Just discovered a strange problem with the DHCP server running at RouterOS 6.38.x preventing another Mikrotik Accesspoint and wireless clients from geting an IP address and rendering them inoperable.
Problem:
Wireless clients on an access point (RB433 - no matter if 6.37.x or 6.38) do not get any DHCP address from the router (RB1200@ROS 6.38 or higher) and even not a dhcp client on the AP gets an address. However, the windows clients connected to the router over the same external switch get IPs assigned.
Router continuously reports: “dhcp1 offering lease 192.168.2.252 for […] without success” for the DHCP client on the AP.
Wireless clients can register to the AP but don’t get any IP.
Configuration:
Router RB1200 ---- ---- Netgear Switch GS108 -------- Accesspoint RB433.
DHCP server running on router (without problems and configuration changes for years now).
Accesspoint:
bridge1 configured with member port wlan1 and eth1,
dhcp client interface set to bridge1
No special filters. Also reset the AP configuration (System/Reset configuration) in case something was misconfigured - same results
Above configuration worked for some years now.
Tests and conclusion so far:
Downgrading to 6.37 and step-by-step upgrading to 6.37.1, 6.37.2, 6.37.3 and at least 6.38 i discovered that until 6.37.3 everything works fine.
I also tried different versions on the access point but the version here doesn’t seem to matter. ROS-6.38.1 on Accesspoint and ROS-6.37.3 on the router is the highest version pair that works with my configuration.
Any clues?
Unfortunately no clues, but I encounter the same issue.
I found that disabling/enabling the dhcp-package or restarting the router can solve it (temporarily)
Just tested bugfix release 6.37.4 - also working fine without DHCP problems. @david123: restarting the router did not work for me. DHCP for Accesspoint and its wireless clients (as described in my OP) doesn’t work from the start with 6.38.
I have to agree with this problem. Since version 6.38 the DHCP Service stops responding and no new IP addresses are issued/renewed. To resolve the problem, one has to disable and re-enable the DHCP service.
Both 6.38 and 6.38.1 are affected with this problem.
Before Restart: (Mitel Phone)
19:19:22 dhcp,warning DHCP offering lease 172.29.40.60 for 08:00:0F:8F:C0:0E without success
After Restart: (Mitel Phone)
19:19:22 dhcp,info DHCP assigned 172.29.40.60 to 08:00:0F:8F:C0:0E
I have a similar problem, which I was about to post before I came across this thread. Heres the description nonetheless…maybe some of my findings are helpful:
Problem receiving DHCP addresses over a VLAN using latest RouterOS v6.38.1. First of all, here is my setup
Network 172.20.20.0/24 is my primary network on an untagged VLAN, coupled with an SSID named “unknown”.
Network 172.20.22.0/24 is my guest network tagged with VLAN 2, coupled with an SSID named “guest”.
A DHCP Server running on the CRS should serve both VLANs with IP addresses, but there is a problem with DHCP für VLAN2 (guest):
WiFi Clients connected to the CRS SSID “unknown” receive IPs from 172.20.20.0/24
WiFi Clients connected to the CRS SSID “guest” receive IPs from 172.20.22.0/24
WiFi Clients connected to the RB951 SSID “unknown” receive IPs from 172.20.20.0/24
WiFi Clients connected to the RB951 “guest” SSID do NOT receive IPs from the CRS at all
What I have found is the following:
dhcp logs on the CRS show that the DHCP request is received and an IP from the correct pool is offered to the client
packet sniffer on the CRS shows that the DHCP reply/offer is sent out on the correct VLAN to the correct MAC
however: packet sniffer on RB951 shows DHCP request leaving the router, but no DHCP reply/offer is being received
The problem occurs both ways, e.g. when I set up DHCP Server for VLAN 2 on the RB951 and connect to the guest SSID on the CRS
at some point in the past it did work with this configuration. I did not notice it stopping to work.
The connection between the two Mikrotik devices is over devolo powerline. But I have confirmed the problem when connecting the two devices directly
I have not configured any L2 security/firewalling e.g. snooping or similar
When I connect a client to the RB951 on the guest SSID and give the client a static IP in the 172.20.22.0/24 range, the client can connect properly.
So basic VLAN tagging seems to be working i.e. its a DHCP problem.
I have rebooted the devices several times, but will try manually disable/enable the dhcp service.
As a workaround I have configured a split scope for VLAN2 on the CRS and RB951. I am happy to provide command outputs, but need the direct commands to enter, since I am not that familiar with the CLI. Since I am not home very often, it might take some time for me to reply.
@kielerjung: At some point I had the issue that my Windows 10 client did not receive an IP, while my Android client did. But I cant say on which version that was. Right now on 6.38.1 none of the clients get an IP.
@AM1: the DHCP Authoritative setting ‘after 2s delay’ was already set in my case, so no luck
same problem here. Since upgrading to RouterOS 6.38 the dhcp service runs unstable. It suddenly stops working and i see “dhcp offering lease without success” errors in the logs.
Version 6.38 had an issue related to STP which was resolved in 6.38.1. Also 6.38 changelog included note which said, that all devices should be upgraded to latest version to implement proper STP functionality in network. Same goes for 6.38.1 We have not seen any actual issues in 6.38.1 version related to software. Usually some devices in network are not upgraded and that is the cause of DHCP problems.
I had this also regarding the “offering DHCP address” but not receiving or accepting it.
Turned our that one of the people, had a very old printer, and bought a print server. But that print server had by default a DHCP server on board to set it up. But the client did not know this, and could not setup the server. And left it for the weekend.
And I had to figure out why the DHCP did not work correctly.
So check if you have a second (unknown) DHCP server running. Just disconnect all devices (physically) from the 1200, and start to connect one by one. And test it thoroughly before connecting the second port.
I can confirm that both my devices are running v6.38.1, and that there is no other DHCP Server running in that VLAN. Like I stated, the problem always occurs with clients connected to the miktorik that bridges the DHCP request to the DHCP Server, and not the mikrotik device which actually is the DHCP Server.
Maybe it was the conflict STP openwrt a mikrotik ??
RB951G-2HnD ver. 6.38.1 DHCP SERVER
3x port and WiFi AP port access VLAN 20
1x port WAN port access internet connect
1x port via RB260GS port trunk VLAN 10, VLAN 20, VLAN 30
DHCP is working - ethernet port and WiFi
RB260GS ver.1.17
1x port via RB951G-2HnD port trunk VLAN 10, VLAN 20, VLAN 30
1x port access VLAN 30
1x port via RB960PGS port trunk VLAN 10,VLAN 20
DHCP not tested
RB960PGS ver. 6.38.1
1x port via RB260GS port trunk VLAN 10, VLAN 20
1x port access VLAN 10
2x port access VLAN 20
1x port via WR1043ND port access VLAN 20
DHCP not tested
WR1043ND-openwrt by LUCI
1x port via RB960PGS port and WiFi AP port access VLAN 20
DHCP is not working - WiFi, Ethernet port not tested
Confirm having issues with DHCP server in 6.38/6.38.1.
Device is CRS125-24G-1S-2HnD.
Problems affected at list one DHCP client.
It is a Virtual Machine with Windows Server 2008 R2.
VM is not able to get address assigned to it in the list of static DHCP leases.
Mikrotik log states:
“dhcp offering lease for without success”.
This VM is running inside physical Server (also with Windows Server 2008 R2).
Physical server gets address from the same DHCP static list just fine.
