One of the key decision you have to make in planning a new Veeam deployment is in sizing the infrastructure. Correctly sizing the infrastructure to ensure that each of the Veeam infrastructure components servers has sufficient resources to perform its role will allow an optimally configured and trouble free environment. During VeeamON I attended a session on Veeam sizing by Tim Smith, I have pulled together my notes from the session and reviewed the Veeam documentation to give the following template for considering Veeam sizing decisions.
Veeam Components
The key components of the Veeam infrastructure you will have to consider in your sizing exercise are as follows:
- Veeam backup server
- Database server
- Proxy Server
- Repository server
- Enterprise Manager server
Information Required
In order to plan your Veeam backup sizing you will need to gather some information
- Size of source data – Required to calculate size of full backup
- Daily rate of change – This is required to understand the size of an incremental backup. Can either be calculated assuming change rate as 10% of your total data size or from existing backup software
- Number of VMs and disks to be protected and number of retention points required will also be needed for disk capacity planning which I will cover in a future post.
Sizing Veeam Considerations
Some considerations when sizing your Veeam environment:
- All sizing information is contained in the user guide and best practice guide
- Some sizing requirements refer to “Per Task”. Each disk that is backed up is a task this is not a per VM measure
- Sizing requirements are specific to each Veeam infrastructure component
- Sizing for each role is cumulative. For example if the server is both a Proxy and Tape server you will need to add the resource requirements for both together
- The sizing information given in the user guide are minimums
- The best practice guide estimates that doubling resources will half the backup window. Although this will of course be dependent on an any other bottlenecks
- Staggering jobs can help to reduce resource requirements since some components are sized on the basis of number of concurrent jobs
- Exact requirements will be specific to your environment, start with the recommended values access if it is meeting your requirements and amend as necessary. Then retest until you are satisfied with the result
Veeam Component Resource Requirements
Backup Server
- 1 CPU core for every 10 actively running jobs
- 4 GB RAM for every 10 actively running jobs
- Minimum is 2 CPU cores and 8 GB RAM.
- Disk space
- Installation – 40GB
- Logs – 3 GB log files generated per 100 protected instances, with a 24 hour RPO
Database Server
SQL Express is bundled with Veeam. Consider using the full blown SQL if:
- Protecting more than 500 VMs
- Using files to tape extensively
When using the full version of SQL:
- Up to 25 concurrent jobs – 2 CPUs, 4GB RAM
- Up to 50 concurrent jobs – 4 CPUs, 8GB RAM
- Up to 25 concurrent jobs – 8 CPUs, 16GB RAM
Backup and Replication Console
- 1 CPU
- 2 GB RAM
Backup Proxy Server
The proxy server is sized per task, a task is defined as a VM hard drive. In effect this setting will determine the number of disks and therefore VMs that can be backed up at once
- 1 CPU core per task
- 2GB per task
- 500MB of disk space per task
Backup Repository Server
This is again sized per task
- 1 CPU core per task
- 4GB per task
- Detailed sizing will
Veeam Backup Enterprise Manager
- Memory 4GB minimum
- 2GB of hard disk space
Final Thoughts
Remember these figures are general guidelines ensure you account for time in your project to test and optimise resource allocation.
I will cover sizing for backup repositories in a future post.