SysAdmin

VMworld Announcements

Matt Patterson

The big day for VMware admins around the world is here with the first day of VMworld 2013. Kicking off the conference (besides the pre-parties) is the keynote given by the CEO of VMware, Pat Gelsinger.

The big announcements from the keynote are:

  • VMware vSphere 5.5 As rumored, we didn’t see vSphere 6, but there were several enhancements here that caught my eye. A new VM version which now supports better vGPU and more device support (more disks per VM). vApp HA which checks for the presence of “heartbeats” from VMware Tools™ as well as I/O activity from the virtual machine. If neither of these is detected in the specified amount of time, vSphere HA resets the virtual machine. You will notice that many of the performance features have increased 2x except the vmdk which overshot and now supports 62TB-512b vmdks attached to a VM. I know several of my customers that will like that new feature. You can read more about what’s new in vSphere 5.5 here.

  • VMware vCloud Hybrid Services – General Availability Not really sure how to feel about this one yet. As a vCloud provider, they appear to be competing directly with me. And based on their 6 data center expansion planned for the next year (4 VMware DCs + 2 provided by Savvis), it makes it hard to think that they are going to throw us a bone anytime soon. At the same time, I do get it if you are an enterprise customer that has always trusted VMware for your infrastructure needs. Moving to the cloud is a bit scary to go with someone you don’t have that relationship with. VMware is your trusted partner here so why not go with them? What I would like to see is VMware say, hey Mr Enterprise, meet my partner here Mr vCloud who can fill this gap for you the same way that we can. Sure you lose the one throat to choke, but you keep partners happy while spreading the operational costs at VMware. Its a pipe dream at this point, but that’s the approach I would have rather seen.

    • VMware NSX – The Platform for Network Virtualization – We’re finally seeing the fruits of the Nicera purchase a year ago by VMware. Essentially, we have a new layer on top of existing physical networks that allows a true software defined network. What ESX did for compute resources, NSX hopes to do for network resources. VMware NSX brings together the best of Nicira NVPTM and VMware vCloud Network and SecurityTM into one unified platform, delivering the entire networking and security model (Layer 2 – Layer 7) in software. I don’t think that there is any avoiding the SDN. Where we used to have just VLANs separating customers, now we can have additional separation and isolation on top of existing networks as well as providing various services (Firewalls and SLB) with the click of a button. Should be exiting to see and I looking forward to turning a few wrenches on this in the lab.
    • VMware Virtual SAN (vSAN) Similar to VMware NSX, VMware Virtual SAN is built on a unique distributed architecture that will enable storage services to scale out linearly with the needs of the application. Through the seamless integration of VMware Virtual SAN with VMware vSphere, VMware has redefined the role of the hypervisor to deliver virtualized compute and storage services in an elastic, flexible fashion. The distributed architecture enables VMware Virtual SAN to deliver I/O performance comparable to mid-range storage arrays while leveraging the economics of direct-attached storage.

Puppet Error: Could not retrieve local facts: undefined method `kernel_fact_value’

Matt Patterson

Recently while patching some CentOS boxes in my lab, I noticed the following error.

May  5 13:48:38 puppet-agent[19924]: Could not run Puppet
configuration client: Could not retrieve local facts: undefined method
 `kernel_fact_value' for Facter::Util::Processor:Module

This was a bit odd as one of the things that I will typically do when I see puppet or facter updates in my update list is to run the agent manually with a

vCloud – Licensing error while accessing vSphere Profile-Driven Storage

Matt Patterson

Recently I was spooling up a dedicate cluster of hardware for a customer in our vCloud environment. It’s a smallish cluster of 5 hosts with a dedicated pool of storage. Nothing out of the ordinary here. Created the DRS/HA Host Cluster. The storage guys had already done the various zoning and disk LUN creation so that was added to the cluster and tagged it with a User Defined Storage Capability.

storage_profile_error2

Puppet Dashboard – Inventory Service

Matt Patterson

As a user of the free version of the Puppet Dashboard, I’m always looking for little tips and tricks that can make it more useful in my day to day job. Some of the features that you find in the Puppet Enterprise (PE) version are available in the free version, it just takes a bit of time and effort to get working for you.

One of those features was recently brought to my attention by a co-worker. In the Puppet Dashboard, its possible to have the facts of your node presented within the web UI. PE does this by default, but in the open source version you have to do a little wrench turning to get it working. My co-worker had already gotten it working on our production environment, but I wanted to see what it took to get running in my lab.

Puppet: Dashboard

Matt Patterson

Earlier in this series I had mentioned the Puppet Dashboard and how we were going to discuss that at a later time. Well guess what kids, its that time.

Server Installation

For those needing a little background on the puppet server in this example, we’re running on a CentOS 6.3 box with both the EPEL and Puppet repositories configured and active. All puppet modules of coarse are coming from the puppet repo itself for the latest and greatest bits.