Best Practices for Agentless VMware Backup

Locate virtualization proxy servers on the network close to the VMs being protected.
Do not use a vCenter node as a virtualization proxy server.
For optimum performance, multiple proxy servers are recommended to enable load balancing. However, the number of proxy servers is limited by the number of allowed connections to vSphere and the ESX server.
DPX automatically chooses proxy servers for a backup operation from available proxy servers and load balances among multiple available proxy servers.
Consider the storage requirements of your storage system based on the number of VMs to protect.
For best performance and reduced load on ESXi servers, back up over a SAN rather than a LAN. Note that the data transfer method for agentless backups is based on availability. If all the datastores of the VM are accessible to the proxy over the SAN, then SAN transfer method is used. If not, Catalogic DPX checks for hotadd and, if available, it will be used. If neither SAN nor hotadd are available, then LAN (NBD) transfer method will be used.

Tip. Try to ensure that the proxy server has SAN or hotadd access to the target datastore, since NBD backup is much slower.

Define backup jobs at the container level, such as resource pool. Supported VMs added to the container are automatically added to the backup.
In a single Agentless backup job, do not combine VMs from more than one vCenter. The VMs from the different vCenters should be backed up using separate backup jobs.
Snapshot verification is available for DPX open storage. By running snapshot verification, you can quickly detect snapshot corruption. For more information, read the knowledge base article .
When VMs are no longer needed, run the Delete VM/VMDK Wizard to remove the VM and associated mappings, LUNs, temporary datastores, and associated cloned volumes. See .
Use Agentless VMware Backup to protect most VMs. Use the Catalogic DPX Block Data Protection (guest-based protection) to provide application-aware protection for VMs running applications such as Microsoft SharePoint, Microsoft Exchange, Microsoft SQL Server, or Oracle Databases, see .
If you install the Catalogic DPX Client on a VM to run Block backups, you can set the option to skip that VM during Agentless VMware Backup. See .
For best performance consider backing up over SAN or HotAdd rather than LAN (NBD).
In a single Agentless backup job, do not combine VMs from more than one vCenter. The VMs from the different VMware vCenters should be backed up using separate backup jobs. That is N servers to 1 job and 1 volume.
Define Agentless VMware backup jobs at the container level, such as VM folders or resource pools. Supported VMs added to the container are automatically added to the backup.
Avoid storage migration while Agentless VMware Backup backups are running.

Last updated 1 year ago

Locate virtualization proxy servers on the network close to the VMs being protected.
Do not use a vCenter node as a virtualization proxy server.
For optimum performance, multiple proxy servers are recommended to enable load balancing. However, the number of proxy servers is limited by the number of allowed connections to vSphere and the ESX server.
DPX automatically chooses proxy servers for a backup operation from available proxy servers and load balances among multiple available proxy servers.
Consider the storage requirements of your storage system based on the number of VMs to protect.
For best performance and reduced load on ESXi servers, back up over a SAN rather than a LAN. Note that the data transfer method for agentless backups is based on availability. If all the datastores of the VM are accessible to the proxy over the SAN, then SAN transfer method is used. If not, Catalogic DPX checks for hotadd and, if available, it will be used. If neither SAN nor hotadd are available, then LAN (NBD) transfer method will be used.

Tip. Try to ensure that the proxy server has SAN or hotadd access to the target datastore, since NBD backup is much slower.

Define backup jobs at the container level, such as resource pool. Supported VMs added to the container are automatically added to the backup.
In a single Agentless backup job, do not combine VMs from more than one vCenter. The VMs from the different vCenters should be backed up using separate backup jobs.
Snapshot verification is available for DPX open storage. By running snapshot verification, you can quickly detect snapshot corruption. For more information, read the knowledge base article .
When VMs are no longer needed, run the Delete VM/VMDK Wizard to remove the VM and associated mappings, LUNs, temporary datastores, and associated cloned volumes. See .
Use Agentless VMware Backup to protect most VMs. Use the Catalogic DPX Block Data Protection (guest-based protection) to provide application-aware protection for VMs running applications such as Microsoft SharePoint, Microsoft Exchange, Microsoft SQL Server, or Oracle Databases, see .
If you install the Catalogic DPX Client on a VM to run Block backups, you can set the option to skip that VM during Agentless VMware Backup. See .
For best performance consider backing up over SAN or HotAdd rather than LAN (NBD).
In a single Agentless backup job, do not combine VMs from more than one vCenter. The VMs from the different VMware vCenters should be backed up using separate backup jobs. That is N servers to 1 job and 1 volume.
Define Agentless VMware backup jobs at the container level, such as VM folders or resource pools. Supported VMs added to the container are automatically added to the backup.
Avoid storage migration while Agentless VMware Backup backups are running.