-
Notifications
You must be signed in to change notification settings - Fork 29
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@er1cthe0ne Thanks for pushing the remaining fixes promptly. I left a few minor comments to the coding part.
Regarding the openflow table doc, let us review it together today.
ACA_LOG_ERROR("PortConfiguration.fixed_ips_size: %d.\n", | ||
current_PortConfiguration.fixed_ips_size()); | ||
throw std::invalid_argument("PortConfiguration.fixed_ips_size is less than zero"); | ||
} | ||
virtual_ip_address = current_PortConfiguration.fixed_ips(0).ip_address(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Although we allow more than fixed IPs, we process the first one only as the primary IP. As a next step, it would be awesome to start thinking about storing the secondary IPs so that when customers add IPs in the VM (manually as limited by current networking stack), Control Plane has the knowledge to grant or reject customers' traffic, depending on whether customers configure the VM in the right way.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
got it. The current port creation code doesn't consume the virtual IP address(s). It would be consumed in security group code to allow traffic come out from the VM with certain virtual IPs and virtual mac (it also should look at the allow_address_pairs in PortConfiguration).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let me rephrase the question: if an interface has more than 1 ports, the current ACA would only process and store the first ip, and other ips are ignored. We will need to handle other ips too.
|
||
ovs-ofctl add-flow br-tun "table=0, priority=1,in_port="patch-int" actions=resubmit(,2)" | ||
ovs-ofctl add-flow br-tun "table=2, priority=0 actions=resubmit(,22)" | ||
if [ "$1" == "delete-bridges" ]; then |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@cj-chung please take a look to confirm this approach is okay for our environment.
This is the change to enable L3 E2E, with this change, we confirmed the L3 routing E2E is working through the horizon UI.
This PR includes the follow changes: