Process ends with "killed" when using brain server
I have just started trying to use a dedicated hashcat brain server when doing slow / difficult hashes to cut down on the rework. Brain is running on a dedicated Ubuntu 14.04 (fully compatabe version for the hp dl380 g7) locally connected via gigabit switch. Server has dual zeons and 48Gb ram with 2.2 tb hdd (overkill on the proc and hhd but it was literally sitting in my attic collecting dust). Hashcat 6.1.1 installed and runs fine in brain mode.

I tried to "pre seed" the database with --brain-session option from the client to add 2 rather long sessions I have already done (4-6 hours each) and I see the connections being made and disconnected on the brain server. When trying to verify the seeding took place I ran the same command without the --brain-session argument; the crack starts on the client but ends with "killed" in under 60 seconds and after disconnecting from the brain server. I get the same result even if I try a new word list any time I am using the brain server .. couldn't find any info on the error and not sure where to start debugging.

In addition I am only going to use a single client and the dedicated brain server, again on lan, can I just use the --brain-client-features-2 argument to bypass the fast hash bottle necks as well?
May have solved my own problem, but in case anyone else runs into this issue in the future, it appears to be the amount of system RAM available on the host (client) PC, I only have 8Gb there and it appears with my current card configuration I needed a little over 10GB, disabling one of the GPU's seems to allow it to run.

Still wonder about the last question though.