how to swap devices to increase mirror size?

how to swap devices to increase mirror size?

[James Ralston]

What is the recommend way to increase the capacity of a mirror on the
fly?

What seemed obvious to me (remove and then replace each part of the
mirror in turn) doesn't work:

E.g. (for simulation purposes):

   Device Boot    Start       End    Blocks   Id  System
/dev/sdb1             1        33    265041   fd  Linux raid autodetect
/dev/sdb2            34        99    530145   fd  Linux raid autodetect

   Device Boot    Start       End    Blocks   Id  System
/dev/sdc1             1        33    265041   fd  Linux raid autodetect
/dev/sdc2            34        99    530145   fd  Linux raid autodetect

$ mdadm --create /dev/md0 --verbose --level=mirror --raid-devices=2 /dev/sdb1 /dev/sdc1
mdadm: array /dev/md0 started.

$ cat /proc/mdstat
Personalities : [raid1] 
read_ahead 1024 sectors
md0 : active raid1 sdc1[1] sdb1[0]
      264960 blocks [2/2] [UU]
unused devices: <none>

$ mdadm --detail /dev/md0
/dev/md0:
        Version : 00.90.00
  Creation Time : Wed Jan 22 18:40:33 2003
     Raid Level : raid1
     Array Size : 264960 (258.75 MiB 271.31 MB)
    Device Size : 264960 (258.75 MiB 271.31 MB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 0
    Persistence : Superblock is persistent
    Update Time : Wed Jan 22 18:40:33 2003
          State : dirty, no-errors
 Active Devices : 2
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 0
    Number   Major   Minor   RaidDevice State
       0       8       17        0      active sync   /dev/sdb1
       1       8       33        1      active sync   /dev/sdc1
           UUID : 441d90a6:0f20b7cf:0ac9ed56:52bee387

Ok, let's remove sdc1 and replace it with sdc2:

$ mdadm /dev/md0 -f /dev/sdc1 -r /dev/sdc1 -a /dev/sdc2
mdadm: set /dev/sdc1 faulty in /dev/md0
mdadm: hot removed /dev/sdc1
mdadm: hot added /dev/sdc2

But no joy; /dev/sdc2 won't take as mirror:

$ mdadm --detail /dev/md0
/dev/md0:
        Version : 00.90.00
  Creation Time : Wed Jan 22 18:40:33 2003
     Raid Level : raid1
     Array Size : 264960 (258.75 MiB 271.31 MB)
    Device Size : 264960 (258.75 MiB 271.31 MB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 0
    Persistence : Superblock is persistent
    Update Time : Wed Jan 22 18:42:13 2003
          State : dirty, no-errors
 Active Devices : 1
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 1
    Number   Major   Minor   RaidDevice State
       0       8       17        0      active sync   /dev/sdb1
       1       0        0        1      faulty removed
       2       8       34        2        /dev/sdc2
           UUID : 441d90a6:0f20b7cf:0ac9ed56:52bee387

$ cat /proc/mdstat 
Personalities : [raid1] 
read_ahead 1024 sectors
md0 : active raid1 sdc2[2] sdb1[0]
      264960 blocks [2/1] [U_]
unused devices: <none>

The mirror won't rebuild /dev/sdc2 to match /dev/sdb1.

What do I have to do in order to get /dev/sdc2 to *replace* /dev/sdc1
in the array?

If the answer is "you can't do that", then what is the recommended
method for doing what I'm trying to do?

Thanks,

-- 
James Ralston, Information Technology
Software Engineering Institute
Carnegie Mellon University, Pittsburgh, PA, USA

Deadlock on mkfs while resyncing the raid1 device

[bmoon]

Hello Raiders!

This is special for ARM linux with Debian package.
I tried with linux2418 on pc without any problem.

I am running linux2.4.18 with debian2.0 on ARM based system.
The system will be hang if I try to run "mkfs" on the md device which
is now resyncing on reboot after crash.

Does the SMP compile option make any difference on this problem?

Here are some clues and symptoms from my test on this hang case;

From " raid1_sync_request" in raid1.c,

It hangs while it is looping in " while (sector_nr >= conf->start_pending)".

If it has 0 on conf->cnt_pending it should be OK. However, it is non-zero
on conf->cnt_pending when we do "mkfs" while it is resyncing.

Do you have any idea to solve or to work around this problem?

Are there anyone who tried the ARM linux with SMP option? How?


THanks in advance,

BO

Re: Deadlock on mkfs while resyncing the raid1 device

[bmoon]

Hello Raiders!

This is special for ARM linux with Debian package.
I tried with linux2418 on pc without any problem.

I am running linux2.4.18 with debian2.0 on ARM based system.
The system will be hang if I try to run "mkfs" on the md device which
is now resyncing on reboot after crash.

Does the SMP compile option make any difference on this problem?

Here are some clues and symptoms from my test on this hang case;

From " raid1_sync_request" in raid1.c,

It hangs while it is looping in " while (sector_nr >= conf->start_pending)".

If it has 0 on conf->cnt_pending it should be OK. However, it is non-zero
on conf->cnt_pending when we do "mkfs" while it is resyncing.

Do you have any idea to solve or to work around this problem?

Are there anyone who tried the ARM linux with SMP option? How?


Thanks in advance,

BO