Closed marrobi closed 1 year ago
Solution is to add the following to the firewall routetable.tf
resource "azurerm_route" "aml" {
name = "AzureMachineLearning"
resource_group_name = data.azurerm_resource_group.rg.name
route_table_name = azurerm_route_table.rt.name
address_prefix = "AzureMachineLearning"
next_hop_type = "Internet"
}
resource "azurerm_route" "batch" {
name = "BatchNodeManagement"
resource_group_name = data.azurerm_resource_group.rg.name
route_table_name = azurerm_route_table.rt.name
address_prefix = "BatchNodeManagement"
next_hop_type = "Internet"
}
The bundle version will need updating, then updating in cosmos, then an update operation carried out on the firewall resource
Fixed.
When using publicly exposed compute instance with Azure Machine learning the following route table entries are required:
Without these deployment of the compute fails