WilliamLam.com

Does ESXi Support DDNS (Dynamic DNS)?

by // 5 Comments

An interesting feature request that was raised internally was for ESXi to support DDNS (Dynamic DNS) which allows a host client to update it's DNS record when using a DHCP Server. In most environments, to assign a hostname from DHCP, a DHCP reservation is used and this is maintained by the DHCP Server versus DDNS, where it is maintained by the client. Thanks to my colleague Eric Wager who did some quick research and found that ESXi does in fact supports DDNS and has been since ESXi 5.0.

I have not worked with DDNS much in the past and I have only seen it used for free/paid online services targeted at consumers to provide a well known address when their public IP Address changes frequently as with most ISPs. If your DHCP Server supports DDNS, this can be a handy feature to have, especially as you add new hosts you no longer have to manually create individual DNS record before hand and great for a lab environment. I did a big more digging to have a better understanding of how DDNS works with ESXi.

To enable support for DDNS on your ESXi host, you just need to set the hostname for the following ESXi Advanced Setting:

/Misc/PreferredHostName

You can do this in a variety of ways using either the vSphere Web/C# Client or using the command-line with ESXCLI.

Here is the syntax for the command:

esxcli system settings advanced set -o /Misc/PreferredHostName -s vesxi04.primp-industries.com

Once you have configured the setting, for the changes to go into effect, you will need to restart the management network. The easiest way to do this is via DCUI which you can run remotely by just typing dcui if you have an SSH session to your ESXi host. If you are using scripted install such as Kickstart, this can easily be automated as part of the post-install and upon first boot, DDNS will be enabled and configured with the proper hostname.

To test this in my lab environment, before enabling DDNS, I performed a reverse lookup of the assigned IP Address of my ESXi host from my DHCP server. In this example, the host received the address 192.168.1.135.

As you can see from the screenshot, a hostname could not be resolved as I would expect. After our changes, if we perform the reverse lookup again, we should now see the hostname that we had configured.

Another useful tidbit is the DHCP Client on ESXi is an ISC BIND implementation and this means if you require advanced things such as authentication keys, you can configured these options in /etc/dhclient-vmkX.conf where X is the specific VMkernel interface. For most deployments, you should not have to edit this file. Also if you want to prevent your DHCP Server from overriding the hostname of your ESXi host, you can add the following entry to the dhclient-vmkX.conf configuration file:

interface vmk0 {
   supersede host-name "vesxi04.primp-industries.com";
}

Just when I thought I knew about all the awesome features ESXi offers, it is a nice surprise to learn about another one!

Automating VCSA Network Configurations For Greenfield Deployments

by // 6 Comments

If you deploy the VCSA (vCenter Server Appliance) or other virtual appliances directly onto an ESXi host,
you will notice the network configuration wizard for the virtual appliance is not available as you would expect when deploying to a vCenter Server.

The reason for this is that ESXi does not support some of the advanced OVF/OVA properties such as the Networking section and you will need to deploy the OVF/OVA to a vCenter Server to be able to configure these advanced options. This poses a problem if you need to deploy the VCSA in a greenfield environment where you will not have an existing vCenter Server running and you will be deploying directly to the ESXi host. Unless you have a DHCP enabled network, you will most likely need to manually go into the vSphere C# Client to change the network configuration as it was unable to obtain an IP Address.

Though this is a one time configuration, it is still not ideal and would require the use of a Windows system to access the vSphere C# Client. You can actually get around this by leveraging the GuestOperations API (previously known as VIX API) which allows you to perform operations within the guestOS that is running VMware Tools. The other nice thing about the GuestOperations API is that it does not require any network connectivity from the virtual machine.

Note: The GuestOperations API can be accessed in variety of ways and in this article I am demonstrating just two methods and does not require a Windows system. You can also access the GuestOperations API using PowerCLI if you are more comfortable with Windows and do not wish to use the vSphere C# Client to manually configure the network settings for the VCSA. I would also like to stress that though this article is about the VCSA, you can easily apply this to any VMware based virtual appliance or virtual appliance running VMware Tools.

The most important thing to identity before using the GuestOperations API is the specific command or program you wish to invoke and the argument it accepts. To configure the network configuration for the VCSA or any other VMware based virtual appliance, you would use /opt/vmware/share/vami/vami_set_network If you just run this command by itself, there are variety of options from IPv4 to IPv6, static or dhcp configuration. In our example, we will be configuring a Static IPv4 address for our VCSA and the command we would run is the following:

/opt/vmware/share/vami/vami_set_network "eth0 STATICV4 192.168.1.150 255.255.255.0 192.168.1.1"

Method 1 - Using RVC (Ruby vSphere Console)

 

RVC is a nice open-source tool for interactively managing and configuring your vSphere infrastructure. RVC can be installed on any platform, in this example, I am running RVC on my Apple OS X laptop.

Step 1 - We first need to deploy the VCSA OVA and we can do so by using the ovftool via the command-line which can also be installed on Mac OS X system.

Step 2 - We then login to our ESXi host using RVC.

Step 3 - Next we will need to "change directory" to the location of our VM, in this example my VCSA is called VCSA-5.1. We can then run the "info ." command to view the summary of our VM. We can see that our VM is powered on from our initial deployment and we are ready to apply our network configurations in the next step.

Step 4 - To be able to run the above command, we will need to first authenticate into the guestOS. To do so, we will run the "vm_guest.authenticate ." and we will be prompted for the VCSA password. By default, the command assumes the username is root but that can also be specified on the command-line. If you are successful, you should not see any errors and then we can run the "vm_guest.start_program" command. Run the following to set a IPv4 static IP Address:

vm_guest.start_program . --program-path /opt/vmware/share/vami/vami_set_network --arguments "eth0 STATICV4 192.168.1.150 255.255.255.0 192.168.1.1"

Note: All commands in RVC can be tabbed out with auto-completion.

If the command was successful, you can quit RVC and you should be able to ping the IP Address that you have just configured.

Method 2 - Using vSphere SDK for Perl Script

 

Awhile back I wrote a script called guestOperations.pl which is a vSphere SDK for Perl script that implements the new GuestOperations API. This is a generic script which can be used to remotely connect to either a vCenter Server or ESXi host and perform operations within a guestOS as long as VMware Tools is installed and running. In this example, I also have the vSphere SDK for Perl installed on my Mac OS X laptop, but you can also install this SDK on any platform as well.

Step 1 - We will first use the "validate" operation to ensure our credentials to the guestOS is correct, but more importantly ensure that VMware Tools is up and running.

If the operation was successful, we should see our guest credentials validated. If not, you may need to wait a minute or two while VMware Tools is still loading up.

Step 2 - To invoke the command to configure the network configuration, we will use the "startprog" operation and run the following:

./guestOpsManagement.pl --server mini --username root --guestusername root --vm VCSA-5.1 --operation startprog --program_path /opt/vmware/share/vami/vami_set_network --program_args "eth0 STATICV4 192.168.1.150 255.255.255.0 192.168.1.1" --working_dir /

If the command was successful, then you should now be able to ping the IP Address that you have just configured.

As you can see, with the use of the GuestOperations API, you can do more than just setup the network configuration for a VM, you can run pretty much any command within the guestOS as you normally would if you were to RDP or SSH in. This is a very powerful interface that you can leverage to help you automate your virtual machine deployment and configurations!

Detecting A Duplicate IP Address For Your ESXi Hosts Using a vCenter Alarm

by // 6 Comments

The motivation for this article was a tweet I noticed from Duncan Epping this morning. Per Duncan's tweet, it looks like he may have accidentally assigned an IP Address to one of his virtual machines which was already being used by an existing ESXi host causing a duplicate IP Address error. We probably have all experienced this once in our lives and it can be quite difficult and frustrating to troubleshoot. Similar to a Windows OS, ESXi can also detect a duplicate IP Addresses but instead of a notification window, it is just logged in the VMkernel logs which looks like the following:

2013-01-21T15:52:35.989Z cpu1:2049)Tcpip_Vmk: 112: arp: 00:50:56:bd:3b:2b is using my IP address 172.30.0.213 on vmk1! 

The biggest challenge of course is to identify which ESXi host actually has a conflict and then taking a look at the logs to find the offending MAC Address and shutting them down yourself or with the help of a Network Administrator. Wouldn't it be great if we had an alarm to automatically notify us when a duplicate IP Address is detected? Well I am glad you asked and the answer is YES! 🙂

In addition to logging to the VMkernel logs, ESXi also logs this "observation" in /var/log/vobd.log which stands for the VMkernel Observation. These "observations" can provide critical identifying information in case of an error and is usually used during troubleshooting. In our case, we are seeing an intermittent network connectivity to our ESXi host which is in result of a duplicate IP Address. The really neat thing about these VOBs is that you can create vCenter Alarms when a specific VOB has been detected. I have shown an example of this before in my Detecting ESXi Remote Syslog Connection Error Using a vCenter Alarm article.

We can do exactly the same for detecting a duplicate IP Address for an ESXi host. The first thing we need to do is identify the VOB ID by looking in /var/log/vobd.log:

2013-01-21T15:02:07.513Z: [netCorrelator] 917174784727us: [esx.problem.net.vmknic.ip.duplicate] Duplicate IP address detected for 172.30.0.83 on interface vmk0, current owner being 00:50:56:bd:3b:2b

We can see the VOB ID for this is esx.problem.net.vmknic.ip.duplicate and this will be used in our vCenter Alarm trigger.

Step 1 - Create a new Alarm and specify a name, the Monitor type will be Hosts and Monitor For will be for a specific event:

Step 2 - Copy the VOB ID that we have identified from above and specify that as our alarm Trigger:

Step 3 - If you wish to receive an email notification or send an SNMP trap go ahead and configure additional actions, else just click next which will just display a vCenter Server alert in the UI.

Now that our alarm has been created, we will want to give this a test drive .... who can we ask? Well it just happens that I have a new user in my environment and I provisioned him a new VM which is already connected to the network. Let's hope he does not try to change the IP Address (because this never happens, right?)

After the user statically assigns the IP Address of an existing ESXi host in the VM, we should see our new alarm trigger in vCenter.

As you can see, we have quickly identified the ESXi host that is impacted and we can then login to DCUI via the console to take a look at the logs to find the offending MAC Address. Hopefully duplicate IP Addresses is not a common problem in your environment but it does happen from time to time and having an alarm to help you quickly narrow down the culprit can be quite useful.