A lot of work goes into the guest virtual machines within my home lab, so it makes sense to protect that time investment through the use of backups. Much of my lab does actual “production” work, such as domain services, View desktop brokers, database, and other necessary infrastructure workloads to provide the foundation of the lab. A while back, there was a comment made on a post asking about how I keep my environment backed up. The short answer is that I’ve been using Veeam Backup and Replication (currently version 6) for the past several years.
I feel that any environment warrants consideration of backups, with a home lab being no different – just think of how much time you’ve put into creating the environment! Or, perhaps you want to have different configuration sets to restore from (beyond just snapshots). This post will go into details on how I maintain a set of backups for my lab, information on how I also ensure that the backups can be restored, and some thoughts on Veeam from a long time customer and NFR (not for resale) license holder.
I should also note that Veeam has recently become a paid sponsor of this blog (am I’m very happy about that!), however the content of this post is entirely of my own creation and expression of their product from the standpoint of someone who has and continues to use it.
Backup That VM Up
My backup configuration is relatively simple. I leverage the included VMware vSphere APIs for Data Protection (VADP) via a third party product – Veeam. I hand pick a number of my virtual machines to be backed up – templates, infrastructure, and View resources.
A list of rollback and backup files created by Veeam
These are all stored as files that Veeam creates on my file server VM, which uses a unique set of disks that are separate (physically and logically) from my other VMs. I then mirror these files to a server that is external to the lab.
The Veeam progress monitor of my Home backup job
Is this a true 3-2-1 style backup strategy? No, but I make the excuse that a home lab has limited budget. 🙂
A fully fledged backup strategy needs:
- (3) Three copies of the files should be maintained, including the primary data
- The backups should be on (2) two types of media (such as disk and tape)
- With (1) one of the copies being off site
While I do have three copies of the data – the original VMs, my backup, and a mirrored copy of the backup – I’m not currently putting them on another media or storing them off site. Because my site is my house, I’m assuming that if I lose my house to a disaster, the lab will be the least of my worries. When doing your backup strategy, weigh the opportunities vs the risks to help determine how far you’ll go to achieve 3-2-1.
The one exception to my backup strategy is SQL server. I use hourly SQL transaction log backups to disk, along with SQL log shipping to a secondary server. If you’d like to learn more about how to do this, check out my post on it here (video).
It’s Not A Backup Until You Can Restore It
Backups are rather worthless bits of binary until they can be restored. Fortunately, Veeam has a pretty unique feature called SureBackup that tests my backups to confirm that they are indeed valid backups. Using a concept called a “virtual lab”, I can schedule certain VMs to automatically restore to a sandbox environment and have verification scripts passed to them to ensure they validate, such as my domain controllers or View desktop brokers. The sandbox is controlled by a special Proxy VM that passes along traffic that provides a masquerade service to avoid duplicate network entries.
One of my SureBackup jobs tests a Domain Controller backup
The nice part here is that it is automated and the results of the tests are emailed out for me to review. In a lab environment this saves me a decent bit of time – for a larger, production environment this was a really handy feature to have.
I started using Veeam in the later half of 2009 to perform backups of my production environment at work. At the time, Veeam was one of the few players in the game who would really leverage vStorage APIs to perform backups. Other products required doing “business as usual” guest level backups via agents, or required tapping into the SAN fabric directly to do crash-consistent backups. I had a really good experience with the product and it served me well.
Since then, the product suite has continued to mature and grow with additional features and functionality such as single file restores, running a VM directly from a backup, and more options for application awareness. Additionally, Veeam has done a great job at connecting with the community at large through social media, informative webcasts, podcasts, vExperts, and releasing a quantity of product licensing for free (initially for vExperts and VCPs). I think this method of doing business is worthy of a nod.
As always, I’m interested in hearing your thoughts on the Veeam products, backing up your home lab, and any experiences you’d care to share. 🙂