Problem when dettaching databases

#1

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#2

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#3

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#4

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#5

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#6

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#7

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#8

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#9

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.

#10

Something is wrong with your cluster. The disk resources are not supposed
to fail when you detach a database. With a single-node cluster, you cannot
test failover correctly. I suspect you have a semi-clustered setup where a
local instance of SQL Server is using "clustered" disks.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Juan Miguel Cervantes" <jmcervantes (AT) hercesa (DOT) com> wrote

Quote:

Hello all,

I have a one-node failover cluster with Windows Server 2008 x64 and SQL
Server 2005 x64. I'm using a HP MSA1000 connected by FC as shared storage
and I have the following issue:

Whenever I detach a database from the SQL Management Studio, the cluster
disks are marked as failed (I have log/DB in different disk), and brougth
off-line, the system automatically runs a chkdsk on both disks that finish
without any error, but fails to bring the disks online again.
In the event log there is an entry that states that the file system may be
corrupt.

If I place the disks in maintenance mode, I can run a chkdsk that ends
successfully and without any error again and I can access the filesystem
without problem, but when I place the disks in normal mode, are brought
off-line again and everything starts over... off-line/chkdsk that finish
without error/disks failed

The only way I have found to "recover" the disk, is to place it in
maintenance mode, move the database/log files to another location, then
FORMAT the disks, move back the database/log files and then, everything
works fine.

Any help or hint on what's going on?

Thank you.