We are running a mikrotik hotspot in a location with many users. They are set to authenticate to a freeradius server, which has been up and running fine for some time now. Recently, we have had two specific problems with users not being able to log in on this specific hotspot router (we have others running fine elsewhere on our network).
The first problem is the most troublesome. At some point users get kicked out of the active hotspot users list, but for some reason the router keeps the simple queues active. When the users who were kicked off try to log in again we then see the following error:
internal error: logion failed: failed to add queue: already have such name
and are not allowed to log in. We haven't been able to figure out what is triggering this, and when it happens our only solution right now is to remove all queues and active users manually to allow people to log in normally.
The other problem is users getting this error often when logging in:
login failed: IP xxx.xxx.xxx.xxx is already logged in
I looked into it and it seems like this is happening because users are being assigned different IP addresses while they are still logged in. Then when they try to get on with the different IP it tells them they are already logged in. What we haven't figured out yet is why they are getting different leases in the middle of a hotspot session. We are running a normal DHCP server on this router and have no address pools set in the hotspot itself. Right now our solution for this is to manually log people out as they call in about it, but this happens so often it's becoming a real burden.
These problems just started occurring recently and have been more and more frequent. We have tried upgrading all the way to 2.9.49 but no upgrade has fixed the problem yet. Right now as a last ditch effort we rolled it back to a very old version, 2.9.28, which was a version we ran it on without problems some time ago. So far we haven't seen it again, but we also just downgraded this morning so it could still happen.
Has anyone else experienced these problems, and if so did you figure out any solution and or workarounds? We could really use some help with this. Also if you could use any other specifics about our configuration I would be happy to post them. Thank you!