Cloud Managed Networks

Hi, 

I have setup gateway cluster of 2 x A9106 gateways and Aruba 735 APs in classic Central. Everything seems to be working fine. The switches are old HPE procurve switches. The ports connecting the Aruba 735 APs have 'Total Rx Errors' increasing at a rate of approximately 1 per min.

Searching through history, this has been mentioned for controller-based AP install where the AP and controller were doing MTU discovery with MTU size of 1578 and DF bit set (Controller Based WLANs - Airheads Community) Though this MTU size is independent of the MTU size on the VLAN, it seems to cause the switch to register as an error on the interface. 

The solution seem to be manually setting the SAP MTU size of the AP instead of leaving it as default setting (Giant packets on switch port connected to Access point). This setting is done under ap system-profile. 

As both describe controller-based design where the AP tunnels to the controller for configuration, I am unsure if this applies to new design where AP get the configuration from Aruba Central but tunnels the SSIDs to the gateway. Appreciate any insight/advice. 

Thanks.

AOS8 and AOS10 APs use a method of MTU discovery - sending increasingly smaller frames with DF=1 until a frame traverses the tunnel to the controller successfully (and reply is seen of course) and this is set as the tunnel MTU ( CLI command show datapath tunnel, and show ap debug system-status) .

Switches that are not configured to handle jumbo frames (9216 bytes) will register the larger test frames as RX error or too large frames.

Manually setting the SAP MTU in the ap system-profile disables the discovery, setting the GRE tunnel MTU at the configured value.

AOS10 APs that tunnel to a Mobility gateway use a gre tunnel profile configuration for the mtu size.

Hi, 

I have setup gateway cluster of 2 x A9106 gateways and Aruba 735 APs in classic Central. Everything seems to be working fine. The switches are old HPE procurve switches. The ports connecting the Aruba 735 APs have 'Total Rx Errors' increasing at a rate of approximately 1 per min.

Searching through history, this has been mentioned for controller-based AP install where the AP and controller were doing MTU discovery with MTU size of 1578 and DF bit set (Controller Based WLANs - Airheads Community) Though this MTU size is independent of the MTU size on the VLAN, it seems to cause the switch to register as an error on the interface. 

The solution seem to be manually setting the SAP MTU size of the AP instead of leaving it as default setting (Giant packets on switch port connected to Access point). This setting is done under ap system-profile. 

As both describe controller-based design where the AP tunnels to the controller for configuration, I am unsure if this applies to new design where AP get the configuration from Aruba Central but tunnels the SSIDs to the gateway. Appreciate any insight/advice. 

Thanks.