RSTP forwarding & flapping Problem

Two S3528G are connected to an UA device through one Ethernet interface (with STP “edge-port enable”) as trunk for a few vlans, G/1/1 is connected to A-Corp Metro Ethernet device and Ethernet from 0/23 to 0/24 are used for link-aggregation between S3528G switches (see attached image), all as trunk for same service vlans. RSTP is use in this scenario and, because of MAC address value of A-Corp devices is lower than Huawei’s, RSTP root selection is always an A-Corp switch. Metro Ethernet network is a layer two backbone for many services and the S3528G switches are use for access to NGN in many access nodes. In some sites the RSTP forwarding and discarding states flap between G1/1 and link-aggregation on both S3528G switches, making service be affected.Alarm Information
#Jul15 11:26:30:088 2011 S3 MSTP/2/PDISC:1.3.6.1.4.1.2011.2.23.1.14.0.2: Instance 0's Port 14.0 has been set to discarding state!
#Jul  15 11:26:30:090 2011 S3 MSTP/2/PFWD:1.3.6.1.4.1.2011.2.23.1.14.0.1: Instance 0's Port 23.0 has been set to forwarding state!
%Jul  15 11:26:30:099 2011 S3 MSTP/3/PDISC:Instance 0's GigabitEthernet1/1 has been set to discarding state!
%Jul  15 11:26:30:109 2011 S3 MSTP/3/PFWD:Instance 0's Ethernet0/23 has been set to forwarding state!
#Jul  15 11:26:31:184 2011 S3 MSTP/2/PDISC:1.3.6.1.4.1.2011.2.23.1.14.0.2: Instance 0's Port 23.0 has been set to discarding state!
#Jul  15 11:26:31:190 2011 S3 MSTP/2/PFWD:1.3.6.1.4.1.2011.2.23.1.14.0.1: Instance 0's Port 14.0 has been set to forwarding state!
%Jul  15 11:26:31:201 2011 S3 MSTP/3/PDISC:Instance 0's Ethernet0/23 has been set to discarding