- Freeradius-Users - lists.freeradius.org

help with xlat
by Matt Zagrabelny 09 Dec '20

09 Dec '20

Greetings FR-users, I've configured my linelog module as such: message { Access-Request = "%S Access Request from calling station %{Calling-Station-Id:-bar}" } but the logging fails: (0) linelog: ERROR: %S Access Request from calling station %{Calling-Station-Id:-bar} (0) linelog: ERROR: ^ Unknown module (0) [linelog] = fail According to the wiki, I feel like the above configuration should work: https://wiki.freeradius.org/config/run_time_variables#conditional-syntax Any ideas what I'm doing wrong? Thanks for the help! -m

2 2

[FreeRadius] Mac Authen with Centos
by yesi 09 Dec '20

09 Dec '20

Hi, I am trying to install FreeRadius to do the Mac Auth. I followed this guide : https://wiki.freeradius.org/guide/mac-auth#mac-auth-or-802-1x <https://wiki.freeradius.org/guide/mac-auth#mac-auth-or-802-1x>. Here are the elements : - server : CentOS Linux release 7.9.2009 (Core) SELinux is disabled. rpm -qa |egrep freeradius freeradius-utils-3.0.13-15.el7.x86_64 freeradius-3.0.13-15.el7.x86_64 - client : a Huawei switch # authentication-profile name ACCESS-MAC mac-access-profile MAC authentication mode multi-authen max-user 100 access-domain toto force # radius-server template TOTO radius-server shared-key cipher tata radius-server authentication 10.x.x.x 1812 vpn-instance management weight 80 # authentication-scheme TOTO authentication-mode radius # domain toto authentication-scheme TOTO accounting-scheme default radius-server TOTO # --- cat /etc/raddb/users bob Auth-Type := Accept, Cleartext-Password := "toto" Reply-Message := "Hello, %{User-Name}" DEFAULT Group == "disabled", Auth-Type := Reject Reply-Message = "Your account has been disabled." DEFAULT Auth-Type := Reject Reply-Message = "\_o< Acces refuse." DEFAULT Framed-Protocol == PPP Framed-Protocol = PPP, Framed-Compression = Van-Jacobson-TCP-IP DEFAULT Hint == "CSLIP" Framed-Protocol = SLIP, Framed-Compression = Van-Jacobson-TCP-IP DEFAULT Hint == "SLIP" Framed-Protocol = SLIP --- cat /etc/raddb/authorized_macs xx-xx-xx-xx-xx-xx Reply-Message = "Device with MAC Address %{Calling-Station-Id} authorized for network access" --- cat /etc/raddb/sites-enabled/default ... authorize { filter_username preprocess rewrite_calling_station_id authorized_macs if (!ok) { reject } else { update control { Auth-Type := Accept } } auth_log chap mschap digest suffix eap { ok = return } files -sql -ldap expiration logintime pap } I did not modified policy.d/canonicalization from the package cat /etc/raddb/mods-available/files files { moddir = ${modconfdir}/${.:instance} filename = ${moddir}/authorize acctusersfile = ${moddir}/accounting preproxy_usersfile = ${moddir}/pre-proxy } files authorized_macs { key = "%{Calling-Station-ID}" usersfile = ${confdir}/authorized_macs # compat = no -------> if not commented, Configuration item "compat" is deprecated } Here is the message from the switch client : Status : Pre-authen There is no more message. From the switch a test with test user "bob" is ok. In the debug mode, "radiusd -X", there is no message when a machine tried to connect to the switch. Any help would be appreciated. y.

3 3

conditional attribute rewrite.
by Benedikt Sveinsson 09 Dec '20

09 Dec '20

I have a particular problem with our service structure We are getting rid of old freeradius servers that have been operational for years and centralizing everything into two radius servers Due to different access networks , Some users are connected to BNG services that can only terminate or forward to a LNS The users (realm) that are forwarded to the LNS first get a response from our prod radius with modified attributes through the legacy attr file were we add the VPDN config, and then proxied to the old legacy Radius. Our Old LNS then receives the session and authenticates with the legacy radius. Client -> BNG - radius to prod Radius – gets VPDN info towards LNS and proxies auth to legacy radius. Client -> LNS - authenticates to the legacy radius This is all and well, but we have new LNS routers that are going to be terminating these sessions. Now the issue is that I need to be able (if possible to only rewrite the attr (insert LNS / VPN info) if the requests comes from the BNG gateways as they can‘t terminate some of the special realms. (only do normal pppoe users) But when the auth requests comes from the LNS – it shoud not get the VPDN config – else it will just loop. Is it possible to conditionally add the attributes I need based on the NAS / NAS-IP or some other identifier ? (basically I‘m using the routers both for LNS and terminate locally) I have searched and I‘m bit lost both between the legacy config and „new“. - Benni

3 3

Re: FR, LDAP (AD) issues
by Matt Zagrabelny 08 Dec '20

08 Dec '20

Ping? -m On Sat, Dec 5, 2020 at 9:58 PM Matt Zagrabelny <mzagrabe(a)d.umn.edu> wrote: > Greetings FR-users! > > I am having trouble getting FR to bind to LDAP (AD). I believe the issue > is with TLS and the CA's, but I am not sure. Any help verifying the problem > and finding a solution would be very appreciated. > > Here's the version: > > $ apt policy freeradius > freeradius: > Installed: 3.0.17+dfsg-1.1 > > According to package dependencies and readelf, FR is using libssl for > crypto library needs: > > $ readelf -d /usr/sbin/freeradius | grep ssl > 0x0000000000000001 (NEEDED) Shared library: [libssl.so.1.1] > > If I use openssl s_client and connect to the AD system without the CA for > the AD system's cert, I get an error: > > $ echo | openssl s_client -connect ad.umn.edu:636 | grep 'Verification: > OK' > depth=0 CN = dc-tc2.ad.umn.edu > verify error:num=20:unable to get local issuer certificate > verify return:1 > depth=0 CN = dc-tc2.ad.umn.edu > verify error:num=21:unable to verify the first certificate > verify return:1 > DONE > > If I add the authority cert of our AD system to the system's list of CA > certs... > > $ sudo cp /path/to/cert /usr/local/share/ca-certificates/ad.umn.edu.crt > $ sudo update-ca-certificates > > ...I do not get the error: > > $ echo | openssl s_client -connect ad.umn.edu:636 | grep 'Verification: > OK' > depth=1 DC = edu, DC = umn, DC = ad, CN = OIT-CA1-ADRCA > verify return:1 > depth=0 CN = dc-tc1.ad.umn.edu > verify return:1 > Verification: OK > DONE > > Also, ldapsearch executes successfully when the authority cert is added to > the ca-certificates for the system, but fails when the cert is not added to > the ca-certificates. ldapsearch is linked against gnutls. > > However, even with the cert added to the system list of ca-certificates, > it appears FR (at least the LDAP component) is failing. Here is a snippet > of "freeradius -X" > > rlm_ldap (ldap): Opening additional connection (0), 1 of 32 pending slots > used > rlm_ldap (ldap): Connecting to ldaps://ad.umn.edu:636 > TLS: can't connect: (unknown error code). > rlm_ldap (ldap): Bind with > CN=d-ins-ldapuser,OU=ServiceAccounts,OU=INS,OU=D,OU=Units,DC=ad,DC=umn,DC=edu > to ldaps://ad.umn.edu:636 failed: Can't contact LDAP server > > rlm_ldap (ldap): Opening connection failed (0) > rlm_ldap (ldap): Removing connection pool > /etc/freeradius/3.0/mods-enabled/ldap[8]: Instantiation failed for module > "ldap" > > The file that the system CA certificates, and hence the AD certificate > authority, reside in is: > > /etc/ssl/certs/ca-certificates.crt > > An strace of the FR shows that it is opening that file: > > $ sudo strace -f /usr/sbin/freeradius 2>&1 | grep '^open' | grep ssl > > openat(AT_FDCWD, "/usr/lib/freeradius/libssl.so.1.1", O_RDONLY|O_CLOEXEC) > = -1 ENOENT (No such file or directory) > openat(AT_FDCWD, "/usr/lib/x86_64-linux-gnu/libssl.so.1.1", > O_RDONLY|O_CLOEXEC) = 3 > openat(AT_FDCWD, "/usr/lib/ssl/openssl.cnf", O_RDONLY) = 3 > openat(AT_FDCWD, "/usr/lib/ssl/openssl.cnf", O_RDONLY) = 3 > openat(AT_FDCWD, "/etc/ssl/certs/ssl-cert-snakeoil.pem", O_RDONLY) = 3 > openat(AT_FDCWD, "/etc/ssl/certs/ca-certificates.crt", O_RDONLY) = 3 > openat(AT_FDCWD, "/etc/ssl/certs/ca-certificates.crt", O_RDONLY) = 3 > openat(AT_FDCWD, "/etc/ssl/private/ssl-cert-snakeoil.key", O_RDONLY) = 3 > > So... part of me thinks that TLS isn't having issues with the CA of the AD > system. > > I've tried increasing the debugging level with extra -x for freeradius, > but the extra logs don't help me determine where the issue is with ldap. > > What do folks think? This is my first foray into LDAP binding. > > Thanks for any help! > > -m >

3 3

Pick up changes in passwd file
by Markus Gaugusch 08 Dec '20

08 Dec '20

Dear freeradius-experts, I'm using version 3.0.21, running in docker (freeradius/freeradius-server:latest-alpine). I want to use freeradius to authenticate users against entries in a file in the form "username:$2a$10$<bcrypt hash>". This works quite well, using the following stanza: passwd my_passwd { filename = /opt/etc/raddb/ext/pwd.txt format = "*User-Name:Crypt-Password:" hash_size = 100 ignore_nislike = no allow_multiple_keys = no } However, when the file gets modified, the changes are not picked up by freeradius. I checked the FAQ [1], which suggests to use kill -HUP <PID>, but that doesn't work either: Do you have any hints for me to get this resolved? It works by restarting the docker container, but I'd rather avoid this if possible. Actually, my users are stored by a 3rd party application in a postgresql DB. This DB is using a different schema than radiusd (I checked [2]), so I decided to generate the file described above. If it would be easy to access the DB directly using custom SQL queries ONLY for usernames/passwords, that would be another option. I only need authentication, no accounting or any logging into the DB. kind regards, Markus [1]: https://wiki.freeradius.org/guide/FAQ#how-do-i_how-do-i-get-radius-to-pick-… [2]: https://wiki.freeradius.org/guide/SQL-HOWTO-for-freeradius-3.x-on-Debian-Ub…

2 1

rlm_ldap: Limit accepted TLS versions on LDAPS
by Robert Hentsch-Jesse 07 Dec '20

07 Dec '20

Hello Freeradius Users, I'm using freeradius with the rlm_ldap module to request users from a OpenLDAP server using the LDAPS protocol. Is there any best practice how to limit the accepted TLS versions to 1.2 and 1.3 on the LDAPS connection? SSL and TLS <= 1.1 should be denied. I found a "tls_min_version" option for the rlm_eap module, but not for rlm_ldap. Are there other possibilities than stripping down the used libssl? Thank you in advance and best regards, Robert Hentsch-Jesse ....................................................................................... PHOENIX CONTACT Cyber Security GmbH Richard-Willstätter-Straße 6, 12489 Berlin, Germany Register Court: AG Charlottenburg, HR B 202908 Geschäftsführer/General Manager: Kilian Golm

2 1

rlm_sql Closing Connections
by Ammad Ali 07 Dec '20

07 Dec '20

Hi Guys, I am having below logs in my radius.log file. The connections are getting closed and some it suddenly got utilized and radius spawns new connections and it takes time to spawn. Also connections are getting closed even the idle timeout isn't reached. Please advise. Below are the logs. Mon Dec 7 12:11:06 2020 : Info: Ready to process requests Mon Dec 7 12:11:31 2020 : Warning: rlm_sql (sql): 22 of 22 connections in use. You probably need to increase "spare" Mon Dec 7 12:11:31 2020 : Error: rlm_sql (sql): Cannot open new connection, connection spawning already in progress Mon Dec 7 12:11:31 2020 : Warning: rlm_sql (sql): 23 of 23 connections in use. You probably need to increase "spare" Mon Dec 7 12:11:31 2020 : Info: rlm_sql (sql): Opening additional connection (23) Mon Dec 7 12:11:31 2020 : Warning: rlm_sql (sql): 23 of 23 connections in use. You probably need to increase "spare" Mon Dec 7 12:11:31 2020 : Error: rlm_sql (sql): Cannot open new connection, connection spawning already in progress Mon Dec 7 12:11:52 2020 : Warning: rlm_sql (sql): 24 of 24 connections in use. You probably need to increase "spare" Mon Dec 7 12:11:52 2020 : Info: rlm_sql (sql): Opening additional connection (24) Mon Dec 7 12:13:31 2020 : Info: rlm_sql (sql): Closing connection (15), from 19 unused connections Mon Dec 7 12:13:46 2020 : Info: rlm_sql (sql): Closing connection (27), from 18 unused connections Mon Dec 7 12:13:53 2020 : Info: rlm_sql (sql): Closing connection (4), from 17 unused connections Mon Dec 7 12:13:56 2020 : Info: rlm_sql (sql): Closing connection (21), from 16 unused connections Mon Dec 7 12:13:57 2020 : Info: rlm_sql (sql): Closing connection (2), from 15 unused connections Mon Dec 7 12:13:58 2020 : Info: rlm_sql (sql): Closing connection (12), from 14 unused connections Mon Dec 7 12:13:59 2020 : Info: rlm_sql (sql): Closing connection (13), from 13 unused connections Mon Dec 7 12:14:00 2020 : Info: rlm_sql (sql): Closing connection (29), from 12 unused connections Mon Dec 7 12:14:01 2020 : Info: rlm_sql (sql): Closing connection (18), from 9 unused connections Mon Dec 7 12:14:02 2020 : Info: rlm_sql (sql): Closing connection (36), from 10 unused connections Mon Dec 7 12:14:03 2020 : Info: rlm_sql (sql): Closing connection (33), from 9 unused connections Mon Dec 7 12:14:04 2020 : Info: rlm_sql (sql): Closing connection (17), from 8 unused connections Mon Dec 7 12:14:05 2020 : Info: rlm_sql (sql): Closing connection (23), from 7 unused connections Mon Dec 7 12:14:06 2020 : Info: rlm_sql (sql): Closing connection (8), from 6 unused connections Mon Dec 7 12:14:07 2020 : Info: rlm_sql (sql): Closing connection (37), from 5 unused connections Mon Dec 7 12:14:08 2020 : Info: rlm_sql (sql): Closing connection (35), from 3 unused connections Mon Dec 7 12:14:09 2020 : Info: rlm_sql (sql): Closing connection (14), from 3 unused connections Mon Dec 7 12:14:10 2020 : Info: rlm_sql (sql): Closing connection (31), from 2 unused connections Mon Dec 7 12:14:11 2020 : Info: rlm_sql (sql): Closing connection (11), from 1 unused connections Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 20 of 20 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Info: rlm_sql (sql): Opening additional connection (40) Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 20 of 20 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Error: rlm_sql (sql): Cannot open new connection, connection spawning already in progress Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 20 of 20 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Error: rlm_sql (sql): Cannot open new connection, connection spawning already in progress Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 21 of 21 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Info: rlm_sql (sql): Opening additional connection (41) Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 21 of 21 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Error: rlm_sql (sql): Cannot open new connection, connection spawning already in progress Mon Dec 7 12:14:20 2020 : Warning: rlm_sql (sql): 22 of 22 connections in use. You probably need to increase "spare" Mon Dec 7 12:14:20 2020 : Info: rlm_sql (sql): Opening additional connection (42) Below are my sql config of radius. sql { driver = "rlm_sql_mysql" dialect = "mysql" server = "localhost" port = 3306 login = "nbbdb" radius_db = "nbbdb" acct_table1 = "radacct" acct_table2 = "radacct" postauth_table = "radpostauth" authcheck_table = "radcheck" groupcheck_table = "radgroupcheck" authreply_table = "radreply" groupreply_table = "radgroupreply" usergroup_table = "radusergroup" delete_stale_sessions = yes pool { min = 20 max = ${thread[pool].max_servers} spare = 20 uses = 0 lifetime = 0 idle_timeout = 900 } client_table = "nas" $INCLUDE ${modconfdir}/${.:name}/main/${dialect}/queries.conf } --

2 3

log proxy auth?
by Matt Zagrabelny 05 Dec '20

05 Dec '20

Greetings, I've looked in both: radiusd.conf proxy.conf for a way to log auth attempts that get routed through a proxy, but I'm not seeing it. I've enabled auth logs via: log { auth = yes } but I'm not seeing my proxied auth attempts in the log. Of course I see details when running radiusd -X: [...] (1) Expecting proxy response no later than 29.666655 seconds from now (1) Received Access-Accept Id 172 from 10.0.0.1:1812 to 192.168.0.1:36687 length 53 [...] (1) Login OK: [bob] (from client foo port 0) (1) Sent Access-Accept Id 222 from 192.168.0.1:1812 to 172.16.0.1:56468 length 0 Is it possible to log proxied auth attempts? Thanks for the help. -m

2 2

Re: Requests being rejected with "Invalid user"
by Dan M 05 Dec '20

05 Dec '20

> Date: Wed, 2 Dec 2020 18:01:15 -0500 > From: Alan DeKok <aland(a)deployingradius.com> > > On Dec 2, 2020, at 5:33 PM, Dan M <dan.red.beard(a)gmail.com> wrote: > > > > We have been running FreeRadius for about 18 months with very light > > traffic and no issues. > > Recently, within the last month, some requests are rejected and a > > message appears in the log Invalid user: [the actual user id] (from > > client <the actual client id> port > > 0) > > That message comes out when the user is rejected. > > > It seems completely random. > > The end user just hits enter again which runs thru the client system > > and generates a new request that succeeds. > > (Well it seems that none of the second requests have failed. I have > > no reports otherwise) The user id seems to me to be identical (at > > least in the log) for both the failure and the success. > > Sure. Sounds like a DB connection issue, TBH. Not using a DB. > > i.e. if FreeRADIUS can't talk to the database, it can't get the user credentials. > So to be safe, it has to reject the user. > > Are there *other* error messages in the logs, around the same time that > user is rejected? Well, that's the thing. There is NOTHING else. I'd have expected something that might identify the reason for the rejection > > > This happened on multiple instances in multiple locations. > > Restarting the instances seemed to clear the issue for a while, but > > it's resurfaced (after about 12 days). > > > > I do not think that the request is getting to the "authorize" phase > > because there is no message from the python plug-in indicating that it > started. > > What "python plugin?" We are using python to make a web service call as part of authorize. Python log has nothing for these requests. So I'm five nines confident it isn't getting that far. Python makes the webservice call which actually does the authentication, and if it succeeds, sets the cleartext password to the password in the request and lets pap (the only thing in authentication section, handle the reply) If it fails, it returns configTuple = (('Auth-Type', "Reject"),)> > i.e. we don't have access to your system. So you shouldn't assume that we > know what you're talking about, when you refer to custom changes you > made. The point was it's getting rejected by something I didn't write and isn't reaching the thing I did. I left all of these in place and haven't changed any of the files. Authorize { filter_username preprocess auth_log suffix expiration logintime python pap } > > > So maybe this comes from some pre-authorize step? > > Maybe. It depends on how you configured it. If the configuration says to > reject the user in the "authorize" stage, then it rejects the user. > > > I scanned all the configuration and don't find that string "Invalid user" > > anywhere, > > so I'm thinking it's generated internally when something else fails. > > It's in the source. > > > Unfortunately, this is in a locked-down, production environment and I > > can't just "start it up in debug mode" > > Is this something you have seen before? > > Is there something I might do to increase the log level? Where, how, > > and of what? > > see raddb/sites-enabled/control-socket, and "radmin". I looked at this and started a different thread about this specifically. > > > We are running version 3.0.17. > > Probably upgrade to 3.0.21. It might not help this issue, but it's not a bad > idea. It's on the long list of maintenance tasks for 2021 > > Alan DeKok. > Lastly, while most requests are generating detail request/reply entries, there are no failures in the detail-reply log. Odd. Dan M Scanned by McAfee and confirmed virus-free. Find out more here: https://bit.ly/2zCJMrO

2 1

Enabling debug level on server
by Dan M 04 Dec '20

04 Dec '20

I need to get debug output without being able to start the server with -X. After reviewing your reply to: Requests being rejected with "Invalid user" and reviewing the related docs I see two options: radmin OR possibly in authorize section Both require config changes and both have challenges in our locked down environment. For using radmin - update control-socket file with socket file name, uid, gid, mode = rw - enable control-socket by creating the symlink - restart Invoke radmin issue useful commands: (possibly using the -f socket_file and -I input_file) show debug level show module list debug file [filename full path] # If I don't set the filename will debug go to the current log file? # e.g. from: log { destination = files file = ${logdir}/radius.log ... # Can I just use the same logfile filename and base logging and debug will go to same file? # Can the debug file name be set permanently in the config somewhere? set debug level 1 # (What is a good/recommended level?) - watch for failures and go read the debug data Alternately, in Vers 3.17 can I just add the below to my authorize section?: update control { Tmp-String-0 = "%{debug:1}" } Does it REQUIRE a conditional? If so would: "if ( "A" == "A") { ... }" suffice? Thanks in advance Dan Mullen Scanned by McAfee and confirmed virus-free. Find out more here: https://bit.ly/2zCJMrO

2 1