* Incremental.c - support --incremental. Part of:
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2006-2009 Neil Brown <neilb@suse.de>
+ * Copyright (C) 2006-2012 Neil Brown <neilb@suse.de>
*
*
* This program is free software; you can redistribute it and/or modify
*/
#include "mdadm.h"
+#include <sys/wait.h>
#include <dirent.h>
#include <ctype.h>
struct supertype *st, int verbose);
static int Incremental_container(struct supertype *st, char *devname,
- char *homehost,
- int verbose, int runstop, int autof,
- int freeze_reshape);
+ struct context *c);
-int Incremental(char *devname, int verbose, int runstop,
- struct supertype *st, char *homehost, int require_homehost,
- int autof, int freeze_reshape)
+int Incremental(char *devname, struct context *c,
+ struct supertype *st)
{
/* Add this device to an array, creating the array if necessary
* and starting the array if sensible or - if runstop>0 - if possible.
struct createinfo *ci = conf_get_create_info();
if (stat(devname, &stb) < 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("stat failed for %s: %s.\n",
devname, strerror(errno));
return rv;
}
if ((stb.st_mode & S_IFMT) != S_IFBLK) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s is not a block device.\n",
devname);
return rv;
}
- dfd = dev_open(devname, O_RDONLY|O_EXCL);
+ dfd = dev_open(devname, O_RDONLY);
if (dfd < 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("cannot open %s: %s.\n",
devname, strerror(errno));
return rv;
if (must_be_container(dfd)) {
if (!st)
st = super_by_fd(dfd, NULL);
+ if (st)
+ st->ignore_hw_compat = 1;
if (st && st->ss->load_container)
rv = st->ss->load_container(st, dfd, NULL);
if (map_lock(&map))
pr_err("failed to get "
"exclusive lock on mapfile\n");
- rv = Incremental_container(st, devname, homehost,
- verbose, runstop, autof,
- freeze_reshape);
+ rv = Incremental_container(st, devname, c);
map_unlock(&map);
return rv;
}
/* 1/ Check if device is permitted by mdadm.conf */
if (!conf_test_dev(devname)) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s not permitted by mdadm.conf.\n",
devname);
goto out;
* version/name from args) */
if (fstat(dfd, &stb) < 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("fstat failed for %s: %s.\n",
devname, strerror(errno));
goto out;
}
if ((stb.st_mode & S_IFMT) != S_IFBLK) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s is not a block device.\n",
devname);
goto out;
have_target = policy_check_path(&dinfo, &target_array);
if (st == NULL && (st = guess_super(dfd)) == NULL) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("no recognisable superblock on %s.\n",
devname);
rv = try_spare(devname, &dfd, policy,
have_target ? &target_array : NULL,
- st, verbose);
+ st, c->verbose);
goto out;
}
+ st->ignore_hw_compat = 1;
if (st->ss->compare_super == NULL ||
st->ss->load_super(st, dfd, NULL)) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("no RAID superblock on %s.\n",
devname);
rv = try_spare(devname, &dfd, policy,
have_target ? &target_array : NULL,
- st, verbose);
+ st, c->verbose);
free(st);
goto out;
}
st->ss->getinfo_super(st, &info, NULL);
/* 3/ Check if there is a match in mdadm.conf */
- match = conf_match(st, &info, devname, verbose, &rv);
+ match = conf_match(st, &info, devname, c->verbose, &rv);
if (!match && rv == 2)
goto out;
if (match && match->devname
&& strcasecmp(match->devname, "<ignore>") == 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("array containing %s is explicitly"
" ignored by mdadm.conf\n",
devname);
* on that. */
if (match)
trustworthy = LOCAL;
- else if (st->ss->match_home(st, homehost) == 1)
+ else if (st->ss->match_home(st, c->homehost) == 1)
trustworthy = LOCAL;
else if (st->ss->match_home(st, "any") == 1)
trustworthy = LOCAL_ANY;
else
trustworthy = FOREIGN;
-
if (!match && !conf_test_metadata(st->ss->name, policy,
(trustworthy == LOCAL))) {
- if (verbose >= 1)
+ if (c->verbose >= 1)
pr_err("%s has metadata type %s for which "
"auto-assembly is disabled\n",
devname, st->ss->name);
* CREATE.
*/
if (match && match->autof)
- autof = match->autof;
- if (autof == 0)
- autof = ci->autof;
+ c->autof = match->autof;
+ if (c->autof == 0)
+ c->autof = ci->autof;
name_to_use = info.name;
if (name_to_use[0] == 0 &&
trustworthy = METADATA;
}
if (name_to_use[0] && trustworthy != LOCAL &&
- ! require_homehost &&
+ ! c->require_homehost &&
conf_name_is_free(name_to_use))
trustworthy = LOCAL;
if (map_lock(&map))
pr_err("failed to get exclusive lock on "
"mapfile\n");
+ /* Now check we can get O_EXCL. If not, probably "mdadm -A" has
+ * taken over
+ */
+ dfd = dev_open(devname, O_RDONLY|O_EXCL);
+ if (dfd < 0) {
+ if (c->verbose >= 0)
+ pr_err("cannot reopen %s: %s.\n",
+ devname, strerror(errno));
+ goto out_unlock;
+ }
+ /* Cannot hold it open while we add the device to the array,
+ * so we must release the O_EXCL and depend on the map_lock()
+ * So now is the best time to remove any partitions.
+ */
+ remove_partitions(dfd);
+ close(dfd);
+ dfd = -1;
+
mp = map_by_uuid(&map, info.uuid);
if (mp)
- mdfd = open_dev(mp->devnum);
+ mdfd = open_dev(mp->devnm);
else
mdfd = -1;
/* Couldn't find an existing array, maybe make a new one */
mdfd = create_mddev(match ? match->devname : NULL,
- name_to_use, autof, trustworthy, chosen_name);
+ name_to_use, c->autof, trustworthy, chosen_name);
if (mdfd < 0)
goto out_unlock;
- sysfs_init(&info, mdfd, 0);
+ sysfs_init(&info, mdfd, NULL);
if (set_array_info(mdfd, st, &info) != 0) {
pr_err("failed to set array info for %s: %s\n",
dinfo.disk.major = major(stb.st_rdev);
dinfo.disk.minor = minor(stb.st_rdev);
if (add_disk(mdfd, st, &info, &dinfo) != 0) {
- pr_err("failed to add %s to %s: %s.\n",
+ pr_err("failed to add %s to new array %s: %s.\n",
devname, chosen_name, strerror(errno));
ioctl(mdfd, STOP_ARRAY, 0);
rv = 2;
goto out_unlock;
}
- sra = sysfs_read(mdfd, -1, (GET_DEVS | GET_STATE |
- GET_OFFSET | GET_SIZE));
-
+ sra = sysfs_read(mdfd, NULL, (GET_DEVS | GET_STATE |
+ GET_OFFSET | GET_SIZE));
+
if (!sra || !sra->devs || sra->devs->disk.raid_disk >= 0) {
/* It really should be 'none' - must be old buggy
* kernel, and mdadm -I may not be able to complete.
}
info.array.working_disks = 1;
/* 6/ Make sure /var/run/mdadm.map contains this array. */
- map_update(&map, fd2devnum(mdfd),
+ map_update(&map, fd2devnm(mdfd),
info.text_version,
info.uuid, chosen_name);
} else {
struct supertype *st2;
struct mdinfo info2, *d;
- sra = sysfs_read(mdfd, -1, (GET_DEVS | GET_STATE |
+ sra = sysfs_read(mdfd, NULL, (GET_DEVS | GET_STATE |
GET_OFFSET | GET_SIZE));
-
+
if (mp->path)
strcpy(chosen_name, mp->path);
else
- strcpy(chosen_name, devnum2devname(mp->devnum));
+ strcpy(chosen_name, mp->devnm);
/* It is generally not OK to add non-spare drives to a
* running array as they are probably missing because
&& (info.disk.state & (1<<MD_DISK_SYNC)) != 0
&& ! policy_action_allows(policy, st->ss->name,
act_re_add)
- && runstop < 1) {
+ && c->runstop < 1) {
if (ioctl(mdfd, GET_ARRAY_INFO, &ainf) == 0) {
pr_err("not adding %s to active array (without --run) %s\n",
devname, chosen_name);
* disk.number. Find and reject any such
*/
find_reject(mdfd, st, sra, info.disk.number,
- info.events, verbose, chosen_name);
+ info.events, c->verbose, chosen_name);
err = add_disk(mdfd, st, sra, &info);
}
+ if (err < 0 && errno == EINVAL &&
+ info.disk.state & (1<<MD_DISK_SYNC)) {
+ /* Maybe it needs to be added as a spare */
+ if (policy_action_allows(policy, st->ss->name,
+ act_force_spare)) {
+ info.disk.state &= ~(1<<MD_DISK_SYNC);
+ err = add_disk(mdfd, st, sra, &info);
+ } else
+ if (c->verbose >= 0)
+ pr_err("can only add %s to %s as a spare, and force-spare is not set.\n",
+ devname, chosen_name);
+ }
if (err < 0) {
- pr_err("failed to add %s to %s: %s.\n",
+ pr_err("failed to add %s to existing array %s: %s.\n",
devname, chosen_name, strerror(errno));
rv = 2;
goto out_unlock;
info.array.working_disks = 0;
for (d = sra->devs; d; d=d->next)
info.array.working_disks ++;
-
+
}
/* 7/ Is there enough devices to possibly start the array? */
/* 7a/ if not, finish with success. */
if (info.array.level == LEVEL_CONTAINER) {
- int devnum = devnum; /* defined and used iff ->external */
+ char devnm[32];
/* Try to assemble within the container */
sysfs_uevent(sra, "change");
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("container %s now has %d device%s\n",
chosen_name, info.array.working_disks,
- info.array.working_disks==1?"":"s");
+ info.array.working_disks == 1?"":"s");
wait_for(chosen_name, mdfd);
if (st->ss->external)
- devnum = fd2devnum(mdfd);
+ strcpy(devnm, fd2devnm(mdfd));
if (st->ss->load_container)
rv = st->ss->load_container(st, mdfd, NULL);
close(mdfd);
sysfs_free(sra);
if (!rv)
- rv = Incremental_container(st, chosen_name, homehost,
- verbose, runstop, autof,
- freeze_reshape);
+ rv = Incremental_container(st, chosen_name, c);
map_unlock(&map);
if (rv == 1)
/* Don't fail the whole -I if a subarray didn't
/* after spare is added, ping monitor for external metadata
* so that it can eg. try to rebuild degraded array */
if (st->ss->external)
- ping_monitor_by_id(devnum);
+ ping_monitor(devnm);
return rv;
}
* things change.
*/
sysfs_free(sra);
- sra = sysfs_read(mdfd, -1, (GET_DEVS | GET_STATE |
+ sra = sysfs_read(mdfd, NULL, (GET_DEVS | GET_STATE |
GET_OFFSET | GET_SIZE));
active_disks = count_active(st, sra, mdfd, &avail, &info);
if (enough(info.array.level, info.array.raid_disks,
info.array.layout, info.array.state & 1,
avail) == 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s attached to %s, not enough to start (%d).\n",
devname, chosen_name, active_disks);
rv = 0;
/* + start the array (auto-readonly). */
if (ioctl(mdfd, GET_ARRAY_INFO, &ainf) == 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s attached to %s which is already active.\n",
devname, chosen_name);
rv = 0;
}
map_unlock(&map);
- if (runstop > 0 || active_disks >= info.array.working_disks) {
+ if (c->runstop > 0 || active_disks >= info.array.working_disks) {
struct mdinfo *dsk;
/* Let's try to start it */
+
+ if (info.reshape_active && !(info.reshape_active & RESHAPE_NO_BACKUP)) {
+ pr_err("%s: This array is being reshaped and cannot be started\n",
+ chosen_name);
+ cont_err("by --incremental. Please use --assemble\n");
+ goto out;
+ }
if (match && match->bitmap_file) {
int bmfd = open(match->bitmap_file, O_RDWR);
if (bmfd < 0) {
rv = sysfs_set_str(sra, NULL,
"array_state", "read-auto");
if (rv == 0) {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s attached to %s, which has been started.\n",
devname, chosen_name);
rv = 0;
rv = 1;
}
} else {
- if (verbose >= 0)
+ if (c->verbose >= 0)
pr_err("%s attached to %s, not enough to start safely.\n",
devname, chosen_name);
rv = 0;
/* count how many devices in sra think they are active */
struct mdinfo *d;
int cnt = 0;
+ int replcnt = 0;
__u64 max_events = 0;
char *avail = NULL;
int *best = NULL;
for (d = sra->devs ; d ; d = d->next)
numdevs++;
- for (d = sra->devs, devnum=0 ; d ; d = d->next, devnum++) {
+ for (d = sra->devs, devnum = 0 ; d ; d = d->next, devnum++) {
char dn[30];
int dfd;
int ok;
st->ss->getinfo_super(st, &info, devmap + raid_disks * devnum);
if (!avail) {
raid_disks = info.array.raid_disks;
- avail = calloc(raid_disks, 1);
- if (!avail) {
- pr_err("out of memory.\n");
- exit(1);
- }
+ avail = xcalloc(raid_disks, 1);
*availp = avail;
- best = calloc(raid_disks, sizeof(int));
- devmap = calloc(raid_disks * numdevs, 1);
+ best = xcalloc(raid_disks, sizeof(int));
+ devmap = xcalloc(raid_disks, numdevs);
st->ss->getinfo_super(st, &info, devmap);
}
else if (info.events == max_events+1) {
int i;
max_events = info.events;
- for (i=0; i < raid_disks; i++)
+ for (i = 0; i < raid_disks; i++)
if (avail[i])
avail[i]--;
avail[info.disk.raid_disk] = 2;
best[info.disk.raid_disk] = devnum;
st->ss->getinfo_super(st, bestinfo, NULL);
}
- }
+ } else if (info.disk.state & (1<<MD_DISK_REPLACEMENT))
+ replcnt++;
st->ss->free_super(st);
}
if (!avail)
}
free(best);
free(devmap);
- return cnt;
+ return cnt + replcnt;
}
/* test if container has degraded member(s) */
int max_degraded = 0;
for(; map; map = map->next) {
- if (!is_subarray(map->metadata) ||
- devname2devnum(map->metadata+1) != me->devnum)
+ if (!metadata_container_matches(map->metadata, me->devnm))
continue;
- afd = open_dev(map->devnum);
+ afd = open_dev(map->devnm);
if (afd < 0)
continue;
/* most accurate information regarding array degradation */
}
free(st2);
}
- sra = sysfs_read(-1, mp->devnum,
+ sra = sysfs_read(-1, mp->devnm,
GET_DEVS|GET_OFFSET|GET_SIZE|GET_STATE|
GET_DEGRADED|GET_COMPONENT|GET_VERSION);
if (!sra) {
/* Probably a container - no degraded info */
- sra = sysfs_read(-1, mp->devnum,
+ sra = sysfs_read(-1, mp->devnm,
GET_DEVS|GET_OFFSET|GET_SIZE|GET_STATE|
GET_COMPONENT|GET_VERSION);
if (sra)
if (st == NULL) {
int i;
st2 = NULL;
- for(i=0; !st2 && superlist[i]; i++)
+ for(i = 0; !st2 && superlist[i]; i++)
st2 = superlist[i]->match_metadata_desc(
sra->text_version);
if (!st2) {
/* true for containers, here we must read superblock
* to obtain minimum spare size */
struct supertype *st3 = dup_super(st2);
- int mdfd = open_dev(mp->devnum);
+ int mdfd = open_dev(mp->devnm);
if (mdfd < 0) {
free(st3);
goto next;
close(mdfd);
}
if ((sra->component_size > 0 &&
- st2->ss->avail_size(st2, devsize) < sra->component_size)
+ st2->ss->avail_size(st2, devsize,
+ sra->devs
+ ? sra->devs->data_offset
+ : INVALID_SECTORS)
+ < sra->component_size)
||
(sra->component_size == 0 && devsize < component_size)) {
if (verbose > 1)
}
if (chosen) {
/* add current device to chosen array as a spare */
- int mdfd = open_dev(devname2devnum(chosen->sys_name));
+ int mdfd = open_dev(chosen->sys_name);
if (mdfd >= 0) {
struct mddev_dev devlist;
char devname[20];
devlist.next = NULL;
devlist.used = 0;
- devlist.re_add = 0;
devlist.writemostly = 0;
devlist.devname = devname;
sprintf(devname, "%d:%d", major(stb.st_rdev),
mdu_array_info_t array;
mdu_bitmap_file_t bmf;
struct mdinfo *sra;
- int mdfd = open_dev(me->devnum);
+ int mdfd = open_dev(me->devnm);
if (mdfd < 0)
continue;
me->path, strerror(errno));
}
}
- sra = sysfs_read(mdfd, 0, 0);
+ /* FIXME check for reshape_active and consider not
+ * starting array.
+ */
+ sra = sysfs_read(mdfd, NULL, 0);
if (sra) {
if (sysfs_set_str(sra, NULL,
"array_state", "read-auto") == 0) {
if (verbose >= 0)
pr_err("started array %s\n",
- me->path ?: devnum2devname(me->devnum));
+ me->path ?: me->devnm);
} else {
pr_err("failed to start array %s: %s\n",
- me->path ?: devnum2devname(me->devnum),
+ me->path ?: me->devnm,
strerror(errno));
rv = 1;
}
if (devname[0] == '/') {
int fd = open(devname, O_RDONLY);
if (fd >= 0) {
- mdname = devnum2devname(fd2devnum(fd));
+ mdname = xstrdup(fd2devnm(fd));
close(fd);
}
} else {
int uuid[4];
struct map_ent *mp, *map = NULL;
-
+
if (!parse_uuid(devname, uuid))
return mdname;
mp = map_by_uuid(&map, uuid);
if (mp)
- mdname = devnum2devname(mp->devnum);
+ mdname = xstrdup(mp->devnm);
map_free(map);
}
}
static int Incremental_container(struct supertype *st, char *devname,
- char *homehost, int verbose,
- int runstop, int autof, int freeze_reshape)
+ struct context *c)
{
/* Collect the contents of this container and for each
* array, choose a device name and assemble the array.
st->ss->getinfo_super(st, &info, NULL);
- if ((runstop > 0 && info.container_enough >= 0) ||
+ if ((c->runstop > 0 && info.container_enough >= 0) ||
info.container_enough > 0)
/* pass */;
else {
- if (verbose)
+ if (c->verbose)
pr_err("not enough devices to start the container\n");
return 0;
}
- match = conf_match(st, &info, devname, verbose, &rv);
+ match = conf_match(st, &info, devname, c->verbose, &rv);
if (match == NULL && rv == 2)
return rv;
/* Need to compute 'trustworthy' */
if (match)
trustworthy = LOCAL;
- else if (st->ss->match_home(st, homehost) == 1)
+ else if (st->ss->match_home(st, c->homehost) == 1)
trustworthy = LOCAL;
else if (st->ss->match_home(st, "any") == 1)
trustworthy = LOCAL;
mp = map_by_uuid(&map, ra->uuid);
if (mp) {
- mdfd = open_dev(mp->devnum);
+ mdfd = open_dev(mp->devnm);
if (mp->path)
strcpy(chosen_name, mp->path);
else
- strcpy(chosen_name, devnum2devname(mp->devnum));
+ strcpy(chosen_name, mp->devnm);
} else {
/* Check in mdadm.conf for container == devname and
free(dn);
/* we have a match */
match = array_list;
- if (verbose>0)
+ if (c->verbose>0)
pr_err("match found for member %s\n",
array_list->member);
break;
if (match && match->devname &&
strcasecmp(match->devname, "<ignore>") == 0) {
- if (verbose > 0)
+ if (c->verbose > 0)
pr_err("array %s/%s is "
"explicitly ignored by mdadm.conf\n",
match->container, match->member);
mdfd = create_mddev(match ? match->devname : NULL,
ra->name,
- autof,
+ c->autof,
trustworthy,
chosen_name);
}
return 2;
}
- assemble_container_content(st, mdfd, ra, runstop,
- chosen_name, verbose, NULL,
- freeze_reshape);
+ assemble_container_content(st, mdfd, ra, c,
+ chosen_name);
close(mdfd);
}
return 0;
}
+static void run_udisks(char *arg1, char *arg2)
+{
+ int pid = fork();
+ int status;
+ if (pid == 0) {
+ execl("/usr/bin/udisks", "udisks", arg1, arg2, NULL);
+ execl("/bin/udisks", "udisks", arg1, arg2, NULL);
+ exit(1);
+ }
+ while (pid > 0 && wait(&status) != pid)
+ ;
+}
+
/*
* IncrementalRemove - Attempt to see if the passed in device belongs to any
* raid arrays, and if so first fail (if needed) and then remove the device.
int IncrementalRemove(char *devname, char *id_path, int verbose)
{
int mdfd;
- int rv;
+ int rv = 0;
struct mdstat_ent *ent;
struct mddev_dev devlist;
+ struct mdinfo mdi;
+ char buf[32];
if (!id_path)
dprintf(Name ": incremental removal without --path <id_path> "
"of any array\n", devname);
return 1;
}
- mdfd = open_dev(ent->devnum);
+ sysfs_init(&mdi, -1, ent->devnm);
+ if (sysfs_get_str(&mdi, NULL, "array_state",
+ buf, sizeof(buf)) > 0) {
+ if (strncmp(buf, "active", 6) == 0 ||
+ strncmp(buf, "clean", 5) == 0)
+ sysfs_set_str(&mdi, NULL,
+ "array_state", "read-auto");
+ }
+ mdfd = open_dev(ent->devnm);
if (mdfd < 0) {
pr_err("Cannot open array %s!!\n", ent->dev);
free_mdstat(ent);
if (id_path) {
struct map_ent *map = NULL, *me;
- me = map_by_devnum(&map, ent->devnum);
+ me = map_by_devnm(&map, ent->devnm);
if (me)
policy_save_path(id_path, me);
map_free(map);
struct mdstat_ent *memb;
for (memb = mdstat ; memb ; memb = memb->next)
if (is_container_member(memb, ent->dev)) {
- int subfd = open_dev(memb->devnum);
+ int subfd = open_dev(memb->devnm);
if (subfd >= 0) {
- Manage_subdevs(memb->dev, subfd,
- &devlist, verbose, 0,
- NULL, 0);
+ rv |= Manage_subdevs(
+ memb->dev, subfd,
+ &devlist, verbose, 0,
+ NULL, 0);
close(subfd);
}
}
free_mdstat(mdstat);
} else
- Manage_subdevs(ent->dev, mdfd, &devlist, verbose, 0, NULL, 0);
- devlist.disposition = 'r';
- rv = Manage_subdevs(ent->dev, mdfd, &devlist, verbose, 0, NULL, 0);
+ rv |= Manage_subdevs(ent->dev, mdfd, &devlist,
+ verbose, 0, NULL, 0);
+ if (rv & 2) {
+ /* Failed due to EBUSY, try to stop the array.
+ * Give udisks a chance to unmount it first.
+ */
+ int devid = devnm2devid(ent->devnm);
+ run_udisks("--unmount", map_dev(major(devid),minor(devid), 0));
+ rv = Manage_runstop(ent->dev, mdfd, -1,
+ verbose, 1);
+ if (rv)
+ /* At least we can try to trigger a 'remove' */
+ sysfs_uevent(&mdi, "remove");
+ if (verbose) {
+ if (rv)
+ pr_err("Fail to stop %s too.\n", ent->devnm);
+ }
+ } else {
+ devlist.disposition = 'r';
+ rv = Manage_subdevs(ent->dev, mdfd, &devlist,
+ verbose, 0, NULL, 0);
+ }
close(mdfd);
free_mdstat(ent);
return rv;
projects / thirdparty / mdadm.git / blobdiff