pfSense

Fix path for relayctl. Fixes #739

only reload if relayd is already running, don't kill and then start it

fix some of the logging for load balancer, still partially broken

couple load balancer fixes, resolves #723

Prevent bootup code from dying on empty load_balancer arrays

Rework includes/require. This saves about 4 megabytes.
Simplify get_memory(). Tested on mips/i386

Do not write out <sitedown></sitedown> if the Fall Back Pool is set to none.. Simply unset() it.

Add pfSense_BUILDER_BINARIES: and pfSense_MODULE: additions

Do not include things just for being able too. Reduce a it to only config.inc

Only attempt starting relayd when the vs array is not empty

oops, changed a line in testing, revert

Comment code on why primary pools port is in use - the front end should
probably only show fallback pools with the same port, but that will be
difficult if not impossible to put in place until we have a better AJAX
framework in place.

Only print out protocols that have actions assigned

Fix generation of protocol line

No need to use a reference to the array, if it doesn't exist already
this tends to make it exist the next time a write_config() occurs
which is now happening during bootup thanks to OpenVPN...yay! ;)

Bring in relay options for inbound load balancer
More work pending, this seems to generate proper configs
needs much more testing though

Ooops, remove leftover debug code

oh how I hate thee php array

• The ability to create your own monitors (tcp, http, https, send/expect)
• A new way of displaying 'standard tables' - MainTable class
• monitor_type array type

add comment with an XXX to remind me to fix

Raise the timeout value to 1 second on checks, setting a global timeout does
no good. See http://readlist.com/lists/openbsd.org/misc/15/78304.html
many thanks to Alex Volkov who pointed me to this thread

allow relayd to actually restart
sitedown isn't required, but may still be set to nothing, don't write out
a bogus config line in this situation

pkill relayd instead of using relayctl due to it losing the control connection (or whatever it is)

Bring in relayd to perform server load balancing
Move gateway load balancing code into gwlb.inc - still uses slbd

TODO: vs and pool status screens are currently broken...and wouldn't work
with the gateway pools anyway, ultimately, the gateway pools need...

Correct is_array() check.

Suggested-by: Bill M

Check for array type before foreach()

Reminded-by: Ryan Wagoner/Seth Mos

Do not foreach through an item that does not exist

Unlink pool file only if it exists

Unlink pool file only if it exists

Remove load balancer pool status file when we reconfigure a pool.

Sleep for 2 seconds between restarting slbd

Handle a very unique edge case when using load balancing plus a dhcp wan. If the DHCP wan goes down, then when we add back the route freebsd will simply prefer the working wan.

• assign interfaces to pool instead of manual IP configuration
• Only show interfaces with a gateway
• Offer list of monitor IPs including interface gateways.

This code is tested for outbound load balancing on a carp cluster and...

backout last commit

Reload filter on boot.

MFC 11474
kill slbd if there are no pools or virtual servers

MFC ougoing lb monitoring fixes

Counter from 1, not 0. Slbd no likie.

MFC from head.
Atleast the code that I wrote prior was pretty much the same.

Commit updated file for inspection

Add ping support. Not tested, not mother approved.

We're in 2006 now, toto

Add round-robin method

Ticket #626

Woops! Use correct field.

Add and honour sitedown slbd field

Only iterate through item if its an array

Allow SLBD to start.

add poolname

$Id -> $Id$

missing paren

only start slbd if we have a load balancer array w/ pools

Change polling interval

Tie load balance front end into the slbd daemon