If you haven’t noticed, Veeam is a very busy company! One of the critical ways that I personally do what I do here (and it is a long list!) is through one of our lab environments. A number of different groups at Veeam have lab functions, and in this blog post, I’ll focus on the SSA lab.
The SSA lab is a joint collection of infrastructure for the US support team, the Strategy department and various Alliance activities. One visible example was at VMworld (See our previous blog post about the featured backup targets), where the SSA lab drove the preview demos of Veeam Backup & Replication version 6 and showcased a number of our different backup storage partners.
…balance the needs of all of the stakeholders
What we do in this lab is wide and varied. Subsequently, it is somewhat tough to balance the needs of all of the stakeholders. For example, the needs of the support staff are very different than the needs of myself and VMDoug. Further, we have to also represent an arrangement that may be at any mix of our customer install base. Specifically, our typical customer doesn’t use vCloud Director or Lab Manager, which are optimized for this type of configuration. And with our Hyper-V support coming with version 6, our usage requirements profile has changed. So, how do we go about designing a lab that services all of our internal stakeholders and fits our customer profile?
Well, the first design element zeroed in on the compute environment. In the planning stages of this build we needed to be able to provision both VMware and Hyper-V infrastructures. While we now can do nested Hyper-V and ESXi (special thanks to Ricky El-Qasem for his excellent guest blog post), at this point we’ll still need to do a lot of on-metal provisioning. So we went the scale-out direction with a (relatively) larger number of smaller hosts. Currently, we have 11 dual-socket servers providing the primary compute environment for the SSA lab, and another two dual-socket servers as a special stakeholder; in the Alliances and activities side. In this mix of servers, we have a core bucket of infrastructure, and a transient allocation. At any given time, the vSphere environments will be at around nine hosts and the Hyper-V environments will be around four hosts.
One of our next challenges was storage. As you would expect, we have no shortage of backup targets in the SSA lab. In terms of primary storage, however, we have a few things in place to address the needs of all of the stakeholders of this lab. In terms of primary storage, our highest tier of storage is a Hitachi Data Systems AMS 2300. This is a pretty sweet array. This unit provisions iSCSI storage to all of the SSA lab and fibre channel to select workloads as well. This device supports VAAI for VMware vSphere and is the most capable storage platform we have in the lab.
The AMS 2300 has 15,000 RPM SAS drives installed, and we provision LUNs on all of the storage resources with a pretty well defined nomenclature. That way, we know what storage protocol is in use, such as a SAN technology like iSCSI or fibre channel, or a NAS protocol like NFS. For example, in the SSA lab, the AMS 2300 has its datastores named as shown below:
In this example, the datastore is named:SAN-AMS2300-LUN1. SAN indicates it is a block storage protocol (in this case, iSCSI). NFS datastores start with “NAS” in our lab. The “AMS2300” indicates which controller is in use for the datastore, and finally “LUN1” indicates which logical unit that allocation corresponds to within the AMS 2300 controller. This configuration is shown below within the Hitachi Storage Navigator Modular 2 management software:
There are other stakeholders in the lab, and the other logical units in this example are provisioned to other environments; such as the Hyper-V cluster and other vSphere environments. Soon, we’ll also be implementing Hitachi NAS Platform (HNAS), this is the technology also known as BlueArc.
The disk nomenclature is one example of simplification of the environment due to the high amount of stakeholders. Anything we can to make the environment self-documenting is a good idea. All of the stakeholders in the lab have pretty well defined permissions and roles to fully test our products for functionality like creating VMs, mounting vPower datastores, deleting VMs, etc. We use the vSphere roles and permissions for Active Directory security groups extensively for this, our Hyper-V roles arrangement of equivalent capability isn't in place yet.
Like all infrastructure environments, we have our challenges
Like all infrastructure environments, we have our challenges: resources, space, cooling and power. Sound familiar? So we leverage a lot of features to help out where we can. vSphere DRS, of course, is the biggest aid, including Distributed Power Management (DPM). Further, we’re planning our migration to vSphere 5 and expanding our Hyper-V footprint with this same equipment, in line with our product support and customer needs. Other lab environments, of course, are already running vSphere 5, such as the product team and development teams.
We would be remiss if we didn’t thank all of our hardware partners to help round out the lab. I’ve mentioned Hitachi Data Systems above with the AMS 2300, but also, for our smaller customers we are leveraging a NETGEAR ReadyNAS 3100 unified storage product. This is a nice unit for the price and provides iSCSI and NAS disk resources. This functions as a different tier of primary storage for the SSA lab. We also have an Iomega StorCenter PX–4 300d running in the lab.
In terms of backup targets, we are using a number of deduplication devices, such as ExaGrid, Data Domain, Quantum and others. Again, being representative of our customer base, we have interests in the entire spectrum of storage technologies. We are also leveraging StarWind iSCSI SAN for a backup target as a virtual storage system.
We hope this peek inside Veeam helps you get a feel for how some of our groups function, while everyone’s needs are different. This is how some groups at Veeam get the job done. Do you have any mixed virtualization environment lab tricks that have been successful in your environment? Please share them below!
GD Star Rating loading...
Veeam on Veeam: The SSA Lab, 5.0 out of 5 based on 1 rating
Author: Rick Vanover
Rick Vanover (MVP, vExpert, Cisco Champion) is the director of Technical Product Marketing & Evangelism for Veeam Software based in Columbus, Ohio. Rick’s IT experience includes system administration and IT management; with virtualization being the central theme of his career recently. Follow Rick on Twitter... More