Administering filesystems

Filesystem mount options (DTFS)

The Filesystem Manager supports the following mount options for non-root DTFS filesystems; for root filesystems, see ``Modifying DTFS root filesystem mount configuration''.

compresses the data on your hard disk, thereby increasing the storage capacity on your disk. The compression ratio depends on the type of file. This feature is enabled by default.

writes files to disk as soon as they are closed. This ensures data integrity and robustness because you can switch the machine off immediately afterward without loss of data.

Maximum number of file versions
sets maximum number of undeletable (versioned) files allowed on the filesystem. A value of 0 disables versioning.

Minimum time before a file is versioned
sets minimum time before a file is versioned. If set to 0, a file is always versioned (as long as Maxiumum number of file versions is greater than 0). If set to a value greater than 0, the file is versioned after it has existed for that number of seconds.

WARNING: The sync-on-close feature can significantly degrade system performance because of the increased writing to disk.

See also:

Data compression

The Compression Filesystem (DTFS) uses transparent data compression and a more efficient disk media format to increase the storage capacity of your disks. The compression is performed prior to the writing of file-data blocks to disk. It is designed primarily for use on systems where disk space is limited, such as personal workstations. The compression ratios are dependent on the data type.

Average disk savings

File type Average savings (%)
Directories 34
Executable programs 36
Program source files 40
Archives and dynamically loaded libraries 42
Symbolic links 50
Binary data (bitmaps, word processing, databases, spreadsheets, and so on) 55
ASCII data (log files, configuration files, and so on) 60

Standard utilities may be used to perform backup and restore operations on this filesystem.

On other filesystems, the number of disk blocks reported by ls -s does not include the amount of space consumed by the disk inode structure. With DTFS, the size of the disk inode is included in the number of blocks reported. This might make a small file appear as if it occupies more space on a DTFS filesystem when compared to other filesystems.

When displaying disk usage information, you can display:

For example, entering:

cd /usr/adm
ls -ls messages

might show:

   82 -rw-r--r--   1 bin    bin    106295 Apr 08 23:01 messages
indicating that the messages file, whose logical size is 106,295 bytes, consumes only 82 disk blocks, or a total of 41,984 bytes of disk space. This represents a 60% savings.

The du, df, and quot utilities display compressed sizes in blocks.

Forced data writes to disk

Sync-on-close ensures that all files modified by a process are written back to the disk when they are closed. This minimizes data loss in the event of power failure.

In addition, the filesystem is transitioned to a quiescent state every second. DTFS does not depend on the standard filesystem flusher functionality (bdflush) to synchronize user data; it does this itself.

In addition, a feature called ``shadow paging'' ensures that new file data is first written to shadow blocks, leaving the original blocks unmodified. If a system failure occurs before the new data is written to disk, the original data is still intact because the blocks were not freed.

These features mean that as soon as you have saved the file, you can switch off the power and the file will be on the disk. This is particularly useful in desktop and laptop environments.

WARNING: Some caching disk controllers use write-back caching techniques that nullify the improved data reliability and integrity provided by DTFS. If this is the case, data integrity cannot be guaranteed. However, controllers that use write-through caching take advantage of DTFS features.

If your controller supports both caching techniques, configure it to use write-through caching.

If sync-on-close is enabled, it will noticeably degrade your overall system performance because of the time spent writing to the disk.

Next topic: Filesystem mount options (High Sierra and ISO9660)
Previous topic: Versioning filesystems (undelete)

© 2003 Caldera International, Inc. All rights reserved.
SCO OpenServer Release 5.0.7 -- 11 February 2003