cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
63195
Views
28
Helpful
11
Replies

“5411 Supplicant stopped responding to ISE". Use EAP-TLS for authentication method and the root CA is window CA

TerryLaw86466
Level 1
Level 1

The background is the end devices PC would like to use EAP-TLS for authentication method and the root CA is window CA.

However the ISE live log show “5411  Supplicant stopped responding to ISE”.

 

Below as the ISE live log.

ise1.pngise2.pngise3.png

 

 

Switch config:

aaa group server radius gp-ISE
server name ISE
!
aaa group server tacacs+ group-ts-ISE
server name ts-ISE
!
aaa authentication login default group group-ts-ISE local
aaa authentication enable default group group-ts-ISE enable
aaa authentication dot1x default group gp-ISE
aaa authorization exec default group group-ts-ISE local
aaa authorization commands 0 default group group-ts-ISE local
aaa authorization commands 1 default group group-ts-ISE local
aaa authorization network default group gp-ISE
aaa accounting auth-proxy default start-stop group gp-ISE
aaa accounting dot1x default start-stop group gp-ISE
aaa accounting exec default start-stop group group-ts-ISE
aaa accounting commands 0 default stop-only group group-ts-ISE
aaa accounting commands 1 default stop-only group group-ts-ISE
aaa accounting commands 15 default start-stop group group-ts-ISE
!


aaa server radius dynamic-author
client 192.168.100.240 server-key 7 xxxxxxxxxxxxxxxxxxx

server-key 7 xxxxxxxxxxxxxxxxx
!
aaa session-id common

!


dot1x system-auth-control
dot1x critical eapol
errdisable recovery cause bpduguard
errdisable recovery cause loopback
errdisable recovery interval 180
license boot level ipservicesk9
diagnostic bootup level minimal
!

interface GigabitEthernet1/0/1
switchport access vlan 100
switchport mode access
ip arp inspection limit rate 100
authentication event fail action next-method
authentication open
authentication order dot1x
authentication priority dot1x
authentication port-control auto
dot1x pae authenticator
spanning-tree portfast
!

ip radius source-interface Vlan100
ip sla enable reaction-alerts
logging history size 50
logging history debugging
logging origin-id ip
logging facility local2
logging source-interface Vlan100
logging host 192.168.100.240

!


!
radius-server attribute 6 on-for-login-auth
radius-server attribute 8 include-in-access-req
radius-server attribute 25 access-request include
!
radius server ISE
address ipv4 192.168.100.240 auth-port 1645 acct-port 1646
key 7 xxxxxxxxxxxxxxxx
!
!

Any hits for this issues?

 

3 Accepted Solutions

Accepted Solutions

Greg Gibbs
Cisco Employee
Cisco Employee

See a similar community post here

This is typically an issue with the client supplicant and/or certificate trust chain.

I've also seen this issue caused by fragmentation due to an MTU mismatch somewhere between the switch and PSN. If your switch MTU is set to jumbo, try setting it to 1500 and testing again. See the document EAP Fragmentation Implementations and Behavior for more info.

You might also want to do a packet capture on the client to see the full EAP conversation and open a TAC case if all else fails.

View solution in original post

Arne Bier
VIP
VIP

Basic checks

It appears that your supplicant is performing Machine Authentication - is it configured for certificate authentication (or user authentication), and does your machine have a machine certificate?

 

if the above is all good (i.e. you r supplicant is configured for Machine Auth and you have a machine cert) then:

In your ISE Wired 802.1X Policy Set Overall Condition, did you include "Allow Protocols EAP-TLS"? 

If your Windows Clients are using TLS 1.0 or TLS 1.1, did you allow this under Admin > System > Settings Security Settings?

 

 

View solution in original post

pnowikow
Level 1
Level 1

ISE teammates,

 

Ok, so after months of working on this issue with "Supplicant Stopped Responding..." errors, I've determined that by updating the Wired AutoConfig GPO we use to include these settings, we got the endpoints connected and authorized as expected.  Out of the box Windows 10 tries once to connect to a Cisco switch via PEAP before giving up and waiting 20 minutes.  Here's the GPO settings we're using.  We opted for public CA certs because I was tired of fooling around with our internal CA.

 

Hope it helps someone,

 

 

GPO_Fix_ISE_802.1x.png

 

 

 

 

View solution in original post

11 Replies 11

Greg Gibbs
Cisco Employee
Cisco Employee

See a similar community post here

This is typically an issue with the client supplicant and/or certificate trust chain.

I've also seen this issue caused by fragmentation due to an MTU mismatch somewhere between the switch and PSN. If your switch MTU is set to jumbo, try setting it to 1500 and testing again. See the document EAP Fragmentation Implementations and Behavior for more info.

You might also want to do a packet capture on the client to see the full EAP conversation and open a TAC case if all else fails.

Arne Bier
VIP
VIP

Basic checks

It appears that your supplicant is performing Machine Authentication - is it configured for certificate authentication (or user authentication), and does your machine have a machine certificate?

 

if the above is all good (i.e. you r supplicant is configured for Machine Auth and you have a machine cert) then:

In your ISE Wired 802.1X Policy Set Overall Condition, did you include "Allow Protocols EAP-TLS"? 

If your Windows Clients are using TLS 1.0 or TLS 1.1, did you allow this under Admin > System > Settings Security Settings?

 

 

"Allow Protocols EAP-TLS" already enabled in  Policy Set in ISE.

And the client cert. and root cert. already imported to client PC.

"certificate authentication (or user authentication)"already configure at network adapter of client PC.

 

Added partial config. of the switch relayed to authenticate configuration. is it possible it caused by switch configuration issues? 

 

Switch config:

aaa group server radius gp-ISE
server name ISE
!
aaa group server tacacs+ group-ts-ISE
server name ts-ISE
!
aaa authentication login default group group-ts-ISE local
aaa authentication enable default group group-ts-ISE enable
aaa authentication dot1x default group gp-ISE
aaa authorization exec default group group-ts-ISE local
aaa authorization commands 0 default group group-ts-ISE local
aaa authorization commands 1 default group group-ts-ISE local
aaa authorization network default group gp-ISE
aaa accounting auth-proxy default start-stop group gp-ISE
aaa accounting dot1x default start-stop group gp-ISE
aaa accounting exec default start-stop group group-ts-ISE
aaa accounting commands 0 default stop-only group group-ts-ISE
aaa accounting commands 1 default stop-only group group-ts-ISE
aaa accounting commands 15 default start-stop group group-ts-ISE
!


aaa server radius dynamic-author
client 192.168.100.240 server-key 7 xxxxxxxxxxxxxxxxxxx

server-key 7 xxxxxxxxxxxxxxxxx
!
aaa session-id common

!


dot1x system-auth-control
dot1x critical eapol
errdisable recovery cause bpduguard
errdisable recovery cause loopback
errdisable recovery interval 180
license boot level ipservicesk9
diagnostic bootup level minimal
!

interface GigabitEthernet1/0/1
switchport access vlan 100
switchport mode access
ip arp inspection limit rate 100
authentication event fail action next-method
authentication open
authentication order dot1x
authentication priority dot1x
authentication port-control auto
dot1x pae authenticator
spanning-tree portfast
!

ip radius source-interface Vlan100
ip sla enable reaction-alerts
logging history size 50
logging history debugging
logging origin-id ip
logging facility local2
logging source-interface Vlan100
logging host 192.168.100.240

!


!
radius-server attribute 6 on-for-login-auth
radius-server attribute 8 include-in-access-req
radius-server attribute 25 access-request include
!
radius server ISE
address ipv4 192.168.100.240 auth-port 1645 acct-port 1646
key 7 xxxxxxxxxxxxxxxx
!

 

 

Is it possible that if the clinet MAC address deleted from WLC and user faces this issue "5411 Supplication issue"? or because of DHCP leases overlapping ?

TerryLaw86466
Level 1
Level 1

Hi All,

 

Finally, it is working fine with the cert. signed by window Standalone CA, but it is not working for windows enterprise ca.

Any requirement or specific setting needed in ISE to compatible to windows enterprise ca.

pnowikow
Level 1
Level 1

ISE teammates,

 

Ok, so after months of working on this issue with "Supplicant Stopped Responding..." errors, I've determined that by updating the Wired AutoConfig GPO we use to include these settings, we got the endpoints connected and authorized as expected.  Out of the box Windows 10 tries once to connect to a Cisco switch via PEAP before giving up and waiting 20 minutes.  Here's the GPO settings we're using.  We opted for public CA certs because I was tired of fooling around with our internal CA.

 

Hope it helps someone,

 

 

GPO_Fix_ISE_802.1x.png

 

 

 

 

It indeed helped me. Thank you for sharing. Appreciate it. 

 

Regards,

Omid

I'm glad it worked out for you. I spent a long time beating my head
against the wall

jonkimsr
Level 1
Level 1

Our Cisco 3850 switch was on 16.09.06. After months of fight to configure EAP-TLS, I upgraded the switch to 16.12.12 today. EAP-TLS just worked without a peep of a complaint. It was highly frustrating journey totally unnecessary and a company like Cisco which produce ISE, Switch, WLC should produce complete validated design and remove all wasted time from these monstrosity. For what it does, ISE infrastructure is too complicated. Well 802.1x Ecosystem and its various actors..  Most organizations just want their own device to connect and there are far easier way to accomplish than pile of mess I had to dig through. Validated Design, Validated configuration which should include all Version numbers in involved should be doable in 30 page doc. If not these don't deserve to exist. Now it is done it is a cinch but journey was too costly. I enjoy reading but I can count some 3000 pages to get here and that was ones I skim through and I have not touched others on the subject.. 

This article was written in 2020 when 16.09.06 is published. I am here from future 16.12.12 a 2024 release will likely fix your issue. My 5411 no longer appears in the log. 

I share your sentiment on suffering simple/stupid bugs for a technology that has been around for a very long time. But you mentioned that there are "far easier way(s) to accomplish ..." - what do you propose as the alternative?

Simplified Canned Deployment. 

what would you like to do today? 
I would like a peap on vlan 10 using v10_user/password. Please authorize anyone with that credential and assign them vlan10. 
What more would you like to today? 
I would like a eap-tls with this CA Key and anyone presenting any Key issued from that CA let them connect. 

 My organization will likely use 20 some peaps and 1 Cert(Maybe 2-3 more if I end up doing BYOD type).

Hi, there is a device with 12:34:56:78:90:12 trying to connect what would you like to do? You can authorize, or put them in vLAN bucket or reject.