Load balancing SSH

See below the tcpdump: We are getting resets and it fails quickly. We are running our LTM in one-arm mode. A few years ago I used telnet going back to a main-frame. It was running routed with an internal and external interface. As a hunch I think the Auto Map SNAT may be giving us problems.

08:44:55.138208 IP 10.2.224.11.802-11-iapp > 10.2.232.35.29418: S 2658976969:2658976969(0) win 512 08:44:55.138407 IP 10.2.232.35.29418 > 10.2.224.11.802-11-iapp: S 1017739285:1017739285(0) ack 2658976970 win 14600 08:44:55.139000 IP 10.2.224.11.802-11-iapp > 10.2.232.35.29418: R 2658976970:2658976970(0) win 0 08:44:56.138089 IP 10.2.224.11.mc3ss > 10.2.232.35.webcache: S 1697615952:1697615952(0) win 512 08:44:56.138279 IP 10.2.232.35.webcache > 10.2.224.11.mc3ss: S 259112963:259112963(0) ack 1697615953 win 14600 08:44:56.139112 IP 10.2.224.11.mc3ss > 10.2.232.35.webcache: R 1697615953:1697615953(0) win 0 08:44:57.137821 IP 10.2.224.11.nssocketport > 10.2.232.35.8085: S 2641368868:2641368868(0) win 512 08:44:57.138019 IP 10.2.232.35.8085 > 10.2.224.11.nssocketport: S 3444298526:3444298526(0) ack 2641368869 win 14600