man pages section 1M: System Administration Commands

Exit Print View

Updated: July 2014
 
 

fdisk(1M)

Name

fdisk - create or modify fixed disk partition table

Synopsis

fdisk [-o offset] [-s size] [-P fill_patt] [-S geom_file] 
     [-w | -r | -d | -n | -I | -B | -t | -T | -g | -G | -R | -E] 
     [-–F fdisk_file] [ [-v] -W {fdisk_file | −}] 
     [-h] [-b masterboot] 
     [-A id : act : bhead : bsect : bcyl : ehead : esect : 
         ecyl : rsect : numsect] 
     [-D id : act : bhead: bsect : bcyl : ehead: esect :
         ecyl : rsect : numsect] rdevice

Description

This command is used to do the following:

  • Create and modify an fdisk partition table on x86 systems

  • Create and modify an fdisk partition table on removable media on SPARC or x86 systems

  • Install the master boot record that is put in the first sector of the fixed disk on x86 systems only

This table is used by the first-stage bootstrap (or firmware) to identify parts of the disk reserved for different operating systems, and to identify the partition containing the second-stage bootstrap (the active Solaris partition). The rdevice argument must be used to specify the raw device associated with the fixed disk, for example, /dev/rdsk/c0t0d0p0.

The program can operate in three different modes. The first is interactive mode. In interactive mode, the program displays the partition table as it exists on the disk, and then presents a menu allowing the user to modify the table. The menu, questions, warnings, and error messages are intended to be self-explanatory.

In interactive mode, if there is no partition table on the disk, the user is given the options of creating a default partitioning or specifying the initial table values. The default partitioning allocates the entire disk for the Solaris system and makes the Solaris system partition active. In either case, when the initial table is created, fdisk also writes out the first-stage bootstrap (x86 only) code along with the partition table. In this mode, (x86 only) when creating an entry for a non-EFI partition on a disk that is larger than 2 TB (terabytes), fdisk warns that the maximum size of the partition is 2 TB. Under these conditions percentages displayed by fdisk are based on 2 TB.

The second mode of operation is used for automated entry addition, entry deletion, or replacement of the entire fdisk table. This mode can add or delete an entry described on the command line. In this mode the entire fdisk table can be read in from a file replacing the original table. fdisk can also be used to create this file. There is a command line option that will cause fdisk to replace any fdisk table with the default of the whole disk for the Solaris system.

The third mode of operation is used for disk diagnostics. In this mode, a section of the disk can be filled with a user-specified pattern and mode sections of the disk can also be read or written.


Note - The third mode of operation is not currently supported for extended partitions

When fdisk creates a partition, the space is allocated in the fdisk partition table, but the allocated disk space is not initialized. newfs(1M) is required to create and write file system metadata to the new partition, and format(1M) is required to write the VTOC or EFI/GPT metadata.

Menu Options

The menu options for interactive mode given by the fdisk program are:

Create a partition

This option allows the user to create a new partition. The maximum number of partitions is 4. The program will ask for the type of the partition (SOLARIS, MS-DOS, UNIX, or other). It will then ask for the size of the partition as a percentage of the disk. The user may also enter the letter c at this point, in which case the program will ask for the starting cylinder number and size of the partition in cylinders. If a c is not entered, the program will determine the starting cylinder number where the partition will fit. In either case, if the partition would overlap an existing partition or will not fit, a message is displayed and the program returns to the original menu.

Change Active (Boot from) partition

This option allows the user to specify the partition where the first-stage bootstrap will look for the second-stage bootstrap, otherwise known as the active partition.

Delete a partition

This option allows the user to delete a previously created partition. Note that this will destroy all data in that partition.

Change between Solaris and Solaris2 Partition IDs

This option allows the user to switch between the current fdisk operating system partition identifier and the previous one. This does not affect any data in the disk partition and is provided for compatibility with older software.

Edit/View extended partitions

This option provides the extended partition menu to the user. Use the extended partition menu to add and delete logical drives, change the sysid of the logical drives, and display logical drive information. To commit the changes made in the extended partition, you must return to the main menu using the extended partition submenu option r. There is also an option to display the list of options that the extended partition submenu supports. Given below is the list:

a

Add a logical drive.

Use this submenu option to add a logical drive. There are three pieces of information that are required: The beginning cylinder, the size (in cylinders or in human readable form - KB, MB, or GB), and the partition ID. While specifying the partition ID, there is an option (I) that you can use to list the supported partitions.

d

Delete a logical drive.

Use this submenu option to delete a logical drive. The only input required is the number of the logical drive that is to be deleted.

h

Display the help menu.

This submenu option displays the supported operations in the extended partition submenu.

i

Change the id of the logical drive.

Use this submenu option to change the system ID of the existing logical drives. A list of supported system IDs is displayed when you use the I option when in this submenu.

p

Display the logical drive layout.

Displays the logical drive information to stdout. This output reflects any changes made during the current run of the fdisk program. The changes are not committed to the disk until return to the main menu (using the submenu r) and choose the option to commit the changes to the disk.

r

Return to the main fdisk menu.

Exit the extended partition submenu and return to the main menu.

Note the dynamic nature of the numbering of extended partitions. For example, consider a Solaris system with the partitions p1, p2, p3, and p4. Following creation of an extended partition, the same system has a logical device node, p5, and successive nodes numbered consecutively up to a maximum of p36. If one logical drive is deleted, say, p8, then all nodes following p8 (p9 up to p36) move up one in the list of partitions, so that p9 becomes p8, p10 becomes p9, and so forth.

Use the following options to include your modifications to the partition table at this time or to cancel the session without modifying the table:

Exit

This option writes the new version of the table created during this session with fdisk out to the fixed disk, and exits the program.

Cancel

This option exits without modifying the partition table.

Options

The following options apply to fdisk:

–A id:act:bhead:bsect:bcyl:ehead:esect:ecyl:rsect:numsect

Add a partition as described by the argument (see the –F option below for the format). Use of this option will zero out the VTOC on the Solaris partition if the fdisk table changes.

–b master_boot

Specify the file master_boot as the master boot program. The default master boot program is /usr/lib/fs/ufs/mboot.

–B

Default to one Solaris partition that uses the whole disk. On an x86 machine, if the disk is larger than 2 TB (terabytes), the default size of the Solaris partition will be limited to 2 TB.

–d

Turn on verbose debug mode. This will cause fdisk to print its state on stderr as it is used. The output from this option should not be used with –F.

–D id:act:bhead:bsect:bcyl:ehead:esect:ecyl:rsect:numsect

Delete a partition as described by the argument (see the –F option below for the format). Note that the argument must be an exact match or the entry will not be deleted! Use of this option will zero out the VTOC on the Solaris partition if the fdisk table changes.

–E

Create an EFI partition that uses the entire disk.

–F fdisk_file

Use fdisk file fdisk_file to initialize table. Use of this option will zero out the VTOC on the Solaris partition if the fdisk table changes.

The fdisk_file contains four specification lines for the primary partitions followed by specification lines for the logical drives. You must have four lines for the primary partitions if there is at least one logical drive. In this case, if the number of primary partitions to be configured is less than four, the remaining lines should be filled with zeros.

Each line is composed of entries that are position-dependent, are separated by whitespace or colons, and have the following format:

id act bhead bsect bcyl ehead esect ecyl rsect numsect

...where the entries have the following values:

id

This is the type of partition and the correct numeric values may be found in fdisk.h.

act

This is the active partition flag; 0 means not active and 128 means active. For logical drives, this flag will always be set to 0 even if specified as 128 by the user.

bhead

This is the head where the partition starts. If this is set to 0, fdisk will correctly fill this in from other information.

bsect

This is the sector where the partition starts. If this is set to 0, fdisk will correctly fill this in from other information.

bcyl

This is the cylinder where the partition starts. If this is set to 0, fdisk will correctly fill this in from other information.

ehead

This is the head where the partition ends. If this is set to 0, fdisk will correctly fill this in from other information.

esect

This is the sector where the partition ends. If this is set to 0, fdisk will correctly fill this in from other information.

ecyl

This is the cylinder where the partition ends. If this is set to 0, fdisk will correctly fill this in from other information.

rsect

The relative sector from the beginning of the disk where the partition starts. This must be specified and can be used by fdisk to fill in other fields. For logical drives, you must make sure that there are at least 63 free sectors before the rsect specified for a logical drive.

numsect

The size in sectors of this disk partition. This must be specified and can be used by fdisk to fill in other fields.

–g

Get the label geometry for disk and display on stdout (see the –S option for the format).

–G

Get the physical geometry for disk and display on stdout (see the –S option for the format).

–h

Issue verbose message; message will list all options and supply an explanation for each.

–I

Forgo device checks. This is used to generate a file image of what would go on a disk without using the device. Note that you must use –S with this option (see above).

–n

Don't update fdisk table unless explicitly specified by another option. If no other options are used, –n will only write the master boot record to the disk. In addition, note that fdisk will not come up in interactive mode if the –n option is specified.

–o offset

Block offset from start of disk. This option is used for –P, –r, and –w. Zero is assumed when this option is not used.

–P fill_patt

Fill disk with pattern fill_patt. fill_patt can be decimal or hex and is used as number for constant long word pattern. If fill_patt is #, then pattern is block # for each block. Pattern is put in each block as long words and fills each block (see –o and –s).

–r

Read from disk and write to stdout. See –o and –s, which specify the starting point and size of the operation.

–R

Treat disk as read-only. This is for testing purposes.

–s size

Number of blocks to perform operation on (see –o).

–S geom_file

Set the label geometry to the content of the geom_file. The geom_file contains one specification line. Each line is delimited by a new-line character (\n). If the first character of a line is an asterisk (*), the line is treated as a comment. Each line is composed of entries that are position-dependent, are separated by white space, and have the following format:

pcyl ncyl acyl bcyl nheads nsectors sectsiz

where the entries have the following values:

pcyl

This is the number of physical cylinders for the drive.

ncyl

This is the number of usable cylinders for the drive.

acyl

This is the number of alt cylinders for the drive.

bcyl

This is the number of offset cylinders for the drive (should be zero).

nheads

The number of heads for this drive.

nsectors

The number of sectors per track.

sectsiz

The size in bytes of a sector.

–t

Adjust incorrect slice table entries so that they will not cross partition table boundaries.

–T

Remove incorrect slice table entries that span partition table boundaries.

–v

Output the HBA (virtual) geometry dimensions. This option must be used in conjunction with the –W flag. This option will work for platforms which support virtual geometry. (x86 only)

–w

Write to disk and read from stdin. See –o and –s, which specify the starting point and size of the operation.

–W

Output the disk table to stdout.

–W fdisk_file

Create an fdisk file fdisk_file from disk table. This can be used with the –F option above.

Files

/dev/rdsk/c0t0d0p0

Raw device associated with the fixed disk.

/usr/lib/fs/ufs/mboot

Default master boot program.

Attributes

See attributes(5) for descriptions of the following attributes:

ATTRIBUTE TYPE
ATTRIBUTE VALUE
Architecture
x86 and SPARC
Availability
system/core-os

See also

uname(1), fmthard(1M), format(1M), newfs(1M), parted(1M), prtvtoc(1M), attributes(5)

Diagnostics

Most messages will be self-explanatory. The following may appear immediately after starting the program:

Fdisk: cannot open <device>

This indicates that the device name argument is not valid.

Fdisk: unable to get device parameters for device <device>

This indicates a problem with the configuration of the fixed disk, or an error in the fixed disk driver.

Fdisk: error reading partition table

This indicates that some error occurred when trying initially to read the fixed disk. This could be a problem with the fixed disk controller or driver, or with the configuration of the fixed disk.

Fdisk: error writing boot record

This indicates that some error occurred when trying to write the new partition table out to the fixed disk. This could be a problem with the fixed disk controller, the disk itself, the driver, or the configuration of the fixed disk.