Im trying to migrate some server farms out of our IOS SLB load balancer to our new CSS11506.
Need some help converting this IOS SLB config into a working CSS 11506 config. So far its not working. With the group command in the CSS config using the add destination the application works but, they see the CSS as the source address because its NATing the traffic to/from the client. This breaks the application, because it needs to build some session and cache information based on the source IP of the client which is being hidden by the NATing. Without the group command add destination service there seems to be a problem with the redirection/communication handoff from the CSS to the server and client. Does anyone see any glaring problems with my conversion of the IOS SLB config to the CSS?
Yes the clients are hitting port 8181 and being translated to port 7011 on the servers. This is also a (redundant pair in active standby) one armed CSS configuration with a single Gigabit FX connection to a pair of 6513's.
I'm not familiar with IOS SLB, but with the CSS, keep in mind that the CSS needs to see the session traffic in both directions to work properly, this generally meands that the CSS is directly in-line from a bridging or routing perspective. If there are any paths that the traffic can take that would bypass the CSS, you're going to have problems.
Considering that the traffic gets through when your services are configured as group destinations, I suspect that there is a non-CSS path that the servers are trying to use when responding to the client with the group disabled.
I've just read your reply to the other message. NAT is a requirement for a true one-armed configuration. One way to make this work without NAT and without changing the physical setup would be to trunk two vlans across the gigabit connection, one VLAN for the front-end virtuals, and a second back-end vlan for your servers. Configure the servers to point to the back end CSS address as their gateway, and the CSS will route between the VLANs for you.
I use this configuration in my environment with a pair of 6509's and it works pretty well.
Thanks for the insight, this is the same conclusion i have come to as well after pouring through sniffer traces. I've added circuit interfaces for both the VIP and the servers. Now i've got to convince our server/app guys to change the default gateway.
I read through the conversation and am missing one detail regarding the communcation between the clients and theservers. is the original IP-Address of the clients necessary to be seen by the server? If not just do a source-nat on the CSS so that the server guys do not have to change der default GW
The unmanaged mode is also known as Network only switching, which is introduced in Brazos release. It adds the flexibility for customer to use only network automation for service appliance.
If a device is configured a...
Usually, we can access ESXi Shell by pressing Alt+F1 from ESXi DCUI (Direct Console User Interface).
But on HyperFlex system, it just shows black window.
This is expected behavior because HyperFlex redirects ESXi Shell output to SoL...
Configuring an Export Policy Using the GUI
This procedure explains how to configure an Export policy using the APIC GUI. Follow these steps to trigger a backup of your data:
On the menu bar, choose Admi...