1.\" Copyright (c) 1983, 1991, 1993 2.\" The Regents of the University of California. All rights reserved. 3.\" 4.\" %sccs.include.redist.man% 5.\" 6.\" @(#)tunefs.8 8.1 (Berkeley) 06/05/93 7.\" 8.Dd 9.Dt TUNEFS 8 10.Os BSD 4.2 11.Sh NAME 12.Nm tunefs 13.Nd tune up an existing file system 14.Sh SYNOPSIS 15.Nm tunefs 16.Op Fl a Ar maxcontig 17.Op Fl d Ar rotdelay 18.Op Fl e Ar maxbpg 19.Op Fl m Ar minfree 20.Bk -words 21.Op Fl o Ar optimize_preference 22.Ek 23.Op Ar special | Ar filesys 24.Sh DESCRIPTION 25.Nm Tunefs 26is designed to change the dynamic parameters of a file system 27which affect the layout policies. 28The parameters which are to be changed are indicated by the flags 29given below: 30.Bl -tag -width Ds 31.It Fl a Ar maxcontig 32This specifies the maximum number of contiguous blocks that will 33be laid out before forcing a rotational delay (see 34.Fl d 35below). 36The default value is one, since most device drivers require 37an interrupt per disk transfer. 38Device drivers that can chain several buffers together in a single 39transfer should set this to the maximum chain length. 40.It Fl d Ar rotdelay 41This specifies the expected time (in milliseconds) 42to service a transfer completion 43interrupt and initiate a new transfer on the same disk. 44It is used to decide how much rotational spacing to place between 45successive blocks in a file. 46.It Fl e Ar maxbpg 47This indicates the maximum number of blocks any single file can 48allocate out of a cylinder group before it is forced to begin 49allocating blocks from another cylinder group. 50Typically this value is set to about one quarter of the total blocks 51in a cylinder group. 52The intent is to prevent any single file from using up all the 53blocks in a single cylinder group, 54thus degrading access times for all files subsequently allocated 55in that cylinder group. 56The effect of this limit is to cause big files to do long seeks 57more frequently than if they were allowed to allocate all the blocks 58in a cylinder group before seeking elsewhere. 59For file systems with exclusively large files, 60this parameter should be set higher. 61.It Fl m Ar minfree 62This value specifies the percentage of space held back 63from normal users; the minimum free space threshold. 64The default value used is 10%. 65This value can be set to zero, however up to a factor of three 66in throughput will be lost over the performance obtained at a 10% 67threshold. 68Note that if the value is raised above the current usage level, 69users will be unable to allocate files until enough files have 70been deleted to get under the higher threshold. 71.It Fl o Ar optimize_preference 72The file system can either try to minimize the time spent 73allocating blocks, or it can attempt minimize the space 74fragmentation on the disk. 75If the value of minfree (see above) is less than 10%, 76then the file system should optimize for space to avoid 77running out of full sized blocks. 78For values of minfree greater than or equal to 10%, 79fragmentation is unlikely to be problematical, and 80the file system can be optimized for time. 81.El 82.Sh SEE ALSO 83.Xr fs 5 , 84.Xr dumpfs 8 , 85.Xr newfs 8 , 86.Xr mkfs 8 87.Rs 88.%A M. McKusick 89.%A W. Joy 90.%A S. Leffler 91.%A R. Fabry 92.%T "A Fast File System for UNIX" 93.%J "ACM Transactions on Computer Systems 2" 94.%N 3 95.%P pp 181-197 96.%D August 1984 97.%O "(reprinted in the BSD System Manager's Manual, SMM:5)" 98.Re 99.Sh BUGS 100This program should work on mounted and active file systems. 101Because the super-block is not kept in the buffer cache, 102the changes will only take effect if the program 103is run on dismounted file systems. 104To change the root file system, the system must be rebooted 105after the file system is tuned. 106.Pp 107You can tune a file system, but you can't tune a fish. 108.Sh HISTORY 109The 110.Nm 111command appeared in 112.Bx 4.2 . 113