1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-21 13:34:40 +03:00
lvm2/lib/raid/raid.c
Jonathan Brassow c87907dcd5 lvconvert: linear -> raid1 upconvert should cause "recover" not "resync"
Two of the sync actions performed by the kernel (aka MD runtime) are
"resync" and "recover".  The "resync" refers to when an entirely new array
is going through the process of initializing (or resynchronizing after an
unexpected shutdown).  The "recover" is the process of initializing a new
member device to the array.  So, a brand new array with all new devices
will undergo "resync".  An array with replaced or added sub-LVs will undergo
"recover".

These two states are treated very differently when failures happen.  If any
device is lost or replaced while "resync", there are no worries.  This is
because any writes created from the inception of the array have occurred to
all the devices and can be safely recovered.  Even though non-initialized
portions will still be resync'ed with uninitialized data, it is ok.  However,
if a pre-existing device is lost (aka, the original linear device in a
linear -> raid1 convert) during a "recover", data loss can be the result.
Thus, writes are errored by the kernel and recovery is halted.  The failed
device must be restored or removed.  This is the correct behavior.

Unfortunately, we were treating an up-convert from linear as a "resync"
when we should have been treating it as a "recover".  This patch
removes the special case for linear upconvert.  It allows each new image
sub-LV to be marked with a rebuild flag and treats the array as 'in-sync'.
This has the correct effect of causing the upconvert to be treated as a
"recover" rather than a "resync".  There is no need to flag these two states
differently in LVM metadata, because they are already considered differently
by the kernel RAID metadata.  (Any activation/deactivation will properly
resume the "recover" process and not a "resync" process.)

We make this behavior change based on the presense of dm-raid target
version 1.9.0+.
2017-06-14 08:35:22 -05:00

676 lines
18 KiB
C

/*
* Copyright (C) 2011-2017 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 "lib.h"
#include "segtype.h"
#include "display.h"
#include "text_export.h"
#include "config.h"
#include "str_list.h"
#include "targets.h"
#include "lvm-string.h"
#include "activate.h"
#include "metadata.h"
#include "lv_alloc.h"
static void _raid_display(const struct lv_segment *seg)
{
unsigned s;
for (s = 0; s < seg->area_count; ++s) {
log_print(" Raid Data LV%2d", s);
display_stripe(seg, s, " ");
}
if (seg->meta_areas)
for (s = 0; s < seg->area_count; ++s)
if (seg_metalv(seg, s))
log_print(" Raid Metadata LV%2d\t%s", s, seg_metalv(seg, s)->name);
log_print(" ");
}
static int _raid_text_import_area_count(const struct dm_config_node *sn,
uint32_t *area_count)
{
uint32_t stripe_count = 0, device_count = 0;
int stripe_count_found, device_count_found;
device_count_found = dm_config_get_uint32(sn, "device_count", &device_count);
stripe_count_found = dm_config_get_uint32(sn, "stripe_count", &stripe_count);
if (!device_count_found && !stripe_count_found) {
log_error("Couldn't read 'device_count' or 'stripe_count' for "
"segment '%s'.", dm_config_parent_name(sn));
return 0;
}
if (device_count_found && stripe_count_found) {
log_error("Only one of 'device_count' and 'stripe_count' allowed for "
"segment '%s'.", dm_config_parent_name(sn));
return 0;
}
*area_count = stripe_count + device_count;
return 1;
}
static int _raid_text_import_areas(struct lv_segment *seg,
const struct dm_config_node *sn,
const struct dm_config_value *cv)
{
unsigned int s;
struct logical_volume *lv;
const char *seg_name = dm_config_parent_name(sn);
if (!seg->area_count) {
log_error("No areas found for segment %s", seg_name);
return 0;
}
for (s = 0; cv && s < seg->area_count; s++, cv = cv->next) {
if (cv->type != DM_CFG_STRING) {
log_error("Bad volume name in areas array for segment %s.", seg_name);
return 0;
}
/* Metadata device comes first. */
if (!(lv = find_lv(seg->lv->vg, cv->v.str))) {
log_error("Couldn't find volume '%s' for segment '%s'.",
cv->v.str ? : "NULL", seg_name);
return 0;
}
if (strstr(lv->name, "_rmeta_")) {
if (!set_lv_segment_area_lv(seg, s, lv, 0, RAID_META))
return_0;
cv = cv->next;
}
if (!cv) {
log_error("Missing data device in areas array for segment %s.", seg_name);
return 0;
}
/* Data device comes second */
if (!(lv = find_lv(seg->lv->vg, cv->v.str))) {
log_error("Couldn't find volume '%s' for segment '%s'.",
cv->v.str ? : "NULL", seg_name);
return 0;
}
if (!set_lv_segment_area_lv(seg, s, lv, 0, RAID_IMAGE))
return_0;
}
/*
* Check we read the correct number of RAID data/meta pairs.
*/
if (cv || (s < seg->area_count)) {
log_error("Incorrect number of areas in area array "
"for segment '%s'.", seg_name);
return 0;
}
return 1;
}
static int _raid_text_import(struct lv_segment *seg,
const struct dm_config_node *sn,
struct dm_hash_table *pv_hash)
{
const struct dm_config_value *cv;
const struct {
const char *name;
uint32_t *var;
} raid_attr_import[] = {
{ "region_size", &seg->region_size },
{ "stripe_size", &seg->stripe_size },
{ "data_copies", &seg->data_copies },
{ "writebehind", &seg->writebehind },
{ "min_recovery_rate", &seg->min_recovery_rate },
{ "max_recovery_rate", &seg->max_recovery_rate },
}, *aip = raid_attr_import;
unsigned i;
for (i = 0; i < DM_ARRAY_SIZE(raid_attr_import); i++, aip++) {
if (dm_config_has_node(sn, aip->name)) {
if (!dm_config_get_uint32(sn, aip->name, aip->var)) {
if (!strcmp(aip->name, "data_copies")) {
*aip->var = 0;
continue;
}
log_error("Couldn't read '%s' for segment %s of logical volume %s.",
aip->name, dm_config_parent_name(sn), seg->lv->name);
return 0;
}
}
}
if (!dm_config_get_list(sn, seg_is_raid0(seg) ? "raid0_lvs" : "raids", &cv)) {
log_error("Couldn't find RAID array for "
"segment %s of logical volume %s.",
dm_config_parent_name(sn), seg->lv->name);
return 0;
}
if (!_raid_text_import_areas(seg, sn, cv)) {
log_error("Failed to import RAID component pairs.");
return 0;
}
if (seg->data_copies < 2)
seg->data_copies = lv_raid_data_copies(seg->segtype, seg->area_count);
if (seg_is_any_raid0(seg))
seg->area_len /= seg->area_count;
return 1;
}
static int _raid_text_export_raid0(const struct lv_segment *seg, struct formatter *f)
{
outf(f, "stripe_count = %u", seg->area_count);
if (seg->stripe_size)
outf(f, "stripe_size = %" PRIu32, seg->stripe_size);
return out_areas(f, seg, seg_is_raid0(seg) ? "raid0_lv" : "raid");
}
static int _raid_text_export_raid(const struct lv_segment *seg, struct formatter *f)
{
int raid0 = seg_is_any_raid0(seg);
if (raid0)
outfc(f, (seg->area_count == 1) ? "# linear" : NULL,
"stripe_count = %u", seg->area_count);
else {
outf(f, "device_count = %u", seg->area_count);
if (seg_is_any_raid10(seg) && seg->data_copies > 0)
outf(f, "data_copies = %" PRIu32, seg->data_copies);
if (seg->region_size)
outf(f, "region_size = %" PRIu32, seg->region_size);
}
if (seg->stripe_size)
outf(f, "stripe_size = %" PRIu32, seg->stripe_size);
if (!raid0) {
if (seg_is_raid1(seg) && seg->writebehind)
outf(f, "writebehind = %" PRIu32, seg->writebehind);
if (seg->min_recovery_rate)
outf(f, "min_recovery_rate = %" PRIu32, seg->min_recovery_rate);
if (seg->max_recovery_rate)
outf(f, "max_recovery_rate = %" PRIu32, seg->max_recovery_rate);
}
return out_areas(f, seg, "raid");
}
static int _raid_text_export(const struct lv_segment *seg, struct formatter *f)
{
if (seg_is_any_raid0(seg))
return _raid_text_export_raid0(seg, f);
return _raid_text_export_raid(seg, f);
}
static int _raid_add_target_line(struct dev_manager *dm __attribute__((unused)),
struct dm_pool *mem __attribute__((unused)),
struct cmd_context *cmd __attribute__((unused)),
void **target_state __attribute__((unused)),
struct lv_segment *seg,
const struct lv_activate_opts *laopts __attribute__((unused)),
struct dm_tree_node *node, uint64_t len,
uint32_t *pvmove_mirror_count __attribute__((unused)))
{
int delta_disks = 0, delta_disks_minus = 0, delta_disks_plus = 0, data_offset = 0;
uint32_t s;
uint64_t flags = 0;
uint64_t rebuilds[RAID_BITMAP_SIZE];
uint64_t writemostly[RAID_BITMAP_SIZE];
struct dm_tree_node_raid_params_v2 params;
memset(&params, 0, sizeof(params));
memset(&rebuilds, 0, sizeof(rebuilds));
memset(&writemostly, 0, sizeof(writemostly));
if (!seg->area_count) {
log_error(INTERNAL_ERROR "_raid_add_target_line called "
"with no areas for %s.", seg->lv->name);
return 0;
}
/*
* 253 device restriction imposed by kernel due to MD and dm-raid bitfield limitation in superblock.
* It is not strictly a userspace limitation.
*/
if (seg->area_count > DEFAULT_RAID_MAX_IMAGES) {
log_error("Unable to handle more than %u devices in a "
"single RAID array", DEFAULT_RAID_MAX_IMAGES);
return 0;
}
if (!seg_is_any_raid0(seg)) {
if (!seg->region_size) {
log_error("Missing region size for raid segment in %s.",
seg_lv(seg, 0)->name);
return 0;
}
for (s = 0; s < seg->area_count; s++) {
uint64_t status = seg_lv(seg, s)->status;
if (status & LV_REBUILD)
rebuilds[s/64] |= 1ULL << (s%64);
if (status & LV_RESHAPE_DELTA_DISKS_PLUS) {
delta_disks++;
delta_disks_plus++;
} else if (status & LV_RESHAPE_DELTA_DISKS_MINUS) {
delta_disks--;
delta_disks_minus++;
}
if (delta_disks_plus && delta_disks_minus) {
log_error(INTERNAL_ERROR "Invalid request for delta disks minus and delta disks plus!");
return 0;
}
if (status & LV_WRITEMOSTLY)
writemostly[s/64] |= 1ULL << (s%64);
}
data_offset = seg->data_offset;
if (mirror_in_sync())
flags = DM_NOSYNC;
}
params.raid_type = lvseg_name(seg);
if (seg->segtype->parity_devs) {
/* RAID 4/5/6 */
params.mirrors = 1;
params.stripes = seg->area_count - seg->segtype->parity_devs;
} else if (seg_is_any_raid0(seg)) {
params.mirrors = 1;
params.stripes = seg->area_count;
} else if (seg_is_any_raid10(seg)) {
params.data_copies = seg->data_copies;
params.stripes = seg->area_count;
} else {
/* RAID 1 */
params.mirrors = seg->data_copies;
params.stripes = 1;
params.writebehind = seg->writebehind;
memcpy(params.writemostly, writemostly, sizeof(params.writemostly));
}
/* RAID 0 doesn't have a bitmap, thus no region_size, rebuilds etc. */
if (!seg_is_any_raid0(seg)) {
params.region_size = seg->region_size;
memcpy(params.rebuilds, rebuilds, sizeof(params.rebuilds));
params.min_recovery_rate = seg->min_recovery_rate;
params.max_recovery_rate = seg->max_recovery_rate;
params.delta_disks = delta_disks;
params.data_offset = data_offset;
}
params.stripe_size = seg->stripe_size;
params.flags = flags;
if (!dm_tree_node_add_raid_target_with_params_v2(node, len, &params))
return_0;
return add_areas_line(dm, seg, node, 0u, seg->area_count);
}
static int _raid_target_status_compatible(const char *type)
{
return (strstr(type, "raid") != NULL);
}
static void _raid_destroy(struct segment_type *segtype)
{
dm_free((void *) segtype);
}
#ifdef DEVMAPPER_SUPPORT
static int _raid_target_percent(void **target_state,
dm_percent_t *percent,
struct dm_pool *mem,
struct cmd_context *cmd,
struct lv_segment *seg, char *params,
uint64_t *total_numerator,
uint64_t *total_denominator)
{
int i;
uint64_t numerator, denominator;
char *pos = params;
/*
* Status line:
* <raid_type> <#devs> <status_chars> <synced>/<total>
* Example:
* raid1 2 AA 1024000/1024000
*/
for (i = 0; i < 3; i++) {
pos = strstr(pos, " ");
if (pos)
pos++;
else
break;
}
if (!pos || (sscanf(pos, FMTu64 "/" FMTu64 "%n", &numerator, &denominator, &i) != 2) ||
!denominator) {
log_error("Failed to parse %s status fraction: %s",
(seg) ? seg->segtype->name : "segment", params);
return 0;
}
*total_numerator += numerator;
*total_denominator += denominator;
if (seg)
seg->extents_copied = (uint64_t) seg->area_len * dm_make_percent(numerator, denominator) / DM_PERCENT_100;
*percent = dm_make_percent(numerator, denominator);
return 1;
}
static int _raid_transient_status(struct dm_pool *mem,
struct lv_segment *seg,
char *params)
{
int failed = 0, r = 0;
unsigned i;
struct lvinfo info;
struct logical_volume *lv;
struct dm_status_raid *sr;
log_debug("Raid transient status %s.", params);
if (!dm_get_status_raid(mem, params, &sr))
return_0;
if (sr->dev_count != seg->area_count) {
log_error("Active raid has a wrong number of raid images!");
log_error("Metadata says %u, kernel says %u.",
seg->area_count, sr->dev_count);
goto out;
}
if (seg->meta_areas)
for (i = 0; i < seg->area_count; ++i) {
lv = seg_metalv(seg, i);
if (!lv_info(lv->vg->cmd, lv, 0, &info, 0, 0)) {
log_error("Check for existence of raid meta %s failed.",
display_lvname(lv));
goto out;
}
}
for (i = 0; i < seg->area_count; ++i) {
lv = seg_lv(seg, i);
if (!lv_info(lv->vg->cmd, lv, 0, &info, 0, 0)) {
log_error("Check for existence of raid image %s failed.",
display_lvname(lv));
goto out;
}
if (sr->dev_health[i] == 'D') {
lv->status |= PARTIAL_LV;
++failed;
}
}
/* Update PARTIAL_LV flags across the VG */
if (failed)
vg_mark_partial_lvs(lv->vg, 0);
r = 1;
out:
dm_pool_free(mem, sr);
return r;
}
/* Define raid feature based on the tuple(major, minor, patchlevel) of raid target */
struct raid_feature {
uint32_t maj;
uint32_t min;
uint32_t patchlevel;
unsigned raid_feature;
const char *feature;
};
/* Return true if tuple(@maj, @min, @patchlevel) is greater/equal to @*feature members */
static int _check_feature(const struct raid_feature *feature, uint32_t maj, uint32_t min, uint32_t patchlevel)
{
return (maj > feature->maj) ||
(maj == feature->maj && min >= feature->min) ||
(maj == feature->maj && min == feature->min && patchlevel >= feature->patchlevel);
}
static int _raid_target_present(struct cmd_context *cmd,
const struct lv_segment *seg __attribute__((unused)),
unsigned *attributes)
{
/* List of features with their kernel target version */
const struct raid_feature _features[] = {
{ 1, 3, 0, RAID_FEATURE_RAID10, SEG_TYPE_NAME_RAID10 },
{ 1, 7, 0, RAID_FEATURE_RAID0, SEG_TYPE_NAME_RAID0 },
{ 1, 9, 0, RAID_FEATURE_SHRINK, "shrinking" },
{ 1, 9, 0, RAID_FEATURE_NEW_DEVICES_ACCEPT_REBUILD, "rebuild+emptymeta" },
{ 1, 10, 1, RAID_FEATURE_RESHAPE, "reshaping" },
};
static int _raid_checked = 0;
static int _raid_present = 0;
static unsigned _raid_attrs = 0;
uint32_t maj, min, patchlevel;
unsigned i;
if (!activation())
return 0;
if (!_raid_checked) {
_raid_checked = 1;
if (!(_raid_present = target_present(cmd, TARGET_NAME_RAID, 1)))
return 0;
if (!target_version("raid", &maj, &min, &patchlevel))
return_0;
for (i = 0; i < DM_ARRAY_SIZE(_features); ++i)
if (_check_feature(_features + i, maj, min, patchlevel))
_raid_attrs |= _features[i].raid_feature;
else
log_very_verbose("Target raid does not support %s.",
_features[i].feature);
/*
* Seperate check for proper raid4 mapping supported
*
* If we get more of these range checks, avoid them
* altogether by enhancing 'struct raid_feature'
* and _check_feature() to handle them.
*/
if (!(maj == 1 && (min == 8 || (min == 9 && patchlevel == 0))))
_raid_attrs |= RAID_FEATURE_RAID4;
else
log_very_verbose("Target raid does not support %s.",
SEG_TYPE_NAME_RAID4);
}
if (attributes)
*attributes = _raid_attrs;
return _raid_present;
}
static int _raid_modules_needed(struct dm_pool *mem,
const struct lv_segment *seg __attribute__((unused)),
struct dm_list *modules)
{
if (!str_list_add(mem, modules, MODULE_NAME_RAID)) {
log_error("raid module string list allocation failed");
return 0;
}
return 1;
}
# ifdef DMEVENTD
static const char *_get_raid_dso_path(struct cmd_context *cmd)
{
const char *config_str = find_config_tree_str(cmd, dmeventd_raid_library_CFG, NULL);
return get_monitor_dso_path(cmd, config_str);
}
static int _raid_target_monitored(struct lv_segment *seg, int *pending)
{
struct cmd_context *cmd = seg->lv->vg->cmd;
const char *dso_path = _get_raid_dso_path(cmd);
return target_registered_with_dmeventd(cmd, dso_path, seg->lv, pending);
}
static int _raid_set_events(struct lv_segment *seg, int evmask, int set)
{
struct cmd_context *cmd = seg->lv->vg->cmd;
const char *dso_path = _get_raid_dso_path(cmd);
return target_register_events(cmd, dso_path, seg->lv, evmask, set, 0);
}
static int _raid_target_monitor_events(struct lv_segment *seg, int events)
{
return _raid_set_events(seg, events, 1);
}
static int _raid_target_unmonitor_events(struct lv_segment *seg, int events)
{
return _raid_set_events(seg, events, 0);
}
# endif /* DMEVENTD */
#endif /* DEVMAPPER_SUPPORT */
static struct segtype_handler _raid_ops = {
.display = _raid_display,
.text_import_area_count = _raid_text_import_area_count,
.text_import = _raid_text_import,
.text_export = _raid_text_export,
.add_target_line = _raid_add_target_line,
.target_status_compatible = _raid_target_status_compatible,
#ifdef DEVMAPPER_SUPPORT
.target_percent = _raid_target_percent,
.target_present = _raid_target_present,
.check_transient_status = _raid_transient_status,
.modules_needed = _raid_modules_needed,
# ifdef DMEVENTD
.target_monitored = _raid_target_monitored,
.target_monitor_events = _raid_target_monitor_events,
.target_unmonitor_events = _raid_target_unmonitor_events,
# endif /* DMEVENTD */
#endif
.destroy = _raid_destroy,
};
static const struct raid_type {
const char name[12];
unsigned parity;
uint64_t extra_flags;
} _raid_types[] = {
{ SEG_TYPE_NAME_RAID0, 0, SEG_RAID0 | SEG_AREAS_STRIPED },
{ SEG_TYPE_NAME_RAID0_META, 0, SEG_RAID0_META | SEG_AREAS_STRIPED },
{ SEG_TYPE_NAME_RAID1, 0, SEG_RAID1 | SEG_AREAS_MIRRORED },
{ SEG_TYPE_NAME_RAID10, 0, SEG_RAID10 | SEG_AREAS_MIRRORED },
{ SEG_TYPE_NAME_RAID10_NEAR,0, SEG_RAID10_NEAR | SEG_AREAS_MIRRORED },
{ SEG_TYPE_NAME_RAID4, 1, SEG_RAID4 },
{ SEG_TYPE_NAME_RAID5, 1, SEG_RAID5 },
{ SEG_TYPE_NAME_RAID5_N, 1, SEG_RAID5_N },
{ SEG_TYPE_NAME_RAID5_LA, 1, SEG_RAID5_LA },
{ SEG_TYPE_NAME_RAID5_LS, 1, SEG_RAID5_LS },
{ SEG_TYPE_NAME_RAID5_RA, 1, SEG_RAID5_RA },
{ SEG_TYPE_NAME_RAID5_RS, 1, SEG_RAID5_RS },
{ SEG_TYPE_NAME_RAID6, 2, SEG_RAID6 },
{ SEG_TYPE_NAME_RAID6_N_6, 2, SEG_RAID6_N_6 },
{ SEG_TYPE_NAME_RAID6_NC, 2, SEG_RAID6_NC },
{ SEG_TYPE_NAME_RAID6_NR, 2, SEG_RAID6_NR },
{ SEG_TYPE_NAME_RAID6_ZR, 2, SEG_RAID6_ZR },
{ SEG_TYPE_NAME_RAID6_LS_6, 2, SEG_RAID6_LS_6 },
{ SEG_TYPE_NAME_RAID6_RS_6, 2, SEG_RAID6_RS_6 },
{ SEG_TYPE_NAME_RAID6_LA_6, 2, SEG_RAID6_LA_6 },
{ SEG_TYPE_NAME_RAID6_RA_6, 2, SEG_RAID6_RA_6 }
};
static struct segment_type *_init_raid_segtype(struct cmd_context *cmd,
const struct raid_type *rt,
uint64_t monitored)
{
struct segment_type *segtype = dm_zalloc(sizeof(*segtype));
if (!segtype) {
log_error("Failed to allocate memory for %s segtype",
rt->name);
return NULL;
}
segtype->ops = &_raid_ops;
segtype->name = rt->name;
segtype->flags = SEG_RAID | SEG_ONLY_EXCLUSIVE | rt->extra_flags;
/* Never monitor raid0 or raid0_meta LVs */
if (!segtype_is_any_raid0(segtype))
segtype->flags |= monitored;
segtype->parity_devs = rt->parity;
log_very_verbose("Initialised segtype: %s", segtype->name);
return segtype;
}
#ifdef RAID_INTERNAL /* Shared */
int init_raid_segtypes(struct cmd_context *cmd, struct segtype_library *seglib)
#else
int init_multiple_segtypes(struct cmd_context *cmd, struct segtype_library *seglib);
int init_multiple_segtypes(struct cmd_context *cmd, struct segtype_library *seglib)
#endif
{
struct segment_type *segtype;
unsigned i;
uint64_t monitored = 0;
#ifdef DEVMAPPER_SUPPORT
# ifdef DMEVENTD
if (_get_raid_dso_path(cmd))
monitored = SEG_MONITORED;
# endif
#endif
for (i = 0; i < DM_ARRAY_SIZE(_raid_types); ++i)
if ((segtype = _init_raid_segtype(cmd, &_raid_types[i], monitored)) &&
!lvm_register_segtype(seglib, segtype))
/* segtype is already destroyed */
return_0;
return 1;
}