PRIMECLUSTER Global File Services Configuration and Administration Guide 4.2 (Solaris(TM) Operating Environment)

F.2.29 sfcswitch(1M) MDS failback

[SYNOPSIS]

sfcswitch [ -w timeout ] special | mount_point

[DESCRIPTION]

sfcswitch fails the primary MDS of the operating GFS Shared File System, specified by special or mount_point, back to the primary server and the secondary MDS back to the secondary server.

sfcswitch does not fail back MDS so as not to affect the operation if any of the following conditions is satisfied:

• The load on sfcfrmd is high
• Mount or unmount is being executed
• There is a node whose file system is locked out
• There is a node on which MDS failed to start

[OPTIONS]

The following options can be specified:

-w timeout

Specifies a timeout interval until a response from the sfcfsd daemon is received, in units of seconds. When the value of timeout is exceeded, processing is terminated. The default value is 3 seconds.

The minimum value that can be specified is 1 and the maximum value is 2147483647.

[OPERAND]

The following options can be operands:

special

Specifies a block special device name or character special device name for the file system representative partition. If you will perform execution from the node on which the file system is not mounted, specify special.

mount_point

You can specify the mount point of the file system.

[EXAMPLE]

The following shows an example of starting the primary MDS on the primary server host1 and the secondary MDS on the secondary server host2 when the secondary MD is running on the primary server host1 and the primary MDS is running on the secondary server host2:

# sfcrscinfo -mi 1
/dev/sfdsk/gfs/dsk/v1:
FSID  MDS/AC  STATE  S-STATE   RID-1   RID-2   RID-N  hostname
  1   MDS(P)   wait     -        0       0       0      host1
  1   AC       run      -        0       0       0      host1
  1   MDS(S)   run      -        0       0       0      host2
  1   AC       run      -        0       0       0      host2
  1   MDS(I)   imag     -        0       0       0      host3
  1   AC       run      -        0       0       0      host3
  1   MDS(I)   imag     -        0       0       0      host4
  1   AC       run      -        0       0       0      host4
# sfcswitch /dev/sfdsk/gfs/rdsk/v1
# sfcrscinfo -mi 1
/dev/sfdsk/gfs/dsk/v1:
FSID  MDS/AC  STATE  S-STATE   RID-1   RID-2   RID-N  hostname
  1   MDS(P)   run      -        0       0       0      host1
  1   AC       run      -        0       0       0      host1
  1   MDS(S)   wait     -        0       0       0      host2
  1   AC       run      -        0       0       0      host2
  1   MDS(I)   imag     -        0       0       0      host3
  1   AC       run      -        0       0       0      host3
  1   MDS(I)   imag     -        0       0       0      host4
  1   AC       run      -        0       0       0      host4

[EXIT STATUS]

Input English.

0 MDS failback terminated normally.

1 The option or the operand is invalid.

2 A problem was detected in the file system setting or Solaris OE.

3 MDS failback was interrupted so as not to adversely affect the operation.

4 A problem was detected in either of the following:

• Failed to read in the management partition.
• Detected that some of the nodes are locked out.
• Detected that startup of MDS failed in some of the nodes.
• A timeout occurred during MDS failback processing.
• The file system is locked out.

5 MDS was not allocated to the location defined when the file system was created.

[FILE]

/etc/mnttab

Table of the file system that is being mounted

/etc/vfstab

Environment file for the file system

[SEE ALSO]

mkfs_sfcfs(1M), sfcnewfs(1M), sfcrscinfo(1M).

mnttab(4), vfstab(4) in the "Solaris X Reference Manual Collection."

[NOTE]

• If the relevant file system is not mounted on either the primary server or the secondary server, sfcswitch cannot return the state of allocation of MDS to the defined allocation.