Sangfor SCP Restore Failed with Sangfor Trial License

Discovered Version：v9.0.0.86562-SP1

Question

Backup jobs for Sangfor SCP in a 10GbE network environment are significantly slower than expected (actual speed ~80MB/s), failing to utilize the available 10GbE bandwidth.



Answer

The slow performance is caused by default transfer speed limits imposed by the Sangfor SCP platform for third-party agentless backups. These default settings are relatively low and need to be manually adjusted within the SCP's "CDP & Backup" configuration to allow for higher throughput and fully leverage the 10GbE network infrastructure.

Resolution

Option 1: SCP Transfer Speed Optimization 

• Access Settings: Log in to the Sangfor SCP management interface and navigate to System Settings → CDP & Backup Settings.

• Adjust Parameters: Update the following configurations to increase transfer efficiency:

  • Single Task/Node Transfer Speed: Set to "Unlimited" or the maximum available value.

  • Transfer Speed Limit: Increase to the platform maximum of 2400Mbps.

  • External Snapshot/Backup Host Limit: Set to "Unlimited" to remove artificial bottlenecks.

• Verification: Apply the changes, wait 1-2 minutes for the configuration to take effect, and re-run the backup task to monitor performance improvements.

Option 2: Comprehensive Network Performance Testing (Diagnostic Solution)

• Network Baseline: Use iperf3 between the Vinchin server and the SCP node to verify raw network throughput and ensure the physical link is not the bottleneck.

• Storage I/O Verification: Use dd commands to test the read/write performance of the underlying storage to rule out disk-level performance issues.

• Localization: Compare the results of the network, storage, and backup task tests to confirm that the bottleneck is indeed the SCP software-defined speed limit.