FREE RADIUS
http://freeradius.org/
http://wiki.freeradius.org/guide/faq
This is the FAQ (Frequently Asked Questions) for the FreeRADIUS Server (freeradius for short) development project. It contains both general and technical information about the FreeRADIUS projects' status, what it is and what it does, how to obtain and configure and run it, and more. Please read this FAQ carefully before you post questions about FreeRADIUS to the mailing lists to see if your question is already answered here first.
The FreeRADIUS Server is a daemon for unix and unix like operating systems which allows one to set up a radius protocol server, which can be used for Authentication and Accounting various types of network access. To use the server, you also need a correctly setup client which will talk to it, including terminal servers, Ethernet Switches, Wireless Access Points or a PC with appropriate software which emulates it (PortSlave, radiusclient etc).
FreeRADIUS is being developed by a group of people who call themselves "the FreeRADIUS project".
First of all, FreeRADIUS is an open-source product, and has all the benefits open-source provides.
Also, it has many features not found in free servers and many commercial ones, like:
FreeRADIUS runs on a large number of 32 and 64bit Unix-like platforms.
Yes, the FreeRADIUS Server site is at http://www.freeradius.org/
It contains the server, documentation, and additional RADIUS programs.
Yes, there are mailing lists for announcements, users and developers, they can all be subscribed to freely, and they all have web archives.
Please see the description of FreeRADIUS mailing lists.
You can find it, along with some useful links and documentation, at the official FreeRADIUS WWW site:
Sources are available at:
Please see the list of available packages.
You may also easily build a package from source.
See the instructions on how to build FreeRADIUS.
Yes - there are several ways to accomplish this.
_-i {IP}
command-line option.listen
directive in radiusd.conf. listen { ipaddr = 192.168.1.250 port = 1817 type = auth }You may specify multiple listen
directives.
Note!
If you have a multi-homed server and specify an explicit address to bind to. It may happen that the server does not respond with the right source ip address. Thus the client will likely ignore the server's response. On Linux and FreeBSD you can solve this by specifying ./configure --with-udpfromto
during compilation.
Then the server will always respond with the correct address.
Yes, you can. Assuming you already have daemontools installed, configured and running in your system (see http://cr.yp.to/daemontools.html), you will have to make two decisions: 1. The log account and group name (_log.log_ is used in this example). Logging programs run under this account.group. If thisaccount.group pair does not exist yet, create it now.
Then perform these steps:
groupadd log
useradd -g log log
mkdir /etc/radiusd
mkdir /etc/radiusd/log
mkdir /etc/radiusd/log/main
chmod +t+s /etc/radiusd /etc/radiusd/log
chown log.log /etc/radiusd/log/main
The supervise program starts radiusd by running a shell script called "run" from /etc/radiusd. Here are the contents of /etc/radiusd/run:
shell# cd /etc/radiusd
shell# cat run
#!/bin/sh
exec 2>&1
exec /usr/sbin/radiusd -fyz -lstderr
It is important to add -f and -l stderr to argument list of radiusd or svc and logging functions will not work properly.
The logging feature is also started by a "run" script. This one is located in /etc/radiusd/log. Here are the contents of /etc/radiusd/log/run
shell# cd /etc/radiusd/log
shell# cat run
#!/bin/sh
exec setuidgid log multilog t ./main
To set the service to start, issue the following command:
ln -sf /etc/radiusd /service
Now you can send signals to radiusd using the svc
program. Here are some interesting ones:
To hang-up (HUP) it, reloading the config, do:
svc -h /service/radiusd
To temporarly disable it (down) do:
svc -d /service/radiusd
To reenable it (up) do:
svc -u /service/radius
The shared secret is incorrect. This is a text string which is a "secret" (in the raddb/clients file) shared by both the NAS and the server. It is used to authenticate and to encrypt/decrypt packets.
Run the server in debugging mode:
radiusd -X
The first password you see will be in a RADIUS attribute:
Password = "dsa2\2223jdfjs"'
The second password will be in a log message, e.g.:
Login failed [user/password] ...
If the text AFTER the slash is garbage then the shared secret is wrong. Delete it on BOTH the NAS and the raddb/clients file and re-enter it. Do NOT check to see if they are the same, as there may be hidden spaces or other characters.
Another cause of garbage passwords being logged is the secret being too long. Certain NAS boxes have limitations on the length of the secret and don't complain about it. FreeRADIUS is limited to 31 characters for the shared secret.
There are two possible causes of this problem.
Some NAS do not send "Gigawords" attributes by default. Read your NAS documentation and configure it to send the attributes Acct-Input-Gigawords and Acct-Output-Gigawords.
For Cisco IOS, this usually achieved by entering
aaa accounting gigawords
(which, by "ingenious" design, requires a reload of the device on certain IOS versions).
Older versions of FreeRADIUS (prior to 1.1.7) include support for logging 64-Bit counters to both the detail file and SQL modules but only the PostgreSQL module had this support configured by default.
The detail files will simply log two distinct Attributes (Acct-Input-Octets + Acct-Input-Gigawords and Acct-Output-Octets + Acct-Output-Gigawords).
The PostgreSQL module stores the data as a 64-bit integer (BIGINT) in one column each: AcctInputOctets and AcctOutputOctets.
FreeRADIUS 1.1.7 and greater supports 64-bit counters in other SQL modules, with the same semantics as PostgreSQL.
The following procedure is recommended to enable proper support for 64-bit counters in FreeRADIUS 1.1.6 and earlier:
Firstly, modify the radacct table schema to be able to store 64bit integers (or 19 digit numeric fields on databases not supporting BIGINT) in the AcctInputOctets and AcctOutputOctets columns using theALTER TABLE
command:
ALTER TABLE radacct CHANGE AcctInputOctets AcctInputOctets BIGINT(20);
ALTER TABLE radacct CHANGE AcctOutputOctets AcctOutputOctets BIGINT(20);
ALTER TABLE radacct MODIFY (AcctInputOctets NUMERIC(19));
ALTER TABLE radacct MODIFY (AcctOutputOctets NUMERIC(19));
Secondly, modify the accounting queries in sql.conf to make the SQL database perform the computation that is required to merge the two values sent as attributes by the NAS into one single 64-bit integer stored in the database.
All occurences of '%{Acct-Input-Octets}'
need to be replaced with:
'%{Acct-Input-Gigawords:-0}' << 32 | '%{Acct-Input-Octets:-0}'
The same thing needs to be done for '%{Acct-Output-Octets}'
:
'%{Acct-Output-Gigawords:-0}' << 32 | '%{Acct-Output-Octets:-0}'
Symptom: you are seeing lots of duplicate requests in radius.log, yet users can not login, and/or you are seeing duplicated accounting messages (up to 50 times the same accounting record as if the NAS doesn't realize you received the packet).
Perhaps your server has multiple IP addresses, perhaps even multiple network cards. If a request comes in on IP address a.b.c.d but the server replies with as source IP address w.x.y.z most NAS won't accept the answer.
The simplest solution is to have radiusd bind to a specific address. It will only listen to that address and replies will always go out with that address as the source address. See #Is there a way to bind FreeRADIUS to a specific IP address?
The above solution is not always possible. If you have multiple IPs and want FreeRADIUS to listen on all of them. Make sure that ./configure --with-udpfromto
was specified during compilation (#Is there a way to bind FreeRADIUS to a specific IP address?). On Linux and FreeBSD this will make FreeRADIUS to respond from the IP the initial request was received to.
Symptom: FreeRADIUS is not including VSA attributes in Access-Reject packets. (but it worked in earlier versions of FreeRADIUS)
According RFC 2865 (section 5.44) Vendor-Specific Attributes aren't allow in Access-Reject packets.
This behaviour was fixed in newer versions of FreeRADIUS
The RADIUS server receives information about user sessions from your NAS but does not actually actively maintain the users sessions. This means that you have to signal your NAS to disconnect sessions.
Depending on your NAS type and it's configuration there are a number of ways to disconnect sessions.
Packet of Disconnect is the standard and recommended method to disconnect users. It is supported by many newer NAS and uses a RADIUS packet (usually sent to port 3799 although some NAS use 1700 (Mikrotik for example)) to signal that a NAS should terminate an active session.
Many NAS support SNMP which can usually (among other things) be used to disconnect users, however implimentation details vary. Read your NAS documentation to find out whether it supports this, and whichMIB to use.
radkill is a TCL program written by Jason Straight for FreeRADIUS users that monitors ISP users' online times and disconnects them if they are over their call limit. It also monitors the number of users online and will disconnect the users with the least time left to always keep lines open. It's very configurable for multiple NAS setups.
The source archive should be available for download at: ftp://ftp.nmo.net/pub/radkill/radkill-latest.tar.gz
You're not using plaintext passwords in the raddb/users file Or other password store).
The CHAP protocol requires a plaintext password on the radius server side, for PAP it doesn't matter.
So, if you're using CHAP, for each user entry you must use:
Auth-Type = Local, Password = "stealme"
If you're using only PAP, you can get away with:
Auth-Type = System
or anything else that tickles your fancy.
Not really.
Q: So doesn't using CHAP really mean I must leave a file lying around with cleartext passwords for the hundreds/thousands of people who'll be using this thing?
A: Yes.
So what do ISP with (tens of?) thousands of customers do?
You have 2 choices:
You allow CHAP and store all the passwords plaintext
You don't allow CHAP, just PAP
Now, people say CHAP is more secure. Now you decide which is more likely:
Right. Still think CHAP is more secure? I thought so.
This is a limitation of the CHAP protocol itself, not the RADIUS protocol. The CHAP protocol requires that you store the passwords in plain-text format.
Commas link lists of attributes together. The general format for a raddb/users file entry is:
name Check-Item = Value, ..., Check-Item = Value
Reply-Item = Value,
.
.
.
Reply-Item = Value
Where the dots means repetition of attributes.
Check-items are used to match attributes in a request packet or to set server parameters. Reply-items are used to set attributes which are to go in the reply packet. So things like Simultaneous-Use go on the first line of a raddb/users file entry and Framed-IP-Address goes on any following line.
All the attribute operators :=,==,+=
and their meanings are listed in man 5 users on machines that have FreeRADIUS installed and at the Operators page in the wiki.
I'm using a 3Com/USR HiPerArc and I keep getting this message on radius.log:
Mon Jul 26 15:18:54 1999: Error: Accounting: logout: entry for NAS tc-if5 port 1 has wrong ID
What should I do to get rid of these messages?
You are using HiPer ARC 4.1.11, right? Version 4.1.11 has a problem reporting NAS-port numbers to Radius. Upgrade the firmware from http://totalservice.usr.com to at least 4.1.59. If you are in Europe you can telephone to 3Com Global Response Center (phone number: 800 879489), and tell them that you have bought it in the last 90 days. They will help you, step by step, to do the upgrade.
Here is a check list:
radiusd -X
and see if it parses the Simultaneous-Use line.checkrad
manually; maybe you may have a wrong version of perl, don't have cmu-snmp installed etc.radwho
. If it says no one is logged in, Simultaneous-Use won't work.The radius server calls the checkrad script when it thinks the user is already logged on on one or more other ports/terminal servers to verify that the user is indeed still online on that other port/server. If Simultaneous-Use > 1, then it might be that checkrad is called several times to verify each existing session.
This method successfully prevents a user from logging in multiple times across multiple NAS boxes.
by Robert Dalton [mailto:support-at-accesswest-dot-com support at accesswest dot com]
Verify if you are using HiPerArc software version V4.2.32 release date 09/09/99
In order for simultaneous logins to be prevented reported port density must be set to 256 using the command :
set pbus reported_port_density 256
Otherwise it changes the calculations of the SNMP object ID's.
There is a bug in effected versions of checkrad namely the line under the subroutine "sub_usrhiper". The line that should be commented out is:
($login) = /^.*\"([^"]+)".*$/;
Q: I am getting the following in radius.log file:
Thu Oct 21 10:59:01 1999: Error: Check-TS: timeout waiting for checkrad
What's wrong?
A: Verify if you have SNMP enabled on your CISCO router, check the existence of the following line:
snmp-server community public RO 97
where 97 is the access-list that specifies who gets access to the SNMP info. You should also have a line like this:
access-list 97 permit A.B.C.D
where A.B.C.D is the ip address of the host running the radius server.
Q: I am getting the following in radius.log file:
Thu Oct 21 10:59:01 1999: Error: Check-TS: timeout waiting for checkrad
What's wrong?
A: Verify that you have the MAX 4048 setup in your naslist as max40xx and that you have Finger turned on.
Ethernet->Mod Config->Finger=Yes
Ascend decided to have the 4000 series NAS boxes retrieve much of their configuration from the RADIUS server. To disable this "feature", set:
Ethernet->Mod Config->Auth->Allow Auth Config Rqsts = No
This is generally caused by an incorrect named configuration. Check your named files and look for invalid entries.
Another file to investigate is raddb/naslist. All entries there must be resolved by a DNS query.
See question 4.9
The most common problem with PEAP is that the client sends a series of Access-Request messages, the server sends an series of Access-Challenge responses, and then... nothing happens. After a little wait, it all starts again.
If you see this happening STOP!
The RAIDUS server certificate has to have special OID's in it, or else the Microsoft clients will silently fail. See the "scripts/xpextensions" file in the server "tar" file for examples, and the following page on Microsoft's site:
http://support.microsoft.com/kb/814394/en-us
If the clients are running Windows XP SP2, see also:
http://support.microsoft.com/kb/885453/en-us
You MUST follow the instructions on the first page, and install the hot fix from the second page for PEAP or EAP-TLS to work with a Windows machine.
Stop right there. Before going any further, be sure that you have included the following items in your request for help:
Too many people post questions saying "something's wrong, how do I fix it?" with NO background information. This is worse than useless, it's annoying.
Now that you have prepared all the information, post your question to the freeradius-users mailing list
If you're REALLY interested in knowing how to debug the RADIUS server yourself, then the following steps will help you:
screen -S radiusd
)screen -r
to see what is going on in real time.)tcpdump udp
radiusd -X
After upgrading to FreeRADIUS, you may discover that some users are unable to fully use the network, but it worked fine with the previous RADIUS server you were using.
The NAS has no idea which RADIUS server you use, and it doesn't care. The entire problem is that the responses to the NAS from the servers are different. Since FreeRADIUS only sends the attributes in a response that you tell it to send, the conclusion is that your local configuration of FreeRADIUS is incomplete.
Use 'tcpdump' http://www.tcpdump.org to snoop the RADIUS responses from each server. Once you discover which attributes are missing from the response of FreeRADIUS, you can add them to it's configuration. Re-start the server, and your users should have full access to the network again.
You may see an error message like the one below, when you try to run the server:
Module: Loaded SQL
rlm_sql: Could not link driver rlm_sql_mysql: file not found
rlm_sql: Make sure it (and all its dependent libraries!) are in the search path of your system's ld.
radiusd.conf[50]: sql: Module instantiation failed.
There are only a few things that can be happening:
1) The 'mysql_config' cannot be found in $PATH. Try running "./configure | grep mysql_config" and see if you get the following:
checking for mysql_config... no
configure: WARNING: mysql libraries not found. Use --with-mysql-lib-dir=<path>.
configure: WARNING: sql submodule 'mysql' disabled
If it is, simply do something like "export $PATH=$PATH:/usr/local/mysql/bin" and "./configure | grep mysql_config" again to make sure you get:
checking for mysql_config... yes
Then you should run "make;make install" again.
2) You put shared libraries into a place where your linker cannot find them.
Everyone blames FreeRADIUS because it's the one printing the error message. But it just gets the error message from your linker.
3) You don't have static libraries for SQL clients on your system. So doing./configure --disable-shared;make
doesn't help.
The 'make' process WILL print out error messages saying it's creating a static library which links to a dynamic one. If your linker is misconfigured (see #1), then FreeRADIUS still won't work.
So you probably ignored the error/warning messages produced during the 'make' stage. That's bad.
And libtool still does dynamic linking when told to do static linking, instead of failing to do the build. It should be taken out and shot.
There is nothing you can do to FreeRADIUS to fix issues with non-FreeRADIUS shared libraries. Fix your SQL libraries so that:
a) your linker can find them
or
b) There are static versions of those libraries available.
You can use the LD_LIBRARY_PATH environment variable in a script which starts the server, to set the paths where these libraries may be found.
One some systems, you can edit /etc/ld.so.conf, ('man ld.so', or 'man ldconfig'), and add the directory containing the dynamic libraries to that list.
See also the 'libdir' configuration directive in the 'radiusd.conf' file which is distributed with the server. It contains additional information.
If none of these solutions work, then your ONLY option is to build FreeRADIUS without dynamic libraries. This may be done via:
./configure --disable-shared
make
make install
Please READ the messages produced during the 'make' and 'make install' stages. While there is a lot of text to wade through, these messages may be the ONLY source of information as to what's wrong your system.
If you have configured FreeRADIUS to use threads, then nothing is wrong. Newer Linux kernels / procps utilities report one thread by default. You can execute ps
in the following ways to see all threads:
For older versions of ps
/ kernel (2.4), use:
ps -efm
ps auxm
For newer versions of ps / kernel (2.6+), you may prefer to use:
ps -efL
ps auxH
On Windows, the short answer is that you don't.
RADIUS defines a Reply-Message attribute, which you can often use to send text messages in a RADIUS reply packet. PPP has provisions for passing text messages back to the user.
Unfortunately, Microsoft decided to ignore that part of the PPP protocol. So you CAN send messages to Windows PPP users. But Windows will throw the message away, and never show it to the user.
If you don't like this behaviour, call Microsoft and complain.
On the Mac side, the only dialer that shows up the server's message is FreePPP athttp://www.rockstar.com.
You need to use the Group check item to match a group. You also need to use the Auth-Type := Reject check item to deny them access. A short message explaining why they were rejected wouldn't hurt, so a Reply-Message reply attribute would be nice. This rule needs to match for all users, so it should be a DEFAULT entry. You want to apply it instead of any other authentication type, so it should be listed BEFORE any other entry which contains an Auth-Type. It doesn't need a Fall-Through, because you're not giving the user any permission to do anything, you're just rejecting them.
The following entry denies access to one specific user. Note that it MUST be put before ANY other entry with an Auth-Type attribute.
foo Auth-Type := Reject
Reply-Msg = "foo is not allowed to dial-in"
The following entry denies access to a group of users. The same restrictions as above on location in the raddb/users file also apply:
DEFAULT Group == "disabled", Auth-Type := Reject
Reply-Message = "Your account has been disabled"
You may run a SQL query each time a user has an access denied. First you need to write your SQL statement in the directive '''postauth_query''' of the module rlm_sql. For example:
postauth_query = "INSERT into radpostauth (user, pass, date) values ('%{User-Name}', '%{User-Password:-Chap-Password}', '%S')"
Then add the sql module to the '''post-auth''' section of radiusd.conf. Since we want to run the SQL query only on failed login, we need to use the sub-section [http://freeradius.org/radiusd/doc/Post-Auth-Type Post-Auth-Type REJECT]. For example:
post-auth {
# Login successful: get an address from the IP pool.
ippool
Post-Auth-Type REJECT {
# Login failed: log to SQL database.
sql
}
}
Note: This option is usable if you want to detect fraud or similar activities from your users. Keep in mind that this table can became a very large in case you disable to much user accounts, in case of ddos attack, etc. Every rejected attempt will be logged.
Limit logons between 08:00am and 08:00pm for Unix group "daysonly"
DEFAULT Group == "daysonly", Login-Time := "0800-2000"
or
DEFAULT Group == "daysonly", Login-Time := "Any0800-2000"
Limit logons between 08:00am and 08:00pm, from Monday to Friday for Unix group "weekdays"
DEFAULT Group == "weekdays", Login-Time := "Wk0800-2000"
Limit logons between 08:00am and 08:00pm, in Saturday and Sunday for Unix group "weekends"
DEFAULT Group == "weekends", Login-Time := "Sa-Su0800-2000"
You can't. A RADIUS server will only log the messages which a NAS sends to it. If your NAS is not sending those messages or attributes, then the RADIUS server will not log them.
You must configure your NAS to send the information you want to the RADIUS server. Once the NAS is sending the information, the server can then log it.
Edit raddb/users, on top of it put
DEFAULT Auth-Type := Accept
That will accept everybody. If you want this to apply to a single user replace DEFAULT with username. You can also add Auth-Type Accept to radcheck or radgroupcheck entries in order to accept that user/group. This only works for PAP, and does NOT work for EAP-TLS, CHAP, MSCHAP, or WIMAX authentication.
Q: I need to limit some users to be able only to use our POP3 and SMTP server.
The most common approach is to just assign non-globally-routable IP addresses to those users, such as RFC1918 addresses. Depending on your internal network configuration, you may need to set up internal routes for those addresses, and if you don't want them to do anything besides SMTP and POP3 within your network, you'll have to set up ACLs on your dialup interfaces allowing only ports 25 and 110 through.
Make sure you have RADIUS authorization enabled on your NAS.
Example user entry in raddb/users file:
foo Auth-Type := System
Framed-Filter-Id += "160.in"
Framed-Filter-Id += "161.out"
Fall-Through = Yes
CISCO's config must have:
aaa authorization network default radius
ip access-list extended 160
permit ip ...
ip access-list extended 161
permit ip ...
The access list 160 gets applied on inbound packets and 161 on outbound packets.
You'll need the redhat/radiusd.pam file from the distribution. It should go into a new file, /etc/pam.d/radiusd.
If you have 100's to 1000's of users in /etc/passwd, you'll want to replace the pam_pwdb.so entries with pam_unix_auth.so, pam_unix_acct.so etc. The pam_pwdb module is INCREDIBLY SLOW for authenticating users from a large /etc/passwd file.
Bruno Lopes F. Cabral also says:
Now I can emulate group behaviour using just PAM and some tricks, like
auth required /lib/security/pam_userdb.so crypt db=/etc/raddb/data/users
auth required /lib/security/pam_listfile.so item=user sense=allow file=/etc/raddb/data/somehunt.allow onerr=fail
auth required /lib/security/pam_nologin.so
account required /lib/security/pam_userdb.so
and
DEFAULT Huntgroup-Name ="somehunt", Auth-Type=PAM, Pam-Auth="radhunt", Simultaneous-Use=1
Fall-Through = Yes
this way I have NO users on /etc/password and NO need for lots of lines on /etc/raddb/users. time to search for a db enabled pam_listfile module
The server reads the config files just once, at startup. This is very efficient, but you need to tell the server somehow to re-read its config files after you made a change. This can be done by sending the server a SIGHUP (signal '1' on almost if not all UNIX systems). The server writes its PID in/var/run/radiusd.pid
, so a simple UNIX command to do this would be:
kill -1 `cat /var/run/radiusd.pid`
Some people would be tempted to do this every 5 minutes so that changes come through automatically. That is not a good idea as it might take some time to re-read the config files and the server may drop a few authentication requests at that time. A better idea is to use a so-called "timestamp file" and only send a SIGHUP if the raddb/users file changed since the last time. For example a script like this, to be run every 5 minutes:
#! /bin/sh
cd /etc/raddb
if [ ! -e .last-reload ] || [ "`find users -nt .last-reload`" ]; then
if radiusd -C > .last-reload 2>&1; then
kill -1 `cat /var/run/radiusd.pid`
else
mail -s "radius reload failed!" root < .last-reload
fi
fi
touch .last-reload
Of course a Makefile is suited perfectly for this kind of stuff.
Some administrators have automated scripts to update the radius servers configuration files. The server can then be signalled via a HUP signal to re-read certain configuration files (see below).
The problem with this approach is that any syntax errors in the configuration file may cause your main radius server to die. Therefore there should be some process of checking the configuration files prior to re-starting the server.
With FreeRADIUS 2.0.0 and up you can use the -C option for radiusd to check certain parts of your configuration files. See the radiusd(8) manpage for further details and limitations of the -C option. Related posts on freeradius-users: 067362 067279
shell# freeradius -XC; echo $?
FreeRADIUS Version 2.0.0-beta, for host i486-pc-linux-gnu, built on Nov 12 2007 at 17:25:45
[...]
/etc/freeradius/users[5]: Syntax error: Previous line is missing a trailing comma for entry DEFAULT
Errors reading /etc/freeradius/users
/etc/freeradius/radiusd.conf[1033]: Instantiation failed for module "files"
[...]
Errors initializing modules
1
shell# freeradius -XC; echo $?
FreeRADIUS Version 2.0.0-beta, for host i486-pc-linux-gnu, built on Nov 12 2007 at 17:25:45
[...]
Configuration appears OK.
0
Note however, that this option is not available in freeradius 1.x. The freeradius distribution contains a check-radiusd-config script which checks the configuration by starting a second server on a different port and waiting for it to crash or not to crash...
Use the following configuration :
Framed-Route := "10.130.1.252/32 0.0.0.0 5",
Framed-Route += "10.130.0.252/32 0.0.0.0 10",</pre>
Which gives : (tcpdump output)
Framed Route Attribute (22), length: 28, Value: 10.130.1.252/32 0.0.0.0 5
0x0000: 3130 2e31 3330 2e31 2e32 3532 2f33 3220
0x0010: 302e 302e 302e 3020 2035
Framed Route Attribute (22), length: 28, Value: 10.130.0.252/32 0.0.0.0 10
0x0000: 3130 2e31 3330 2e30 2e32 3532 2f33 3220
0x0010: 302e 302e 302e 3020 3130
The whole netmask business is a complicated one. An IP interface has an IP address and usually a netmask associated with it. Netmasks on point-to-point interfaces like a PPP link are generally not used.
If you set the Framed-IP-Netmask attribute in a radius profile, you are setting the netmask of the interface on the side of the NAS. The Framed-IP-Netmask attribute is NOT something you can set to influence the netmask on the side of the dialin user. And usually, that makes no sense anyway even if you could set it.
The result of this on most NAS is that they start to route a subnet (the subnet that contains the assigned IP address and that is as big as the netmask indicates) to that PPP interface and thus to the user. If that is exactly what you want, then that's fine, but if you do not intend to route a whole subnet to the user, then by all means do NOT use the Framed-IP-Netmask attribute.
Many NAS interpret a left-out Framed-IP-Netmask as if it were set to 255.255.255.255, but to be certain you should set the Framed-IP-Netmask to 255.255.255.255.
For example, the following entries do almost the same on most NAS:
user Cleartext-Password := "blegh"
Service-Type = Framed-User,
Framed-Protocol = PPP,
Framed-IP-Address = 192.168.5.78,
Framed-IP-Netmask = 255.255.255.240
user Cleartext-Password := "blegh"
Service-Type = Framed-User,
Framed-Protocol = PPP,
Framed-IP-Address = 192.168.5.78,
Framed-Route = "192.168.5.64/28 0.0.0.0 1"
The result is that the end user gets IP address 192.168.5.78 and that the whole network with IP addresses 192.168.5.64 - 195.64.5.79 is routed over the PPP link to the user (see the RADIUS RFCs for the exact syntax of the Framed-Route attribute).
The ldap module can only work with PAP passwords since it needs to send the clear text user password to the LDAP server to authenticate the user. There are however provisions to extract the user password from the LDAP and make it available to the server core and the chap module. See rlm_ldap for more details on how to configure the ldap module to do that.
There are a few things that the administrator should watch out for though:
If you have control over NAS, then set it to accept only PAP authentication. If you do that, all clients will "listen" and use only PAP. In that case there is no need to configure anything on the client(s).
Acct-{Input|Output}-Gigawords?
Please see: Why do Acct-Input-Octets and Acct-Output-Octets wrap at 4 GB?
FreeRADIUS 1.1.x does not particularly care if the host it runs on is dual-stack. It will work just fine, but only use the IPv4 stack of the machine. It will also transport IPv6 RADIUS attributes just fine but will NOT send packets over IPv6.
FreeRADIUS Server 2.0.0 and greater has full support for both IPv6 attributes and IPv6 network packets.
RADIUS related RFCs and Drafts are archived at:
FreeRADIUS is the result of the work done by a large number of people.
The major contributors are listed on the Acknowledgments page.