I'm facing PXE-boot issues within a school network. I'm posting here to make sure I'm not missing something on switches configurations. The issue is that PXE-boot is performing systematically with success on switches directly connected to the core switch stack. On cascaded switches, ie switches connected to another switch but not to the core switch stack, pxe-boot is randomly performed: sometimes it does, sometimes it doesn't. If I temporarily connect a cascaded switch to the core switch stack, pxe-boot from this switch becomes systematically successfull (and randomly performed on the previously attached-to-the-core-switch-stack one).
Both client computers and servers are on the same default VLAN #1. I have a DHCP server together with a dedicated WDS server that delivers ISO images.
I have made a synthetic view of the situation giving switches' interfaces configuration (see network-schema). I can add additional informations if necessary.
Thanks in advance for your help and technical tips on that subject.
Thomas
--
EDIT: additional informations asked for in the comments.
When failing, the client computer remains stuck on Start PXE over ipv4 message. The DHCP server shows no sign of the computer. So it seems like no DHCP address is allocated to the client.
Switch ports facing client computers are all configured as follow:
interface <if>
tagged vlan 12
untagged vlan 1
spanning-tree admin-edge-port
spanning-tree loop-guard bpdu-protection
loop-protect
exit
