mdadm - is not able to properly assemble multipath device

[Debian Bug Importer]

Automatically imported from Debian bug report #271033 http://bugs.debian.org/271033

[Philipp Kolmann (philipp-kolmann)]

Hi!

we are having the same problem here. I am just curious if you have solved this
issue already.

The error on my side reads as follows:
nodenfs:~# mdadm -A /dev/md2 /dev/sdc1 /dev/sdd1
mdadm: device 1 in /dev/md2 has wrong state in superblock, but /dev/sdd1 seems
ok
mdadm: failed to add /dev/sdc1 to /dev/md2: Device or resource busy
mdadm: /dev/md2 has been started with 2 drives and -1 spares.

Thanks for any help
Philipp Kolmann

--
Life is not fair,
but the root password helps!

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #271033 http://bugs.debian.org/271033

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 10 Sep 2004 20:27:44 +0200
From: Bastian Blank <email address hidden>
To: <email address hidden>
Subject: mdadm - is not able to properly assemble multipath device

--qMm9M+Fa2AknHoGS
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Package: mdadm
Version: 1.6.0-3
Severity: important

mdadm is not able to assemble multipath device except of its own
creation.

| # mdadm --create -n 2 -l multipath /.dev/md0 /dev/zfcp/0x2*/*/*
| mdadm: /dev/zfcp/0x2100002037470047/0x0000000000000000/disc appears to be=
 part of a raid array:
| level=3D-4 devices=3D2 ctime=3DFri Sep 10 18:15:48 2004
| mdadm: /dev/zfcp/0x2200002037470047/0x0000000000000000/disc appears to be=
 part of a raid array:
| level=3D-4 devices=3D2 ctime=3DFri Sep 10 18:15:48 2004
| Continue creating array? y
| mdadm: array /.dev/md0 started.
| # mdadm --examine --scan >> /etc/mdadm/mdadm.conf
| # cat /etc/mdadm/mdadm.conf
| DEVICE /dev/zfcp/*/*/*
| ARRAY /dev/md0 level=3Dmultipath num-devices=3D2 UUID=3Dac88c560:ad6b1e3d=
:211de7b0:5d562f61
| devices=3D/dev/zfcp/0x2200002037470047/0x0000000000000000/disc,/dev/zf=
cp/0x2100002037470047/0x0000000000000000/disc
| # mdadm --stop /dev/md0
| # mdadm -A -s -v =20
| mdadm: looking for devices for /dev/md0
| mdadm: /dev/zfcp/0x2200002037470047/0x0000000000000000/disc is identified=
 as a member of /dev/md0, slot 1.
| mdadm: /dev/zfcp/0x2100002037470047/0x0000000000000000/disc is identified=
 as a member of /dev/md0, slot 1.
| mdadm: added /dev/zfcp/0x2100002037470047/0x0000000000000000/disc to /dev=
/md0 as 1
| mdadm: failed to add /dev/zfcp/0x2200002037470047/0x0000000000000000/disc=
 to /dev/md0: Device or resource busy
| mdadm: /dev/md0 has been started with 2 drives and -1 spares.
| # mdadm -D /dev/md0
| /dev/md0:
| Version : 00.90.01
| Creation Time : Fri Sep 10 18:19:40 2004
| Raid Level : multipath
| Array Size : 35566400 (33.92 GiB 36.42 GB)
| Raid Devices : 2
| Total Devices : 1
| Preferred Minor : 0
| Persistence : Superblock is persistent
|=20
| Update Time : Fri Sep 10 18:20:49 2004
| State : clean, degraded
| Active Devices : 1
| Working Devices : 1
| Failed Devices : 0
| Spare Devices : 0
|=20
|=20
| Number Major Minor RaidDevice State
| 0 8 16 0 active sync /dev/zfcp/0x2100002=
037470047/0x0000000000000000/disc
| 1 0 0 -1 removed
| UUID : ac88c560:ad6b1e3d:211de7b0:5d562f61
| Events : 0.9
| # dmesg
| [...]
| md: bind<sdb>
| md: bind<sda>
| multipath: array md0 active with 2 out of 2 IO paths
| md: md0 stopped.
| md: unbind<sda>
| md: export_rdev(sda)
| md: unbind<sdb>
| md: export_rdev(sdb)
| md: md0 stopped.
| md: bind<sdb>
| md: export_rdev(sda)
| multipath: array md0 active with 1 out of 2 IO paths
| # l /dev/zfcp/*/*/*
| brw-rw---- 1 root root 8, 16 Sep 10 17:56 /dev/zfcp/0x2100002037470047/0=
x0000000000000000/disc
| brw-rw---- 1 root root 8, 0 Sep 10 17:56 /dev/zfcp/0x2200002037470047/0=
x0000000000000000/disc

-- System Information:
Debian...

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 17 Jan 2005 18:18:36 +0100
From: Philipp Kolmann <email address hidden>
To: <email address hidden>
Cc: Bastian Blank <email address hidden>
Subject: Same Problem here, have you solved this issue?

Hi!

we are having the same problem here. I am just curious if you have solved this
issue already.

The error on my side reads as follows:
nodenfs:~# mdadm -A /dev/md2 /dev/sdc1 /dev/sdd1
mdadm: device 1 in /dev/md2 has wrong state in superblock, but /dev/sdd1 seems
ok
mdadm: failed to add /dev/sdc1 to /dev/md2: Device or resource busy
mdadm: /dev/md2 has been started with 2 drives and -1 spares.

Thanks for any help
Philipp Kolmann

--
Life is not fair,
but the root password helps!

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Tue, 1 Feb 2005 10:54:44 +0100
From: Bastian Blank <email address hidden>
To: <email address hidden>
Subject: severity of 271033 is grave

# Automatically generated email from bts, devscripts version 2.8.10
severity 271033 grave

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 2 Feb 2005 10:30:51 +1100
From: Neil Brown <email address hidden>
To: =?ISO-8859-15?Q?Mario_Jou=DFen?= <email address hidden>
cc: <email address hidden>, Bastian Blank <email address hidden>
Subject: Re: mdadm and multipath device

On Tuesday February 1, <email address hidden> wrote:
> Hi Neil,
>
> do you know something about the problem described in the following bug
> report?
>
> <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=271033>

The main problem is a kernel bug (2.6.X), not an mdadm bug.

in drivers/md/md.c, in function super_90_load, it says:

 if (sb->level == MULTIPATH)
  rdev->desc_nr = -1;
 else
  rdev->desc_nr = sb->this_disk.number;

That should be
 if (sb->level == LEVEL_MULTIPATH)
  rdev->desc_nr = -1;
 else
  rdev->desc_nr = sb->this_disk.number;

This causes the failure when trying to add the second device.
I'll try to get the patch into 2.6.11.

The message
  mdadm: /dev/md4 has been started with 2 drives and -1 spares.

is wrong (it should be "started with 1 drive"). I'll try to get that
fixed for mdadm 1.9.0

NeilBrown

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 6 Feb 2005 19:41:20 +0100
From: Mario Joussen <email address hidden>
To: <email address hidden>
Subject: Downgrade severity to important

severity #271033 important

stop

It's at most an important bug, because it is caused by a kernel problem.

Cheers, Mario

[Matt Zimmerman (mdz)]

No point in tracking this indedpenently of Debian; it isn't severe and we won't
pursue it