projects / unix-history / blame
| Commit | Line | Data |
|---|---|---|
| 868d1828 C |
1 | .TH TUNEFS 8 "20 February 1983" |
| 2 | .UC 4 | |
| 3 | .SH NAME | |
| 4 | tunefs \- tune up an existing file system | |
| 5 | .SH SYNOPSIS | |
| 6 | .B /etc/tunefs | |
| 7 | .I tuneup-options | |
| 8 | .IR special | filesys | |
| 9 | .SH DESCRIPTION | |
| 10 | .I Tunefs | |
| 11 | is designed to change the dynamic parameters of a file system | |
| 12 | which affect the layout policies. | |
| 13 | The parameters which are to be changed are indicated by the flags | |
| 14 | given below: | |
| 15 | .IP "\fB\-a\fP maxcontig" | |
| 16 | .br | |
| 17 | This specifies the maximum number of contiguous blocks that will | |
| 18 | be laid out before forcing a rotational delay (see \-d below). | |
| 19 | The default value is one, since most device drivers require | |
| 20 | an interrupt per disk transfer. | |
| 21 | Device drivers that can chain several buffers together in a single | |
| 22 | transfer should set this to the maximum chain length. | |
| 23 | .IP "\fB\-d\fP rotdelay" | |
| 24 | .br | |
| 25 | This specifies the expected time (in milliseconds) | |
| 26 | to service a transfer completion | |
| 27 | interrupt and initiate a new transfer on the same disk. | |
| 28 | It is used to decide how much rotational spacing to place between | |
| 29 | successive blocks in a file. | |
| 30 | .IP "\fB\-e\fP maxbpg" | |
| 31 | .br | |
| 32 | This indicates the maximum number of blocks any single file can | |
| 33 | allocate out of a cylinder group before it is forced to begin | |
| 34 | allocating blocks from another cylinder group. | |
| 35 | Typically this value is set to about one quarter of the total blocks | |
| 36 | in a cylinder group. | |
| 37 | The intent is to prevent any single file from using up all the | |
| 38 | blocks in a single cylinder group, | |
| 39 | thus degrading access times for all files subsequently allocated | |
| 40 | in that cylinder group. | |
| 41 | The effect of this limit is to cause big files to do long seeks | |
| 42 | more frequently than if they were allowed to allocate all the blocks | |
| 43 | in a cylinder group before seeking elsewhere. | |
| 44 | For file systems with exclusively large files, | |
| 45 | this parameter should be set higher. | |
| 46 | .IP "\fB\-m\fP minfree" | |
| 47 | .br | |
| 48 | This value specifies the percentage of space held back | |
| 49 | from normal users; the minimum free space threshold. | |
| 50 | The default value used is 10%. | |
| 51 | This value can be set to zero, however up to a factor of three | |
| 52 | in throughput will be lost over the performance obtained at a 10% | |
| 53 | threshold. | |
| 54 | Note that if the value is raised above the current usage level, | |
| 55 | users will be unable to allocate files until enough files have | |
| 56 | been deleted to get under the higher threshold. | |
| 57 | .SH "SEE ALSO" | |
| 58 | fs(5), | |
| 59 | newfs(8), | |
| 60 | mkfs(8) | |
| 61 | .PP | |
| 62 | McKusick, Joy, Leffler; "A Fast File System for Unix", | |
| 63 | Computer Systems Research Group, Dept of EECS, Berkeley, CA 94720; | |
| 64 | TR #7, September 1982. | |
| 65 | .SH BUGS | |
| 66 | This program should work on mounted and active file systems. | |
| 67 | Because the super-block is not kept in the buffer cache, | |
| 68 | the program will only take effect if it is run on dismounted file systems. | |
| 69 | (if run on the root file system, the system must be rebooted) | |
| 70 | .PP | |
| 71 | You can tune a file system, but you can't tune a fish. |