Duplicate ARP entries after IP change

I am using RB1100AH2 with OS 6.10 and CRS125-24G OS 6.9

I notice on my two routers that when a device reboots or changes ip due to DHCP lease expiration the ARP list shows the MAC assigned to two different IPS (old and new).

The ARP list is not automatically updated. The device is not removing the old ARP entry automatically. This is causing lots of problems for me.

I have ARP set to enabled on all the interfaces.

Dave

Hello

You should be run the following command

/ip settings print

and check arp-timeout.
And if necessary run:

/ip settings set arp-timeout=30s

On DHCP server, if there are flagged, remove the [not default on] flag “Add ARP for Leases”.

“Add ARP for leases” is not checked. Should I check?

Timeout is already set to 30 sec.

[admin@AISLECOMMKTOWN] > /ip setting print
ip-forward: yes
send-redirects: yes
accept-source-route: no
accept-redirects: no
secure-redirects: yes
rp-filter: no
tcp-syncookies: no
arp-timeout: 30s
icmp-rate-limit: 10
icmp-rate-mask: 0x1818
allow-fast-path: yes


Dave

I am starting to believe this is a bug on 6.x. I have encountered this a few times now and previously thought it was a ASUS router firmware causing the problem. This time I just downgraded to 5.26 on the affected segment and the problem went away. Can you downgrade to 5.26 and see if that helps? We’ve deployed 6.7 to about 20 routers as of a few months ago and the problem is occurring on a regular basis now.

Was this ever resolved?
I’m seeing this on 6.3
BTW, downgrading really isn’t an option.

saludos
Bernardo

I am seeing it with 5.26 as well actually. I am starting to think its a mikrotik bug in the dhcp server but I can’t confirm. I am trying to get better pcaps so I can submit them.

we do not use DHCP at all
Here it’s the IP-alias, we use for CPE management, sitting in the ARP tables forever.
Looks like it is ignoring arp-timeout or something.

saludos
Bernardo

Same problem here, rb1100ahx2, seems like ignoring the ARP timeout. To make things even worse this bug leads the bridge to flood all the ports with traffic destined for the expired mac addreses.

I am running 6.10 and will upgrade to 6.13 - but I don’t see in the change logs something related to this BUG.