VIRL on bare metal – lessons learned

So this weekend, I made the jump to getting VIRL working on a bare metal server that I was able to get my hands on. The server in question is a Dell Poweredge R710. Nothing too new, but it has a dual Xeon and 32GB RAM, so I figured I’d see what it could do. Mind you, I have never installed anything onto a server before in my life, but I like a little bit of a challenge and to learn some new things, so I figured I’d give a little rundown of some things I encountered.

First things first, I downloaded and burned a DVD with the ISO file to a disc and booted it up on my server. The install process is very easy to follow along at http://virl-dev-innovate.cisco.com/iso.bm.php, so I highly suggest using that if you, like me, have never done anything like this before. After the installation was completed, I restarted and booted up to the VIRL server. I should, however, note that before this step I wasn’t even seeing my HDD’s, as the SAS controller was disabled. This was easily fixed and installation itself was easy.

Once logged in, I started to sense that some things weren’t right. First off, I couldn’t even ping out on my server. I didn’t see any IP information set up on eth0, then I realized that the server I was using used the em nomenclature for ports, rather than eth. I quickly modified the ‘/etc/network/interfaces’ file and changed the naming to match what my server used. Restart, and success, I can ping out. (This was not success, as I learned later).

I started to go through the checks in the installation process and everything looked good, except I noticed linux-bridge-agent under the neutron agent-list wasn’t happy or alive. I ignored this (big mistake) and moved on, figuring that it’d work later. I was able to get through the activation process of the SALT server just fine, but I had to modify my ntp daemon to point to em1 again instead of eth0 before it could talk to the salt server and authenticate.

I updated my VIRL server to run the latest VMMaestro and other components, loaded up VMMaestro, and started up a design from INE.  Succe….fail. None of the routers would load up. They kept indicating “state changed from BUILD to ERROR with message: No valid host was found”. I rebooted the server a few times, no luck. I then started searching Google for this error. Finally, I stumbled on this link that laid out my solution: http://community.dev-innovate.com/t/linuxbridge-agent-down-vms-fail-with-no-valid-host-was-found/3685. I had to go through the virl.ini file and edit all of the interface pointers to eth and modify them to my controller addresses. After going through these steps and a couple reboots, I’m happy to say that my VIRL server is working just fine.

This was definitely a learning experience, and one I’m mainly documenting for my own records, so when I attempt this again I hopefully have something to reference when I inevitably run into these same issues. I’ve been enjoying running the server at home, labs run with no problem and memory isn’t an issue yet. I have a feeling when I start loading up some larger IOS-XR topologies in the future that I will probably need to kick up the memory, but so far, everything is working well.

~Dustin

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s