ufs - UFS file system
UFS is an optional disk-based file system for the Oracle Solaris environment. The UFS file system is hierarchical, starting with its root directory (/ ) and continuing downward through a number of directories. The root of a UFS file system is inode 2. A UFS file system's root contents replace the contents of the directory upon which it is mounted.
Subsequent sections of this manpage provide details of the UFS file systems.
UFS uses state flags to identify the state of the file system. fs_state is FSOKAY - fs_time. fs_time is the timestamp that indicates when the last system write occurred. fs_state is updated whenever fs_clean changes. Some fs_clean values are:
Indicates an undamaged, cleanly unmounted file system.
Indicates a mounted file system that has modified data in memory. A mounted file system with this state flag indicates that user data or metadata would be lost if power to the system is interrupted.
Indicates an idle mounted file system. A mounted file system with this state flag indicates that neither user data nor metadata would be lost if power to the system is interrupted.
Indicates that this file system contains inconsistent file system data.
Indicates that the file system has logging enabled. A file system with this flag set is either mounted or unmounted. If a file system has logging enabled, the only flags that it can have are FSLOG or FSBAD. A non-logging file system can have FSACTIVE, FSSTABLE, or FSCLEAN.
It is not necessary to run the fsck command on unmounted file systems with a state of FSCLEAN, FSSTABLE, or FSLOG. mount(2) returns ENOSPC if an attempt is made to mount a UFS file system with a state of FSACTIVE for read/write access.
As an additional safeguard, fs_clean should be trusted only if fs_state contains a value equal to FSOKAY - fs_time, where FSOKAY is a constant integer defined in the /usr/include/sys/fs/ufs_fs.h file. Otherwise, fs_clean is treated as though it contains the value of FSACTIVE.
Extended Fundamental Types (EFT) provide 32-bit user ID (UID), group ID (GID), and device numbers.
If a UID or GID contains an extended value, the short variable (ic_suid , ic_sgid) contains the value 65535 and the corresponding UID or GID is in ic_uid or ic_gid. Because numbers for block and character devices are stored in the first direct block pointer of the inode (ic_db) and the disk block addresses are already 32 bit values, no special encoding exists for device numbers (unlike UID or GID fields).
A multiterabyte file system enables creation of a UFS file system up to approximately 16 terabytes of usable space, minus approximately one percent overhead. A sparse file can have a logical size of one terabyte. However, the actual amount of data that can be stored in a file is approximately one percent less than one terabyte because of file system overhead.
On-disk format changes for a multiterabyte UFS file system include:
The magic number in the superblock changes from FS_MAGIC to MTB_UFS_MAGIC. For more information, see the /usr/include/sys/fs/ufs_fs file.
The fs_logbno unit is a sector for UFS that is less than 1 terabyte in size and fragments for a multiterabyte UFS file system.
UFS logging bundles the multiple metadata changes that comprise a complete UFS operation into a transaction. Sets of transactions are recorded in an on-disk log and are applied to the actual UFS file system's metadata.
UFS logging provides two advantages:
A file system that is consistent with the transaction log eliminates the need to run fsck after a system crash or an unclean shutdown.
UFS logging often provides a significant performance improvement. This is because a file system with logging enabled converts multiple updates to the same data into single updates, thereby reducing the number of overhead disk operations.
The UFS log is allocated from free blocks on the file system and is sized at approximately 1 Mbyte per 1 Gbyte of file system, up to 256 Mbytes. The log size may be larger (up to a maximum of 512 Mbytes), depending upon the number of cylinder groups present in the file system. The log is continually flushed as it fills up. The log is also flushed when the file system is unmounted or as a result of a lockfs(8) command.
You can mount a UFS file system in various ways using syntax similar to the following:
Use mount from the command line:
# mount -F ufs /dev/dsk/c0t0d0s7 /export/home
Include an entry in the /etc/vfstab file to mount the file system at boot time:
/dev/dsk/c0t0d0s7 /dev/rdsk/c0t0d0s7 /export/home ufs 2 yes -
For more information on mounting UFS file systems, see mount_ufs(8).
See attributes(7) for a description of the following attributes:
For information about internal UFS structures, see newfs(8) and mkfs_ufs(8). For information about the ufsdump and ufsrestore commands, see ufsdump(8), ufsrestore(8), and /usr/include/protocols/dumprestore.h.
If you experience difficulty in allocating space on the ufs filesystem, it may be due to fragmentation. Fragmentation can occur when you do not have sufficient free blocks to satisfy an allocation request even though df(8) indicates that enough free space is available. (This may occur because df only uses the available fragment count to calculate available space, but the file system requires contiguous sets of fragments for most allocations). If you suspect that you have exhausted contiguous fragments on your file system, you can use the fstyp(8) utility with the -v option. In the fstyp output, look at the nbfree (number of blocks free) and nffree (number of fragments free) fields. On unmounted filesystems, you can use fsck(8) and observe the last line of output, which reports, among other items, the number of fragments and the degree of fragmentation. To correct a fragmentation problem, run ufsdump(8) and ufsrestore(8) on the ufs filesystem.