We'd love to hear about it! Click here to go to the product suggestion community
If I want to do maintenance on a server I change the weight to "0". After the persistence timeout and even on the next day I still get new connections to this server.
The documentation and logging is very limited on this topic. Is there any further information available on how the Server Load Balancing works? And can we see via CLI how the connections are handled? What software component is used on the UTM for that task?
If you set the weight to "0", it means that the corresponding server only receives data when no other server is available. As long as your other servers are available and Health check is green for them, UTM would not forward the traffic to the "0" weight server. You might see the traffic flow using TCPDUMP in the backend but I don't think you can monitor service or check how connections are handled.
Hallo and welcome to the UTM Community!
If you don't want to temporarily remove the server's Host definition from the 'Real Servers' box in the Load Balancing setup, you can temporarily change the Host definition's IP to something that doesn't exist.
Cheers - Bob
In reply to BAlfson:
it would be a great enhancement if there was a possibility to monitor or log the load balancing actions. Like I said even after 24h I see some clients being redirected to the "0"-weigh server.
I didn't think about that, but perhaps this could be the solution to our problem. As soon as the server isn't reachable all clients should get redirected to other server, right?
In reply to regnor_01:
Unfortunately, there is no Live monitoring tool to check this. You may raise a feature request here.
Bob has suggested a good way to do that. Once you change the IP of the Server Host, it will not be reachable by the UTM and that will force UTM to use other servers from the configuration.