Using mkfs to create a VxFS filesystem

When used to make a VxFS filesystem, the mkfs(ADM) command builds a filesystem with a root directory and a lost+found directory. VxFS supports three disk layout versions. By default, mkfs creates a Version 4 layout filesystem which supports filesets, dynamic inode allocation, ACLs, large files, and quotas. mkfs does not assign a fixed number of inodes to the filesystem unless the Version 1 disk layout is specified (using the version=1 option). However, if the filesystem is to be used by applications that demand System V Release 3 compatibility (inode numbers <64K), mkfs allows you to specify an upper bound on inode allocation using the -C option.

To create a new filesystem or convert an old one to a new logical block size, use the following procedure:

1. If the new filesystem is to be created on a disk partition that contains an old filesystem, back up the old filesystem.

   If the new filesystem is to be created from an old filesystem, run the labelit(ADM) command, which reports the mounted filesystem name and the physical volume name of the old filesystem. See volcopy(ADM). These labels are destroyed when you make the new filesystem, so you must restore them.

2. Use the mkfs_vxfs(ADM) command to make the new filesystem. with the options described in the manual page.

3. Run the labelit(ADM) command to restore the filesystem and volume names.

4. Populate the new filesystem -- for example, do a restore from a filesystem backup, or, if your system has two hard disks, run cpio(C) from a mounted filesystem. (The volcopy(ADM) and dd(C) commands copy a filesystem image; they cannot convert logical block size.)