host1 is running boinc version BOINC client version 7.2.42 for x86_64-pc-linux-gnu (IP address 10.x.x.9)

host1 had an empty remote_hosts.cfg

host2 is running an old boinc client version 6.10.17 (IP address 10.x.x.2/24)

Attempting to use boincmgr on host2 to manage host1 caused repeated errors to appear in the event log on host1

Sun 05 Jul 2015 19:15:28 BST |  | GUI RPC request from non-allowed address 2.0.137.55
Sun 05 Jul 2015 19:15:28 BST |  | 786 connections rejected in last 10 minutes
Sun 05 Jul 2015 19:25:29 BST |  | GUI RPC request from non-allowed address 2.0.140.82
Sun 05 Jul 2015 19:25:29 BST |  | 787 connections rejected in last 10 minutes
Sun 05 Jul 2015 19:35:29 BST |  | GUI RPC request from non-allowed address 2.0.143.110
Sun 05 Jul 2015 19:35:29 BST |  | 787 connections rejected in last 10 minutes
Sun 05 Jul 2015 19:45:30 BST |  | GUI RPC request from non-allowed address 2.0.146.131
Sun 05 Jul 2015 19:45:30 BST |  | 787 connections rejected in last 10 minutes
Sun 05 Jul 2015 19:55:30 BST |  | GUI RPC request from non-allowed address 2.0.149.157
Sun 05 Jul 2015 19:55:30 BST |  | 786 connections rejected in last 10 minutes

I ran a network trace on host1 which showed no activity from these IP addresses, but repeated connections from host2.

Updating remote_hosts.cfg on host1 (with host2 details) and reloading config files, stopped the errors.

There are two issues.

a) The error handler on host1 should log a correct IP address.
b) boincmgr on host2 should not keep trying hundreds of times.

I have not tried out the reverse.

That one needs re-confirming with a host1 running - ideally - alpha BOINC v7.6.3, or at the very least recommended v7.4.42

I doubt it'll make any difference, but it has to be tested.

There's nothing you can do about the retry rate from obsolete v6.10.17 - except note that all Managers of all versions have always updated every second. More than 600 retries in 600 seconds probably implies that the Manager requires multiple RPCs to update all pages.

I'm getting the same issue on my Debian based clients (VMs)

My computers are on an internal subnet 192.168.x.x

Why are ip addresses listed as from france being shown instead of the real ip address? I use BoincTasks to monitor my clients and before I added the IP of the computer runing BT, I would get these errors