How to fix a loop in tree-type topology?

[Chingiz]

Hello everyone! We have a subnet, described on the picture, what contents a main Layer 3 Cisco switch, what have a Gateway of Vlan 10 on it, and 5 Layer 2 switches with the same only vlan 10 (its vlan interfaces is vlan 10 too). Each of these 5 switches have a 8 port stupid switches on each port, what use as access layer switches, and has end device joined to it. Everything in one vlan 10. So the topology is tree-type with layer 3 switch as a root, and end devices as "leaves".  So what is the problem - sometimes we accept that kind logs on Layer 3 switch: 

Host 7483.c243.d6e4 in vlan 10 is flapping between port Po4 and port Po1

Host 7483.c243.d6e4 in vlan 10 is flapping between port Po4 and port Po2

Host 7483.c243.d6e4 in vlan 10 is flapping between port Po4 and port Po3

 Or, the other time:

Host 000b.6701.d6e4 in vlan 10 is flapping between port Po5 and port Po1

Host 000b.6701.d6e4 in vlan 10 is flapping between port Po5 and port Po2

Host 000b.6701.d6e4 in vlan 10 is flapping between port Po5 and port Po3

It's always ONE MAC-address in the same time flapping between po4 and other 1-3, or po5 and other 1-3 (never between 4 and 5). Flapping MAC is  mac of some device on  the one of 8 port access switches on 4th or 5th LAYER 2 switches. I can find its port and when I shut it down, MAC flapps only between 1,2 and 3 switches. When I write "show mac add add ..." on each of these 3 switches, 2 of them shows uplink, and one shows some port or also uplink, it always change. I find 3 ports by this way, one of the each switch, and successively shut them down. After that log dissapear. Then I write "no shut" to  real device port on 4th switch, and then to all these 3 ports on 3 switch, that I shuted before. So I return all to first state, when I saw log. But now, after I unlocked these 4 ports, log didnt arrive. But after 1-2 days (no matter) this happens again with other MAC. Seems like there is some loop between 1-3 switches, and MAC of devices from 4 and 5 switches get into this loop and flapps there. But if that so, why does it always only ONE mac in the SAME time, and why it dissapear after I shut-no shut ports?

 And more, I wrote bpduguard enable command on all ports of 1-5 switches, but problem still happens, and no one port have blocked. Can it be something else than loop? Or, if that 100% loop, so how I can fix it? 

   There is a picture of our topology. I noticed suspicious ports by "x". (After blocking these ports log dissapears)

[Flavio Miranda]

Hi

  Can´t see the picture. For loop problem, it is adivesed to review all the connection and fix it accordingly. 

 The mac address you have shown 000b.6701.d6e4 belongs to a camera and 7483.c243.d6e4 belongs to an Access Point  as per mac address lookup. Make sure this device is not doing anything wrong and any unecessary feature enabled.

[Chingiz]

I have just added picture, you can see it. Yes, each of these 8 port access switches is located on its own pillar, and have several  videocontrol cameras on it. There can be UPS device, but access-point isn't. I maybe antenna, but it is no matter, because 9 of 10 flapping MAc is MAc of just 1 camera. When I fix this problem, next time this log arrive with other camera's MAC. Only this time it was an antenna, next time it could be MAC of UPS, or other CAMERA. I wrote these 2 log only to explain that MAC always flapping between 4th and 1-3, and 5 and 1-3 switches.