Situation:

I deployed 4 RSA VM's across my multiple subnets, configured syncing, and started syncing my images. I noticed that a 7gb image took about 24 hours to sync and came to the realization that it's not, in fact, utilizing gigabit. I changed the network speed from auto-negotiated to 1000mbit full duplex and the device became totally unresponsive. I tried removing the RSA and creating another one with the same results. This is pretty hard to troubleshoot without real access to a command line, so I had to call KACE.

Problem identification:

I engaged the KACE support team and eventually determined that when you manually force the speed, it removes the IP address (verified via ifconfig). It would work properly after that point if you manually assigned an IP using ifconfig. The issue was taken to engineering and the determination was made that the current OVF downloaded with a k2000 running 3.4 has the vmware nic set to "Flexible" instead of the normal E1000.

The fix:

The fix was to remove the Flexible NIC and replace it with an e1000. After making the change (and the requisite reboot) the RSA booted in about 30% the previously required amount of time, and in Konfig reports the currently negotiated speed (i.e. Speed: Auto-negotiated (1000mbits full duplex)).

Our configuration:

HP C7000 blade enclosure, 8 of 16 blades installed, 10gig backplane. VMware ESXi 5, Veeam.

Feel free to post any questions and I'll answer as best I can.