FreeBSD uses two different methods. I am going to focus on the Ports collection since this way seems to be the most flexible, I must warn you though, compiling everything can become a pain in the neck if you have a slow machine. Fortunately for me, all of the servers I am using in the Server Migration are new and quite fast.

The Ports collection is relatively easy to use. Simply go into the ports collection.


server# cd /usr/ports

Now you will need to find the package you want to install (I will forgo the instructions on how to do this, as I am assuming that you know how to do that already.) In my case I started with MySQL 5

server# cd databases/mysql50-server
server# make install clean
server# mysql_install_db
server# chown -R mysql:mysql /var/db/mysql
server# /usr/local/bin/mysqld_safe -user=mysql&
server# mysqladmin -u root password <password>
server# vi /etc/rc.conf

Now you will need to tell the server to automatically start the MySQL daemon when the computer restarts. Add the following line to your rc.conf file (Which you will be editing if you have followed the previous steps.)

mysql_enable=”yes”

There you go, as long as no error messages show up, your MySQL 5.0 server will be running and will be usable.

I had previously decided to purchase some servers (rather than use dedicated servers spread around the nation) and move into a local data center. Since I work for a company that provides web hosting and other data center services, (such as colocation, which is what I needed) I got some rack space in the Fibernet & Nethosting data center and ordered my servers.

The servers have finally started to come in and I am slowly setting them up and moving them into the data center. I am planning on sharing all of my wo’s involved in migrating all of my servers and I am also planning on sharing my solutions to the problems

Return often to find out about my trials and solutions.

Although this frame work provides a ton of flexibility and power over the database associations and the layout of the site, I quickly discovered a large number of things that are either incredibly difficult to do, or that are impossible to do. The sad thing about this, is that these are basic database types of things that I ran into. I could understand if it were something very complex that no one does, but it is stuff that is an everyday kind of thing to do. Let me put forth an example:

Here is a very basic database to use with our example.

Videos
--------
id as int(11)
name as varchar(255)
icon as varchar(255)
path as varchar(255)
user_id as int(11)

Users
-------
id as int(11)
first_name as varchar(255)
last_name as varchar(255)
username as varchar(32)
password as varchar(32)
company_id as int(11)

Companies
--------------
id as int(11)
name as varchar(255)
description as text

As you can see, we have a very basic example that consists of Videos that are attached to users. Users are then attached to a company. I simply wanted to use cake to read in each video and use the model association of belongsTo to attach both a company and a user. (This is because each video only has one user, and each user only has one company.) This should be easy in CakePHP right? Wrong!

After beating myself in the head for hours trying to figure out how to get the company to attach properly, I decided to dig into the code of CakePHP itself. I quickly discovered that the framework is there to allow for this type of thing, is simply doesn’t work. After a few minutes of hacking on the core CakePHP sources, I was able to devise a solution. My solution even though it is not graceful, works.

The worst part of my experience was not running into this issue though. The thing that drove me nuts was that I tried to talk to the developers in the IRC channel for CakePHP and was quickly snubbed by the developers after trying to get them to give me an answer on how to get this to work properly I was banned from the channel. WOW. Way to go open source community. Want to know why the corporate world will never take you serious? It is because you can’t handle constructive criticism.

Anyway, overall I really do like CakePHP, but there is an issue here that simply drives me nuts. HELP OUT YOUR USERS!

mysql> ALTER tablename AUTO_INCREMENT=new_number

There you go, it is that easy to change the auto increment value for MySQL.

NOTE: I have discovered that MySQL will not allow you to set the auto_increment value to a number smaller that what is currently set. The data must be truncated and then set to get it to work properly.

These games are designed to support a gigantic user base; in fact, the developers of these games do an amazing job of setting up an environment capable of supporting thousands of users. Unfortunately, the current networking system used by these games is limited in two ways. First, an enormous number of servers are required to allow these simulated environments to function. Second, an exuberant amount of Internet bandwidth is consumed by these games. Considering these factors, these games do not scale well at all.
I am writing a research paper designed to address both of these issues. It will discuss different ways of load balancing the actual game play, allowing a minimal internet connection, and using just a few servers to support the gaming environment.

In addition, it will also include code snippets from a small sample game that will be used to show example code of these load balancing algorithms and practices at work. My goal with this project is to increase my knowledge of how these types of games work as well as create an environment that is not limited to only a few hundred thousand poeple. I want a place where millions can interact with each other. Come back for more information. This project is under a huge amount of work right now, and I will post more as it is available.

The firstthing to do is to set up the terminal to function properly. Open something like Hyperterminal and use the following settings:

• 9600 baud rate
• No parity
• 8 data bits
• 1 stop bit
• No flow control

Ok, now that the terminal is set. Connect to the Cisco router, and power cycle it. Within the first 60 seconds, press the Break key on the key board. This will put the router in ROMMON. At the prompt that comes up type confreg 0×2124 to boot the flash.

That step skips the startup configuration, ingoring the passwords.

Now type reset and wait for the router to reboot. This will start the router up without loading the saved configuration. It will act like a brand new router. Once you are at the prompt type in enable you will now be in enable mode. The prompt should be Router#.

Now type in configure terminal and use the enable secret <password> to set the password for your router. The next thing that you should type is config-register 0×2102.

Now type end or Ctrl-z this will leave configuration mode. Type in write memory or copy running-config startup-config to commit and save changes. You will now be able to set up your router like you know (or want to know)

To reset your FreeBSD password simple boot the system into single user mode (done in one of two ways; first, if you are using FreeBSD 5.4 or earlier, you simple press the space bar at the appropirate time, which the computer will tell you. Then simple type in boot -s; second, simply select option#four from the boot up menu, it says boot into single user mode.

The system will prompt for the shell you want to use, simple choose the default /bin/sh this will boot the system to a command prompt. Now enter the following commands in sequence:

mount -u /

mount -a

passwd (It will prompt for a new root password, enter it twice here.)

sync; sync

reboot

You will now have root access to you system once more. Good luck

First I added a helo restriction using this:

smtpd_helo_required = yes
smtpd_helo_restrictions =
permit_mynetworks,
check_helo_access hash:/etc/postfix/helo_access,
reject_non_fqdn_hostname,
reject_invalid_hostname,
permit

This eliminated a ton of my spam. The helo_access file allows me
to make exceptions for my clients that have broken networks. In
addition, this file includes rejects for anything coming from itself.
(Handy since most Spammers try to pretend they are you, hoping to get
around your relay restrictions) It looks similar to this:
mydomain.com REJECT You are not me!
localhost REJECT You are not me!
127.0.0.1 REJECT You are not me!
localhost.localdomain REJECT You are not me!

Next I added this to my main.cf
smtpd_sender_restrictions =
permit_sasl_authenticated,
permit_mynetworks,
reject_non_fqdn_sender,
reject_unknown_sender_domain,
permit
Forcing everything to use a fully qualified domain name helped
eliminate a ton of spam. The next item I did was the last of the light
weight stuff, this catches almost everything else:
smtpd_recipient_restrictions =
reject_unauth_pipelining,
reject_non_fqdn_recipient,
reject_unknown_recipient_domain,
permit_mynetworks,
permit_sasl_authenticated,
reject_unauth_destination
check_sender_access
hash:/etc/postfix/sender_access,
check_recipient_access
hash:/etc/postfix/recipient_access,
check_helo_access
hash:/etc/postfix/secondary_mx_access,
reject_rbl_client list.dsbl.org
reject_rbl_client sbl-xbl.spamhaus.org,
permit

The big stuff here is the reject_rbl_client lines. These line check
the incoming server against relay databases. I used these lists because they
seemed relatively fair (I don’t want valid email getting rejected) and
people are able to get off of these lists since they don’t seem to be
Nazi’s about it.

After adding these items, pretty much the only email getting to ClamAV
and SpamAssassin are valid email messages. This brought the load of my
sever back into a managable place (4-5% CPU load) and makes email
delivery fast again.