Part two of a six part series looking at ClearOS, one of the major commercial alternatives to Small Business Server. In this post I look at the initial installation, setup, and configuration of the operating system. Be sure to read part one for an overview of ClearOS.
Part Two: Installation
In part one we looked at the overall platform functions for ClearOS, and what it can offer a small business as a collaboration platform. Let us take a look at its installation and configuration process to see what is required to get it up and running.
Linux, and ClearOS by extension can run on very minimal hardware, but of course, that’s all dependent on the configuration and applications that are running. ClearOS recommends that if you have 50 users you should have 2GB of RAM, and their web page advises that messaging and firewall applications “following software modules are processor and memory intensive”. Zarafa messaging recommends 4GB of RAM for 50GB of mail. Therefore a server with 8 GB of RAM and a modern processor should meet the requirements of most small businesses.
Installation is straight forward, and if you are familiar with installing any server based operating system you are not going to be in uncharted territory. When Windows Admins hear the words ‘Linux Server’ often their face goes pale at the thought of a complicated command line, this is no longer the case. The installer is a nice, visually appealing GUI that you click through. Boot from the installer media, choose your language, name your server, configure networking, and choose your disk layout. For a test server deployment, the layout does not matter much to me, but in production this can be a big deal. As a Windows admin, setting up my partitions is important to me. The ClearOS installer does not offer a clear default choice that I should choose for a new setup. For those or you that are used to working with Windows systems, you have implemented best practices where the operating system is on the C: drive, file shares are on the D: drive, etc… You have done this so that the OS can be on a redundant Mirrored volume, while storage is on cheaper RAID 5 volumes. You have done this to keep a clean file system. You have done this for faster recovery in a disaster recovery solution. Linux has a completely different partition scheme. There are three partitions created by default, the / root directory (think C: in Windows), the /Swap directory, and the /home directory. The Installer does give you the option to modify these upfront, but if you are not anticipating it or if you have to add storage later this can be complicated to change. The /home directory will be used when you create a user. This is where the users data will reside.
After choosing my disk layout, the setup continues and installs all the packages and binaries and I am presented with a “Setup is Complete” message. The server is rebooted and you are presented with the initial set up screen:
Verify that your network settings are proper, then connect to your server using a web browser. The rest of your server configuration will be done through the web browser. Connecting to the server with a browser prompts you to complete a setup wizard. You are afforded a vary large and compelling set of options that you can configure and install. The first phase of the setup wizard is the network mode: Private server, secured by and External Firewall, Public Server for when the server is installed in a hostile environment, and Gateway Mode which includes a firewall application that requires 2 NICs, and has monitoring and policies reminiscent of ISA server.
After configuring the network setup portion, you are prompted to install security updates and register your system. If you have a ClearOS account and serial number, this is where you enter it. If you need to create an account, the installer lets you do it there. Creating an account from the installer automatically grants you a 30 day trial.
The next setup phase is titled Configuration. The wizard prompts you for an Internet Domain Name, a hostname, and Internet Host Name. The steps are straight forward to configure the settings, but at this point you need to have some familiarity with DNS domain names but there is nothing here that is confusing for an experienced Systems Administrator. After configuring DNS and your systems’ Hostname you are prompted to select your account synchronization method: Standalone, Master or Slave. The installer advises you to “choose wisely”, you only have one chance, so get it right the first time. For this deployment I chose Standalone. Standalone configures the server so that it will be the “ONLY” server in the network, Master indicates that it is the “PDC” and slave configures it to pull directory replication from another ClearOS Server.
The server then launched into the Marketplace. The Marketplace is where you choose to install the productivity applications that you want to make your server do things. At this point you could skip to the end and not install anything else, but you would have a very bare bones server. The MarketPlace is the killer feature for me so far in ClearOS. A complete and organized list of the Applications that you can install on the server, a detailed description of what the application does, its cost and its EULA. The MarketPlace makes it very easy to know what you need to choose, even if going into the server install you don’t know what you need. I choose to install the Directory Server, Password Policies, Print Server, Windows Networking, and the messaging solution Zarafa. This provides what I believe to be the primary functions of a complete server solution for a small business; access Control file and print sharing, and a messaging solution.
For many Windows admins out there, the concept of having to add a Directory Server AND Windows networking is an unfamiliar concept. The Directory Server is just that, an OpenLDAP deployment. This is where your directory, users and groups are stored. A directory server provides information on users, groups and systems. The Windows Networking portion provides authentication services, file and print services, and Windows Domain functionality.
The final phase of the setup has you configure the DHCP sever for your network. You can also make changes to the running DNS server, and allow SSH access to your server. After these final pieces are configured, the setup wizard completes and you have an up and running ClearOS domain controller. You are returned to the Dashboard where you can get an overview of your system.
Overall the process of the installation is painless. There was never any point where I felt like I didn’t understand what I was choosing, or why I was choosing it. I am not a fan of the disk partitioning wizard at the beginning, and there will be further discussion on that in the domain and file sharing section. I did encounter a hiccup where the Windows networking would not start. I had to Google for an answer and discovered that I needed to SSH into the server and delete a lock file, and re-initialize the Windows Networking. There seems to be a race condition between OpenLdap and SAMBA that can result in initialization issues. This is not an unknown issue in the Windows world either, so I cannot bring major fault to ClearOS for this. ClearOs has developed a smooth, process based, wizard driven installation that really shines when it comes to setting up the product for the first time.