Hi, For a few weeks some devices have had their IP allocation "lost" en route. They authenticate, and IP is allocated but not received by the device. Afterwards a device continuously asking for an IP address, being allocated, and then asking again. This has happened intermittently with some devices on Vodafone APN. I did not see devices on other APNs affected. Please see the three Radius requests in debug mode at the end of this email in attachments. They are small files of approx 10Kb and were collected with options to radiusd -xxx /etc/raddb/users ( with comments and empty lines stripped but willing to resend if empty lines and # comments are needed ) # grep -v ^# users | grep -v ^$ DEFAULT NAS-IP-Address == 172.200.100.1, Pool-Name := vodafone DEFAULT NAS-IP-Address == 10.19.255.254, Pool-Name := proximus-carli DEFAULT NAS-IP-Address == 10.17.255.254, Pool-Name := proximus-mechelen DEFAULT Framed-IP-Netmask = 255.255.255.255 O/S and packages Centos 6.2 freeradius-mysql-2.1.12-1.el6.x86_64 freeradius-utils-2.1.12-1.el6.x86_64 freeradius-2.1.12-1.el6.x86_64 Example from logs: Thu Oct 29 00:00:15 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 594301 cli 31655418538) Thu Oct 29 00:00:15 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 594301 user 000006156) Thu Oct 29 00:01:12 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 146587 cli 31655418538) Thu Oct 29 00:01:12 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 146587 user 000006156) Thu Oct 29 00:02:39 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 983132 cli 31655418538) Thu Oct 29 00:02:39 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 983132 user 000006156) Thu Oct 29 00:05:01 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 928184 cli 31655418538) Thu Oct 29 00:05:01 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 928184 user 000006156) Thu Oct 29 00:07:33 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 575003 cli 31655418538) Thu Oct 29 00:07:33 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 575003 user 000006156) Thu Oct 29 00:08:27 2015 : Auth: Login OK: [000006156] (from client vodafone-nas port 728168 cli 31655418538) Thu Oct 29 00:08:27 2015 : Info: Allocated IP: 10.26.0.61 from vodafone (did vodafone.exampleapn.com cli 31655418538 port 728168 user 000006156) This happens continuously amounting to 1000s of IP allocations daily. I do not see Info: Released IP entries for this device. ( there were some earlier this month. ) I have tcpdump running with : tcpdump -s0 -ni eth0 port 1812 or port 1813 or port 3799 -vw /tmp/rad.pcap Some questions, - Could my ippool run out of addresses because the device is continuously requesting IP addresses? - Would you offer any improvements on the tcpdump parameters? - How could I check if Radius really sent the packet back to the client? - (wireshark/tcpdump question: Is there an expression I could use to find the radius packet sent back for Info: Allocated IP? e.g I could radius.User_Name == "00000123" in wireshark to find User-Name we used for authentication. Is there any useful info in the attached debugs indicating problems? Like why the device did not receive an IP. If not then I can start looking into networks/vpn and vodafone. Kind regards.