Link in LAG group failed, could still access switch but DHCP fails. Why?

So I have a switch LAGed (2 links) to the main switch. The switch provides POE to a bunch of Access Points that are on trunk ports with a native vlan. The other day, when all of those APs went offline it was pretty clear that the issue was with that single switch or at least it's line of communication back to our core switches. I could SSH into the switch and could ping around internally an externally, as well as saw in 'show cdp neighbors detail' that the APs weren't getting IP addresses. At some point I saw that one of the two links in the LAG group was dark, but the assumption that the LAG group was meant to continuing operating if a link failed led me to look else where. After some amount of dead end troubleshooting, i decided to just fix the dead link in the LAG group and voila, all the Access Points received DHCP and all was well. Can someone explain this to me?