
* Copyright (C) 2004-2016 Red Hat, Inc. All rights reserved.
*
2004-03-30 23:35:44 +04:00
* 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.
2004-03-30 23:35:44 +04:00
*
* You should have received a copy of the GNU Lesser General Public License
2004-03-30 23:35:44 +04:00
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
2002-11-18 17:04:08 +03:00
#include "lib.h"
#include <defaults.h>
2002-01-10 21:12:26 +03:00
#include "metadata.h"
2003-09-15 22:22:50 +04:00
#include "lv_alloc.h"
#include "pv_alloc.h"
2004-03-08 20:19:15 +03:00
#include "str_list.h"
2004-09-16 22:40:56 +04:00
#include "segtype.h"
/*
* Attempt to merge two adjacent segments.
2004-05-05 01:25:57 +04:00
* Currently only supports striped segments on AREA_PV.
2008-01-30 17:00:02 +03:00
* Returns success if successful, in which case 'first'
* gets adjusted to contain both areas.
*/
static int _merge(struct lv_segment *first, struct lv_segment *second)
{
2004-05-05 01:25:57 +04:00
if (!first || !second || first->segtype != second->segtype ||
!first->segtype->ops->merge_segments)
return 0;
2004-05-05 01:25:57 +04:00
return first->segtype->ops->merge_segments(first, second);
}
int lv_merge_segments(struct logical_volume *lv)
{
struct dm_list *segh, *t;
pvmove: Enable all-or-nothing (atomic) pvmoves pvmove can be used to move single LVs by name or multiple LVs that lie within the specified PV range (e.g. /dev/sdb1:0-1000). When moving more than one LV, the portions of those LVs that are in the range to be moved are added to a new temporary pvmove LV. The LVs then point to the range in the pvmove LV, rather than the PV range. Example 1: We have two LVs in this example. After they were created, the first LV was grown, yeilding two segments in LV1. So, there are two LVs with a total of three segments. Before pvmove: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- After pvmove inserts the temporary pvmove LV: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- Each of the affected LV segments now point to a range of blocks in the pvmove LV, which purposefully corresponds to the segments moved from the original LVs into the temporary pvmove LV. The current implementation goes on from here to mirror the temporary pvmove LV by segment. Further, as the pvmove LV is activated, only one of its segments is actually mirrored (i.e. "moving") at a time. The rest are either complete or not addressed yet. If the pvmove is aborted, those segments that are completed will remain on the destination and those that are not yet addressed or in the process of moving will stay on the source PV. Thus, it is possible to have a partially completed move - some LVs (or certain segments of LVs) on the source PV and some on the destination. Example 2: What 'example 1' might look if it was half-way through the move. --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | | ------------------------- source PV | | 256 - 511 | 512 - 767 | | ------------------------- | || ------------------------- dest PV | 000 - 255 | 256 - 511 | ------------------------- This update allows the user to specify that they would like the pvmove mirror created "by LV" rather than "by segment". That is, the pvmove LV becomes an image in an encapsulating mirror along with the allocated copy image. Example 3: A pvmove that is performed "by LV" rather than "by segment". --------- --------- | LV1s0 | | LV2s0 | --------- --------- | | ------------------------- pvmove0 | * LV-level mirror * | ------------------------- / \ pvmove_mimage0 / pvmove_mimage1 ------------------------- ------------------------- | seg 0 | seg 1 | | seg 0 | seg 1 | ------------------------- ------------------------- | | | | ------------------------- ------------------------- | 000 - 255 | 256 - 511 | | 000 - 255 | 256 - 511 | ------------------------- ------------------------- source PV dest PV The thing that differentiates a pvmove done in this way and a simple "up-convert" from linear to mirror is the preservation of the distinct segments. A normal up-convert would simply allocate the necessary space with no regard for segment boundaries. The pvmove operation must preserve the segments because they are the critical boundary between the segments of the LVs being moved. So, when the pvmove copy image is allocated, all corresponding segments must be allocated. The code that merges ajoining segments that are part of the same LV when the metadata is written must also be avoided in this case. This method of mirroring is unique enough to warrant its own definitional macro, MIRROR_BY_SEGMENTED_LV. This joins the two existing macros: MIRROR_BY_SEG (for original pvmove) and MIRROR_BY_LV (for user created mirrors). The advantages of performing pvmove in this way is that all of the LVs affected can be moved together. It is an all-or-nothing approach that leaves all LV segments on the source PV if the move is aborted. Additionally, a mirror log can be used (in the future) to provide tracking of progress; allowing the copy to continue where it left off in the event there is a deactivation.
2014-06-18 07:59:36 +04:00
struct lv_segment *seg, *current, *prev = NULL;
2014-06-20 14:41:20 +04:00
/*
* Don't interfere with pvmoves as they rely upon two LVs
* having a matching segment structure.
*/
if (lv_is_locked(lv) || lv_is_pvmove(lv))
return 1;
if (lv_is_mirror_image(lv) &&
pvmove: Enable all-or-nothing (atomic) pvmoves pvmove can be used to move single LVs by name or multiple LVs that lie within the specified PV range (e.g. /dev/sdb1:0-1000). When moving more than one LV, the portions of those LVs that are in the range to be moved are added to a new temporary pvmove LV. The LVs then point to the range in the pvmove LV, rather than the PV range. Example 1: We have two LVs in this example. After they were created, the first LV was grown, yeilding two segments in LV1. So, there are two LVs with a total of three segments. Before pvmove: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- After pvmove inserts the temporary pvmove LV: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- Each of the affected LV segments now point to a range of blocks in the pvmove LV, which purposefully corresponds to the segments moved from the original LVs into the temporary pvmove LV. The current implementation goes on from here to mirror the temporary pvmove LV by segment. Further, as the pvmove LV is activated, only one of its segments is actually mirrored (i.e. "moving") at a time. The rest are either complete or not addressed yet. If the pvmove is aborted, those segments that are completed will remain on the destination and those that are not yet addressed or in the process of moving will stay on the source PV. Thus, it is possible to have a partially completed move - some LVs (or certain segments of LVs) on the source PV and some on the destination. Example 2: What 'example 1' might look if it was half-way through the move. --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | | ------------------------- source PV | | 256 - 511 | 512 - 767 | | ------------------------- | || ------------------------- dest PV | 000 - 255 | 256 - 511 | ------------------------- This update allows the user to specify that they would like the pvmove mirror created "by LV" rather than "by segment". That is, the pvmove LV becomes an image in an encapsulating mirror along with the allocated copy image. Example 3: A pvmove that is performed "by LV" rather than "by segment". --------- --------- | LV1s0 | | LV2s0 | --------- --------- | | ------------------------- pvmove0 | * LV-level mirror * | ------------------------- / \ pvmove_mimage0 / pvmove_mimage1 ------------------------- ------------------------- | seg 0 | seg 1 | | seg 0 | seg 1 | ------------------------- ------------------------- | | | | ------------------------- ------------------------- | 000 - 255 | 256 - 511 | | 000 - 255 | 256 - 511 | ------------------------- ------------------------- source PV dest PV The thing that differentiates a pvmove done in this way and a simple "up-convert" from linear to mirror is the preservation of the distinct segments. A normal up-convert would simply allocate the necessary space with no regard for segment boundaries. The pvmove operation must preserve the segments because they are the critical boundary between the segments of the LVs being moved. So, when the pvmove copy image is allocated, all corresponding segments must be allocated. The code that merges ajoining segments that are part of the same LV when the metadata is written must also be avoided in this case. This method of mirroring is unique enough to warrant its own definitional macro, MIRROR_BY_SEGMENTED_LV. This joins the two existing macros: MIRROR_BY_SEG (for original pvmove) and MIRROR_BY_LV (for user created mirrors). The advantages of performing pvmove in this way is that all of the LVs affected can be moved together. It is an all-or-nothing approach that leaves all LV segments on the source PV if the move is aborted. Additionally, a mirror log can be used (in the future) to provide tracking of progress; allowing the copy to continue where it left off in the event there is a deactivation.
2014-06-18 07:59:36 +04:00
(seg = get_only_segment_using_this_lv(lv)) &&
(lv_is_locked(seg->lv) || lv_is_pvmove(seg->lv)))
pvmove: Enable all-or-nothing (atomic) pvmoves pvmove can be used to move single LVs by name or multiple LVs that lie within the specified PV range (e.g. /dev/sdb1:0-1000). When moving more than one LV, the portions of those LVs that are in the range to be moved are added to a new temporary pvmove LV. The LVs then point to the range in the pvmove LV, rather than the PV range. Example 1: We have two LVs in this example. After they were created, the first LV was grown, yeilding two segments in LV1. So, there are two LVs with a total of three segments. Before pvmove: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- After pvmove inserts the temporary pvmove LV: --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | ------------------------------------- PV | 000 - 255 | 256 - 511 | 512 - 767 | ------------------------------------- Each of the affected LV segments now point to a range of blocks in the pvmove LV, which purposefully corresponds to the segments moved from the original LVs into the temporary pvmove LV. The current implementation goes on from here to mirror the temporary pvmove LV by segment. Further, as the pvmove LV is activated, only one of its segments is actually mirrored (i.e. "moving") at a time. The rest are either complete or not addressed yet. If the pvmove is aborted, those segments that are completed will remain on the destination and those that are not yet addressed or in the process of moving will stay on the source PV. Thus, it is possible to have a partially completed move - some LVs (or certain segments of LVs) on the source PV and some on the destination. Example 2: What 'example 1' might look if it was half-way through the move. --------- --------- --------- | LV1s0 | | LV2s0 | | LV1s1 | --------- --------- --------- | | | ------------------------------------- pvmove0 | seg 0 | seg 1 | seg 2 | ------------------------------------- | | | | ------------------------- source PV | | 256 - 511 | 512 - 767 | | ------------------------- | || ------------------------- dest PV | 000 - 255 | 256 - 511 | ------------------------- This update allows the user to specify that they would like the pvmove mirror created "by LV" rather than "by segment". That is, the pvmove LV becomes an image in an encapsulating mirror along with the allocated copy image. Example 3: A pvmove that is performed "by LV" rather than "by segment". --------- --------- | LV1s0 | | LV2s0 | --------- --------- | | ------------------------- pvmove0 | * LV-level mirror * | ------------------------- / \ pvmove_mimage0 / pvmove_mimage1 ------------------------- ------------------------- | seg 0 | seg 1 | | seg 0 | seg 1 | ------------------------- ------------------------- | | | | ------------------------- ------------------------- | 000 - 255 | 256 - 511 | | 000 - 255 | 256 - 511 | ------------------------- ------------------------- source PV dest PV The thing that differentiates a pvmove done in this way and a simple "up-convert" from linear to mirror is the preservation of the distinct segments. A normal up-convert would simply allocate the necessary space with no regard for segment boundaries. The pvmove operation must preserve the segments because they are the critical boundary between the segments of the LVs being moved. So, when the pvmove copy image is allocated, all corresponding segments must be allocated. The code that merges ajoining segments that are part of the same LV when the metadata is written must also be avoided in this case. This method of mirroring is unique enough to warrant its own definitional macro, MIRROR_BY_SEGMENTED_LV. This joins the two existing macros: MIRROR_BY_SEG (for original pvmove) and MIRROR_BY_LV (for user created mirrors). The advantages of performing pvmove in this way is that all of the LVs affected can be moved together. It is an all-or-nothing approach that leaves all LV segments on the source PV if the move is aborted. Additionally, a mirror log can be used (in the future) to provide tracking of progress; allowing the copy to continue where it left off in the event there is a deactivation.
2014-06-18 07:59:36 +04:00
return 1;
dm_list_iterate_safe(segh, t, &lv->segments) {
current = dm_list_item(segh, struct lv_segment);
if (_merge(prev, current))
dm_list_del(&current->list);
else
prev = current;
}
return 1;
}
#define ERROR_MAX 100
#define inc_error_count \
if (error_count++ > ERROR_MAX) \
goto out
/*
* RAID segment property checks.
*
* Checks in here shall catch any
* bogus segment structure setup.
*/
#define raid_seg_error(msg) { \
log_error("LV %s invalid: %s for %s segment", \
seg->lv->name, (msg), lvseg_name(seg)); \
if ((*error_count)++ > ERROR_MAX) \
return; \
}
#define raid_seg_error_val(msg, val) { \
log_error("LV %s invalid: %s (is %u) for %s segment", \
seg->lv->name, (msg), (val), lvseg_name(seg)); \
if ((*error_count)++ > ERROR_MAX) \
return; \
}
/* Check raid0 segment properties in @seg */
static void _check_raid0_seg(struct lv_segment *seg, int *error_count)
{
if (seg_is_raid0_meta(seg) &&
!seg->meta_areas)
raid_seg_error("no meta areas");
if (!seg->stripe_size)
raid_seg_error("zero stripe size");
if (!is_power_of_2(seg->stripe_size))
raid_seg_error_val("non power of 2 stripe size", seg->stripe_size);
if (seg->region_size)
raid_seg_error_val("non-zero region_size", seg->region_size);
if (seg->writebehind)
raid_seg_error_val("non-zero write behind", seg->writebehind);
if (seg->min_recovery_rate)
raid_seg_error_val("non-zero min recovery rate", seg->min_recovery_rate);
if (seg->max_recovery_rate)
raid_seg_error_val("non-zero max recovery rate", seg->max_recovery_rate);
}
/* Check RAID @seg for non-zero, power of 2 region size and min recovery rate <= max */
static void _check_raid_region_recovery(struct lv_segment *seg, int *error_count)
{
if (!seg->region_size)
raid_seg_error("zero region_size");
if (!is_power_of_2(seg->region_size))
raid_seg_error_val("non power of 2 region size", seg->region_size);
/* min/max recovery rate may be zero but min may not be larger than max*/
if (seg->min_recovery_rate > seg->max_recovery_rate)
raid_seg_error_val("min recovery larger than max recovery larger", seg->min_recovery_rate);
}
/* Check raid1 segment properties in @seg */
static void _check_raid1_seg(struct lv_segment *seg, int *error_count)
{
if (!seg->meta_areas)
raid_seg_error("no meta areas");
if (seg->stripe_size)
raid_seg_error_val("non-zero stripe size", seg->stripe_size);
_check_raid_region_recovery(seg, error_count);
}
/* Check raid4/5/6/10 segment properties in @seg */
static void _check_raid45610_seg(struct lv_segment *seg, int *error_count)
{
/* Checks applying to any raid4/5/6/10 */
if (!seg->meta_areas)
raid_seg_error("no meta areas");
if (!seg->stripe_size)
raid_seg_error("zero stripe size");
if (!is_power_of_2(seg->stripe_size))
raid_seg_error_val("non power of 2 stripe size", seg->stripe_size);
_check_raid_region_recovery(seg, error_count);
/* END: checks applying to any raid4/5/6/10 */
/* Specific checks per raid level */
if (seg_is_raid4(seg) ||
seg_is_any_raid5(seg)) {
/*
* To allow for takeover between the MD raid1 and
* raid4/5 personalities, exactly 2 areas (i.e. DataLVs)
* can be mirrored by all raid1, raid4 and raid5 personalities.
* Hence allow a minimum of 2 areas.
*/
if (seg->area_count < 2)
raid_seg_error_val("minimum 2 areas required", seg->area_count);
} else if (seg_is_any_raid6(seg)) {
/*
* FIXME: MD raid6 supports a minimum of 4 areas.
* LVM requests a minimum of 5 due to easier
* processing of SubLVs to replace.
*
* Once that obstacle got removed, allow for a minimum of 4.
*/
if (seg->area_count < 5)
raid_seg_error_val("minimum 5 areas required", seg->area_count);
} else if (seg_is_raid10(seg)) {
/*
* FIXME: raid10 area_count minimum has to change to 2 once we
* support data_copies and odd numbers of stripes
*/
if (seg->area_count < 4)
raid_seg_error_val("minimum 4 areas required", seg->area_count);
if (seg->writebehind)
raid_seg_error_val("non-zero writebehind", seg->writebehind);
}
}
/* Check any non-RAID segment struct members in @seg and increment @error_count for any bogus ones */
static void _check_non_raid_seg_members(struct lv_segment *seg, int *error_count)
{
if (seg->origin) /* snap and thin */
raid_seg_error("non-zero origin LV");
if (seg->indirect_origin) /* thin */
raid_seg_error("non-zero indirect_origin LV");
if (seg->merge_lv) /* thin */
raid_seg_error("non-zero merge LV");
if (seg->cow) /* snap */
raid_seg_error("non-zero cow LV");
if (!dm_list_empty(&seg->origin_list)) /* snap */
raid_seg_error("non-zero origin_list");
if (seg->log_lv)
raid_seg_error("non-zero log LV");
if (seg->segtype_private)
raid_seg_error("non-zero segtype_private");
/* thin members */
if (seg->metadata_lv)
raid_seg_error("non-zero metadata LV");
if (seg->transaction_id)
raid_seg_error("non-zero transaction_id");
if (seg->zero_new_blocks)
raid_seg_error("non-zero zero_new_blocks");
if (seg->discards)
raid_seg_error("non-zero discards");
if (!dm_list_empty(&seg->thin_messages))
raid_seg_error("non-zero thin_messages list");
if (seg->external_lv)
raid_seg_error("non-zero external LV");
if (seg->pool_lv)
raid_seg_error("non-zero pool LV");
if (seg->device_id)
raid_seg_error("non-zero device_id");
/* cache members */
if (seg->cache_mode)
raid_seg_error("non-zero cache_mode");
if (seg->policy_name)
raid_seg_error("non-zero policy_name");
if (seg->policy_settings)
raid_seg_error("non-zero policy_settings");
if (seg->cleaner_policy)
raid_seg_error("non-zero cleaner_policy");
/* replicator members (deprecated) */
if (seg->replicator)
raid_seg_error("non-zero replicator");
if (seg->rlog_lv)
raid_seg_error("non-zero rlog LV");
if (seg->rlog_type)
raid_seg_error("non-zero rlog type");
if (seg->rdevice_index_highest)
raid_seg_error("non-zero rdevice_index_highests");
if (seg->rsite_index_highest)
raid_seg_error("non-zero rsite_index_highests");
/* .... more members? */
}
/*
* Check RAID segment sruct members of @seg for acceptable
* properties and increment @error_count for any bogus ones.
*/
static void _check_raid_seg(struct lv_segment *seg, int *error_count)
{
uint32_t area_len, s;
/* General checks applying to all RAIDs */
if (!seg_is_raid(seg))
raid_seg_error("erroneous RAID check");
if (!seg->area_count)
raid_seg_error("zero area count");
if (!seg->areas)
raid_seg_error("zero areas");
if (seg->extents_copied > seg->area_len)
raid_seg_error_val("extents_copied too large", seg->extents_copied);
/* Default still 8, change! */
if (seg->area_count > DEFAULT_RAID_MAX_IMAGES) {
log_error("LV %s invalid: maximum supported areas %u (is %u) for %s segment",
seg->lv->name, DEFAULT_RAID_MAX_IMAGES, seg->area_count, lvseg_name(seg));
if ((*error_count)++ > ERROR_MAX)
return;
}
if (seg->chunk_size)
raid_seg_error_val("non-zero chunk_size", seg->chunk_size);
/* FIXME: should we check any non-RAID segment struct members at all? */
_check_non_raid_seg_members(seg, error_count);
/* Check for any DataLV flaws like non-existing ones or size variations */
for (area_len = s = 0; s < seg->area_count; s++) {
if (seg_type(seg, s) != AREA_LV)
raid_seg_error("no DataLV");
if (!lv_is_raid_image(seg_lv(seg, s)))
raid_seg_error("DataLV without RAID image flag");
if (area_len &&
area_len != seg_lv(seg, s)->le_count) {
raid_seg_error_val("DataLV size variations",
seg_lv(seg, s)->le_count);
} else
area_len = seg_lv(seg, s)->le_count;
}
/* Check for any MetaLV flaws like non-existing ones or size variations */
if (seg->meta_areas)
for (area_len = s = 0; s < seg->area_count; s++) {
if (seg_metatype(seg, s) != AREA_LV)
raid_seg_error("no MetaLV");
if (!lv_is_raid_metadata(seg_metalv(seg, s)))
raid_seg_error("MetaLV without RAID metadata flag");
if (area_len &&
area_len != seg_metalv(seg, s)->le_count) {
raid_seg_error_val("MetaLV size variations",
seg_metalv(seg, s)->le_count);
} else
area_len = seg_metalv(seg, s)->le_count;
}
/* END: general checks applying to all RAIDs */
/* Specific segment type checks from here on */
if (seg_is_any_raid0(seg))
_check_raid0_seg(seg, error_count);
else if (seg_is_raid1(seg))
_check_raid1_seg(seg, error_count);
else if (seg_is_raid4(seg) ||
seg_is_any_raid5(seg) ||
seg_is_raid10(seg))
_check_raid45610_seg(seg, error_count);
else
raid_seg_error("bogus RAID segment type");
}
/* END: RAID segment property checks. */
/*
* Verify that an LV's segments are consecutive, complete and don't overlap.
*/
2005-10-28 01:51:28 +04:00
int check_lv_segments(struct logical_volume *lv, int complete_vg)
{
2005-10-28 01:51:28 +04:00
struct lv_segment *seg, *seg2;
uint32_t le = 0;
unsigned seg_count = 0, seg_found;
2005-06-01 20:51:55 +04:00
uint32_t area_multiplier, s;
struct seg_list *sl;
struct glv_list *glvl;
int error_count = 0;
struct replicator_site *rsite;
struct replicator_device *rdev;
/* Check LV flags match first segment type */
if (complete_vg) {
if (lv_is_thin_volume(lv)) {
if (dm_list_size(&lv->segments) != 1) {
log_error("LV %s is thin volume without exactly one segment.",
lv->name);
inc_error_count;
} else if (!seg_is_thin_volume(first_seg(lv))) {
log_error("LV %s is thin volume without first thin volume segment.",
lv->name);
inc_error_count;
}
}
if (lv_is_thin_pool(lv)) {
if (dm_list_size(&lv->segments) != 1) {
log_error("LV %s is thin pool volume without exactly one segment.",
lv->name);
inc_error_count;
} else if (!seg_is_thin_pool(first_seg(lv))) {
log_error("LV %s is thin pool without first thin pool segment.",
lv->name);
inc_error_count;
}
}
if (lv_is_pool_data(lv) &&
(!(seg2 = first_seg(lv)) || !(seg2 = find_pool_seg(seg2)) ||
seg2->area_count != 1 || seg_type(seg2, 0) != AREA_LV ||
seg_lv(seg2, 0) != lv)) {
log_error("LV %s: segment 1 pool data LV does not point back to same LV",
lv->name);
inc_error_count;
}
if (lv_is_pool_metadata(lv)) {
if (!(seg2 = first_seg(lv)) || !(seg2 = find_pool_seg(seg2)) ||
seg2->metadata_lv != lv) {
log_error("LV %s: segment 1 pool metadata LV does not point back to same LV",
lv->name);
inc_error_count;
}
if (lv_is_thin_pool_metadata(lv) &&
!strstr(lv->name, "_tmeta")) {
log_error("LV %s: thin pool metadata LV does not use _tmeta",
lv->name);
inc_error_count;
} else if (lv_is_cache_pool_metadata(lv) &&
!strstr(lv->name, "_cmeta")) {
log_error("LV %s: cache pool metadata LV does not use _cmeta",
lv->name);
inc_error_count;
}
}
}
dm_list_iterate_items(seg, &lv->segments) {
seg_count++;
if (seg_is_raid(seg))
_check_raid_seg(seg, &error_count);
if (seg->le != le) {
log_error("LV %s invalid: segment %u should begin at "
"LE %" PRIu32 " (found %" PRIu32 ").",
lv->name, seg_count, le, seg->le);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
2008-01-30 17:00:02 +03:00
area_multiplier = segtype_is_striped(seg->segtype) ?
2005-06-01 20:51:55 +04:00
seg->area_count : 1;
if (seg->area_len * area_multiplier != seg->len) {
log_error("LV %s: segment %u has inconsistent "
"area_len %u",
lv->name, seg_count, seg->area_len);
inc_error_count;
}
if (lv_is_error_when_full(lv) &&
!seg_can_error_when_full(seg)) {
log_error("LV %s: segment %u (%s) does not support flag "
"ERROR_WHEN_FULL.", lv->name, seg_count, seg->segtype->name);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
if (complete_vg && seg->log_lv &&
!seg_is_mirrored(seg) && !(seg->status & RAID_IMAGE)) {
log_error("LV %s: segment %u log LV %s is not a "
"mirror log or a RAID image",
lv->name, seg_count, seg->log_lv->name);
inc_error_count;
}
/*
* Check mirror log - which is attached to the mirrored seg
*/
if (complete_vg && seg->log_lv && seg_is_mirrored(seg)) {
if (!lv_is_mirror_log(seg->log_lv)) {
log_error("LV %s: segment %u log LV %s is not "
"a mirror log",
lv->name, seg_count, seg->log_lv->name);
inc_error_count;
}
2005-10-28 16:48:50 +04:00
if (!(seg2 = first_seg(seg->log_lv)) ||
find_mirror_seg(seg2) != seg) {
log_error("LV %s: segment %u log LV does not "
"point back to mirror segment",
lv->name, seg_count);
inc_error_count;
}
}
2005-10-28 01:51:28 +04:00
if (complete_vg && seg->status & MIRROR_IMAGE) {
if (!find_mirror_seg(seg) ||
!seg_is_mirrored(find_mirror_seg(seg))) {
log_error("LV %s: segment %u mirror image "
"is not mirrored",
lv->name, seg_count);
inc_error_count;
}
}
/* Check the various thin segment types */
if (complete_vg) {
if (seg_is_thin_pool(seg)) {
if (!lv_is_thin_pool(lv)) {
log_error("LV %s is missing thin pool flag for segment %u",
lv->name, seg_count);
inc_error_count;
}
if (lv_is_thin_volume(lv)) {
log_error("LV %s is a thin volume that must not contain thin pool segment %u",
lv->name, seg_count);
inc_error_count;
}
}
2015-08-11 16:11:45 +03:00
if (seg_is_cache_pool(seg) &&
!dm_list_empty(&seg->lv->segs_using_this_lv)) {
switch (seg->cache_mode) {
case CACHE_MODE_WRITETHROUGH:
case CACHE_MODE_WRITEBACK:
case CACHE_MODE_PASSTHROUGH:
break;
default:
log_error("LV %s has invalid cache's feature flag.",
lv->name);
inc_error_count;
2015-08-11 16:11:45 +03:00
}
if (!seg->policy_name) {
log_error("LV %s is missing cache policy name.", lv->name);
inc_error_count;
}
}
if (seg_is_pool(seg)) {
if (seg->area_count != 1 ||
seg_type(seg, 0) != AREA_LV) {
log_error("LV %s: %s segment %u is missing a pool data LV",
lv->name, seg->segtype->name, seg_count);
inc_error_count;
} else if (!(seg2 = first_seg(seg_lv(seg, 0))) || find_pool_seg(seg2) != seg) {
log_error("LV %s: %s segment %u data LV does not refer back to pool LV",
lv->name, seg->segtype->name, seg_count);
inc_error_count;
}
if (!seg->metadata_lv) {
log_error("LV %s: %s segment %u is missing a pool metadata LV",
lv->name, seg->segtype->name, seg_count);
inc_error_count;
} else if (!(seg2 = first_seg(seg->metadata_lv)) ||
find_pool_seg(seg2) != seg) {
log_error("LV %s: %s segment %u metadata LV does not refer back to pool LV",
lv->name, seg->segtype->name, seg_count);
inc_error_count;
}
if (!validate_pool_chunk_size(lv->vg->cmd, seg->segtype, seg->chunk_size)) {
log_error("LV %s: %s segment %u has invalid chunk size %u.",
lv->name, seg->segtype->name, seg_count, seg->chunk_size);
inc_error_count;
}
} else {
if (seg->metadata_lv) {
log_error("LV %s: segment %u must not have pool metadata LV set",
lv->name, seg_count);
inc_error_count;
}
}
if (seg_is_thin_volume(seg)) {
if (!lv_is_thin_volume(lv)) {
log_error("LV %s is missing thin volume flag for segment %u",
lv->name, seg_count);
inc_error_count;
}
if (lv_is_thin_pool(lv)) {
log_error("LV %s is a thin pool that must not contain thin volume segment %u",
lv->name, seg_count);
inc_error_count;
}
2011-09-16 15:53:14 +04:00
if (!seg->pool_lv) {
log_error("LV %s: segment %u is missing thin pool LV",
lv->name, seg_count);
inc_error_count;
} else if (!lv_is_thin_pool(seg->pool_lv)) {
log_error("LV %s: thin volume segment %u pool LV is not flagged as a pool LV",
lv->name, seg_count);
inc_error_count;
}
if (seg->device_id > DM_THIN_MAX_DEVICE_ID) {
log_error("LV %s: thin volume segment %u has too large device id %u",
lv->name, seg_count, seg->device_id);
inc_error_count;
}
if (seg->external_lv && (seg->external_lv->status & LVM_WRITE)) {
log_error("LV %s: external origin %s is writable.",
lv->name, seg->external_lv->name);
inc_error_count;
}
2013-11-29 18:51:28 +04:00
if (seg->merge_lv) {
if (!lv_is_thin_volume(seg->merge_lv)) {
log_error("LV %s: thin volume segment %u merging LV %s is not flagged as a thin LV",
lv->name, seg_count, seg->merge_lv->name);
inc_error_count;
}
if (!lv_is_merging_origin(seg->merge_lv)) {
log_error("LV %s: merging LV %s is not flagged as merging.",
lv->name, seg->merge_lv->name);
inc_error_count;
}
}
} else if (seg_is_cache(seg)) {
if (!lv_is_cache(lv)) {
log_error("LV %s is missing cache flag for segment %u",
lv->name, seg_count);
inc_error_count;
}
if (!seg->pool_lv) {
log_error("LV %s: segment %u is missing cache_pool LV",
lv->name, seg_count);
inc_error_count;
}
} else {
2011-09-16 15:53:14 +04:00
if (seg->pool_lv) {
log_error("LV %s: segment %u must not have pool LV set",
lv->name, seg_count);
inc_error_count;
}
}
}
if (seg_is_snapshot(seg)) {
if (seg->cow && seg->cow == seg->origin) {
log_error("LV %s: segment %u has same LV %s for "
"both origin and snapshot",
lv->name, seg_count, seg->cow->name);
inc_error_count;
}
}
if (seg_is_replicator(seg) && !check_replicator_segment(seg))
inc_error_count;
2005-06-01 20:51:55 +04:00
for (s = 0; s < seg->area_count; s++) {
if (seg_type(seg, s) == AREA_UNASSIGNED) {
log_error("LV %s: segment %u has unassigned "
"area %u.",
lv->name, seg_count, s);
inc_error_count;
} else if (seg_type(seg, s) == AREA_PV) {
2005-06-01 20:51:55 +04:00
if (!seg_pvseg(seg, s) ||
seg_pvseg(seg, s)->lvseg != seg ||
seg_pvseg(seg, s)->lv_area != s) {
log_error("LV %s: segment %u has "
"inconsistent PV area %u",
lv->name, seg_count, s);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
} else {
if (!seg_lv(seg, s) ||
seg_lv(seg, s)->vg != lv->vg ||
seg_lv(seg, s) == lv) {
log_error("LV %s: segment %u has "
"inconsistent LV area %u",
lv->name, seg_count, s);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
2005-10-28 01:51:28 +04:00
if (complete_vg && seg_lv(seg, s) &&
lv_is_mirror_image(seg_lv(seg, s)) &&
(!(seg2 = find_seg_by_le(seg_lv(seg, s),
seg_le(seg, s))) ||
find_mirror_seg(seg2) != seg)) {
log_error("LV %s: segment %u mirror "
"image %u missing mirror ptr",
lv->name, seg_count, s);
inc_error_count;
}
/* FIXME I don't think this ever holds?
2005-06-01 20:51:55 +04:00
if (seg_le(seg, s) != le) {
log_error("LV %s: segment %u has "
"inconsistent LV area %u "
"size",
lv->name, seg_count, s);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
*/
seg_found = 0;
dm_list_iterate_items(sl, &seg_lv(seg, s)->segs_using_this_lv)
if (sl->seg == seg)
seg_found++;
if (!seg_found) {
log_error("LV %s segment %u uses LV %s,"
" but missing ptr from %s to %s",
lv->name, seg_count,
seg_lv(seg, s)->name,
seg_lv(seg, s)->name, lv->name);
inc_error_count;
} else if (seg_found > 1) {
log_error("LV %s has duplicated links "
"to LV %s segment %u",
seg_lv(seg, s)->name,
lv->name, seg_count);
inc_error_count;
}
2005-06-01 20:51:55 +04:00
}
if (complete_vg &&
seg_is_mirrored(seg) && !seg_is_raid(seg) &&
seg_type(seg, s) == AREA_LV &&
seg_lv(seg, s)->le_count != seg->area_len) {
log_error("LV %s: mirrored LV segment %u has "
"wrong size %u (should be %u).",
lv->name, s, seg_lv(seg, s)->le_count,
seg->area_len);
inc_error_count;
}
}
le += seg->len;
}
dm_list_iterate_items(sl, &lv->segs_using_this_lv) {
seg = sl->seg;
seg_found = 0;
for (s = 0; s < seg->area_count; s++) {
if (seg_type(seg, s) != AREA_LV)
continue;
if (lv == seg_lv(seg, s))
seg_found++;
if (seg_is_raid_with_meta(seg) && (lv == seg_metalv(seg, s)))
seg_found++;
}
if (seg_is_replicator_dev(seg)) {
dm_list_iterate_items(rsite, &seg->replicator->rsites) {
dm_list_iterate_items(rdev, &rsite->rdevices) {
if (lv == rdev->lv || lv == rdev->slog)
seg_found++;
}
}
if (lv == seg->replicator)
seg_found++;
}
if (seg_is_replicator(seg) && lv == seg->rlog_lv)
2016-05-24 02:57:48 +03:00
seg_found++;
if (seg->log_lv == lv)
seg_found++;
if (seg->metadata_lv == lv || seg->pool_lv == lv)
seg_found++;
if (seg_is_thin_volume(seg) && (seg->origin == lv || seg->external_lv == lv))
seg_found++;
2016-05-24 02:57:48 +03:00
if (!seg_found) {
log_error("LV %s is used by LV %s:%" PRIu32 "-%" PRIu32
", but missing ptr from %s to %s",
lv->name, seg->lv->name, seg->le,
seg->le + seg->len - 1,
seg->lv->name, lv->name);
inc_error_count;
} else if (seg_found != sl->count) {
log_error("Reference count mismatch: LV %s has %u "
"links to LV %s:%" PRIu32 "-%" PRIu32
", which has %u links",
lv->name, sl->count, seg->lv->name, seg->le,
seg->le + seg->len - 1, seg_found);
inc_error_count;
}
seg_found = 0;
dm_list_iterate_items(seg2, &seg->lv->segments)
2016-05-24 02:57:48 +03:00
if (seg == seg2) {
seg_found++;
break;
}
2016-05-24 02:57:48 +03:00
if (!seg_found) {
log_error("LV segment %s:%" PRIu32 "-%" PRIu32
" is incorrectly listed as being used by LV %s",
seg->lv->name, seg->le, seg->le + seg->len - 1,
lv->name);
inc_error_count;
}
}
dm_list_iterate_items(glvl, &lv->indirect_glvs) {
if (glvl->glv->is_historical) {
if (glvl->glv->historical->indirect_origin != lv->this_glv) {
log_error("LV %s is indirectly used by historical LV %s"
"but that historical LV does not point back to LV %s",
lv->name, glvl->glv->historical->name, lv->name);
inc_error_count;
}
} else {
if (!(seg = first_seg(glvl->glv->live)) ||
seg->indirect_origin != lv->this_glv) {
log_error("LV %s is indirectly used by LV %s"
"but that LV does not point back to LV %s",
lv->name, glvl->glv->live->name, lv->name);
inc_error_count;
}
}
}
2005-06-01 20:51:55 +04:00
if (le != lv->le_count) {
log_error("LV %s: inconsistent LE count %u != %u",
lv->name, le, lv->le_count);
inc_error_count;
2005-06-01 20:51:55 +04:00
}
out:
return !error_count;
}
2003-09-15 22:22:50 +04:00
/*
* Split the supplied segment at the supplied logical extent
* NB Use LE numbering that works across stripes PV1: 0,2,4 PV2: 1,3,5 etc.
2003-09-15 22:22:50 +04:00
*/
static int _lv_split_segment(struct logical_volume *lv, struct lv_segment *seg,
uint32_t le)
{
struct lv_segment *split_seg;
uint32_t s;
uint32_t offset = le - seg->le;
uint32_t area_offset;
2003-09-15 22:22:50 +04:00
if (!seg_can_split(seg)) {
2004-05-05 01:25:57 +04:00
log_error("Unable to split the %s segment at LE %" PRIu32
" in LV %s", lvseg_name(seg), le, lv->name);
2003-09-15 22:22:50 +04:00
return 0;
}
/* Clone the existing segment */
if (!(split_seg = alloc_lv_segment(seg->segtype,
2005-04-22 19:44:00 +04:00
seg->lv, seg->le, seg->len,
seg->status, seg->stripe_size,
seg->log_lv,
2005-04-22 19:44:00 +04:00
seg->area_count, seg->area_len,
2005-06-01 20:51:55 +04:00
seg->chunk_size, seg->region_size,
seg->extents_copied, seg->pvmove_source_seg))) {
2005-04-22 19:44:00 +04:00
log_error("Couldn't allocate cloned LV segment.");
2003-09-15 22:22:50 +04:00
return 0;
}
if (!str_list_dup(lv->vg->vgmem, &split_seg->tags, &seg->tags)) {
2004-03-08 20:19:15 +03:00
log_error("LV segment tags duplication failed");
return 0;
}
2003-09-15 22:22:50 +04:00
/* In case of a striped segment, the offset has to be / stripes */
area_offset = offset;
if (seg_is_striped(seg))
area_offset /= seg->area_count;
2003-09-15 22:22:50 +04:00
split_seg->area_len -= area_offset;
seg->area_len = area_offset;
split_seg->len -= offset;
seg->len = offset;
split_seg->le = seg->le + seg->len;
2003-09-15 22:22:50 +04:00
/* Adjust the PV mapping */
for (s = 0; s < seg->area_count; s++) {
2005-06-01 20:51:55 +04:00
seg_type(split_seg, s) = seg_type(seg, s);
2003-09-15 22:22:50 +04:00
/* Split area at the offset */
2005-06-01 20:51:55 +04:00
switch (seg_type(seg, s)) {
2003-09-15 22:22:50 +04:00
case AREA_LV:
if (!set_lv_segment_area_lv(split_seg, s, seg_lv(seg, s),
seg_le(seg, s) + seg->area_len, 0))
return_0;
log_debug_alloc("Split %s:%u[%u] at %u: %s LE %u", lv->name,
seg->le, s, le, seg_lv(seg, s)->name,
seg_le(split_seg, s));
2003-09-15 22:22:50 +04:00
break;
case AREA_PV:
if (!(seg_pvseg(split_seg, s) =
assign_peg_to_lvseg(seg_pv(seg, s),
2005-06-01 20:51:55 +04:00
seg_pe(seg, s) +
seg->area_len,
2005-06-01 20:51:55 +04:00
seg_pvseg(seg, s)->len -
seg->area_len,
2008-01-30 16:19:47 +03:00
split_seg, s)))
return_0;
log_debug_alloc("Split %s:%u[%u] at %u: %s PE %u", lv->name,
seg->le, s, le,
dev_name(seg_dev(seg, s)),
seg_pe(split_seg, s));
2003-09-15 22:22:50 +04:00
break;
case AREA_UNASSIGNED:
log_error("Unassigned area %u found in segment", s);
2003-09-15 22:22:50 +04:00
return 0;
}
}
/* Add split off segment to the list _after_ the original one */
dm_list_add_h(&seg->list, &split_seg->list);
2003-09-15 22:22:50 +04:00
return 1;
}
/*
* Ensure there's a segment boundary at the given logical extent
*/
int lv_split_segment(struct logical_volume *lv, uint32_t le)
{
struct lv_segment *seg;
if (!(seg = find_seg_by_le(lv, le))) {
log_error("Segment with extent %" PRIu32 " in LV %s not found",
le, lv->name);
return 0;
}
/* This is a segment start already */
if (le == seg->le)
return 1;
2008-01-30 16:19:47 +03:00
if (!_lv_split_segment(lv, seg, le))
return_0;
2003-09-15 22:22:50 +04:00
2008-01-30 16:19:47 +03:00
if (!vg_validate(lv->vg))
return_0;
2003-09-15 22:22:50 +04:00
return 1;
}