The Cisco UCS (Unified Computing System) is an extensive system in terms of size and cost. Because of this, it is difficult to get hands-on experience within the home. Thankfully, Cisco has released an emulator, UCSPE (UCS Platform Emulator), which runs happily on a laptop!
There are some limitations to UCSPE when compared to the whole system. We cannot install operating systems on the blades and rack mount computers, we cannot connect to other networking equipment, such as switches and SAN storage, and we cannot perform tasks such as setting up LDAP authentication.
These limitations will not stop us from having some fun though, as we can learn a lot about how to operate and maintain a UCS, without even having to leave the house!
So, let’s start by downloading and setting up UCSPE.
Setting Up UCSPE
UCSPE is a free download from Cisco.com. You will need a Cisco ID, so sign up if you do not have one already (https://id.cisco.com/signin/register). You can easily find UCSPE by searching for it on the main Cisco page (Figure 1-1).
Click on the link (as shown in Figure 1-1) and you will be taken to the main page for UCSPE where you can click the link for the OVA and Zip file downloads (Figure 1-2).
You will then be prompted to sign in using your Cisco ID. Once you have signed in, you can download the UCSPE software (Figure 1-3).
The OVA file is the easiest option to use. You will need to accept the Cisco license agreement to download it. At the time of writing, the current version is 4.2(2a).
Importing UCSPE into VMWare
To run UCSPE, you will need the following available hardware:
1 CPU
2048MB memory (2GB)
This is hardly resource-intensive, so it should run happily on most modern computers. Our platform will primarily use VMWare (Fusion), but UCSPE will also run fine on VirtualBox or other hypervisors. Installing the VMWare software is not covered in the book.
Firstly, start VMWare, and select the option to import (Figure 1-4).
In the window that pops up, click on “Choose File…” (Figure 1-5).
Select the UCSPE file you downloaded earlier and click on “Open.” Click Continue again (Figure 1-6).
In the next window, you can rename the VM and select where to store it (Figure 1-7). Click “Save.”
The virtual machine will start to import into VMWare (Figure 1-8).
Once the import has been completed, you can customize the settings or click Finish (Figure 1-9).
Starting UCSPE
Start the UCSPE virtual machine by right-clicking on it and selecting Start.
When you start UCSPE for the first time, it runs through a quick installation and then starts the services, which can take a few minutes. Subsequent starts do not trigger an installation. The VM will then do some basic tasks, like using DHCP to gain an IP address (well, three actually; one for each of the Fabric Interconnects and one for the virtual IP, or VIP) (Figure 1-10).
In the next stage, UCSPE will complete the backend tasks, such as setting up SSH and GUI access, as well as generating the hardware catalog, which we will look at in the next chapter (Figure 1-11).
Once everything has been completed, you will see the login details, which are the IP address of the VIP and the username and password (which are both “ucspe”) (Figure 1-12).
You can login and use the UCSPE VM console to show you more network details, the general status, and to perform functions such as resetting the system, rebooting, or shutting down (Figure 1-13).
You can now browse to the GUI using HTTPS (in this instance, it would be https://172.16.2.134) (Figure 1-14).
You can see that we have a list of our equipment on one side, and on the other we can log into UCS Manager (using the default username and password “ucspe”).
When UCSPE starts up, it will generate a fairly random environment, and that is where we will start in the next chapter after we look at how a UCS would be set up in the real world.
Real-World UCS Setup
If you are setting up a physical UCS, you will need to allocate three IP addresses; one for the management of each of the Fabric Interconnects (FICs), and one for the virtual IP (VIP) that will be used for the cluster management.
Once you have racked the FICs, we need to do the essential cabling for them, which will be management (LAN) interfaces that will connect to your upstream or management switches, and then we need to connect the cluster interfaces. These are the L1 and L2 interfaces, and we cable L1 on the first FIC to L1 on the other FIC and L2 to L2.
In this section, we will be configuring two FICs. FIC A will have the IP address 10.99.1.10, with a /24 subnet mask (255.255.255.0). FIC B will use the IP address 10.99.1.11. The VIP will be 10.99.1.200, and the default gateway will be 10.99.1.1. We will configure a DNS server IP address of 10.99.1.5. The cluster will be called “Mastering-UCS.”
Only power up the first FIC at first. We only power up the second one once the first FIC has been configured.
Connect your computer to the first FIC using a console cable, open your terminal software (like PuTTY or SecureCRT) and get on to the serial port using the following settings:
9600 baud
8 data bits
No parity
1 stop bit
After connecting, you will see the following prompt:
—- Please read the following carefully —-
At the request of the publisher, I have had to change
the output shown in the Cisco console, because
it is also shown on a website. Even though what you would
Have seen would have come straight from the Cisco device.
This was done to avoid any copyright issues.
So, I hope you understand that (for this chapter only)
Things are going to get super weird.
Everything goes back to normal in the next chapter though, so please bear with the incoming strangeness, but they wouldn't listen to me.
We are going to use the console for our configuration, so type “console” and press enter.
Choose a config technique. (console/gui) ? console
Type in “setup” next, as we are not restoring from a backup.
Here you choose a setup mode; New or rebuild from backup. (setup/restore) ? setup
This is a new Fabric Interconnect, so type “Yes” to start a fresh setup.
Thou hast elected to make a new Fabric interconnect. Resume? (y/n): y
Type in a new password for the admin user when prompted. The complexity needs to be at least eight characters.
Choose a p@ssw0rd for "admin": Admin123
Type it in again "admin": Admin123
We will set up a cluster, so type in “yes” when prompted.
Might this interconnect device be half of a cluster(select "ney" for lonesome-mode)? (yes/no) [n]: yes
Because this is the primary switch that we are setting up, it will be the “A” of the pair. The secondary switch will be “B.”
Pop in the switch fabric (A/B) []: A
Next, we get to name our cluster. The name we chose will also be used on each Fabric Interconnect, with them getting either -A or -, depending on if they are the first FI, or the second.
Enter the system name: Mastering-UCS
Set the IP address, subnet, and gateway for the fabric interconnect.
Now we type in the cluster IP address. It is this address that will be used by the primary Fabric Interconnect and is the IP you use when you fire up the UCS Manager.
Gathering IPv4 addy : 10.99.1.200
You can then configure your DNS servers. This is optional.
Wanna use a DNS Server? (yes/no) [n]: y
Pop in its numbers : 10.99.1.5
We are not going to configure a default domain name, so choose “no” for the next section.
Probably don't need a default domain name? (yes/no) [n]:n
You will see a summary of the settings we have entered so far. Either type “y” to save the settings and restart the fabric interconnect, or select “n” to go back and make any modifications that you need to.
The following configuration will be pertained:
Switcharoo Role=A
UCS Designation=Mastering-UCS
Corporeal Switch Management0 IP Addy=10.99.1.10
Corporeal Switch Management0 IP Netmask=255.255.255.0
Gateway=10.99.1.1
Nameserver=10.99.1.5
Gathering Enabled=yes
Gathering IP Addy=10.99.1.200
Save it and use this config (select "no" if you want to re-enter)? (yes/no): yes
Sorting this out for you. Standby.
Configuration file – Ok
Our first fabric interconnect is now completed. Connect to the console on the second device and start it up.
From this point, it’s just a matter of following the same steps.
—- Please read the following carefully —-
Still with us? Marvelous. Not much more to do now,
And we can get back to how things actually look when
you configure a UCS. Again, I'd like to stress that I did try
to keep everything as you'd see it on screen, but it was a
case of either change it, or not get it published.
I hope you are well. Been to any good music gigs recently? How's the family? I wish you a bright and happy future.
Let's finish this off. Cheers.
Which do you prefer? (console/gui) ? console
The installation will detect the presence of Fabric Interconnect 1 (assuming you have connected the heartbeat ports together), and this switch will be added to the existing cluster. Type “y” to do this.
The setup is like a Jedi and has felt the aura of another Fabric interconnect. Wanna add this switch to the cluster? Carry on (y/n) ? y
Pop in the admin p@ssw0rd of the other Switcharoo thing: Admin123
Attaching to the other switch… finito
Stealing info from other switch… finito
Other switch Management0 IP Addy: 10.99.1.10
Other switch Management0 subnet: 255.255.255.0
Gathering IP addy : 10.99.1.200
We now set the IP address for this fabric interconnect.
Wanna keep the config (type "negative" if you don't like it)? (yes/no): yes
Saving config. Give it a second or three.
Finally, we're done – Thanks
Navigate to https://10.99.1.200 using a browser to log into the UCS.
Summary
We started this chapter by downloading the latest version of UCSPE from the Cisco website, importing it into VMWare, and starting it up. We then compared this to how a real UCS installation works in a clustered environment.
In the next chapter, we will configure UCSPE according to a topology (instead of the random assortment of hardware that it generates) and, as we do this, we will look in greater depth at the various components of a UCS environment.