For modern hardware and networking gear, MTU's above 9000 are common now. For instance, we have switches that support 9216, and intel 10gbit adapters that support 15500. There are a few reasons to want this:
- Being able to offer the default fabric networks to end users at 9000 (9216 would be plenty, as it should be 9000 + 50byte overhead)
- Being able to offer end users a flat VLAN for their use. Allowing them to utilize VXlans with a 9000 MTU
# dladm show-linkprop -p mtu ixgbe0 LINK PROPERTY PERM VALUE DEFAULT POSSIBLE ixgbe0 mtu rw 9000 1500 1500-15500
I am testing an update to the constants.js to provision compute nodes with an MTU of 9216. I will update with the results here.
I think bumping up the max to 9216 would suffice for our use case, but still provide the ability to keep users from setting it to something horribly detrimental.
For modern hardware and networking gear, MTU's above 9000 are common now. For instance, we have switches that support 9216, and intel 10gbit adapters that support 15500. There are a few reasons to want this:
# dladm show-linkprop -p mtu ixgbe0 LINK PROPERTY PERM VALUE DEFAULT POSSIBLE ixgbe0 mtu rw 9000 1500 1500-15500I am testing an update to the constants.js to provision compute nodes with an MTU of 9216. I will update with the results here.
I think bumping up the max to 9216 would suffice for our use case, but still provide the ability to keep users from setting it to something horribly detrimental.