WAN Accelerator Sizing
To ensure correct work of remote jobs via WAN accelerators, you must provide enough free space for service data on the source and target WAN accelerators.
Source WAN Accelerator
When you run a remote job via WAN accelerators, Veeam Backup & Replication analyses data blocks going to target and calculates digests for these data blocks. Digests data is stored on the source WAN accelerator, in the VeeamWAN folder on the disk with the most amount of free disk space.
It is strongly recommended that you provide enough disk space on the source WAN accelerator to store digest data. The amount of required space depends on the capacity of source VM disks. For 1 TB of VM disks capacity, you must provide 20 GB of disk space. For example, if you plan to process 10 VMs whose capacity is 2 TB, you must allocate 40 GB of disk space for digest data on the source WAN accelerator.
Note: |
Global cache is stored only on the target WAN accelerator. You do not have to provide space for global cache on the source WAN accelerator. |
Target WAN Accelerator
Global cache is stored on the target WAN accelerator, in the VeeamWAN folder on the disk with the most amount of free disk space. The size of global cache is specified in the properties of the target WAN accelerator.
You must provide enough free space for global cache data. It is recommended that you provide 10 GB per every type of OS on VMs that you plan to process. By default, Veeam Backup & Replication allocates 100 GB for the global cache size.
For example, you want to process the following VMs:
- 1 VM that runs Microsoft Windows 7
- 3 VMs that run Microsoft Windows 2008 R2
- 2 VMs that run Microsoft Windows 2012 R2
There are 3 types of OSes so you must allocate 30 GB for the global cache on the target WAN accelerator.
In some cases, Veeam Backup & Replication may require more space on the target WAN accelerator than specified in the WAN accelerator properties. This can happen if digest data on the source WAN accelerator are missing or cannot be used. For example:
- You have performed the Clear Cache operation on the source WAN accelerator and it no longer contains digest data. For more information, see Clearing Global Cache.
- Veeam Backup & Replication has attempted to resume operation of backup data transfer but the backup file was not prepared for the operation in a proper way. The digest data must be re-calculated.
In such situations, the target WAN accelerator will have to calculate digest data on its own, which will require additional space.
For safety reasons, it is recommended that you provide additional 20 GB of free per 1 TB of source VM data on the target WAN accelerator. This amount of space is required for digest data recalculation. If you do not provide this amount of space and a situation when Veeam Backup & Replication needs to recalculate digest data occurs, the remote job will work in the limited mode. Veeam Backup & Replication will not deduplicate data against the previous restore points copied to target. For more information, see Global Data Deduplication.
For example:
- You have allocated 100 GB for global cache on the target WAN accelerator.
- The size of VMs to be processed is 2 TB.
In this case, the total size of space that you must allocate for global cache on the target WAN accelerator is:
100 GB + 40 GB = 140 GB
Many-to-One WAN Acceleration Scenario
Global cache size is calculated per 1 source WAN accelerator working with the target WAN accelerator. If you plan to use several source WAN accelerators with 1 target WAN accelerator, you must increase the size of the global cache proportionally. The global cache size is calculated by the following formula:
Total Target WAN Accelerator Cache Size = (# of Source WAN Accelerators) * (Size Configured in the Target WAN Accelerator Properties) + 20 GB per 1 TB of Source Data
For example:
- You have 4 source WAN accelerators in the source side working with 1 target WAN accelerator in the DR site.
- The global cache size configured in properties of the target WAN accelerator is 100 GB.
- The size of VMs to be processed is 2 TB.
In this case, the total size of space that you must allocate for the global cache on the target WAN accelerator is:
4*100 GB + 40 GB = 440 GB