.Xr disklabel 8

.Xr disklabel 8

to lay out your filesystems on a hard disk it is important to remember

to lay out your filesystems on a hard disk it is important to remember

that hard drives can transfer data much more quickly from outer tracks

that hard drives can transfer data much more quickly from outer tracks

try to pack your smaller filesystems and swap closer to the outer tracks,

try to pack your smaller filesystems and swap closer to the outer tracks,

follow with the larger filesystems, and end with the largest filesystems.

follow with the larger filesystems, and end with the largest filesystems.

It is also important to size system standard filesystems such that you

It is also important to size system standard filesystems such that you

in the partition table) will increase I/O performance in the partitions 

in the partition table) will increase I/O performance in the partitions 

where you need it the most.  Now it is true that you might also need I/O

where you need it the most.  Now it is true that you might also need I/O

performance in the larger partitions, but they are so large that shifting

performance in the larger partitions, but they are so large that shifting

performance improvement whereas moving /var to the edge can have a huge impact.

performance improvement whereas moving /var to the edge can have a huge impact.

Finally, there are safety concerns.  Having a small neat root partition that

Finally, there are safety concerns.  Having a small neat root partition that

is essentially read-only gives it a greater chance of surviving a bad crash

is essentially read-only gives it a greater chance of surviving a bad crash

recovery times after a crash.  Do not use this option

recovery times after a crash.  Do not use this option

unless you are actually storing large files on the partition, because if you

unless you are actually storing large files on the partition, because if you

overcompensate you can wind up with a filesystem that has lots of free

overcompensate you can wind up with a filesystem that has lots of free

32768, 65536, or 262144 bytes/inode is recommended.  You can go higher but

32768, 65536, or 262144 bytes/inode is recommended.  You can go higher but

it will have only incremental effects on fsck recovery times.  For

it will have only incremental effects on fsck recovery times.  For

example, 

example, 

Softupdates drastically improves meta-data performance, mainly file

Softupdates drastically improves meta-data performance, mainly file

creation and deletion.  We recommend turning softupdates on on all of your

creation and deletion.  We recommend turning softupdates on on all of your

filesystems.  There are two downsides to softupdates that you should be

filesystems.  There are two downsides to softupdates that you should be

case of a crash but could very easily be several seconds (even a minute!)

case of a crash but could very easily be several seconds (even a minute!)

behind updating the physical disk.  If you crash you may lose more work

behind updating the physical disk.  If you crash you may lose more work

blocks.  If you have a filesystem (such as the root filesystem) which is 

blocks.  If you have a filesystem (such as the root filesystem) which is 

close to full, doing a major update of it, e.g.

close to full, doing a major update of it, e.g.

.Em make installworld,

.Em make installworld,

time.  You should only stripe partitions that require serious I/O performance...

time.  You should only stripe partitions that require serious I/O performance...

typically /var, /home, or custom partitions used to hold databases and web

typically /var, /home, or custom partitions used to hold databases and web

pages.  Choosing the proper stripe size is also 

pages.  Choosing the proper stripe size is also 

means you want to use a large off-center stripe size such as 1152 sectors

means you want to use a large off-center stripe size such as 1152 sectors

so sequential I/O does not seek both disks and so meta-data is distributed

so sequential I/O does not seek both disks and so meta-data is distributed

you really need to get sophisticated, we recommend using a real hardware

you really need to get sophisticated, we recommend using a real hardware

raid controller from the list of

raid controller from the list of

.Fx

.Fx

the VM Page Cache to cache the directories.  The advantage is that all of

the VM Page Cache to cache the directories.  The advantage is that all of

memory is now available for caching directories.  The disadvantage is that

memory is now available for caching directories.  The disadvantage is that

the minimum in-core memory used to cache a directory is the physical page

the minimum in-core memory used to cache a directory is the physical page

on if you are running any services which manipulate large numbers of files.

on if you are running any services which manipulate large numbers of files.

Such services can include web caches, large mail systems, and news systems.

Such services can include web caches, large mail systems, and news systems.

Turning on this option will generally not reduce performance even with the

Turning on this option will generally not reduce performance even with the

improve bandwidth utilization by increasing the default at the cost of 

improve bandwidth utilization by increasing the default at the cost of 

eating up more kernel memory for each connection.  We do not recommend

eating up more kernel memory for each connection.  We do not recommend

increasing the defaults if you are serving hundreds or thousands of

increasing the defaults if you are serving hundreds or thousands of

out of memory due to stalled connections building up.  But if you need

out of memory due to stalled connections building up.  But if you need

high bandwidth over a fewer number of connections, especially if you have

high bandwidth over a fewer number of connections, especially if you have

gigabit ethernet, increasing these defaults can make a huge difference.

gigabit ethernet, increasing these defaults can make a huge difference.

without eating too much kernel memory.  Note that the route table, see

without eating too much kernel memory.  Note that the route table, see

.Xr route 8 ,

.Xr route 8 ,

can be used to introduce route-specific send and receive buffer size

can be used to introduce route-specific send and receive buffer size

firewall rules, see

firewall rules, see

.Xr ipfw 8 ,

.Xr ipfw 8 ,

to limit the bandwidth going to or from particular IP blocks or ports.

to limit the bandwidth going to or from particular IP blocks or ports.

We recommend that you turn on (set to 1) and leave on the 

We recommend that you turn on (set to 1) and leave on the 

.Em net.inet.tcp.always_keepalive

.Em net.inet.tcp.always_keepalive

control.  The default is usually off.  This introduces a small amount of

control.  The default is usually off.  This introduces a small amount of

will eventually be recognized and cleared.  Dead tcp connections are a

will eventually be recognized and cleared.  Dead tcp connections are a

because users often disconnect their modems without properly closing active

because users often disconnect their modems without properly closing active

connections.

connections.

.Pp

.Pp

willing to allocate.  Each cluster represents approximately 2K of memory,

willing to allocate.  Each cluster represents approximately 2K of memory,

so a value of 1024 represents 2M of kernel memory reserved for network

so a value of 1024 represents 2M of kernel memory reserved for network

buffers.  You can do a simple calculation to figure out how many you need.

buffers.  You can do a simple calculation to figure out how many you need.

and each connection eats a 16K receive and 16K send buffer, you need

and each connection eats a 16K receive and 16K send buffer, you need

approximate 32MB worth of network buffers to deal with it.  A good rule of

approximate 32MB worth of network buffers to deal with it.  A good rule of

thumb is to multiply by 2, so 32MBx2 = 64MB/2K = 32768.  So for this case

thumb is to multiply by 2, so 32MBx2 = 64MB/2K = 32768.  So for this case

.Sh CPU, MEMORY, DISK, NETWORK

.Sh CPU, MEMORY, DISK, NETWORK

The type of tuning you do depends heavily on where your system begins to

The type of tuning you do depends heavily on where your system begins to

bottleneck as load increases.  If your system runs out of cpu (idle times

bottleneck as load increases.  If your system runs out of cpu (idle times

an SMP motherboard (multiple cpu's), or perhaps you need to revisit the

an SMP motherboard (multiple cpu's), or perhaps you need to revisit the

programs that are causing the load and try to optimize them.  If your system

programs that are causing the load and try to optimize them.  If your system

is paging to swap a lot you need to consider adding more memory.  If your

is paging to swap a lot you need to consider adding more memory.  If your

.Xr firewall 7

.Xr firewall 7

we describe a firewall protecting internal hosts with a topology where

we describe a firewall protecting internal hosts with a topology where

the externally visible hosts are not routed through it.  Use 100BaseT rather

the externally visible hosts are not routed through it.  Use 100BaseT rather

Most bottlenecks occur at the WAN link (e.g. modem, T1, DSL, whatever).

Most bottlenecks occur at the WAN link (e.g. modem, T1, DSL, whatever).

If expanding the link is not an option it may be possible to use ipfw's

If expanding the link is not an option it may be possible to use ipfw's

.Sy DUMMYNET

.Sy DUMMYNET