/*
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2001-2013 Neil Brown <neilb@suse.de>
+ * Copyright (C) 2001-2016 Neil Brown <neilb@suse.com>
*
*
* This program is free software; you can redistribute it and/or modify
#include "mdadm.h"
#include <ctype.h>
-static int name_matches(char *found, char *required, char *homehost)
+mapping_t assemble_statuses[] = {
+ { "but cannot be started", INCR_NO },
+ { "but not safe to start", INCR_UNSAFE },
+ { "and started", INCR_YES },
+ { NULL, INCR_ALREADY }
+};
+
+
+/**
+ * struct assembly_array_info - General, meaningful information for assembly.
+ * @name: Array name.
+ * @new_cnt: Count of drives known to be members, recently added.
+ * @preexist_cnt: Count of member drives in pre-assembled array.
+ * @exp_cnt: Count of known expansion targets.
+ *
+ * FIXME: @exp_new_cnt for recently added expansion targets.
+ */
+struct assembly_array_info {
+ char *name;
+ int new_cnt;
+ int preexist_cnt;
+ int exp_cnt;
+};
+
+/**
+ * set_array_assembly_status() - generate status of assembly for an array.
+ * @c: Global settings.
+ * @result: Pointer to status mask.
+ * @status: Status to be set/printed.
+ * @arr: Array information.
+ *
+ * Print status message to user or set it in @result if it is not NULL.
+ */
+static void set_array_assembly_status(struct context *c,
+ int *result, int status,
+ struct assembly_array_info *arr)
+{
+ int raid_disks = arr->preexist_cnt + arr->new_cnt;
+ char *status_msg = map_num_s(assemble_statuses, status);
+
+ if (c->export && result)
+ *result |= status;
+
+ if (c->export || c->verbose < 0)
+ return;
+
+ pr_err("%s has been assembled with %d device%s", arr->name,
+ raid_disks, raid_disks == 1 ? "":"s");
+ if (arr->preexist_cnt > 0)
+ fprintf(stderr, " (%d new)", arr->new_cnt);
+ if (arr->exp_cnt)
+ fprintf(stderr, " ( + %d for expansion)", arr->exp_cnt);
+ fprintf(stderr, " %s.\n", status_msg);
+}
+
+static int name_matches(char *found, char *required, char *homehost, int require_homehost)
{
/* See if the name found matches the required name, possibly
* prefixed with 'homehost'
*/
- char fnd[33];
+ char *sep;
+ unsigned int l;
- strncpy(fnd, found, 32);
- fnd[32] = 0;
if (strcmp(found, required)==0)
return 1;
- if (homehost) {
- int l = strlen(homehost);
- if (l < 32 && fnd[l] == ':' &&
- strcmp(fnd+l+1, required)==0)
+ sep = strchr(found, ':');
+ if (!sep)
+ return 0;
+ l = sep - found;
+ if (strncmp(found, "any:", 4) == 0 ||
+ (homehost && strcmp(homehost, "any") == 0) ||
+ !require_homehost ||
+ (homehost && strlen(homehost) == l &&
+ strncmp(found, homehost, l) == 0)) {
+ /* matching homehost */
+ if (strcmp(sep+1, required) == 0)
return 1;
}
return 0;
static int ident_matches(struct mddev_ident *ident,
struct mdinfo *content,
struct supertype *tst,
- char *homehost,
- char *update, char *devname)
+ char *homehost, int require_homehost,
+ enum update_opt update, char *devname)
{
- if (ident->uuid_set && (!update || strcmp(update, "uuid")!= 0) &&
+ if (ident->uuid_set && update != UOPT_UUID &&
same_uuid(content->uuid, ident->uuid, tst->ss->swapuuid)==0 &&
memcmp(content->uuid, uuid_zero, sizeof(int[4])) != 0) {
if (devname)
pr_err("%s has wrong uuid.\n", devname);
return 0;
}
- if (ident->name[0] && (!update || strcmp(update, "name")!= 0) &&
- name_matches(content->name, ident->name, homehost)==0) {
+ if (ident->name[0] && update != UOPT_NAME &&
+ name_matches(content->name, ident->name, homehost, require_homehost)==0) {
if (devname)
pr_err("%s has wrong name.\n", devname);
return 0;
struct mdinfo *content = NULL;
int report_mismatch = ((inargv && c->verbose >= 0) || c->verbose > 0);
struct domainlist *domains = NULL;
+ dev_t rdev;
tmpdev = devlist; num_devs = 0;
while (tmpdev) {
tmpdev = tmpdev ? tmpdev->next : NULL) {
char *devname = tmpdev->devname;
int dfd;
- struct stat stb;
struct supertype *tst;
struct dev_policy *pol = NULL;
int found_container = 0;
pr_err("cannot open device %s: %s\n",
devname, strerror(errno));
tmpdev->used = 2;
- } else if (fstat(dfd, &stb)< 0) {
- /* Impossible! */
- pr_err("fstat failed for %s: %s\n",
- devname, strerror(errno));
- tmpdev->used = 2;
- } else if ((stb.st_mode & S_IFMT) != S_IFBLK) {
- pr_err("%s is not a block device.\n",
- devname);
+ } else if (!fstat_is_blkdev(dfd, devname, &rdev)) {
tmpdev->used = 2;
} else if (must_be_container(dfd)) {
if (st) {
pr_err("%s is a container, but we are looking for components\n",
devname);
tmpdev->used = 2;
-#if !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO)
} if (!tst && (tst = super_by_fd(dfd, NULL)) == NULL) {
if (report_mismatch)
pr_err("not a recognisable container: %s\n",
devname);
tmpdev->used = 2;
-#endif
- } else if (!tst->ss->load_container
- || tst->ss->load_container(tst, dfd, NULL)) {
+ } else if (!tst->ss->load_container ||
+ tst->ss->load_container(tst, dfd, NULL)) {
if (report_mismatch)
pr_err("no correct container type: %s\n",
devname);
tmpdev->used = 2;
} else if (auto_assem &&
- !conf_test_metadata(tst->ss->name, (pol = devid_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
if (report_mismatch)
- pr_err("%s has metadata type %s for which "
- "auto-assembly is disabled\n",
+ pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
devname, tst->ss->name);
tmpdev->used = 2;
} else
pr_err("no recogniseable superblock on %s\n",
devname);
tmpdev->used = 2;
- } else if (tst->ss->load_super(tst,dfd, NULL)) {
+ } else if ((tst->ignore_hw_compat = 0),
+ tst->ss->load_super(tst, dfd,
+ report_mismatch ? devname : NULL)) {
if (report_mismatch)
pr_err("no RAID superblock on %s\n",
devname);
tst->ss->name, devname);
tmpdev->used = 2;
} else if (auto_assem && st == NULL &&
- !conf_test_metadata(tst->ss->name, (pol = devid_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
if (report_mismatch)
- pr_err("%s has metadata type %s for which "
- "auto-assembly is disabled\n",
+ pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
devname, tst->ss->name);
tmpdev->used = 2;
}
if (auto_assem || !inargv)
/* Ignore unrecognised devices during auto-assembly */
goto loop;
- if (ident->uuid_set || ident->name[0] ||
+ if (ident->name[0] ||
ident->super_minor != UnSet)
/* Ignore unrecognised device if looking for
* specific array */
goto loop;
+ if (ident->uuid_set)
+ /* ignore unrecognized device if looking for
+ * specific uuid
+ */
+ goto loop;
pr_err("%s has no superblock - assembly aborted\n",
devname);
st->ss->free_super(st);
dev_policy_free(pol);
domain_free(domains);
+ if (tst) {
+ tst->ss->free_super(tst);
+ free(tst);
+ }
return -1;
}
content = content->next) {
if (!ident_matches(ident, content, tst,
- c->homehost, c->update,
+ c->homehost, c->require_homehost,
+ c->update,
report_mismatch ? devname : NULL))
/* message already printed */;
else if (is_member_busy(content->text_version)) {
st = tst; tst = NULL;
if (!auto_assem && inargv && tmpdev->next != NULL) {
- pr_err("%s is a container, but is not "
- "only device given: confused and aborting\n",
+ pr_err("%s is a container, but is not only device given: confused and aborting\n",
devname);
st->ss->free_super(st);
dev_policy_free(pol);
domain_free(domains);
+ free(st);
return -1;
}
if (c->verbose > 0)
tmpdev = NULL;
goto loop;
} else {
- int rv = 0;
- struct mddev_ident *match;
-
content = *contentp;
tst->ss->getinfo_super(tst, content, NULL);
if (!ident_matches(ident, content, tst,
- c->homehost, c->update,
+ c->homehost, c->require_homehost,
+ c->update,
report_mismatch ? devname : NULL))
goto loop;
- match = conf_match(tst, content, devname,
- report_mismatch ? c->verbose : -1,
- &rv);
- if (!match && rv == 2)
- goto loop;
- if (match && match->devname &&
- strcasecmp(match->devname, "<ignore>") == 0) {
- if (report_mismatch)
- pr_err("%s is a member of an explicitly ignored array\n",
- devname);
- goto loop;
- }
- if (match && !ident_matches(match, content, tst,
- c->homehost, c->update,
- report_mismatch ? devname : NULL))
- /* Array exists in mdadm.conf but some
- * details don't match, so reject it
+ if (auto_assem) {
+ /* Never auto-assemble things that conflict
+ * with mdadm.conf in some way
*/
- goto loop;
+ struct mddev_ident *match;
+ int rv = 0;
+
+ match = conf_match(tst, content, devname,
+ report_mismatch ? c->verbose : -1,
+ &rv);
+ if (!match && rv == 2)
+ goto loop;
+ if (match && match->devname &&
+ strcasecmp(match->devname, "<ignore>") == 0) {
+ if (report_mismatch)
+ pr_err("%s is a member of an explicitly ignored array\n",
+ devname);
+ goto loop;
+ }
+ if (match && !ident_matches(match, content, tst,
+ c->homehost, c->require_homehost,
+ c->update,
+ report_mismatch ? devname : NULL))
+ /* Array exists in mdadm.conf but some
+ * details don't match, so reject it
+ */
+ goto loop;
+ }
/* should be safe to try an exclusive open now, we
* have rejected anything that some other mdadm might
if (st->ss != tst->ss ||
st->minor_version != tst->minor_version ||
- st->ss->compare_super(st, tst) != 0) {
+ st->ss->compare_super(st, tst, 1) != 0) {
/* Some mismatch. If exactly one array matches this host,
* we can resolve on that one.
* Or, if we are auto assembling, we just ignore the second
st->ss->free_super(st);
dev_policy_free(pol);
domain_free(domains);
+ free(tst);
return -1;
}
tmpdev->used = 1;
/* Collect domain information from members only */
if (tmpdev && tmpdev->used == 1) {
if (!pol)
- pol = devid_policy(stb.st_rdev);
+ pol = devid_policy(rdev);
domain_merge(&domains, pol, tst?tst->ss->name:NULL);
}
dev_policy_free(pol);
pol = NULL;
- if (tst)
+ if (tst) {
tst->ss->free_super(tst);
+ free(tst);
+ }
}
/* Check if we found some imsm spares but no members */
if (dfd < 0 ||
st->ss->load_super(st, dfd, NULL))
tmpdev->used = 2;
- if (dfd > 0)
- close(dfd);
+ close_fd(&dfd);
}
}
/* Now reject spares that don't match domains of identified members */
for (tmpdev = devlist; tmpdev; tmpdev = tmpdev->next) {
- struct stat stb;
if (tmpdev->used != 3)
continue;
- if (stat(tmpdev->devname, &stb)< 0) {
- pr_err("fstat failed for %s: %s\n",
- tmpdev->devname, strerror(errno));
+ if (!stat_is_blkdev(tmpdev->devname, &rdev)) {
tmpdev->used = 2;
} else {
- struct dev_policy *pol = devid_policy(stb.st_rdev);
+ struct dev_policy *pol = devid_policy(rdev);
int dt = domain_test(domains, pol, NULL);
if (inargv && dt != 0)
/* take this spare as domains match
};
static int load_devices(struct devs *devices, char *devmap,
- struct mddev_ident *ident, struct supertype *st,
+ struct mddev_ident *ident, struct supertype **stp,
struct mddev_dev *devlist, struct context *c,
struct mdinfo *content,
int mdfd, char *mddev,
struct mddev_dev *tmpdev;
int devcnt = 0;
int nextspare = 0;
-#ifndef MDASSEMBLE
int bitmap_done = 0;
-#endif
int most_recent = -1;
int bestcnt = 0;
int *best = *bestp;
+ struct supertype *st = *stp;
for (tmpdev = devlist; tmpdev; tmpdev=tmpdev->next) {
char *devname = tmpdev->devname;
struct stat stb;
+ struct supertype *tst;
int i;
+ int dfd;
+ int disk_state;
if (tmpdev->used != 1)
continue;
/* looks like a good enough match to update the super block if needed */
-#ifndef MDASSEMBLE
if (c->update) {
- int dfd;
/* prepare useful information in info structures */
struct stat stb2;
- struct supertype *tst;
int err;
fstat(mdfd, &stb2);
- if (strcmp(c->update, "uuid")==0 &&
- !ident->uuid_set) {
- int rfd;
- if ((rfd = open("/dev/urandom", O_RDONLY)) < 0 ||
- read(rfd, ident->uuid, 16) != 16) {
- *(__u32*)(ident->uuid) = random();
- *(__u32*)(ident->uuid+1) = random();
- *(__u32*)(ident->uuid+2) = random();
- *(__u32*)(ident->uuid+3) = random();
- }
- if (rfd >= 0) close(rfd);
+ if (c->update == UOPT_UUID && !ident->uuid_set)
+ random_uuid((__u8 *)ident->uuid);
+
+ if (c->update == UOPT_PPL && ident->bitmap_fd >= 0) {
+ pr_err("PPL is not compatible with bitmap\n");
+ close(mdfd);
+ free(devices);
+ free(devmap);
+ return -1;
}
+
dfd = dev_open(devname,
tmpdev->disposition == 'I'
? O_RDWR : (O_RDWR|O_EXCL));
close(mdfd);
free(devices);
free(devmap);
+ tst->ss->free_super(tst);
+ free(tst);
+ *stp = st;
return -1;
}
tst->ss->getinfo_super(tst, content, devmap + devcnt * content->array.raid_disks);
strcpy(content->name, ident->name);
content->array.md_minor = minor(stb2.st_rdev);
- if (strcmp(c->update, "byteorder") == 0)
+ if (c->update == UOPT_BYTEORDER)
err = 0;
+ else if (c->update == UOPT_HOME_CLUSTER) {
+ tst->cluster_name = c->homecluster;
+ err = tst->ss->write_bitmap(tst, dfd, NameUpdate);
+ } else if (c->update == UOPT_NODES) {
+ tst->nodes = c->nodes;
+ err = tst->ss->write_bitmap(tst, dfd, NodeNumUpdate);
+ } else if (c->update == UOPT_REVERT_RESHAPE && c->invalid_backup)
+ err = tst->ss->update_super(tst, content,
+ UOPT_SPEC_REVERT_RESHAPE_NOBACKUP,
+ devname, c->verbose,
+ ident->uuid_set,
+ c->homehost);
else
- err = tst->ss->update_super(tst, content, c->update,
+ err = tst->ss->update_super(tst, content,
+ c->update,
devname, c->verbose,
ident->uuid_set,
c->homehost);
if (err < 0) {
if (err == -1)
- pr_err("--update=%s not understood"
- " for %s metadata\n",
- c->update, tst->ss->name);
+ pr_err("--update=%s not understood for %s metadata\n",
+ map_num(update_options, c->update), tst->ss->name);
tst->ss->free_super(tst);
free(tst);
close(mdfd);
close(dfd);
free(devices);
free(devmap);
+ *stp = st;
return -1;
}
- if (strcmp(c->update, "uuid")==0 &&
+ if (c->update == UOPT_UUID &&
!ident->uuid_set) {
ident->uuid_set = 1;
memcpy(ident->uuid, content->uuid, 16);
if (tst->ss->store_super(tst, dfd))
pr_err("Could not re-write superblock on %s.\n",
devname);
- close(dfd);
- if (strcmp(c->update, "uuid")==0 &&
+ if (c->update == UOPT_UUID &&
ident->bitmap_fd >= 0 && !bitmap_done) {
if (bitmap_update_uuid(ident->bitmap_fd,
content->uuid,
else
bitmap_done = 1;
}
- tst->ss->free_super(tst);
- } else
-#endif
- {
- struct supertype *tst = dup_super(st);
- int dfd;
+ } else {
dfd = dev_open(devname,
tmpdev->disposition == 'I'
? O_RDWR : (O_RDWR|O_EXCL));
+ tst = dup_super(st);
if (dfd < 0 || tst->ss->load_super(tst, dfd, NULL) != 0) {
pr_err("cannot re-read metadata from %s - aborting\n",
close(mdfd);
free(devices);
free(devmap);
+ tst->ss->free_super(tst);
+ free(tst);
+ *stp = st;
return -1;
}
tst->ss->getinfo_super(tst, content, devmap + devcnt * content->array.raid_disks);
- tst->ss->free_super(tst);
- close(dfd);
}
- stat(devname, &stb);
+ fstat(dfd, &stb);
+ close(dfd);
if (c->verbose > 0)
pr_err("%s is identified as a member of %s, slot %d%s.\n",
devices[devcnt].i.disk.major = major(stb.st_rdev);
devices[devcnt].i.disk.minor = minor(stb.st_rdev);
- if (devices[devcnt].i.disk.state == 6) {
+ disk_state = devices[devcnt].i.disk.state & ~((1<<MD_DISK_FAILFAST) |
+ (1<<MD_DISK_WRITEMOSTLY));
+ if (disk_state == ((1<<MD_DISK_ACTIVE) | (1<<MD_DISK_SYNC))) {
if (most_recent < 0 ||
devices[devcnt].i.events
- > devices[most_recent].i.events)
+ > devices[most_recent].i.events) {
+ struct supertype *tmp = tst;
+ tst = st;
+ st = tmp;
most_recent = devcnt;
+ }
}
+ tst->ss->free_super(tst);
+ free(tst);
if (content->array.level == LEVEL_MULTIPATH)
/* with multipath, the raid_disk from the superblock is meaningless */
i = devcnt;
else
i = devices[devcnt].i.disk.raid_disk;
- if (i+1 == 0) {
+ if (i+1 == 0 || i == MD_DISK_ROLE_JOURNAL) {
if (nextspare < content->array.raid_disks*2)
nextspare = content->array.raid_disks*2;
i = nextspare++;
bestcnt = newbestcnt;
}
if (best[i] >=0 &&
- devices[best[i]].i.events
- == devices[devcnt].i.events
- && (devices[best[i]].i.disk.minor
- != devices[devcnt].i.disk.minor)
- && st->ss == &super0
- && content->array.level != LEVEL_MULTIPATH) {
+ devices[best[i]].i.events ==
+ devices[devcnt].i.events &&
+ (devices[best[i]].i.disk.minor !=
+ devices[devcnt].i.disk.minor) &&
+ st->ss == &super0 &&
+ content->array.level != LEVEL_MULTIPATH) {
/* two different devices with identical superblock.
* Could be a mis-detection caused by overlapping
* partitions. fail-safe.
*/
- pr_err("WARNING %s and %s appear"
- " to have very similar superblocks.\n"
- " If they are really different, "
- "please --zero the superblock on one\n"
- " If they are the same or overlap,"
- " please remove one from %s.\n",
+ pr_err("WARNING %s and %s appear to have very similar superblocks.\n"
+ " If they are really different, please --zero the superblock on one\n"
+ " If they are the same or overlap, please remove one from %s.\n",
devices[best[i]].devname, devname,
inargv ? "the list" :
"the\n DEVICE list in mdadm.conf"
close(mdfd);
free(devices);
free(devmap);
+ free(best);
+ *stp = st;
return -1;
}
- if (best[i] == -1
- || (devices[best[i]].i.events
- < devices[devcnt].i.events))
+ if (best[i] == -1 || (devices[best[i]].i.events
+ < devices[devcnt].i.events))
best[i] = devcnt;
+ else if (st->ss == &super_imsm)
+ best[i+1] = devcnt;
}
devcnt++;
}
*most_recentp = most_recent;
*bestcntp = bestcnt;
*bestp = best;
+ *stp = st;
return devcnt;
}
int okcnt = 0;
while (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1,
- avail)
- ||
+ avail) ||
(content->reshape_active && content->delta_disks > 0 &&
!enough(content->array.level, (content->array.raid_disks
- content->delta_disks),
- content->new_layout, 1,
- avail)
- )) {
+ content->new_layout, 1, avail))) {
/* Choose the newest best drive which is
* not up-to-date, update the superblock
* and add it.
int chosen_drive = -1;
int i;
- for (i = 0; i < content->array.raid_disks && i < bestcnt; i++) {
+ for (i = 0;
+ i < content->array.raid_disks * 2 && i < bestcnt;
+ i += 2) {
int j = best[i];
- if (j>=0 &&
- !devices[j].uptodate &&
- devices[j].i.recovery_start == MaxSector &&
- (chosen_drive < 0 ||
+ if (j < 0)
+ continue;
+ if (devices[j].uptodate)
+ continue;
+ if (devices[j].i.recovery_start != MaxSector) {
+ int delta;
+ if (!devices[j].i.reshape_active ||
+ devices[j].i.delta_disks <= 0)
+ continue;
+ /* When increasing number of devices, an
+ * added device also appears to be
+ * recovering. It is safe to include it
+ * as long as it won't be a source of
+ * data.
+ * For now, just allow for last data
+ * devices in RAID4 or last devices in RAID4/5/6.
+ */
+ delta = devices[j].i.delta_disks;
+ if (is_level456(devices[j].i.array.level) &&
+ i/2 >= content->array.raid_disks - delta)
+ /* OK */;
+ else if (devices[j].i.array.level == 4 &&
+ i/2 >= content->array.raid_disks - delta - 1)
+ /* OK */;
+ else
+ continue;
+ } else if (devices[j].i.reshape_active !=
+ content->reshape_active ||
+ (devices[j].i.reshape_active &&
+ devices[j].i.reshape_progress !=
+ content->reshape_progress))
+ /* Here, it may be a source of data. If two
+ * devices claim different progresses, it
+ * means that reshape boundaries differ for
+ * their own devices. Kernel will only treat
+ * the first one as reshape progress and
+ * go on. It may cause disaster, so avoid it.
+ */
+ continue;
+ if (chosen_drive < 0 ||
devices[j].i.events
- > devices[chosen_drive].i.events))
+ > devices[chosen_drive].i.events)
chosen_drive = j;
}
if (chosen_drive < 0)
current_events = devices[chosen_drive].i.events;
add_another:
if (c->verbose >= 0)
- pr_err("forcing event count in %s(%d) from %d upto %d\n",
+ pr_err("forcing event count in %s(%d) from %d up to %d\n",
devices[chosen_drive].devname,
devices[chosen_drive].i.disk.raid_disk,
(int)(devices[chosen_drive].i.events),
continue;
}
content->events = devices[most_recent].i.events;
- tst->ss->update_super(tst, content, "force-one",
+ tst->ss->update_super(tst, content, UOPT_SPEC_FORCE_ONE,
devices[chosen_drive].devname, c->verbose,
0, NULL);
avail[chosen_drive] = 1;
okcnt++;
tst->ss->free_super(tst);
-
/* If there are any other drives of the same vintage,
* add them in as well. We can't lose and we might gain
*/
- for (i = 0; i < content->array.raid_disks && i < bestcnt ; i++) {
+ for (i = 0;
+ i < content->array.raid_disks * 2 && i < bestcnt ;
+ i += 2) {
int j = best[i];
if (j >= 0 &&
!devices[j].uptodate &&
devices[j].i.recovery_start == MaxSector &&
- devices[j].i.events == current_events) {
+ devices[j].i.events == current_events &&
+ ((!devices[j].i.reshape_active &&
+ !content->reshape_active) ||
+ (devices[j].i.reshape_active ==
+ content->reshape_active &&
+ devices[j].i.reshape_progress ==
+ content->reshape_progress))) {
chosen_drive = j;
goto add_another;
}
unsigned int okcnt,
unsigned int sparecnt,
unsigned int rebuilding_cnt,
+ unsigned int journalcnt,
struct context *c,
int clean, char *avail,
int start_partial_ok,
int i;
unsigned int req_cnt;
+ if (content->journal_device_required && (content->journal_clean == 0)) {
+ if (!c->force) {
+ pr_err("Not safe to assemble with missing or stale journal device, consider --force.\n");
+ return 1;
+ }
+ pr_err("Journal is missing or stale, starting array read only.\n");
+ c->readonly = 1;
+ }
+
+ if (content->consistency_policy == CONSISTENCY_POLICY_PPL)
+ clean = 1;
+
rv = set_array_info(mdfd, st, content);
if (rv && !err_ok) {
pr_err("failed to set array info for %s: %s\n",
}
/* First, add the raid disks, but add the chosen one last */
- for (i=0; i<= bestcnt; i++) {
+ for (i = 0; i <= bestcnt; i++) {
int j;
if (i < bestcnt) {
j = best[i];
j = chosen_drive;
if (j >= 0 && !devices[j].included) {
- int dfd = dev_open(devices[j].devname,
- O_RDWR|O_EXCL);
+ int dfd;
+
+ dfd = dev_open(devices[j].devname, O_RDWR|O_EXCL);
if (dfd >= 0) {
remove_partitions(dfd);
close(dfd);
rv = add_disk(mdfd, st, content, &devices[j].i);
if (rv) {
- pr_err("failed to add "
- "%s to %s: %s\n",
- devices[j].devname,
- mddev,
+ pr_err("failed to add %s to %s: %s\n",
+ devices[j].devname, mddev,
strerror(errno));
- if (i < content->array.raid_disks * 2
- || i == bestcnt)
+ if (errno == EINVAL && content->array.level == 0 &&
+ content->array.layout != 0) {
+ cont_err("Possibly your kernel doesn't support RAID0 layouts.\n");
+ cont_err("Please upgrade.\n");
+ }
+ if (i < content->array.raid_disks * 2 ||
+ i == bestcnt)
okcnt--;
else
sparecnt--;
- } else if (c->verbose > 0)
+ } else if (c->verbose > 0) {
pr_err("added %s to %s as %d%s%s\n",
devices[j].devname, mddev,
devices[j].i.disk.raid_disk,
devices[j].uptodate?"":
" (possibly out of date)",
- (devices[j].i.disk.state & (1<<MD_DISK_REPLACEMENT))?" replacement":"");
+ (devices[j].i.disk.state &
+ (1<<MD_DISK_REPLACEMENT)) ?
+ " replacement":"");
+ }
} else if (j >= 0) {
if (c->verbose > 0)
pr_err("%s is already in %s as %d\n",
devices[j].devname, mddev,
devices[j].i.disk.raid_disk);
- } else if (c->verbose > 0 && i < content->array.raid_disks*2
- && (i&1) == 0)
+ } else if (c->verbose > 0 &&
+ i < content->array.raid_disks * 2 && (i & 1) == 0)
pr_err("no uptodate device for slot %d of %s\n",
- i, mddev);
+ i/2, mddev);
}
- if (content->array.level == LEVEL_CONTAINER) {
+ if (is_container(content->array.level)) {
+ sysfs_rules_apply(mddev, content);
if (c->verbose >= 0) {
- pr_err("Container %s has been "
- "assembled with %d drive%s",
- mddev, okcnt+sparecnt, okcnt+sparecnt==1?"":"s");
+ pr_err("Container %s has been assembled with %d drive%s",
+ mddev, okcnt + sparecnt + journalcnt,
+ okcnt + sparecnt + journalcnt == 1 ? "" : "s");
if (okcnt < (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)",
+ fprintf(stderr, " (out of %d)\n",
content->array.raid_disks);
- fprintf(stderr, "\n");
+ else
+ fprintf(stderr, "\n");
+ }
+
+ if (st->ss->validate_container) {
+ struct mdinfo *devices_list;
+ struct mdinfo *info_devices;
+ unsigned int count;
+
+ devices_list = NULL;
+ info_devices = xmalloc(sizeof(struct mdinfo) *
+ (okcnt + sparecnt));
+ for (count = 0; count < okcnt + sparecnt; count++) {
+ info_devices[count] = devices[count].i;
+ info_devices[count].next = devices_list;
+ devices_list = &info_devices[count];
+ }
+ if (st->ss->validate_container(devices_list))
+ pr_err("Mismatch detected!\n");
+ free(info_devices);
}
+
st->ss->free_super(st);
sysfs_uevent(content, "change");
+ if (err_ok && okcnt < (unsigned)content->array.raid_disks)
+ /* Was partial, is still partial, so signal an error
+ * to ensure we don't retry */
+ return 1;
return 0;
}
if (c->runstop == 1 ||
(c->runstop <= 0 &&
- ( enough(content->array.level, content->array.raid_disks,
- content->array.layout, clean, avail) &&
- (okcnt + rebuilding_cnt >= req_cnt || start_partial_ok)
- ))) {
+ (enough(content->array.level, content->array.raid_disks,
+ content->array.layout, clean, avail) &&
+ (okcnt + rebuilding_cnt >= req_cnt || start_partial_ok)))) {
/* This array is good-to-go.
* If a reshape is in progress then we might need to
* continue monitoring it. In that case we start
* it read-only and let the grow code make it writable.
*/
int rv;
-#ifndef MDASSEMBLE
+
if (content->reshape_active &&
!(content->reshape_active & RESHAPE_NO_BACKUP) &&
content->delta_disks <= 0) {
pr_err("%s: Need a backup file to complete reshape of this array.\n",
mddev);
pr_err("Please provided one with \"--backup-file=...\"\n");
- if (c->update &&
- strcmp(c->update, "revert-reshape") == 0)
+ if (c->update == UOPT_REVERT_RESHAPE)
pr_err("(Don't specify --update=revert-reshape again, that part succeeded.)\n");
return 1;
}
"array_state", "readonly");
if (rv == 0)
rv = Grow_continue(mdfd, st, content,
- c->backup_file,
+ c->backup_file, 0,
c->freeze_reshape);
} else if (c->readonly &&
- sysfs_attribute_available(
- content, NULL, "array_state")) {
+ sysfs_attribute_available(content, NULL,
+ "array_state")) {
rv = sysfs_set_str(content, NULL,
"array_state", "readonly");
} else
-#endif
rv = ioctl(mdfd, RUN_ARRAY, NULL);
+ reopen_mddev(mdfd); /* drop O_EXCL */
if (rv == 0) {
+ sysfs_rules_apply(mddev, content);
if (c->verbose >= 0) {
pr_err("%s has been started with %d drive%s",
mddev, okcnt, okcnt==1?"":"s");
if (okcnt < (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)", content->array.raid_disks);
+ fprintf(stderr, " (out of %d)",
+ content->array.raid_disks);
if (rebuilding_cnt)
- fprintf(stderr, "%s %d rebuilding", sparecnt?",":" and", rebuilding_cnt);
+ fprintf(stderr, "%s %d rebuilding",
+ sparecnt?",":" and",
+ rebuilding_cnt);
if (sparecnt)
- fprintf(stderr, " and %d spare%s", sparecnt, sparecnt==1?"":"s");
+ fprintf(stderr, " and %d spare%s",
+ sparecnt,
+ sparecnt == 1 ? "" : "s");
+ if (content->journal_clean)
+ fprintf(stderr, " and %d journal",
+ journalcnt);
fprintf(stderr, ".\n");
}
if (content->reshape_active &&
- content->array.level >= 4 &&
- content->array.level <= 6) {
+ is_level456(content->array.level)) {
/* might need to increase the size
* of the stripe cache - default is 256
*/
- if (256 < 4 * (content->array.chunk_size/4096)) {
- struct mdinfo *sra = sysfs_read(mdfd, NULL, 0);
+ int chunk_size = content->array.chunk_size;
+
+ if (content->reshape_active &&
+ content->new_chunk > chunk_size)
+ chunk_size = content->new_chunk;
+ if (256 < 4 * ((chunk_size+4065)/4096)) {
+ struct mdinfo *sra;
+
+ sra = sysfs_read(mdfd, NULL, 0);
if (sra)
sysfs_set_num(sra, NULL,
"stripe_cache_size",
- (4 * content->array.chunk_size / 4096) + 1);
+ (4 * chunk_size / 4096) + 1);
sysfs_free(sra);
}
}
if (content->array.level == 6 &&
okcnt + 1 == (unsigned)content->array.raid_disks &&
was_forced) {
- struct mdinfo *sra = sysfs_read(mdfd, NULL, 0);
+ struct mdinfo *sra;
+
+ sra = sysfs_read(mdfd, NULL, 0);
if (sra)
sysfs_set_str(sra, NULL,
"sync_action", "repair");
}
return 0;
}
- pr_err("failed to RUN_ARRAY %s: %s\n",
- mddev, strerror(errno));
+ pr_err("failed to RUN_ARRAY %s: %s\n", mddev, strerror(errno));
+ if (errno == 524 /* ENOTSUP */ &&
+ content->array.level == 0 && content->array.layout == 0)
+ cont_err("Please use --update=layout-original or --update=layout-alternate\n");
if (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1, avail))
- pr_err("Not enough devices to "
- "start the array.\n");
+ pr_err("Not enough devices to start the array.\n");
else if (!enough(content->array.level,
content->array.raid_disks,
- content->array.layout, clean,
- avail))
- pr_err("Not enough devices to "
- "start the array while not clean "
- "- consider --force.\n");
+ content->array.layout, clean, avail))
+ pr_err("Not enough devices to start the array while not clean - consider --force.\n");
return 1;
}
if (c->runstop == -1) {
pr_err("%s assembled from %d drive%s",
- mddev, okcnt, okcnt==1?"":"s");
+ mddev, okcnt, okcnt == 1 ? "" : "s");
if (okcnt != (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)", content->array.raid_disks);
+ fprintf(stderr, " (out of %d)",
+ content->array.raid_disks);
fprintf(stderr, ", but not started.\n");
return 2;
}
if (c->verbose >= -1) {
- pr_err("%s assembled from %d drive%s", mddev, okcnt, okcnt==1?"":"s");
+ pr_err("%s assembled from %d drive%s",
+ mddev, okcnt, okcnt == 1 ? "" : "s");
if (rebuilding_cnt)
- fprintf(stderr, "%s %d rebuilding", sparecnt?",":" and", rebuilding_cnt);
+ fprintf(stderr, "%s %d rebuilding",
+ sparecnt ? "," : " and", rebuilding_cnt);
if (sparecnt)
- fprintf(stderr, " and %d spare%s", sparecnt, sparecnt==1?"":"s");
+ fprintf(stderr, " and %d spare%s", sparecnt,
+ sparecnt == 1 ? "" : "s");
if (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1, avail))
fprintf(stderr, " - not enough to start the array.\n");
else if (!enough(content->array.level,
content->array.raid_disks,
- content->array.layout, clean,
- avail))
- fprintf(stderr, " - not enough to start the "
- "array while not clean - consider "
- "--force.\n");
+ content->array.layout, clean, avail))
+ fprintf(stderr, " - not enough to start the array while not clean - consider --force.\n");
else {
if (req_cnt == (unsigned)content->array.raid_disks)
- fprintf(stderr, " - need all %d to start it", req_cnt);
+ fprintf(stderr, " - need all %d to start it",
+ req_cnt);
else
fprintf(stderr, " - need %d to start", req_cnt);
fprintf(stderr, " (use --run to insist).\n");
* START_ARRAY
*
*/
- int rv;
- int mdfd;
+ int rv = -1;
+ int mdfd = -1;
int clean;
int auto_assem = (mddev == NULL && !ident->uuid_set &&
- ident->super_minor == UnSet && ident->name[0] == 0
- && (ident->container == NULL || ident->member == NULL));
- struct devs *devices;
+ ident->super_minor == UnSet && ident->name[0] == 0 &&
+ (ident->container == NULL || ident->member == NULL));
+ struct devs *devices = NULL;
char *devmap;
int *best = NULL; /* indexed by raid_disk */
int bestcnt = 0;
int devcnt;
- unsigned int okcnt, sparecnt, rebuilding_cnt, replcnt;
+ unsigned int okcnt, sparecnt, rebuilding_cnt, replcnt, journalcnt;
+ int journal_clean = 0;
int i;
int was_forced = 0;
int most_recent = 0;
mddev ? mddev : "further assembly");
content = &info;
- if (st)
+ if (st && c->force)
st->ignore_hw_compat = 1;
num_devs = select_devices(devlist, ident, &st, &content, c,
inargv, auto_assem);
*/
if (map_lock(&map))
pr_err("failed to get exclusive lock on mapfile - continue anyway...\n");
- mp = map_by_uuid(&map, content->uuid);
+ if (c->update == UOPT_UUID)
+ mp = NULL;
+ else
+ mp = map_by_uuid(&map, content->uuid);
if (mp) {
struct mdinfo *dv;
/* array already exists. */
pr_err("Found some drive for an array that is already active: %s\n",
mp->path);
pr_err("giving up.\n");
- return 1;
+ goto out;
}
for (dv = pre_exist->devs; dv; dv = dv->next) {
/* We want to add this device to our list,
name = content->name;
break;
}
+ if (mddev && map_by_name(&map, mddev) != NULL) {
+ pr_err("Cannot create device with %s because is in use\n", mddev);
+ goto out;
+ }
if (!auto_assem)
/* If the array is listed in mdadm.conf or on
* command line, then we trust the name
*/
trustworthy = LOCAL;
- if (name[0] == 0 &&
- content->array.level == LEVEL_CONTAINER) {
+ if (!name[0] && is_container(content->array.level)) {
name = content->text_version;
trustworthy = METADATA;
}
name = strchr(name, ':')+1;
mdfd = create_mddev(mddev, name, ident->autof, trustworthy,
- chosen_name);
+ chosen_name, 0);
}
if (mdfd < 0) {
st->ss->free_super(st);
if (auto_assem)
goto try_again;
- return 1;
+ goto out;
}
mddev = chosen_name;
- if (get_linux_version() < 2004000 ||
- md_get_version(mdfd) < 9000) {
- pr_err("Assemble requires Linux 2.4 or later, and\n"
- " md driver version 0.90.0 or later.\n"
- " Upgrade your kernel or try --build\n");
- close(mdfd);
- return 1;
- }
if (pre_exist == NULL) {
if (mddev_busy(fd2devnm(mdfd))) {
pr_err("%s already active, cannot restart it!\n",
st->ss->free_super(st);
if (auto_assem)
goto try_again;
- return 1;
+ goto out;
}
/* just incase it was started but has no content */
ioctl(mdfd, STOP_ARRAY, NULL);
}
-#ifndef MDASSEMBLE
if (content != &info) {
/* This is a member of a container. Try starting the array. */
int err;
err = assemble_container_content(st, mdfd, content, c,
- chosen_name);
+ chosen_name, NULL);
close(mdfd);
return err;
}
-#endif
+
/* Ok, no bad inconsistancy, we can try updating etc */
devices = xcalloc(num_devs, sizeof(*devices));
devmap = xcalloc(num_devs, content->array.raid_disks);
- devcnt = load_devices(devices, devmap, ident, st, devlist,
+ devcnt = load_devices(devices, devmap, ident, &st, devlist,
c, content, mdfd, mddev,
&most_recent, &bestcnt, &best, inargv);
- if (devcnt < 0)
- return 1;
+ if (devcnt < 0) {
+ mdfd = -3;
+ /*
+ * devices is already freed in load_devices, so set devices
+ * to NULL to avoid double free devices.
+ */
+ devices = NULL;
+ goto out;
+ }
if (devcnt == 0) {
pr_err("no devices found for %s\n",
mddev);
if (st)
st->ss->free_super(st);
- close(mdfd);
- free(devices);
free(devmap);
- return 1;
+ goto out;
}
- if (c->update && strcmp(c->update, "byteorder")==0)
+ if (c->update == UOPT_BYTEORDER)
st->minor_version = 90;
st->ss->getinfo_super(st, content, NULL);
okcnt = 0;
replcnt = 0;
sparecnt=0;
+ journalcnt=0;
rebuilding_cnt=0;
for (i=0; i< bestcnt; i++) {
int j = best[i];
/* note: we ignore error flags in multipath arrays
* as they don't make sense
*/
- if (content->array.level != LEVEL_MULTIPATH)
- if (!(devices[j].i.disk.state & (1<<MD_DISK_ACTIVE))) {
+ if (content->array.level != LEVEL_MULTIPATH) {
+ if (devices[j].i.disk.state & (1<<MD_DISK_JOURNAL)) {
+ if (content->journal_device_required)
+ journalcnt++;
+ else /* unexpected journal, mark as faulty */
+ devices[j].i.disk.state |= (1<<MD_DISK_FAULTY);
+ } else if (!(devices[j].i.disk.state & (1<<MD_DISK_ACTIVE))) {
if (!(devices[j].i.disk.state
& (1<<MD_DISK_FAULTY))) {
devices[j].uptodate = 1;
}
continue;
}
+ }
/* If this device thinks that 'most_recent' has failed, then
* we must reject this device.
*/
devices[most_recent].i.events
) {
devices[j].uptodate = 1;
+ if (devices[j].i.disk.state & (1<<MD_DISK_JOURNAL))
+ journal_clean = 1;
if (i < content->array.raid_disks * 2) {
if (devices[j].i.recovery_start == MaxSector ||
(content->reshape_active &&
- ((i >= content->array.raid_disks - content->delta_disks) ||
- (i >= content->array.raid_disks - content->delta_disks - 1
- && content->array.level == 4)))) {
+ i >= content->array.raid_disks - content->delta_disks)) {
if (!avail[i/2]) {
okcnt++;
avail[i/2]=1;
replcnt++;
} else
rebuilding_cnt++;
- } else
+ } else if (devices[j].i.disk.raid_disk != MD_DISK_ROLE_JOURNAL)
sparecnt++;
}
}
*/
chosen_drive = -1;
st->ss->free_super(st);
- for (i=0; chosen_drive < 0 && i<bestcnt; i++) {
+ for (i=0; chosen_drive < 0 && i<bestcnt; i+=2) {
int j = best[i];
int fd;
: (O_RDONLY|O_EXCL)))< 0) {
pr_err("Cannot open %s: %s\n",
devices[j].devname, strerror(errno));
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (st->ss->load_super(st,fd, NULL)) {
close(fd);
pr_err("RAID superblock has disappeared from %s\n",
devices[j].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
close(fd);
}
if (st->sb == NULL) {
pr_err("No suitable drives found for %s\n", mddev);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
st->ss->getinfo_super(st, content, NULL);
-#ifndef MDASSEMBLE
- sysfs_init(content, mdfd, NULL);
-#endif
+ if (sysfs_init(content, mdfd, NULL)) {
+ pr_err("Unable to initialize sysfs\n");
+ goto out;
+ }
+
+ /* after reload context, store journal_clean in context */
+ content->journal_clean = journal_clean;
for (i=0; i<bestcnt; i++) {
int j = best[i];
unsigned int desired_state;
- if (i >= content->array.raid_disks * 2)
+ if (j < 0)
+ continue;
+ if (devices[j].i.disk.raid_disk == MD_DISK_ROLE_JOURNAL)
+ desired_state = (1<<MD_DISK_JOURNAL);
+ else if (i >= content->array.raid_disks * 2)
desired_state = 0;
else if (i & 1)
desired_state = (1<<MD_DISK_ACTIVE) | (1<<MD_DISK_REPLACEMENT);
else
desired_state = (1<<MD_DISK_ACTIVE) | (1<<MD_DISK_SYNC);
- if (j<0)
- continue;
+ desired_state |= devices[j].i.disk.state & ((1<<MD_DISK_FAILFAST) |
+ (1<<MD_DISK_WRITEMOSTLY));
+
if (!devices[j].uptodate)
continue;
if (!(devices[j].i.array.state & 1))
clean = 0;
- if (st->ss->update_super(st, &devices[j].i, "assemble", NULL,
+ if (st->ss->update_super(st, &devices[j].i, UOPT_SPEC_ASSEMBLE, NULL,
c->verbose, 0, NULL)) {
if (c->force) {
if (c->verbose >= 0)
i, mddev, devices[j].devname);
}
}
-#if 0
- if (!(super.disks[i].i.disk.state & (1 << MD_DISK_FAULTY))) {
- pr_err("devices %d of %s is not marked FAULTY in superblock, but cannot be found\n",
- i, mddev);
- }
-#endif
}
- if (c->force && !clean &&
+ if (c->force && !clean && !is_container(content->array.level) &&
!enough(content->array.level, content->array.raid_disks,
- content->array.layout, clean,
- avail)) {
- change += st->ss->update_super(st, content, "force-array",
+ content->array.layout, clean, avail)) {
+ change += st->ss->update_super(st, content, UOPT_SPEC_FORCE_ARRAY,
devices[chosen_drive].devname, c->verbose,
0, NULL);
was_forced = 1;
if (fd < 0) {
pr_err("Could not open %s for write - cannot Assemble array.\n",
devices[chosen_drive].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (st->ss->store_super(st, fd)) {
close(fd);
pr_err("Could not re-write superblock on %s\n",
devices[chosen_drive].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (c->verbose >= 0)
pr_err("Marking array %s as 'clean'\n",
* that was moved aside due to the reshape overwriting live data
* The code of doing this lives in Grow.c
*/
-#ifndef MDASSEMBLE
if (content->reshape_active &&
!(content->reshape_active & RESHAPE_NO_BACKUP)) {
int err = 0;
int *fdlist = xmalloc(sizeof(int)* bestcnt);
if (c->verbose > 0)
- pr_err(":%s has an active reshape - checking "
- "if critical section needs to be restored\n",
+ pr_err("%s has an active reshape - checking if critical section needs to be restored\n",
chosen_name);
+ if (!c->backup_file)
+ c->backup_file = locate_backup(content->sys_name);
enable_fds(bestcnt/2);
for (i = 0; i < bestcnt/2; i++) {
int j = best[i*2];
c->backup_file, c->verbose > 0);
if (err && c->invalid_backup) {
if (c->verbose > 0)
- pr_err("continuing"
- " without restoring backup\n");
+ pr_err("continuing without restoring backup\n");
err = 0;
}
}
pr_err("Failed to restore critical section for reshape, sorry.\n");
if (c->backup_file == NULL)
cont_err("Possibly you needed to specify the --backup-file\n");
- close(mdfd);
- free(devices);
- return err;
+ goto out;
}
}
-#endif
/* Almost ready to actually *do* something */
/* First, fill in the map, so that udev can find our name
* as soon as we become active.
*/
- if (c->update && strcmp(c->update, "metadata")==0) {
+ if (c->update == UOPT_METADATA) {
content->array.major_version = 1;
content->array.minor_version = 0;
strcpy(content->text_version, "1.0");
rv = start_array(mdfd, mddev, content,
st, ident, best, bestcnt,
chosen_drive, devices, okcnt, sparecnt,
- rebuilding_cnt,
+ rebuilding_cnt, journalcnt,
c,
clean, avail, start_partial_ok,
pre_exist != NULL,
if (rv == 1 && !pre_exist)
ioctl(mdfd, STOP_ARRAY, NULL);
free(devices);
+out:
map_unlock(&map);
if (rv == 0) {
wait_for(chosen_name, mdfd);
break;
close(mdfd);
}
- usleep(usecs);
+ sleep_for(0, USEC_TO_NSEC(usecs), true);
usecs <<= 1;
}
}
- } else
+ } else if (mdfd >= 0)
close(mdfd);
+ free(best);
/* '2' means 'OK, but not started yet' */
+ if (rv == -1) {
+ free(devices);
+ return 1;
+ }
return rv == 2 ? 0 : rv;
}
-#ifndef MDASSEMBLE
int assemble_container_content(struct supertype *st, int mdfd,
struct mdinfo *content, struct context *c,
- char *chosen_name)
+ char *chosen_name, int *result)
{
- struct mdinfo *dev, *sra;
- int working = 0, preexist = 0;
- int expansion = 0;
- struct map_ent *map = NULL;
+ struct mdinfo *dev, *sra, *dev2;
+ struct assembly_array_info array = {chosen_name, 0, 0, 0};
int old_raid_disks;
int start_reshape;
+ char *avail;
+ int err;
+ int is_clean, all_disks;
+ bool is_raid456;
- sysfs_init(content, mdfd, NULL);
-
- sra = sysfs_read(mdfd, NULL, GET_VERSION);
- if (sra == NULL || strcmp(sra->text_version, content->text_version) != 0) {
- if (content->array.major_version == -1 &&
- content->array.minor_version == -2 &&
- c->readonly &&
- content->text_version[0] == '/')
- content->text_version[0] = '-';
- if (sysfs_set_array(content, md_get_version(mdfd)) != 0) {
- if (sra)
- sysfs_free(sra);
+ if (sysfs_init(content, mdfd, NULL)) {
+ pr_err("Unable to initialize sysfs\n");
+ return 1;
+ }
+
+ sra = sysfs_read(mdfd, NULL, GET_VERSION|GET_DEVS);
+ if (sra == NULL) {
+ pr_err("Failed to read sysfs parameters\n");
+ return 1;
+ }
+
+ /* Fill sysfs properties only if they are not set. Determine it by checking text_version
+ * and ignoring special character on the first place.
+ */
+ if (strcmp(sra->text_version + 1, content->text_version + 1) != 0) {
+ if (sysfs_set_array(content) != 0) {
+ sysfs_free(sra);
return 1;
}
}
if (st->ss->external && content->recovery_blocked && start_reshape)
block_subarray(content);
- if (sra)
- sysfs_free(sra);
+ for (dev2 = sra->devs; dev2; dev2 = dev2->next) {
+ for (dev = content->devs; dev; dev = dev->next)
+ if (dev2->disk.major == dev->disk.major &&
+ dev2->disk.minor == dev->disk.minor)
+ break;
+ if (dev)
+ continue;
+ /* Don't want this one any more */
+ if (sysfs_set_str(sra, dev2, "slot", STR_COMMON_NONE) < 0 && errno == EBUSY) {
+ pr_err("Cannot remove old device %s: not updating %s\n", dev2->sys_name, sra->sys_name);
+ sysfs_free(sra);
+ return 1;
+ }
+ sysfs_set_str(sra, dev2, "state", "remove");
+ }
old_raid_disks = content->array.raid_disks - content->delta_disks;
- for (dev = content->devs; dev; dev = dev->next)
+ avail = xcalloc(content->array.raid_disks, 1);
+ for (dev = content->devs; dev; dev = dev->next) {
+ if (dev->disk.raid_disk >= 0)
+ avail[dev->disk.raid_disk] = 1;
if (sysfs_add_disk(content, dev, 1) == 0) {
if (dev->disk.raid_disk >= old_raid_disks &&
content->reshape_active)
- expansion++;
+ array.exp_cnt++;
else
- working++;
+ array.new_cnt++;
} else if (errno == EEXIST)
- preexist++;
- if (working + expansion == 0)
- return 1;/* Nothing new, don't try to start */
+ array.preexist_cnt++;
+ }
+ sysfs_free(sra);
+
+ all_disks = array.new_cnt + array.exp_cnt + array.preexist_cnt;
- map_update(&map, fd2devnm(mdfd),
- content->text_version,
+ map_update(NULL, fd2devnm(mdfd), content->text_version,
content->uuid, chosen_name);
- if (c->runstop > 0 ||
- (working + preexist + expansion) >=
- content->array.working_disks) {
- int err;
+ if (content->consistency_policy == CONSISTENCY_POLICY_PPL &&
+ st->ss->validate_ppl) {
+ content->array.state |= 1;
+ err = 0;
- if (start_reshape) {
- int spare = content->array.raid_disks + expansion;
- if (restore_backup(st, content,
- working,
- spare, c->backup_file, c->verbose) == 1)
- return 1;
+ for (dev = content->devs; dev; dev = dev->next) {
+ int dfd;
+ char *devpath;
+ int ret;
- err = sysfs_set_str(content, NULL,
- "array_state", "readonly");
- if (err)
- return 1;
+ ret = st->ss->validate_ppl(st, content, dev);
+ if (ret == 0)
+ continue;
- if (st->ss->external) {
- if (!mdmon_running(st->container_devnm))
- start_mdmon(st->container_devnm);
- ping_monitor(st->container_devnm);
- if (mdmon_running(st->container_devnm) &&
- st->update_tail == NULL)
- st->update_tail = &st->updates;
+ if (ret < 0) {
+ err = 1;
+ break;
}
- err = Grow_continue(mdfd, st, content, c->backup_file,
- c->freeze_reshape);
- } else switch(content->array.level) {
- case LEVEL_LINEAR:
- case LEVEL_MULTIPATH:
- case 0:
- err = sysfs_set_str(content, NULL, "array_state",
- c->readonly ? "readonly" : "active");
+ if (!c->force) {
+ pr_err("%s contains invalid PPL - consider --force or --update-subarray with --update=no-ppl\n",
+ chosen_name);
+ content->array.state &= ~1;
+ avail[dev->disk.raid_disk] = 0;
break;
- default:
- err = sysfs_set_str(content, NULL, "array_state",
- "readonly");
- /* start mdmon if needed. */
- if (!err) {
- if (!mdmon_running(st->container_devnm))
- start_mdmon(st->container_devnm);
- ping_monitor(st->container_devnm);
- }
+ }
+
+ /* have --force - overwrite the invalid ppl */
+ devpath = map_dev(dev->disk.major, dev->disk.minor, 0);
+ dfd = dev_open(devpath, O_RDWR);
+ if (dfd < 0) {
+ pr_err("Failed to open %s\n", devpath);
+ err = 1;
break;
}
- if (!err)
- sysfs_set_safemode(content, content->safe_mode_delay);
- /* Block subarray here if it is not reshaped now
- * It has be blocked a little later to allow mdmon to switch in
- * in to R/W state
- */
- if (st->ss->external && content->recovery_blocked &&
- !start_reshape)
- block_subarray(content);
+ err = st->ss->write_init_ppl(st, content, dfd);
+ close(dfd);
- if (c->verbose >= 0) {
if (err)
- pr_err("array %s now has %d device%s",
- chosen_name, working + preexist,
- working + preexist == 1 ? "":"s");
- else
- pr_err("Started %s with %d device%s",
- chosen_name, working + preexist,
- working + preexist == 1 ? "":"s");
- if (preexist)
- fprintf(stderr, " (%d new)", working);
- if (expansion)
- fprintf(stderr, " ( + %d for expansion)",
- expansion);
- fprintf(stderr, "\n");
+ break;
}
- if (!err)
- wait_for(chosen_name, mdfd);
- return err;
- /* FIXME should have an O_EXCL and wait for read-auto */
- } else {
- if (c->verbose >= 0) {
- pr_err("%s assembled with %d device%s",
- chosen_name, preexist + working,
- preexist + working == 1 ? "":"s");
- if (preexist)
- fprintf(stderr, " (%d new)", working);
- fprintf(stderr, " but not started\n");
+
+ if (err) {
+ free(avail);
+ return err;
}
+ } else if (c->force) {
+ /* Set the array as 'clean' so that we can proceed with starting
+ * it even if we don't have all devices. Mdmon doesn't care
+ * if the dirty flag is set in metadata, it will start managing
+ * it anyway.
+ * This is really important for raid456 (RWH case), other levels
+ * are started anyway.
+ */
+ content->array.state |= 1;
+ }
+
+ is_raid456 = is_level456(content->array.level);
+ is_clean = content->array.state & 1;
+
+ if (enough(content->array.level, content->array.raid_disks,
+ content->array.layout, is_clean, avail) == 0) {
+ set_array_assembly_status(c, result, INCR_NO, &array);
+
+ if (c->verbose >= 0 && is_raid456 && !is_clean)
+ pr_err("Consider --force to start dirty degraded array\n");
+
+ free(avail);
+ return 1;
+ }
+ free(avail);
+
+ if (c->runstop <= 0 && all_disks < content->array.working_disks) {
+
+ set_array_assembly_status(c, result, INCR_UNSAFE, &array);
+
+ if (c->verbose >= 0 && c->force)
+ pr_err("Consider --run to start array as degraded.\n");
return 1;
}
+
+ if (is_raid456 && content->resync_start != MaxSector && c->force &&
+ all_disks < content->array.raid_disks) {
+
+ content->resync_start = MaxSector;
+ err = sysfs_set_num(content, NULL, "resync_start", MaxSector);
+ if (err)
+ return 1;
+
+ pr_err("%s array state forced to clean. It may cause data corruption.\n",
+ chosen_name);
+ }
+
+ /*
+ * Before activating the array, perform extra steps required
+ * to configure the internal write-intent bitmap.
+ */
+ if (content->consistency_policy == CONSISTENCY_POLICY_BITMAP &&
+ st->ss->set_bitmap)
+ st->ss->set_bitmap(st, content);
+
+ if (start_reshape) {
+ int spare = content->array.raid_disks + array.exp_cnt;
+ if (restore_backup(st, content,
+ array.new_cnt,
+ spare, &c->backup_file, c->verbose) == 1)
+ return 1;
+
+ if (content->reshape_progress == 0) {
+ /* If reshape progress is 0 - we are assembling the
+ * array that was stopped, before reshape has started.
+ * Array needs to be started as active, Grow_continue()
+ * will start the reshape.
+ */
+ sysfs_set_num(content, NULL, "reshape_position",
+ MaxSector);
+ err = sysfs_set_str(content, NULL,
+ "array_state", "active");
+ sysfs_set_num(content, NULL, "reshape_position", 0);
+ } else {
+ err = sysfs_set_str(content, NULL,
+ "array_state", "readonly");
+ }
+
+ if (err)
+ return 1;
+
+ if (st->ss->external) {
+ if (!mdmon_running(st->container_devnm))
+ start_mdmon(st->container_devnm);
+ ping_monitor(st->container_devnm);
+ if (mdmon_running(st->container_devnm) &&
+ st->update_tail == NULL)
+ st->update_tail = &st->updates;
+ }
+
+ err = Grow_continue(mdfd, st, content, c->backup_file,
+ 0, c->freeze_reshape);
+ } else switch(content->array.level) {
+ case LEVEL_LINEAR:
+ case LEVEL_MULTIPATH:
+ case 0:
+ err = sysfs_set_str(content, NULL, "array_state",
+ c->readonly ? "readonly" : "active");
+ break;
+ default:
+ err = sysfs_set_str(content, NULL, "array_state",
+ "readonly");
+ /* start mdmon if needed. */
+ if (!err) {
+ if (!mdmon_running(st->container_devnm))
+ start_mdmon(st->container_devnm);
+ ping_monitor(st->container_devnm);
+ }
+ break;
+ }
+ if (!err)
+ sysfs_set_safemode(content, content->safe_mode_delay);
+
+ /* Block subarray here if it is not reshaped now
+ * It has be blocked a little later to allow mdmon to switch in
+ * in to R/W state
+ */
+ if (st->ss->external && content->recovery_blocked &&
+ !start_reshape)
+ block_subarray(content);
+
+ if (err)
+ set_array_assembly_status(c, result, INCR_NO, &array);
+ else {
+ set_array_assembly_status(c, result, INCR_YES, &array);
+ wait_for(chosen_name, mdfd);
+ sysfs_rules_apply(chosen_name, content);
+ }
+
+ return err;
+ /* FIXME should have an O_EXCL and wait for read-auto */
}
-#endif