/*
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2001-2009 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)
+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 is_member_busy(char *metadata_version)
{
/* check if the given member array is active */
- struct mdstat_ent *mdstat = mdstat_read(1, 0);
+ struct mdstat_ent *mdstat = mdstat_read(0, 0);
struct mdstat_ent *ent;
int busy = 0;
static int ident_matches(struct mddev_ident *ident,
struct mdinfo *content,
struct supertype *tst,
- char *homehost,
+ char *homehost, int require_homehost,
char *update, char *devname)
{
return 0;
}
if (ident->name[0] && (!update || strcmp(update, "name")!= 0) &&
- name_matches(content->name, ident->name, homehost)==0) {
+ name_matches(content->name, ident->name, homehost, require_homehost)==0) {
if (devname)
pr_err("%s has wrong name.\n", devname);
return 0;
return 0;
}
if (ident->raid_disks != UnSet &&
+ content->array.raid_disks != 0 && /* metadata doesn't know how many to expect */
ident->raid_disks!= content->array.raid_disks) {
if (devname)
pr_err("%s requires wrong number of drives.\n",
struct mdinfo **contentp,
struct context *c,
int inargv, int auto_assem)
-
{
struct mddev_dev *tmpdev;
int num_devs;
struct supertype *st = *stp;
struct mdinfo *content = NULL;
- int report_missmatch = ((inargv && c->verbose >= 0) || c->verbose > 0);
+ 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;
if (tmpdev->used > 1)
continue;
- if (ident->devices &&
- !match_oneof(ident->devices, devname)) {
- if (report_missmatch)
+ if (ident->container) {
+ if (ident->container[0] == '/' &&
+ !same_dev(ident->container, devname)) {
+ if (report_mismatch)
+ pr_err("%s is not the container required (%s)\n",
+ devname, ident->container);
+ continue;
+ }
+ } else if (ident->devices &&
+ !match_oneof(ident->devices, devname)) {
+ /* Note that we ignore the "device=" identifier if a
+ * "container=" is given. Checking both is unnecessarily
+ * complicated.
+ */
+ if (report_mismatch)
pr_err("%s is not one of %s\n", devname, ident->devices);
continue;
}
dfd = dev_open(devname, O_RDONLY);
if (dfd < 0) {
- if (report_missmatch)
+ if (report_mismatch)
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) {
/* already found some components, this cannot
* be another one.
*/
- if (report_missmatch)
+ if (report_mismatch)
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_missmatch)
+ 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)) {
- if (report_missmatch)
+ } 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 = devnum_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
- if (report_missmatch)
- pr_err("%s has metadata type %s for which "
- "auto-assembly is disabled\n",
+ if (report_mismatch)
+ pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
devname, tst->ss->name);
tmpdev->used = 2;
} else
found_container = 1;
} else {
if (!tst && (tst = guess_super(dfd)) == NULL) {
- if (report_missmatch)
+ if (report_mismatch)
pr_err("no recogniseable superblock on %s\n",
devname);
tmpdev->used = 2;
- } else if (tst->ss->load_super(tst,dfd, NULL)) {
- if (report_missmatch)
+ } 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);
tmpdev->used = 2;
} else if (tst->ss->compare_super == NULL) {
- if (report_missmatch)
+ if (report_mismatch)
pr_err("Cannot assemble %s metadata on %s\n",
tst->ss->name, devname);
tmpdev->used = 2;
} else if (auto_assem && st == NULL &&
- !conf_test_metadata(tst->ss->name, (pol = devnum_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
- if (report_missmatch)
- pr_err("%s has metadata type %s for which "
- "auto-assembly is disabled\n",
+ if (report_mismatch)
+ pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
devname, tst->ss->name);
tmpdev->used = 2;
}
goto loop;
pr_err("%s has no superblock - assembly aborted\n",
- devname);
+ devname);
if (st)
st->ss->free_super(st);
dev_policy_free(pol);
*/
dfd = dev_open(devname, O_RDONLY | O_EXCL);
if (dfd < 0) {
- if (report_missmatch)
+ if (report_mismatch)
pr_err("%s is busy - skipping\n", devname);
goto loop;
}
close(dfd);
- if (ident->container) {
- if (ident->container[0] == '/' &&
- !same_dev(ident->container, devname)) {
- if (report_missmatch)
- pr_err("%s is not the container required (%s)\n",
- devname, ident->container);
+ if (ident->container && ident->container[0] != '/') {
+ /* we have a uuid */
+ int uuid[4];
+
+ content = *contentp;
+ tst->ss->getinfo_super(tst, content, NULL);
+
+ if (!parse_uuid(ident->container, uuid) ||
+ !same_uuid(content->uuid, uuid, tst->ss->swapuuid)) {
+ if (report_mismatch)
+ pr_err("%s has wrong UUID to be required container\n",
+ devname);
goto loop;
}
- if (ident->container[0] != '/') {
- /* we have a uuid */
- int uuid[4];
-
- content = *contentp;
- tst->ss->getinfo_super(tst, content, NULL);
-
- if (!parse_uuid(ident->container, uuid) ||
- !same_uuid(content->uuid, uuid, tst->ss->swapuuid)) {
- if (report_missmatch)
- pr_err("%s has wrong UUID to be required container\n",
- devname);
- goto loop;
- }
- }
}
/* It is worth looking inside this container.
*/
if (c->verbose > 0)
pr_err("looking in container %s\n",
- devname);
+ devname);
for (content = tst->ss->container_content(tst, NULL);
content;
content = content->next) {
if (!ident_matches(ident, content, tst,
- c->homehost, c->update,
- report_missmatch ? devname : NULL))
+ c->homehost, c->require_homehost,
+ c->update,
+ report_mismatch ? devname : NULL))
/* message already printed */;
else if (is_member_busy(content->text_version)) {
- if (report_missmatch)
+ if (report_mismatch)
pr_err("member %s in %s is already assembled\n",
- content->text_version,
- devname);
+ content->text_version,
+ devname);
} else if (content->array.state & (1<<MD_SB_BLOCK_VOLUME)) {
/* do not assemble arrays with unsupported configurations */
pr_err("Cannot activate member %s in %s.\n",
- content->text_version,
- devname);
+ content->text_version,
+ devname);
} else
break;
}
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",
- devname);
+ 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);
}
if (c->verbose > 0)
pr_err("found match on member %s in %s\n",
- content->text_version, devname);
+ content->text_version, devname);
/* make sure we finished the loop */
tmpdev = NULL;
goto loop;
} else {
-
content = *contentp;
tst->ss->getinfo_super(tst, content, NULL);
if (!ident_matches(ident, content, tst,
- c->homehost, c->update,
- report_missmatch ? devname : NULL))
+ c->homehost, c->require_homehost,
+ c->update,
+ report_mismatch ? devname : NULL))
goto loop;
-
+
+ if (auto_assem) {
+ /* Never auto-assemble things that conflict
+ * with mdadm.conf in some way
+ */
+ 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
* be looking at
*/
dfd = dev_open(devname, O_RDONLY | O_EXCL);
if (dfd < 0) {
- if (report_missmatch)
+ if (report_mismatch)
pr_err("%s is busy - skipping\n", devname);
goto loop;
}
(first == 1 || last == 1)) {
/* We can do something */
if (first) {/* just ignore this one */
- if (report_missmatch)
+ if (report_mismatch)
pr_err("%s misses out due to wrong homehost\n",
- devname);
+ devname);
goto loop;
} else { /* reject all those sofar */
struct mddev_dev *td;
- if (report_missmatch)
+ if (report_mismatch)
pr_err("%s overrides previous devices due to good homehost\n",
- devname);
+ devname);
for (td=devlist; td != tmpdev; td=td->next)
if (td->used == 1)
td->used = 0;
}
}
pr_err("superblock on %s doesn't match others - assembly aborted\n",
- devname);
+ devname);
tst->ss->free_super(tst);
st->ss->free_super(st);
dev_policy_free(pol);
/* Collect domain information from members only */
if (tmpdev && tmpdev->used == 1) {
if (!pol)
- pol = devnum_policy(stb.st_rdev);
+ pol = devid_policy(rdev);
domain_merge(&domains, pol, tst?tst->ss->name:NULL);
}
dev_policy_free(pol);
/* 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 = devnum_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 = 0;
+ 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;
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 (strcmp(c->update, "uuid") == 0 && !ident->uuid_set)
+ random_uuid((__u8 *)ident->uuid);
+
+ if (strcmp(c->update, "ppl") == 0 &&
+ ident->bitmap_fd >= 0) {
+ pr_err("PPL is not compatible with bitmap\n");
+ 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);
if (strcmp(c->update, "byteorder") == 0)
err = 0;
+ else if (strcmp(c->update, "home-cluster") == 0) {
+ tst->cluster_name = c->homecluster;
+ err = tst->ss->write_bitmap(tst, dfd, NameUpdate);
+ } else if (strcmp(c->update, "nodes") == 0) {
+ tst->nodes = c->nodes;
+ err = tst->ss->write_bitmap(tst, dfd, NodeNumUpdate);
+ } else if (strcmp(c->update, "revert-reshape") == 0 &&
+ c->invalid_backup)
+ err = tst->ss->update_super(tst, content,
+ "revert-reshape-nobackup",
+ devname, c->verbose,
+ ident->uuid_set,
+ c->homehost);
else
err = tst->ss->update_super(tst, content, c->update,
devname, c->verbose,
ident->uuid_set,
c->homehost);
if (err < 0) {
- pr_err("--update=%s not understood"
- " for %s metadata\n",
- c->update, tst->ss->name);
+ if (err == -1)
+ pr_err("--update=%s not understood for %s metadata\n",
+ 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 (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 &&
ident->bitmap_fd >= 0 && !bitmap_done) {
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.\n",
- devname, mddev, content->disk.raid_disk);
+ pr_err("%s is identified as a member of %s, slot %d%s.\n",
+ devname, mddev, content->disk.raid_disk,
+ (content->disk.state & (1<<MD_DISK_REPLACEMENT)) ? " replacement":"");
devices[devcnt].devname = devname;
devices[devcnt].uptodate = 0;
devices[devcnt].included = (tmpdev->disposition == 'I');
devices[devcnt].i = *content;
devices[devcnt].i.disk.major = major(stb.st_rdev);
devices[devcnt].i.disk.minor = minor(stb.st_rdev);
- if (most_recent < devcnt) {
- if (devices[devcnt].i.events
- > devices[most_recent].i.events)
+
+ if (devices[devcnt].i.disk.state == 6) {
+ if (most_recent < 0 ||
+ devices[devcnt].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 (nextspare < content->array.raid_disks)
- nextspare = content->array.raid_disks;
+ if (i+1 == 0 || i == MD_DISK_ROLE_JOURNAL) {
+ if (nextspare < content->array.raid_disks*2)
+ nextspare = content->array.raid_disks*2;
i = nextspare++;
} else {
- if (i >= content->array.raid_disks &&
+ /* i is raid_disk - double it so there is room for
+ * replacements */
+ i *= 2;
+ if (devices[devcnt].i.disk.state & (1<<MD_DISK_REPLACEMENT))
+ i++;
+ if (i >= content->array.raid_disks*2 &&
i >= nextspare)
nextspare = i+1;
}
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);
+ *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;
}
devcnt++;
}
- *most_recentp = most_recent;
+ if (most_recent >= 0)
+ *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 (devices[j].i.array.level >= 4 &&
+ devices[j].i.array.level <= 6 &&
+ 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;
+ }
+ 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)
add_another:
if (c->verbose >= 0)
pr_err("forcing event count in %s(%d) from %d upto %d\n",
- devices[chosen_drive].devname,
- devices[chosen_drive].i.disk.raid_disk,
- (int)(devices[chosen_drive].i.events),
- (int)(devices[most_recent].i.events));
+ devices[chosen_drive].devname,
+ devices[chosen_drive].i.disk.raid_disk,
+ (int)(devices[chosen_drive].i.events),
+ (int)(devices[most_recent].i.events));
fd = dev_open(devices[chosen_drive].devname,
devices[chosen_drive].included ? O_RDWR
: (O_RDWR|O_EXCL));
if (fd < 0) {
pr_err("Couldn't open %s for write - not updating\n",
- devices[chosen_drive].devname);
+ devices[chosen_drive].devname);
devices[chosen_drive].i.events = 0;
continue;
}
if (tst->ss->load_super(tst,fd, NULL)) {
close(fd);
pr_err("RAID superblock disappeared from %s - not updating.\n",
- devices[chosen_drive].devname);
+ devices[chosen_drive].devname);
devices[chosen_drive].i.events = 0;
continue;
}
content->events = devices[most_recent].i.events;
tst->ss->update_super(tst, content, "force-one",
- devices[chosen_drive].devname, c->verbose,
- 0, NULL);
+ devices[chosen_drive].devname, c->verbose,
+ 0, NULL);
if (tst->ss->store_super(tst, fd)) {
close(fd);
pr_err("Could not re-write superblock on %s\n",
- devices[chosen_drive].devname);
+ devices[chosen_drive].devname);
devices[chosen_drive].i.events = 0;
tst->ss->free_super(tst);
continue;
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 &&
return okcnt;
}
+static int start_array(int mdfd,
+ char *mddev,
+ struct mdinfo *content,
+ struct supertype *st,
+ struct mddev_ident *ident,
+ int *best, int bestcnt,
+ int chosen_drive,
+ struct devs *devices,
+ 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 err_ok,
+ int was_forced
+ )
+{
+ int rv;
+ 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",
+ mddev, strerror(errno));
+ return 1;
+ }
+ if (ident->bitmap_fd >= 0) {
+ if (ioctl(mdfd, SET_BITMAP_FILE, ident->bitmap_fd) != 0) {
+ pr_err("SET_BITMAP_FILE failed.\n");
+ return 1;
+ }
+ } else if (ident->bitmap_file) {
+ /* From config file */
+ int bmfd = open(ident->bitmap_file, O_RDWR);
+ if (bmfd < 0) {
+ pr_err("Could not open bitmap file %s\n",
+ ident->bitmap_file);
+ return 1;
+ }
+ if (ioctl(mdfd, SET_BITMAP_FILE, bmfd) != 0) {
+ pr_err("Failed to set bitmapfile for %s\n", mddev);
+ close(bmfd);
+ return 1;
+ }
+ close(bmfd);
+ }
+
+ /* First, add the raid disks, but add the chosen one last */
+ for (i = 0; i <= bestcnt; i++) {
+ int j;
+ if (i < bestcnt) {
+ j = best[i];
+ if (j == chosen_drive)
+ continue;
+ } else
+ j = chosen_drive;
+
+ if (j >= 0 && !devices[j].included) {
+ 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,
+ strerror(errno));
+ if (i < content->array.raid_disks * 2 ||
+ i == bestcnt)
+ okcnt--;
+ else
+ sparecnt--;
+ } 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":"");
+ }
+ } 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)
+ pr_err("no uptodate device for slot %d of %s\n",
+ i/2, mddev);
+ }
+
+ if (content->array.level == LEVEL_CONTAINER) {
+ if (c->verbose >= 0) {
+ 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)",
+ content->array.raid_disks);
+ 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;
+ }
+
+ /* Get number of in-sync devices according to the superblock.
+ * We must have this number to start the array without -s or -R
+ */
+ req_cnt = content->array.working_disks;
+
+ 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)))) {
+ /* 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;
+
+ if (content->reshape_active &&
+ !(content->reshape_active & RESHAPE_NO_BACKUP) &&
+ content->delta_disks <= 0) {
+ if (!c->backup_file) {
+ 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)
+ pr_err("(Don't specify --update=revert-reshape again, that part succeeded.)\n");
+ return 1;
+ }
+ rv = sysfs_set_str(content, NULL,
+ "array_state", "readonly");
+ if (rv == 0)
+ rv = Grow_continue(mdfd, st, content,
+ c->backup_file, 0,
+ c->freeze_reshape);
+ } else if (c->readonly &&
+ sysfs_attribute_available(content, NULL,
+ "array_state")) {
+ rv = sysfs_set_str(content, NULL,
+ "array_state", "readonly");
+ } else
+ rv = ioctl(mdfd, RUN_ARRAY, NULL);
+ reopen_mddev(mdfd); /* drop O_EXCL */
+ if (rv == 0) {
+ 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);
+ if (rebuilding_cnt)
+ fprintf(stderr, "%s %d rebuilding",
+ sparecnt?",":" and",
+ rebuilding_cnt);
+ if (sparecnt)
+ 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) {
+ /* might need to increase the size
+ * of the stripe cache - default is 256
+ */
+ 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 * chunk_size / 4096) + 1);
+ sysfs_free(sra);
+ }
+ }
+ if (okcnt < (unsigned)content->array.raid_disks) {
+ /* If any devices did not get added
+ * because the kernel rejected them based
+ * on event count, try adding them
+ * again providing the action policy is
+ * 're-add' or greater. The bitmap
+ * might allow them to be included, or
+ * they will become spares.
+ */
+ for (i = 0; i < bestcnt; i++) {
+ int j = best[i];
+ if (j >= 0 && !devices[j].uptodate) {
+ if (!disk_action_allows(&devices[j].i, st->ss->name, act_re_add))
+ continue;
+ rv = add_disk(mdfd, st, content,
+ &devices[j].i);
+ if (rv == 0 && c->verbose >= 0)
+ pr_err("%s has been re-added.\n",
+ devices[j].devname);
+ }
+ }
+ }
+ if (content->array.level == 6 &&
+ okcnt + 1 == (unsigned)content->array.raid_disks &&
+ was_forced) {
+ struct mdinfo *sra;
+
+ sra = sysfs_read(mdfd, NULL, 0);
+ if (sra)
+ sysfs_set_str(sra, NULL,
+ "sync_action", "repair");
+ sysfs_free(sra);
+ }
+ return 0;
+ }
+ pr_err("failed to RUN_ARRAY %s: %s\n", mddev, strerror(errno));
+
+ if (!enough(content->array.level, content->array.raid_disks,
+ content->array.layout, 1, avail))
+ 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");
+
+ return 1;
+ }
+ if (c->runstop == -1) {
+ pr_err("%s assembled from %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, ", but not started.\n");
+ return 2;
+ }
+ if (c->verbose >= -1) {
+ 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);
+ if (sparecnt)
+ 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");
+ else {
+ if (req_cnt == (unsigned)content->array.raid_disks)
+ 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");
+ }
+ }
+ return 1;
+}
+
int Assemble(struct supertype *st, char *mddev,
struct mddev_ident *ident,
struct mddev_dev *devlist,
* START_ARRAY
*
*/
+ int rv;
int mdfd;
int clean;
int auto_assem = (mddev == NULL && !ident->uuid_set &&
- ident->super_minor == UnSet && ident->name[0] == 0
- && (ident->container == NULL || ident->member == NULL));
- int old_linux = 0;
- int vers = vers; /* Keep gcc quite - it really is initialised */
+ ident->super_minor == UnSet && ident->name[0] == 0 &&
+ (ident->container == NULL || ident->member == NULL));
struct devs *devices;
char *devmap;
int *best = NULL; /* indexed by raid_disk */
int bestcnt = 0;
int devcnt;
- unsigned int okcnt, sparecnt, rebuilding_cnt;
- unsigned int req_cnt;
+ unsigned int okcnt, sparecnt, rebuilding_cnt, replcnt, journalcnt;
+ int journal_clean = 0;
int i;
+ int was_forced = 0;
int most_recent = 0;
int chosen_drive;
int change = 0;
struct map_ent *map = NULL;
struct map_ent *mp;
- if (get_linux_version() < 2004000)
- old_linux = 1;
-
/*
* If any subdevs are listed, then any that don't
* match ident are discarded. Remainder must all match and
else if (mddev)
inargv = 1;
- try_again:
+try_again:
/* We come back here when doing auto-assembly and attempting some
* set of devices failed. Those are now marked as ->used==2 and
* we ignore them and try again
mddev ? mddev : "further assembly");
content = &info;
+ if (st && c->force)
+ st->ignore_hw_compat = 1;
num_devs = select_devices(devlist, ident, &st, &content, c,
inargv, auto_assem);
if (num_devs < 0)
return 1;
-
+
if (!st || !st->sb || !content)
return 2;
*/
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 && strcmp(c->update,"uuid") == 0)
+ mp = NULL;
+ else
+ mp = map_by_uuid(&map, content->uuid);
if (mp) {
struct mdinfo *dv;
/* array already exists. */
- pre_exist = sysfs_read(-1, mp->devnum, GET_LEVEL|GET_DEVS);
+ pre_exist = sysfs_read(-1, mp->devnm, GET_LEVEL|GET_DEVS);
if (pre_exist->array.level != UnSet) {
pr_err("Found some drive for an array that is already active: %s\n",
mp->path);
strcmp(mddev, chosen_name) != 0)
pr_err("Merging with already-assembled %s\n",
chosen_name);
- mdfd = open_dev_excl(mp->devnum);
+ mdfd = open_dev_excl(mp->devnm);
} else {
int trustworthy = FOREIGN;
name = content->name;
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);
return 1;
}
mddev = chosen_name;
- vers = md_get_version(mdfd);
- if (vers < 9000) {
- pr_err("Assemble requires 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(fd2devnum(mdfd))) {
+ if (mddev_busy(fd2devnm(mdfd))) {
pr_err("%s already active, cannot restart it!\n",
mddev);
for (tmpdev = devlist ;
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)
if (devcnt == 0) {
pr_err("no devices found for %s\n",
- mddev);
+ mddev);
if (st)
st->ss->free_super(st);
close(mdfd);
*/
avail = xcalloc(content->array.raid_disks, 1);
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.
*/
- if (j != most_recent &&
+ if (j != most_recent && !c->force &&
content->array.raid_disks > 0 &&
devices[most_recent].i.disk.raid_disk >= 0 &&
devmap[j * content->array.raid_disks + devices[most_recent].i.disk.raid_disk] == 0) {
if (c->verbose > -1)
pr_err("ignoring %s as it reports %s as failed\n",
- devices[j].devname, devices[most_recent].devname);
+ devices[j].devname, devices[most_recent].devname);
best[i] = -1;
continue;
}
+ /* Require event counter to be same as, or just less than,
+ * most recent. If it is bigger, it must be a stray spare and
+ * should be ignored.
+ */
if (devices[j].i.events+event_margin >=
- devices[most_recent].i.events) {
+ devices[most_recent].i.events &&
+ devices[j].i.events <=
+ devices[most_recent].i.events
+ ) {
devices[j].uptodate = 1;
- if (i < content->array.raid_disks) {
+ 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)))) {
- okcnt++;
- avail[i]=1;
+ i >= content->array.raid_disks - content->delta_disks)) {
+ if (!avail[i/2]) {
+ okcnt++;
+ avail[i/2]=1;
+ } else
+ replcnt++;
} else
rebuilding_cnt++;
- } else
+ } else if (devices[j].i.disk.raid_disk != MD_DISK_ROLE_JOURNAL)
sparecnt++;
}
}
free(devmap);
- if (c->force)
- okcnt += force_array(content, devices, best, bestcnt,
- avail, most_recent, st, c);
-
+ if (c->force) {
+ int force_ok = force_array(content, devices, best, bestcnt,
+ avail, most_recent, st, c);
+ okcnt += force_ok;
+ if (force_ok)
+ was_forced = 1;
+ }
/* Now we want to look at the superblock which the kernel will base things on
* and compare the devices that we think are working with the devices that the
* superblock thinks are working.
*/
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;
devices[j].included ? O_RDONLY
: (O_RDONLY|O_EXCL)))< 0) {
pr_err("Cannot open %s: %s\n",
- devices[j].devname, strerror(errno));
+ devices[j].devname, strerror(errno));
close(mdfd);
free(devices);
return 1;
if (st->ss->load_super(st,fd, NULL)) {
close(fd);
pr_err("RAID superblock has disappeared from %s\n",
- devices[j].devname);
+ devices[j].devname);
close(mdfd);
free(devices);
return 1;
return 1;
}
st->ss->getinfo_super(st, content, NULL);
-#ifndef MDASSEMBLE
- sysfs_init(content, mdfd, 0);
-#endif
+ if (sysfs_init(content, mdfd, NULL)) {
+ pr_err("Unable to initialize sysfs\n");
+ close(mdfd);
+ free(devices);
+ return 1;
+ }
+
+ /* 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)
- desired_state = (1<<MD_DISK_ACTIVE) | (1<<MD_DISK_SYNC);
- else
+ 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;
if (c->force) {
if (c->verbose >= 0)
pr_err("clearing FAULTY flag for device %d in %s for %s\n",
- j, mddev, devices[j].devname);
+ j, mddev, devices[j].devname);
change = 1;
} else {
if (c->verbose >= -1)
pr_err("device %d in %s has wrong state in superblock, but %s seems ok\n",
- i, mddev, devices[j].devname);
+ 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);
+ i, mddev);
}
#endif
}
content->array.layout, clean,
avail)) {
change += st->ss->update_super(st, content, "force-array",
- devices[chosen_drive].devname, c->verbose,
+ devices[chosen_drive].devname, c->verbose,
0, NULL);
+ was_forced = 1;
clean = 1;
}
O_RDWR : (O_RDWR|O_EXCL));
if (fd < 0) {
pr_err("Could not open %s for write - cannot Assemble array.\n",
- devices[chosen_drive].devname);
+ devices[chosen_drive].devname);
close(mdfd);
free(devices);
return 1;
if (st->ss->store_super(st, fd)) {
close(fd);
pr_err("Could not re-write superblock on %s\n",
- devices[chosen_drive].devname);
+ devices[chosen_drive].devname);
close(mdfd);
free(devices);
return 1;
}
if (c->verbose >= 0)
pr_err("Marking array %s as 'clean'\n",
- mddev);
+ mddev);
close(fd);
}
* 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",
- chosen_name);
- for (i=0; i<bestcnt; i++) {
- int j = best[i];
+ 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];
if (j >= 0) {
fdlist[i] = dev_open(devices[j].devname,
devices[j].included
? O_RDWR : (O_RDWR|O_EXCL));
if (fdlist[i] < 0) {
pr_err("Could not open %s for write - cannot Assemble array.\n",
- devices[j].devname);
+ devices[j].devname);
err = 1;
break;
}
if (st->ss->external && st->ss->recover_backup)
err = st->ss->recover_backup(st, content);
else
- err = Grow_restart(st, content, fdlist, bestcnt,
+ err = Grow_restart(st, content, fdlist, bestcnt/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;
}
}
return err;
}
}
-#endif
- /* count number of in-sync devices according to the superblock.
- * We must have this number to start the array without -s or -R
- */
- req_cnt = content->array.working_disks;
/* Almost ready to actually *do* something */
- if (!old_linux) {
- int rv;
-
- /* First, fill in the map, so that udev can find our name
- * as soon as we become active.
- */
- map_update(&map, fd2devnum(mdfd), content->text_version,
- content->uuid, chosen_name);
-
- rv = set_array_info(mdfd, st, content);
- if (rv && !pre_exist) {
- pr_err("failed to set array info for %s: %s\n",
- mddev, strerror(errno));
- ioctl(mdfd, STOP_ARRAY, NULL);
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- }
- if (ident->bitmap_fd >= 0) {
- if (ioctl(mdfd, SET_BITMAP_FILE, ident->bitmap_fd) != 0) {
- pr_err("SET_BITMAP_FILE failed.\n");
- ioctl(mdfd, STOP_ARRAY, NULL);
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- }
- } else if (ident->bitmap_file) {
- /* From config file */
- int bmfd = open(ident->bitmap_file, O_RDWR);
- if (bmfd < 0) {
- pr_err("Could not open bitmap file %s\n",
- ident->bitmap_file);
- ioctl(mdfd, STOP_ARRAY, NULL);
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- }
- if (ioctl(mdfd, SET_BITMAP_FILE, bmfd) != 0) {
- pr_err("Failed to set bitmapfile for %s\n", mddev);
- close(bmfd);
- ioctl(mdfd, STOP_ARRAY, NULL);
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- }
- close(bmfd);
- }
-
- /* First, add the raid disks, but add the chosen one last */
- for (i=0; i<= bestcnt; i++) {
- int j;
- if (i < bestcnt) {
- j = best[i];
- if (j == chosen_drive)
- continue;
- } else
- j = chosen_drive;
-
- if (j >= 0 && !devices[j].included) {
- int 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,
- strerror(errno));
- if (i < content->array.raid_disks
- || i == bestcnt)
- okcnt--;
- else
- sparecnt--;
- } else if (c->verbose > 0)
- pr_err("added %s to %s as %d%s\n",
- devices[j].devname, mddev,
- devices[j].i.disk.raid_disk,
- devices[j].uptodate?"":
- " (possibly out of date)");
- } 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)
- pr_err("no uptodate device for slot %d of %s\n",
- i, mddev);
- }
+ /* 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) {
+ content->array.major_version = 1;
+ content->array.minor_version = 0;
+ strcpy(content->text_version, "1.0");
+ }
- if (content->array.level == LEVEL_CONTAINER) {
- if (c->verbose >= 0) {
- pr_err("Container %s has been "
- "assembled with %d drive%s",
- mddev, okcnt+sparecnt, okcnt+sparecnt==1?"":"s");
- if (okcnt < (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)",
- content->array.raid_disks);
- fprintf(stderr, "\n");
- }
- st->ss->free_super(st);
- sysfs_uevent(content, "change");
- map_unlock(&map);
- wait_for(chosen_name, mdfd);
- close(mdfd);
- free(devices);
- return 0;
- }
+ map_update(&map, fd2devnm(mdfd), content->text_version,
+ content->uuid, chosen_name);
- 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)
- ))) {
- /* 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.
+ rv = start_array(mdfd, mddev, content,
+ st, ident, best, bestcnt,
+ chosen_drive, devices, okcnt, sparecnt,
+ rebuilding_cnt, journalcnt,
+ c,
+ clean, avail, start_partial_ok,
+ pre_exist != NULL,
+ was_forced);
+ if (rv == 1 && !pre_exist)
+ ioctl(mdfd, STOP_ARRAY, NULL);
+ free(devices);
+ map_unlock(&map);
+ if (rv == 0) {
+ wait_for(chosen_name, mdfd);
+ close(mdfd);
+ if (auto_assem) {
+ int usecs = 1;
+ /* There is a nasty race with 'mdadm --monitor'.
+ * If it opens this device before we close it,
+ * it gets an incomplete open on which IO
+ * doesn't work and the capacity is
+ * wrong.
+ * If we reopen (to check for layered devices)
+ * before --monitor closes, we loose.
+ *
+ * So: wait upto 1 second for there to be
+ * a non-zero capacity.
*/
- int rv;
-#ifndef MDASSEMBLE
- if (content->reshape_active &&
- !(content->reshape_active & RESHAPE_NO_BACKUP) &&
- content->delta_disks <= 0) {
- rv = sysfs_set_str(content, NULL,
- "array_state", "readonly");
- if (rv == 0)
- rv = Grow_continue(mdfd, st, content,
- c->backup_file,
- c->freeze_reshape);
- } else if (c->readonly &&
- sysfs_attribute_available(
- content, NULL, "array_state")) {
- rv = sysfs_set_str(content, NULL,
- "array_state", "readonly");
- } else
-#endif
- rv = ioctl(mdfd, RUN_ARRAY, NULL);
- if (rv == 0) {
- 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);
- if (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, ".\n");
- }
- if (content->reshape_active &&
- content->array.level >= 4 &&
- content->array.level <= 6) {
- /* 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, 0, 0);
- if (sra)
- sysfs_set_num(sra, NULL,
- "stripe_cache_size",
- (4 * content->array.chunk_size / 4096) + 1);
- sysfs_free(sra);
- }
- }
- if (okcnt < (unsigned)content->array.raid_disks) {
- /* If any devices did not get added
- * because the kernel rejected them based
- * on event count, try adding them
- * again providing the action policy is
- * 're-add' or greater. The bitmap
- * might allow them to be included, or
- * they will become spares.
- */
- for (i = 0; i < bestcnt; i++) {
- int j = best[i];
- if (j >= 0 && !devices[j].uptodate) {
- if (!disk_action_allows(&devices[j].i, st->ss->name, act_re_add))
- continue;
- rv = add_disk(mdfd, st, content,
- &devices[j].i);
- if (rv == 0 && c->verbose >= 0)
- pr_err("%s has been re-added.\n",
- devices[j].devname);
- }
- }
+ while (usecs < 1000) {
+ mdfd = open(mddev, O_RDONLY);
+ if (mdfd >= 0) {
+ unsigned long long size;
+ if (get_dev_size(mdfd, NULL, &size) &&
+ size > 0)
+ break;
+ close(mdfd);
}
- map_unlock(&map);
- wait_for(mddev, mdfd);
- close(mdfd);
- if (auto_assem) {
- int usecs = 1;
- /* There is a nasty race with 'mdadm --monitor'.
- * If it opens this device before we close it,
- * it gets an incomplete open on which IO
- * doesn't work and the capacity is
- * wrong.
- * If we reopen (to check for layered devices)
- * before --monitor closes, we loose.
- *
- * So: wait upto 1 second for there to be
- * a non-zero capacity.
- */
- while (usecs < 1000) {
- mdfd = open(mddev, O_RDONLY);
- if (mdfd >= 0) {
- unsigned long long size;
- if (get_dev_size(mdfd, NULL, &size) &&
- size > 0)
- break;
- close(mdfd);
- }
- usleep(usecs);
- usecs <<= 1;
- }
- }
- free(devices);
- return 0;
- }
- pr_err("failed to RUN_ARRAY %s: %s\n",
- mddev, strerror(errno));
-
- if (!enough(content->array.level, content->array.raid_disks,
- content->array.layout, 1, avail))
- 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");
-
- if (auto_assem)
- ioctl(mdfd, STOP_ARRAY, NULL);
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- }
- if (c->runstop == -1) {
- pr_err("%s assembled from %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, ", but not started.\n");
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 0;
- }
- if (c->verbose >= -1) {
- 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);
- if (sparecnt)
- 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");
- else {
- if (req_cnt == (unsigned)content->array.raid_disks)
- fprintf(stderr, " - need all %d to start it", req_cnt);
- else
- fprintf(stderr, " - need %d of %d to start", req_cnt, content->array.raid_disks);
- fprintf(stderr, " (use --run to insist).\n");
+ usleep(usecs);
+ usecs <<= 1;
}
}
- if (auto_assem)
- ioctl(mdfd, STOP_ARRAY, NULL);
+ } else
close(mdfd);
- free(devices);
- map_unlock(&map);
- return 1;
- } else {
- /* The "chosen_drive" is a good choice, and if necessary, the superblock has
- * been updated to point to the current locations of devices.
- * so we can just start the array
- */
- unsigned long dev;
- dev = makedev(devices[chosen_drive].i.disk.major,
- devices[chosen_drive].i.disk.minor);
- if (ioctl(mdfd, START_ARRAY, dev)) {
- pr_err("Cannot start array: %s\n",
- strerror(errno));
- }
- }
- close(mdfd);
- free(devices);
- map_unlock(&map);
- return 0;
+ /* '2' means 'OK, but not started yet' */
+ 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;
+ struct mdinfo *dev, *sra, *dev2;
int working = 0, preexist = 0;
int expansion = 0;
- struct map_ent *map = NULL;
int old_raid_disks;
int start_reshape;
+ char *avail;
+ int err;
- sysfs_init(content, mdfd, 0);
+ if (sysfs_init(content, mdfd, NULL)) {
+ pr_err("Unable to initialize sysfs\n");
+ return 1;
+ }
- sra = sysfs_read(mdfd, 0, GET_VERSION);
+ sra = sysfs_read(mdfd, NULL, GET_VERSION|GET_DEVS);
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_set_array(content, 9003) != 0) {
+ sysfs_free(sra);
return 1;
}
}
* Check if metadata requests blocking container wide reshapes
*/
start_reshape = (content->reshape_active &&
- !((content->reshape_active == CONTAINER_RESHAPE) &&
- (content->array.state & (1<<MD_SB_BLOCK_CONTAINER_RESHAPE))));
+ !((content->reshape_active == CONTAINER_RESHAPE) &&
+ (content->array.state & (1<<MD_SB_BLOCK_CONTAINER_RESHAPE))));
/* Block subarray here if it is under reshape now
* Do not allow for any changes in this array
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", "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)
working++;
} else if (errno == EEXIST)
preexist++;
- if (working + expansion == 0)
+ }
+ sysfs_free(sra);
+ if (working + expansion == 0 && c->runstop <= 0) {
+ free(avail);
return 1;/* Nothing new, don't try to start */
-
- map_update(&map, fd2devnum(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 (ret < 0) {
+ err = 1;
+ break;
+ }
- if (st->ss->external) {
- if (!mdmon_running(st->container_dev))
- start_mdmon(st->container_dev);
- ping_monitor_by_id(st->container_dev);
- if (mdmon_running(st->container_dev) &&
- st->update_tail == NULL)
- st->update_tail = &st->updates;
+ 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;
}
- err = Grow_continue(mdfd, st, content, c->backup_file,
- c->freeze_reshape);
- } else switch(content->array.level) {
+ /* 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;
+ }
+
+ err = st->ss->write_init_ppl(st, content, dfd);
+ close(dfd);
+
+ if (err)
+ break;
+ }
+
+ if (err) {
+ free(avail);
+ return err;
+ }
+ }
+
+ if (enough(content->array.level, content->array.raid_disks,
+ content->array.layout, content->array.state & 1, avail) == 0) {
+ if (c->export && result)
+ *result |= INCR_NO;
+ 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");
+ }
+ free(avail);
+ return 1;
+ }
+ free(avail);
+
+ if (c->runstop <= 0 &&
+ (working + preexist + expansion) <
+ content->array.working_disks) {
+ if (c->export && result)
+ *result |= INCR_UNSAFE;
+ 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 safe to start\n");
+ }
+ return 1;
+ }
+
+
+ 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;
+
+ 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:
break;
default:
err = sysfs_set_str(content, NULL, "array_state",
- "readonly");
+ "readonly");
/* start mdmon if needed. */
if (!err) {
- if (!mdmon_running(st->container_dev))
- start_mdmon(st->container_dev);
- ping_monitor_by_id(st->container_dev);
+ 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);
+ 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);
+ /* 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 (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");
- }
- 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");
- }
- return 1;
+ if (c->export && result) {
+ if (err)
+ *result |= INCR_NO;
+ else
+ *result |= INCR_YES;
+ } else 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");
}
+ if (!err)
+ wait_for(chosen_name, mdfd);
+ return err;
+ /* FIXME should have an O_EXCL and wait for read-auto */
}
-#endif
-
projects / thirdparty / mdadm.git / blobdiff