Upgrading Raid1 array to have 2. disks.

DISCUSSION Anders Fugmann 4 participants 8 messages May 6, 2003 – May 30, 2003 MD RAID

lore

Share & tools

lore

Anders Fugmann

author

2003-05-06 23:24 UTC #0 lore raw

Hi,

Sorry if this is the wrong list to post this on.

I have made an error when creating a new raid1 array as I specified that
the raid should consist of only one disk. Now, I would like to add another disk, but mdadm insists on adding it as a spare disk, offering no data redundancy.

Before:
md0 : active raid1 ide/host0/bus0/target0/lun0/part1[0]

      1048704 blocks [1/1] [U]

After running

# mdadm -add /dev/md/0
    /dev/ide/host0/bus0/target0/lun0/part1[0]

md0 : active raid1 ide/host0/bus1/target0/lun0/part1[1] ide/host0/bus0/target0/lun0/part1[0]

      1048704 blocks [1/1] [U]

I have read the manpages, and tried googling without success.
I do not wish to destroy the data on the discs, and recreate the array but It might be a last resort. Can anyone help me convince mdadm to add the 2. partition as an active disk.

Regards
Anders Fugmann

P.s.
Please CC me, as I'm not subscribed to the list.

Mads Peter Bach

2003-05-06 23:46 UTC #1 lore raw

Quoted 1 line · Anders Fugmann

Sorry if this is the wrong list to post this on.

It's the right list.

Quoted 9 lines

I have made an error when creating a new raid1 array as I specified that the raid should consist of only one disk. Now, I would like to add another disk, but mdadm insists on adding it as a spare disk, offering no data redundancy.
I have read the manpages, and tried googling without success. I do not wish to destroy the data on the discs, and recreate the array but It might be a last resort. Can anyone help me convince mdadm to add the 2. partition as an active disk.

You could keep your data by doing this:

Remove the current spare:

mdadm --remove /dev/md/0 /dev/ide/host0/bus0/target0/lun0/part1[0]

Create new array with

mdadm --create /dev/md/1 --level=1 --raid-devices=/dev/ide/host0/bus0/target0/lun0/part1[0] missing

Then create fs on new array, mount & copy data to the new array. Stop the old array, add disk from old array to the new array.

Signature · 9 lines

Mads Peter Bach
Systemadministrator,  Det Humanistiske Fakultet, Aalborg Universitet
Kroghstræde 3 - 5.111, DK-9220 Aalborg Øst - (+45) 96358062
# whois [email protected]

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Neil Brown

2003-05-07 00:00 UTC #2 lore raw

In reply to Anders Fugmann

Quoted 19 lines · On Wednesday May 7, [email protected] wrote:

Hi,
Sorry if this is the wrong list to post this on.
I have made an error when creating a new raid1 array as I specified that the raid should consist of only one disk. Now, I would like to add another disk, but mdadm insists on adding it as a spare disk, offering no data redundancy.
Before:
md0 : active raid1 ide/host0/bus0/target0/lun0/part1[0]
       1048704 blocks [1/1] [U]
After running
# mdadm -add /dev/md/0
     /dev/ide/host0/bus0/target0/lun0/part1[0]
md0 : active raid1 ide/host0/bus1/target0/lun0/part1[1] ide/host0/bus0/target0/lun0/part1[0]

Quoted 7 lines

       1048704 blocks [1/1] [U]
I have read the manpages, and tried googling without success. I do not wish to destroy the data on the discs, and recreate the array but It might be a last resort. Can anyone help me convince mdadm to add the 2. partition as an active disk.

You cannot add a new data drive to a raid1 array.
If you shut down the array and re-create it you should not loose data. The correct command to use would be something like:
mdadm --create /dev/md0 --level=1 -n 2 /dev/ide/host0/bus0/target0/lun0/part1 missing

Note the word "missing" at the end. This tells mdadm to create a 2-drive array with one missing drive (and so only one active drive). This will have the same data as the old md0.
Now "mdadm --add" should do what you want.

NeilBrown

Quoted 12 lines

Regards
Anders Fugmann
P.s.
Please CC me, as I'm not subscribed to the list.
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Anders Fugmann

author

2003-05-08 19:11 UTC #3 lore raw

Quoted 9 lines · Neil Brown

You cannot add a new data drive to a raid1 array.
If you shut down the array and re-create it you should not loose data. The correct command to use would be something like:
mdadm --create /dev/md0 --level=1 -n 2 /dev/ide/host0/bus0/target0/lun0/part1 missing
Note the word "missing" at the end. This tells mdadm to create a 2-drive array with one missing drive (and so only one active drive). This will have the same data as the old md0.
Now "mdadm --add" should do what you want.

Thanks for the replies.
The method above worked like a charm. The data on the array even surived the operation.

Quoted 2 lines

NeilBrown

Regards
Anders Fugmann

3tcdgwg3

2003-05-20 00:33 UTC #4 lore raw

raid5, 2 drives dead at same time,kernel will Oops?

In reply to Neil Brown

Hi,

I am trying to simulate a case that two drives in an array fail ad same time.
I use two ide drives, I try to create a
raid 5 array with 4 arms, created as following:

/dev/hdc1
/dev/hde1
/dev/hdc2
/dev/hde2

This is just for test, I know create two arms on one hard drive doesn't make much sense.

Anyway, when I run this array, if I power off one of hard drive (/dev/hde) to simulate two arms failing at same time in an array, I got system Oops. I am using 2.4-18 kernel.

Anyone can tell me if this is normal? or if there is a fix for this?

Thanks in advance.

Here are the output:

=================================
hde: drive not ready for command

md: updating md0 RAID superblock on device

md: (skipping faulty hde2 )

md: hdc2 [events: 00000002]<6>(write) hdc2's sb offset: 3469952

md: (skipping faulty hde1 )

md: hdc1 [events: 00000002]<6>(write) hdc1's sb offset: 16064896

md0: no spare disk to reconstruct array! -- continuing in degraded mode

md: recovery thread finished ...

md: recovery thread got woken up ...

md0: no spare disk to reconstruct array! -- continuing in degraded mode

md: recovery thread finished ...

sector=561a8 i=0 00000000 00000000 8132e840 0

kernel BUG at raid5.c:309!

kupdated(7): Oops

================================================

Neil Brown

2003-05-21 02:42 UTC #5 lore raw

Re: raid5, 2 drives dead at same time,kernel will Oops?

Quoted 16 lines · On Monday May 19, [email protected] wrote:

Hi,
I am trying to simulate a case that two drives in an array fail ad same time.
I use two ide drives, I try to create a
raid 5 array with 4 arms, created as following:
/dev/hdc1
/dev/hde1
/dev/hdc2
/dev/hde2
This is just for test, I know create two arms on one hard drive doesn't make much sense.

Quoted 7 lines

Anyway, when I run this array, if I power off one of hard drive (/dev/hde) to simulate two arms failing at same time in an array, I got system Oops. I am using 2.4-18 kernel.
Anyone can tell me if this is normal? or if there is a fix for this?

Congratulations and thanks. You have managed to trigger a bug that no-one else has found.

The following patch (against 2.4.20) should fix it. If you can test and confirm I would really appreciate it.

NeilBrown

------------------------------------------------------------
Handle concurrent failure of two drives in raid5

If two drives both fail during a write request, raid5 doesn't cope properly and will eventually oops.

With this patch, blocks that have already been 'written' are failed when double drive failure is noticed, as well as blocks that are about to be written.

----------- Diffstat output ------------

 ./drivers/md/raid5.c |   10 +++++++++-
 1 files changed, 9 insertions(+), 1 deletion(-)

drivers/md/raid5.c renamed from drivers/md/raid5.c~current~ +9 -1

diff ./drivers/md/raid5.c~current~ ./drivers/md/raid5.c
@@ -882,7 +882,7 @@ static void handle_stripe(struct stripe_
 	/* check if the array has lost two devices and, if so, some requests might
 	 * need to be failed
 	 */
-	if (failed > 1 && to_read+to_write) {
+	if (failed > 1 && to_read+to_write+written) {
 		for (i=disks; i--; ) {
 			/* fail all writes first */
 			if (sh->bh_write[i]) to_write--;

@@ -891,6 +891,14 @@ static void handle_stripe(struct stripe_
 				bh->b_reqnext = return_fail;
 				return_fail = bh;
 			}
+			/* and fail all 'written' */
+			if (sh->bh_written[i]) written--;
+			while ((bh = sh->bh_written[i])) {
+				sh->bh_written[i] = bh->b_reqnext;
+				bh->b_reqnext = return_fail;
+				return_fail = bh;
+			}
+
 			/* fail any reads if this device is non-operational */
 			if (!conf->disks[i].operational) {
 				spin_lock_irq(&conf->device_lock);

3tcdgwg3

2003-05-22 01:04 UTC #6 lore raw

Re: raid5, 2 drives dead at same time,kernel will Oops?

Neil,
Preliminary test looks good, will test more
when have time.

Thanks,
-Will.
----- Original Message -----
From: "Neil Brown" <[email protected]>
To: "3tcdgwg3" <[email protected]>
Cc: <[email protected]>
Sent: Tuesday, May 20, 2003 7:42 PM
Subject: Re: raid5, 2 drives dead at same time,kernel will Oops?

Quoted 52 lines · On Monday May 19, [email protected] wrote:

Quoted 16 lines
Hi,
I am trying to simulate a case that two drives in an array fail ad same time.
I use two ide drives, I try to create a
raid 5 array with 4 arms, created as following:
/dev/hdc1
/dev/hde1
/dev/hdc2
/dev/hde2
This is just for test, I know create two arms on one hard drive doesn't make much sense.
Quoted 7 lines
Anyway, when I run this array, if I power off one of hard drive (/dev/hde) to simulate two arms failing at same time in an array, I got system Oops. I am using 2.4-18 kernel.
Anyone can tell me if this is normal? or if there is a fix for this?
Congratulations and thanks. You have managed to trigger a bug that no-one else has found.
The following patch (against 2.4.20) should fix it. If you can test and confirm I would really appreciate it.
NeilBrown
------------------------------------------------------------
Handle concurrent failure of two drives in raid5
If two drives both fail during a write request, raid5 doesn't cope properly and will eventually oops.
With this patch, blocks that have already been 'written' are failed when double drive failure is noticed, as well as blocks that are about to be written.
----------- Diffstat output ------------
 ./drivers/md/raid5.c |   10 +++++++++-
 1 files changed, 9 insertions(+), 1 deletion(-)
diff ./drivers/md/raid5.c~current~ ./drivers/md/raid5.c

drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 static void handle_stripe(struct stripe_ · line 882

Quoted 1 line

drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 renamed from drivers/md/raid5.c~current~ 2003-05-21 12:42:07.000000000 +1000 +0 -0
@@ -882,7 +882,7 @@ static void handle_stripe(struct stripe_
  /* check if the array has lost two devices and, if so, some requests

might

Quoted 26 lines

drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 renamed from drivers/md/raid5.c~current~ 2003-05-21 12:42:07.000000000 +1000 +9 -2

  * need to be failed
  */
- if (failed > 1 && to_read+to_write) {
+ if (failed > 1 && to_read+to_write+written) {
  for (i=disks; i--; ) {
  /* fail all writes first */
  if (sh->bh_write[i]) to_write--;

@@ -891,6 +891,14 @@ static void handle_stripe(struct stripe_
  bh->b_reqnext = return_fail;
  return_fail = bh;
  }
+ /* and fail all 'written' */
+ if (sh->bh_written[i]) written--;
+ while ((bh = sh->bh_written[i])) {
+ sh->bh_written[i] = bh->b_reqnext;
+ bh->b_reqnext = return_fail;
+ return_fail = bh;
+ }
+
  /* fail any reads if this device is non-operational */
  if (!conf->disks[i].operational) {
  spin_lock_irq(&conf->device_lock);
-

To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

3tcdgwg3

2003-05-30 20:33 UTC #7 lore raw

Re: raid5, 2 drives dead at same time,kernel will Oops?

Hi,

I have some other issues under this "more than 1
arm broken in a raid5 array" condition. The next
important one is this:

If I have two arrays, if first one is a raid5 and resync is going on, second is another raid5, resync is schedule to start after the first raid5 array synced.
At this time, if I kill to arms in the first raid5 array, the resync will stop, but never aborted, consequently, second raid5 array never get a chance to start the resync.
Is there a fix for this?

Thanks

-W

----- Original Message -----
From: "3tcdgwg3" <[email protected]>
To: "Neil Brown" <[email protected]>
Cc: <[email protected]>
Sent: Wednesday, May 21, 2003 6:04 PM
Subject: Re: raid5, 2 drives dead at same time,kernel will Oops?

Quoted 99 lines

Neil,
Preliminary test looks good, will test more
when have time.
Thanks,
-Will.
----- Original Message -----
From: "Neil Brown" <[email protected]>
To: "3tcdgwg3" <[email protected]>
Cc: <[email protected]>
Sent: Tuesday, May 20, 2003 7:42 PM
Subject: Re: raid5, 2 drives dead at same time,kernel will Oops?
Quoted 52 lines · On Monday May 19, [email protected] wrote:
Quoted 16 lines
Hi,
I am trying to simulate a case that two drives in an array fail ad same time.
I use two ide drives, I try to create a
raid 5 array with 4 arms, created as following:
/dev/hdc1
/dev/hde1
/dev/hdc2
/dev/hde2
This is just for test, I know create two arms on one hard drive doesn't make much sense.
Quoted 7 lines
Anyway, when I run this array, if I power off one of hard drive (/dev/hde) to simulate two arms failing at same time in an array, I got system Oops. I am using 2.4-18 kernel.
Anyone can tell me if this is normal? or if there is a fix for this?
Congratulations and thanks. You have managed to trigger a bug that no-one else has found.
The following patch (against 2.4.20) should fix it. If you can test and confirm I would really appreciate it.
NeilBrown
------------------------------------------------------------
Handle concurrent failure of two drives in raid5
If two drives both fail during a write request, raid5 doesn't cope properly and will eventually oops.
With this patch, blocks that have already been 'written' are failed when double drive failure is noticed, as well as blocks that are about to be written.
----------- Diffstat output ------------
 ./drivers/md/raid5.c |   10 +++++++++-
 1 files changed, 9 insertions(+), 1 deletion(-)
diff ./drivers/md/raid5.c~current~ ./drivers/md/raid5.c
drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 static void handle_stripe(struct stripe_ · line 882
Quoted 1 line
drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 renamed from drivers/md/raid5.c~current~ 2003-05-21 12:42:07.000000000 +1000 +0 -0
@@ -882,7 +882,7 @@ static void handle_stripe(struct stripe_
  /* check if the array has lost two devices and, if so, some requests
might
Quoted 23 lines
drivers/md/raid5.c 2003-05-21 12:37:37.000000000 +1000 renamed from drivers/md/raid5.c~current~ 2003-05-21 12:42:07.000000000 +1000 +9 -2
  * need to be failed
  */
- if (failed > 1 && to_read+to_write) {
+ if (failed > 1 && to_read+to_write+written) {
  for (i=disks; i--; ) {
  /* fail all writes first */
  if (sh->bh_write[i]) to_write--;
@@ -891,6 +891,14 @@ static void handle_stripe(struct stripe_
  bh->b_reqnext = return_fail;
  return_fail = bh;
  }
+ /* and fail all 'written' */
+ if (sh->bh_written[i]) written--;
+ while ((bh = sh->bh_written[i])) {
+ sh->bh_written[i] = bh->b_reqnext;
+ bh->b_reqnext = return_fail;
+ return_fail = bh;
+ }
+
  /* fail any reads if this device is non-operational */
  if (!conf->disks[i].operational) {
  spin_lock_irq(&conf->device_lock);
-
Quoted 3 lines
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html