I have a custom system that allows to manage whole zoo of stabdalone APs (up to 600 per building). Everything, including SAP1601 works fine except the ten recent SAP1602I[1].
SAP1602 BVI 1 can't get it's ip address from dhcp server with a message "DHCP: QScan: Timed out Selecting state%Unknown DHCP problem.. No allocation possible" [2]
Further debugging shows me that the AP accepts DHCP_OFFER message that comes to local broadcast (255.255.255.255) but not the one that comes to subnet broadcast (like 10.1.1.255). It's for sure IOS behavior change, not a firewall or network config issue.
Is it expected behaviour or IOS bug?
Is this change in behaviour documented somewhere?
Is there a way to restore old behaviour on AP side?
[1]
Cisco IOS Software, C1600 Software (AP1G2-K9W7-M), Version 15.2(2)JB2, RELEASE SOFTWARE (fc1)
Technical Support: http://www.cisco.com/techsupport
Copyright (c) 1986-2013 by Cisco Systems, Inc.
Compiled Mon 29-Jul-13 12:16 by prod_rel_team
[2]
*Mar 1 00:03:27.861: Temp IP addr: 0.0.0.0 for peer on Interface: BVI1
*Mar 1 00:03:27.861: Temp sub net mask: 0.0.0.0
*Mar 1 00:03:27.861: DHCP Lease server: 0.0.0.0, state: 3 Selecting
*Mar 1 00:03:27.861: DHCP transaction id: 163
*Mar 1 00:03:27.861: Lease: 0 secs, Renewal: 0 secs, Rebind: 0 secs
*Mar 1 00:03:27.861: Next timer fires after: 00:00:04
*Mar 1 00:03:27.861: Retry count: 3 Client-ID: 18e7.28a8.42fc
*Mar 1 00:03:27.861: Client-ID hex dump: 18E728A842FC
*Mar 1 00:03:27.861: Hostname: ap1600-018
*Mar 1 00:03:27.861: DHCP: SDiscover placed class-id option: 436973636F204150206331363030
*Mar 1 00:03:27.861: DHCP: SDiscover: sending 294 byte length DHCP packet
*Mar 1 00:03:27.861: DHCP: SDiscover 294 bytes
*Mar 1 00:03:27.861: B'cast on BVI1 interface from 0.0.0.0
*Mar 1 00:03:31.861: DHCP: QScan: Timed out Selecting state%Unknown DHCP problem.. No allocation possible
*Mar 1 00:03:40.957: DHCP: Waiting for 30 seconds on interface BVI1
