Hi,
Not sure if anyone else is having this issue, we have had it with about 5 routers int he last 4 weeks.
On powerpc, mips and x86 boards
On 3.13,3.17,3.19 ROS versions.
See attached jpg for error from log.
The only thing that fixes it is a replacement of hardware.
The router is fairly simple setup, dhcp server, hotspot, nothing fancy.
We have tried:
1.) Log shows errors as per screengrab attached
2.) Rebooting DOES NOT FIX
3.) Upgrading firmware (3.17 – 3.19 or 3.9 – 317 or 3.13 – 3.17 does not fix)
4.) Downgrading firmware to 3.9 or 3.13 does not fix
5.) Removing and re-adding dhcp server DOES NOT fix
6.) Removing DHCP package (uninstalling) then reinstalling package DOES NOT fix
7.) Link light disappears from affected port (not 100% confirmed on all routers)
8.) Restoring a backup from affected router to NEW HARDWARE fixes problem
I have submitted s supout to mikrotik 4 weeks ago with no real response, just submitted another one now. Does anyone have similar problem?
Most of those mac addresses belong to Senao Int’l. 00:02:6F:xx:xx Do you have any Senao radios on your system?
Maybe a DoS or hack attack? It could be a hacker/spammer trying to find a mac that will work without logging in. Looks like it is going through the whole series of mac addresses.
all those mac’s are valid aps on our system (yes I know there are heaps of them)
At that site about 40 radios.
But all clients are being rejected, not just senao cards or our mikrotik routers, just couldnt capture them all in one screengrab at that site.
That site has been functional for about 9 months now with no issues until this started happnening, and htis is the latest of about 5 different sites getting exactly same issue.
Just checking. It seemed like a lot of radios on one connection is why I asked. I will presume then that those requests are client dhcp requests from your APs then? They will all show up as the mac of the AP at the dhcp server. How many leases do you have issued?
/ip dhcp-server lease print
OK: UPDATE
Thanks to some help from mikrotik, it would appear at least in this instance that the ether port is faulty.
The fact that clients were still trying to get leases and ROS boards were appearing in neighbors is weird to me, however it does seem to be the case.
Still cant believe I have had 4 or 5 other boards do exactly the same thing in the last 4 weeks though, but a swapout did always fix it.
Well AFAIK this message appiers when server (software) sends out necessary data, but client still asks the same thing again - something is happening to packet in between.
It can be Ethernet port of device, it can be wire related problems (wire is not shielded and run near power line for example) , it can be any device (such as switch) in between.
It is easy to eliminate all these possibilities, by simple swapping.
From my experience it is usually related to wires - in some mystical way packets with proper size get corrupted and Ethernet drops them with checksum error.