Here at our little shop, I have been using Veeam for more than 5 years. It was already implemented when I got here, but I quickly decided to completely rebuild it from scratch and have been pretty happy with it. Of course, the build-out changes over time, as new features become available and/or requirements change.
I remember how excited I was when Veeam announced Agents a while back! Finally, I was able to use one product to back up virtual and physical workloads. I called my rep that day and told him to get me some licenses as soon as it goes GA.
Now, since we only are backing up about 200 VMs and a handful of physical machines, our environment is not overly complicated. But there is always a benefit in automation, so today I’m going to write about tags in an SMB environment when you don’t have different tier storage.
In our set up we are currently all using spindle drives in our SANs. This will probably change in 2020, but this is what we’ve got for now. I have on primary backup in HQ and replicate to DR to a different SAN. I’m following the 3-2-1 rule, having 3 copies of data. One is your actual data, the second on a backup locally and 3rd copy off-site. Having the backup locally allows me to very quickly restore files or do an Instant VM Recovery in Veeam to have that server back up in no time.
Before using Tags, I created a spreadsheet and created Server Groups (SG). I would put similar Servers in one Group (e.g. Domain Controllers) and have the application-specific settings set for this group. I would then manually add the server on the Veeam Server in HQ to a backup job and again in DR for a replication job.
Since I have converted to Tags, I still use my spreadsheet to track servers (we are using this for all kinds of other information too), but once I set up my tags in vSphere, I just pick the appropriate tags for the server I want to add. The next time the job runs Veeam will pick up the newly added machine and I don’t have to worry about a thing.
The reason I’m still utilizing Server Groups (SG) for our environment is I found it hard to get the jobs to an even number of machines and to apply settings across machines. Let me explain what I mean. I usually prefer my jobs to have about the same amount of machine to backup. I completely understand depending on the size and load of one particular VM, a job with 6 machines might take longer than one with 7. In terms of resource planning, especially to the DR site with bandwidth limitation, I need to time my jobs carefully. Here is an example of my SG:
HQ-BU-SG04 is the Tag I use at HQ. BU indicates it is a Backup job, not replication. SG04 stands for the server group. DR Tag is DR-REP-SG04. REP indicates a replication job. I have similar Tags for Agents, like HQ-SABU-Jarvis (HQ=Location, SABU=Server Agent Backup, Jarvis=Servername)
I’m sure there are a thousand ways how folks configured their Veeam Backup. This approach is working well for us, but I’m curious how others are doing it? Drop a comment and let me know. Always happy to learn and improve.