On Thursday September 8, babydr@xxxxxxxxxxxxxxxx wrote:
> > What happens if you then
> > mdadm /dev/md_d0 -a /dev/sda[pqrs]
> > ??
>
> Getting stranger & stranger .
>
> root@devel-0:~ # mdadm /dev/md_d0 -a /dev/sda[pqrs]
> mdadm: re-added /dev/sdap
>
Hmm.. mdadm bug.
> root@devel-0:~ # cat /proc/mdstat
> Personalities : [linear] [raid0] [raid1] [raid5] [multipath] [raid6] [raid10]
> md_d0 : active raid5 sdap[36] sdc[0] sdao[40] sdan[34] sdam[33]
> sdal[32] sdak[31] sdaj[30] sdah[29] sdag[28] sdaf[27] sdae[26]
> sdad[25] sdac[24] sdab[23] sdaa[22] sdz[21] sdy[20] sdw[19] sdv[18]
> sdu[17] sdt[16] sds[15] sdr[14] sdq[13] sdp[12] sdo[11] sdn[10] sdl[9]
> sdk[8] sdj[7] sdi[6] sdh[5] sdg[4] sdf[3] sde[2](F) sdd[1]
> 1244826240 blocks level 5, 64k chunk, algorithm 2 [36/35]
> [UU_UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU]
Hmm.. obviously hot-add isn't enough to trigger the rebuild in that
kernel.
Attached are three patches.
The first two are needed by 2.6.12.5 to make sure resync happens (this
is particularly a problem for version-1 superblocks) or just upgrade
to 2.6.13.
The last fixes mdadm-v2.0 so that when you add /dev/sda[pqrs] it
actually adds all of them, and so that when you --assemble a version-1
array with spares, the spares actually get included.
NeilBrown
Status: ok
Make sure recovery happens when add_new_disk is used for hot_add
Currently if add_new_disk is used to hot-add a drive to a degraded
array, recovery doesn't start ... because we didn't tell it to.
Signed-off-by: Neil Brown <neilb@xxxxxxxxxxxxxxx>
### Diffstat output
./drivers/md/md.c | 2 ++
1 files changed, 2 insertions(+)
diff ./drivers/md/md.c~current~ ./drivers/md/md.c
--- ./drivers/md/md.c~current~ 2005-05-31 13:40:35.000000000 +1000
+++ ./drivers/md/md.c 2005-05-31 13:40:34.000000000 +1000
@@ -2232,6 +2232,8 @@ static int add_new_disk(mddev_t * mddev,
err = bind_rdev_to_array(rdev, mddev);
if (err)
export_rdev(rdev);
+
+ set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
if (mddev->thread)
md_wakeup_thread(mddev->thread);
return err;
Status: ok
Make sure resync gets started when array starts.
We weren't actually waking up the md thread after setting
MD_RECOVERY_NEEDED when assembling an array, so it is possible to
lose a race and not actually start resync.
So add a call to md_wakeup_thread, and while we are at it, remove
all the "if (mddev->thread)" guards as md_wake_thread does its own
checking.
Signed-off-by: Neil Brown <neilb@xxxxxxxxxxxxxxx>
### Diffstat output
./drivers/md/md.c | 7 +++----
1 files changed, 3 insertions(+), 4 deletions(-)
diff ./drivers/md/md.c~current~ ./drivers/md/md.c
--- ./drivers/md/md.c 2005-08-26 17:00:30.000000000 +1000
+++ ./drivers/md/md.c~current~ 2005-08-26 17:00:39.000000000 +1000
@@ -256,8 +256,7 @@ static inline void mddev_unlock(mddev_t
{
up(&mddev->reconfig_sem);
- if (mddev->thread)
- md_wakeup_thread(mddev->thread);
+ md_wakeup_thread(mddev->thread);
}
mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr)
@@ -1726,6 +1725,7 @@ static int do_md_run(mddev_t * mddev)
mddev->in_sync = 1;
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
+ md_wakeup_thread(mddev->thread);
if (mddev->sb_dirty)
md_update_sb(mddev);
@@ -2255,8 +2255,7 @@ static int add_new_disk(mddev_t * mddev,
export_rdev(rdev);
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
- if (mddev->thread)
- md_wakeup_thread(mddev->thread);
+ md_wakeup_thread(mddev->thread);
return err;
}
diff ./Assemble.c~current~ ./Assemble.c
--- ./Assemble.c~current~ 2005-09-05 10:55:01.000000000 +1000
+++ ./Assemble.c 2005-09-09 16:24:50.000000000 +1000
@@ -119,6 +119,7 @@ int Assemble(struct supertype *st, char
struct mdinfo info;
struct mddev_ident_s ident2;
char *avail;
+ int nextspare = 0;
vers = md_get_version(mdfd);
if (vers <= 0) {
@@ -320,6 +321,11 @@ int Assemble(struct supertype *st, char
i = devcnt;
else
i = devices[devcnt].raid_disk;
+ if (i+1 == 0) {
+ if (nextspare < info.array.raid_disks)
+ nextspare = info.array.raid_disks;
+ i = nextspare++;
+ }
if (i < 10000) {
if (i >= bestcnt) {
unsigned int newbestcnt = i+10;
diff ./Manage.c~current~ ./Manage.c
--- ./Manage.c~current~ 2005-09-05 10:54:55.000000000 +1000
+++ ./Manage.c 2005-09-09 16:04:12.000000000 +1000
@@ -288,7 +288,7 @@ int Manage_subdevs(char *devname, int fd
if (ioctl(fd, ADD_NEW_DISK, &disc) == 0) {
if (verbose >= 0)
fprintf(stderr, Name ": re-added %s\n", dv->devname);
- return 0;
+ continue;
}
/* fall back on normal-add */
}
