18.5 The HighPerformance File System | HP Certified: HP-UX System Administration

HFS is the legacy file system used with HP-UX. Although JFS is used in the newer versions of HP-UX, HFS is still in use for the /stand file system and is supported on all HP-UX releases.

The first 8 Kbytes of all HFS file systems contain the HFS superblock, which contains general information and pointers to the metadata area as already discussed. HFS contains more than one copy of the superblock, and the locations of these redundant copies are recorded in the /var/adm/sbtab file. If the main superblock is damaged, it can be recovered from one of the backup copies. If you use the cat command, this file looks like the following.

 /dev/vg00/rlvol1: super-block backups (for fsck -b#) at:     16,   2504,   4992,   7480,   9968,  12456,  14944,  17432,  19728,  22216,  24704,  27192,  29680,  32168,  34656,  37144,  39440,  41928,  44416,  46904,  49392,  51880,  54368,  56856,  59152,  61640,  64128,  66616,  69104,  71592,  74080,  76568,  78864,  81352,  83840,  86328,  88816,  91304,  93792,  96280,  98576, 101064

The remaining HFS area is divided into one or more cylinder groups. Each cylinder group contains a data area as well as inodes. You can specify the number of inodes when you create the file system. The disadvantage is that if you create too few inodes, you can run short. It may happen that the data area space is still vacant, but you can't create new files as no free inode is available. There is no way to increase the number of inodes without allocating more disk space to the file system or recreating the file system, damaging all of the data stored on it.

HFS is easy to manage. However, you can't reduce its size . Another disadvantage is that you must unmount the file system if you want to increase its size.

HFS Blocks

The HFS block is the minimum data that are read in one read request. The default HFS block size is 8 Kbytes. You can use block sizes of 4, 8, 16, 32, or 64 Kbytes when you create a file system. So, in case you are using a 64-Kbyte block size, a request for reading even 512 bytes will need 64 Kbytes to be read; a large block size for small files may be inefficient to use. Similarly, if you are using large files on a file system and the block size is small, the system may have to make many read requests when reading a large amount of data. And if these blocks are scattered across the disk, it may take a longer time for the read head to locate the data. You must be careful when selecting the block size, keeping in mind the size of the files that will be stored on the file system.

The block size cannot be changed once a file system is created.

HFS Fragments

A fragment is the smallest unit of data that can be allocated to a file. HFS blocks are divided into fragments. A block may contain one, two, four or eight fragments. The number of fragments and the fragment size are also set at the time of file system creation and cannot be changed later. A block may contain fragments related to different files.

Top