I am stumped with an issue and I really hope someone can step in and solve it. I have done google searches and have not been able to find a hit that provides the solution.
I have a setup of VMs that are effectively Linux-based VMs that talk to eachother over a network that make up a product that I support (I work for Cisco) called QvPC-DI (DI = distributed Instance), which is basically a Mobile Gateway used by Service providers, and which I am trying to get to run on this platform for internal testing and playing around. I have it up and working actually 100% of the time when only deploying a single VM (actually two VMs but the first VM is a controller/management VM that does not handle call traffic) where ALL THE PROCESSING OF THE CALL BEING MADE IS ON THAT VM. BUT when I try with 2 VMs, where traffic needs to go between processes located on both VMs through the life of the call setup (and teardown), then what I am seeing is that packets are getting sent from one VM but I am pretty sure they are not reaching the other VM, and this is 100% reproducible as well.
It has been mentioned to me about the idea of jumbo frames - well, I can't tell you for sure what the sizes are of these packets being sent back and forth, but they could easily be larger than 1500 bytes. My working theory is that VMware Workstation is dropping these packets due to their size, but I have no solid proof of that even with all the debug logging I have turned on in my product, and I have not been able to find the name of any parameter that can control the size of packets allowable.
Qs are:
Is VMware dropping the packets and if so how can I actually see that?
Is there a "name" for this internal network? (i.e. OpenVSwitch, ????)?
Is there an actual limit to the size of packets allowable on the network?
Is there a configurable that can increase the size of packets sent over the network?
Again, this is VMware Workstation Pro, this is NOT VSphere or any other product from VMware.
thanks a lot!
dd