MSA1000 Disks changing count

[UNIX Admin Group_1]

We have 2 RP3440 (Server A and B) with one dual port fibre card (Ports 1,2). Each server is connected to a MSA1000 with 1 dual port controller. Identical setup and patches on both servers. We have 11 Luns configured on each MSA. Here is the configuration: Server A FC Port 1 goes to local MSA port 1. FC Port 2 goes to Server B MSA port 2. Server B FC port 1 goes to Server A MSA port 2 and FC Port 2 goes to local MSA port 1. This gives us 4 paths for our one cluster and similar disks designations for recovery. All hard addresses are set. Last week, I saw 22 luns on Server A and 113 luns on Server B. I deleted the connections on Server B and re-entered them. Server B now sees 22 luns and Server A now sees 113 luns. As if they failed over. I have a similar set up on 2 other servers and see 22 luns on each server.